xvEPA
United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/600/R-00/015
April 2000
Fish Physiology,
Toxicology, and Water
Quality
Proceedings of the Fifth
International Symposium,
Hong Kong,
November 10-13, 1998
-------
-------
EPAJ600/R-00/015
April 2000
FISH PHYSIOLOGY, TOXICOLOGY,
AND WATER QUALITY
Proceedings of the Fifth International
Symposium, Hong Kong,
November 10-13, 1998
Edited By
Robert V. Thurston
Fisheries Bioassay Laboratory
Montana State University
Bozeman, Montana 59717
Published By
Ecosystems Research Division
Athens, Georgia 30605
National Exposure Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Printed on Recycled Paper
-------
NOTICE
The views expressed in these Proceedings are those of the individual authors and do not
necessarily reflect the views and policies of the U.S. Environmental Protection Agency (EPA).
Scientists in EPA's Office of Research and Development have authored or coauthored papers
presented herein; these papers have been reviewed in accordance with EPA's peer and
administrative review policies and approved for presentation and publication. Mention of trade
names or commercial products does not constitute endorsement or recommendation for use by
EPA.
11
-------
FOREWORD
Joint ecological research involving scientists and environmental managers from every
country in the world is essential if global environmental problems are to be solved. Recognition
of this international aspect of environmental protection is reflected in the joint activities
undertaken under Annex 3, Item 4 of the United States of America-People's Republic of China
Protocol for Environmental Protection. This component of the protocol provides for cooperative
research on the environmental processes and effects of pollution on freshwater organisms, soils,
surface water and groundwater, and on the application of transport and transformation models.
Specific areas of cooperation in environmental research include: inorganic chemical
characterization and measurement; inorganic chemical transport and transformation process
characterization; biological degradation process characterization; oxidation/reduction process
characterization; field evaluation of selected transport, exposure and risk models; and application
of models for environmental decision-making concerning organic pollution in semi-arid
conditions, heavy metal pollution, and permissible loading of conventional and toxic pollutants
in rivers. Activities include seminars, workshops, joint symposia, training programs, joint
research, and publications exchange.
This fifth symposium presented under the protocol was held on the campus of the
City University of Hong Kong on November 10-13,1998. Scientists from ten countries
presented papers at the symposium, which was sponsored by the U.S. Environmental Protection
Agency, the American Fisheries Society, the City University of Hong Kong, and Montana State
University. The four earlier symposia were held in Guangzhou, PRC, on September 14-16,
1988; in Sacramento, California USA, on September 18-19,1990; in Nanjing, PRC, on
November 3-5, 1992, and in Bozeman, Montana USA, on September 19-21, 1995.
Symposia are an effective means of fostering cooperation among scientists from different
countries as environmental organizations seek to gain the information necessary to predict the
effects of pollutants on ecosystems and apply the results on a global scale. The symposia
provide a forum through which distinguished scientists from laboratories and institutions from
several countries can exchange scientific knowledge on environmental problems of concern to
the U.S. Environmental Protection Agency and the international environmental community.
Rosemarie C. Russo, Ph.D.
Director
Ecosystems Research Division
Athens, Georgia
in
-------
ABSTRACT
Scientists from ten countries presented papers at the Fifth International Symposium on
Fish Physiology, Toxicology, and Water Quality, which was held on the campus of the City
University of Hong Kong on November 10-13,1998. These Proceedings include 23 papers
presented in sessions convened over 4 days. Papers addressed effects of metals on the
physiology of fishes and aquatic invertebrates, effects of ammonia on fishes, effects on fishes of
toxicants from oil shale and coal gasification effluents, thermal effects on fishes, effects of
pollutants on reproduction of fishes, bioaccumulation and physiological effects on fishes of
xenobiotics, and the use of semi-permeable membrane devices to monitor xenobiotics. Water
quality papers included discussions on hypoxia, metal ecotoxicology, metals on sediments,
methodologies to evaluate the health of riparian and wetland environments, risk management of
metal pollution, remedial strategies to reduce impacts on a watershed of metal mine wastes,
watershed modelling, and strategies for developing nutrient and sediment load allocations for
water quality protection.
IV
-------
CONTENTS
Page
FOREWORD iii
ABSTRACT iv
ACKNOWLEDGMENTS vii
PHYSIOLOGY AND TOXICOLOGY
Thermal effects on the adrenergic cardiac system in teleost fishes.
M.C. Cerra, R, Mazza, and B. Tota 1
Effects of pollutants on the reproduction of fishes.
LinH.R , 17
Identification of ammonia and volatile phenols as primary toxicants in a coal
gasification effluent H. Jin, X. Yang, H. Yu, andD. Yin 31
Fishes: ammonia production, excretion, and toxiciry.
D.J. Randall and B.J. Wicks 41
Tissue ammonia levels and swimming performance of brown trout exposed to copper
in soft, acidic water. M. W. Beaumont, P.J. Butler, and E. W. Taylor 51
The mudskippers: ammonia toxicity and tolerance.
Y.K. Ip, K. W. Peng, S.F. Chew, W.K. Kok, J. Wilson, and D.J. Randall 69
Effects of cadmium on nitrogen and phosphorus turnover in a tropical freshwater
snail, Brotia hainanensis. P.K.S. Lam 87
Metal bioavailability to marine invertebrates: significance of trophic transfer.
W-X. Wang 95
Bioaccumulation and metabolism in fishes of complex mixtures from oil shale
contamination in Estonia. A. Tuvikene, S. Huuskonen, and
P. Lindstrom-Seppd 105
Physiological studies on the toxicity of silver to freshwater fishes: implications for
environmental regulations. C.M. Wood 119
-------
Bioaccumulation of xenobiotic chemicals by aquatic organisms.
D.Connell 137
A physiological model to predict xenobiotic concentrations in fishes.
R. Yang, D.J. Randall, C. Brauner, J. F. Neuman, and R. V. Thurston 145
Monitoring of trace metals in the aquatic environment by artificial mussels.
R.S.S. Wu, T.C. Lau, and W.K.MFung 159
A comparison of mussels and SPMDs as monitors for trace organic pollutants in
Hong Kong waters. B. J. Richardson, G.J. Zheng, and E.S. C. Tse 169
Use of semipermeable membrane devices to measure xenobiotics in Lithuanian
rivers. A. Cetkauskaite, D. Sabaliunas, J. Ellington, I Sabaliuniene,
and A. Sodergren 179
Hypoxia in coastal waters: pressing problems worldwide and their scientific
challenges. R.S.S. Wu. 203
WATER QUALITY
Towards predicting metal ecotoxicology: applying coordination theory, surface
chemistry, and simulation models. G. W. Bailey and Z. Z. Zhang 215
Development of a methodologies to evaluate the health of riparian and wetland areas.
P.L. Hansen, W.H. Thompson, R.C. Ehrhart, D.K. Hinckley, W. Haglan,
and K. Rice 233
Trace metals in surface sediments from the Baja California-California border area
and the Sea of Cortez. J. V. Macias-Zamora 245
Risk management of mercury pollution in China.
Lin Y.H. andChenJ.H. 259
Remedial strategies to reduce impacts from metal mine wastes on a western
United States watershed. D.R. Neuman 267
Watershed modeling - where are we heading?
Y.D.Chen 277
Strategy for developing nutrient and sediment load allocations for water quality
protection. R.C. Russo andR.F. Carousel 291
VI
-------
ACKNOWLEDGMENTS
Organizing and presenting a symposium and preparing the proceedings is frequently a
complex task, particularly when participants represent organizations in several counties. For
this reason the cooperative work of those involved from the sponsoring agencies-the U.S.
Environmental Protection Agency, the American Fisheries Society, the City University of
Hong Kong, and Montana State University-is very much appreciated. The work of members
of the organizing committee, Dr. Rudolf Wu and Ms. Mandy Dung of the City University and
Dr. David J. Randall of the University of British Columbia, is gratefully acknowledged. The
scientists, engineers, and environmental managers who prepared papers and participated in the
symposium are, of course, deserving of primary recognition. In particular, recognition is
accorded to the session chairpersons, who assured efficient functioning of the symposium: Drs.
Wu, Randall, Chris Wood of McMaster University, Canada, Des Connell of Griffith University,
Australia, and George Bailey of the U.S. Environmental Protection Agency. Assistance in
preparing this document was provided by Linda Heydon, Steve Potter, and Norma Hamilton.
Sheila Walker prepared the final draft.
Robert V. Thurston
Chairman of the Symposium
vn
-------
-------
THERMAL EFFECTS ON THE ADRENERG1C CARDIAC SYSTEM
IN TELEOST FISHES
Maria C. Cerra1'2, Rosa Mazza1, and Bruno Tota1'3
ABSTRACT
In fishes, generalized endocrine responses to temperature involve the activation of the
adrenergic system, i.e. non-neural chromaffin cells and adrenergic neurons, causing release of
catecholamines that in turn trigger a number of homeostatic readjustments. Although with large
intra- and interspecific differences in the responsiveness of the target tissues, the cardiovascular
system plays a major role in such readjustments. Direct cardiovascular actions of
catecholamines include an increase in stroke, volume, heart rate, and blood pressure, with
consequent beneficial effects on oxygen transport. Fishes, as do others ectotherms, also
compensate for temperature-induced cellular disturbances through biochemical and functional
remodelling of their cell machinery, including mechanisms of homeoviscous adaptation of the
membranes. At the level of whole-organism, it is difficult to assess either the potential
independent effects of the temperature-dependent cell remodelling and those of the temperature-
induced adrenergic activation, or to which extent these two processes may be mechanistically
linked. In contrast, experiments at organ, tissue, and cell levels can be better used to illustrate
how adrenergic responses affect the plasticity of the fish heart so that it can better face the short
term and long term thermal challenges.
INTRODUCTION
The best documented driving force in the evolution of the vertebrate cardiovascular
function is the need for an efficient transport of respiratory gases between the gas exchanger
and the tissues (Nilsson 1983). The fulfilment of this need becomes even more critical under
stress conditions, including temperature changes, calling for immediate adjustments of the
cardiocirculatory function, the perfusion of the gas exchanger, the blood flow to the tissues,
and the content of oxygen and glucose carried in it. In the classical adaptation stress syndrome
described by Selye (1956), the activation of the adrenergic vasomotor control (circulating
catecholamines and adrenergic neurones) plays a major role in these cardiovascular
readjustments.
1 Department of Cell Biology, University of Calabria, Arcavacata di Rende, Cosenza, Italy.
2 Department of Pharmaco Biology, University of Calabria, Arcavacata di Rende, Cosenza, Italy.
3"A.Dohrn" Zoological Station of Naples, Naples, Italy.
-------
The most relevant cardiovascular actions of both these nervous and hormone effectors
appear generally synergistic and include modulation of cardiac output, i.e. the product of heart
rate (HR) and stroke volume (SV), and increase of systemic blood pressure. As a consequence,
beneficial effects on both gas exchanges and oxygen transport are attained. Despite the
remarkable similarities in the adrenergic cardiac stress response that have been conserved across
the vertebrate lineage, in teleosts this response, extensively studied by fish physiologists, shows
a high degree of complexity and remarkable species-specific differences. In comparison,
relatively little is known in teleosts regarding the involvement of the cardiac adrenergic system
in response to acute and chronic thermal stress conditions. Research in this area can highlight
important ecophysiological questions and at the same time can reveal fundamental aspects of
cardiac plasticity. After mentioning very briefly aspects of piscine catecholamine metabolism
and release that can be particularly affected by thermal challenges, this review will consider
myocardial aspects of (3-adrenergic sensitivity of thermally-acclimated fish and the modulation
of the cascade from adrenoceptors to components of the excitation-contraction coupling or
Ca*"1" channels. It is well known that in some vertebrate groups, adrenocortical steroids are of
paramount importance in the stress response in relation to the adrenergic system. However,
this aspect is not considered here.
CATECHOLAMINE METABOLISM
Biosynthesis
In contrast to mammals, in fishes the catecholamines adrenaline (epinephrine) and
noradrenaline (nor-epinephrine) are synthesized and stored in both chromaffin cells and
adrenergic neurones by identical metabolic pathways (the "Blaschko pathway"), therefore both
catecholamines can act as neurotransmitters. Molecular oxygen is a co-factor in both the
hydroxylation reactions catalyzed by the enzyme tyrosine hydroxylase (TH), which converts
tyrosine to dihydroxyphenylalanine (DOPA), and the enzyme dopamine-p-hydroxylase (DBH),
which converts dopamine to L-noradrenaline. Therefore, limited oxygen availability may also
limit catecholamines biosynthesis. This aspect must be taken into consideration when analysing
catecholamine function under thermal stress. Temperature and pH optima of the enzymes of the
biosynthetic pathway in chromaffin tissue of Atlantic cod (Gadus morhud) and dogfish (Squalus
acanthias) may illustrate putative biochemical targets of thermal stress (see Randall and Perry
1992, Table 1).
Degradation
Monoamine oxidase (MAO) and catechol-O-methyl transferase (COMT) are the major
enzymes that catabolize the catecholamines. The predominantly extraneural COMT is
principally involved in the catabolism of circulating catecholamines, while the mitochondrial
enzyme MAOis more important in neuronal degradation. The gills, receiving 100% of cardiac
output and endowed with an extensive cell-plasma interface well suited for hormone extraction,
appear as the major site of plasma catecholamine degradation (Colletti and Olson 1988).
-------
This task could be especially accomplished by the chloride cells, the most abundant components
of the gill filamental epithelia. Since the oxidative deamination of catecholamines requires
molecular oxygen, the rate of catecholamine metabolism will be affected by the oxygen status of
the blood. In rainbow trout (Oncorhynchusfaykiss), the increased activity of chromaffiri tissue
DBH after administration of cortisol was presumably the result of a reduced enzyme degradation
(Jonsson et al. 1983).
Release
The two potential sources of circulating catecholamines are the release from chromaffin
tissue and the "overflow" from adrenergic nerve terminals. In teleosts, the chromaffin tissue
resides mostly in the wall of the large veins within the anterior part of the kidney (head kidney),
thus the mechanism of catecholamine release in fishes has been best studied in this group.
Several organs contain chromaffin cells. The heart itself is a site of chromaffin tissue of
particular importance in cyclostomes and dipnoans (Nilsson 1983). In both elasmobranch and
teleost fishes the primary mechanism causing the secretion of catecholamines from the
chromaffin cells is their stimulation by the sympathetic system. Acetylcholine released from the
pre-ganglionic sympathetic fibres stimulates the chromaffin cells by interacting with their
muscarinic and nicotinic receptors, the latter likely being predominant. The involvement of
these receptors in the cholinergic control of catecholamine secretion is variable, not only
amongst fish species, but also amongst different stressors (Reid et al. 1998). In this light, it is of
interest to emphasize the higher thermal sensitivity of the muscarinic receptors. Recent evidence
indicates that, as in mammals, the control of chromaffin tissue also involves, in addition to
cholinergic stimuli, humoral signals and peptidergic neural signals. For example, Angiotensin II
(Ang II) has been shown to act as a potent non-cholinergic secretagogue of humoral adrenaline
in the trout so that some of the cardiovascular effects of exogenous Ang II can be attributed to
increased levels of plasma adrenaline (Bernier and Perry 1999).
The ratio of adrenaline/noradrenaline at rest varies among and between species (being
slightly predominant adrenaline in teleosts and noradrenaline in elasmobranchs). Several studies
indicate that in the trout resting levels of catecholamines are in the range of 1-9 nniol/L,
conceivably providing an adrenergic cardiac tonus; for this reason adrenaline has been often
included in the perfusion fluid in experiments with isolated perfused heart preparations (Graham
and Farrell 1989). In response to various types of stress, marked interspecific differences in the
levels of plasma catecholamines have been reported, with the measured increases ranging from
minimal to several orders of magnitude (Thomas and Perry 1992). This variation, which in some
cases is also critically dependent from the experimental protocol used, makes it difficult to
notice any generalized pattern. However, there is strong evidence that during severe stress, in
representatives of all the fish groups studied, the circulating levels of catecholamines rise, with
adrenaline usually being the predominant plasma catecholamine. As remarked by Randall and
Perry (1992), during stress the adrenergic neural overflow seems to represent a minor component
of the circulating catecholamines. However, in regions of densely adrenergically innervated
organs or tissues including the heart, "overflow" of catecholamines may significantly contributes
to increase the local hormone levels. So, in the perfused heart of the European eel (Anguilla
-------
anguilla), a spontaneous "overflow" of neural catecholamines was detected into the perfusion
fluid and was increased by stimulation of the vagus nerve, thereby affecting cardiac performance
(Pennec and Le Bras 1984). Similarly, in the Atlantic cod stimulation of the autonomic nerves
of the head kidney caused a catecholamine release into the veins, which was large enough to
affect cardiac function and branchial vasculature (Nilsson et al. 1976, Wahlqvist 1981).
Noradrenalin was the dominant neurotransmitter released into perfusate following vagal
stimulation of perfused hearts of the goldfish (Camssius auratus) (Cameron and O'Connor
1979). Different chromaffin cell types may respond differently to potential catecholamine-
releasing stimuli (Randall and Perry 1992). Hypoxia is a powerful stressor eliciting
catecholamine release (Gamperl et al. 1998). Recent evidence indicates that the depression
of blood oxygen content, rather than the partial pressure of oxygen in the arterial blood, is the
stimulus triggering the mobilization of catecholamines into the circulation (Reid et al. 1998).
CATECHOLAMINES AND CARDIAC FUNCTION
All teleost hearts studied so far show a cholinergic inhibitory and an adrenergic excitatory
innervation. The relative tonic influence of the two systems varies among species. The
adrenergic nerves enter the sinus venosus through cardiac branches of the vagus nerve and
converge on the sinoatrial plexus and to some extent along the atrial canal and the atrio-
ventricular funnel (Laurent et al 1983). As in other species, in the rainbow trout the heart is
shown to receive a dense innervation of the sinus venosus and the atrium, especially the
sinoatrial and the atrioventricular region, while the innervation of the ventricle is sparse (see for
review Santer 1985, Farrell and Jones 1992). Adrenergic innervation is also associated with the
coronary vasculature, being more dense in the outer compact myocardium of the trout. The
overflow of catecholamines from the adrenergic nerve endings to the circulation is not fully
elucidated (Randall and Perry 1992). The heart is the major target organ for catecholamines
released from the chromaffin cells in the head kidney into the circulation. In addition to
adrenergic innervation and circulating catecholamines, adrenergic stimulation of the fish heart
can also result from cardiac cells containing catecholamines.
Generally, the adrenergic stimulation in teleosts causes positive chronotropy and inotropy,
hence enhancing cardiac output (CO). In contrast to the 100% increase reported in some
mammals, catecholamines increase the intrinsic HR only by 15-20%. This is not surprising,
since different from mammals which increase HR rather than SV, HR in teleosts is a poor
indicator of integrated cardiac performance. Available knowledge indicates that the positive
chronotropism (tachycardia) is mediated by p-adrenoceptors, while negative chronotropism
(bradycardia) is mediated by a-adrenoceptors. As mentioned later, a low reliance on adrenergic
mechanisms in the control of the HR has been found in Antarctic teleosts (Axelsson et al 1998).
As classically documented in cardiac strips of fishes, frogs, and mammals, p-adrenergic
stimulation increases inotropism severalfold. Therefore, catecholamines enhance the intensity
with which the heart responds to both preload and afterload, so that consequent changes in SV,
CO, power output and stroke work are attained. In addition, catecholamines can protect the
myocardium against extracellular acidosis. This situation, which often occurs following
-------
stressful exercise, is harmful to cardiac performance since it decreases the force of contraction.
In some species, such as sea raven (Hemitripterus americanus) contractile deterioration is
prevented by adrenaline in physiological concentrations (Farrell et al. 1986).
Adrenergic responses are mediated by a- and (3- receptors, both of these with several
subtypes differing in tissue distribution and relative importance. The cardiac responses to
adrenergic stimulation depend on the relative density of adrenoceptor subtypes, the density of
the adrenoceptor, and the concentration and type of ligand. Both a- and p-adrenoceptors have
been detected in the teleost hearts, but both receptor types may not occur in a given species
(Farrell and Jones 1992). The teleost adrenoceptors of both the a- and (3-type exhibit
similarities with, but also differences from, those of mammals (Ask 1983, Nilsson 1983). In the
rainbow trout the catecholamines are found to mediate their effects on the myocardium via
p2-adrenoceptors (Ask et al. 1980, 1981). In general, the p-adrenergic regulation of the teleost
heart appears a major mechanism in maintaining normal cardiac performance and in adapting
the myocardium to changing environmental conditions. The |3-adrenergic pathway, following
the binding of the hormone to the membrane receptor, includes activation of a guanosine
triphosphate (GTP) binding protein (G protein), the stimulation of adenylyl cyclase activity,
the production of cyclic adenosine monophosphate (cAMP), and the activation by cAMP of
the cAMP-dependent protein kinase (PKA). The breaking down of cAMP by cAMP
phosphodiesterase (PDE), which limits the degree of cAMP-dependent phosphorylation,
provides a fine tuning of this cascade. Perhaps the most important p-adrenergic mechanism in
regulating myocardial inotropism is the stimulation of the L-type Ca++ current (ICa), whose
amplitude and kinetics are crucial determinants of myocardial contractile activity (Hove-Madsen
etal 1996).
TEMPERATURE AND CARDIAC FUNCTION
Temperature has important effects on both HR and SV, the two determinants of CO.
The relative contribution of FDR and SV to changes in CO in fishes varies among species. For
example, during.acute warming winter flounder (Pseudopleuronectes americanus) increase
resting CO primarily by increasing HR, whereas during thermal acclimation they increase resting
CO by increasing both FIR and SV (Cech et al. 1976). The force of contraction is dependent on
the length of contraction and relaxation periods and on the rate of contraction. Changes in HR
can change the contraction-relaxation cycles, thereby affecting myocardial contractility and,
ultimately SV (an index of contraction force) and pressure generation. The relationships
between HR and SV is illustrated by the staircase effect: at low HR, an increase in HR is usually
associated with an increase in S V (positive staircase), while at higher HR, the relationships
between HR and SV is negative (negative staircase). Since HR changes are associated with
changes in intracellular calcium levels, those species that are highly dependent on intracellular
calcium for contractile activation generally exhibit a particularly evident negative staircase
(Driedzic and Gesser 1994).
-------
Chronotropy
In teleosts the HR is minimally influenced by preload and afterload changes, but is
remarkably affected by temperature (Driedzic and Gesser 1994), as shown by in vitro isolated
cardiac preparations (allowing evaluation of the intrinsic components of temperature-induced
changes in HR) and in vivo studies, including telemetry analysis, on intact fish (allowing
information on the extrinsic neuro-hormonal components). In general, an acute increase in
temperature increases HR. The increased HR results from the direct effect of temperature on the
intrinsic rate of the pacemaker cells (i.e. direct action on membrane permeability) and from the
extrinsic neuro-humoral control, which can be also influenced by thermal changes. Thus, it is
unlikely that there is little compensation of resting HR in fish during an acute thermal stress
(Farrell and Jones 1992). The factor by which rates change for an increased temperature of 10°C
(Qio) for intrinsic HR is 2.0 or greater, in agreement with conformer-type responses to acute
temperature changes. The increased HR is often paralleled by increased cardiac contractile
activity and metabolism, hence matching the predictable higher oxygen demands during
warming. In contrast, compensation of resting HR has been reported under conditions of chronic
thermal changes.
An inverse or non-compensatory acclimation is identified when the cold-acclimated fish shows
lower HR, contractile activity and Chothan warm-acclimated crucian carp (Carassius carassius)
(Matikainen and Vornanen 1992). Opposite to this pattern is the positive compensation in HR
(and heart size) at low temperatures for rainbow trout, sole (Solea vulgaris), European eel, and
goldfish. For references see Axelsson et al. 1998). It is possible that such thermal compensation
can be the result of extrinsic modulation rather than a direct effect on the pacemaker per se. In
this view, changes in the cholinergic tone of the heart can be of particular importance, as shown
by Wood et al. (1979) who identified a higher cholinergic tone in rainbow trout acclimated for
2 weeks to 5°C, compared to rainbow trout acclimated to 20°C.
Inotropy
On the basis of acclimation studies, temperature is considered to exert negative inotropy
on the teleost heart (Matikainen and Vornanen 1992). This can result from the thermally-
induced positive chronotropy, which can reduce the filling time, in combination with the
negative staircase.
Several components of the myocardial cell machinery can be targets of the thermal
modulation of inotropy. Myocardial contraction in fishes is activated mainly by sarcolemmal
calcium entry, which represents the primary source of Ca attaching to the binding sites of
troponin of the myofilaments. Accordingly, the duration of the sarcolemmal action potential
(AP) is considered as a major determinant of the myocardial contraction in most teleosts (Tibbits
et al. 1992). Thermal acclimation can alter calcium activation of contraction at the level of
sarcolemma in the crucian carp. In fact, the ventricular myocardium of cold-acclimated fish
(2°C) showed contraction activated at much lower external calcium concentrations and longer
duration of contraction and relaxation than in their warm-acclimated (22°C) counterparts.
A notable prolongation of contraction duration by low calcium levels was present only in the
cold-acclimated fish. In contrast to homeotherm heart, in fish heart the delivery of calcium
-------
to the myofilaments from the sarcoplasmic reticulum (SR) may be less important, particularly at
low temperatures (for references see Driedzic and Gesser 1994). In rainbow trout ventricle
strips, at 5°C the force was correlated with time-to-peak tension and appeared regulated by a
ryanodine-insensitive mechanism, while at 25°C the force was correlated with the maximal rate
of force development, and the SR contributed significantly to excitation-contraction coupling
(Hove-Madsen 1992). In contrast, a significant increase of SR was detected in the heart
ventricle of cold-acclimated (5°C) perch (Percafluviatilis) in comparison with their warm-
acclimated counterparts (Bowler and Tirri 1990). The present hypothesis suggests that the
function of the SR in excitation-contraction coupling could become progressively more
important with higher temperature and also during increased cardiac demands such as exercise.
This is the case of yellowfin tuna (Thunnus albacares), and perhaps other scombroid fish, which
modulate CO primarily through changes in HR that are high relative to other active teleost
species. In isometric atrial muscle preparation of yellowfin tuna, Shiels et al. (1999) have
demonstrated that SR is active in contributing Ca*"1" to force development at physiological
contraction frequencies, as indicated by reduction of contractile force of about 50% and rates
of relaxation by 60% after ryanodine exposure. High levels of adrenaline were unable to
ameliorate the effects of ryanodine. At low pacing frequencies, the temperature-dependency
of SR-Ca""" utilization in atrium was similar to that observed for trout ventricle, with a greater
ryanodine response at warmer temperature than those tested at 15°C. Instead, within the
physiological range of pacing rates, percentage SR-Ca""" utilization was greater at colder
temperatures.
CARDIAC EFFECTS OF TEMPERATURE AND CATECHOJLAMINES
In view of the combined effects of thermal changes and catecholamines on the teleost
heart, oxygen and homeoviscous adaptation are two aspects of particular concern that reveal
a further layer of complexity.
Oxygen
Since the tissue oxygen status can influence both catecholamine metabolism and release,
the action of thermal stress on the adrenergic response can involve the combined effects of
temperature, oxygen transport (i.e. haemoglobin-oxygen binding), and metabolism. For
example, temperature acclimation can lead to an enhanced haemoglobin-oxygen binding affinity
at lower temperature, changing simultaneously also the threshold at which catecholamines are
released into the circulatory system. Given the interrelationships between these variables it is
difficult to assess, at the organismic level in vivo and .in vitro, the independent effects of
temperature on the activation of the adrenergic system. This can explain in part the disparity
among the various data. These methodological aspects have been reviewed by Randall and Perry
(1992), and Reid et al. (1998).
-------
Homeoviscous Adaptation
Among other environmental/experimental stressors temperature is the major modulator
of membrane architecture. In fact, following a temperature change, cells compensate for stress-
induced cellular disturbances through biochemical and physiological mechanisms of
"homeoviscous" or "homeophasic" adaptation (Hazel 1995) that enable them to attain a new
homeostatic equilibrium. This response is characterized by changes in membrane lipid
composition, including modification in fatty acid unsaturation, changes in the lipid/ protein ratio
and changes in the proportion of lipid classes, that in turn lead to changes in the structural and
functional properties of membrane lipids and membrane-associated transporters and enzymes.
Following such reshaping of the plasma membrane and other organelle membranes of the cells,
temperature can modulate cellular responses and intracellular signaling. This means that the
thermal stress, more than other stressors, not only affects the organism at the different
hierarchical levels of its organization, but also affects the adrenergic system, namely the system
responsible for the stress response itself. For example, thermal acclimation could alter the
structure of the diffusion-restricted microdomains that regulate the activity of the Na+/Ca "H"
exchanger of the sarcolemma or the Ca^-ATPase activity of the SR in acclimated rainbow trout
and crucian carp myocytes, possibly explaining their species-specific different responses during
cold wintering periods (Vornanen 1998a, 1998b). This situation should be kept in mind when
analyzing the thermal effects on the adrenergic cardiac system.
Acute Thermal Changes
It is generally accepted that adrenergic activation exerts specific positive effects on
myocardial performance in teleosts as part of the alert reaction at low body temperatures.
For example, in sockeye salmon (Oncorhynchns nerka), exposed to thermal stress the plasma
adrenaline increased during the first 10 minutes and remained thereafter at steady levels also
during a longer (3-hour) stress period (Mazeaud et al. 1977). Interestingly, in the Atlantic
salmon (Salmo salar) the three-fold increase in the adrenaline content in the atrium following a
3-hour pre-slaughter stress period in vivo was paralleled in vitro after adrenaline exposure, at the
same temperature (8°C). The measured contents of intact adrenaline reflect the net sum of the
total uptake minus the metabolized fractions. The net atrial content and/or metabolism of
adrenaline seems to be temperature dependent, with a higher net uptake of adrenaline both at
2°C and at 14°C than at 8°C.
By comparing the effects of temperature and adrenaline on the atrial myocardium of the
Atlantic salmon (Floysand and Helle 1994) and the rainbow trout (Ask et al. 1981), subjected to
the same experimental design, species differences in their adrenergic activation were identified.
Adrenaline led to a pronounced inotropic response in the Atlantic salmon atrium, being larger at
2°C than at 8°C and 14°C , similar to, but not as marked as in the rainbow trout atrium. The
apparent affinities for adrenaline for the HR at 8°C and 14°C and for the maximal tension
responses at 2°C and 8°C were higher in the Atlantic salmon than in the rainbow trout,
indicating a higher affinity for circulating adrenaline in the Atlantic salmon than in the rainbow
trout myocardium in vivo. In any case, in both salmonids the enhanced adrenergic effect on
-------
myocardial inotropism (maximal tension response) at low temperature may point to a functional
role of the circulating catecholamines in adjusting the heart to maximal performance at sudden
declines in environmental temperature.
For the cold-adapted Antarctic notothenioid teleosts, high temperature represents one of
the most severe stressors, but these teleosts seem either barely sensitive or totally insensitive to
cardiac adrenergic stimulation. When the active cryopelagic Pagothenia borchgrevinki and the
benthic Trematomus bernacchii wereover-stressed by changing seawater temperature from
-1°C to 10°C, and held for 10 minutes, they released catecholamines into their circulation.
HR was unchanged in P. borchgrevinki but doubled in T. bernacchii, although ventral aorta
blood pressure remained constant in both fishes. Interestingly, despite high levels of circulating
catecholamines, P. borchgrevinki was able to maintain control of cardiac function relying on a
neural, cholinergic tone of HR. Similarly, infusion of adrenaline into Notothenia coriceps and
icefish (Chaenocephalus aceratus) raised blood pressure but not HR (Egginton 1997).
T. bernacchii is able to control its HR up to about 2.5°C, after which control is lost and HR
follows a Qio response. Available evidence suggests that cholinergic control of the heart had
been lost at the high experimental temperature, the increased HR being a result of temperature
and possibly also of elevated levels of catecholamines (Forster et al. 1998).
In view of these different patterns of adrenergic activation of cardiac function among
teleost species, we must be cautious in seeking generalised adaptive explanations. So far, the
available evidence is consistent with a poor or absent compensation of resting HR in intact fish
during acute thermal changes. The documented direct effects of temperature on HR indicate that
the acute Cho for HR is about 2.0 for isolated hearts and for intact fish (Graham and Farrell
1989).
Temperature Acclimation
The adrenergic responses obtained in the heart of acclimated fish can be different from
those obtained in acute conditions. For example, Graham and Farrell (1989) found that in the
in situ perfused hearts of rainbow trout at two acclimation temperatures, 5° C and 15°C, CO was
apparently affected more by temperature-dependent changes in HR than it was by temperature-
dependent effects on inotropism. Adrenaline increased HR at 5°C and 15°C, cardiac
preparations were more sensitive to filling pressure at 15°C in the presence of high adrenaline
concentrations, and under conditions of volume loading adrenaline significantly increased HR
at 15°C, while under pressure loading conditions adrenaline increased SV at both 5°C and 15°C
(Ibid.). In conclusion, the resting levels of catecholamines seemed to confer an important
cardiac tonus in intact rainbow trout. At 5°C this tonus appeared to be more relevant in
preserving cardiac rhythmicity, while at 15°C it stimulated inotropy to a greater degree. The
larger ventricle mass of rainbow trout acclimated at 5°C was associated with a higher absolute
stroke work at 5°C than at 15 °C.
The adrenergic modulation of cardiac performance at various temperatures in each species
depends on the type of myocardial adrenoreceptor involved. For example, in rainbow trout
exposed to colder temperatures, in which an adrenergic tone can stimulate the atrial pacemaker,
9
-------
preventing at the same time atrioventricular desynchronization, the positive chronotropy and
inotropy are mediated by P-adrenoceptors (Ask et al 1981; Graham and Farrell 1989) without
an involvement of a-adrenoceptors. A similar situation has been reported in plaice
(Pleuronectes platessd), goldfish, and common carp, but not in Eurasian perch (Perca
fluviatilus) (Tirri and Lehto 1984). In the European eel an a-adrenergic stimulation was
associated with opposite adrenergic effects (Pennec and Peyraud 1983). Adrenaline is generally
10 times more effective than noradrenaline in stimulating P-adrenoceptors in the heart of
rainbow trout (Ask et al. 1981).
There is evidence that a significant degree of plasticity exists in the density of myocardial
cell-surface P-adrenoceptors. An inverse relationships between myocardial p-adrenoceptor Bmax
and temperature for rainbow trout has been identified (Keen et al. 1993, Gamperl et al. 1994),
the density of myocardial cell-surface P-adrenoceptors being significantly higher in trout
acclimated at 8°C than in trout at 14°C5 with Bmax increasing by 11% with each 1°C decrease in
water temperature. Such temperature dependence might diminish in rainbow trout the ability of
catecholamines to stimulate myocardial performance at high water temperatures, as suggested by
a decreased myocardial sensitivity to catecholamines concomitant with the 180% reduction in
cell-surface p-adrenoceptor Bmax of trout between 8 and 18°C (Keen et al. 1993). Indeed, Farrell
et al. (1996) showed that in situ maximum cardiac performance of rainbow trout at high
acclimation temperatures was not improved by increasing the perfusate adrenaline concentration
from 30 to 200 nM.
In ventricular myocytes of rainbow trout acclimated at 4 and 17°C and in crucian carp
acclimated at 4 and 24°C, a patch-clamp analysis of L-type Ca"1"1" revealed marked species-
specific differences both in calcium current density and its p-adrenergic modulation (Vornanen
1998b). The density of Ica was almost twice as high in crucian carp as in rainbow trout
ventricular myocytes under control conditions, but, under maximal P-adrenergic stimulation,
because of stronger stimulation in rainbow trout myocytes, this difference disappeared. This can
suggest that the phosphorylation of sarcolemmal Ca""" channels may be differently regulated in
the two fishes. Thermal acclimation, while it did not change either the density or the kinetics of
the L-type Ca"1"1" current in crucian carp, did accelerate the rate of current activation in cold-
adapted rainbow trout myocytes, without changing the density of the Ca** current. As a
consequence of faster current decay, the contribution of sarcolemmal Ca"^ current to total
cellular calcium concentration was smaller in cold-acclimated than in warm-acclimated rainbow
trout. In either fish species the responses of Ca"1"1" current to maximal P-adrenergic activation or
direct activation of adenylate cyclase by forskolin were not changed by thermal acclimation.
Two important conclusions can be drawn from these data: (1) the acclimation to cold does
not increase the density of sarcolemmal Ca^ current; and (2) the P-adrenergic stimulation of
Ca*"*' current is more important in regulating myocardial contractility in rainbow trout than in
crucian carp. These responses may be correlated to the different ecophysiology of the two
teleosts during the cold wintering period. The rainbow trout is a cold-active species and thus
may require efficient P-adrenergic stimulation to maintain adequate cardiac performance.
10
-------
In contrast, the crucian carp tolerates longer winter anoxia by metabolic down-regulation,
which is associated with lowering of cardiac activity. Accordingly, for the crucian carp heart
(3-adrenergic stimulation could be unnecessary and even metabolically detrimental.
COLD-INDUCED CARDIAC ENLARGEMENT:
ANY ROLE FOR CATECHOLAMINES ?
Cold-induced cardiac enlargement, particularly at the ventricular level, has been
documented in several teleost species and has been interpreted as an adaptive response
compensating for direct effects of low temperature on rate of enzyme catalysis and cardiac
contractility (for references see Driedzic and Gesser 1994, Driedzic et al. 1996, and Tota et al.
1998). This trend reaches an extreme in the case of some Antarctic stenothermal species, the
Antarctic icefish (Tota et al. 1998). Such complex program of cardiac remodelling involves an
active phase of myocyte proliferation and/or hypertrophy (both mechanisms being active in the
fish heart), extracellular matrix deposition and, in the case of vascularized type of hearts such as
that of the salmonids, angiogenesis. The humoral and neural changes participating to such
remodelling are unknown, but a putative role of the adrenergic system is suggestive. Recent
studies in mammals indicate that ot-adrenoceptors, particularly the cti -adrenergic receptors, play
a role in long term-adaptation processes of cardiac adjustment, including "abnormal" or
pathological development. For example, it has been suggested that aragonists have long term
effects on cardiac structure and function by enhancing expression of several proto-oncogenes,
c-myc, c-fos, and myosin light chain-2 and cardiac a-actin (Hwang et al. 1996). Seasonal and
thermal-induced changes in myosin heavy chain composition have been identified in the heart of
crucian carp (Vornanen 1994). To what extent such cold-induced restructuring of the fish heart
may be related to the hitherto unexplored actions of the adrenergic system remains an open
question of great ecophysiological interest.
ACKNOWLEDGEMENTS
This study was supported in part by Programma Nazionale di Ricerche in Antartide
(PNRA) 1977 (B. Tota).
REFERENCES
Ask, J.A. 1983. Comparative aspects of adrenergic receptors in the hearts of lower vertebrates.
Comparative Biochemistry and Physiology 76A:543-552.
Ask, J.A,, G. Stene-Larsen, and K.B. Helle. 1980. Atrial p2 -adrenoceptors in the trout. Journal
of Comparative Physiology 139:109-115.
Ask, J.A., G. Stene-Larsen, and K.B. Helle. 1981. Temperature effects on the (32 -adrenoceptors
of the trout atrium. Journal of Comparative Physiology 143:161-168.
11
-------
Axelsson, M., C. Agnisola, S. Nilsson, and B. Tota. 1998. Fish Cardio-circulatory Function in
the Cold. 7n:H.O. Portner and R.C. Playle, (Eds.). Cold Ocean Physiology. Society of
Experimental Biology, Seminar series: 66. Cambridge University Press 327-364.
Bernier, N.J. and S.F. Perry. 1999. Cardiovascular effects of angiotensin-II-mediated adrenaline
release in rainbow trout Oncorhynchus mykiss. The Journal of Experimental Biology
202:55-66.
Bowler, K., and R. Tirri. 1990. Temperature dependence of the heart isolated from the cold or
warm acclimated perch (Percafluviatilis). Comparative Biochemistry and Physiology
96A: 177-180.
Cameron, J.S., and E.F. O' Connor. 1979. Liquid chromatographic demonstration of
catecholamine release in fish heart. Journal of Experimental Zoology 209:473-479.
Cech, J.J., R.W. Bridges, D.M. Rowell, and P.J. Balzer. 1976. Cardiovascular responses of
winter flounder, Pseudopleuronectes americanus, to acute temperature increase.
Canadian Journal of Zoology 54:1383-1388.
Colletti, A.E., and K.R. Olson. 1988. Catecholamine metabolism by the perfused rainbow trout
gill. Journal of Experimental Zoology 248:177-184.
Driedzic, W.R., and H. Gesser. 1994. Energy metabolism and contractility in hearts of
ectothermic vertebrates: impact of hypoxia, acidosis and low temperature. Physiological
Reviews 74:221-258.
Driedzic, W.R., J.R. Bailey, and D.H. Sephton. 1996. Cardiac adaptation to low temperature in
non-polar teleost fish. Journal of Experimental Zoology 275:186-195.
Egginton, S. 1997. A comparison of the response to induced exercise in red- and white-blooded
Antarctic fishes. Journal of Comparative Physiology B:129-134.
Farrell, A.P., K.R. MacLeod, and B. Chancey. 1986. Intrinsic mechanical properties of the
perfused rainbow trout heart and the effects of catecholamines and extracellular calcium
under control and acidotic conditions. Journal of Experimental Biology 125:319-345.
Farrell, A.P., and D.R. Jones. 1992. The Heart. In: W.S. Hoar, D.J. Randall, and A.P. Farrell,
(Eds.). Fish Physiology, Vol. XIIB. Academic Press, New York, pp. 1-88.
Farrell, A.P., A.K. Gamperl, J.M.T. Hicks, H.A. Shiels, and K.E. Jain. 1996. Maximum cardiac
performance of rainbow trout (Oncorhynchus mykiss) at temperatures approaching their
upper lethal limit. Journal of Experimental Biology 199:663-672.
12
-------
Floysand, R., and K.B. Helle. 1994. Effects of temperature and adrenaline on the atrial
myocardium of the cultured Atlantic salmon. Fish Physiology and Biochemistry
13:23-30.
Forster, M. E., W, Davison, M. Axelsson, L. Sundin, C.E. Franklin, and S. Gieseg. 1998.
Catecholamine release in heat-stressed Antarctic fish causes proton extrusion by the red
cells. Journal of Comparative Physiology 168: 345-352.
Gamperl, A.K., M.M. Vijayan, C. Preira, and A.P. Farrell. 1998. p-receptors and stress protein
70 expression in hypoxic myocardium of rainbow trout and chinook salmon. American
Journal of Physiology 274:R428-R436.
Gamperl, A.K., Wilkinson, M., and R. G. Boutilier. 1994. B-adrenoreceptors in the trout
Oncorhynchus mykiss heart: characterization, quantification, and effects of repeated
catecholamine exposure. General and Comparative Endocrinology 95:259-272.
Graham, M.S., and A.P. Farrell. 1989. The effect of temperature acclimation and adrenaline on
the performance of a perfused trout heart. Physiological Zoology 62:38-61.
Hazel, J.R. 1995. Thermal adaptation in biological membranes: Is homeoviscous adaptation
the explanation? Annual Review Physiology 57:19-54.
Hove-Madsen, L. 1992. The influence of temperature on ryanodine sensitivity and the force-
frequency relationship in the myocardium of rainbow trout. Journal of Experimental
Biology 167:47-60.
Hove-Madsen, L., P.P. Mery, J. Jurevicius, A.V. Skeberdis, and R. Fischmeister. 1996.
Regulation of myocardial calcium channels by cyclic AMP metabolism. Basic Research
in Cardiology 91, Supplement 2:1-8.
Hwang, K., C. Gray, W.E. Sweet, C.S. Moravec, and M. Im. 1996. ai -adrenergic receptor
coupling with Gh in the failing human heart. Circulation 94:718-726.
Jonsson A.C., I. Wahlqvist, and T. Hansson. 1983. Effects of hypophysectomy and cortisol on
the catecholamine biosynthesis and catecholamine content in chromaffin tissue from
rainbow trout, Salmo gairdneri. General Comparative Endocrinology 51:278- 285.
Keen, J.E., D.M. Vianzon, A.P. Farrell, and G.F. Tibbits. 1993. Thermal acclimation alters both
adrenergic sensitivity and adrenoceptor density in cardiac tissue of rainbow trout.
Journal of Experimental Biology 181:27-47.
Laurent, P., S. Holmgren, and S. Nilsson. 1983. Nervous and humoral control of the fish heart:
structure and function. Comparative Biochemistry and Physiology 76A. 525-542.
13
-------
Matikainen, N., and M. Vornanen. 1992. Effect of season and temperature acclimation on the
function of crucian carp (Carassius carassius) heart. Journal of Experimental Biology
167:203-220.
Mazeaud, MM, F. Mazeaud, and E.M. Donaldson. 1977. Primary and secondary effects of
stress in fish: some new data with a general review. Transaction of the American
Fisheries Society 106:201-212.
Nilsson, S. 1983. Autonomic Nerve Function in Vertebrates. Zoophysiology Vol. 13. Springer-
Verlag, Berlin.
Nilsson, S., T. Abrahamsson, and DJ. Grove. 1976. Sympathetic nervous control of adrenaline
release from the head kidney of the cod, Gadus morhua. Comparative Biochemistry and
Physiology 055:123-127.
Pennec, J.P., Y.M. Le Bras. 1984. Storage and release of catecholamines by nervous endings in
the isolated heart of the eel (Anguilla anguilla). Comparative Biochemistry and
Physiology C 77:167-172.
Pennec, J.P., and C. Peyraud. 1983. Effects of adrenaline on isolated heart of the eel (Anguilla
anguilla) during winter. Comparative Biochemistry and Physiology C 74:477-480
Randall, D.J., and S.F.Perry. 1992. Catecholamines. In : W.S. Hoar, DJ. Randall, A.P. Farrell,
(Eds.). Fish Physiology, Vol. XIIB. Academic Press, New York. pp. 255-300.
Reid, S.G., N.J. Bernier, and S.F. Perry. 1998. The adrenergic stress response in fish: control of
catecholamine storage and release. Comparative Biochemistry and Physiology
C 120:1-27.
Santer, R. 1985. Morphology and innervation of the fish heart. Advances in Embryology and
Cell Biology 89:1-102.
Selye,H. 1956. The Stress of Life. McGraw-Hill, NewYork.
Shiels, H.A., E.V. Freund, A.P. Farrell, and B.A. Block. 1999. The sarcoplasmic reticulum
plays a major role in isometric contraction in atrial muscle of yellowfin tuna. Journal of
Experimental Biology 202:881-890.
Thomas, S., and S.F. Perry. 1992. Control and consequences of adrenergic activation of red
blood cell Na+/H+ exchange on blood oxygen and carbon dioxide transport in fish.
Journal of Experimental Zoology 263:160-175.
Tibbits, G.F., C.D. Moyes, and L. Hove-Madsen. 1992. Excitation-contraction coupling in the
teleost heart. In: W.S. Hoar, D.J. Randall, and A.P. Farrell, (Eds.). Fish Physiology. Vol.
XIIA. Academic Press, New York. pp. 267-304.
14
-------
Tirri, R., and H. Lehto. 1984. Alpha and beta adrenergic control of contraction force of perch
heart. Comparative Biochemistry and Physiology A 77:301-304.
Tota, B., M.C. Cerra, R. Mazza, D. Pellegrino, and J. Icardo. 1998. The heart of the Antarctic
icefish as paradigm of cold adaptation. Journal of Thermal Biology 22:409-417.
Vornanen, M. 1994. Seasonal and temperature-induced changes in myosin heavy chain
composition of the crucian carp hearts. American Journal of Physiology 267:R1567-
1573.
Vornanen, M. 1998a. Ca^-ATPase activity and Ca4"*" uptake by sarcoplasmic reticulum in fish
heart: effects of thermal acclimation. Journal of Experimental Biology 201:525- 532.
Vornanen, M. 1998b. L-type Ca44" current in fish cardiac myocytes: effects of thermal
acclimation and p-adrenergic stimulation. Journal of Experimental Biology 201:533-
547. .
Wahlqvist, I. 1981. Branchial vascular effects of catecholamines released from the head kidney,
Gadus morhua. Molecular Physiology 1:235-241.
Wood, C.M., P. Pieprzak, and J.N. Trott. 1979. The influence of temperature and anaemia on
the adrenergic and cholinergic mechanisms controlling heart rate in the rainbow trout.
Canadian Journal of Zoology 57:2440-2447.
15
-------
-------
EFFECTS OF POLLUTANTS ON THE REPRODUCTION OF FISHES
Lin Hao-Ran1
ABSTRACT
Pollutants can impair reproductive neuroendocrine/endocrine function in fishes by
exerting their effects at a variety of sites on the brain-pituitary-gonadal axis, including effects on
hypothalamo-hypophysial complex, steroidogenesis, gonadal development, and maturation as
well as on ovulation and spawning. During the past several decades, the rapid development of
industrial activity and the modernization of agriculture and aquaculture has resulted in increased
levels of pollutants in aquatic environments. More and more evidence indicates that heavy metals,
organophosphorus pesticides, organochlorine pollutants, carbamate pesticides, and industrial
pollutants can seriously interfere with the normal functioning offish reproduction and decrease the
fecundity of fish populations. However, the mechanisms of toxic disruption have not been
elucidated for the majority of reproductive toxicants because of the overall complexity of the
reproductive neuroendocrine/endocrine system in vertebrates, including fishes.
INTRODUCTION
During the past several decades, the rapid development of industrial activities and the
modernization of agriculture and aquaculture have resulted in increased levels of contaminants
in aquatic environments, including rivers, lakes, ponds, reservoirs, and coastal ocean regions
(Akerman et al. 1996, Browne al. 1996,Ericson<#a/. 1996, Hose et al. 1996,Kocane£a/. 1996a,
1996b, Leblanc et al. 1997, McGurk and Brown 1996, Middaugh et al. 1998, Tanner and Knuth
1995, 1996, Wilson and Tillitt 1996). Meanwhile, in association with the contaminants that come
from agriculture, aquaculture, and industrial activities, there is widespread occurrence of
endocrine-disrupting chemicals (Stone 1994). Many of these compounds act as anti-estrogens
by interfering with the activity of the estrogen receptor or by reducing the number of receptors
in the organism (Goldberg 1995). Colbom et al. (1993) have identified a number of potential
endocrine-disrupting chemicals that include herbicides, fungicides, insecticides, nematocides, and
other chemicals (Table 1). The reproductive alterations in fishes exposed to these contaminants
are widespread (Dhawan and Kaur 1996, Ensenbach and Nagel 1997, Foster and Berlin 1997,
Guiney et al. 1996, Helmstetter and Alden 1995, Hwang et al. 1995, Stoulhart et al. 1998, Zabel
et al. \ 995). Proper chemical names for those chemicals reported by trade or common name in
Table 1 are listed in Table 2.
'institute of Aquatic Economic Animals, School of Life Sciences, Zhongshan University, Guangzhou, China
17
-------
Table 1. Chemicals with widespread distribution in the environment reported to have reproductive
and endocrine-disrupting effects.1
Herbicides
2,4-D
2,4,5-T
Alachlor
Amitrole
Alra/ine
Mctribuzin
Nitrofen
Trifluralin
Lindane(y-HCH)
Fungicides
Benomyl
Hexaclilorobenzene
Mancozeb
Maneb
Meliram-complex.
Triburyl tin
Zineb
Ziram
Biocides
Industrial Chemicals
Insecticides Nematocides
P-HCH
Carbaryl
Chlordane
Dicofol
Dieldria
DDT
Endosulfan
Heptachlor
& H-epoxide
Melhomyl Aldicarb
Methoxychlor DBCP
Mirex
Oxychlordane
Paralhion
Synthetic
pyrethroids
Toxaphene
Transnonachlor
Cadmium
Dioxin
Lead
Mercury
Polybrominated biphenyls
Polychlorinated biphenyls
Pentaclilorophenol
Phthalates
Styrenes
Colbornetal. 1993.
In recent years, the potential risk of environmental pollutants on fish reproduction has
become an issue of increasing concern and interest (Johnson et al. 1995, Matta et a/. 1997). Lin
(1990) published the first review which covers most of the literature up to 1989. Kime (1995)
published the next review citing most of the papers up to 1993. The present review will summarize
more recent research progress in this area. To date, the effects of more than 80 pollutants on the
reproduction of 60 species of fishes have been examined, including the effects of heavy metals,
organophosphorus pesticides, organochlorine pollutants, carbamate pesticides and other industrial
pollutants on hypothalamo-hypophysial complex, steroidogenesis, gonadal development and
maturation.
EFFECTS OF POLLUTANTS ON NEUROENDOCRINE/ENDOCRINE
FUNCTIONS IN RELATION TO REPRODUCTION
Fishes, like all other vertebrates, integrate their reproductive activities with seasonal
environmental cycles; certain environmental factors, such as temperature, photoperiod, and
rainfall act as cues for the approaching season which is favorable for reproduction. Signals from
environmental cues and endogenous physiological cycles provide input to the neuroendocrine
system, which in turn regulates pituitary and gonadal function (Lin and Peter 1996). This is the
neuroendocrine/endocrine regulatory function of the hypothalamo-hypophyseal-gonadal axis on
fish reproduction. Environmental contaminants, like other detrimental environmental factors,
can impair reproductive neuroendocrine/endocrine functions in fishes by exerting their effects at a
18
-------
Table 2. Proper chemical names for trade names used in this paper.
Alachlor: 2-chloro-2',6'-diethyl-N- (methoxymethyl) acetanalide
Aldicarb: 2-methyl-2- (methylthio) propionaldehyde o-(methylcarbamoyl)oxime
Amitrole: 3-amino-l,2,4-triazole .
Arochlor: a PCS mixture
Atrazine: 2-chloro-4-ethylamino-6-isopropylamino-5-triazine
Benotnyl: methyl 1- (butylcarbamoyi) -2-benzimidazoi carbamate
Carbaryl: 1-naphthyl-l-N-methylcarbamate
Chlordane: 1,2,4,5,6,7,8,8-octachloro-3a,4,7,7a-tetrahydro-4,7-methanoindane
Chlordecone: 1,la,3,3a,4,5a,5b>6-decachtorooctahydro-l,3,4-metheno-2H-cyclobuta[c,d] pentalen-2-one
Cythion: dicarboethoxyethyl O,O-dimethyl phosphorodithioate
2,4-D: 2,4-dichlorophenoxyacetic acid
o,/>'-DDD: 1- (2-chlorophenyl) -1- (4-cWorophenyl) -2,2-dichloroethane
DDT: 2,2-bis (p-chlorophenyl) -1,1,1 -trichloroethane
DBCP: 1,2 dibromo-3-chloropropane
Dicofol: 2,2,2-trichloro-l,l-bis (4-chlorophenyl) ethanol
Dieldrin: 1,2,3,4,10,10-hexachloro-6,7-epoxy-l,4,4a,5,6,7,8,8a-octahydro-exo-l,4-eHofe>-5,8-dimethanonaphthalene
Dioxin: 2, 3", 7',8-tetrachlorodibenzodioxin
Endosulfan: 1,2,3,4,7,7-hexachlorobicyclo [2.2.1]-2-heptene-5,6-bisoxymethylene sulphite
y-HCH: y-hexachlorocyclohexane
Heptachlor: l,4,5,6,7,8,8-heptachloro-3a,4,7,7a-tetrahydro-4,7-methanoindene
Heptachlor epoxide: l,4,5,6,7,8,8-heptachloro-2,3-epoxy-3a,4,7,7a-tetrahydro-4,7-methanoindan
Kepone: Alternate name for Chlordecone
Lindane: y-l,2,3,4,5,6-hexachlorocyclohexane
Malathion: diethyl [ (dimethoxyphosphinothioyl) thio ] butanedioate
Mancozeb: manganese ethylene bisdithiocarbamate
Maneb: manganese [ [1,2-ethanediylbis [carbamodithioato] ] (2-)]
Methomyl: methyl-N- [ [ (methylamino) carbonyl] oxy] ethanimidothioate
Methoxychlor: l,l,l-trichloro-2,2-bis [p-methoxyphenyl] ethane
Metiram: ethylenebis (dithiocarbamic acid), polymer with ammonia complex of zinc ebdc
Metribuzin: 4-amino-6-tert-butyl-3-methylthio-as-triazin-5-(4n)-one
Mirex: l,la,2,2,3,3a,4,5,5,5a,5b,6-dodecachlorooctahydro-i,3,4-metheno-lH-cyclobuta [cd] pentalene
Nitrofen: 2,4-dichloro-4'-nitrodiphenyl ether
Oxychlordane: 2,3,4,5,6,6a,7,7-octachloro-la,lb,5,5a,6,6a-hexahydro-2,5-methano-2H-indeno [1,2-b] oxirene
Parathion: O,O-diethyl O,/?-nitrophenyl phosphorothioate
2,4,5-T: 2,4,5-trichlorophenoxyacetic acid
TCB: 1,3,5-trichlorobenzene
Toxaphene: a chlorinated camphene containing 67 to 69% chlorine
Transnonachlor: 1,2,3,4,5,6,7,8,8-nonachlor-2,3,3a,4,7,7a-hexahydro-4,7-methano-lH-indene
Trifluralin: a,a,a-trifluoro-2,6-dinitro-N,N-dipropyl-jy-toluidine
Zineb: zinc [ [1,2-ethanediylbis [carbamodithioato] ] (2-)]
Ziram: zinc dimethyldithiocarbamate
19
-------
variety of sites on the brain-pituitary-gonadal axis (Celius and Walther 1998, Chatterjee et al. 1997,
Friedmann et al, 1996, Ghosh and Thomas 1995, Hazarika and Das 1998, Holm et al \ 993,1994,
Thomas et al. 1995, Ungerer and Thomas 1996). However, the mechanisms of neuroendocrine/
endocrine disruption have not been well elucidated for the majority of reproductive toxicants
because of the overall complexity of the reproductive neuroendocrine/endocrine system in fishes
and all other vertebrates. Recent evidence suggests that the brain is an important site of chemical
interference with reproductive neuroendocrine functions (Wilson and Leigh 1992). In the brain of
teleosts, there are multifactorial controls of gonadotropin secretion from the pituitary, a
stimulatory influence of one or more gonadotropin-releasing hormones (GnRHs), a stimulatory
or inhibitory influence of monoamine neurotransmitters, e.g. dopamine (DA), norepinephrine
(ME), serotonin (5-HT), and some other neuroendocrine factors.
In general, histological and hormone assays were used to examine the effects of pollutants
on the hypothalamic-pituitary system. In fishes exposed to heavy metals (e.g. cadmium, mercury,
lead) or the organophosphorus pesticides (e.g. cythion, malathion, endosulfan), the neurones
of both nucleus preopticus and nucleus lateralis tuberis which are involved in neuroendocrine
regulation of reproduction are deformed and degeneratively changed, with decreased
neurosecretory materials; the pituitary gonadotrophs and thyrotrophs are small, vacuolized,
inactive and few in number. These histological damages are simultaneous with the decreasing
of a hypothalamic GnRH-like substance and pituitary and plasma gonadotropin concentrations.
These effects, in turn, can result in inhibition of gonadal development and maturation. For
example, in Atlantic croaker (Micropogonia.fi undulatus), 5-HT has been shown to potentiate the
action of LHRH analogs in inducing gonadotropin hormone (GtH) release, and the close proximity
and co-localization of the GnRH and 5-HT in the preoptic-anterior hypothalamic area, a brain
region involved in regulating the reproduction in vertebrates, provides the morphological basis for
the possible interaction of these two hormones in the neuroendocrine control of reproduction
(Khan and Thomas 1992). Polychlorinated biphenyls (PCBs) have been shown to influence
monoamine neurotransmitter metabolism in discrete areas of the brain in mammals. Recent
studies have shown that there was a significant decline in 5-HT and DA concentrations in
hypothalamus of Atlantic croaker exposed to Arochlor 1254 (a PCB mixture). The fish were fed
a diet containing Arochlor 1254,0.1 mg/lOOg body weight/day for 30 days. Arochlor 1254
exposure also resulted in the loss of the in vitro GtH response to stimulation by a LHRH analog,
and this indicated that the decline in pituitary GtH release by Arochlor 1254-exposure may be
mediated, at least partially, by impaired hypothalamic serotonegic activity (Khan and Thomas
1996). The decrease observed in GtH secretion from the pituitary will subsequently result in the
lower circulating levels of sex steroids observed in the Atlantic croaker exposed to Arochlor 1254
(Ibid.).
Since the process of reproductive physiology is controlled mainly by GtH and sex steroid
hormones, it is possible that pollutant-induced reproductive dysfunction could be due to impaired
endocrine function of pituitary/gonad. Singh et al. (1994) investigated the impact of an
20
-------
organocblorine pesticide y-hexachlorocyclohexane (y-HCH) exposure on plasma GtH levels
in goldfish (Carassius auratm). Male and female fish were exposed to safe (SC, 0.01 ppm)
and sublethal (SL, 0.1 ppm) concentrations of y-HCH for 4 weeks during the prespawning phase
(June) of the annual reproductive cycle. GtH levels and gonadosomatic index (GSI) were
significantly lower after exposure to both safe and sublethal y-HCH concentrations than were the
controls. After 4 weeks, gonadal tissue from both exposed fish and the controls was incubated
with pituitary homogenate from carp (Cyprinm carpio). Tn both sexes, testosterone and
11-ketotestosterone production Was greatly decreased in fish exposed to y-HCH compared to the
controls; 17a, 20p-dihydroxyprogesterone production was low in all fish and was not significantly
affected by y-HCH. This is probably due to the fact that it is secreted for only a very short period
and for a few hours, so it is difficult to obtain samples at exactly comparable stages, or time of
secretion. These results indicate that y-HCH inhibited gonadal recrudescence by decreasing
plasma concentrations of GtH and possibly its gonadal receptors, and by altering GtH potential
for stimulation of steroidogenesis.
A variety of organochlorines such as dichlorodiphenyltrichloroethane (DDT),
methoxychlor, PCB mixtures, and kepone have estrogenic actions and disrupt reproductive
function in mammals by binding to nuclear estrogenic receptor. In Atlantic croaker, Thomas et al.
(1995) have tested the ability of these xenobiotics to bind the hepatic estrogen receptor. Several of
the DDT derivatives, kepone, and PCB mixtures also bound to the Atlantic croaker estrogen
receptor over a range of 10"5-10"3 M. These estrogenic compounds were also tested for their
ability to bind the ovarian membrane, which is the receptor for the maturation-inducing steroid
17a, 20p-21-trihydroxy-4-pregnen-3-one (20(3-8), kepone, methoxychlor, and DDT caused
concentration-dependent displacement of 3H-20f3-S from its receptor site and inhibition of
20p-S-induced final oocyte maturation in an in vitro bioassay over the range of 10"7-10"3 M.
Therefore, these studies indicate that a variety of organochlorines with estrogenic actions can
also bind to other sex steroid receptors to influence their normal reproductive processes.
In general, measures of plasma concentrations of hormone will indicate overall effects
on circulating hormones but do not give information as to the site or sites at which pollutants exert
their effects. Decreased plasma steroids could result from inhibition of hypothalamic GnRH or
pituitary GtH, decreased ovarian cholesterol, decreased activity of any of the enzymes in the
biosynthetic sequence from cholesterol to the hormone, or increased rate of hepatic catabolism and
excretion, etc. In addition, it is always unclear whether the gonadal damage described is the cause
or effect of altered steroid synthesis. Tn this respect, in vitro incubations of tissue using endogenous
or exogenous precursors may yield valuable information. For example, testicular fragments from
spermiating roach (Rutilus rutilus) were incubated with 0, 1,10, and 20 mg/L of y-HCH, together
with carp pituitary homogenate or 3H-17-hydroxyprogesterone (3H-17P) (Singh and Kime 1995).
Endogenous production of testosterone (T) was stimulated by 1 mg/L y-HCH but inhibited at
higher doses, while its glucuronide was increased in all incubations containing y-HCH;
17-hydroxyprogesterone (17-P) and 11-ketotestosterone (KT) production was decreased at all
21
-------
concentrations of y-HCH. But the low endogenous 17a, 20f3-dihydroxyprogesterone (17a, 20(3P)
production was unaffected. y-HCH had a much lower impact on the metabolism of exogenous
17P than on endogenous precursors. When y-HCH was incubated with 3H-17P, the major
metabolite of 17P, 17a, 20PP was significantly increased together with a decrease of its
glucuronide; T was also significantly increased at some concentrations, but none of the other
metabolites was affected by y-HCH.
These results show that the balance of steroid production by testes of spermiating roach is
significantly perturbed by y-HCH and suggests that the pesticide predominantly affects the stage
of steroidogenesis leading to 17-P production. McMaster ttt al. (1995) used an in vitro gonadal
incubation procedure to examine reproductive function in white sucker (Catostomus commersoni)
and brown bullhead (Iclalurus nebulosus) collected at sites exposed to effluents of bleached kraft
pulp mill, sulphite pulp mill, and steel mill. Studies on the effects of bleached kraft pulp mill
effuent demonstrated that the reduced production of steroid hormones by ovarian follicles
paralleled the reduction found in circulating steroid levels. These reductions also corresponded to
other reproductive alterations in white sucker populations such as decreased gonad size, reduced
expression of secondary sexual characteristics, and increased age to maturation. Studies on the
mechanisms of bleached kraft pulp mill effluent action on steroid production indicate that multiple
sites within the steroid biosynthetic pathway are disrupted. Brown bullheads, collected at a site
heavily contaminated with polycyclic aromatic hydrocarbons, demonstrated similar reductions in
gonad size, circulating steroid levels, and in vitro steroid production. However, brown bullheads
exposed to bleached kraft pulp mill effluent failed to show signs of reproductive alteration due to
effluent exposure. White suckers collected from this same site showed reduced gonad size, but
failed to show a consistent correlation between reduced circulating steroids and in vitro steroid
production. These studies indicate that reproductive alterations in feral fishes exposed to organic
pollutants are widespread. Although some sites exhibit identical reproductive responses,
differences within species between sites suggest that different mechanisms may be responsible for
the reproductive changes. The lack of reproductive responses in brown bullheads exposed to
bleached kraft pulp mill effuent indicate species differences exist in reproductive responsiveness
to effluent exposure. Therefore, multiple species should be evaluated for reproductive alterations
when conducting field assessment.
EFFECTS OF POLLUTANTS ON GONADAL DEVELOPMENT AND MATURATION
The teleost ovary undergoes a seasonal reproductive cycle which may be divided into
four main phases: (1) vitellogenesis; (2) oocyte maturation; (3) ovulation and spawning;
(4) postspawning. Long-term exposure to pollutants, which generally begins in the long period of
vitellogenesis, almost invariably leads to a decrease in GSi, a greater number of smaller and
less-developed oocytes, fewer large mature oocytes, and an increase in the number of atretic
follicles. Oocytes frequently contained less yolky granules, ruptured oocyte walls, damaged yolk
22
-------
vesicles, nucleoli and cytoplasmic changes, and finally, suppression or delay of ovarian maturation,
as well as ovulation and spawning (Barry et al, 1995, Chatterjee and Ghosh 1995, Lin and Hwang
1998a, 1998b, Miller and Amrhein 1995, Welsh et al 1996). Since the incorporation of
vitellogenin into the developing oocytes is the major part of the growth during ovarian
recrudescence, the lower GSI in fish exposed to pollutants is probably due to inhibition of
vitellogenesis, and/or to inhibition of the uptake of vitellogenin into oocytes. For example, white
perch (Morone americana) exposed to 3, 3', 4, 4'-tetrachlorobiphenyl (TCB) during late
vitellogenesis, produced fewer mature and maturing females as compared to control fish
(Monosson et al. 1994, 1996). Of those fish that did mature, GSI was inversely related to TCB
dose; the mean GSI of the control fish was approximately two-fold higher than that of
TCB-exposed fish; oocyte diameters were also significantly reduced. In fish close to ovulation,
there were decreasing estradiol (E2) and viellogenin (VTG) levels with increasing TCB dose, and
significantly lower EI and VTG levels in immature fish compared with mature or maturing fish.
This suggests that the pollutant inhibited estrogen synthesis and hepatic production of VTG, and
possibly inhibited uptake of VTG into the oocyte. There were no detectable effects of TCB on
hatching success, body length of larvae at hatching, or larval growth up to 5 days after hatching;
there was, however, a decline in survival near the end of yolk sac absorption in larvae from
TCB-exposed females.
As with the female, one of the most common measures of the effects of pollutants in male
fishes is the GSi. The GSI is decreased by many pollutants in male fishes. Inhibition of
spermatogenesis, with large number of spermatogonia and spermatocytes and few spermatids and
mature sperm, atrophy or necrosis of the interstitial cells and changes in Sertoli cell structure are
common effects of heavy metal and pesticides (organophosphorus, organochlorine, and carbamate)
on testes. For example, extensive damage in the testes was caused by both acute as well as short-
and long-term sub-acute exposure of catfish (Heteropneustes fossilis) to chlordecone (Srivastava
and Srivastava 1994). Seminiferous tubules showed a flattened, degenerated and desquammated,
germinal epithelium, spermatids and sperms showed cytolysis, interstitial tissues were atrophied
and Leydig's cells vacuolized, and leucocyte infiltration was observed in various part of the testes.
In addition, exposure of male fishes to pollutants may result in malformed sperm and a decrease of
sperm motility. Sperm cells may be a good measure of toxicity because they are more susceptible
to pollutants than the oocyte in female fishes; they require only a short exposure time, and their
mobility is rapidly and easily measured.
Final maturation of oocytes in teleosts is initiated by GtH, which induces both migration
of the geminal vesicle to the periphery of the oocyte and the follicular synthesis of a
maturation-inducing steroid (e.g. 1 la, 20PP; 20p-S). The maturation-inducing steroid then causes
germinal vesicle breakdown (GVBD) which is followed by ovulation. Therefore, oocyte
maturation lasts a very short period, less than one day in cyprinids and up to 7 days in salmonids,
and is very susceptible to pollution. As noted above, the in vitro induction of GVBD by luteinizing
hormone (LH) was inhibited by both organophosphorus and organochlorine pesticides, but it is not
23
-------
clear whether this is due to inhibition of LH-induced germinal vesicle (GV) migration or to
inhibition of the synthesis of maturation-inducing steroid. Recently Thomas and Budiantara (1995)
reported that there were fewer fully grown primed oocytes from water-soluble fractions of
oil- and naphthalene-exposed Altantic croaker when they underwent GVBD in vitro
in response to 20p-S. Direct addition of naphthalene in vitro to primed oocytes also blocked
20p~S-induced GVBD, which suggest that some of the toxic effect of naphthalene on reproduction
may be due to membrane damage or interference with hormone membrane receptors in the target
cell. Furthermore, Ghosh and Thomas (1995) examined the antagonistic effects of kepone and
o,p'-DDD on the maturation-inducing steroid (20p-S) action on oocyte final maturation of
Atlantic croaker. Exposure to kepone or o,p'-DDD (100 nM-100|oM) prevented the majority of
the oocytes from completing GVBD in response to 20(3-S; many of them were arrested at the lipid
coalescence or GV migration stage. In addition, clearing of the ooplasm, oil droplet formation,
and hydration were incomplete in those xenobiotic-exposed oocytes that underwent GVBD. The
xenobiotics inhibited GVBD in a dose-dependent manner, washing the follicle-enclosed oocytes
after xenobiotic exposure restored their ability to undergo GVBD in response to 20|3-S. These
results suggest that kepone and o, p'-DDD exert antagonistic effects on 20p-S action, possibly by
competing for the ovarian 20p-S membrane receptor.
SUMMARY
To date, the effects of more than 80 pollutants on the reproduction of approximately
60 species of fishes have been examined. More and more evidence indicates that heavy metals,
organophosphorus pesticides, organochlorine pollutants, carbamate pesticides, and industrial
pollutants can seriously interfere with the normal functioning offish reproduction and decrease
the fecundity offish populations.
Pollutants can impair reproductive neuroendocrine/endocrine function in fishes by
exerting their effects at a variety of sites on the brain-pituitary-gonad axis. The brain is an
important site of chemical interference with reproductive neuroendocrine function. In Atlantic
croaker exposed to Arochlor 1254 the decline of pituitary GtH release is mediated, at least partially,
by impaired hypomalamic serotoninergic activity; the decrease in GtH secretion from the pituitary
will subsequently result in a lower circulating sex steroid level. Long-term exposure of goldfish to
y-HCH causes inhibition of gonadal recrudescence by decreasing plasma concentrations of
GtH and possibly its gonadal receptors, and by altering GtH potential for stimulation of
steroidogenesis. In vitro studies further indicate that the balance of steroid production by testes
of spermiating roach is significantly perturbed by y-HCH and suggests that the pesticide
predominantly affects the stages of steroidogenesis leading to 17P production. A variety of
organochlorine pollutants with estrogenic actions (e.g. DDT derivatives, kepone, PCB) can
also bind to the hepatic estrogen receptor and other steroid receptors in ovarian membranes
(the receptors for the maturation-inducing steroid) and suppress the effects of those steroids
for stimulating ovarian development and maturation.
24
-------
Long-term exposure of fishes to pollutants generally leads to a decrease in ovary GS1
and suppresses ovarian development. In TCB-exposed white perch during late vitellogenesis, GSI,
estradiol, and vitellogenin levels were inversely related to increasing TCB doses, suggesting that
the pollutant inhibited estrogen synthesis and hepatic production of vitellogenin, and possibly
inhibited uptake of vitellogenin into the oocyte. Long-term exposure to pollutants causes a
decrease in testis GS1 and suppresses spermatogenesis, as well as the numbers of mature sperm
and sperm motility. Sperm may be a good measure of toxicity because they are more susceptible to
pollutants than the oocyte in female fishes. Final maturation of oocytes is very susceptible to
pollutants. Naphthalene blocked Atlantic croaker oocytes from undergoing GVBD in vitro in
response to 20p-S. This may be due to oocyte membrane damage or interference with the binding
of hormone (20p-S) on the membrane receptor of oocyte.
ACKNOWLEDGEMENT
This work was supported by grants from the National Natural Sciences Foundation
of China and the Provincial Natural Sciences Foundation of Guangdong Province, China.
REFERENCES
Akerman, G., U. Tjarnlund, D. Broman, C. Naf, L. Eestin, and L. Balk. 1996. Comparison of
reproductive success of cod, Gadus morhua, from the Barents Sea and Baltic Sea. Marine
Environmental Research 42:139-144.
Barry, M.J., K. O'Halloran, D.C. Logan, J.T. Ahokas, and D.A. Holdway. 1995. Sublethal effects
of esfenvalerate pulse-exposure on spawning australian crimson-spotted rainbowfish
(Melanotaeniu fluviatilis). Archives of Environmental Contamination and Toxicology
28:459-463.
Brown, E.D., B.L. Norcross, and J. W. Short. 1996. An introduction to studies on the effects of the
Exxon Valdez oil spill on early life history stages of Pacific herring, Clupea pallasi, in
Prince William Sound, Alaska. Canadian Journal of Fisheries and Aquatic Science
53:2337-2342.
Celius, T., and B.T. Walther. 1998. Differential sensitivity of zonagenesis and vitellogenesis in
Atlantic salmon (Salmo salar L.) to DDT pesticides. Journal of Experimental Zoology
281:346-353.
Chatterjee, S., A.B. Dutla, and R. Ghosh. 1997. Impact of carbofuran in the occyte maturation of
catfish, Heteropneustes fossil'is (Bloch). Archives of Environmental Contamination and
Toxicology 32:426-430.
25
-------
Chatterjee, S., and R, Ghosh. 1995. Toxicity of carbofuran technical 75 DB to the fertilization of
eggs of catfish, Heteropneustesfossilis (Bloch). Bulletin of Environmental Contamination
and Toxicology 55:111-115.
Colborn, T., F.S. Vom Saal, and A.M. Solo. 1993. Developmental effects of endocrine-disrupting
chemicals in wildlife and humans. Environmental Health Perspectives 101:378-384.
Dhawan, A., and K. Kaur. 1996. Toxic effects of synthetic pyrethroids on Cyprinus carpio Linn
eggs. Bulletin of Environmental Contamination and Toxicology 57:999-1002.
Ensenbach, U., and R. Nagel. 1997. Toxicity of binary chemical mixtures: Effects on
reproduction of Zebrafish (Brachydanio rerio*). Archives of Environmental Contamination
and Toxicology 32:204-210.
Ericson, G., G. Akerman, B. Liewenborg, and L. Balk. 1996. Comparison of DNA damage in the
early life stage of cod, Gadus morhua, originating from the Barents Sea and Baltic Sea.
Marine Environmental Research 42:119-123.
Foster, N.R., and W.H. Berlin. 1997. Fertilization of eggs of Lake Michigan lake trout Salvelinus
namaycush in lake water: effect of PCBs (Aroclor 1254). Bulletin of Environmental
Contamination and Toxicology 59:160-466.
Friedmann, A.S., M.C. Watzin, J.C. Letter, and B. Johnson. 1996. Effects of environmental
mercury on gonad function in Lake Champlain northern pike (Esox Indus). Bulletin of
Environmental Contamination and Toxicology 56:485-492.
Ghosh, P., and P. Thomas. 1995. Binding of metals to red drum vitellogenin and incorporation
into oocytes. Marine Environmental Research 39:165-168.
Ghosh, S., and P. Thomas. 1995. Antagonistic effects of xenobiotics on steroid-induced final
maturation of Atlantic croaker oocytes in vitro. Marine Environmental Research
39:159-163.
Goldberg, E.D. 1995. Emerging problems in the coastal zone for twenty first century. Marine
Pollution Bulletin 31:152-158.
Guiney, P.D., P.M. Cook, J.M. Casselman, J.D. Fitzsimmons, H.A. Simonin, E.W. Zabel, andR.E.
Peterson. 1996. Assessment of 2, 3, 7, 8-tetrachlorodibenzo-p-dioxin induced sac fry
mortality in lake trout (Salvelinus namaycush) from different regions of the Great Lakes.
Canadian Journal of Fisheries and Aquatic Science 53:2080-2092.
26
-------
Hazarika, R., and M. Das. 1998. lexicological impact of BHC on the ovary of the air-breathing
catfish, Heteropneustes fossilis (Bloch). Bulletin of Environmental Contamination and
Toxicology 60:16-21.
Helmstetter, M.F., and R.W. Alden III. 1995. Toxic responses of Japanese medaka (Oryzias
latipes) eggs following topical and immersion exposures to pentachlorophenol. Aquatic
Toxicology 32:15-29.
Holm, G., J. Lundstrom, T. Anderson, and L. Norrgren. 1994. Influences of halogenerated
organic substances on ovarian development and hepatic EROD activity in the three-spined
stickleback, Gasterosteus aculeatus, and rainbow trout, Oncorhynchus mykiss, Aquatic
Toxicology 29:241-256.
Holm, G., L. Narrgren, T. Anderson, and A. Thuren. 1993. Effects of exposure to food
contaminated with PBDE, PCN or PCB on reproduction, liver morphology and
cytochrome P450 activity in the three-spined stickleback, Gasterosteus aculeatus. Aquatic
Toxicology 27:33-50.
Hose, J.E.,M.D. McGurk, G.D. Marty, D.E. Hinton, E.D. Brown, andT.T. Baker. 1996. Sublethal
effects of the Exxon Valdez oil spill on herring embryos and larvae: morphological,
cytogenetic and histopathological assessments (1.989-1991), Canadian Journal of Fisheries
and Aquatic Science 53:2355-2365.
Hwang, P.P., S.W. Lin, and H.C. Lin. 1995. Different sensitivities to cadmium in Tilapia larvae
(Oreochromis mossamhicus; Teleostei). Archives of Environmental Contamination and
Toxicology 29:1-7.
Johnson, L.L., S.Y. Sol, D.P. Lomax, andK. Collier. 1995. Effects of endocrine-disrupting
chemicals on marine flatfish: An approach to environmental risk assessment. In:
Proceedings of the 5th International Symposium on the Reproductive Physiology of Fish.
F.W. Goetz and P. Thomas, (Eds.), p. 188, Fish Symposium '95. Austin, Texas, USA.
Khan, LA., and P. Thomas. 1992. Stimulatory effects of serotonin on maturational gonadotropin
release in the Atlantic croaker, Micropogonias undulatus. General and Comparative
Endocrinology 88:388-396.
Khan, LA., and P. Thomas. 1996. Disruption of neuroendocrine function in Atlantic croaker
exposed to Aroclor 1254. Marine Environmental Research 42:145-149.
Kime, D.E. 1995. The effects of pollution on reproduction in fish. Reviews in Fish Biology
and Fisheries 5(1 ):52-95.
27
-------
Kocan, R.M., J.E. Hose, E.D. Brown, and T.T. Baker. 1996a. Pacific herring (Clupeapallets!)
embryo sensitivity to Prudhoe Bay petroleum hydrocarbons: laboratory evaluation and
in situ exposure at oiled and unoiled sites in Prince William Sound. Canadian Journal
of Fisheries and Aquatic Science 53:2366-2375.
Kocan, R.M., G.D. Marty, M.S. Okihiro, E.D. Brown, and T.T. Baker. 1996b. Reproductive
success and histopathology of individual Prince William Sound Pacific herring 3 years
after the Exxon Valdez oil spill. Canadian Journal of Fisheries and Aquatic Science
53:2388-2393.
Leblanc, J., C.M. Couillard, and J.E Brethes. 1997. Modification of the reproductive period in
mummichog (Fundulm heteroclitus) living downstream from a bleached kraft pulp mill in
the Miramichi Estuary, New Brunswick. Canadian Journal of Fisheries and Aquatic
Science 54:2564-2573.
Lin, H.R. 1990. Effects of some pollutants on reproduction offish: A review. In: Proceeding
of an International Symposium on Fish Physiology, Toxicology, and Water Quality
Management. Sacramento, California, U.S.A., September 18-19, 1990. U.S.
Environmental Protection Agency Environmental Research Laboratory, Athens, Georgia
(EPA/600/R-93/157). pp. 1-11.
Lin, H.R., and R.E. Peter. 1996. Hormones and spawning in fish. Asian Fisheries Science
9:21-33.
Lin, H.C., and P.P. Hwang. 1998a. Acute and chronic effects of gallium cholride (GaCl3) on
tilapia (Oreochromis mossambicus) larvae. Bulletin of Environmental Contamination and
Toxicology 60:931-935.
Lin, H.C., and P.P. Hwang. 1998b. Acute and chronic effects of indium chloride (InCl3) on tilapia
(Oreochromis mossambicus) larvae. Bulletin of Environmental Contamination and
Toxicology 61:123-128.
Matta, M.B., C. Calmcross, and R.M. Kocan. 1997. Effects of a polychlorinated biphenyl
metabolite on early life stage survival of two species of trout. Bulletin of Environmental
Contamination and Toxicology 59:146-151.
McGurk, M.D., and E.D. Brown. 1996. Egg-larval mortality of Pacific herring in Prince William
Sound, Alaska, after the Exxon Valdez oil spill. Canadian Journal of Fisheries and Aquatic
Science 53:2343-2354.
28
-------
McMaster, M.E., G. Van Der Kraak, and K.R. Munkittrick. 1995. Differential modes of hormonal
disruption in fish exposed to various organic contaminants. In: Proceedings of the 5th
International Symposium on the Reproductive Physiology of Fish. F. W. Goetz and P.
Thomas (Eds.), pp. 191. Fish Symposium '95, Austin, Texas, USA.
Middaugh, D.P., M.E. Shelton, C.L. McKenney, G. Cherr, Jr., P.J. Chapman, andL.A. Courtney.
1998. Preliminary observations on responses of embryonic and larval Pacific herring,
Clupea pallasi, to neutral fraction biodegradation products of weathered Alaska North
Slope oil. Archives of Environmental Contamination and Toxicology134:188-196.
Miller, M.A., and J.F. Amrhein. 1995. Maternal transfer of organochlorine compounds in Lake
Superior siscowet (Salvelinus namaycush siscourf) to their eggs. Bulletin of
Environmental Contamination and Toxicology 55:96-103.
Monosson, E., W.J. Fleming, and C.V. Sullivan. 1994. Effects of the planar PCB 3,
3', 4,4'-tetrachlorobiphenyl (TCB) on ovarian development, plasma levels of sex steroid
hormones and vitellogenin, and progeny survival in the white perch (Morone americand).
Aquatic Toxicology 29:1-19.
Monosson, E., R.G. Hogson, W.J. Fleming, and C.V. Sullivan. 1996. Blood plasma levels of sex
steroid hormones and vitellogenin in striped bass (Morone saxatilis) exposed to 3, 3', 4,
4'-tetrachlorobiphenyl (TCB). Bulletin of Environmental Contamination and Toxicology
56:782-787.
Singh, P.B., D.E. Kime, P. Epler, and J. Chyb. 1994. Impact of y-hexachlorocyclohexane
exposure on plasma gonadotropin levels and in vitro stimulation of gonadal steroid
production by carp hypophyseal homogenate in Carassius auratus. Journal of Fish
Biology 44:195-204.
Singh, P.B., and D.E. Kime. 1995. Impact of y-hexachlorocyclohexane on the in vitro production
of steroids from endogenous and exogenous precursors in the spermiating roach, Rutilus
rutilus. Aquatic Toxicology 31:231 -240.
Srivastava, A.K., and A.K. Srivastava. 1994. Effects of chordecone on the gonads of freshwater
catfish, Heteropneustesfossilis. Bulletin of Environmental Contamination and Toxicology
53:186-191.
Stone, R. 1994. Environmental estrogens stir debate. Science 265:308-310.
29
-------
Stoulhart, X.J., M.A.A. Huijbregts, P.H.M. Bolm, R.A.C. Lock, and S.E.W. Bonga. 1998.
Endocrine stress response and abnormal development in carp (Cyprinux carpio) larvae
after exposure of the embryos to PCB 126. Fish Physiology and Biochemistry 18:321 -329.
Tanner, D.K., and M.L. Knuth. 1995. Effects of azinphos-methyl on the reproductive success of
the bluegill sunfish, Lepomis macrochirus, in littoral enclosures. Ecotoxicolgy and
Environmental Safety 32:184-193.
Tanner, D.K., and M.L. Knuth. 1996. Effects of esfenvalerate on the reproductive success of the
bluegill sunfish, Lepomis macrochirm in littoral enclosures. Archives of Environmental
Contamination and Toxicology 31:244-251.
Thomas, P., and L. Budiantara. 1995. Reproductive life history stages sensitive to oil and
naphthalene in Atlantic croaker. Marine Environmental Research 39:147-150.
Thomas, P., S. Gosh, and J.S. Smith. 1995. Interaction of organochlorines with steroid hormone
receptors in Atlantic Croaker. Marine Environmental Research 39:176.
Ungerer, J., and P. Thomas. 1996. Transport and accumulation of organochlorines in the ovaries
of Atlantic croaker (Micropogonias undulatus). Marine Environmental Research
42:167-171.
Welsh, P.O., J.L. Parrotl, D.G. Dixon, P.V. Hodson, D.J. Spry, and G. Mierle. 1996. Estimating
acute copper toxicity to larval fathead minnow (Pimephales promelas) in soft water from
measurements of dissolved organic carbon, calcium and pH. Canadian Journal of Fisheries
and Aquatic Science 53:1263-1271.
Wilson, C.A., and A.J. Leigh. 1992. In: Endocrine Toxicology. C.K. Atlerwill and J.D. Flack
(Eds.). Cambridge University Press, Cambridge, pp. 313-398.
Wilson, P. J., and D.E. Tillitt. 1996. Rainbow trout embryotoxicity of a complex contaminant
mixture extracted from Lake Michigan lake trout. Marine Environmental Research
42:129-134.
Zabel, E.W., P.M. Cook, and R.E. Peterson. 1995. Toxic equivalency factors of polychlorinated
dibenzo-p-dioxin, dibenzofuran and biphenyl congeners based on early life stage mortality
in rainbow trout (Oncorhynchus mykiss). Aquatic Toxicology 31:315-328.
30
-------
IDENTIFICATION OF AMMONIA AND VOLATILE PHENOLS AS PRIMARY
TOXICANTS IN A COAL GASIFICATION EFFLUENT
Hongjun JinVXuan Yang1, Hongxia Yu1, Daqiang Yin1
ABSTRACT
A toxicity identification evaluation (TIE) was conducted using Daphnia magna in an
effort to identify the toxicity cause of effluent discharged from the Coal Gasification Plant in
Nanjing, China. Phase I procedures suggested that ammonia and non-polar organics were the
primary suspect toxicants causing whole effluent toxicity. In Phase II, a zeolite test and
ammonia analysis of the effluent indicated that ammonia was responsible for a portion of the
whole effluent toxicity, and gas chromatography-mass spectroscopy (GC-MS) of the effluent's
toxic fraction revealed that six volatile phenols, 2,4-dimethylphenol, 3-ethylphenol,
3-methylphenol, 2-methylphenol, 2,6-dimethylphenol, and 2-ethylphenol were the major
non-polar organics present at sufficient concentrations to cause the toxicity. In Phase III,
mass balance tests and the fact that both ammonia and volatile phenols were present at toxic
concentrations in three samples of the effluent confirmed that ammonia and volatile phenols
accounted for the toxicity of the coal gasification effluent.
INTRODUCTION
In response to the increasing use of toxicity limits for discharges monitored under the
National Pollutant Discharge Elimination System (NPDES) in the United States, a series of
toxicity-based procedures to isolate and identify toxic compounds in complex effluents has
been developed (U.S. EPA 1991, 1993a, 1993b). These procedures for toxicity identification
evaluation (TIE) are manipulations designed to characterize classes of toxicants (Phase I),
identify specific toxic compounds within these classes (Phase II), and confirm that these
compounds actually are the true toxicants (Phase III) (Burkhard and Ankley 1989). The TIE
approach has been applied successfully to a variety of toxic aqueous samples including effluents,
sediment fractions, leachates, and ambient water (Ankley and Burkhard 1992). In recent years,
the impacts of complex effluents to aquatic ecosystems have increased with the development of
industry in China, especially in the coastal southeast. The effluent discharge limits in China now
emphasize general physico-chemical parameters as well as a few priority toxic pollutants among
thousands of chemicals. Effective control of toxicant discharge is difficult, and some treated
effluents which were once considered to have met the requirements of the national discharge
'State Key Laboratory of Pollution Control and Resource Reuse, and the Department of Environmental Science and
Engineering, Nanjing University, Nanjing, China.
31
-------
limits are still highly toxic to aquatic organisms based on bioassay results (Jin et al. 1994).
Therefore, it can be very useful to use the TIE approach as a way of identifying the real
toxicants causing whole effluent toxicity in China. This approach can also be helpful to develop
cost-effective effluent treatment and resource recovery technology for reduction of toxicants.
This paper describes use of the TIE approach to investigate a recent fish kill incident caused by
effluent discharged from the Coal Gasification Plant in Nanjing, China. Daphnia magna were
used to identify the major toxicants causing the whole effluent toxicity in this study.
MATERIALS AND METHODS
Effluent Samples
Three grab samples were collected on different dates from the Coal Gasification Plant
in Nanjing, China and shipped in polyethylene or glass containers to the laboratory at Nanjing
University. Samples were stored in the shipping containers at approximately 4 C during the
course of TIE analyses. Upon sample arrival at the laboratory, dissolved oxygen (DO), chemical
oxygen demand (COD), suspended solids (SS), pH, conductivity, ammonia, and temperature
were measured.
Control/Dilution Water
Control/dilution water used for all tests was tap water which was dechlorinated with
activated carbon. The general characteristics of the water were: pH 7.2-8.5, DO 6.9-8.0 mg/L,
CODMnl-8 mg/L, temperature 20 ± 2 C, conductivity 294 usem/cm, and hardness 129 mg/L
as CaCO3.
Toxicity Tests and Statistical Analyses
The toxicity tests were performed as described in detail elsewhere (U.S. EPA 1991).
Static 24-hour and/or 48-hour acute toxicity tests were conducted with D. magna, no more than
24-hours old, in a test volume of 10 mL at 20 ± 2 C. The test end point was death. To establish
LCso values, all tests were set up using 50% serial dilutions. The LC50 values for toxicity tests
were calculated with the Trimmed Spearman-Karber (TSK) Program Version 1.5 (U.S. EPA
1990). For certain comparisons, effluent toxicity was converted to toxic units (TU) by dividing
100% by the LCso of the effluent. For a specific chemical, TU was calculated by dividing the
chemical's concentration in the sample by its LCso to D. magna which was cited from relevant
literature or determined from the laboratory toxicity test.
Phase I Toxicity Characterization Procedures
Various manipulations and toxicity tests aimed at characterizing classes of potential
toxicants in effluent were conducted with effluent Sample 1 (U.S. EPA 1991). These
manipulation tests included initial toxicity test, baseline test, pH adjustment test, pH
adjustment/aeration test, pH adjustment/filtration test, pH adjustment/Ci8 solid-phase
32
-------
extraction (SPE) test, EDTA chelation test, oxidant reduction test, and graduated pH test. By
observing the alteration of the effluent toxicity via each manipulation, the classes of chemicals
causing the effluent toxicity could be realized.
Phase n Toxicity Identification Procedures
Phase II procedures were conducted with effluent Sample 1 to isolate and identify
specific suspect toxicants within the classes of compounds characterized in Phase I (U.S.EPA,
1993a). A zeolite test was performed to determine if ammonia was a toxicant in the effluent.
Concentrations of total ammonia in original and post-zeolite effluent samples were measured
with a titration method (China NEPA 1989) at 19 C, and then were converted into their
concentrations of un-ionized ammonia (Seager et al. 1988). Toxicity of un-ionized ammonia
was tested at the same temperature.
Fractionation and toxicity testing procedures were employed to identify non-polar
organic toxicants. Filtered effluent Sample 1 was passed through a preconditioned C18 SPE
column, and two post-column samples were collected in sequence and tested for toxicity.
The column was then sequentially eluted with 25, 50,75, 80, 85,90,95 and 100% methanol/
water solutions. Eight SPE fractions were obtained relative to the effluent sample and each
fraction was tested separately for toxicity at 2x the original sample concentration. The toxic
fraction was then concentrated into 100% methanol using another C18 SPE column, and the
concentrate was tested for toxicity at 4x the original sample concentration and analyzed by gas
chromatography-mass spectroscopy (GC-MS). Following the GC-MS analysis, suspect non-
polar organic toxicants were identified by comparing the concentrations of compounds in the
whole effluent with their corresponding LC50 cited from relevant literature or determined from
the laboratory test.
Concentration of total volatile phenols was measured by the 4-aminoantipyrine
calorimetric method (China NEPA 1989) to ensure the identification of non-polar organic
toxicants.
Phase in Toxicity Confirmation Procedures
After the toxicity tests on individual SPE fractions were completed in Phase II, mass
balance tests consisting of a "toxic-fraction test", a "non-toxic-fraction test", and an "all-fraction
test" were initiated to determine whether the non-polar organic toxicants identified in the SPE
fractions accounted for the toxicity removed from the original effluent sample by the Qg SPE
column (U.S. EPA 1993b).
Concentrations of ammonia and total volatile phenols in effluent Samples 2 and 3 were
measured with the same methods as Sample 1. Comparison between the toxicity and
concentrations of ammonia and total volatile phenols in three samples was used to confirm
further if ammonia and volatile phenols were responsible for the toxicity of the coal gasification
effluent.
33
-------
RESULTS AND DISCUSSION
General Toxicity Characteristics
Results of Phase I tests on effluent Sample 1 are summarized in Table 1. Baseline tests
were conducted on succeeding days to provide a baseline toxicily value for comparing the
toxicity of manipulated effluent on that same day. The baseline tests revealed that the effluent
toxicity decreased day by day to a certain extent, and therefore some toxicants which existed in
the effluent were easily degraded. EDTA chelation and sodium thiosulfate oxidant reduction
manipulations had no effect on the toxicity of effluent, and so the cationic metals and oxidants
were not the major toxicants in the effluent.
Table 1. Phase I test results for effluent Sample 1.
LCSO (95% C.I.) (mg/L)
Test
PH
24-hour
48-hour
Initial
8.63
13.4 (7.29-24.6)
<12.5
Baseline
EDTA chelation
Oxidant reduction
pH adjustment
pH adjustment
Baseline
pH adjustment/filtration
pH adjustment/filtration
pH adjustment/filtration
pH adjustment/aeration
pH adjustment/aeration
pH adjustment/aeration
Baseline
Graduated pH
Graduated pH
Graduated pH
Baseline
pH adjustment/filtration
pH adjustment/filtration
pH adjustment/ Cig SPE 1*
pH adjustment/ Ci8 SPE 2*
pH adjustment/ CIS SPE l[
pH adjustment/ C]8 SPE 2*
8.63
8.63
8.63
3
11
8.63
3
8.63
11
3
8.63
11
8.63
6.5
7.5
8.5
8.63
3
8.63
3
3
8.63
8.63
14.3(11.0-18.6)
No effect
No effect
17.9(13.2-23.7)
18.9 (14.0-25.3)
15.4(12.9-18.3)
20.3 (15.9-26.0)
17.7 (NC)
20.7 (16.0-27.0)
18.9 (14.0-25.3)
18.6 (13.9-25.0)
31.2(22.5-43.4)
17.7 (NC)
26.8 (19.8-36.3)
21.0(16.1-27.1)
17.7 (NC)
17.7 (NC)
14.9 (8.24-26.8)
17.7(6.04-51.7)
70.7 ( NC )
59.5 (NC)
70.7 (NC)
29.7 ( NC )
14.3(11.0-18.6)
No effect
No effect
8.84 (NC)
17.0 (NC)
15.4(12.9-18.3)
14.0 (NC)
17.7 (NC)
17.7 (NC)
18.9(14.0-25.3)
18.6(13.9-25.0)
18.9(14.0-25.3)
15.6(12.4-19.7)
Test not done
Test not done
Test not done
15.6(12.4-19.7)
14.0 (NC)
<12.5
26.8 ( NC )
<12.5
70.7 ( NC )
29.7 (NC)
NC » 95% C.I. not calculable due to lack of partial mortality.
*Post-Cig column samples: No. 1 taken after passage of 25 mL, No. 2 after 100 mL; total passage 150 mL.
34
-------
Adjustment of pH and filtration manipulations at pH 3, pH 11, and/or pH i did not
markedly reduce the toxicity of the effluent. Aeration at pH 11 partly reduced the effluent
toxicity, which indicated the possible presence of volatile and/or oxidizable chemicals in
the effluent.
In the graduated pH tests, effluent toxicity at pH 8.5 was greater than that at pH 7.5, and
the toxicity at pH 7.5 was greater than that at pH 6.5. These test results suggested that ammonia,
which is more toxic at higher pH values, or some metals, might have contributed to the effluent
toxicity. Combined with the results of baseline, EDTA and aeration tests, ammonia was a
suspect toxicant in the effluent.
In the pH adjustment/Ci8 SPE tests, post-column effluent samples were significantly less
toxic than the pre-column filtrated effluent samples at pH 3 and pH i. This result suggested that
non-polar organic toxicants were present in the effluent.
The Phase I test results indicated that ammonia and non-polar organic compounds were
possibly key toxicants causing the effluent toxicity.
Toxicity Identification
The results of toxicity tests performed in Phase II are reported in Table 2, and the results
of determining the concentrations of ammonia and total volatile phenols in effluent Sample 1 are
reported in Table 3. Zeolite manipulation removed 0.6 TU from the original effluent sample
(Table 2). The analyses of ammonia in the original effluent and post-zeolite effluent indicated
that zeolite removed 7.4 mg NH3-N/L un-ionized ammonia (67% of total ammonia) from the
original effluent (Table 3), which had been responsible for 1.12 TU based on its LCs0 (Table 2).
The removal of ammonia could interpret the decrease of whole effluent toxicity, and therefore
ammonia was identified as one of the toxicants causing the effluent toxicity.
Table 2. Phase II toxicity test results for effluent Sample 1.
LC50 (95% C.I.) (mg/L)
Toxicity test
24-hour TU
24-hour
48-hour
Original effluent
Post-zeolite effluent
Cig SPE la
Cig SPE 2a
50% SPE fraction
50% SPE fraction concentrate
Un-ionized ammonia
22.8 (16.9-30.8)
26.5(13.1-53.5)
70.7 (NC)
29.7 (NC)
35.4 (NC)
159(129-196)
6.59mg/LNH3-N
18.4(15.7-21.5)
22.3 (12.4-40.0)
70.7 (NC)
29.7 (NC)
ND
141 ( NC )
4.38
3.78
1.41
3.36
2.83
0.63
NC = 95% C.I. not calculable due to lack of partial mortality.
aPost-Cig column samples: No. 1 taken after passage of 25ml, No. 2 after 150 ml; total passage 200 ml.
35
-------
Table 3. Concentrations of ammonia and volatile phenols in effluent Sample 1.
Sample
Original sample
Post-zeolite sample
Total ammonia
(mg/L NH3-N)
81.7
27.0
Un-ionized ammonia
(mg/L NH3-N)
11.1 '
3.7
Total volatile phenols
(mg/L)
34.4
From Table 2, the No. 1 post-C18 column sample was 2.97 TU less than the original
effluent sample. The toxicity of the No. 2 post-Cjg column sample only had a slight decrease
from the No. 1 post Ci8 column sample, which demonstrated that the Qg column was overloaded
and the toxicants might have broken through (U.S. EPA 1993b). Only the 50% SPE fraction
among eight fractions was toxic and it measured at 2.83 TU. The 50% SPE fraction concentrate
was 0.63 TU, which was lower than the toxicity of the 50% SPE fraction. This is characteristic
of this concentration step and the absolute recovery was not as important as the fact that toxicity
was retained in the concentrate that was analyzed via GC-MS (Norberg-King et al. 1991).
The 50% fraction concentrate was analyzed by GC-MS, and the results of chemical
analysis and toxicity of each constitute in the concentrate are dejpicted in Table 4. There were
ten volatile phenols, n-phenylisopropylamine, benzeneacetonitrile, 4-methylpyridine, and two
dimethylbenzaldehydes existing in the effluent.
Table 4. Constituents analyzed by GC-MS in the 50% fraction of effluent Sample 1
and their toxicity values.
Constituent
2,4-dimethylphenol
3-ethylphenol
3-methylphenol
2-methylphenol
2,6-dimethylphenol
2-ethylphenol
3-ethyl-5-methylphenol
3,4-dimethylphenol
2-ethyl-6-methylphenol
N-phenylisopropylamine
Benzeneacetonitrile
4-methylpyridine
2,5-dimethyihenzaldehyde
3,4-dimethylbenzaldehyde
Phenol
Total Phenols
Concentration
(mg/L)
11.0
7.45
11.6
7.09
2.08
1.22
0.772
0.899
0.228
0.098
0.071
0.091
0.105
0.177
0.074
42.4
Type of
test
24-hour LC50
48-hour EC50
24-hour LC50
24-hour LC5o
24-hour LC50
48-hour EC50
48-hour EC50
24-hour LC50
48-hour EC50
24-hour LC50
MATC
48-hour LC50
24-hour LC50
24-hour LC50
24-hour LC50
(mg/L)
7.01
10.0
19.2
22.6
14.0
10.0
10.0
19.7
10.0
8.00
5.00
30.0
50.0
50.0
64.4
Reference
1
2
3
1
3
2
2
1
2
4
4
5
4
4
1
TU
1.57
0.75
0.60
0.31
0.15
0.12
0.08
0.05
0.02
0.01
0.01
O.01
<0.01
<0.01
O.01
3.65
'Laboratory test; 2Sheedy elal. 1991; 3Devillers 1988; 4Grushko 1982; 5Phippsefa/. 1984.
36
-------
Assuming the toxicity of all these 15 compounds were additive, the estimated total
toxicity of all the detected volatile phenols was 3.65 TU, which might account for the observed
2.83 TU of the 50% SPE fraction. Therefore, volatile phenols were identified as suspect
non-polar organic toxicants in the effluent. Six phenols, i.e., 2,4-dimethylphenol, 3-ethylphenol,
3-methylphenol, 2-methylphenol, 2,6-dimethylphenol, and 2-ethylphenol were the major
toxicants with more than 0.1 TU among the ten phenols. In addition, the concentration of total
phenols in the effluent was 42.4 mg/L which was close to the 34.4 mg/L measured by the
4-aminoantipyrine calorimetric method (Table 3). This result supported that volatile phenols
were responsible for the effluent toxicity.
Toxicity Confirmation
Results of mass balance tests are reported in Table 5. The toxic fraction, non-toxic
fraction, and all fractions were 2.83,0, and 3.75 TU, respectively. The toxicity removed by
the Cig SPE column from the original sample (2.97 TU, see Table 2) was recovered in the
toxic-fraction and all-fraction. This result confirmed that the suspect toxicants, volatile phenols,
identified from the 50% toxic fraction were the real toxicants causing the effluent toxicity.
Table 5. Mass balance test results for effluent Sample 1.
Toxicity test
24-hour LC50(mg/L)
(95% C.I.)
24-hour TU
Toxic-fraction
Non-toxic-fraction
All-fraction
35.4 (NC)
70.7 ( NC )
17.7(13.0-23.7)
2.83
0
3.75
NC = 95% C.I. not calculable due to lack of partial mortality. 48-hour LC50 not conducted.
The toxicity and concentrations of volatile phenols and ammonia in three samples of the
effluent are reported in Table 6. All the three samples collected at different dates were toxic to
D. magna and both ammonia and total volatile phenols in these samples were present at toxic
concentrations. These results further confirmed that ammonia and volatile phenols were the
major chemicals responsible for the toxicity of whole effluent discharged from the Coal
Gasification Plant.
Table 6. Toxicity and concentrations of volatile phenols and ammonia in three samples
of the effluent.
Sample
1
2
3
pH
8.63
9.12
9.30
24-hour TU
4.38
5.66
19.2
Volatile phenols
(mg/L)
34.4
22.1
56.7
Total ammonia
(mg/L NH3-N)
81.7
74.6
254
Un-ionized ammonia
(mg/L NH3-N)
11.1
23.7
108
37
-------
ACKNOWLEDGEMENT
This work was jointly supported by the Jiangsu Environmental Protection Bureau (Grant
No. 9726) and the Jiangsu Science and Technology Commission (Grant No. BK 9704) in China.
The authors also thank personnel at the Nanjing Municipal Environmental Monitoring Center for
assistance.
REFERENCES
Ankley, G.T., and L.P. Burkhard. 1992. Identification of surfactants as toxicants in primary
effluent. Environmental Toxicology and Chemistry 11:1235-1248.
Burkhard, L.P., and G.T. Ankley. 1989. Identifying toxicants: NETA's toxicity-based
approach. Environmental Science and Technology 23:143 8-1443.
China NEPA (National Environmental Protection Agency). 1989. Analytical Methods for the
Monitoring of Water and Wastewater, Third Edition. China Environmental Science Press.
Devillers, J. 1988. Acute toxicity of cresols, xylenols, and trimethylphenols to Daphnia magna
Straus 1820. The Science of the Total Environment 76:79:83.
Grushko, Ya.M. 1982. Manual of Toxic Organic Chemicals in Industrial Wastewaters. In
English Translation by Y.L. Geng, Hydrocarbon Processing Press, PRC.
Jin, H.J., X. Lou, Z.H. Zhang, and G.X. Wang. 1994. Ecotoxicological monitoring of major
industrial effluents in Nanjing, China. In: Proceedings of the Third International Symposium
of Fish Physiology, Toxicology, and Water Quality. Nanjing, PRC., November 1992,
D.J. Randall (Ed.). U.S. EPA/600/R-94/138. Office of Research and Development, U.S.
Environmental Protection Agency, Athens, Georgia, USA.
Norberg-King, T.J., E.J. Durham, and G.T. Ankley. 1991. Application of toxicity identification
evaluation procedures to the ambient waters of the Colusa Basin drain, California.
Environmental Toxicology and Chemistry 10:891-900.
Phipps, G.L., M.J. Harden, E.N. Leonard, T.H. Rousch, D.L. Spehar, C.B. Stephan,
Q.H. Pickering, and A.L. Buikema Jr. 1984. Effects of pollution on freshwater organisms.
Journal of the Water Pollution Control Federation 56:725-758.
Seager, J., E.W. Wolff, and V.A. Cooper. 1988. Proposed Environmental Quality Standards for
List Il'substances in Water, Ammonia. TR260, Water Research Centre, Marlow, U.K.
Sheedy, B.R., J.M. Lazorchak, and DJ. Grunwald. 1991. Effects of pollution on freshwater
organisms. Research Journal of the Water Pollution Control Federation 63:619-696.
38
-------
U.S. EPA (U.S. Environmental Protection Agency). 1990. Trimmed Spearman-Karber (TSK)
Program Version 1.5. Ecological Monitoring Research Division, Environmental Monitoring
Systems Laboratory, U.S. Environmental Protection Agency, Cincinnati, Ohio, USA.
U.S. EPA (U.S. Environmental Protection Agency). 1991. Methods for Aquatic Toxicity
Identification Evaluations: Phase I Toxicity Characterization Procedures. U.S. EPA
600/6-91/003. Environmental Research Laboratory-Durum, Minnesota, USA.
U.S. EPA (U.S. Environmental Protection Agency). 1993a. Methods for Aquatic Toxicity
Identification Evaluations: Phase II Toxicity Identification Procedures for Samples
Exhibiting Acute and Chronic Toxicity. U.S. EPA 600/R-92/080. Environmental Research
Laboratory-Duluth, Minnesota, USA.
U.S. EPA (U.S. Environmental Protection Agency). 1993b. Methods for Aquatic Toxicity
Identification Evaluations: Phase III Toxicity Confirmation Procedures for Samples
Exhibiting Acute and Chronic Toxicity. U.S. EPA 600/R-92/081. Environmental Research
Laboratory-Duluth, Minnesota, USA.
39
-------
-------
FISHES: AMMONIA PRODUCTION, EXCRETION AND TOXICITY
David J. Randalll, and Beverley J. Wicks'
ABSTRACT
Ammonia is toxic to fishes as well as other vertebrates. Most fishes continually produce
and excrete ammonia by diffusion of NH3(g> across the gills. Fishes avoid ammonia toxicity by
utilizing several physiological mechanisms. Under conditions of elevated ambient ammonia, the
mudskipper, Periophthalmus schlosseri, can continue to excrete ammonia by active transport of
ammonium ions. This has been shown only in this single species of teleost fish. In order to
avoid toxic levels of ammonia in their body, some fishes reduce their own production of
ammonia, whereas others convert excess to less toxic compounds including glutamine and other
amino acids for storage. A few species have active ornithine urea cycles and convert ammonia
to urea for both storage and excretion. The Lake Magadi tilapia can convert ammonia to urea via
the ornithine urea cycle at the same rate as a rat (Rattus sp.) and, unlike most other fishes, is able
to survive in very alkaline waters. Most teleost fishes, however, are ammonotelic and are very
susceptible to elevated water ammonia levels. Temperature has only minor effects on ammonia
toxicity, ionic strength of the water can influence ammonia toxicity, and pH has a very marked
effect on toxicity. Acid waters ameliorate whereas alkaline waters exacerbate ammonia toxicity.
Fishes have difficulty surviving even in ammonia free alkaline waters because of impaired
ammonia excretion under alkaline conditions. Toxicity studies are usually carried out on unfed
and resting fish in order to facilitate comparison of results. Based on recent studies, however,
environmental stresses, including swimming, can have dramatic effects on ammonia toxicity.
It is also clear that feeding will result in elevated post-prandial body ammonia levels. Thus,
feeding probably will also exacerbate ammonia toxicity. Fishes may be more susceptible to
elevated ammonia levels during and following feeding or when swimming. Thus, present
ammonia criteria may fail to protect migrating salmonids and may be inappropriate for fishes
that feed on a regular basis.
AMMONIA PRODUCTION AND EXCRETION
The two main waste products of metabolism are COa and NHs, with ammonia production
being between one tenth to one third the rate of carbon dioxide production. The main internal
source of ammonia in fishes is through catabolism of proteins. Ingested proteins are hydrolyzed
producing ammonia which is available for use in metabolism or, if in excess, removal.
Ammonia can become toxic in fishes if accumulation occurs, making excretion or conversion
to less toxic compounds critical.
'University of British Columbia, Zoology Department, Vancouver, British Columbia, Canada
41
-------
In aqueous solution total ammonia has two species, NH3 and
NH3 + H+ <-»• NH4+
The pK of this ammonia/ammonium ion reaction is around 9.5. The NH3/NH4+
equilibrium both internally and in ambient water depends on temperature, ionic strength and pH,
of these pH is of greatest significance to the fish. The gill membranes are permeable to NH3,
but not NHt"1". Ammonia is excreted across the gills by diffusion of NH3 in most fishes, although
there may be some carrier mediated excretion of NtV" in some fish species. The rate of NH3
excretion is determined by the magnitude of the NH3 gradient between blood and water.
Ammonia excretion is augmented by acidic conditions in the water because any NH3 excreted
into the water is rapidly converted and trapped as NH4+, maintaining the NH3 gradient across the
gills and, therefore, augmenting ammonia excretion. Many freshwater fishes actively excrete
protons forming an acid boundary layer next to the gill surface (Lin and Randall 1991) and this
augments ammonia excretion (Wright et al. 1989). Ammonia excretion in water above pH 9 is
reduced because of the absence of trapping of ammonium ions (Ibid) resulting in elevated plasma
ammonia levels (Yesaki and Iwama 1992). Thus, many fishes have difficulty excreting
ammonia when exposed to alkaline conditions.
AMMONIA ACCUMULATION AND TOXICITY IN THE BODY
Accumulation of ammonia in the body can be due either to inability to excrete or convert
nitrogenous wastes or to a net influx of NH3 from the environment. Externally, the concentration
of NHs, rather than NlV", is of concern because NH3 diffuses readily through the fish gills. As a
result, if NH3 levels are high in the environment, ammonia enters the fish as NH3, but is then
converted to the more toxic form, NlV". In acid water most of the ammonia is as NH4+ and the
rate of ammonia entry into the fish is low. As pH increases to more alkaline conditions and
water pH approaches the pK of the ammonia/ammonium ion reaction, toxicity increases
significantly for many species due to the shift in equilibrium to the more diffusable NH3 form.
Water of pH above 9.5 can be toxic, even though it contains little or no ammonia, because
ammonia levels rise to toxic levels in the fish as a result of impaired excretion.
Elevated body ammonia levels are toxic and have both acute and chronic impacts,
which vary with the species. At non-lethal levels elevated ammonia can interfere with energy
metabolism through impairment of the TCA cycle. The decrease is due to inhibition of some key
enzymes including isocitrate dehydrogenase, a-ketoglutarate dehydrogenase, and pyruvate
dehydrogenase. It has also been suggested that ammonia affects the ionic balance in fishes,
reducing Na^infiux and K+ loss through substitution of NtV" for K+ in NaV K+-ATPase and/or
Na+/K+/2Cl" co-transport. NH3 is lipid soluble and can cross the mammalian blood-brain barrier
which separates the brain and cerebrospinal fluid from the blood (Stabenau et al. 1958).
42
-------
High ammonia levels modify many aspects of the blood brain barrier in mammals (reviewed by
Breen and Schenker 1972, Cooper and Plum 1987). In addition, NH4+ can substitute for K
affecting membrane potential and it can also impair glutamate metabolism. As a result, ammonia
acts on the central nervous system of vertebrates, including fishes, causing hyperventilation,
hyper-excitability, coma, convulsions, and finally death.
ENVIRONMENTAL CONDITIONS AND AMMONIA TOXICITY
It might be expected that temperature, ionic strength and pH would effect ammonia
toxicity because of their action on the NH3/NH4+ equilibrium. Compiled normalized data on
acute toxicity in various species of fishes, expressed as total ammonia ([NH3] + [NH4 ])
indicates, however, that the effect of increased temperature on toxicity is minor between
3 and 30°C (U.S. EPA 1998). Experimental evidence, however, does indicate that the hardness
of ambient water can influence ammonia toxicity (Tomasso et al 1980, Soderberg and Meade
1992). Recent acute flow-through toxicity tests conducted on fry of rainbow trout
(Onchorynchus my kiss) suggest that addition of calcium to water with pH ranging from 7.8 to 9.0
decreased acute toxicity of ammonia (Figure 1, Tang and Randall, Zoology Department,
University of British Columbia, unpublished observations). Calcium addition had no effect at
pH 6.0. Elevated calcium levels activate apical membrane proton ATPase in the gills (Lin and
Randall 1995) and reduce gill permeability to ions. The former could increase the acid boundary
layer around the gills and enhance ammonia excretion. In addition, elevated calcium levels
appear to reduce the cortisol response to stressful situations (Yesaki and Iwama 1992). Cortisol
stimulates protein catabolism, thus reduced cortisol release will reduce ammonia production by
the fish and this may be the major way in which elevated calcium ameliorates ammonia toxicity
(Wilson etal. 1998).
CHRONIC AND ACUTE TOXICITY VALUES
Chronic toxicity of total ammonia to fishes based on 30-day exposures and based on
weight loss, total biomass, and hatchability, indicate that chronic effects can be observed at one
half to one twentieth of the acute toxicity level, the ratio of chronic to acute levels varying with
the species. Chronic toxicity is very pH dependent, increasing with pH in much the same way
as acute toxicity. Thus the U.S. Environmental Protection Agency's Criterion Continuous
Concentration (CCC), the threshold value resulting in unacceptable effects (more than a
20% reduction in survival, growth and/or reproduction), and the Criterion Maximum
Concentrations (CMC), defined as half of the final value derived from acute toxicity tests are
both pH-dependent, approaching zero environmental ammonia levels above pH 9.5. At pH 7
the revised CCC recommended by the U.S. EPA (1998) was 3.48 mg N/L total ammonia,
whereas the new CMC was 48.8 mgN/L total ammonia. Tests conducted in an experimental
stream (Zisehke and Arthur 1987) indicated little change in biomass for numerous test species
unless the 4-day mean ammonia level exceeded the CCC. Therefore, based on the available data,
the U.S. EPA standards for ammonia levels in fresh water appear to be appropriate.
43
-------
i
o
v>
U
1.
0 -
50
45
40
35
30
25.
20 .
15-
10.
160
150
140 -
130 -
120 -
110 -
100 -
90
B
012345
Calcium addition (mmol/L)
Figure 1. Effect of total ammonia (mg/L N) on rainbow trout at pH 6.5, 7.8, and 9.0 and with
calcium addition of 0, 2 and 5 mmol/L. (The 96-hour LC50 estimates are calculated
according to the trimmed Spearman-Karber method, Hamilton et al. 1977. Significant
differences in the 95% confidence intervals are indicated by a change in letters.)
44
-------
PROBLEMS WITH PRESENT AMMONIA TOXICITY CRITERIA
Methods designated by the U.S. EPA for toxicity studies, however, follow standard
guidelines that require toxicity tests to be conducted under static conditions on starved fish.
These conditions are rare in the natural environment. Recent studies indicate that coho salmon
(Oncorhynchus kisutch) show a linear relationship between critical swimming velocity (Uorit),
a measure of swimming performance, and ambient ammonia concentration. According to
Joensen, Tang and Randall, Zoology Department, University of British Columbia (Unpublished
data), swimming performance in coho salmon decreased with increasing concentrations of
ammonia in the water. There was a linear increase in plasma ammonia with increasing ambient
water ammonia concentrations, again, Ucrit decreased with increasing plasma ammonia levels.
However, there was no relationship found between Ucrit and tissue ammonia levels in fish
exposed to a range of ambient ammonia concentrations. These swimming fish not only had a net
influx of ammonia from the water, but also had to cope with ammonia accumulation in muscle
due to deamination of adenylates during exercise (Mommsen and Hochachka 1988). Since the
increase in environmental ammonia results in increased plasma ammonia and this decreases the
swimming ability, it is probable that elevated ammonia could reduce swimming performance and
decrease the survival of migrating salmon. Beaumont et al. (1999) has shown that elevated
ammonia levels in the body result in muscle depolarization, and this will in turn reduce
swimming performance. Increased acute ammonia toxicity when swimming was evident in a set
of experiments testing the effect of total ambient ammonia on the mortality of rainbow trout
under resting and exercise conditions (60% Uctit) (Tang and Randall, Zoology Department,
University of British Columbia, Unpublished observations). The calculated LCso for the resting
fish under the conditions of these experiments, that is in the swimming respirometer but not
swimming, was approximately 207 mg N/L total ammonia (Table 1 and Figure 2). LCso values
were calculated using the trimmed Spearman-Karber method (Hamilton et al. 1977). This value
is well above the U.S. EPA ammonia acute toxicity value of 48 mg N/L total ammonia at pH 7.0.
The LCso for swimming fish was below the U.S. EPA's acute value at 32 mg N/L total ammonia.
Thus the present U.S. EPA ammonia standards will not protect migrating salmon.
Table 1. Final acute values for ammonia toxicity (mg/L N) to various fishes at pH 7.0
Species tested and conditions
Final acute value for ammonia (mg/L of N)
*U.S EPA value for salmonids
'U.S. EPA value for non-salmonids
2Resting rainbow trout
2Swimming rainbow trout
48
74
207
32
'From U.S. EPA 1998.
2From Tang and Randall, Zoology Department, University of British Columbia, Unpublished data
45
-------
I
120
110 -
100 -
90 -
80 -
70-
60 -
50 -
40 -
30 -
20 -
10 -
0 -
-10 -
-20
96 h LC50
resting= 207 mg/L N
swimming= 32 mg/L N
100 200
Total Ammonia (mg/L N)
300
400
Figure 2. Effect of total ammonia (mg/L N) on rainbow trout mortality under resting conditions
(circles) and exercise conditions (60% Ucrit.) (triangles). (The 96-hour LCso estimates
calculated according to the the trimmed Spearman-Karber method, Hamilton et al. 1977.)
Ammonia is an unusual toxicant in that it is a product of metabolism as well as being a
toxicant present in the environment. Standard 96-hour LCso acute tests for aquatic toxicants,
using starved unstressed fish, are not appropriate in this case. Both feeding and stress, such as
exercise, result in increased ammonia production by the fish and these must be taken into
account when promulgating criteria for ammonia. The extent of increase in ammonia production
following feeding will undoubtedly be related to the protein content of the food. Thus diets
could be designed to reduce the magnitude of the post-prandial ammonia pulse and, therefore,
reduce the impact of feeding on ammonia toxicity. Other stresses, such as crowding or rapid
shifts in water temperature, could also exacerbate ammonia toxicity due to increased ammonia
production by the fish following cortisol release.
PHYSIOLOGICAL MECHANISMS OF AMMONIA TOLERANCE
Since nitrogen excretion is critical and niches do exist where it is difficult to excrete
s, or where ambient ammonia levels are elevated, some species of fishes have adapted
physiologically to allow them to survive under these conditions. Some fishes convert ammonia
to less toxic compounds in an attempt to reduce ammonia levels and avoid toxic accumulation.
46
-------
Increased internal NH4+ concentrations increase glutamine synthetase activity, which converts
NHU"1" to glutamine, a less toxic compound. Glutamine can be stored until ammonia levels in the
ambient environment decrease. The mudskipper, Periophthalmus schlosseri, for example,
converts ammonia to free amino acids, including glutamine, and stores them in the brain, muscle
and liver, when exposed to high environmental ammonia levels (Peng et al. 1998). Exposure of
rainbow trout to water of pH 10 did not result in any conversion of accumulated ammonia to urea
or glutamine, but there was a reduction in ammoniagenesis (Wilson et al. 1998). Presumably
this reduced ammonia production ameliorated the rise in body ammonia levels in the face of
decreased excretion rates associated with exposure to alkaline conditions.
A single species of tilapia, Oreochromis alcalicus grahami lives in Lake Magadi, Kenya,
which has a pH of around 10. Lake Magadi tilapia can tolerate very high levels of ammonia in
their environment (Walsh et al. 1993). Another closely related tilapia, O. nilotica, dies in less
than 60 minutes if placed in Lake Magadi waters. The reason for the survival of O. alcalicus
grahmi in Lake Magadi is evident when ammonia excretion is compared in these species.
O. alcalicus grahami excretes urea and little or no ammonia, hence nitrogen excretion is not
impaired by high environmental pH, whereas O. nilotica, excretes ammonia. O. alcalicus
grahami has the enzymes of the ornithine-urea cycle (OUC), namely carbamoyl phosphate
synthetase I & II, and carbamyl transferase. Several of these enzymes were not detectable in
O. nilotica, thus this species does not have a functional ornithine urea cycle (Randall et al. 1989).
The presence of an OUC allows the Lake Magadi tilapia to tolerate high ambient ammonia levels
by converting any ammonia entering the fish to urea, and then excreting the urea produced.
This fish has the same capacity to produce urea as a rat.
There are two pathways that produce urea in fish. The uricolytic pathway produces
urea from purines and the ornithine urea cycle that converts ammonia and bicarbonate to urea.
Most teleosts produce some urea via uricolysis. The OUC is found in elasmobranchs, the Lake
Magdi tilapia, some fish embryos, the marine toadfish (Opsanus beta), some airbreathing fish
(Heteropneutus fossilis) and the largemouth bass (Micropterus salmoides) (Wright 1995).
This cycle is uncommon in other teleosts and aquatic invertebrates. They simply excrete
ammonia directly into the water and avoid converting it to urea. It costs energy to convert
ammonia to urea, thus if water is in abundance, ammonia rather than the more costly urea is
the choice for excretion.
Most ammonia excretion in ammontelic organisms occurs by diffusion of NHs. It is
possible, however, that some excretion may be coupled to either Na+/K+(NH4+)/2CF cotransport,
Na+/H+CNH4+) exchange, Na+/K+(NH4+) ATPase, and/or K+/NH4+ exchange (Amlal and
Soleimani 1997). Periophthalamoden schlosseri, a mudskipper found in Mangrove swamps in
Singapore and Malaysia, can tolerate high ambient ammonia levels and shows no conversion of
nitrogenous wastes to urea. However, it is able to actively excrete this excess ammonia across
the gills against large gradients (Ip et al. 1999), probably using Na+/NH4+ ATPase on the basal
lateral membrane and Na+/NH4+ antiporters on the apical membrane of the gills (Randall et al.
1999). This active transport allows the mudskipper to tolerate high ambient ammonia.
Mudskippers in general convert ammonia to glutamine and store amino acid during periods of
ammonia accumulation, reducing ammonia toxicity. To what extent other mudskippers can
actively excrete ammonium ions has yet to be determined.
47
-------
REFERENCES
Amlal, H., and M. Soleimani. 1997. K+/NH4 + antiporter: a unique ammonium carrying
transporter in the kidney inner medulla. Biochimica et Biophysica Acta 1323:319-333.
Beaumont, M.W., P.J. Butler, and E.W. Taylor, 1999. Tissue ammonia levels and swimming
performance of brown trout exposed to copper in soft, acidic water. In: Proceedings of
the Fifth International Symposium on Fish Physiology, Toxicology and Water Quality,
November 1998, City University of Hong Kong. R.V. Thurston (Ed.). In Press.
Breen, K.J., and S. Schenker. 1972. Hepatic coma: present concepts of pathogenesis and
therapy. Progress in Liver Disease 4:301-332.
Cooper, J.L., and F. Plum. 1987. Biochemistry and physiology of brain ammonia.
Physiological Reviews 67:440-519.
Hamilton, M.A., R.C. Russo, and R.V. Thurston. 1977. Trimmed Spearman-Karber method
for estimating median lethal concentrations in toxicity bioassays. Environmental Science
& Technology 11:714-719. Correction: Environmental Science & Technology 12:417
Ip, Y.K., K. W. Peng, S.F. Chew, W.K. Kok, J. Wilson, and D. Randall. 1999. The
mudskippers: ammonia toxicity and tolerance. In: Proceedings of the Fifth International
Symposium on Fish Physiology, Toxicology and Water Quality, November 1998, City
University of Hong Kong. R.V. Thurston (Ed.). In Press.
Lin, H., and D.J. Randall. 1991. Evidence for the presence of an eletrogenic proton pump on the
trout gill epithelium. Journal of Experimental Biology 161:119-134.
Lin, H., and D.J. Randall. 1995. Proton pumps in fish gills. Cellular and molecular approaches
to fish ionic regulation. In: Fish Physiology, Volume XIV. E.C.M. Wood, and T.J.
Shuttleworth (Eds.). Academic Press, Inc., Orlando, Florida, USA. pp. 229-255.
Mommsen, T.P., and P.W. Hochachka. 1988. The purine nucleotide cycle as two temporally
separated metabolic units: a study on trout muscle. Metabolism: Clinical and
Experimental 37:552-556. ,-
Peng, K.W., S.F. Chew, C.B. Lim, S.S.L. Kuah, W.K. Kok, and Y.K. Ip. 1998. The
mudskippers Periophthalmodon schlosseri and Boleophthalmus boddaerti can tolerate
environmental NHs concentrations of 446 and 36 |^M, respectively. Fish Physiology
and Biochemistry 19:59-69.
48
-------
Randall, D.J., J.M. Wilson, K.W. Peng, T.W.K. Kok, S.S.L. Kuah, S.F. Chew, TJ. Lam,
and Y.K. Ip. 1999. The mudskipper, Periophthalmodon schlosseri, actively transports
t + against a concentration gradient. American Journal of Physiology. In press.
Randall, D.J., C.M. Wood, S.F. Perry, H. Bergman, G.M. Maloiy, T.P. Mommsen, and
P. A. Wright. 1989. Urea excretion as a strategy for survival in a fish living in a very
alkaline environment. Nature 337:165-166.
Soderberg, R.W., and J.W. Meade. 1992. Effects of sodium and calcium on acute toxicity of
un-ionized ammonia to Atlantic salmon and lake trout. Journal of Applied Aquaculture
1:83-92.
Stabenau, J., K.S. Warren, and D.P. Rail. 1958. The role of pH gradients in the distribution of
ammonia between blood and cerebrospinal fluid, brain, and muscle. Journal of Clinical
Investigations 37:933-934.
Tommaso, J.R., C.A. Goudie, B.A. Simco, and K.B. Davis. 1980. Effects of environmental
pH and calcium on ammonia toxicity in channel catfish. Transactions of the American
Fisheries Society 109:229-234.
U.S. EPA (U.S. Environmental Protection Agency). 1998. Addendum to Ambient Water
Quality Criteria for Ammonia - 1984. National Technical Information Service,
Springfield, Virginia, USA.
Walsh, P.J., H.L. Bergman, A. Narahara, C.M. Wood, P. A. Wright, D.J. Randall, J.N. Maina,
and P. Laurent. 1993. Effects of ammonia on survival, swimming and activities of
enzymes of nitrogen metabolism in the Lake Magadi tilapia Oreochromis alcalicus
grahami. Journal of Experimental Biology 180:323-387.
Wilson, J.M., K. Iwata, O.K. Iwama, and D.J. Randall. 1998. Inhibition of ammonia excretion
and production in rainbow trout during severe alkaline exposure. Comparative
Biochemistry and Physiology 121:99-109.
Wright, P. A. 1995. Nitrogen excretion: three end products, many physiological roles. Journal
of Experimental Biology 198:273-281.
Wright, P. A., D.J. Randall, and S.F. Perry. 1989. Fish gill boundary layer: a site of linkage
between carbon dioxide and ammonia excretion. Journal of Comparative Physiology
158:627-635.
49
-------
Yesaki, T.Y., and G.K. Iwama. 1992. Some effects of water hardness on survival, acid-base
regulation, ion regulation and ammonia excretion in rainbow trout in highly alkaline
water. Physiological Zoology 65:763-787.
Zischke, J.A., and J.W. Arthur. 1987. Effects of elevated ammonia levels on the fingernail
clam, 'Musculium transversuni, in outdoor experimental streams. Archives of
Environmental Contamination 16:225-232.
50
-------
TISSUE AMMONIA LEVELS AND SWIMMING PERFORMANCE
OF BROWN TROUT EXPOSED TO COPPER IN SOFT, ACIDIC WATER
Matthew W. Beaumont1, Patrick J. Butler1, and Edwin W. Taylor1
ABSTRACT
The supply of oxygen to the tissues has often been thought of as the limiting factor in
exercising fish exposed to metals or to low pH, either due to the effects of these pollutants upon
the ability offish to extract oxygen from their environment or to transport oxygen in the
circulatory system. However, in brown trout (Salmo truttd) exposed to a sub-lethal combination
of low pH and copper in soft water, swimming performance is impaired and yet such a limitation
to oxygen supply does not seem to exist. Although both red and white muscle showed some
metabolic disruptions consistent with hypoxia, namely a high lactate concentration at rest and, in
the white muscle, depletion of glycogen and phosphocreatine, a putative role of increased blood
viscosity following haematological changes in reducing the supply of oxygen to the tissues is
unsupported. Haematocrit, haemoglobin and plasma protein concentration were not affected by
copper and low pH exposure, and a lack of further change in variables such as lactate at the
onset of exercise lead us to look for an alternative explanation for the effects of copper and low
pH upon tissue metabolites. Ammonia concentration, both in the plasma and muscles, is
significantly higher in trout exposed to copper and low pH. Ammonia plays a role in the
regulation of a number of metabolic pathways and could account for the altered metabolic status
of these fish. In addition to a role in the regulation of a number of metabolic pathways,
ammonium ions can cause electrophysiological disruptions that might lead to the observed loss
of swimming performance, particularly the displacement of K+ in ion exchange mechanisms.
Using the distribution of ammonia between intracellular and extracellular compartments to
estimate resting muscle membrane potential, a significant depolarisation is predicted in both
red and white muscle offish exposed to copper and low pH. These predictions are supported
by estimates of white muscle membrane potential derived from ion distribution and from direct
measurement using microelectrodes.
COPPER AND ACIDITY IN THE AQUATIC ENVIRONMENT
While Robert Angus Smith recognised acid rain as a component of the polluted air
around Manchester (UK) as early as 1852 (Franks 1983), it has been the media interest
surrounding the decline and death of Scandinavian, Scottish, and North American forests and
destruction offish stocks that has more recently brought the issue to the attention of the public.
There has, of course, also been considerable scientific interest in the topic, and the toxicity of
'School of Biological Sciences, University of Birmingham, Birmingham, United Kingdom
51
-------
acidification to aquatic organisms in particular has been well documented (e.g. Spry et al.
1981, Morris et al. 1989). The modifying effects of other variables such as ambient calcium
level, salinity, and temperature have also been investigated (McDonald 1983, Powell and
McKeown 1986, Butler et al. 1992) providing the insight that has enabled the mechanisms of
acid toxicity to be characterised (Wood 1989). However, although pH is probably the crucial
toxic factor at extremes of acidity, at less exceptional and more environmentally common levels,
aluminium and heavy metals such as copper, zinc, and cadmium are mobilised and contribute
significantly to toxicity (Spry et al. 1981, Wood 1989). Therefore, these too have been well
studied and their toxic mechanisms have, in general, also been characterised. For fishes, lethal
concentrations of these pollutants tend to have as their primary target the gill where they may
disrupt either or both the ionoregulatory and gas exchange functions. Cupric ions, for example,
inhibit branchial transport ATPases, probably by binding to sulphydryl groups (Stagg and
Shuttleworth 1982, Lauren and McDonald 1987) and displace calcium ions which are important
in controlling the integrity and permeability of the branchial epithelium. These two effects, in
freshwater, disrupt ionoregulation, decreasing the ability of the fish to resist ion efflux. At
higher copper concentrations, the branchial disruption becomes more severe leading to
significant ultrastructural histopathologies (Kirk and Lewis 1993, Wilson and Taylor 1993a)
and consequently to inefficiencies in gas exchange (Figure 1).
Binding to -SH groups
of transport ATPases.
Competitive replacement
of Ca2* in tight junctions.
Further replacement of Ca2
from intracellular matrix
Decreased active ion uptake
Increased passive ion efflux
Mucus production (to maintain Ca2
nvicro -environment?)
Epithelial disruption
Ionoregulation
O Respiration
Figure 1. Mechanisms of copper toxicity.
SWIMMING PERFORMANCE
Bearing in mind the respiratory and ionoregulatory consequences of exposure to lethal
concentrations of low pH and heavy metals, a reduction in the swimming performance offish
following exposure to a pollutant is often attributed to disruption of oxygen transfer at the gill
or to a decrease in the oxygen carrying capacity of the blood. In previous studies by the current
authors, the effects of sub lethal doses of both low pH alone (LP) and of low pH and copper
together (CLP) upon brown trout (Salmo truttd) have been investigated (Butler et al. 1992,
Beaumont et al. 1995a). Critical swimming speed (Uotit) (Brett 1964) was found to be lower in
both cases. However, no significant differences in pH of either the plasma or red blood cells
52
-------
were observed following acid exposure (Butler and Day 1993). More importantly, arterial
oxygen content was not lower (Butler et al. 1992). In fact, at 15°C, oxygen content was
significantly higher in LP exposed trout. CLP exposure, caused considerable gill damage,
which, at some copper concentrations, resulted in a two to four times expansion of the
blood/water diffusion barrier (Taylor et al. 1996), but here, too, there was little indication that
oxygen transfer ever became limiting. Indeed, as for acid exposure, the oxygen content of
arterial blood was unchanged or even elevated after exposure to copper at low pH (Beaumont et
al. 1995a).
HAEMOCONCENTRATION / OXYGEN TRANSFER
While neither branchial oxygen transfer nor the apparent oxygen carrying capacity of
the blood seemed to have been limiting, oxygen transport to the tissues may have still been
disrupted. Both acid and copper exposure have a number of effects such as the osmotic loss
of plasma water and an increase in number and size of red cells which may result in
haemoconcentration (Milligan and Wood 1982, Beckman and Zaugg 1988). Indeed, at least
during acid exposure, haemoconcentration and the concomitant increase in blood viscosity are
believed to be the causes of death (Milligan and Wood 1982, Wood 1989). Randall and Brauner
(1991) suggested that sub-lethal changes in blood viscosity during acid exposure may affect
exercise through changes to cardiac output, principally reduced stroke volume, that reduce
oxygen transport to the tissues. Changes in blood viscosity may also affect the local circulation
of blood through the peripheral capillaries (Wells and Weber 1991). In our earlier studies
(Butler et al. 1992, Beaumont et al. 1995a), there were changes caused by both LP and CLP
exposure to support this hypothesis. While heart rate was not significantly lower at Ucrit in either
case, when compared to fish in water at neutral pH, mean arterial blood pressure (MABP^
haemoglobin concentration ([Hb]), and protein concentration ([Pro]) were higher, indicating the
possibility of a lower plasma volume (Butler et al. 1992, Beaumont et al. 1995a). However,
these alterations were not consistent between treatments; e.g. sub-lethal pH at 15°C caused
elevations in [Hb] and [Pro] but no change in MABP, while at 5°C there were smaller changes to
[Hb] and no change to [Pro] but MABP was significantly elevated (Butler et al. 1992).
As a result of these inconsistencies, further measurements have been taken from CLP-
exposed brown trout at 10°C, which included analysing samples of muscle tissue for a number of
metabolites in order to investigate the possibility of localised hypoxia (Beaumont et al. In Prep.).
In highly aerobic tissues such as brain and red muscle, there is normally no change in the
carbohydrate stores or energy charge during exposure to environmental hypoxia (Dunn and
Hochachka 1986). However, when challenged with a level of hypoxia that is severe or nearly
lethal, these tissues are unable to maintain their energy status, and glycogen, ATP, and
phosphocreatine (PCr) concentrations all fall (Johnston 1975, van Raaij et al 1994). Since all of
these factors were unchanged in the red muscle of resting CLP trout (Table 1), we conclude that
this tissue did not face a severe hypoxia under these conditions. However, red muscle lactate
was higher in CLP trout in comparison to control trout (Figure 2). Since there was no
corresponding increase in circulating lactate concentration, this is most likely endogenously
derived as a result of increased anaerobic metabolism and could indicate the presence of some
moderate constraint upon the delivery of oxygen to the red muscle.
53
-------
Resting lactate concentration was also elevated in white muscle of CLP trout and, in
addition, glycogen and PCr were depleted. White muscle has a high capacity for anaerobiosis,
and during hypoxia both ATP and energy charge are maintained through elevated glycogenolysis
and the use of PCr energy stores (Dunn and Hochachka 1986). Even when the level of hypoxia
is near fatal, rainbow trout (Oncorhynchus mykiss) can maintain white muscle energy status (van
Raaij et al. 1994). A moderate level of tissue hypoxia could, therefore, be the cause of the
differences between the white muscle of CLP trout and control trout at rest, i.e., elevated lactate,
lowered glycogen and PCr, but no change in ATP.
Table 1. Glycogen, phosphocreatine, and ATP concentrations in red and white muscle
of brown trout at rest or after exercise to Ucrit and following 96 hours of exposure to either
circum-neutral, copper-free water or to water containing 0.08 umol/L Cu + at pH 5 and at 10°C.
Control
(pH 7, no copper)
CLP ,.
(pH 5, 0.08 umol/L Cu )
Rest
Exercise
Rest
Exercise
Red muscle
Glycogen
PCr
ATP
Glycogen
PCr
ATP
26.3 ± 1.8
2.8 ± 0.6
7.3 ± 0.7
31.4 ±3.5
16.6 ± 2.4
12.6 ± 0.6
19.5 ± 1.5
1.6 ± 0.6
5.9 ± 0.9
27.4 ±8.3
4.9 ± 1.2
9.8 ± 1.5
White muscle
26.2 ±1.8
2.7 ± 0.7
6.4 ±0.8
15.2 ± 2.5*
9.0 ±2.1*
11.5 ±0.3
23.9 ± 1.9
2.1 ±0.5
6.1 ±0.7
21.0 ± 1.8
6.6 ±3.1
10.2 ± 1.1
1 Values are means ± S.E.M (n = 6) and units are umol g"1 tissue wet weight. Symbols used in this table:
'indicates a significant effect of exercise in a given water quality, and * a significant effect of copper and low pH
upon resting or exercised trout. One, two or three symbols signify PO.05, 0.01 or 0.001 respectively. (From
Beaumont et al. In Prep.).
It seems that the muscle metabolite data from resting brown trout support the hypothesis
that CLP trout are subject to localised tissue hypoxia. However, under control conditions,
exercise caused lactate to be elevated in both muscles (Figure 2). A limitation in the delivery
of oxygen in CLP trout ought to have been revealed by further increases in lactate following
exercise, i.e., when a 'functional hypoxia' is superimposed on any existing problem. This,
however, was not the case. Lactate concentration was no higher in either white or red muscle
of CLP trout despite, in the case of the white muscle, being below that normally seen at Ucrit
in control trout. The red muscle lactate concentration of the resting CLP fish was already at a
similar level to that of control fish at Ucrit (6.2 ± 0.8, and 5.2 ± 0.9 umol/L respectively) but
"exhaustive" excercise may cause considerably greater increases (Day and Butler 1996). There
were also no changes in PCr in either muscle of CLP trout, other than those which also occurred
in CLP trout at rest.
54
-------
White Muscle
Red Muscle
Control CLP Control CLP
I Rest O Exercise
Figure 2. (A) Lactate and (B) pyruvate concentration measured in white and red muscle of
brown trout sampled at rest or after exercise to Ucrit and following 96 hours of exposure
to either circum-neutral, copper-free water (control) or to water containing 0.08 umol/L
Cu2+ at pH 5 (CLP). All experiments were conducted at 10°C. Values are means
± S.E.M (n = 6). Symbols indicating significant differences as described in Table 1.
(From Beaumont et til. In Prep.).
More significantly than this lack of a further increase in lactate, the haematological data
from this latest set of experiments raise a question regarding the basis of tissue hypoxia induced
by copper and low pH. Elevated blood viscosity as a factor in the reduced performance of trout
from polluted water was initially proposed in reference to low pH alone (Randall and Brauner
1991); Butler et al. (1992) found clear changes in haematology of LP exposed trout to support
this. In our initial study of copper and low pH (Beaumont et al. 1995a), while some alterations
in haematological variables did occur, the changes were not consistent and may have been
complicated by seasonal and temperature-related effects. The results from the current study are
less equivocal. There are no changes in haematocrit (Hct), [Hb] or [Pro] associated with
exposure to copper at low pH and it seems unlikely that blood viscosity increased significantly.
Moreover, Gallaugher et al. (1995) have now shown that, despite a positive relationship between
Hct and viscosity, polycythemia of up to 55% Hct does not reduce the UCnt of rainbow trout, but
rather increases it.
55
-------
ELEVATED AMMONIA AND METABOLISM
A consistent effect of exposure to copper or acid water, either alone or in combination,
is the elevation of plasma ammonia (Wilson and Taylor 1993a, 1993b, Beaumont et al. 1995b,
Day and Butler 1996). Fish exposed to copper at low pH showed a significant negative
correlation between UCIit and total plasma ammonia concentration ([Tamm]), with an r2 value of
over 0.6 (Beaumont et al. 1995b). Corresponding to the changes which occur in the plasma,
JTamro] in both the red and white muscle are also higher in CLP-exposed fish (Beaumont et al.
In Prep.).
A major effect of ammonia may be to stimulate glycolytie flux by activating the rate
limiting enzyme phosphofructokinase-1 (PFK-1), and this may occur in spite of the absence of
any change in tissue ATP concentration (Raabe and Lin 1984). In the white muscle, which
normally uses carbohydrate fuel, one consequence could be the depletion of glycogen at rest,
particularly if, as some authors have suggested (Zaleski and Bryla 1977), gluconeogenesis is also
impaired in the presence of elevated ammonium ions. With its high capacity for anaerobiosis
and concomitant high lactate dehydrogenase (LDH) activity (Bostrom and Johansson 1972), an
increase in flux through glycolysis would account for the higher lactate concentration in the
white muscle of CLP trout. PCr concentration in the white muscle of CLP trout is lower than
that of control trout at rest. This may represent the cost of maintaining ATP concentration in the
face of some futile cycling in the glycolytic/gluconeogenic pathway.
In red muscle, lactate concentration displays a similar pattern of change to that of white
muscle, in that it is greater in resting CLP trout in comparison to their counterparts in neutral
water while that of exercised CLP trout was not significantly higher than similarly exposed fish
at rest. Unlike white muscle, however, glycogen concentration is not significantly lowered by
exposure to copper at low pH. What, therefore, is the source of the lactate in red muscle and
why is glycogen not depleted by an ammonia-stimulated increase in glycolytie flux?
It is possible that, despite an increase in [Tamm], glycolytie flux may not be increased
in red muscle. Under normal conditions, the PFK-1 activity of red muscle is some 2-3 times
lower than that of white muscle (Knox et al. 1980, Moon and Foster 1995). In addition, while
the activity of PFK-1 from white muscle is influenced by NHt4" (Su and Storey 1994), Storey
(1991) found that PFK-1 extracted from the liver of rainbow trout was not significantly activated
by NHj*. Perhaps red muscle PFK-1 has more similarity to the liver isozyme than that of
white muscle in this respect, and so glycolysis in red muscle is relatively insensitive to
hyperammonaemia. A second, not necessarily exclusive, possibility is that red muscle is using
an alternative to stored glycogen as a fuel source. Red muscle has a greater hexokinase activity
than that of white muscle (Knox et al. 1980), and West et al. (1993) showed the red muscle of
rainbow trout to have a considerable scope for an increase in uptake and utilisation of circulating
glucose, above that normally found in unstressed resting fish. CLP trout do show a significant
plasma hyperglycaemia and it is possible that at rest, the red muscle of these fish is using this
exogenous source of glucose to maintain glycogen stores and fuel an increased flux through the
glycolytie pathway.
56
-------
Under aerobic conditions, the products of glycolysis in the red muscle would normally be
expected to be completely oxidised to CO2 by the TCA cycle and not to accumulate as lactate.
However, the first step in this process is the oxidative decarboxylation of pyruvate to acetyl CoA
by the pyruvate dehydrogenase complex (PDHC) which, in the cat cerebral cortex, is inhibited in
the presence of high ammonia (Katunuma et al. 1966). Fitzpatrick et al. (1988) produced
circumstantial evidence for a similar decrease in the activity of this enzyme in cultured
astrocytes and, refer to unpublished observations of Lai and Cooper in which the presence of
5 mM ammonium reduced PDHC activity by 20%. In the current study, CLP exposed fish had
significantly higher red muscle pyruvate and lactate concentrations than control trout, which may
indeed reflect an impaired transfer into the TCA cycle.
Additionally, carbohydrate is likely to be a relatively minor fuel for oxidative metabolism
in red muscle of fishes (Moyes and West 1995). Substrates such as amino acids, and fatty acids
in particular, are oxidized directly through the TCA cycle without involvement of the glycolytic
pathway. However, hyperammonaemia can additionally impair the function of the TCA cycle,
not only as mentioned earlier by impeding the transfer of substrate into the mitochondria through
its inhibitory effect on PDHC, but also by limiting the activity of its enzymes. In particular,
isocitrate and (3-ketoglutarate dehydrogenases have been shown to be inhibited by ammonium
ions in the mitochondria of liver and brain (Katunuma et al. 1966, Lai and Cooper 1986,1991).
Similarly, the enzymes which form the malate-aspartate shuttle, responsible for the provision of
reducing equivalents for the TCA cycle and the regeneration of cytosolic NAD+, may be
disrupted in the presence of elevated ammonia (Hindfelt et al. 1977, Ratna Kumari et al. 1986,
Lai and Cooper 1991, Kosenko et al. 1997). The degree to which, or even whether, ammonia
affects the malate-aspartate shuttle may be tissue-specific (Fitzpatrick et al. 1988, Faff-Michalak
and Albrecht 1991).
While none of the aforementioned authors has examined the effect of ammonium ions on
the metabolism of fishes, Arillo et al. (1981), in their study of the effects of ammonia on
rainbow trout, refer to unpublished data showing an impaired TCA cycle. In CLP-exposed
brown trout, red muscle citrate is slightly, but significantly, higher than that of control fish at rest
(Beaumont et al. In Prep.) and this accumulation could be a consequence of an impaired
functioning of the TCA cycle. Thus, elevated ammonium ions may reduce the performance of
red muscle by uncoupling and slowing oxidative metabolism, lowering the efficiency of aerobic
metabolism.
ELEVATED AMMONIA AND ELECTROPHYSIOLOGY
In addition to possible metabolic effects, ammonium ions may also interfere with central
or peripheral nervous activity, with transmission at the neuromuscular junction and with
excitation/contraction coupling and muscle electrophysiology (see Beaumont et al. 1995b for
discussion). One particular effect with consequences for muscle function may derive from the
ability of ammonium ions to replace potassium ions in exchange mechanisms (Skou 1965,
Binstock and Lecar 1969) which in the long term can result in the depolarisation of neurons and
57
-------
muscle fibres. In a study of the effect of ammonium ions on frog sartorius muscle, Heald (1975)
attributed a decrease in the twitch tension of the whole muscle to a progressive loss of fibres
which had become electrically inexcitable. This inexcitability, he concluded, was due to the
depolarisation of muscle membrane potential (EM) which could be reversed by the application of
anodal current to the muscle preparation with the consequential recovery of twitch tension. In a
study of the effects of low pH alone, Day and Butler (1996) described the failure of brown trout
to recruit white muscle when swimming at speeds at which it would normally be used and this
may be related to a loss of excitability.
In recent experiments, we have used three methods to investigate the electrophysiological
status of the muscle membrane of CLP-exposed fish (Beaumont et al. In Prep.). Initially, the
distribution of ammonia between the tissues and extracellular fluids of trout exposed to copper
and low pH was used to make some deductions about the muscle membrane potential (EM). The
distribution of ammonia between the extracellular and intracellular tissue compartments offish
has been examined by several authors and was most recently reviewed by Wood (1993). In
general it is assumed that biological membranes are much more permeable to NH3 than to NEU"1"
(Jacobs 1940, Milne et al. 1958, Kleiner 1981). In the case that NHU"1" permeability is negligible,
the movement of ammonia between the extracellular fluid and the tissues will be determined by
the gradient of NH3 partial pressures according to the theory of non-ionic diffusion (Jacobs 1940,
Milne et al. 1958). In this situation, the distribution is established according to the gradient of
pH across the membrane (Equation 1).
m
[Tamm]e
Where [Tamn,^ intracellular total ammonia concentration, and [Tamm]e = extracellular total
ammonia concentration.
With a typical extra- to intracellular pH gradient of about 0.5 units, the ratio of
i/ [TaJnm]e is relatively low, approximately 5, and this is the situation that has been
observed in most mammalian muscle (e.g. Stabenau et al. 1959, Meyer et al. 1980, MacLean
et al. 1995, Bangsbo et al. 1996). However, as a result of its relatively high pK of approximately
9-10, most ammonia will be present in physiological systems as NH4+. This large difference in
the relative availability of the two forms means that NH4+ permeability can be significant while
still not necessarily being greater than that of NH3. In this case, the distribution of ammonia
across the membrane is influenced by the gradient of voltage across it in addition to that of pH.
In the extreme situation, [Tamnji/ [Tamm]e is determined by the Nernst equation (Equation 2), and
muscle with EM of -80 to -90 m V would have a [Tamm] ratio of 30 or more.
EM=~
(2)
Where R = the gas constant, T = temperature °K, z = the valency, and F = the Faraday constant.
58
-------
It is more likely that distribution is governed both by pH and electrical gradient as
described by Equation 3 (Roos and Boron 1981, see Wood 1993) in which the ratio of relative
permeabilities of the muscle membrane to NH3 and NH4+ (pNH3 / pNHU+) determines the relative
significance of the pH gradient and the membrane potential.
(PNH3/pNH;)-(FxEM/RxT(1-r))x([H*]e+K)
(pNH3/PNH;)-(FxEM
C3)
EMF
where r = exP -M— , and K is the dissociation constant for NH3/NH4+.
RT
In white muscle of fish, most authors have found a high ratio of [Tainm]i / [Tamm]e
(e.g. Wright et al. 1988a, 1988b, Wright and Wood 1988, Saha and Ratha 1989, Tang et al
1992). Resting brown trout under control conditions had a [Tamnji/ [Tamm]e ratio of 28.4 ± 6.1
and 33.6 ± 7.8 for white and red muscles respectively (Beaumont et al. In Prep.). These values
suggest that the ratio of membrane permeability in both tissues is relatively low and hence that
the distribution of ammonia is mainly determined by EM. Indeed, predictions of the [Tamnji/
[Tammje ratio made using the hypothesis of non-ionic diffusion (Equation 1) fall far short of the
measured ratio, while those calculated from the Nerast equation (Equation 2, with EM -85 mV)
are close to it (Table 2). Conversely, using the Nernst equation to calculate EM from the
observed distribution of NBU+, provides estimates of -83.0 ± 5.9 and -79.3 + 5.5 mV for red and
white muscle respectively. These estimates fall within the range of vertebrate and more
specifically fish muscle measured under resting, steady state conditions (Hikada and Toida 1969,
Stanfield 1972, Eugene and Barets 1983, Altringham and Johnston 1988).
Table 2. The ratio [Tamnji / [Tamm]e at rest in white and red muscle of brown trout following
96 hours of exposure to either circum-neutral, copper-free water or to water containing
0.08 umol/L Cu2+ at pH.5.'
Tissue
White muscle
Red muscle
Treatment
No copper, pH 7.
0.08 umol/L Cu^pH 5.
No copper, pH 7.
0.08 umol/L Cu2+, pH 5.
Observed
28.4 ±6.1
10.5 ±1.2
33.6 ± 7.8
15.3 ±2.1
[Tammji / [Tammje
Predicted from
non-ionic
diffusion (pH)
5.3+0.5
5.9 ± 1.3.
6.3 ± 0.9
5.9 ± 0.4
Predicted from
Nernst equation
(EM=-85mV)
32.2 ± 0.05
32.3 ± 0.06
32.2 ± 0.05
32.3 ± 0.04
1 The ratios presented are calculated from the measured values of [Tamm], from the measured pH gradient
(Equation 1) or from the Nernst equation (equation 2) assuming a muscle membrane potential of-85 mV.
Values are mean ± S.E.M (n = 6).
59
-------
The mean [Tamm]i/ [Tamm]e ratios of the red and white muscle of CLP trout at rest are also
shown in Table 2. Both are significantly lower than the observed ratios of control trout. Since
there are no differences in plasma, red or white muscle pH of control or CLP trout, there can be
only two explanations for this decrease in [Tan^] ratio; either pNH3 / pNH4+ has increased or the
voltage gradient across the membrane decreased. In the latter case, one might try to estimate the
size of the depolarisation from the distribution of ammonia. The use of Equation 3 requires an
estimate of pNHs / pNH4+ which, although it applies a somewhat circular approach, could be
usefully estimated for control fish, using a value for EM from the literature (e.g. Hikada and
Toida 1969, Altringham and Johnston 1988). Alternatively, one could assume that the
contribution of the pH gradient is insignificant and predict white and red muscle EMs of CLP
trout using the Nernst equation. The results from both approaches are presented in Table 3 and
indicate a depolarisation of some 17-25 mV.
Table 3. The resting muscle membrane potential (EM) of white and red muscle from brown
trout following 96 hours of exposure to either circum-neutral, copper-free water or to
water containing 0.08 umol/L Cu2+ at pH 5.
Tissue
White muscle
Red muscle
Treatment
No copper, pH 7.
0.08 umol/L Cu2+, pH 5.
No copper, pH 7.
0.08 umol/L Cu2+, pH 5.
Calculated
From Nernst
equation
-79.3 ± 5.5
-57.4 ± 3.2
-83.0 ± 5.9
-65.6 ± 3.0
EM (mV)
From Roos &
Boron equation
-60 ±4
-67 ±3
Calculated using permeability ratios (pNEb / pNH/) of 2 and 1.1 for white and red muscle respectively estimated
from the control trout, assuming an EM of-85 mV. Values are mean ± S.E.M (n = 6).
However, there has been some theoretical criticism of the hypothesis of distribution
according to EM due to the additional stress placed upon cellular pH regulation by proton
cycling (Heisler 1990). Moreover, not all studies are in agreement with the finding of a high
[Tamm]i/ [Tamm]e (e.g. Mommsen and Hochachka 1988), and recent studies have suggested that
the situation may be more plastic in that the pH gradient has the dominant effect upon ammonia
distribution at rest, and the effect of muscle membrane potential is dominant following exercise
(Wang et al. 1994,1996). This analysis also makes a number of assumptions and it should be
noted in particular that, as the studies of Wang et al. (1994, 1996) imply, membrane
permeability to NH3 and NHU"1" may change under some circumstances. Bearing these criticisms
in mind, we have attempted to determine muscle membrane potential using two alternative
60
-------
methods. Where we have had sufficient tissue remaining (n = 4 for both control and CLP), white
muscle K+, Na+ and Cl~ concentrations were measured and used, with the plasma concentrations
of these ions and with membrane permeabilities from Hodgkin and Horowicz (1959), to estimate
EM from the Goldman-Hodgkin-Katz equation. These results also indicate that this muscle is
depolarised after copper/acid exposure (Table 4). In a second series of experiments (Beaumont
et al In Prep.), direct measurements of white muscle EM of trout from both control and CLP
conditions have been made. These also support the predictions made from the distribution of
ammonia. In three trout taken from control conditions, the mean membrane potential was -83.4
± 1.9 mV while that of five CLP trout was -52.5 ± 4.6 mV (Table 4).
Table 4. Estimates of the resting membrane potential (EM) of white muscle from brown trout
following 96 hours of exposure to either circum-neutral, copper-free water or to water
containing 0.08 umol/L Cu2+ at pH 5 (From Beaumont et al. In Prep.).1
Treatment
No copper, pH 7.
0.08 umol/L Cu2+, pH 5.
Calculated from [Timm,]
ratio and Nernst
equation
-79.3 ± 5.5 (n=6)
-57.4±3.2(n=6)
EM(mV)
Calculated from
Goldman-Hodgkin-
Katz equation
-83.9 ± 1.6 (n=4)
-64.2 ± 2.7 (n=4)
Measured using
microelectrodes
-83. 4 ± 1.9(n=3)
-52.5 ± 4.6 (n=5)
1 Measurements obtained from three sources, the distribution of ammonia, the distribution of ions and direct
measurement using microelectrodes.
Thus, despite the problems inherent in estimating membrane potential from ammonia
distribution, the predictions made using it are supported by the data from two other methods of
analysis, including the direct measurement of EM. Trout exposed to copper at low pH do indeed
seem to have muscle fibres that are significantly depolarised with respect to those of control
trout and this is probably due to an elevation in ammonia in CLP-exposed trout. Jenerick (1959)
reports the onset of loss of excitability of frog muscle fibres above -60 mV and complete loss
from -55 to -45 mV (Jenerick 1956), and this probably arises from the inactivation of the
voltage-gated sodium channel. Using rat skeletal muscle, Heald (1975) demonstrated the
functional significance of such a depolarisation showing that there was a reduction in twitch
tension as fibres became electrically inexcitable. That this was a membrane phenomenon was
demonstrated by the presence of contractures in fibres treated with caffeine or in which outward
current was injected. In the brown trout, neither of the two methods of calculating red muscle
voltage gradient gave values of less than -65 mV, while white muscle was estimated to be below
-60 mV and measured as -52.5 ± 4.6 mV. This may provide some explanation of why the red
muscle of trout exposed to sub-lethal acid continues to function, despite an increase in ammonia,
while white muscle does not (Day and Butler 1996).
61
-------
From our studies of brown trout exposed to copper and low pH, we have now cause to
believe that one of the major physiological effects of sub-lethal exposure is an increase in
ammonia both in the circulation and tissues. This has effects upon the metabolism and the status
of the muscle membrane potential, and consequently upon the ability of the animal to swim.
It is likely that ammonia also has significant effects upon other tissues. For example, Bubien and
Meade (1986) report that isolated hearts of the brook trout (Salvelinus fontinalis) show
arrhythmias and abnormal ventricular electrograms when exposed to ammonium ions. Nervous
tissue is also likely to be affected in such a way as to disrupt the co-ordination of exercise.
Although the brain is in some way protected against ammonia intoxication by high levels of
glutarnine synthetase activity (Wilson and Fowlkes 1976), an elevation in circulating ammonia
may cause both central and peripheral effects through depolarisation and through the inhibition
of glutaminase and hence the production of the neurotransmitters glutamate, aspartate, and
GAB A (O'Neill and O'Donovan 1979). Ammonia can also cause the inactivation of Cl" extrusion
from neurons which decreases the effectiveness of postsynaptic inhibition and may occur at
levels of ammonia that have no other effect upon metabolism or EM (Raabe and Lin 1984).
REFERENCES
Altringham, J.D., and LA. Johnston. 1988. Activation of multiply innervated fast and slow
myotomal muscle fibres of the teleost Myoxocephalm scorpius. Journal of Experimental
Biology 140:313-324.
Arillo, A., C. Margiocco, F. Melodia, and P. Mensi. 1981. Effects of ammonia on liver
lysosomal functionality in Salmo gairdneri Richardson. Journal of Experimental
Zoology 218:321-326.
Bangsbo, J., B. Kiens, and E.A. Richter. 1996. Ammonia uptake in inactive muscles during
exercise in humans. American Journal of Physiology 270:E101-E106.
Beaumont, M.W., P.J. Butler, and E.W. Taylor. 1995a. Exposure of brown trout, Salmo trutta,
to sub-lethal copper concentrations in soft acidic water and its effect upon sustained
swimming performance. Aquatic Toxicology 33:45-63.
Beaumont, M.W., P.J. Butler, and E.W. Taylor. 1995b. Plasma ammonia concentration in
brown trout (Salmo trutta) exposed to sub-lethal copper concentrations and its
relationship to decreased swimming performance. Journal of Experimental Biology
198:2213-2220.
Beaumont, M.W., P.J. Butler, and E.W. Taylor. In Prep. Exposure of brown trout (Salmo trutta)
to a sub-lethal concentration of copper in soft, acidic water: effects upon muscle
metabolism and membrane potential.
62
-------
Beckman, B.R., and W.S. Zaugg. 1988. Copper intoxication in chinook salmon (Oncorhynchus
tshawytscha) induced by natural spring water: Effects on gill Na+, K+-ATPase,
haematocrit and plasma glucose. Canadian Journal of Fisheries and Aquatic Sciences
45:1430-1435.
Binstock, L., and H. Lecar. 1969. Ammonium ion currents in the squid giant axon. Journal of
General Physiology 53:342-361.
Bostrom, S.L. and R.G. Johansson. 1972. Enzyme activity patterns in white and red muscle of
the eel (Anguilla anguilld) at different development stages. Comparative Biochemistry
and Physiology 42B:533-542.
Brett, J.R. 1964. The respiratory metabolism and swimming performance of young sockeye
salmon. Journal of the Fisheries Research Board of Canada 21:1183-1226.
Bubien, J.K., and T.L. Meade. 1986. Electrophysiological abnormalities produced by
ammonium in isolated perfused brook trout, Salvelinusfontinalis, hearts. Journal of Fish
Biology 28:47-53.
Butler, P.J., and N. Day. 1993. The relationship between intracellular pH and swimming
performance of brown trout exposed to neutral and sub-lethal pH. Journal of
Experimental Biology 176:271-284.
Butler, P.J., N. Day, and K. Namba. 1992. Interactive effects of seasonal temperature and low
pH on resting oxygen uptake and swimming performance of adult brown trout (Salmo
truttd). Journal of Experimental Biology 165:195-212.
Day, N., and P.J. Butler. 1996. Environmental acidity and white muscle recruitment during
swimming in the brown trout (Salmo truttd). Journal of Experimental Biology 199:
1947-1959.
Dunn, J.F., and P.W. Hochachka. 1986. Metabolic response of trout (Salmo gairdneri) to acute
environmental hypoxia. Journal of Experimental Biology 123:229-242.
Eugene, D., and A. Barets. 1983. Electrical activity of trout skeletal muscle fibers. Journal De
Physiologic 78:814-820.
Faff-Michalak, L., and J. Albrecht. 1991. Aspartate-aminotransferase, malate-dehydrogenase
and pyruvate carboxylase activities in rat cerebral synaptic and non-synaptic
mitochondria effects of in vitro treatment with ammonia, hyperammonemia and hepatic
encephalopathy. Metabolic Brain Disease 6:187-197.
63
-------
Fitzpatrick, S.M., A.J.L. Cooper, and L. Hertz. 1988. Effects of ammonia and I3-methylene-DL-
aspartate on oxidation of glucose and pyruvate by neurons and astrocytes in primary
culture. Journal of Neurochemistry 51:1197-1203.
Franks, J. 1983. Acid rain. Chemistry in Britain 19:504-509.
Gallaugher, P., H. Thorarensen, and A.P. Farrell. 1995. Haematocrit in oxygen transport and
swimming in rainbow trout (Oncorhynchus mykiss). Respiratory Physiology 102:
279-292.
Heald, D.E. 1975. Influence of ammonium ions on mechanical and electrophysiological
responses of skeletal muscle. American Journal of Physiology 229:1174-1179.
Heisler, N. 1990. Mechanisms of ammonia elimination in fishes. In: Animal Nutrition and
Transport Processes, Vol. 6. J.P. Truchot and B. Lahlou (Eds). Karger, Bassel.
pp. 137-151.
Hikada, T., and N. Toida. 1969. Biophysical and mechanical properties of red and white muscle
fibres in fish. Journal of Physiology (London) 201:49-59.
Hindfelt, B., F. Plum, and T.E. Duffy. 1977. Effect of acute ammonia intoxication on cerebral
metabolism in rats with portaclaval shunts. Journal of Clinical Investigation 59:386-396.
Hodgkin, A.L., and P. Horowicz. 1959. The influence of potassium and chloride ions on the
membrane potential of single muscle fibres. Journal of Physiology (London) 148:
127-160.
Jacobs, M.H. 1940. Some aspects of cell permeability to weak electrolytes. Cold Spring
Harbour Symposium of Quantitative Biology 8:30-39.
Jenerick, H.P. 1956. The relations between prepotential, resting potential and latent period in
frog muscle fibres. Journal of General Physiology 39:773-787.
Jenerick, H.P. 1959. The control of membrane ionic currents by the membrane potential of
muscle. Journal of General Physiology 42:923-930.
Johnston, I. A. 1975. Studies on the swimming musculature of the rainbow trout. II. Muscle
metabolism during severe hypoxia. Journal of Fish Biology 7:459-467.
Katunuma, N., M. Okada, and Y. Nishii. 1966. Regulation of the urea cycle and TCA cycle by
ammonia. In: Advances in Enzyme Regulation, Vol. 4. G. Weber (Ed.). Pergamon
Press, London, pp. 317-335.
64
-------
Kirk, R.S., and J.W. Lewis. 1993. An evaluation of pollutant induced changes in the gills of
rainbow trout using scanning electron microscopy. Environmental Technology 14:
577-585.
Kleiner, D. 1981. The transport of NH3 and NH4+ across biological membranes. Biochimica
Biophysica Acta 639:41-52.
Knox, D., M.J. Walton, and C.B. Cowey. 1980. Distribution of enzymes of glycolysis and
gluconeogenesis in fish tissues. Marine Biology 56:7-10.
Kosenko, E., V. Felipo, C. Montolin, S. Grisolia, and Y. Kaminsky. 1997. Effects of acute
hyperammonemia in vivo on oxidative metabolism in non-synaptic rat brain
mitochondria. Metabolic Brain Disease 12:69-82.
Lai, J.C.K., and A.J.L. Cooper. 1986. Brain a-ketoglutarate dehydrogenase complex: kinetic
properties, regional distribution and effects of inhibitors. Journal of Neurochemistry
47:1376-1386.
Lai, J.C.K., and AJ.L. Cooper. 1991. Neurotoxicity of ammonia and fatty acids- differential
inhibition of mitochondrial dehydrogenases by ammonia and fatty acyl coenzyme A
derivatives. Neurochemical Research 16:795-803.
Lauren, D. J., and D.G. McDonald. 1987. Acclimation to copper by rainbow trout Salmo
gairdneri: biochemistry. Canadian Journal of Fisheries and Aquatic Sciences 44:
105-111.
MacLean, D. A., J.K. Barclay, and T.E. Graham. 1995. Electrical stimulation and amino acid
and ammonia metabolism in the canine gastrocnemius muscle. American Journal of
Physiology 268:R759-R770.
McDonald, D.G. 1983. The interaction of environmental calcium and low pH on the physiology
of the rainbow trout, Salmo gairdneri. 1. Branchial and renal net ion and H+ fluxes.
Journal of Experimental Biology 102:123-140.
Meyer, R.A., G.A. Dudley, and R.L. Terjung. 1980. Ammonia and IMP in different skeletal
muscle fibres after exercise in rats. Journal of Applied Physiology 49:1037-1041.
Milligan, C.L., and C.M. Wood. 1982. Disturbances in hematology, fluid volume distribution
and circulatory function associated with low environmental pH in the rainbow trout,
Salmo gairdneri. Journal of Experimental Biology 99:397-415.
Milne, M.D., B.H. Scribner, and M.A. Crawford. 1958. Non-ionic diffusion and the excretion
of weak acids and bases. American Journal of Medicine 24:709-729.
65
-------
Mommsen, T.P., and P.W. Hochachka. 1988. The purine nucleotide cycle as two temporally
separated metabolic units: a study on trout muscle. Metabolism 37:552-556.
Moon, T.W., and G.D. Foster. 1995. Tissue carbohydrate metabolism, gluconeogenesis and
hormonal and environmental influences. In: Biochemistry and Molecular Biology of
Fishes: Vol. 4. P.W. Hochachka and T.P. Mommsen (Eds.). Elsevier Science,
Amsterdam, pp. 65-100.
Morris, R., E.W. Taylor, D.J.A. Brown, and J.A. Brown. 1989. Acid toxicity and aquatic
animals. SEB Seminar Series. Cambridge University Press, Cambridge, New York.
Moyes, C.D., and T.G. West. 1995. Exercise metabolism offish. In: Biochemistry and
Molecular Biology of Fishes: Vol.4. P.W. Hochachka and T.P. Mommsen (Eds.).
Elsevier Science, Amsterdam, pp. 367-392.
O'Neill, B.B., and D.J. ODonovan. 1979. Cerebral energy metabolism in hyperammonaemia.
Biochemical Society Transactions 7:35-36.
Powell, J.F., and B.A. McKeown. 1986. The effects of acid exposure on the ion regulation and
seawater adaptation of coho salmon (Oncorhynchus kisutcK) parr and smolts.
Comparative Biochemistry and Physiology 83C:45-52.
Raabe, W., and S. Lin. 1984. Ammonia, post-synaptic inhibition and CNS energy state. Brain
Research 303:67-76.
Randall, D.J., and C. Brauner. 1991. Effects of environmental factors on exercise in fish.
Journal of Experimental Biology 160:113-126.
RatnaKumari, L., G.Y.C. Subbalakshmi, and C.R.K. Murthy. 1986. Acute effects of ammonia
on the enzymes of the citric acid cycle in rat brain. Neurochemistry International
8:115-120.
Roos, A., and W.F. Boron. 1981. Intracellular pH. Physiological Reviews 61:296-434.
Saha, N., and B.K. Ratha. 1989. Comparative study of ureogenesis in freshwater, air-breathing
teleosts. Journal of Experimental Zoology 252:1-8.
Skou, J.C. 1965. Enzymatic basis for active transport of Na+ and K+ across cell membrane.
Physiological Reviews 45:596-617.
Spry, D.J., C.M. Wood, and P.V. Hodson. 1981. The effects of environmental acid on
freshwater fish with particular reference to the soft water lakes in Ontario and the
modifying effects of heavy metals. A literature review. Canadian Technical Report
Fisheries and Aquatic Sciences 999.
66
-------
Stabenau, J.R., K.S. Warren, and D.P. Rail. 1959. The role of pH gradient in the distribution of
ammonia between blood and cerebrospinal fluid. Journal of Clinical Investigation
38:373-383.
Stagg, R.M., and TJ. Shuttleworth. 1982. The effects of copper on ionic regulation by the gills
of the seawater adapted flounder (Platychthys flesus L.). Journal of Comparative
Physiology 149:83-90.
Stanfield, P.R. 1972. Electrical properties of white and red muscle fibres of the elasmobranch
fish Scyliorhinus canicula. Journal of Physiology (London) 222:161-186.
Storey, K.B. 1991. Metabolic consequences of exercise in organs of rainbow trout. Journal of
Experimental Zoology 260:157-164.
Su, S.Y., and Storey, K.B. 1994. Regulation of rainbow trout white muscle
phosphofructokinase during exercise. International Journal of Biochemistry 26:519-528.
Tang, Y., H. Lin, and D.J. Randall. 1992. Compartmental distributions of carbon dioxide and
ammonia in rainbow trout at rest and following exercise, and the effect of bicarbonate
infusion. Journal of Experimental Biology 169:235-249.
Taylor, E.W., M.W. Beaumont, P.I Butler, J. Mair, and M.S. Mujallid. 1996. Lethal and sub-
lethal effects of copper upon fish: a role for ammonia toxicity? In: Toxicology of
Aquatic Pollution. E.W. Taylor (Ed.). Cambridge University Press, Cambridge,
New York, pp.85-113.
van Raaij, M.T.M., E. Bakker, M.C. Nieveen, H. Zirkzee, and G.E. Van den Thillart. 1994.
Energy status and free fatty acid patterns in tissues of common carp (Cyprinus carpio, L.)
and rainbow trout (Oncorhynchus mykiss, L.) during severe oxygen restriction.
Comparative Biochemistry and Physiology 109A:755-767.
Wang, Y.X., G.J.F. Heigenhauser, and C.M. Wood. 1994. Integrated responses to exhaustive
exercise and recovery in rainbow-trout white muscle — acid-base, phosphogen,
carbohydrate, lipid, ammonia, fluid volume and electrolyte metabolism. Journal of
Experimental Biology 195:227-258.
Wang, Y.X., G.J.F. Heigenhauser, and C.M. Wood. 1996. Ammonia movement and distribution
after exercise across white muscle cell membranes in rainbow trout. American Journal
of Physiology 40:R738-R750.
Wells, R.M.G., and R.E. Weber. 1991. Is there an optimal haematocrit for rainbow trout,
Oncorhynchus mykiss (Walbaum)? An interpretation of recent data based on blood
viscosity measurements. Journal of Fish Biology 38:53-65.
67
-------
West, T.G., P.O. Arthur, R.K. Suarez, CJ. Doll, and P.W. Hochachka. 1993. In vivo utilisation
of glucose by heart and locomotory muscles of exercising rainbow trout (Oncorhynchus
mykiss). Journal of Experimental Biology 177:63-79.
Wilson, R.P., and P.L. Fowlkes. 1976. Activity of glutamine synthetase in channel catfish
tissues determined by an improved tissue assay method. Comparative Biochemistry and
Physiology 54B:365-368.
Wilson, R.W., and E.W. Taylor. 1993a. Differential responses to copper in rainbow trout
{Oncorhynchus mykiss) acclimated to sea water and brackish water. Journal of
Comparative Physiology 1638:239-246.
Wilson, R.W., and E.W. Taylor. 1993b. The physiological responses of freshwater rainbow
trout, Oncorhynchus mykiss, during acutely lethal copper exposure. Journal of
Comparative Physiology 1638:38-47.
Wood, C.M. 1989. The physiological problems offish in acid waters. In: Acid Toxicity and
Aquatic Animals. R. Morris, E.W. Taylor, DJ.A. Brown, and J.A. Brown (Eds.).
Cambridge University Press, Cambridge, New York. pp. 124-152.
Wood, C.M. 1993. Ammonia and urea metabolism and excretion. In: The Physiology of Fishes.
D.H. Evans (Ed.). CRC Press, Boca Raton, pp. 379-425^
Wright, P.A., and C.M. Wood. 1988. Muscle ammonia stores are not determined by pH
gradients. Fish Physiology and Biochemistry 5:159-162.
Wright, P. A., D.J. Randall, and C.M. Wood. 1988a. The distribution of ammonia and
H between tissue compartments in lemon sole (Parophrys velulus) at rest, during
hypercapnia and following exercise. Journal of Experimental Biology 136:149-175.
Wright, P.A., C.M. Wood, and D.J. Randall. 1988b. An in vitro and in vivo study of the
distribution of ammonia between plasma and red-cells of rainbow-trout (Salmo
gairdnerf). Journal of Experimental Biology 134:423-428.
Zaleski, J., and J. Bryla. 1977. Effects of oleate, palmitate and octanoate on gluconeogenesis
in isolated rabbit liver cells. Archiv Biochimica Biophysica 183:553-562.
68
-------
THE MUDSKIPPERS: AMMONIA TOXICITY AND TOLERANCE
Y. K. Ip1, K. W. Peng1, S. F. Chew2, W. K. Kok1, J. Wilson3, and D.J. Randall3
ABSTRACT
Ammonia is toxic to mudskippers, although they have higher ammonia tolerance than
other fishes. One species of the local mudskippers, Periophthalmodon schlosseri, can tolerate
exceptionally high concentration of environmental ammonia. The 24-, 48-, and 96-hour median
lethal concentration (LCso) of un-ionized ammonia (NH3) for P. schlosseri are 643, 556, and
536 jj,M, respectively. The LCso values for those same time periods for another mudskipper,
Boleophthalmus boddaerti, are 77.1,64.0, and 60.2 juM. Urea production and excretion are
not utilized as a mean for environmental ammonia detoxification in these mudskippers.
At sub-lethal concentrations of NHs (446 juM for P. schlosseri and 36 ^M for B. boddaerti), both
mudskippers detoxify ammonia to free amino acids (FAA), leading to increases in concentrations
of total free amino acids (TFAA) in the brains, livers, and muscles. In the brains, glutamine is
the major free amino acid involved. The high ammonia tolerance in P. schlosseri is due to its
ability of actively excreting ammonia against a concentration gradient. It can maintain low
levels of ammonia in its tissues and organs when exposed to 36 or 135 juM NH3. The branchial
Na+, K+-ATPase activity in this mudskipper is very high relative to other fishes, and it can be
activated by physiological levels of NEit+ instead of K+. Ammonia excretion by P. schlosseri
against a concentration gradient is inhibited by the addition of ouabain to the external medium.
Amiloride added to the water also reduces ammonia excretion by this mudskipper. It is
concluded that Na+, K+-ATPase and a Na+/Ffl" exchanger are involved in the active excretion of
ammonia across its gills. This unique ability of P. schlosseri to excrete ammonia actively seems
related to the special structure of its gills, and allows the fish to continue to excrete ammonia in
air and in its burrow, which may contain water of high ammonia concentrations.
INTRODUCTION
Periophthalmodon schlosseri and Boleophthalmus boddaerti are two species of
mudskippers which inhabit the intertidal zone of the mudflats at the Pasir Ris estuary, Singapore.
Although they are mudskippers, their branchial morphologies and morphometries are very
different (Low et al. 1988, 1990). The gills of P. schlosseri exhibit several terrestrial adaptations.
They have thick gill rods, branched gill filaments, and intrafilamentary secondary lamellar
fusions, which prevent the filaments from collapsing and the lamellae from coalescing. The gill
filaments of B. boddaerti are comparatively longer and larger in number. The secondary
lamellae are aligned parallel to the respiratory water current, suggesting that B. boddaerti is well
adapted to aquatic respiration.
Department of Biological Sciences, National University of Singapore, Republic of Singapore. 2Biology Division,
School of Science, National Institute of Education, Nanyang Technological University, Republic of Singapore.
Department of Zoology, University of British Columbia, Vancouver, Canada.
69
-------
Indeed, these two mudskippers have very different behaviors in their natural habitats.
B. boddaerti has a lesser affinity to land and makes burrows at the lower regions of the mudflat.
At low tide, it is usually found on the mud and enters water only periodically. As the tide
rises, it retreats into a burrow and remains submerged until the tide ebbs (Ip et al. 1990). In
comparison, P. schlosseri has relatively greater affinity to land. It builds its burrow at the higher
regions of the mudflat and, during high tides, is usually found swimming along the water's edge
with its snout above water (Ip et al. 1990).
For fishes exposed to a terrestrial condition, branchial ammonia excretion would be
inefficient as no external water current is available to irrigate the gills. Hence, it is possible that
P. schlosseri, having developed a greater affinity to land than B. boddaerti, has also acquired a
greater capacity to detoxify ammonia. Indeed, P. schlosseri can survive better than B. boddaerti
under conditions of ammonia-loading in the laboratory (Ip et al. 1993), indicating that
P. schlosseri would exhibit an environmental ammonia tolerance higher than that of other fishes.
Experiments were undertaken to obtain the median lethal concentration (LC5o) of ammonia for
P. schlosseri and B. boddaerti to enable a comparison with information available in the
literature. In addition, attempts were made to examine the mechanisms involved in enabling
P. schlosseri the capability of tolerating high levels of environmental ammonia.
MATERIALS AND METHODS
Collection and Maintenance of Mudskippers
P. schlosseri (56-110 g) and B. boddaerti (6-23 g) were collected from the estuarine
canal at Pasir Ris, Singapore between 1991 and 1998, and maintained at 25°C in the laboratory
as described by Peng and Ip (1998).
Determination of LCso
Fifty percent seawater was prepared by dissolving Wimexhw-Marinemix (Wiegandt,
Germany) seasalt in tapwater 24 hours in advance. Salinity was measured using a YSI Model 33
S-C-T meter (Yellow Springs Instrument Co. Inc., USA). Trizma Base (Sigma Chemical Co.,
USA) at a final concentration of 10 mM was added to the 50% seawater and the pH of the Tris-
50% seawater was adjusted to 7 with concentrated HC1 using a Sigma Tris electrode (E-4878).
Groups of 10 individuals of B. boddaerti (6-8 g, 6-8 cm) were immersed in 3 L of
Tris (pH 7.0)-50% seawater at different concentrations of reagent grade NHUCl (Merck, West
Germany) at 25.0 ± 0.5°C with aeration. The M^Cl concentrations tested were 5,6, 8,10,12,
17, and 20 mM. P. schlosseri (50-65 g, 18-25 cm) were individually exposed to 3 L of Tris
(pH 7.0)-50% seawater containing various concentrations of NHUCl in plastic aquaria.
The NHjCl concentrations tested were 100, 110, 120, 130, 140 and 150 mM. Ten P. schlosseri
were tested at each concentration of NHUCl. B. boddaerti and P. schlosseri exposed to Tris-50%
seawater served as controls. The experiments were repeated twice to confirm the mortality of
fishes at each NHUCl concentration tested. The test solutions were renewed with freshly
dissolved NH^C! every 24 hours. After the fishes had been exposed to the test solutions
70
-------
for 24,48 or 96 hours, mortality was recorded and dead fishes were removed. Fish were
considered dead if they were immobile and exhibited no respiratory movements. The 24-, 48-,
and 96-hour LC50 values were estimated according to the methods of Litchfield andWilcoxon
(1949) on log-probit paper.
Exposure of Mudskippers to Sublethal Concentrations of NH4C1 and Collection of Tissues
Based on the results obtained from LC50 studies, the sublethal concentrations of NH4C1
for P. schlosseri and B. boddaerti were chosen at 100 mM (446 jjM NHa) and 8 mM (36 juM
NHs), respectively. To allow for comparison between the two mudskippers, P. schlosseri was
also exposed to 8 mM NH^Cl. After 6 days of exposure, P. schlosseri and B. boddaerti were
anaesthetized with 3-aminobenzoic acid ethylesther (MS 222; pH adjusted to 7) (Sigma
Chemical Co.) at final concentrations (w/v) of 0.125% and 0.08%, respectively, for 10 minutes.
To collect blood, the caudal peduncle was severed and exuding blood was collected from the
caudal artery in heparinized micro-hematocrit capillary tubes. The tubes were centrifuged at
4,000 g for 10 minutes at 4°C. The resulting plasma was deproteinized immediately with two
volumes (v/v) of 2% ice-cold perchloric acid and centrifuged at 10,000 g for 15 minutes.
The resulting supernatant fluid was used for analyses of urea, ammonia, and FAA.
For the collection of other tissues and organs, the anaesthetized fish was killed with a
blow on the head. The lateral muscle, liver, and brain were excised as fast as possible. No
attempt was made to separate the red and white muscle. The samples were immediately freeze-"
clamped in liquid nitrogen with pre-cooled aluminum tongs (Faupel et al. 1972). The whole
procedure was carried out within 1 minute. Samples were stored at -80°C until analysis.
Determination of Urea, Ammonia, and FAA Concentrations
For urea and ammonia analyses, the frozen samples were weighed and homogenized in
three volumes (w/v) of 6% perchloric acid over three time intervals of 20 seconds each with
10 seconds between each interval using an Ultra-Turrax homogenizer (Janke and Kundel Gmbh
and Co., Germany) at maximum speed (24,000 rpm). For the determination of FAA
concentrations, the frozen sample was homogenized in 7 volumes (w/v) of 6% trichloroacetic
acid. The homogenates were centrifuged at 10,000 g for 20 minutes. The supernatant fluids
obtained were used for the analysis of urea or FAA.
For the analysis of FAA, the sample was adjusted to pH 2.2 with 4 M LiOH and diluted
appropriately with 0.2 M lithium citrate buffer (pH 2.2). Amino acids were analyzed using a
Shimadzu LC-6A Amino Acid Analysis System with a Shim-pack ISC-07/S1504 Li type
column. Ammonia was determined according to the method of Kun and Kearney (1974). Urea
assay was performed with a Sigma Urea Assay Kit Procedure 535. The sample was neutralized
with 4 M K2CO3 and 0.3 ml of the sample was used for the assay. To another 0.3 ml of the same
sample, similar analysis was performed after incubating for 15 minutes at 25°C with 0.2 ml of
20 mM Tris-HCl (pH 7.0) containing 5 IU urease (Sigma Chemical Co., USA). The difference
in absorbance obtained from samples in the presence and absence of urease was used for the
71
-------
estimation of urea concentration in the sample. Urea obtained from Sigma Chemical Co. was
used as a standard for comparison. Concentrations of ammonia, urea and FAA were expressed
as jamol/g wet weight for tissues and umol/ml for the plasma.
Determination of Rates of Urea Excretion
Individual P. schlosseri (60-70 g) and groups of three B. boddaerti (a total of 50-60 g)
were exposed to 400 ml of 50% seawater containing 10 mM Tris HC1 (pH 7.0) for 1 day. At the
end of the 24-hour period, a water sample was collected for the determination of urea. The
individual P. schlosseri and groups of B. boddaerti were then exposed to the same volume of
Tris-seawater containing 100 mM NtttCl and 8 mM NHjCl, respectively, with daily renewal of
test solutions. After 1 day or 6 days exposure to NHUCl, a water sample was collected for urea
determination as mentioned above.
Determination of Branchial Na+, K^-ATPase and Na+, NH4+-ATPase Activities
Mudskippers were exposed to 50% seawater containing 10 mM Tris (pH 7.0) for 6 days.
Gill samples were collected according to the method of Zaugg (1982) and stored at -80°C until
analysis. The gill samples in frozen buffer were allowed to thaw on ice and processed following
the method of Chew et al (1998). Their methods of determining ATPase activities was also
adopted in this study. The specific ATPase activity was expressed as umol inorganic phosphate
released/20 minutes per mg protein. Na+,K+-ATPase was determined in the presence of 100 mM
NaCl, 20 mM KC1, and 5 mM MgCl2, and its activity was obtained as the difference between
enzyme activities assayed in the presence and absence of 1 mM ouabain. Na+,NH4+- ATPase was
determined in the presence of 20 mM NKUCl, 10 mM NaCl, and 5 mM MgCl2, and its activity
was obtained as the difference between enzyme activities assayed in the presence of
NaCl+NHtCl and MgCl2.
Effects of Ammonia Exposure on Ammonia Excretion
Muskippers were placed in 5 volumes (w/v) of 50% seawater containing 10 mM Tris
(pH 7.0) in conical flasks with continuous aeration, and left to stabilize for 12 hours overnight.
Prior to experimentation the seawater in the flask was replaced with the same volume of Tris-
50% seawater, or Tris-50% seawater containing 30 mM NI^Cl (135 uM NH3). A water sample
was taken immediately at start of test, and again at 24 hours. Water samples were analyzed for
ammonia concentrations using a Tecator Aquatec Analyzer equipped with an ammonia cassette.
Statistical Analysis
Results were presented as means ± SE. Student's 't' test and one way analysis of variance
followed by Duncan's multiple range test were used to compare differences between means
where applicable. Differences with PO.05 were regarded as statistically significant.
72
-------
RESULTS
The 24-, 48-, and 96-hour LC50 values (mM with 95% confidence intervals, n=3) of
NH^ClforS. boddaerti were 17.3 (14.7-20.1), 14.2 (11.1-17.4), and 13.5 (10.7-16.2),
respectively. The values for P. schlosseri were much higher; the 24-, 48-, and 96-hour LC50
values were 144 (137-151), 128 (111-146) and 120 (105-135), respectively. None of the
mudskippers died when exposed to Tris-50% seawater (control) for 96 hours. Using the
Henderson-Hasselbach equation, the 24-, 48-, and 96-hour LC5o values of NH3 for B. boddaerti
were calculated to be 77. 1, 64.0, and 60. 1 uM, respectively. For P. schlosseri for the same time
periods, the LCso values of NH3 were 643, 556, and 536
Exposure to 36 jjM NH3 had no significant effect on the levels of urea in the muscle,
liver, and plasma of B. boddaerti (Figure 1 ). Similarly, the levels of urea in the muscle, liver
and plasma of P. schlosseri exposed to 446 jxM NH3 were comparable to those of the control
(Figure 2). In addition, ammonia exposure has no significant effect on the rate of urea excretion
by B. boddaerti, but significantly reduced the excretion of urea by P. schlosseri.
5.00 '->
•2 4.00 -
J=
• 3.00 -
o a.
j= 2.00
si
i= o
1.00 -
0.00 J
Muscle
Liver
Plasma
Figure 1. Concentration of urea in the muscle, liver, and plasma of B. boddaerti
(N=5) exposed to Tris-50% seawater with 36 jaM NH3 for 6 days (Black bars).
Control fish were exposed to Tris-50% seawater (White bars).
73
-------
1.00 -i
0.80 -
.C
O>
o
ft
^ - 0-40
1
o
O
0.20 -
0.00 J
Muscle
Liver
Plasma
Figure 2. Concentration of urea in the muscle, liver, and plasma of P. schlosseri
(N=5) exposed to Tris-50% seawater with 446 joM NH3 for 6 days (Black bars).
Control fish were exposed to Tris-50% seawater (White bars).
Exposure of P. schlosseri to 36 juM NHs exhibited no significant effect on the
concentrations of various FAA and total FAA (TFAA) in the brain, liver, muscle and plasma
(Figure 3). However, concentrations of TFAA in the brain, liver, muscle and plasma of
P. schlosseri exposed to 446 jjM NHs increased significantly compared to the control values
(Figure 4). Glutamine contributed 91.4% to the increase in TFAA (excluding taurine)
concentration in the brain of this mudskipper. In the liver of P. schlosseri exposed to 446 ^M
NHs, glutamine, glutamate, and alanine contributed 72.5,14, and 8.4% to the increase in the
concentration of TFAA. In the muscle, the increase in the concentration of TFAA was mainly
due to glutamine (25%), alanine (17.3%) and lysine (14.5%). In the plasma, glutamine,
alanine, and glutamate contributed 20.5%, 11.1% and 10.3% to the increase in the concentration
of TFAA.
74
-------
40
30 -
-
I
10
o -i
Brain Liver Muscle Plasma
Figure 3. Concentration of total free amino acids (T.FAA) in the brain, liver, muscle, and
plasma of P. schlosseri (N=5) exposed to Tris-50% seawater with 36 LiM NH3 for 6 days
(Black bars). Control fish were exposed to Tris-50% seawater (White bars).
60
50
o
£ 40
30 J
20
10
I
Brain
Liver Muscle Plasma
Figure 4. Concentration of total free amino acids (TFAA) in the brain, liver, muscle,
and plasma of P. schlosseri (N=5) exposed to Tris-50% seawater with 446 jaM NH3
for 6 days (Black bars). Control fish were exposed to Tris-50% seawater (White bars).
* = Significant difference from corresponding control value.
75
-------
The concentrations of TFAA in the brain and muscle of B. boddaerti increased
significantly after exposure to 36 juM NH3 (Figure 5). Glutamine contributed the greatest to the
increase in the TFAA concentration in the brain. In the liver of B. boddaerti exposed to 36 jaM
NHs, glutamine and glutamate contributed 34.3 and 36.8% to the increase in TFAA. In the
muscle of B. boddaerti, glycine, glutamine, and alanine contributed 53.8, 15.45, and 10.2% to
the increase in TFAA concentration. There was no significant increase in the TFAA
concentration in the plasma of B. boddaerti exposed to 36 fiM
60
20
10
Brain Liver Muscle
Plasma
Figure 5. Concentration of total free amino acids (TFAA) in the brain, liver, muscle,
and plasma of B. boddaerti (N=5) exposed to Tris-50% seawater with 36 \\M NH3
for 6 days (Black bars). Control fish were exposed to Tris-50% seawater (White bars).
* = Significant difference from corresponding control value.
76
-------
Ammonia concentrations in the plasma, brain, liver, and muscle of P. schlosseri was
unaffected by exposure to 36 fxM NH3 for 6 days (Figure 6). In contrast, the ammonia
concentrations in the plasma, brain, liver, and muscle of B. boddaerti exposed to 36 jaM NH3
were approximately 3.8, 3.7, 5.0, and 5.4 times higher than the control values (Figure 7).
Similar results were obtained for P. schlosseri when it was exposed to 446 |_iM NH3 (Figure 8).
The blood pH values of P. schlosseri (7.38 + 0.03, n=6) and B. boddaerti (7.35 + 0.03, n=6) in
the control condition were comparable. Exposure to ammonia exerted no significant effect on
the blood pHs of these two mudskippers.
6.00 i
§ 5.00 -
a.
«1
§ i 4.00 -
•
o .
3.00 -
2.00-
1.00
0.00 J
Brain
Liver
Muscle Plasma
Figure 6. Concentration of ammonia in the brain, liver, muscle, and plasma of
P. schlosseri (N=5) exposed to Tris-50% seawater with 36 juM NH3 for 6 days
(Black bars). Control fish were exposed to Tris-50% seawater (White bars).
77
-------
15.00 1
I
5.00 -
0.00
J
Brain Liver
Muscle Plasma
Figure 7. Concentration of ammonia in the brain, liver, muscle, and plasma of
B. boddaerti (N=5) exposed to Tris-50% seawater with 36 joM NHs for 6 days
(Black bars). Control fish were exposed to Tris-50% seawater (White bars).
* = Significant difference from corresponding control value.
30.00
o § 20.00
a *
I fe
8 s>
8C yt
,w
.& •£
I* J
| S 10.00 -
O)
a
0.00 J
Brain Liver Muscle Plasma
Figure 8. Concentration of ammonia in the brain, liver, muscle, and plasma of
P. achloaseri (N=5) exposed to Tris-50% seawater with 446 |uM NHs for 6 days
(Black bars). Control fish were exposed to Tris-50% seawater (White bars).
* = Significantly different from corresponding control value.
78
-------
The branchial Na+, K+-ATPase (2.75 ± 0.32, n = 4) and Na+,NH4+-ATPase (2.05 ± 0.30,
n = 4) activities (jimol inorganic phosphate 720 min per mg protein) from P. schlosseri were
approximately three times higher than those from B. boddaerti (0.88 ± 0.08 and 0.69 ± 0.06,
respectively, n = 4). Ouabain inhibited the activities of branchial Na+,K+-ATPase and Na+,
NH4+-ATPase from both mudskippers.
Table 1. Concentrations of ammonia during ammonia excretion/absorption tests on
Periophthalmodon schlosseri and Boleophthalmus boddaerti.
Day
Time
(hours)
NH3mmol/L
Control tank
Test tank1
P. schlosseri
1
1
6
6
0
24
0
24
0.080 ±0.002
3.38 ±0.026
0.127 ±0.008
3.78 ±0.483
29.4 ±0.215
32.8 ±0.218
30.1 ±0.400
32.1 ±0.900
B. boddaerti
1
1
6
6
0
24
0
24
0.090 ± 0.007
2.61 ±0.253
0.152 ±0.020
2.63 ± 0.203
29.7 ± 0.347
28.3 ±0.193
**
**
1 Test tank concentration 50% seawater with 10 mM Tris and 30 mM NHUC1 (135nM NH3). Results reported are
mean value ±SE (n=3). ** = Dead by 24 hours.
P. schlosseri was capable of maintaining normal ammonia excretion in up to 30 mM
NHUC1 (135 nMNH3) (Table 1). Inhibition of ammonia excretion (Figure 9) was observed when
ouabain was added to the seawater containing 2 mM NH^d. In addition, ammonia excretion by
P. schlosseri could be inhibited by the addition of amiloride to the external medium (Figure 10).
79
-------
800 -
Figure 9. Ammonia excretion rate of P. schlosseri (N=5) exposed to Tris-50% seawater
with 2 mM NH4C1 and 0.1 mM ouabain for 3 hours (Black bars). Control fish were
exposed to Tris-50% seawater with 2mM NH4 Cl (White bars).
* = Significant difference from corresponding control value.
Control 0.1 mM Amiloride Recovery
b
I
ab
1
I
Hgure 10. Ammonia excretion rate of P. schlosseri (N=3) exposed to
Tris-50% seawater or Tris-50% seawater with 0.1 mM amiloride for 3 hours.
80
-------
DISCUSSION
The LCso of ammonia for both mudskippers, especially those for P. schlosseri, were
higher than the values for many fishes reported elsewhere. Calamari et a/.(1981) reported that
the lowest lethal concentration of NHs for salmonids was 11.8 juM (0.2 mg/L). For other fishes,
the 96-hour LC50 ranged from 8.2-247 \iM (0.14-4.2 mg/L) (Thurston et al. 1983). In comparison
with other gobies, both P. schlosseri and B. boddaerti could tolerate much higher levels of
ammonia. Iwata (1988) reported that the water-breathing gobioid fishes, Tridentiger obscurus,
Acanthogobius flavimanus (estuary), Bathygobim fuscus, Chasmichthys dolichognathus
dolichognathus (tide pool) and Rhinogobim brunneus (freshwater) died within 24 hours when
they were exposed to 10 mM NtLtCl dissolved in 20% seawater.
In the presence of high levels of ambient ammonia, there would presumably be a net
influx of ammonia into the fish. In addition, excretion of nitrogenous waste as NHs becomes
difficult as a result of a decrease or reversal in the blood-to-water diffusion gradient of NFI?.
Ammonia affects, among other things, the nervous system, intracellular pH and integrity of the
cell membrane (Campbell 1991), leading to convulsions and death of the fish (Olson and Fromm
1971). To survive such periods of ammonia loading, the fish must have the ability to convert
toxic ammonia into less or non-toxic forms.
Accumulated ammonia can be converted to urea, provided the fish has the ability to
synthesize urea via the ornithine-urea cycle (Mommsen and Walsh 1992). Two well-known
examples of fishes tolerant of high levels of ammonia (75 mM NHiCl and above) are the
air-breathing catfish (Heteropneustes fossilis) and the Lake Magadi tilapia (Oreochromis
alcalicus grahami). Their high tolerance to ammonia could be due to the presence of the
ornithine-urea cycle, which increased urea production under hyperammonia stress (Saha and
Ratha 1987, 1989; Walsh et al. 1993). Our results indicate that the high tolerance of P. schlosseri
and B. boddaerti to ammonia were unrelated to urea formation and excretion.
Both species of mudskippers coped with ammonia loading by converting the toxic
ammonia into FAA, though the sublethal concentrations of ammonia to which they were exposed
were very different. Approximately 91% of the increases in the TFAA (excluding taurine)
concentrations in the brains of these two mudskippers were due to glutamine. For the liver of
P. schlosseri exposed to 446 |jM NH3, other than glutamine (62%), glutamate (12%), and alanine
(12%) also contributed significantly to the increase in TFAA concentration. In the liver of
B. boddaerti exposed to 36 jiMNHb, glutamine and glutamate contributed 34.3% and 36.8%,
respectively, to the increase in TFAA concentration. Glutamine is likely to be released from the
brain under conditions of chronic ammonia exposure (Mommsen and Walsh 1991). The organ
likely to deal with glutamine is the liver. Glutamine could either be deaminated by hepatic
glutaminase for the generation and excretion of ammonia when the insult has subsided, or for the
formation and excretion of urea should the capability exist (Mommsen and Walsh 1992). When
B. boddaerti and P. schlosseri were exposed to their respective sublethal concentrations of NH3,
the level of TFAA in the muscle increased significantly. Muscle might be an important storage
site for various FAA, as it is the largest tissue of the fish by weight.
81
-------
There were drastic differences in the levels of ammonia between the tissues of
P. schlosseri and B. boddaerti exposed to 36 jaM NH3. The ammonia levels in the tissues of
P. schlosseri exposed to 36 juMNHs remained comparable to those of the control. The most
intriguing results obtained in this study were the lack of effect on any of the parameters studied
when P. schlosseri was exposed to 36 yM NH3. This is very different from the situation when
the same mudskipper was subjected to 446 uM NH3. In 36 j^M NH3, B. boddaerti was
apparently unable to prevent the influx of ammonia, and ammonia detoxification mechanisms
were invoked, leading to the increases in TFAA concentrations. These results suggest that
P. schlosseri might have special mechanisms in maintaining a low steady state level of ammonia
in its body when it was exposed to an NH3 concentration which is lethal to other fishes.
If the blood pH of an animal is kept at a relatively high pH, less of the NH3 which
diffuses into the animal would be protonated. Thus, less ammonia should enter the blood.
We tried to examine if the blood pH of P. schlosseri was relatively more alkaline, however,
results obtained did not support such a hypothesis. Hence, input of ammonia from the
environment is likely to be the same for P. schlosseri and B. boddaerti exposed to 36 jjM NH3.
Since output due to conversion to FAA was not observed in the P. schlosseri, the only possible
way for this mudskipper to keep the steady state level of ammonia low would be to excrete it out
to the environment, despite the unfavorable concentration gradient. Indeed, ammonia excretion
was maintained when P. schlosseri was exposed up to 135 pM. NH3 for 24 hours. Since the
excretion rate of ammonia was constant and independent of the external ammonia concentration,
it can be assumed that the efflux of ammonia was increased to offset the increase in ammonia
influx.
In the presence of Na+ and Mg+, NHU+could stimulate branchial ATPase activities
from P. schlosseri and B. boddaerti. The Na+, NH4+-ATPase activities in these mudskippers
could be inhibited by ouabain, a well known inhibitor for Na+, K+-ATPase, indicating that the
Na4, NH4+-ATPase and Na+, K+-ATPase were probably the same enzyme. The activities of these
branchial ATPases from P. schlosseri were approximately three times higher than those from
B. boddaerti. Substitution of NtV" for K+ln Na+, K+-ATPase has been proposed as a method of
branchial ammonia excretion in fish (Evans and Cameron, 1986), and NH4+ can be a more
effective counter ion for Na+, K+-ATPase than K+ (Mallery, 1983). Therefore the high activities
of Na*,K+-ATPase in the gills of P. schlosseri could probably explain the high ammonia
tolerance of this fish.
Indeed, inhibition of ammonia excretion by P. schlosseri was observed when ouabain
was added to the seawater containing 2 mM NHjCl. One problem that basolateral uptake of
K* in the branchial cells as NHU* was a slow depletion of K+ in the branchial cells as N1HU+ was
transported into the cell instead of K+. This decrease in intracellular K+ concentration is harmful
to the cell as K+ is an important factor for many enzymes and for the potassium gradient over the
cytoplasmic membrane (Martinelle and Maggstrom 1993). However, it is possible that the high
activity of Na+,K+-ATPase present in the gills of P. schlosseri was adequate in maintaining the
K* gradients between the branchial cells and the blood when part of the ATPase activity was
operating to remove NH4+ during ammonia exposure.
82
-------
Substitution of NKLf1" for H* has been proposed as a method of branchial ammonia
excretion across the apical surfaces of chloride cells in fish (Claiborn et al. 1982). The excretion
of ammonia by this mudskipper was inhibited by the addition of amiloride to the external
medium, indicating that, indeed, a Na+/H+ exchanger might be involved in the excretion process.
Hence, NtV", which was pumped from the blood across the basolateral membrane into the cells,
could be removed by the apical Na+/H+ (NHU"*) exchange mechanism into the external medium.
In conclusion, P. schlosseri might have evolved to have mechanisms to transport
ammonia actively against an electrochemical gradient, thereby maintaining low tissue ammonia
levels and enabling it to tolerate high environmental ammonia concentrations.
REFERENCES
Calamari, D., R. Marchetti, and G. Vailati. 1981. Effects of long term exposure to ammonia on
the developmental stages of rainbow trout (Salmo gairdneri Richardson}. In: The Early
Life History of Fish: Recent Studies. R. Lasker, and K. Sherman (Eds.). Rapports et
Proces-verbaux des Reunions Conseil International pour 1'Exploration de la Mer.
Copenhagen, Denmark.
Campbell, J.W. 1991. Excretory nitrogen metabolism. In: Environmental and Metabolic
Animal Physiology. C.L. Prosser (Ed.). Wiley Liss, New York. pp. 277-324.
Chew, S.F., E. Goh, C.B. Lim, and Y.K. Ip. 1998. Cyanide exposure affected the producation
and excretion of ammonia in the mudskipper Boleophthalmus boddaerti. Comparative
Biochemistry and Physiology 120C:441-448.
Dabrowska, H., and T. Wlasow. 1986. Sublethal effect of ammonia on certain biochemical and
haematological indicators in common carp (Cyprinus carpio L.}. Comparative
Biochemistry and Physiology 83C: 179-184.
Evans, D.H., and J.N. Cameron. 1986. Gill ammonia transport. Journal of Experimental
Zoology 239:17-23.
Faupel, R.P., H.J. Seitz, W. Taraowsk, V. Thiemann, and C. Weiss. 1972. The problem of
tissue sampling from experimental animals with respect to freezing technique, anoxia,
stress and narcosis. A new method for sampling rat liver tissue and physiological values
of glycolytic intermediates and related compounds. Archives of Biochemistry and
Biophysics 148:509-522.
83
-------
Ip, Y. K., S.F. Chew, and R.W.L. Lim. 1990. Ammoniagenesis in the mudskipper
Periophthalmus chrysospilos. Zoological Science 7:187-194.
Ip, Y. K., C.Y. Lee, S.F. Chew, W.P. Low, and K.W. Peng. 1993. Differences in responses
of two mudskippers to terrestrial exposure. Zoological Science 10:51 1-5 19.
Iwata, K. 1988. Nitrogen metabolism in the mudskipper, Periophthalmus cantonensis: changes
in free amino acids and related compounds in various tissues under conditions of
ammonia loading, with special reference to its high ammonia tolerance. Comparative
Biochemistry and Physiology 91A:499-508.
Kun, E., and E.B. Kearney. 1974. Ammonia. In: Method of Enzymatic Analysis. H.U.
Bergmeyer and K. Gawehn (Eds.). Academic Press, New York. pp. 1802-1806.
Litchfleld, J.T. Jr., and F. Wilcoxon. 1949. A simplified method of evaluating dose-effect
experiments. Journal of Pharmacology and Experimental Therapeutics 96:99-1 13.
Low, W.P., Y.K. Ip, and D.J.W. Lane. 1990. A comparative study of the gill morphometry
in the mudskippers Periophthalmus chrysospilos, Boleophthalmus boddaerti and
Periophthalmodon schlosseri. Zoological Science 7:29-38.
Low, W.P., D.J.W. Lane, and Y.K. Ip. 1988. A comparative study of terrestrial adaptation of
the gills in three mudskippers - Periophthalmus chrysospilos, Boleophthalmus boddaerti
and Periophthalmodon schlosseri. Biological Bulletin 175:434-438.
Mallery, C.H. 1983. A carrier enzyme basis for ammonium excretion in teleost gill -
stimulated Na+-dependent ATPase activity in Opsanus beta. Comparative Biochemistry
and Physiology 74:889-897.
Mommsen, T.P., and P.J. Walsh. 1991. Urea synthesis in fishes: evolutionary and biochemical
perspectives. In: Biochemistry and Molecular Biology of Fishes. Vol.1. Phylogenetic
and Biochemical Perspectives. P.W. Hochachka and T.P. Mommsen (Eds.). Elsevier,
New York. pp. 137-163.
Mommsen, T.P., and P.J. Walsh. 1992. Biochemical and environmental perspectives on
nitrogen metabolism in fishes. Experentia 48:583-593.
Olson, K.R., and P.O. Fromm. 1971. Excretion of urea by two teleosts exposed to different
concentrations of ambient ammonia. Comparative Biochemistry and Physiology 40 A:
999-1007.
Peng, K.W., S.F. Chew, C.B. Lim, T.W.K. Kok, S.S.L. Kuah, and Y.K. Ip. 1998. The
mudskippers Periophthalmodon schlosseri and Boleophthalmus boddaerti can tolerate
environmental NHs concentrations of 446 uM and 36 jaM, respectively. Fish Physiology
and Biochemistry 19:59-69.
84
-------
Saha, N., and B.K. Ratha. 1987. Active ureogenesis in a freshwater air-breathing teleost,
Heteropneustesfossilis. Journal of Experimental Zoology 241:137-141.
Saha, N., and B.K. Ratha. 1989. Comparative study of ureogeneis in freshwater air breathing
teleosts. Journal of Experimental Zoology 252:1-8.
Thurston, R.V., and R.C. Russo. 1983. Acute toxicity of ammonia to rainbow trout.
Transactions of the American Fisheries Society 112:696-704.
Walsh, P.J., H.L. Bergman, A. Narahara, C.M. Wood, P.A. Wright, D.J. Randall, J.N. Maina,
and P. Laurent. 1993. Effects of ammonia on survival swimming and activities of
enzymes of nitrogen metabolism in the Lake Magadi tilapia Oreochromis alcalicus
grahami. Journal of Experimental Biology 180:323-327.
Zaugg, W.S. 1982. A simplified preparation for adenosine triphosphatase determination in gill
tissue. Canadian Journal of Fisheries and Aquatic Sciences 39:215-217.
85
-------
-------
EFFECTS OF CADMIUM ON NITROGEN AND PHOSPHORUS TURNOVER
IN A TROPICAL FRESHWATER SNAIL, BROTIA HAINANENSIS
i
Paul K.S. Lam
ABSTRACT
Sublethal effects of cadmium on the turnover of nitrogen and phosphorus in Brotia
hainanensis (Brot, 1872) were investigated. An increase in cadmium concentration caused a
significant reduction in nitrogen consumption and retention by B. hainanensis. There was, however,
no significant change in the amount of nitrogen lost through faecal production. The "no observable
effect concentration" and "lowest observable effect concentration" were 0.8 and 1.0 mg/L,
respectively, when nitrogen consumption was used as the end point, while corresponding values
were 0.4 and 0.8 mg/L if nitrogen retention was used. In contrast, an increase in cadmium
concentration caused no significant change in the amount of phosphorus consumed, egested and
retained by B, hainanensis. The potential effects of toxic chemical stress on the ecological role of
B, hainanensis as important shredders in tropical streams are discussed.
INTRODUCTION
Traditionally, environmental monitoring in Hong Kong and its nearby region relied almost
exclusively on chemical determination of concentrations of specific pollutants/toxicants in the
environment under surveillance. This approach failed to take into account the potential biological
impact of the pollutants on local systems. The lack of this type of ecotoxicological information
precluded the application of the more powerful and useful risk assessment procedure. From the
mid-1990s, there has been a move by the environmental protection authorities in Hong Kong to
incorporate ecotoxicological investigations into their routine monitoring programmes in sites of
special ecological importance, and into risk assessment exercises prior to the discharge of potentially
hazardous effluents, e.g. cooling water and chemically treated sewage.
Initially, environmental samples were shipped overseas, mainly North America, for testing
using North American species. While acknowledging that the testing laboratories in North America
probably have more experience in ecotoxicity testing work, the use of non-native species as test
organisms put into question the ecological relevance of the test results. Recently, universities in
Hong Kong have been allowed to undertake some of the marine ecotoxicological tests using local
species including larvae of the barnacle Balanus amphitrite Darwin, juveniles of the amphipod
Melita longidactyla Hirayama, and fingerlings of the mangrove snapper Lutjanus
argentimaculatus (Forsk.) (Wu et al. 1997a). So far, these tests have involved the laboratory
determinations of the mortality of test organisms exposed to varying concentrations of specific
toxicants over a fixed period of time. Although these acute tests are, in general, fairly reproducible
1 Department of Biology and Chemistry, City University of Hong Kong, Hong Kong.
87
-------
and relatively easy to perform, they tend to lack ecological realism and are less useful in predicting
environmental effects of the toxicants at sublethal concentrations (Sprague 1976, Anderson and
D'Apollonia 1978, McGreer 1979).
To counter these problems, work has been undertaken to develop cost-effective chronic
sublethal tests using local species. For example, a number of tests have been developed using larvae
of the barnacle, B. amphitrite (Wu et al. 1997a, 1997b, 1997c). In addition, a sublethal bioassay
based on changes in the patterns of energy allocation (scope for growth) has been developed for a
freshwater snail, Brotia hainanensis (Prosobrachia: Thiaridae) (Lai and Lam 1994, Lam 1996a,
1996b). A similar assay based on food consumption rate is also available for a freshwater
pulmonate, Radix plicatulus Benson (Lam 1996c).
It should be noted that although ecotoxicological tests often involve the use of only one
or two test organisms, the ultimate purpose of the exercise is to study the effect of toxic
chemicals on the ecological systems so that the systems concerned can be effectively protected.
B. hainanensis has been selected as a test organism mainly because of its potential ecological
importance as a shredder (sensu Cummins 1973). These animals perform important ecological
functions in the processing of allochthonous leaf materials, and converting coarse organic matter to
particulate organic matter in tropical streams (Dudgeon 1982).
Since energy and nutrients are the two main limiting factors in natural ecosystems, the
analysis of the dynamics of energy and nutrient turnover would enable us to understand the
efficiencies of these processes and hence the functioning of the ecosystem. To supplement previous
studies on energetics (Lam 1996a), this investigation aimed to investigate the effect of sublethal
concentrations of cadmium on nutrient turnover in B. hainanensis, and to examine the possible
impact of toxic stress on nutrient turnover in the stream ecosystem via its influence on the shredder
populations.
MATERIALS AND METHODS
Test Species
B. hainanensis (Gastropoda: Prosobranchia: Thiaridae) is a common freshwater snail that
generally occurs in high abundance in clear, stony hill streams and the upper course of rivers in Hong
Kong and tropical Asia (Dudgeon 1982,1989). These snails play an important ecological role in the
transfer of energy and nutrients to other organisms that feed on fine particulscte organic matter.
Field Collection
Only animals with an aperture width less than 14.5 mm (non-breeding animals; Dudgeon
1989, Lam 1996a) were collected from the upper course of Lam Tsuen Riveir, New Territories, Hong
Kong. A description of the collecting site is given by Dudgeon (1982) and Lai and Lam (1994).
88
-------
Food
Freshly fallen Bauhinia sp. leaves were collected adjacent to the stream where
B. hainanensis was taken. In the laboratory the leaves were carefully brushed and rinsed to remove
the silt on the surface, and then cut into 2 cm x 4 cm pieces. They were oven-dried at 70°C and
stored for subsequent use. The stored leaves were rehydrated and presented to the snails following
the same procedure described in Lam (1996a).
Range-Finding Test
Previous studies have indicated that different populations of B. hainanensis can exhibit
significant variations in cadmium tolerance (Lam 1996b). Consequently, a standard 96-hour acute
test was carried out on B. hainanensis to determine the range of sublethal cadmium concentrations to
be used for subsequent experimentation. The experimental procedure followed that described in
Lam (1996a). All snails were acclimated in aerated, artificial pond water (APW; Lam and Calow
1989) (conductivity = 600 ofT1, pH = 7.4 and [Ca2+] = 80 mg/L) at 20 (±1)°C under a 12-hour light
and 12-hour dark cycle for 2 weeks prior to experimentation. The experimental procedure was the
same as that described in Lam (1996a). All subsequent experiments were conducted under the same
laboratory conditions outlined above. The nominal concentrations of cadmium (CdCl2) in the test
containers were 0,4, 8,10,15,20,25, 30 and 40 mg/L. All snails were starved during the entire
exposure period. The 96-hour median lethal concentration (96-hour LC50) for cadmium was
estimated following Finney (19,89).
Nitrogen and Phosphorus Budgets of Individual Snails
The nutrient budget of an individual snail may be summarized by the following mass
balance equation, in which all budget items are expressed in terms of mass of a specific nutrient
for an individual snail:
C=P+E+F
where C is the amount of nutrient consumed as food; P is the amount of nutrient retained for
growth; E is the amount of nutrient lost through excretion; and F is the amount of nutrient lost
through egestion (or defaecation).
Theoretically, the equation should be balanced, and thus any one budget item can be
estimated if all the other items are known. On this basis, P can be estimated from C - E - F.
Assuming that nutrient losses due to excretion and leaching are negligible in comparison to the
nutrient contents of the leaves and faeces, the amount of nutrient retained for growth can be
calculated from the difference between the amount of nutrient consumed and the amount of
nutrient lost through egestion, i.e.:
P = C-F
89
-------
To estimate the various budget items of individual snails exposed to different concentrations
of cadmium, each snail was reared individually in a 400 mL experimental chamber containing
250 mL of the test solution. The nominal concentrations of Cd2+ in the test solutions were 0,0.2,0.4,
0.8 and 1.0 mg/L. These concentrations were selected based on results of the acute range-finding
tests (also see Lam 1996a). Each treatment group consisted of five snails. A known oven-dried
weight ofBauhinia sp. leaves was supplied ad libitum to individual snails. A perforated platform
was placed at the bottom of each chamber so as to facilitate the collection of faeces. Test solutions
were aerated throughout the whole experimental period which lasted 8 days. Over the 8-day
experimental period, unconsumed leaves and faeces were filtered, collected, oven-dried at 80°C, and
weighed every 2 days when the test media were replaced with fresh solution and food replenished.
The somatic (flesh) weights of individual snails were determined by dissolving the snail shells in
10% nitric acid and then oven drying the soft tissues to constant weight at 80°C.
Total nitrogen and phosphorus contents of leaf tissue and faeces were determined using an
automatic flow injection analyzer (Aquatec 5400 Analyzer; Tecator) following the sulphuric acid-
hydrogen peroxide digestion procedure (Allen 1989). The amount of nitrogen or phosphorus
consumed, excreted and egested can thus be expressed in terms of mg N or P per g dry weight of
snail over the experimental period of 8 days.
Statistical Analysis
One-way analysis of variance (ANOVA) was used to test the null hypothesis that the
different levels of cadmium do not affect the dynamics of nutrient turnover of B. hainanensis.
Pair-wise comparisons between the control and individual treatment groups were performed using
Dunnett's tests (Zar 1984) to determine the "no observable effect concentration" (NOEC) and the
"lowest observable effect concentration" (LOEC) for cadmium.
RESULTS
Range-Finding (LC50) Test
The 96-hour LC50 for cadmium was 16 mg/L, which was similar to the value of 14.5 mg/L
reported in Lam (1996a).
Cadmium Effect on Nitrogen and Phosphorus Dynamics
An increase in cadmium concentration caused no significant change In the amount of
phosphorus consumed, egested and retained by B. hainanensis (ANOVA: F4>20 < 2.30, P > 0.05;
Figure 1). In contrast, increasing cadmium concentrations resulted in a significant reduction in
nitrogen consumption (ANOVA: F4 20 = 6.68, P < 0.05; Figure 2), and nitrogen retention
(ANOVA: /xao = 13.17, P < 0.001;'Figure 2).
90
-------
£j 0.5.
1
w
i_ no
o "•"•
Q.
(0
o
&
°- -0.5,
-1.0
• Consumption
A Egestion
T Retention
i A A-
o'.a o!4 oie 0.8 1.0
T
I
I
Cadmium (mg/L)
Figure 1. Effect of various concentrations of cadmium on phosphorus consumption, egestion
and retention (mg P per g dry weight of snail) in Brotia hainanensia. Vertical lines are ± 1 S.E.
No significant difference was observed between control and individual treatment groups.
1-5.1
O)
I
-1.5.
-3.0.
Consumption
Egestion
Retention I
0.6
0.9
*;*
Cadmium (mg/L)
Figure 2. Effect of various concentrations of cadmium on nitrogen consumption, egestion and
retention (mg N per g dry weight of snail) in Brotia hainanensis. Vertical lines are ± 1 S.E.
** = significantly different from control at P < 0.01.
91
-------
There was, however, no significant change in the amount of nitrogen lost through faecal
production (ANOVA: F4)2o = 0.89, P > 0.05). Pair-wise comparisons using Dunnett's tests revealed
that the NOEC and LOEC were 0.8 and 1.0 mg/L, respectively, when nitrogen consumption was
used as the end point, while NOEC and LOEC were 0.4 and 0.8 mg/L, respectively, when nitrogen
retention was used.
DISCUSSION
A number of aquatic species (e.g. the European water flea Daphnia magna) have been used
as standard organisms for toxicity testing all over the world. It is often assumed that data derived
from these tests can be widely applied to many different types of ecosystems in other parts of the
world. However, Wu (1981) pointed out that such an assumption ;might not be appropriate,
particularly when the test organism does not naturally occur in the receiving system. It was argued
that toxicity tests should be carried out on species which are ecologically important in identified
receiving environments. Furthermore, tests should also include ecological processes, over and above
the organisms per se, so that test results can be related in a meaningful way to impact assessment at
the ecosystem level. In order to assess accurately the effects of toxicants on ecological systems in
Hong Kong and south China, standardized toxicity testing protocols on representative local species
are urgently needed. Despite some recent progress, such experimental protocols have not yet been
fully established (Lam 1996a, 1996c).
Previous investigations reported that scope for growth in B. hainctnensis was sensitive to
cadmium (Lam 1996a). The present study indicates that an increase in cadmium concentration
causes no significant change in the amount of phosphorus consumed, egested and retained by
B. hainanensis. However, exposure to sublethal concentrations of cadmium can result in a
significant reduction in nitrogen consumption and retention by B. hainanensis, while no significant
change in the amount of nitrogen lost through faecal production is observed.
Although our results suggest that cadmium stress at levels investigated in this study do not
result in a reduced amount of nitrogen entering into the stream system via faecal production, the
amount of nitrogen consumed by B. hainanensis is significantly reduced. The level of nitrogen lost
through faecal production is probably maintained by metabolism of the animal's body reserve.
In the long term, this would probably have a negative impact on the snail population, leading to a
smaller amount of allochthonous material, in the form of leaf litter, being processed by the snails.
Consequently, less nitrogen would be available to animals, particularly collectors, downstream
leading conceivably to an impairment of the productivity of the stream ecosystem.
In the case of phosphorus, it is conceivable that the relatively stable levels of consumption,
egestion and retention may at least partly be maintained by the animals extracting a sufficiently
larger amount of phosphorus even with a reduced amount of leaf material consumed.
Notwithstanding, the effect of toxic stress on phosphorus turnover may still be manifested through
a suppression of the snail population upon prolonged exposure to the toxic stress.
92
-------
A characterisation of the exact impact of these factors on the stream invertebrate
communities and the integrity of the stream ecosystem would require a detailed study of the input of
allochthonous organic matter into the stream and the processing and utilisation of this resource by
various invertebrate taxa in the system. Only with this type of information, can a rigorous risk
assessment on toxic stressors be carried out.
ACKNOWLEDGEMENTS
The author acknowledges the support of research grants from the City University of Hong
Kong and the Hong Kong Research Grants Council (Number: CityU0952/96M).
REFERENCES
Allen, S.E. 1989. Chemical Analysis of Ecological Materials. Second Edition. Blackwell
Scientific Publications, Oxford. 368pp.
Anderson, P.D., and S. DApollonia. 1978. Aquatic Animals. In: Principles of Ecotoxicology.
Scientific Committee on Problems of the Environment (SCOPE) Report No. 12, John Wiley
&Sons. pp. 187-221.
Cummins, K.W. 1973. Trophic relations of aquatic insects. Annual Review of Entomology 18:
183-206.
Dudgeon, D. 1982. The life history of Brotia hainanensis (Brot, 1872) (Gastropoda: Prosobranchia:
Thiaridae) in a tropical forest stream. Zoological Journal of the Linnean Society 76:141-154.
Dudgeon, D. 1989. Ecological strategies of Hong Kong Thiaridae (Gastropoda: Prosobranchia ).
Malacological Review 22: 39-53.
Finney,D.J. 1989. Probit Analysis. Third Edition. Cambridge University Press, Cambridge,
333pp.
Lai, P.C.C., and P.K.S. Lam. 1994. Scope for growth in a tropical freshwater snail Brotia
hainanensis (Brot, 1872): implications for monitoring sublethal toxic stressor,
In: Conservation and Management of Inland Waters in Tropical Asia and Australia,
The International Association of Theoretical and Applied Limnology, Schweizerbart,
Germany, pp. 315-320.
Lam, P.K.S. 1996a. Sublethal effects of cadmium on the energetics of a tropical freshwater snail,
Brotia hainanensis (Brot, 1872). Environmental Toxicology and Water Quality 11:345-349.
93
-------
Lam, P.K.S. 1996b. Intel-population differences in acute response ofBrotia hainanensis
(Gastropoda, Prosobranchia) to cadmium: genetic or environmental variance?
Environmental Pollution 94:1-7.
Lam, P.K. S. 1996c. Effects of cadmium on the consumption and absorption rates of a tropical
freshwater snail, Radixplicatulus. Chemosphere 32:2127-2132.
Lam, P.K.S., and P. Calow. 1989. Intraspecific life-history variation in Lymnaeaperegra
(Gastropoda: Pulmonata). H Environmental or genetic variance? Journal of Animal
Ecology 58:589-602.
McGreer, E.R. 1979. Sublethal effects of heavy metal contaminated sediments on the bivalve,
Macoma balthica (L). Marine Pollution Bulletin 10:259-262.
I
Sprague,J.B. 1976. Current status of sublethal tests of pollutants on aquatic organisms. Journal of
the Fisheries Research Board of Canada 33:1988-1992.
Wu, R.S.S. 1981. Difference in the toxicities of an oil dispersant and a surface active agent to some
marine animals, and their implications in the choice of species in toxicity testing. Marine
Environmental Research 5:157-163.
Wu, R.S.S., P.K.S. Lam, and B.S. Zhou. 1997a. A settlement inhibition assay with cyprid larvae of
the barnacle Balanns amphitrite. Chemosphere 35:1867-1874.
Wu, R.S.S., P.K.S. Lam, and B.S. Zhou. 1997b. A phototaxis inhibition assay using barnacle larvae.
Environmental Toxicology and Water Quality 12:231-236.
Wu, R.S.S., P.K.S. Lam, and B.S. Zhou. 1997c. Effects of two oil dispersants on phototaxis and
swimming behaviour of barnacle larvae. Hydrobiologia 352:9-16.
Zar, J.H. 1984. Biostatistical Analysis, Second Edition. Prentice-Hall International, Englewood
Cliffs, New Jersey, USA.
94
-------
METAL BIOAVAELABDLJTY TO MARINE INVERTEBRATES:
SIGNIFICANCE OF TROPHIC TRANSFER
Wen-Xiong Wang1
ABSTRACT
Recent studies on the uptake of metals by aquatic invertebrates have revealed a few
important findings that may have significant implications for the settings of water and sediment
quality criteria by regulatory agencies. Trophic transfer has been shown to be a major source
for metal accumulation in several aquatic invertebrates including both suspension feeders
and deposit feeders. These studies demonstrate that a complete understanding of metal
bioavailability must include the knowledge of the significance of trophic transfer and the
physiological processes governing metal uptake. Previous studies are mainly concerned with
the controls of aquatic chemistry on metal bioavailability. It is equally important to understand
the controls of physiological processes on metal accumulation and availability. The implication
of using kinetic models quantitatively to assess the exposure pathways of metals is discussed.
In the barnacle Balanus amphitrite, >60% of the Cd and >95% of the Zn are obtained from
ingested food source. This study stresses the significance of trophic transfer in predicting metal
concentrations in aquatic animals.
INTRODUCTION
In aquatic environments, metals are partitioned into the dissolved and the particulate
phases. Dissolved metal is operationally defined as any metal passing through a certain size
of pore membrane (0.2 or 0.45 jum). Recent studies have, however, shown that a significant
fraction of traditionally defined dissolved metals may indeed be associated with colloids
(e.g., Martin et al, 1995, Wen et al. 1999). Metals in the particulate phase may be associated
with biogenic and abiotic particles, or the sediments. Aquatic organisms are constantly exposed
to metals in both dissolved and particulate phases, which behavior can vary spatially and
temporally and is driven by physical, chemical, and biological processes in the water. Both the
dissolved and particulate metals have been considered in studies of metal accumulation and
availability in the animals. However, the mechanistic understanding of metal bioavailability
appears to be limited to dissolved metals. The development of the free ion speciation model
(Sunda and Guillard 1976, Campbell 1995), for example, has had a major impact on the study of
metal bioavailability. There have been numerous experimental verifications of this model in
different animal and metal systems. In a recent study, Hare and Tessier (1996) demonstrate that
the free ion speciation model can be used to predict Cd concentrations in a freshwater insect
Chaoborus punctipennis collected from different Canadian lakes, when competition of Cd with
proton or humic substances for biological uptake sites is taken into account.
'Department of Biology, The Hong Kong University of Science and Technology, Kowloon, Hong Kong.
95
-------
In contrast to the free ion speciation model, uptake from ingested food sources was not
well studied until a few years ago. In many previous studies, uptake from the dissolved phase
was considered to be responsible for the overall metal accumulation in aquatic animals. For
example, current toxicity tests require that the animals should not be fed during the exposure
period. Thus, only metals in the dissolved phase are considered to be taken up by the animals
and exert their potential toxicity. Water quality criteria are then calculated based on these
toxicity assays by applying a safety factor. Negligence of metal uptake from the particulate
phase is presumably a result of the lack of a conceptual model to describe the control of metal
bioavailability from ingested food. In a few empirical studies, mainly on deposit feeders, metal
concentrations in the animals are found to be directly correlated with the concentrations in
several geochemical fractions in the sediments (Bryan and Langston 1992).
Recent progress in delineating metal exposure pathways stems mainly from the
development of realistic approaches in quantifying important physiological parameters
governing metal uptake, and the development of a bioenergetic-based kinetic model. Kinetic
models have been developed for a long time in radioecological literature, but only Thomann
(1981) emphasized the significance of trophic transfer. Key physiological parameters described
in the model are the metal assimilation efficiency from ingested food sources, and metal uptake
rate constant from the dissolved phase. Overall, the relative significance of each exposure
pathway is determined by the influx rate from each uptake pathway. The influx rate is, in turn,
a function of the uptake rate constant and the metal concentration in that phase. Further, the
uptake rate constant is a function of the absorption efficiency from the dissolved phase or the
assimilation efficiency from the particular phase times the animal's filtration rate or ingestion
rate. In this paper I discuss the most recent progress in understanding the processes controlling
metal bioavailability from food sources. The application of a kinetic model in delineating metal
exposure pathways is illustrated using barnacles as an example.
METAL BIOAVAILABILITY FROM F0OD SOURCE
Realistic measurements of metal physiological parameters, including assimilation
efficiency and uptake rate constants, can be greatly facilitated by a gamma-emitting radiotracer
technique and the development of a pulse-chase feeding technique for assimilation efficiency
measurements (Wang and Fisher In Press a), or a short-term exposure technique for absorption
efficiency measurements. Over the past few years there have been an increasing number of
reports on metal assimilation efficiencies from ingested food source by a diverse number of
aquatic invertebrates, especially in bivalves and zooplankton (Reinfelder and Fisher 1991, Wang
et al. 1996, Wang and Fisher 1997, 1998, In Press a, Reinfelder et a/. 1998). There is no generic
value of metal assimilation efficiency (AE) for different metals and aquatic species. Even for a
given species, various biological and environmental factors may considerably influence the metal
assimilation (Wang and Fisher 1996). Because assimilation quantifies the extent to which a
metal can be retained by the animals following digestive process, it represents an excellent
index for quantifying metal bioavailability from food sources.
An important question remains regarding the process controlling metal assimilation
in aquatic animals. Whether there is a metal species analogous to the free ion species in the
96
-------
dissolved phase that can be used objectively to assess metal bioavailability remains to be
vigorously tested. Reinfelder and Fisher (1991) suggest that the distribution of metals in prey
cells (e.g., phytoplankton) is the only factor controlling metal assimilation in marine copepods,
the dominant zooplankton in marine systems. They show that metal AEs are directly
proportional to their distributions in the algal cytoplasm (i.e., a 1:1 relationship). This study
illustrates that metal bioavailability may simply be predicted based on the measurements of
metal partitioning in an algal cell. It thus has important implications in identifying the binding
phase of a particle that is most bioavailable to the animals. Later, Hutchins et al. (1995)
confirmed such a relationship in copepods for Fe by presenting algae with different cellular
distributions of Fe to the copepods. However, the 1:1 relationship has so far only been shown in
marine copepods characterized by simple gut architecture and a short gut passage time. The
digestion in this group of marine animals is mainly extracellular, which allows a quick passage
of food materials and metals associated with the particles (Wang and Fisher 1998).
In another group of marine suspension feeders, namely the bivalves, the relationship
between metal AE and metal cytoplasmic distribution in the algal cells is often more subtle.
Generally, metals that distribute more in the algal cytoplasm are assimilated at a higher
efficiency (Wang et al. 1996, Reinfelder etal. 1997, Lee and Luoma 1998). However, only
Wang and Fisher (1996) investigated such a relationship for a given metal in the mussel Mytilus
edulis. In that study, a significant relationship (though not 1:1) was found for several metals
such as Co and Se, but not for other metals such as Ag, Cd, and Zn. Recent studies on barnacles
(Balanus amphitrite) reveal that the distribution of metals in algal cytoplasm can also account for
the assimilation of Cd, Cr, and Se (Wang et al. Submitted a). However, barnacles appear to
assimilate a considerable fraction of metals associated with the algal cell wall, in contrast to
marine copepods that only assimilate the cytoplasmic pool of metals. For example, only 9-40%
of Zn is distributed in the algal cytoplasm, and its AE in barnacles can be as high as 88%
(Wang et al. Submitted b). Although the distribution of metals in algal cytoplasm provides a
simple tool for predicting metal assimilation in several aquatic invertebrates, it should be noted
that for many aquatic animals many metals do not follow this general trend. For example, there
is no relationship between the distribution of Zn in algal cytoplasm and its assimilation in
mussels. It is clear that other mechanisms such as metal association with different biochemical
pools (protein, lipid, and polysaccharides) should be sought in order to develop a more realistic
model for the prediction of metal bioavailability from ingested food.
Other parameters that can influence metal assimilation include metal passage time in the
gut (Wang and Fisher 1996), and partitioning of metals in different digestive phases such as the
extracellular and intracellular digestion in bivalves (Decho and Luoma, 1994, 1996, Wang et al.
1995, and Roditi and Fisher In press). These studies emphasize that the physiological processes
are critical for the understanding of metal bioavailability to the animals. However, because the
physiology of the animals varies greatly from one species to another, it is difficult to identify a
common mechanism to predict metal bioavailability from ingested food.
In contrast to our current understanding of marine herbivores, processes governing the
metal bioavailability in higher trophic levels are not well understood. The assimilation of metals
in these animals can be higher or lower than the AEs observed for herbivores. For example,
97
-------
Fowler and Teyssie (1997) report that the AEs of metals in the seastar Marthasterias glacialis
for Ag, Cd, Co, and Zn are 69,73,73, and 78% respectively. In fishes, Reinfelder and Fisher
(1994) find that metal AEs are generally lower than those in most invertebrates. They show that
the AEs in fishes are determined by the distribution of metals in the soft tissues of copepods
upon which the fishes prey. In contrast, Wang et al. (Submitted a) show that the AEs of metals
in the barnacle B. amphitrite are not related to their distributions in the soft tissues of copepods.
Other mechanisms must also be identified to account for the processes controlling metal
bioavailability in higher trophic levels.
Other physiological processes, particularly the ingestive activities of the animals, have
been largely ignored, presumably due to the difficulty in realistically quantifying this
physiological parameter under a complex field condition. Both the quantity and quality of food
available to the animals can vary greatly in natural environments. The overall influx rate of
metals is a function of the metal AE and the animal's feeding activity. Thus a higher
bioavailability is a result of both a higher AE and a higher ingestion rate. A food particle with
a high AE does not necessarily indicate a high bioavailability if its ingestion rate is maintained
at a low level due to its low concentration in ambient waters.
KINETIC MODEL TO DELINEATE METAL EXPOSURE PATHWAYS
Kinetic models incorporating the rate constant of metal uptake from each exposure
pathway are invaluable in delineating the exposure pathways of metals in aquatic invertebrates
(Thomann 1981, Landrum et al. 1992, Wang and Fisher In Press b). Empirical approaches such
as the measurement of metal concentrations in animals following exposure to metals of different
phases have limited implication in real environmental conditions. Models can incorporate the
variation of each physiological and geochemical parameter and thus have greater applicability in
real situations.
Kinetic models have been validated in a few marine invertebrates where extensive
measurements of various physiological and chemical factors are available. For example, Wang
el al. (1996,1997) demonstrate that metal concentrations of Ag, Cd, Se, Cr, and Zn in mussels
(M. edulis) predicted by the bioenergetic-based kinetic model are very close to the actual metal
concentrations measured independently by the National Status and Trend Program. Fisher et al.
(Submitted) recently showed that the metal concentrations in marine copepods predicted by the
kinetic model are comparable to the metal concentrations measured in field collected samples.
These studies unequivocally demonstrate that kinetic models can be readily applied to predict
metal concentrations in aquatic organisms, and that the physiological parameters such as metal
AE, metal influx rate, and metal efflux rate can account for metal accumulation in these animals.
A recent study by Wang et al. (Submitted b) show that the model-predicted Zn concentrations
in barnacle B. amphitrite (2610 - 11,560 ng g"1) are the same as the field measurements
(3100 - 11,000 |^g g"1). Barnacles are known to accumulate Zn and Cu to a phenomenal level,
largely because these metals can be sequestered in the numerous granules (phosphate) in the cell
layer beneath the midgut epithelium (stratum perintestinale) (Walkers al. 1975a, 1975b,
Rainbow 1987).
98
-------
Wang and Fisher (In Press b) review the current understanding of the exposure pathways
in marine invertebrates. For suspension feeders such as mussels and copepods, uptake from
both the dissolved phase and food ingestion can be equally important to the overall metal
accumulation. For marine deposit feeding polychaetes such as Nereis succinea, however, nearly
all metals are obtained from the ingestion of sediments, largely because of their high ingestion
rates and their low rates of uptake from the dissolved phase. The relative importance of each
exposure pathway depends on the rates of uptake from each source. Most recently, a
bioenergetic model has been developed for the barnacles B. amphitrite, which are very abundant
in the Indo-Pacific regions and have been used as a biomonitor for coastal contamination in
Hong Kong and Southern China (Phillips and Rainbow 1988, Rainbow 1993, Blackmore et al.
1998). In this model
Iw = ku • Cw, (1)
and If=AE«IR»Cf,
(2)
where Iw is the influx rate from the dissolved phase, If is the influx rate from the ingested food
source, ku is the uptake rate constant from the dissolved phase, Cwis the metal concentration in
the dissolved phase, AE is the metal assimilation efficiency from ingested food source, IR is the
ingestion rate of the animals, Cf is the metal concentration in ingested food source.
Assuming that phytoplankton is the primary food for the barnacles, metal concentration
in phytoplankton can simply be calculated by Cw and the metal partition coefficient (Kd) in the
phytoplankton (Wang et al. 1996):
Cf=Cw«Kd (3)
Thus the relative importance of metals accumulated from the food source (R) can be
calculated as:
R = If /(Iw + If) = (AE • IR) / [(AE • IR) + (ku / Kd)] (4)
R is, therefore, dependent on four parameters, including AE, IR, ku, and Kd. The numeric values
for each parameter described in Equation 4 for Cd and Zn in the barnacle B. amphitrite as
summarized in Table 1.
Table 1. Numeric values of parameters used in modeling Cd and Zn accumulation
in the barnacle Balanus amphitrite.
Parameters
Cd
Zn
AE
KdCLg1)
0.35 - 0.85a
0.4a
0.174°
2-20d
0.75 - 0.95"
0.4b
1.519°
10 - 100d
Data sources: "Wang et al. Submitted a; ^Mangetal. Submitted b; cRainbowand White 1989; Fisher and
Reinfelder 1995.
99
-------
7"!
\*J
O
o
H—
0
•*— '
E
2
fc»—
0
^£
co
•4—'
Q.
xjO
CP*
\\J\J
80
60
40
T^W
20
1
1
inn
I UU
80
60
A f\
40 -
20 •
Cd ^^-^-^^^^
^ ^^^
-^
AE=0.35
AC r\ cr\
AC— U.DU
AE=0.85
10
Kd(lg'1)
"
Zn
AI~ n 7"=;
AC r\ oc
AC — U.OO
AE=0.95
10 , 10C
KdOg'1)
Figure 1. Predicted percentages of Cd and Zn in the barnacle Balanus amphitrite due to uptake
from ingested food. Kd = metal partition efficient; AE = metal assimilation efficiency.
100
-------
The model calculations indicate that the majority of Zn (>95%) is derived from the
ingestion of food particles; uptake from the dissolved phase contributes negligibly to the overall
Zn concentration in B. amphitrite (Figure 1). This is consistent with the observation that the
majority of Zn is associated with the midgut epithelium. Trophic transfer thus solely accounts
for the phenomenal Zn concentration in B. amphitrite (up to 1.1% of their dry tissue weight).
Furthermore, the relative importance of trophic transfer is independent of the variation of Zn Kd
in phytoplankton and its AE in B. amphitrite. For Cd, the model predicts that 60-100% of metals
is derived from the ingestion of food particles, and dissolved uptake contributes <40% to the
overall Cd accumulation (Figure 1). The relative significance of trophic transfer is more
dependent on Cd Kd than on Cd AE, largely because of the greater variation of Cd Kd in natural
environments (by > 1 order of magnitude).
These results clearly illustrate that kinetic models can be valuable in separating the
uptake pathways of metals in aquatic invertebrates. Of the two metals considered here (Cd and
Zn), Zn has consistently been shown to be mostly derived from ingested food for several aquatic
invertebrates such as mussels, copepods, polychaetes, and barnacles (Wang and Fisher 1998,
In Press b). The significance of Zn trophic transfer is at least a result of the high Zn
concentration in food particles and the high Zn AE, despite the fact that the uptake rate constant
of Zn from the dissolved phase is often high. Uptake of Zn can be greatly facilitated by its
binding with SH-containing ligands. For Cd, the relative importance of trophic transfer varies
among the animals. For example, dissolved uptake is a dominant source for Cd accumulation in
both mussels and copepods, but is less significant in polychaetes and barnacles. Cd
concentrations in food particles are often >2 orders of magnitude lower than Zn concentrations.
RECOMMENDATION
Understanding exposure pathways has important implication for the establishment of
water quality and sediment quality criteria. Bioassays that only consider metal toxicity from the
dissolved phase underestimate metal toxicity to aquatic animals for metals that are mostly
derived from ingested food in natural environments. Existing criteria for water or sediment
quality may therefore be insufficient for the protection of aquatic life due to the unrealistic
conditions of toxicity testing. It is therefore critical for us to understand the exposure pathways
of metals (and other pollutants) in aquatic animals before we develop a new generation of water
and sediment quality criteria. It is advisable that toxicity tests should consider metal exposure
from both dissolved and food sources. Under many circumstances, the latter can be a dominant
source for metal accumulation in aquatic invertebrates, as shown here for Cd and Zn in the
barnacle B. amphitrite.
ACKNOWLEDGEMENT
This study was supported by a Direct Allocation Grant / Research Grant Council.
101
-------
REFERENCES
i
Blackmore, G., B. Morton, and Z. Huang. 1998. Heavy metals in Balanus amphitrite and
Tetradita squamosa (Crustacea: Cirripedia) collected from the coastal waters of Xiamen,
China. Marine Pollution Bulletin 36:32-40.
Bryan, G.W., and W.J. Langston. 1992. Bioavailability, accumulation and effects of heavy
metals in sediments with special reference to United Kingdom estuaries: a review.
Environmental Pollution 76:89-131.
Campbell, P.G.C. 1995. A critique of the free-ion activity model. In: Metal Speciation and
Bioavailability in Aquatic Systems. A. Tessier and D.R. Turner (Eds.). John Wiley
& Sons, Chichester. pp. 45-102.
Decho, A.W., and S.N. Luoma. 1994. Humic and fulvic acid: sink or source in the availability
of metals to the marine bivalve Macoma balthica and Potamocorbula amuremisl
Marine Ecology Progress Series 108:133-145.
i
Decho, A.W., and S.N. Luoma. 1996. Flexible digestive strategies and trace metal assimilation
in marine bivalves. Limnology and Oceanography 41:568-572.
Fisher, N.S., and J.R. Reinfelder. 1995. The trophic transfer of metals in marine systems.
In: Metal Speciation and Bioavailability in Aquatic Systems. A. Tessier and D.R. Turner
(Eds.). John Wiley & Sons, Chichester. pp. 363-406.
Fisher, N.S., I. Stupakoff, S. Sanudo-Wilhelmy, W,-X. Wang, J.-L. Teyssie, S.W. Fowler, and
J. Crusius. (Submitted) Modeling and measuring trace metal concentrations in marine
copepods. Limnology and Oceanography.
Fowler, S.W., and J.-L. Teyssie. 1997. Assimilation and excretion of selected heavy metals
and radionuclides ingested by seastars. Radioprotection - Colloques 32:317-322.
Hare, L, and A. Tessier. 1996. Predicting animal cadmium concentrations in lakes. Nature
38:430-432.
Hutchins, D.A., W.-X. Wang, and N.S. Fisher. 1995. Copepod grazing and the biogeochemical
fate of phytoplankton iron. Limnology and Oceanography 40:989-994.
I
Landrum, P.P., H. Lee, and M.J. Lydy. 1992. Toxicokinetics in aquatic systems: model
comparisons and use in hazard assessments. Environmental Toxicology and Chemistry
11:1709-1725.
Lee, B.-G., and S.N. Luoma. 1998. Influence of microalgal biomass on absorption efficiency of
Cd, Cr, and Zn by two bivalves from San Francisco Bay. Limnology and Oceanography
43:1455-1466.
102
-------
Martin, J.-M., M.-H. Dai, and G. Cauwet. 1995. Significance of colloids in the biogeochemical
cycling of organic carbon and trace metals in a coastal environment - example of the
Venice Lagoon (Italy). Limnology and Oceanography 40:119-131.
Phillips, D.J.H., and P. Rainbow. 1988. Barnacles and mussels as biomonitors of trace
elements: a comparative study. Marine Ecology Progress Series 49:83-93.
Rainbow, P.S. 1987. Heavy metals in barnacles. In: Barnacle Biology. A.J. Southward (Ed.).
A. A. Balkema, Rotterdam, United Kingdom, pp. 405-517.
Rainbow, P.S. 1993. Biomonitoring of marine heavy metal pollution and its application in
Hong Kong waters. In: The Marine Biology of the South China Sea. B. Morton (Ed.).
Hong Kong University Press, Hong Kong. pp. 235-250.
Rainbow, P.S., and S.L. White. 1989. Comparative strategies of heavy metal accumulation
by crustaceans: zinc, copper and cadmium in a decapod, an amphipod and a barnacle.
Hydrobiologia 174:245-262.
Reinfelder, J.R., and N.S. Fisher. 1991. The assimilation of elements ingested by marine
copepods. Science 251:794-796.
Reinfelder, J.R., and N.S. Fisher. 1994. Retention of elements absorbed by juvenile fish
(Menidia menidia, Menidia beryllind) from zooplankton prey. Limnology and
Oceanography 39:1783-1789.
Reinfelder, J.R., N.S. Fisher, S.N. Luoma, J.W. Nichols, and W.-X. Wang. 1998. Trace
element trophic transfer in aquatic organisms: a critique of the kinetic model approach.
Science of Total Environment 219:117-135.
Reinfelder, J.R., W.-X. Wang, N.S. Fisher, and S.N. Luoma. 1997. Assimilation efficiencies
and turnover rates of trace elements in marine bivalves: a comparison of oysters, clams,
and mussels. Marine Biology 129:434-452.
Roditi, H., and N.S. Fisher. (In Press) Rates and routes of trace element uptake in zebra
mussels. Limnology and Oceanography.
Sunda, W.G., and R.R.L. Guillard. 1976. The relationship between cupric ion activity and
the toxicity of copper to phytoplankton. Journal of Marine Research 34:511-529.
Thomann, R.V. 1981. Equilibrium model of fate of microcontaminants in diverse aquatic
food chains. Canadian Journal of Fisheries and Aquatic Sciences 38:280-296.
Walker, G., P.S. Rainbow, P. Foster, and D.L. Holland. 1975a. Zinc phosphate granules in
tissues surrounding the midgut of the barnacle Balanus balanoides. Marine Biology
33:161-166.
103
-------
Walker, G., P.S. Rainbow, P. Foster, and D.J. Crisp. 1975b. Barnacles: possible indicators of
zinc pollution. Marine Biology 30:57-65.
Wang, W.-X., and N.S. Fisher. 1996. Assimilation of trace elements and carbon by the mussel
Mytilus edulis: effects of food composition. Limnology and Oceanography 42:197-207.
Wang, W.-X., and N.S. Fisher. 1997. Modeling metal bioavailability for marine mussels.
Reviews for Environmental Contamination and Chemistry. 151:39-65.
Wang, W.-X., and N.S. Fisher. 1998. Accumulation of trace elements in a marine copepod.
Limnology and Oceanography 43:273-283.
i
Wang, W.-X., and N.S. Fisher. (In Press a) Assimilation of contaminants in aquatic
invertebrates. Environmental Toxicology and Chemistry.
Wang, W.-X., and N.S. Fisher. (In Press b) Delineating metal accumulation pathways for
marine invertebrates. Science of Total Environment.
Wang, W.-X., N.S. Fisher, and S.N. Luoma. 1995. Assimilation of trace elements ingested by
the mussel, Mytilus edulis: effects of algal food abundance. Marine Ecology Progress
Series 129:165-176.
Wang, W.-X., N.S. Fisher, and S.N. Luoma. 1996. Kinetic determinations of trace element
bioaccumulation in the mussel, Mytilus edulis. Marine Ecology Progress Series 140:
91-113.
Wang, W.-X., S.B. Griscom, and N.S. Fisher. 1997. Bioavailability of Cr(ffl) and Cr(VI) to
marine mussels from solution and particulate pathways. Environmental Science and
Technology 31:603-611.
Wang, W.-X., J.-W. Qiu, and P.-Y. Qian. (Submitted a) The trophic transfer of Cd, Cr, and Se
in barnacle Balanus amphitrite from planktonic food. Marine Ecology Progress Series.
Wang, W.-X., J.-W. Qiu, and P.-Y. Qian. (Submitted b) Significance of trophic transfer in
predicting the high concentration of zinc in barnacles. Environmental Science and
Technology.
Wen, L.S., P. Santschi, G. Gill, and C. Paternostro. 1999. Estuarine trace metal distributions
in Galveston Bay: importance of colloidal forms in the speciation of the dissolved phase.
Marine Chemistry 63:185-212.
104
-------
BIOACCUMULATION AND METABOLISM IN FISHES OF COMPLEX MIXTURES
FROM OIL SHALE CONTAMINATION IN ESTONIA
Arvo Tuvikene1, Sirpa Huuskonen2, and Pirjo Lindstrom-Seppa2
ABSTRACT
This paper overviews our studies conducted in the River Narva, in northeast Estonia.
Based on our findings, the use of a multiparameter-based approach, containing both in vivo
and in vitro tests, is suggested for assessing the risk caused by a complex pollution in an aquatic
environment. Xenobiotics in the River Narva basin originate mainly from municipal sewage
and from oil shale mining and processing. Material collected from this area was used to
determine the sensitivity of various biomarkers in feral fish and a fish liver cell line (PLHC-1).
Bioaccumulations of selected polycyclic aromatic hydrocarbons (PAHs) and heavy metals
(HMs), biotransformation enzyme activities in roach (Rutilus rutilits) and perch (Perca
fluviatilis), and cytotoxicity and biotransformation activity in the PLHC-1 cells were used to
study aquatic pollution. Bioaccumulations of PAHs and HMs from fish muscle and liver tissues
were compared with contents in sediments, from which biota-sediment accumulation factors
(BSAFs) were calculated. To assess whether the pollution affected the balance between
bioactivation and detoxication, the ratio between phase I (7-ethoxyresorufin O-deethylase,
EROD, or aryl hydrocarbon hydroxylase, AHH) and phase II (uridinediphosphate-
glucuronosyltransferase, UDP-GT, or glutathione S-transferase, GST) enzyme activities was
calculated. To study the potential risk of compounds present in sediments, the PLHC-1 cells
were exposed to sediment extracts.
Overall, the accumulation of PAHs in fishes from exposed sites was high. The lack of
clear induction of cytochrome P4501A (CYP1A) in fishes may have affected the accumulation
and the metabolism of PAHs. BSAFs of PAHs appeared differently in roach and perch. The
sediment extracts sensitively affected the PLHC-1 cells, and this suggested the use of PLHC-1
bioassays in prescreening and comparing of study areas. When the biological changes in fishes
from the polluted sites of the River Narva were compared to pollutant background at the study
area, the biological parameters seemed to underestimate the chemical loading in terms of
organic pollution. Differences in the responses of the studied parameters provide future
evidence that assessment of pollutant effects in fishes requires a multiparameter approach.
'Institute of Zoology and Botany, Estonian Agricultural University, 51014 Tartu, Estonia;
Department of Physiology, P.O. Box 1627, University of Kuopio, FIN-70211 Kuopio, Finland.
105
-------
INTRODUCTION
During the past 4 years we have been studying the comparative sensitivity of various
tests indicating aquatic pollution from areas of oil shale mining and processing. Mixed pollution
is characteristic for these areas. The production and utilization of Estonian oil shale have been
reported to be an important source of PAHs, heavy metals, as well as phenols and sulphates
(Veldre et al. 1979, Liblik and Ratsep 1994, Trapido et al. 1995, Tuvikene et al. 1997,
Huuskonen et al. 1998b). While chemical pollution in aquatic ecosystems often occurs as a
complex mixture, the prediction of the effects of pollution may be difficult and may require the
use of multiple biomarkers. Relatively few studies have coupled measurements of biochemical
and physiological responses with chemical analyses in measuring the level of exposure of fishes
to chemical contaminants. The advantage of this approach was observed, for example, in the
studies of Goks0yr et al. (1994), Bogovski et al. (1997), Willett et al, (1997), and Collier et al.
(1998).
The fate of PAHs and HMs is largely determined by sorption to the suspended
particulates and sediments. Sorption depends on the characteristics of both the sediments and
the chemicals involved, such as the organic matter content and particle size of the sediment
(Kukkonen 1991). Furthermore, various abiotic and biotic factors may affect bioavailability
and bioaccumulation of the sediment-associated pollutants. PAHs are a major class of organic
contaminants in the aquatic environment and may affect the productivity of organisms (Adams
et al. 1989, Tuvikene 1995). PAHs are hydrophobic and thus rapidly become associated with
particles and are deposited in aquatic sediments. They are degraded slowly in sediments and in
lower animals (Livingstone 1998). In contrary, fishes usually metabolize PAHs rapidly (Djomo
etal. 1996, Livingstone 1998) via different enzymatic biotransformation processes. Another
group of pollutants, HMs, are ubiquitous, readily dissolved in and transported by water, readily
taken up by aquatic organisms, and strongly bound by sulfhydryl groups of protein (Hodson
1988). Sulfhydryl binding changes the structure and enzymatic activities of proteins and causes
toxic effects (Pent and Stegeman 1993).
Biomarkers can be powerful tools in detecting the exposure of sublethal concentrations
of xenobiotics in fishes. However, evaluation of the results should be done with caution.
Van der Cost et al. (1998) suggested the use of a biotransformation index (BTI), the ratio
between phase I and phase II activities, to assess effects on the balance between bioactivation
and detoxication. Possible limitations of biomarkers are the lack of specif icity, low sensitivity
and/or the development of resistance. As alternatives for in vivo animal studies, fish cell
cultures are increasingly being used in lexicological research (Hightower and Renfro 1988).
A fish liver cell line, PLHC-1, has been used in studying the cytotoxicity of different types of
contaminants, and CYP1A induction and inhibition caused by PAH-type compounds (Hahn et al.
1993, 1996, Briischweiler et al. 1995, 1996). Comparisons between in vivo and in vitro data are
needed to facilitate the aquatic risk assessment (Huuskonen et al. 1998b).
The health status of fishes living in waters of the highly industrialized oil shale
processing area in northeast Estonia is relatively unknown. Some studies in this Estonian oil
shale basin have shown that the pollution from oil shale processing has a negative impact on
fishes. Palm et al. (1992) demonstrated that the content of blood electrolytes and frequencies
106
-------
of micronuclei in erythrocytes in rainbow trout (Oncorhynchus mykiss) caged in drainage water
of an oil shale mine differed from control fish. In another study rainbow trout caged in these
same waters had high bioaccumulation of PAHs (Tuvikene et al. 1997).
The initial objectives of our studies conducted in the River Narva in northeast Estonia
(Huuskonen et al. 1998b, Tuvikene et al. 1999) were to detect the effects of pollutants with
biomarkers in fishes, and to evaluate the sediment pollution with a PLHC-1 fish cell line and a
midge (Chironomus riparius) test. This paper focuses on comparing the accumulation and
BSAFs of selected PAHs and HMs in roach and perch, discussing the dose-effect relationships,
and shedding light on some aspects of using fish in vivo and in vitro in evaluating the biological
effects of a complex mixture of pollutants.
MATERIALS AND METHODS
The River Narva has a drainage area of 56,200 km2 including some highly industrial
areas, and has an average run-off volume of 400 mVsecond at the mouth of the river. The river
receives its pollution mainly with drainage water from oil shale mines and with leachate from
oil shale ash plateaus of Estonian and Baltic thermal power plants (TPPs). The region receives
pollution from TPPs also via air. Study areas were Mustajogi (proposed reference area), and
Baltic TPP and Riigikula (exposed areas) (Figure 1).
Figure 1. Map of the River Narva, northeast Estonia, with the location of the study areas:
Mustajogi as a reference area, and Baltic TPP and Riigikula as exposed areas.
107
-------
Perch and roach were selected to be the test species because they have widespread
distribution and abundance in Estonian water bodies. Fishes were caught with gillnets during
the fall. Fish tissue (muscle and liver) and sediment contents of selected PAHs and HMs have
been reported in Tuvikene et al. (1999) and Huuskonen et al. (1998b), respectively. To measure
the bioavailability, the biota sediment accumulation factors (BSAFs) were calculated for selected
PAHs and HMs. Data on EROD, AHH, GST, and UDP-GT for feral fish were taken from
Tuvikene et al. (1999). To assess effects on the balance between bioactivation and
detoxification, the biotransformation indices (BTIs), ratios between phase I (EROD, AHH)
and phase II (GST, UDP-GT), were calculated. The biochemical data were tested with a
nonparametric Kruskal-Wallis one-way analyses of variance. The effects of organic lipophilic
compounds in the aquatic sediments were detected by exposing the PLHC-1 ceils to sediment
extracts. Cytotoxicity and EROD data for PLHC-1 cells has been presented in Huuskonen et al.
(1998b).
RESULTS
|
The accumulation of selected groups of PAHs in sediments is presented in Figure 2.
The dominating compounds were the 4-ring PAHs, and this was mainly due to high content of
pyrene. The contents of PAHs in muscle and liver of roach and perch are presented in Figure 3.
There were no big differences in PAH profiles between sediments and fish tissues. The
dominating compounds were pyrene and phenanthrene. The total contents of PAHs in the
muscles and liver of perch were higher at exposure sites Baltic TPP and Riigikiila (Figure 3B
and 3D). On the contrary, the total content of PAHs in muscles of roach was highest at the
reference site Mustajogi (Figure 3A). In most cases fish liver accumulated more PAHs than
muscles.
PAHs in sediment
100-
I 10-
I
0.1 -
Mustajogi Baltic TPP Riigikula
Figure 2. Polycyclic aromatic hydrocarbon (PAH) groups in sediments of the River Narva.
(3-Ring: phenanthrene, anthracene: 4-Ring: fluoranthene, pyrene, benzo(#)anthracene,
chrysene: 5-Ring: benzo(e)pyrene, benzo(6)fiuoranthene, benzo(a)pyrene).
108
-------
A. PAHs in roach muscle
B. PAHs in perch muscle
2500
2000 •
1500-I
j
1000 •
500
4-ring i
5-ring
Mustajogi Baltic TPP Riigikula
tuuu •
3500
3000
2500
2000 -
1500-
1000 -
500 -
0 -
1
sL i
I
Mustajogi Baltic TPP RiigikCIa
C. PAHs in roach liver
Mustajogi Baltic TPP
D. PAHs in perch liver
9000 T
8000 -j
7000-1
6000
5000
4000 j
3000 •
2000-
1X0 -
Mustajogi Baltic TPP RiigikUla
Figure 3. Polycyclic aromatic hydrocarbon (PAH) ring groups
in muscle and liver tissues of roach and perch.
According to the calculation of BSAFs, the 3- and 4-ring PAHs seemed to be more
bioavailable than 5-ring (Figure 4). The BSAFs for roach were decreased at the most polluted
sites. However, BSAFs for 4-ring PAHs in perch were slightly increased at the most polluted
sites.
109
-------
A. Biota/sodimsnt accumulation factors in roach
3-nng
[ ssa 4-ring
j K:;;;;;J S-ring
Mustajogi Baltic TPP RiigikUia
B. Biota/sediment accumulation factors in perch
40-
Mustajogi Baltic TPP RiigikUia
Figure 4. Biota-sediment accumulation factors (BSAFs) for PAH groups measured
in livers of (A) roach and (B) perch from the River Narva.
There were no exact trends in HM concentrations in fish between study sites (not
illustrated). No remarkable differences were seen in BSAF of HMs between roach and perch
(Table 1). BSAFs for HMs showed that the most bioavailable metal was mercury and the least
cadmium. Also copper seemed to be very little bioavailable for fish. This was the trend
especially at more polluted sites (Baltic TPP and Riigikula), compared to the upstream reference
area (Mustaj5gi).
Table 1. Biota-sediment accumulation factors (BSAFs) of selected metals from muscles of
roach and perch sampled at different locations in the River Narva.
Metal
Cadmium
Copper
Mercury
Lead
BSAF ([Fish imiscle]/[sediment])
Mustajogi
Baltic TPP
Perch
1.4x10"
0.15
29.1
0.06
2.2 x 10
4.2 x
21.5
0.007
,-5
RiigikUia
Cadmium
Copper
Mercury
Lead
5.8 x 10'5
0.4
15.8
0.003
Roach
l.lxlO'6
6.7 x 1Q-4
13.5
0.019
4.9 xlO"5
3.1X10-4
5.9
0.005
9.7 x lO'5
3.5 xlO"4
4.1
0.01
110
-------
The trend of CYPlA-dependent hepatic EROD and AHH activities in perch and roach
varied, but mostly showed decreased values at exposed areas (Table 2). GST activities were
increased at exposed sites (Table 2). The BTI of perch calculated as a ratio between EROD and
UDP-GT activities increased at exposed sites (Figure 5B). The other BTIs for perch did not
show any trends. On the contrary, for roach all the calculated BTIs showed lower values at more
polluted sites (Figure 5A). The most potent sediments to show EROD activity in the PLHC-1
cells were collected from Riigikula where the total PAH content was also the highest (Table 3).
Mustajogi sediments also contained bioactive compounds. Riigikula and MustajSgi sediment
extracts also showed cytotoxicity at higher doses of sediment extracts (not shown).
Table 2. Biotransformation enzyme activities in roach and perch from the River Narva.
(Data are modified from Tuvikene et al 1999).
Site
EROD
AHH
UDP-GT
GST
N
Mustajogi
Baltic TPP
K-W
MustajSgi
Baltic TPP
Riigikula
K-W
4.07±9.82
1.05±1.12
0.076
34.1±13.1
32.3±20.2
61.9±37.6
0.163
Roach
8.41±3.53
5.64±2.00
0.002
Perch
7.79±4.52
3.49*1.38
5.72±2.69
0.026
116±62
124+69
0.223
174±33
113±46
143±61
0.068
518±173
661 ±168
0.031
239±73
250+83
379±115
0.052
19
21
5
10
6
EROD = 7-ethoxyresorufin 0-deethylase (pmol/min/mg protein); AHH = arylhydrocarbon hydroxylase (pmoymin/
mg protein); UDP-GT = UDP-glucuronosyltransferase (pmol/min/mg protein); GST = glutathione S-transferase
(nmol/niin/mg protein); K-W = Kruskal-Wallis one-way analysis of variance
A. BTI in roach
0.08 T-
0,07 -
0.06 •
0.05 •
JJJO.04-
I
0.03 -i
0.02 -;
i
0.01 4
0,45
B. BTI in perch
0,40 -
0,35 -
0,30 -
0,25-
' 0,20 •
0,15 -
0,10 -
0,05
0,00
Mustajogi Baltic TPP
Mustaogi Baltic TPP Riigikula
Figure 5. Biotransformation indexes (BTIs) for (A) roach and (B) perch from the River Narva.
Ill
-------
Table 3. Maximum EROD activity in PLHC-1 cells exposed to River Narva sediment extracts,
and the dose where it was reached, as well as PAH content in the sediments. (Data
adapted from Huuskonen el a!. 1998b)
Collection
Site
Maximum EROD
(pmol/min/mg protein)
Mustajogi
Baltic TPP
Riigikiila
5
19
1
1
20
Dose at maximum EROD
(extract from mg dry
sediment in ml medium)
Total PAH content
(ng/g dry weight)
80
58
79
81
31
52
69
744
DISCUSSION
The oil shale mining and processing area in northeast Estonia is polluted with PAHs and
other compounds. In the present study, roach and perch from the River Narva were heavily
contaminated with PAHs. According to Neff (1985), the relative concentration of PAHs in
aquatic ecosystems are generally highest in the sediments, intermediate in aquatic biota, and
lowest in the water. Since the elimination of PAHs by fishes is very efficient, no
bioaccumulation of these compounds has generally been demonstrated. PAH levels alone in
fishes are, therefore, not indicative of the levels to which the fishes were exposed and cannot be
used as bioaccumulation markers for exposure assessment. Based on this knowledge the
accumulation of PAHs in fishes from the River Narva was much higher than would have been
expected from the content of PAHs in the water which has been reported to range from 42 to
1,400 ng/L (Tuvikene et a/., unpublished data) or from 52 to 755 ng/g (dw) in sediment
(Huuskonen et a/. 1998b). Concentrations of PAHs in fishes in the River Narva were generally
higher than those reported for fishes at other polluted areas, for example in hardhead catfish
(Ariusfelis) and Atlantic croaker (Micropogonias undulatus) from Galveston Bay Texas USA
(Willett et al. 1997).
Van der Cost et al. (1998) did not observe differences in contaminant levels, or in levels
and activities of biochemical parameters in carps (Cyprimts carpio) which were caged with or
without sediment contact. They concluded that the uptake through sediment is of minor
importance and that the main route of uptake for carp was via water and food. In the present
study, roach collected from a reference site accumulated more PAHs than those collected from
exposed sites. Differences in bioaccumulation of PAHs between study sites could be explained
with the difference in bioavailability. The reference sediments had lower organic carbon content
than the exposed sediments, and the particle size of the reference sediment was bigger
(Huuskonen et al. 1998b). This suggests that the hydrophobic PAHs were more bioavailable
112
-------
at the reference site than at the exposed sites. Bioavailability of PAHs may have a bigger role
especially with benthic feeders, such as roach. Bioaccumulation in different fish species may
also be partly affected by feeding habitats.
Biaccumulation of HMs in fishes from the River Narva was relatively low. The most
bioavailable HM was Hg, but Cd and Cu had limited bioavailability. Many factors may affect
the toxicity and bioavailability of HMs. For example, increased hardness can reduce the toxicity
of Cd. Part et al. (1985) studied the availability of Cd to perfused rainbow trout gills, and found
a strong inverse relationship between the external calcium concentration and the Cd transfer via
gills. Water of the River Narva has relatively high alkalinity and hardness. This may cause the
limited bioavailability of Cd to fishes of the river.
There was a trend towards decreased AHH activities in exposed fishes in the River
Narva. Van der Oost et al. (1994) reported similar observations in feral roach from lakes in
Amsterdam. In that study, none of the phase I enzymes (total CYP, EROD) was induced in the
liver of roach from the polluted sites, and the CYP was inhibited from the more polluted lake
also. The suggestion was that the pollutant levels in the lakes that were studied were most likely
too low to cause a significant induction of hepatic MO enzymes. Another explanation was that
the control fish may have been exposed to MO-inducing xenobiotics. However^ in the present
study the PAH contents were high. ,
Reason for the lack of clear CYP IA induction in fish from the River Narva, in spite of
the heavy loading of PAHs, was suggested to be the elevated content of heavy metals, some of
which can act as inhibitors for MOs. In the study by George (1989), Cd injection into plaice
(Pleuronectes platessd) strongly reduced CYP1A dependent EROD activity. This was proposed
to be mainly due to a decrease in enzyme protein rather than a direct inhibition of activity. In
our study the Cd content was approximately one order of magnitude lower, but due to chronic
exposure it may have had some inhibitory effect. Viarengo et al. (1997) observed that the
EROD activity in the liver of bass (Dicentrarchus labrax), which had previously been induced
by in vitro treatment with^-naphthofiavone or benzo(«)pyrene, was significantly inhibited by
nano- to micro-molar concentrations of ionic Cu, or Hg, or methyl Hg, while treatments with
mixtures of these compounds had additive effects. In the present study, the most bioavailable
HM was Hg and chronic exposure may have had some influence on the xenobiotic metabolism.
Another possibility is that through adaptation processes the fish had lost some of their sensitivity
to PAHs. That kind of adaptation for xenobiotics was described by Stegeman et al. (1992) and
Wall era/. (1998).
From BTIs of perch, the ratio between EROD and UDP-GT showed the clearest trend,
being larger at exposed sites than at the reference site. According to Van der Oost (1998), the
elevated BTI showed the ratio between bioactivation and detoxification processes. All
calculated BTIs in roach were lower at exposed sites. This may be explained with the finding
that PAHs at the reference site were more bioavailable than those at exposed sites. The possible
obstruction of MO induction has to be taken into account, too.
113
-------
The most potent sediments for the PLHC-1 cells were collected from Riigikula. In
addition, the total PAH content in the sediment was highest at that site. Even more, in vitro
studies with the PLHC-1 cells showed that the sediments from the proposed reference area
(MustajSgi) were cytotoxic and caused EROD induction. This indicates that the MustajSgi site
may not have been the best choice for a reference site. The cells proved to be sensitive for the
lipophilic compounds in the sediment extracts. The use of PLHC-1 bioassays in comparing
different water areas and in detecting the potential hazard of PAH-type pollution has been
suggested in other studies as well (Huuskonen et al 1998a, 1999). A limitation of cellular
assays is the difficulty of extrapolating results from these to changes at a higher organizational
level. Such extrapolation can often lead to incorrect evalutaions of environmental pollution.
CONCLUSIONS
Despite of the relatively moderate levels of contamination of sediments of the River
Narva, significant amounts of PAHs tended to accumulate in fishes. This accumulation can be
explained by low metabolism through MOs. Based on our present fish data, the only responding
biomarker was the GST activity. Indeed, when the biological changes of fishes in waters of the
River Narva were compared to the pollution background in the study areas, the biological
parameters seemed to underestimate the chemical loading in terms of PAHs. The PLHC-1 cells,
instead, were more sensitive to showing the EROD-inducing capacity and cytotoxicity caused by
PAH-type compounds. However, since only lipophilic sediment extracts were studied in the in
vitro cell culture tests, tests providing data of the effects of other pollutants, including the effects
of HMs, would have given a more holistic view of the hazards of contaminants.
This study shows that a multibiomarker-based approach provides better information than
that obtained using a single biomarker, especially in the case of mixed contamination. Future
studies in the River Narva are needed to confirm the observed associations between chemical
contamination and biomarker responses.
ACKNOWLEDGEMENTS
This study was supported in part by the Estonian Science Foundation (Grant No. 3694).
REFERENCES
Adams, S.M., K.L. Shepard, M.S. Greeley Jr., B.I). Jimenez, M.G. Ryon, L.R. Shugart, J.F.
McCarthy, and D.E. Hinton. 1989. The use of bioindicators for assessing the effects of
pollutant stress on fish. Marine Environmental Research 28:459-464.
114
-------
Bogovski, S., I. Veldre, B. Sergejev, V. Muzoka, and S. Karlova. 1997. Dynamics of
carcinogenic pollution in Estonian coastal waters. In: Problems of Contemporary
Ecology. Temporal Changes in Estonian Nature and Environment. Short
Communications of Estonian VII Conference in Ecology., T. Frey (Ed.). Tartu,
May 8-9,1997: pp 13-18.
Bruschweiler, B.J., F.E. Wiirgler, and K. Pent. 1995. Cytotoxicity in vitro of organotin
compounds to fish hepatoma cells PLHC-1 (Poeciliopsis lucida). Aquatic Toxicology
32:143-160.
Bruschweiler, B.J., F.E. Wurgler, and K. Pent 1996. An ELISA assay for cytochrome
P4501A in fish liver cells. Environmental Toxicology and Chemistry 15:592-596.
Collier, T.K., L.L. Johnson, C.M. Stehr, M.S. Myers, and I.E. Stein. 1998. A comprehensive
assessment of the impact of contaminants on fish from urban waterway. Marine
Environmental Research 46:243-247.
Djomo, J.E., P. Garrigues, and J.F. Narbonnes. 1996. Uptake and depuration of polycyclic
aromatic hydrocarbons from sediment by the zebrafish (Brachydanio rerio).
Environmental Toxicology and Chemistry 15:1177-1181.
Pent, K., and JJ. Stegeman. 1993. Effects of tributyltin in vivo on hepatic cytochrome P450
forms in marine fish. Aquatic Toxicology 24: 219-240.
George, S.G. 1989. Cadmium effects on plaice liver xenobiotic and metal detoxication
systems: dose-response. Aquatic Toxicology 15:303-310.
Goks0yr, A., J. Beyer, A.-M. Hus0y, H. Larsen, K. Westrheim, S. Wilhelmsen, and J. Klungs0yr.
1994. Accumulation and effects of aromatic and chlorinated hydrocarbons in juvenile
Atlantic cod (Gadus morhua) caged in polluted fjord (S0rfjorden, Norway). Aquatic
Toxicology 29: 21-35.
Hahn, M.E., T.M. Lamb, M.E. Schultz, R.M. Smolowitz, and J.J. Stegeman. 1993.
Cytochrome P4501A induction and inhibition by 3,3',4,4'-tetrachlorobiphenyl in an Ah
receptor-containing fish hepatoma cell line (PLHC-1). Aquatic Toxicology 26:185-208.
Hahn, M.E., B.L. Woodward, J.J. Stegeman, and S.W. Kennedy. 1996. Rapid assessment of
induced cytochrome P4501A protein and catalytic activity in fish hepatoma cells grown
in multiwell plates: response to TCDD, TCDF, and two planar PCBs. Environmental
Toxicology and Chemistry 15:582-591.
Hightower, L.E., and J.L. Renfro. 1988. Recent applications offish cell culture to
biochemical research. Journal of Experimental Zoology 248:290-302.
115
-------
Hodson, P.V. 1988. The effect of metal metabolism on uptake, disposition, and toxicity in
fish. Aquatic Toxicology 11:3-18.
Huuskonen, S., K. Koponen, O. Ritola, M. Hahn, and P. Lindstrom-Seppa. 1998a. Induction
of CYP1A and porphyrin accumulation in fish hepatoma cells (PLHC-1) exposed to
sediment or water from a PCB-contaminated lake (Lake Keraaala, Finland). Marine
Environmental Research 46:379-384.
Huuskonen, S.E., T.E. Ristola, A. Tuvikene, M.E. Hahn, J. V.K. Kukkonen, and P. Lindstrom-
Seppa. 1998b. Comparsion of two bioassays, a fish liver cell line (PLHC-1) and midge
(Chironomus riparius), in monitoring freshwater sediments. Aquatic Toxicology
44:47-67.
Huuskonen, S.E., A. Tuvikene, M. Trapido, K. Pent, and M.E. Hahn. 1999. Cytochrome
P4501A induction and porphyrin accumulation in PLHC-1 fish cells exposed to sediment
and oil shale extracts. Archives of Environmental Contamination and Toxicology.
(In Press)
Kukkonen, J. 1991. Effects of dissolved organic material in fresh water on the binding and
bioavailability of organic pollutants. Doctoral thesis. University of Joensuu: 39 pp.
j
Liblik, V. and A. Ratsep. 1994. Pollution sources and the distribution of pollutants. In: The
Influence of Natural and Anthropogenic Factors on the Development of Landscapes.
The Results of a Comprehensive Study in NE Estonia., J.-M. Punning (Ed.). Institute
of Ecology, Estonian Academy of Sciences: pp 70-93.
!
Livingstone, D.R. 1998. The fate of organic xenobiotics in aquatic ecosystems: quantitative
and qualitative differences in biotransformation by invertebrates a.nd fish. Comparative
Biochemistry and Physiology 120A:43-49.
Neff, J.M. 1985. Polycyclic aromatic hydrocarbons. In: Fundamentals of Aquatic
Toxicology. G.M. Rand, and S.R. Petrocelle (Eds.). Taylor and Frances Group,
Hemisphere Publishing Corporation, New York: pp 416-454.
i
Palm A., A. Tuvikene, and T. Krause. 1992. Changes of hematological characteristics of the
rainbow trout (Oncorhynchus mykiss Walb.) reared in the mixture of natural and oil-shale
mine drainage water. Proceedings of Estonian Academy of Sciences. Biology 44:69-77.
Part, P., O. Svanberg, and A. Kiessling. 1985. The availability of cadmium to perfused
rainbow trout gills in different water qualities. Water Research 19:427-434.
116
-------
Stegeman, J.J., M. Brouwer, R.T. Di Giulio, L. Forlin, B.A. Fowler, B.M. Sanders, and
P.A. Van Veld. 1992. Molecular responses to environmental contamination: enzyme
and protein systems as indicators of chemical exposure and effects. In: Biomarkers.
Biochemical, Physiological, and Histological Markers of Anthropogenic Stress.
R.J. Huggett, R.A. Kimerle, P.M. Mehrle Jr. and H.L. Bergman (Eds.). Lewis
Publishers, Michigan: pp 253-336.
Trapido, M., Y. Veressinina and R. Munter. 1995. Ozonation of phenols containing in
wastewaters from oil shale chemical treatment. Environmental Technology
16:233-241.
Tuvikene, A. 1995. Responses offish to polycyclic aromatic hydrocarbons (PAHs).
Annales Zoologici Fennici 32:295-309.
Tuvikene, A., S. Huuskonen, S. Roy, and P. Lindstrom-Seppa. 1997. Xenobiotic-
metabolizing enzymes in rainbow trout as water pollution biomonitors. In: Fish
Physiology, Toxicology, and Water Quality. Proceedings of the Fourth International
Symposium, Bozeman, Montana, USA, September 1995. R. V. Thurston (Ed.).
U.S. Environmental Protection Agency, Ecosystems Research Division, Athens,
Georgia, USA. EPA/600/R-97/098, pp. 43-49.
Tuvikene, A., S. Huuskonen, K. Koponen, O. Ritola, U. Mauer, and P. Lindstrom-Seppa.
1999. Oil shale processing as a source of aquatic pollution: monitoring of the
biological effects in caged and feral freshwater fish. Environmental Health
Perspectives. 107:745-752.
Van der Oost, R., L. van Gastel, D. Worst, M. Hanraads, K. Satumalay, F.-J. van Schooten,
H. Heida, and N.P.E. Vermeulen. 1994. Biochemical markers in feral roach (Rutilus
rutilus) in relation to the bioaccumulstion of organic trace pollutants. Chemosphere
29:801-817.
Van der Oost, R., S.C.C. Lopes, H. Komen, K. Satumalay, R. van den Bos, H. Heida, and
N.P.E. Vermeulen. 1998. Assessment of environmental quality and inland water
pollution using biomarker responses in caged carp (Cyprinus carpio): use of a
bioactivation:detoxication ratio as a biotransformation index (BTI). Marine
Environmental Research 46:315-319.
•
Veldre, L, A. Itra, and L. Paalme. 1979. Levels of benzo(a)pyrene in oil shale industry wastes,
some bodies of water in the Estonian S.S.R. and water organisms. Environmental Health
Perspectives 30:211-216.
•--.
Viarengo, A., E. Bettella, R. Fabbri, B. Burlando, and M. Lafaurie. 1997. Heavy metal
inhibition of EROD activity in liver microsomes from the bass Dicentrarchm labrax
exposed to organic xenobiotics: role of GSH in the reduction of heavy metal effects.
Marine Environmental Research 44:1-11.
117
-------
Wall, K.L., K. Jessen-Eller, and J.F. Crivello. 1998. Assessment of various biomarkers in
winter flounder from costal Massachusetts, USA. Environmental Toxicology and
Chemistry 17:2504-2511.
I
Willett, K.L., S.J. McDonald, M.A. Steinberg, K.B. Beatty, M.C. Kennicutt, and S.H. Safe.
1997. Biomarker sensitivity for polynuclear aromatic hydrocarbon contamination in
two marine fish species collected in Galveston Bay, Texas. Environmental Toxicology
and Chemistry 16:1472-1479.
118
-------
PHYSIOLOGICAL STUDIES ON THE TOXICITY OF SILVER TO
FRESHWATER FISHES: IMPLICATIONS FOR ENVIRONMENTAL REGULATIONS
Chris M. Wood1
ABSTRACT
Over the past 6-7 years, a concerted research effort on the toxicology of silver in the
aquatic environment has confirmed the free silver ion (Ag"*) as the agent causing acute toxicity to
freshwater fishes. However, in most natural and polluted environments, the concentration of Ag+
is well below levels causing acute toxicity because most silver is complexed by ligands such as
chloride, thiosulfate, dissolved organic carbon, sulfide, and anionic particles. Current ambient
water quality criteria (AWQC) for silver (U.S. EPA 1980, 1985, B.C. MOELP 1995, CCME
1995), which focus on total silver fail to recognize this critical importance of geochemical
speciation, and place undue emphasis on the importance of hardness (i.e. [Ca2+]) as a protective
agent. The physiological mechanism of toxic action of Ag+to freshwater fishes is a non-
competitive inhibition of active Na+ and Cl" uptake at the gills through blockade of the key
transport enzyme Na+,K+-ATPase. Fishes appear to die from cardiovascular collapse secondary
to ionoregulatory failure. Silver is readily accumulated on the gills and internally (especially in
liver) from both toxic (free Ag+) and non-toxic forms (silver complexes). Ag+uptake occurs via
Na+ channels coupled to H^-ATPase in the apical membranes of gill ionocytes, and basolateral
extrusion is active, via a P-type ATPase. The mechanism of silver uptake from complexes is
unknown. Bioaccumulation of silver in itself does not appear to cause toxicity, and is a powerful
inducer of metallothionein in the liver, emphasizing the importance of distinguishing between
bioavailability for acute toxicity, and bioavailability for bioaccumulation. Further work is
needed to evaluate several studies which suggest that silver may cause chronic toxicity at very
low levels during prolonged exposure. Sufficient knowledge is now available to reformulate
AWQC as geochemical models which will yield more accurate and cost-effective site-specific
criteria based on local water chemistry. Such models should not be based on gill silver-burden,
but rather should predict either the availability of free Ag+ or the inhibition of gill
Na+,K+-ATPase as appropriate endpoints for acute toxicity.
BACKGROUND
The origins, history, distribution, and economic impact of silver in the environment have
recently been reviewed (Eisler 1996, Purcell and Peters 1998,1999). Silver enters surface
waters from natural leaching and from anthropogenic activities such as mining, jewellery and
silverware manufacture, and photographic processing. The anti-microbial benefits of silver
Department of Biology, McMaster University, Hamilton, Ontario, Canada
119
-------
nitrate have long been known in the public health professions, but only since the 1960s has the
presence of silver in the aquatic environment attracted regulatory attention. Initial regulations
focused on drinking water standards, which were set at 50 -100 ug-L'W more for total silver in
most jurisdictions, because orally ingested silver has low toxicity to humans and other mammals.
These relatively high levels had little economic impact, but the situation changed in the late
1970s and early 1980s with the development of Ambient Water Quality Criteria (AWQC) which
were guidelines designed to protect freshwater organisms. At that time, available aquatic
toxicity data were sparse and almost entirely based on nominal rather than measured silver
concentrations. No information existed on the mechanism(s) by which waterborne silver caused
toxicity, and there was little understanding of the chemical forms of silver which caused toxicity.
In general AWQC advocated much lower levels (a few ng-L"1 or less for total silver) than
drinking water standards. This change, together with the development of analytical methods
(e.g. "clean" techniques plus graphite furnace atomic absorption analysis with preconcentration)
which could reliably measure silver concentrations at or below lug-L"1, put steadily increasing
pressure on both silver-discharging industries and sewage treatment plants to improve silver
recovery from effluents, and to reduce silver discharges to the environment. Economic and legal
impacts escalated, and industry started to question the validity of the criteria (Cooley et al. 1988,
Dufficy etal. 1993).
Present AWQC still numerically reflect those implemented 20 years ago. For example,
in the United States, the AWQC (acute criterion) promulgated in 1980 (U.S. EPA 1980) remains
in force, and is written as a numerical equation based on water hardness (in mg-L"1 of CaCO3
equivalents): "e i'72^1™^6^-6-52)'' which yields total silver concentrations of 1.2,4.1, and 13.4
ug-L"1 at hardness values of 50,100, and 200 mg-L"1 respectively. There is no chronic criterion.
In Canada (where acute and chronic values are not separated), the national guideline for
protection of freshwater life is still a total silver concentration of 0.1 ug-L"1 (CCME 1995), the
same number implemented in 1980 (Taylor et al. 1980). Notably, such regulations consider only
the total concentration of the metal (though "translation" to "dissolved" silver based on 0.45 um
filtration has been allowed in the United States since 1995), and not the chemical species present.
In the face of increasing regulatory pressures and with a general acceptance that the knowledge
base on which these silver regulations were founded was inadequate, a concerted research effort
on silver in the environment was started in the early 1990s. This research has been supported by
the photographic industry and government agencies, and has been greatly stimulated by the
annual Argentwn conference series. Much of the background information summarized here can
be found in the Argentum conference proceedings (Andren and Sober 1994-1998).
In this overview paper, I summarize the contributions of my research program and those
of my colleagues in this area, focusing on the mechanism(s) of silver toxicity to freshwater
fishes, the mechanisms and impacts of silver bioccumulation, and the influence of water
chemistry and silver speciation on these processes. Two recent review articles (Hogstrand and
Wood 1998, Wood et al. 1999) provide additional information, and also summarize a parallel but
less extensive research program on seawater fishes. In addition, Eisler (1996) and Ratte (1999)
provide exhaustive summaries of toxicity and bioaccumulation data in fishes as well as a wide
variety of organisms. Over the past 6-7 years our knowledge base has increased markedly, and
we are now close to the point where we can sensibly reformulate environmental regulations
using a new paradigm.
120
-------
IMPORTANCE OF WATER CHEMISTRY AND SILVER SPECIATION
IN SILVER TOXICITY
A critical issue in the regulatory debate has been whether or not it is valid to relate the
toxicity of "total" silver, or even "dissolved" silver, in a natural water sample to laboratory
toxicity test results. The recent availability of aquatic geochemical speciation programs such as
MINEQL+ (Schecher and McAvoy 1994) and MINTEQA2/PRODEFA2 (U.S. EPA 1991) has
been extremely useful in addressing this issue. Most laboratory toxicity tests are performed
using AgNO3, often in simplified synthetic freshwater. Because NO3" is an extremely weak
ligand (log K = -0.3), substantial amounts of free Ag+ will usually be present during such tests.
However, in most natural and polluted waters, both geochemical speciation modelling and the
very few available measurements indicate that free Ag+ levels are extremely small (low ng-L"1
range), because of the presence of stronger ligands such as chloride (log K = 3.3-5.5), miosulfate
(8.8-14.2), dissolved organic carbon (DOC; 8.0-10.0), and sulfide (19.2), as well as anionic
particulates which complex silver in natural waters.
The first indication that only free Ag+, and not silver complexed to various anions
(chloride, thiosulfate, sulfide), was acutely toxic to fishes was provided by LeBlanc et al. (1984),
working on fathead minnows (Pimephales promelas). Our initial toxicity studies with juvenile
rainbow trout Oncorhynchus mykiss (our standard test organism throughout this research
program) confirmed these results (Hogstrand et al. 1996). AgNO3 was at least three orders of
magnitude more toxic than the other forms of silver tested (Table 1).
Table 1. Acute toxicity to juvenile rainbow trout and fathead minnows of different silver salts
expressed as 96-hour LC50 values in ug-L"1 of total silver.
Salt
Rainbow trout
(Hogstrand et al. 1996)
Fathead minnow
(LeBlanc etal 1984)
AgNO3*
Ag(S203)NI'N
Ag2S
12
>100,000
161,000
16
>4,600
>280,000
>240,000
* Only AgNOs yields significant amounts of free ionic Ag in solution.
121
-------
Subsequently, Galvez and Wood (1997) used experimental alterations of water chloride
concentration at constant concentration of total silver, together with geochemical modelling, to
demonstrate that acute toxicity was solely a function of the concentration of free Ag+ (Figure 1).
The 7-day LC50 value obtained in this manner (approximately 30 nmol.L"1 Ag+ or 3.2 ug-L"1 as
free Ag+) was virtually identical to that obtained in standard tests with manipulations of total
silver concentration at several different concentrations of Cl" (see also Figure 2 in Hogstrand and
Wood 1998). In addition, manipulations of concentrations of Ca2+ in water (i.e. hardness)
proved curiously ineffective in protecting trout against Ag+ toxicity in our tests, contrary to the
predictions of the EPA (1980) "hardness" equation, but in agreement with some earlier studies
which were not used in the original formulation of this equation (Goettl and Davies 1978, Davies
et al. 1978, Nebeker et al. 1983). Subsequent independent tests by Bury et al. (1999a)
confirmed these results with respect to both the marked protective effect of Cl" concentration and
the relative ineffectiveness of Ca2+ concentration, and further demonstrated that DOC, which
strongly binds Ag"1", was also extremely protective against silver toxicity in rainbow trout. Again,
these results fitted the "free [Ag+]" hypothesis. In that same study, simultaneous experiments on
fathead minnows reinforced the findings with respect to Ca and DOC concentrations, but did
not show clear protection by Cl". Studies by other laboratories (Erickson et al. 1998, Karen et al.
1999) were in general accord with these results, suggesting that fathead minnows arid rainbow
trout respond differently with respect to Cl" concentration.
999 |JN/1 Cr -
2230 JJM Cl
2800 uM Cr—•-•
ionic [Ag*] (* 10'8 M)
Figure 1. Toxicity curve for 7-day lethality bioassay on juvenile rainbow trout at:
total silver concentration (0.92 umol.L = 100 ug-L"1) but variable total chlorid
fixed
(0.92 umoLL"1 = 100 ug-L'1) but variable total chloride
concentration, the latter being used to vary the level of free [Ag+]. Note that the 7-day
LC50 value is 30 nmol.L"1 as free [Ag+]. (Data from Galvez and Wood 1997).
122
-------
In order to understand the discrepancy between our results and the U.S. EPA "hardness
equation" (U.S. EPA 1980), we re-analysed the original data set (Lemke 1981) upon which the
derivation of this equation was largely based. Our conclusion, elaborated in detail elsewhere
(Hogstrand et al 1996, Galvez and Wood 1997, Hogstrand and Wood 1998)+is that the data were
misinterpreted and that Cl" concentration rather than hardness itself (i.e. Ca2+ concentration) was
the effective protective agent. Furthermore, in the original document (U.S. EPA 1980) those
data available at the time which indicated only a small role for hardness were explicitly
excluded. Comparison of several acute data sets for rainbow trout and fathead minnow (cited
above) with the output of the hardness equation suggests that it is fundamentally flawed, being
underprotective at high hardness and overprotective at low hardness.
PHYSIOLOGICAL MECHANISMS OF TOXIC ACTION
The physiological responses of rainbow trout to AgNO3 at concentrations around the
7-day LC50 level (total Ag -10 ug-L"1; ionic Ag+ ~ 30 nmol.L"1) in moderately hard freshwater
from Lake Ontario have now been extensively documented (Wood et al. 1996a, Morgan el al.
1997, Webb and Wood 1998, McGeer and Wood 1998). Overall, the toxic syndrome appears
remarkably similar to that seen during low pH exposure (Wood 1989). In brief, there is a severe
and progressive loss of Na+ and Cl" from the blood plasma which sets in motion a complex series
of events which eventually kills the fish by circulatory failure. Osmotic imbalance between the
extracellular and intracellular fluid compartments causes a net fluid shift out of the blood plasma
into the tissues. Plasma protein and red blood cell (hematocrit) concentrations increase, the latter
compounded by discharge of stored erythrocytes from the spleen, and there is a marked fall in
blood volume (Table 2). A severe stress response occurs, manifested in several-fold increases in
plasma concentrations of the stress hormone cortisol, of glucose (likely due to mobilization of
other stress hormones, catecholamines), and of ammonia (likely due to increased proteolysis
driven by the stress hormones). Cardioaccelatory and vasoconstrictor actions of catecholamines,
in addition to the greatly increased blood viscosity, likely cause a large rise in blood pressure at
the same time as plasma volume declines to critically low levels.
Table 2. The effects of 6 days exposure to 10 ug-L"1 total silver (~ 1/3 as Ag+) on circulatory
parameters in adult rainbow trout. Mean value ± 1 S.E.M. (N = 7-10).
Parameter
Plasma protein (g/100 ml)
Hemoglobin (g/1 00 ml)
Relative plasma volume (%)
[Spleen weight/Body weight] x 100
Spleen hemoglobin (g/spleen)
Mean value ±1
Control
3.19 + 0.22
4.46 ± 0.66
95.2 ±6.8%
0.46 ±0.06
0.101+0.022
S.E.M.
Ag+
4.65 ±0.19*
7.69 + 0.65*
60.8 ±2.6%*
0.28 ±0.06*
0.044 ±0.014*
* p < 0.05
123
-------
The fish dies of "hypovolemic cardiovascular collapse", but the proximate cause can be
traced back to a "surface active" effect of the free Ag+ ion on or in the gills. None of these
responses occurs when silver is presented as silver thiosulfate (with negligible concentrations
of free Ag+), even at more than three orders of magnitude higher total concentration (Wood et al.
1996b). These physiological findings, therefore, support the arguments made earlier about the
importance of speciation of silver.
This surface action of Ag+ is a highly potent blockade of the active uptake "pumping"
of Na+ and Cl" from the water into the blood plasma by specialized salt transport cells in the gills
(Morgan et al. 1997, Webb and Wood 1998). Normally, active branchial uptake of Na+ and Cl"
slightly exceeds the combined rate of diffusive and urinary effluxes; its elimination results in the
progressive net loss of Na+ and Cl" from the plasma. Again this is very similar to the effects of
low pH (Wood 1989), but unlike environmental acidity, Ag+ does not appear to alter
substantially the passive diffusive effluxes of Na+ and Cl" across the gills. Ag+ also has no effect
on branchial Ca2 fluxes (Wood et al. 1996b), so it likely does not displace Ca2+ from tight
junctions, the mechanism by which acidity is thought to increase passive diffusive effluxes.
Kinetic analyses revealed that the inhibition of Na+ and Cl" uptake is due to non-competitive
inhibition, manifested as a decrease in maximum transport rate with an unchanged vertical
position of the kinetic curve (Figure 2). This represents a loss of transport sites rather than a
change in affinity of the transport system.
Control
3000
Figure 2. Kinetic influx curves showing that silver is a non-competitive inhibitor of Na+ uptake
at the gills of rainbow trout. Trout were exposed for 48 hours to a total silver
concentration of 2 ug-L"1 added as AgNOa, approximately one third of which was ug
calculated to exist as free [Ag+] in solution. (Data from Morgan et al. 1997).
124
-------
Morgan et al. (1997) also demonstrated that this loss of transport sites is due to a potent
in vivo blockade of the key enzyme, Na+, K+-ATPase, which energizes the NaCl "pumping"
functions of the gill salt transport cells. Carbonic anhydrase, another transport-related enzyme
which is present in great excess of requirements in the gills, was only moderately inhibited. In
mammals, it is well known that AgNO3 added in vitro to crude or purified Na+, K+-ATPase is a
potent non-competitive blocker of the enzyme (Nechay and Saunders 1984, Hussain et al.
1994), but this is the first time that inhibition in vivo has been shown. Ferguson et al. (1997)
have shown that Ag+ competitively inhibits the binding of Mg2+ to a key activation locus on the
enzyme, thereby explaining the loss of transport capacity and the phenomenon of non-
competitive inhibition with respect to Na+.
Further evidence that this action is a specific effect of Ag+ and not of other forms of
silver has been provided by experiments on trout exposed to sublethal levels of AgNO3 (McGeer
and Wood 1998, Bury et al. 1999b). Elevations of water hardness again had little protective
effect. However, elevations of water Cl" concentration and DOC concentration protected against
the inhibition of gill Na+, K+-ATPase, branchial Na+ uptake, and plasma NaCl regulation. These
protective effects were related in a concentration-dependent fashion to the reduction in the level
of free Ag+ in the exposure water, and not to the total amount of Ag on the gills. The latter
apparently reflects the fact that silver complexes may accumulate on the gills to high levels
without causing toxicity under certain conditions e.g., when high levels of silver chloride,
silver-DOC, or silver thiosulfate complexes are present.
BIOAVAILABILITY VERSUS BIO ACCUMULATION OF SILVER
This finding emphasizes that for silver, we must distinguish between bioavailability for
acute toxicity (where only the concentrations of free Ag+ is important), and bioavailability for
bioaccumulation (where a variety of different chemical species may contribute). Depending on
speciation, silver may undergo considerable bioaccumulation on the gills and/or internally in the
fish without causing acute toxicity, and alternately, may cause acute toxicity without substantial
bioaccumulation. This reflects the fact that waterborne Ag+ is a "surface-active toxicant" at the
gills, doing its damage on or in the branchial epithelium, whereas internal bioaccumulation may
occur without apparent damage to these key toxic sites. While the major internal fate of
bioaccumulated silver appears to be storage in the liver, considerable amounts may also build up
in gills, kidney, and blood plasma (Hogstrand et al. 1996, Wood et al. 1996a, 1996b, Webb and
Wood 1998, Galvez et al. 1998). In contrast to acute toxicity, bioaccumulation may occur from
(or as) many different forms of waterborne silver, not just the free ion. Indeed, the greatest
internal levels of silver are seen when trout are exposed to extremely high levels of silver
thiosulfate, which is relatively non-toxic (Hogstrand et al. 1996, Woodet al. 1996b). However,
to compare the potential for bioaccumulation of different species of silver on a more equal basis,
Hogstrand and Wood (1998) calculated the rate of accumulation of total silver in the fish liver
per day per ppb (ug-L"1) of external silver concentration in trout exposed for 6-7 days in different
water chemistries. These results, when correlated with speciation analysis, indicated that both
the free Ag+ ion (which is highly toxic) and the neutral dissolved AgCl0(? complex (which is not
125
-------
acutely toxic) are readily taken up, whereas silver thiosulfate complexes are not easily absorbed.
Silver may also be bioaccumulated from the diet, but the actual rates per unit concentration are
many orders of magnitude lower than from the water (Galvez and Wood 1999).
At present, we know nothing about the mechanism(s) by which the "non-rtoxic" forms of
silver, such as the AgCla? complex, cross the gills. One possibility is that the whole complex is
taken up, in such a way as to avoid reaction with branchial Na+, K+-ATPase. If we assume, in
accord with conventional belief, that the cells for active ion uptake account for less than 10% of
total gill surface area (e.g. Perry 1997), then 90% or more is available as a route for passive
diffusive entry without toxic consequence. Neutral lipophilic complexes may be much more
efficient at this than free Ag+. Alternately, Ag+ could be "off-loaded" from ligands such as
chloride and thiosulfate to stronger anionic ligands on the gills, again avoiding reaction with
branchial Na^K^-ATPase. However, it is difficult to see why this latter route would not also
serve as the major entry pathway for waterborne Ag+ directly, which is apparently not the case,
as outlined in the next section.
THE MECHANISM OF BRANCHIAL Ag+ UPTAKE
FOR BIOACCUMULATION AND TOXICITY
Since Ag is the only silver species clearly associated with acute aquatic toxicity as well
as a major source for bioaccumulation, our recent research has focussed on the branchial uptake
mechanism for Ag+(Bury and Wood 1999, Bury et al., In Press). In particular, we have
evaluated to what extent Ag+ uptake follows the Na+ uptake pathway via the branchial ionocytes.
Currently, this is thought to involve inward Na+ entry via specific channels, energized via
outwardly directed H+ pumps (V-type ATPases) in the apical membrane (e.g. Lin and Randall
1991). Basolateral Na+,K+-ATPase (a P-type ATPase) extrudes Na+ from the ionocytes to the
blood, thereby helping to maintain the electrochemical gradient for apical Na+ entry, In these
experiments, which were mainly conducted at sublethal concentrations, very dilute, synthetic
softwater was used so as to maximize the contribution of the free Ag+ and minimize the
contribution of silver complexes. Radioisotopically labelled UOmAg was employed to increase
analytical sensitivity.
I
Silver concentrations in the gills during constant waterborne exposures first increased
then decreased over time, suggesting that transfer across the basolateral membrane of the gill
cells is the rate-limiting step (Bury and Wood 1999). Just as with Na+ (Figure 2), Ag+ uptake
exhibited concentration-dependent, Michaelis-Menten kinetics, and waterborne Na+ acted as a
non-competitive inhibitor of Ag+ uptake (Figure 3). Other cations (K+, Ca24) were ineffective,
but the specific Na+ channel-blocker phenamil (10"4M), and the specific V-type ATPase blocker
bafilomycin AI (2 x lO^M; Figure 4) were both highly effective in reducing Na+ and Ag+ uptake.
These results strongly suggest that Ag+ is taken up across the apical membranes of ion-
transporting cells via the H+-coupled Na+ channels.
126
-------
140-
120-
100-
O 80-
£3
J3 60-
o,
OJD 40-
20-
0-
!3mM
50
100
150
200
Water [Ag] (nM)
Figure 3. Kinetic influx curves showing that Na+ is a non-competitive inhibitor of silver uptake
at the gills of rainbow trout. Trout were acutely exposed to the test solutions for 2 hours.
Silver was present largely as free [Ag+]. (Data from Bury and Wood 1999).
Bury et al. (In Press) used basolateral membrane vesicle preparations (BLMVs) of gills
of rainbow trout to study the transport mechanism from cytosol to blood. The process was
dependent upon ATP and followed Michaelis-Menten kinetics, indicating carriers-mediation.
Ag+ transport across BLMV's was inhibited by sodium orthovanadate (10~4 M), which is
diagnostic of P-type ATPases. Furthermore, Ag+, in the presence of ATP, was capable of
activating an enzyme subunit of the correct size for a P-type ATPase, as shown by the formation
of an acylphosphate intermediate which ran at 104 Kd on a polyacrilamide gel. These results
suggest but do not prove that basolateral transport of Ag+occurs via interaction at the Na site
on Na+,K+-ATPase, the same enzyme which Ag+ non-competitively inhibits at the Mg2+ site.
Alternatively, they may indicate that a different, as yet unidentified, P-type ATPase is
responsible for basolateral Ag+ transport.
127
-------
600 Na+ uptake
T
400 -
o
| 200
Ag* uptake
T
- 25
- 20
-15 |
100), whereas ACR for most metals are usually quite low. For example, the ACR value
is about 3 for copper (U.S. EPA 1985), which is thought to have a similar physiological
128
-------
mechanism of action as silver. At present, there does not appear to be sufficient information on
which to construct reliable chronic criteria. There is an urgent need for new studies in which
total and dissolved silver levels are directly measured and water chemistry is fully characterized,
thereby allowing speciation analysis. If such studies confirm the toxicity of silver at extremely
low levels, then chronic criteria should be developed.
While we have learned a great deal about the mechanisms of acute toxicity of silver to
freshwater fishes in the past 6-7 years, we are only now starting to investigate the mechanisms of
chronic toxicity. In the only physiological study to date, Galvez et al. (1998) detected sublethal
disturbances in plasma Na+ concentration and Ci" concentration which were largely corrected
over a 28-day exposure in juvenile trout exposed to 0.5 or 2.0 ug-L"1 total silver, suggesting a
similar mechanism as seen during acute toxicity (Figure 5).
140
T 130
•5 120
E
E 110
100
130
- 120
i
•5 110
E
E 100
90
Plasma [Na*]
Control
* Ag (2.0
Plasma [CD
CO 5 10 15 20 25 30
Time (Days)
Figure 5. Changes in plasma [Na+] and [Cl~] of juvenile rainbow trout during chronic exposure
to a total silver concentration of 2 ng-L"1 for 28 days. Note the recovery from
ionoregulatory disturbance during prolonged exposure. (Data from Galvez et al. 1998).
129
-------
The fact that large amounts of silver accumulate internally in fish during sublethal
exposures raises the question whether this will cause deleterious chronic effects on growth,
reproduction, or long term survival. At present, the question cannot be answered, but to date we
have no evidence that bioaccumulated silver is in any way harmful to the fish. The only
documented effect is an induction of metallothionein in the gills, kidney, and especially the liver
(Hogstrand et al. 1996, Wood et al. 1996a, 1996b, Galvez et al. 1998). The greater the tissue
burden of total silver, the greater appears to be the metallothionein induction, regardless of the
source of the silver (Figure 6). This low molecular weight, cysteine-rich protein is thought to be
important in the internal immobilization and detoxification of metals. For example, it has been
shown to protect Na+,K+-ATPase against inhibition by silver in vitro (Hussain et al. 1995,
Ferguson et al. 1997). Indeed silver appears to be a more potent inducing agent than other
metals (e.g., Cd, Cu, Zn) to which metallothionein binds, which may help explain why
internalized silver appears to be relatively benign (Hogstrand et al. 1996). Undoubtedly,
metallothionein synthesis must encumber some subtle costs, but it remains to be seen whether
they are important for fitness.
Ag+
AgCl2-
Ag(S203)2*-
AgS203-
0.0 3.8 0.0 0.1 0.0
0.5 6.6 0.0 16.8 0.0
0.0 0.5 0.0 91.2 0.0
0.0 0.0 29431 0.0 95158
0.0 0.0 879 0.0 2842
Figure 6. Total silver (Ag) and metalliothionein (MT) concentrations in liver of juvenile
rainbow trout exposed for 6-7 days to the silver species and concentrations (in ug-L"1)
shown below each pair of bars. The bars on the far left represent non-exposed control
fish. (Data from Hogstrand et al. 1996, and Wood et at. 1996a, 1996b).
130
-------
IMPLICATIONS FOR WATER QUALITY CRITERIA: A NEW PARADIGM.
Current guidelines and criteria vary considerably between jurisdictions, but none appears
to recognize the current state of knowledge in the field. Most are based on total silver
concentration, some greatly (and erroneously) emphasize the modifying role of water "hardness",
and many appear to be idiosyncratic, e.g., the massive "breakpoint" in the allowable silver
concentrations at a hardness of 100 mg-L"1 CaCO3 in the British Columbia criteria (B.C. MOELP
1995). The most important implication from the knowledge acquired over the past few years is
that an understanding of the geochemical speciation of silver in the water is critical to sensible
environmental regulation. Clearly Ag+ is the agent of concern, at least for acute toxicity, and it is
now possible to predict its concentration with reasonable certainty using geochemical models if
local water chemistry is known. An important first step in reforming water quality criteria for
silver in light of current knowledge will be to reformulate the U.S EPA (1980) "hardness"
equation in the form of a geochemical model. In addition to the very minor role of Ca2+
concentration, more important and environmentally relevant variables which directly affect the
availability of free [Ag+] should be incorporated, particularly chloride, DOC, and sulfide. This
approach will allow more accurate and cost-effective site-specific criteria rather than "blanket"
nation-wide criteria.
A promising new development, based on the original ideas of Pagenkopf (1983), is the
incorporation of anionic ligands on or in the fish gill itself into such geochemical models
(Bergman and Dorward-Kmg 1997, Playle 1998). The idea here is that a particular degree of
saturation of "toxic receptor sites" (anionic ligands) on or in the gills by a metal ion will cause
death, and can be predicted if water chemistry and gill chemistry are known. As originally
formulated for silver (Janes and Playle 1995, DiToro et al. 1998), these "Biotic Ligand Models"
considered a particular gill total silver burden as the appropriate endpoint for toxicity. However
as outlined earlier, it is now clear that silver may accumulate on the gills in either a toxic or non-
toxic form. What matters for toxicity is the Ag+ bound to branchial Na+,K+-ATPase. A further
limitation is that such models do not have kinetic components, but rather assume thermodynamic
equilibrium, whereas it is now clear that total gill silver burden changes markedly over time
during constant exposures. However it should be relatively easy to recalibrate the Biotic Ligand
Models to use Ag+ inhibition of gill Na+,K+-ATPase (i.e. saturation of "toxic receptor sites") as
the appropriate endpoint, rather than simply gill total silver burden.
ACKNOWLEDGMENTS
Original research cited here was supported by grants from the Industrially Oriented
Research Program of the Natural Science and Engineering Research Council of Canada, Kodak
Canada, Inc., the National Association of Photographic Manufacturers (now the Photographic
and Imaging Manufacturers Association), and the Silver Coalition. I thank Christer Hogstrand,
Steve Munger, Fernando Galvez, Ian Morgan, Jim McGeer, Nic Bury, Martin Grosell, Nathan
Webb, and Colin Brauner for their many contributions to this research program, and Bob Cappel,
Tom Dagon, and Joe Gorsuch (Eastman Kodak Company) for their long-term support of our
work. Joe Gorsuch, Tom Dagon, and Tom Purcell (the Silver Council) provided valuable
comments on the manuscript.
131
-------
REFERENCES
Andren, A., and T.W. Sober, (Eds.). 1994-98. Argentum: Proceedings of the 1st, 2nd, 3rd, 4th, &
5th International Conferences on Transport, Fate & Effects of Silver in the Environment.
University of Wisconsin Sea Grant Institute, Madison, Wisconsin, USA.
B.C. MOELP. (British Columbia Ministry of Environment, Lands, and Parks). 1995. Ambient
Water Quality Criteria for Silver. Environmental Protection Department, Water Quality
Branch, Victoria, British Columbia, Canada.
Bergman, H.L., and E.J. Dorward-King. 1997. Reassessment of Metals Criteria for Aquatic
Life Protection. Society of Environmental Toxicology and Chemistry, Technical
Publication Series. SETAC Press, Pensacola, Florida, USA.
j
Bury, N.R., F. Galvez, and C.M. Wood. 1999a. Effects of silver chloride, calcium, and
dissolved organic carbon on silver toxicity: comparison between rainbow trout and
fathead minnows. Environmental Toxicology and Chemistry 18:56-62.
Bury, N.R., J.C. McGeer, and C.M. Wood. 1999b. Effects of altering freshwater chemistry on
physiological responses of rainbow trout to silver exposure. Environmental Toxicology
and Chemistry 18:49-55.
Bury, N.R., and C.M. Wood. 1999. The mechanism of branchial apical silver uptake by
rainbow trout is via the proton-coupled Na+-channel. American Journal of Physiology R.
In Press.
Bury, N.R., M. Grosell, A.K. Grover, and C.M. Wood. 1999. ATP-dependent silver
transport across the basolateral membrane of rainbow trout gills. Toxicology and
Applied Pharmacology 159:1-8.
CCME (Canadian Council of Ministers of the Environment). 1995. Canadian Water Quality
Guidelines. Ottawa, Ontario, Canada.
,
Cooley, A.C., T.J. Dagon, P.W. Jenkins, and K.A. Robillard. 1988. Silver and the environment.
Journal of Imaging Technology 14:183-189.
Davies, P.H., J.P. Goettl Jr., and J.R. Sinley. 1978. Toxicity of silver to rainbow trout (Salmo
gairdneri). Water Research 12:113-117.
DiToro, D., P. Paquin, R. Santore, and B. Wu. 1998. Chemistry of silver bioavailability: a
model of acute silver toxicity to fish. In: Proceedings of the 5th International Conference
on Transport, Fate & Effects of Silver in the Environment. A.W. Andren and
T.W. Bober (Eds.). University of Wisconsin Sea Grant Institute, Madison, Wisconsin,
USA pp. 191-202.
132
-------
Dufficy, T.J., R. Cappel, and S.M. Summers. 1993. Silver discharge regulations questioned.
Water Environment & Technology. April:52-56.
Eisler, R. 1996. Silver hazards to fish, wildlife, and invertebrates: a synoptic review.
U.S. National Biological Service, National Biological Service Report 32:1-44
Erickson, R.J., L.T. Brooke, M.D. Kahl, F.V. Venter, S.L. Harting, T.P. Markee, and
R.L. Spehar. 1998. Effects of laboratory test conditions on the toxicity of silver to
aquatic organisms. Environmental Toxicology and Chemistry 17:572-578.
Ferguson, E.A., D.A. Leach, and C. Hogstrand. 1997. Metallothionein protects against silver
blockage of the Na+/K+-ATPase. In: Proceedings of the 4th International Conference on
Transport, Fate & Effects of Silver in the Environment. A. W. Andren and T. W. Bober
(Eds.). University of Wisconsin Sea Grant Institute, Madison, Wisconsin, USA.
Galvez, F., and C.M. Wood. 1997. The relative importance of water hardness (Ca) and chloride
levels in modifying the acute toxicity of silver to rainbow trout. Environmental
Toxicology and Chemistry 16:2363-2368.
Galvez, F., and C.M. Wood. 1999. Physiological effects of dietary silver sulfide exposure in
rainbow trout. Environmental Toxicology and Chemistry 18:84-88.
Galvez, F., C. Hogstrand, and C.M. Wood. 1998. Physiological responses of juvenile rainbow
trout to chronic low level exposures to waterborne silver. Comparative Biochemistry and
Physiology 119C: 131-137.
Goettl, J.P. Jr., and P.M. Davies. 1978. Water pollution studies. Job 6. Study of the effects of
metallic ions on fish and aquatic organisms. Job Progress Report for Federal Aid Project.
F-33-R-13. Colorado Division of Wildlife, Fort Collins, Colorado, USA p.6.
Hogstrand, C., F. Galvez, and C.M. Wood. 1996. Toxicity, silver accumulation and
metallothionein induction in freshwater rainbow trout during exposure to different silver
salts. Environmental Toxicology and Chemistry 15:1102-1108.
Hogstrand, C., and C.M. Wood. 1998. Toward a better understanding of the bioavailability,
physiology, and toxicity of silver in fish: implications for water quality criteria.
Environmental Toxicology and Chemistry 17:547-561.
Holcombe, G.W., G.L. Phipps, and J.T. Fiandt. 1983. Toxicity of selected priority pollutants to
various aquatic organisms. Ecotoxicology and Environmental Safety 7:400-409.
Hussain, S., R.M. Anner, and B.M. Anner. 1995. Metallothionein protects purified
Na+,K+-ATPase from metal toxicity in vitro. In Vitro Toxicology 8:25-30.
133
-------
Hussain, S., E. Meneghini, M. Moosmayer, D. Lacotte, and B.M. Anner. 1994. Potent and
reversible interaction of silver with pure Na+,K+-ATPase and Na^,K+-ATPase liposomes.
Biochimica Biophysica Acta 1190:402-408.
Janes, N.} and R.C. Playle. 1995. Modeling silver binding to gills of rainbow trout.
Environmental Toxicology and Chemistry 14:1847-1858.
Karen, D.J., D.R. Ownby, B.L. Forsythe, T.P. Bills, T.W. La Point, G.P. Cobb, and S.J. Klaine
1999. Influence of water quality on silver toxicity to rainbow trout (Oncorhynchus
mykiss), fathead minnows (Pimephalespromelas) and water fleas (Daphnia magna).
Environmental Toxicology and Chemistry 18:71-83.
LeBlanc, G.A., J.D. Masone, A.P. Paradice, B. Wilson, B. Lockhart, and K.A. Robillard. 1984.
The influence of speciation on the toxicity of silver to fathead minnow (Pimephales
promelas). Environmental Toxicology and Chemistry 3:37-46.
Lemke, A.E. 1981. Interlaboratory comparison: acute testing set. Office of Pesticides and
Toxic Substances, U.S. Environmental Protection Agency, Washington, D.C., USA.
EPA 600/3-81-005.
Lin, H., and D.J. Randall. 1991. Evidence for the presence of an electrogenic proton pump on
the trout gill epithelium. Journal of Experimental Biology 161:119-134.
McGeer, J.C., and C.M. Wood. 1998. Protective effects of water Cl'ancl physiological
responses to waterborne silver in rainbow trout. Canadian Journal of Fisheries and
Aquatic Sciences 55:2447-2554.
Morgan, I.J., R.P. Henry, and C.M. Wood. 1997. The mechanism of acute silver nitrate toxicity
in freshwater rainbow trout (Oncorhynchus mykiss) is inhibition of gill Na+ and Cl"
transport. Aquatic Toxicology 38:143-163.
Nebeker, A.V., C.K. McAuliffe, R. Mshar, and D.G. Stevens. 1983. Toxicity of silver to
steelhead and rainbow trout, fathead minnows, and Daphnia magna. Environmental
Toxicology and Chemistry 2:95-104.
Nechay, B.R., and J.P. Saunders. 1984. Inhibition of adenosine triphosphatases in vitro by
silver nitrate and silver sulfadiazine. Journal of the American College of Toxicology
3:37-42.
1 ' !
Pagenkopf, G.K. 1983. Gill surface interaction model for trace metal toxicity to fishes: role of
complexation, pH, and water hardness. Environmental Science and Technology
17:342-347.
134
-------
Perry, S.F. 1997. The chloride cell: structure and function in the gills of freshwater fishes.
Annual Review of Physiology 59:325-347.
Playle, R.C. 1998. Modelling metal interactions at fish gills. Science of the Total Environment
219:147-163.
Purcell, T.W., and J.J. Peters. 1998. Sources of silver in the environment. Environmental
Toxicology and Chemistry 17:539-546.
Purcellj T.W., and J.J. Peters. 1999. Historical impacts of environmental regulation of silver.
Environmental Toxicology and Chemistry 18:3-8.
Ratte, H.T. 1999. Bioaccumulation and toxicity of silver compounds: a review. Environmental
Toxicology and Chemistry 18:89-108.
Schecher, W.D., and D.C. McAvoy. 1994. MINEQL+. User's Manual, Version 3.01.
Hallowell, Maine, USA.
Taylor, M.C., A. Demayo, and S. Reeder. 1980. Guidelines for Surface Water Quality.
Inorganic Chemical Substances. Silver. Environment Canada, Inland Waters
Directorate, Water Quality Branch, Ottawa, Volume 1. 14pp.
U.S. EPA (United States Environmental Protection Agency). 1980. Ambient Water Quality
Criteria for Silver - 1980. Office of Water Regulations and Standards, Criteria and
Standards Division, Washington, D. C., USA. EPA-440/5-80-071.
U.S. EPA (United States Environmental Protection Agency). 1985. Ambient Water Quality
Criteria for Copper -1984. Office of Water Regulations and Standards, Criteria and
Standards Division, Washington, D.C., USA. EPA 440/5-84-031.
U.S. EPA (United States Environmental Protection Agency). 1991. MINTEQA2/PRODEFA2.
A Geochemical Assessment Model for Environmental Systems: Version 3.0 User's
Manual. Center for Exposure Assessment Modelling, U.S. Environmental Protection
Agency, Athens, Georgia, USA. EPA/600/3-91/021.
Webb, N.A., and C.M. Wood. 1998. Physiological analysis of the stress response associated
with acute silver nitrate exposure in freshwater rainbow trout (Oncorhynchus mykiss).
Environmental Toxicology and Chemistry 17:579-588.
Wood, C.M. 1989. The physiological problems offish in acid waters. In: Acid Toxicity and
Aquatic Animals. R. Morris, E.W. Taylor, D.J.A.Brown, and J.A. Brown, (Eds.).
Society of Experimental Biology Seminar Series. Cambridge University Press,
Cambridge, U.K
135
-------
Wood, C.M., C. Hogstrand, F. Galvez, and R.S. Munger. 1996a. The physiology of waterborne
silver toxicity in freshwater rainbow trout (Oncorhynchus mykiss). 1. The effects of
ionic Ag+. Aquatic Toxicology 35:93-109.
Wood, CM., C. Hogstrand, F. Galvez, and R.S. Munger. 1996b. The physiology of waterborne
silver toxicity in freshwater rainbow trout (Oncorhynchus mykiss}. 2. The effects of
silver thiosulphate. Aquatic Toxicology 35:111-125.
Wood, C.M., R.C. Playle, and C. Hogstrand. 1999. Physiology and modeling of mechanisms of
silver uptake and toxicity in fish. Environmental Toxicology and Chemistry 18:71-83.
136
-------
BIOACCUMULATION OF XENOBIOTIC CHEMICALS BY AQUATIC ORGANISMS
Des Cornell1
ABSTRACT
The bioaccumulation of persistent organic chemicals by aquatic organisms can occur
from food or from the ambient water, but the latter source is generally dominant in most
situations. Bioaccumulation of chemicals from water, described as bioconcentration, occurs by
passive diffusion from the ambient water across the gills into the circulatory fluid to be deposited
in lipid tissues. The chemicals most suscceptible to bioaccumulation are the chlorohydrocarbons
and polyaromatic hydrocarbons with log KOW values between 2 and 6.5. Octanol provides a
reasonable surrogate for biota lipid in most situations; thus, bioconcentration behaviour with
aquatic organisms can generally be predicted by relationships with the octanol/water partition
coefficient. Bioaccumulation should be seen as a part of the process of uptake and deposition of
chemicals within an organism leading to the development of toxic effects. The use of the
internal biotic concentration provides an additional perspective on toxicity as compared to
ambient water concentrations, the internal lethal concentrations approach a constant value for
many sets of related compounds providing a basis for the prediction of toxic effects.
INTRODUCTION
Current evidence suggests that some aquatic biota bioaccumulate compounds as a result
of direct partitioning with ambient water, which is usually described by the term
"bioconcentration" (Moriarity 1975). For example, aquatic systems include biota/water
partitioning, filter feeders/water partitioning, and infauna/interstitial water partitioning which can
usually be considered to be direct partition processes involving three phases. These systems
would then be expected to be characterized by partition coefficients which could be used to
interpret the bioconcentration process. On the other hand the transfer of chemicals from low
trophic level aquatic biota to higher trophic level biota cannot be considered to be a simple direct
partitioning process, although partitioning would most likely be involved. This transfer occurs
during the consumption of low trophic level aquatic biota as food by higher trophic level aquatic
biota and is usually described by the term "biomagnification." An example of this would be the
consumption of phytoplankton by some species offish. In many situations the means whereby
an organism acquires a chemical is not known and the term "bioaccumulation" is used.
Both bioconcentration and biomagnification must operate with aquatic organisms.
Air-breathing aquatic mammals, such as whales, dolphins, etc. lack an organism/water exchange
interface, since exchange cannot occur through the skin; thus, biomagnification is the only
School of Public Health, Griffith University, Brisbane, Australia
137
-------
mechanism involved. In contrast, autotrophic organisms, such as phytoplankton and some
bacteria, draw their food, as well as other chemical components, directly from dissolved
substances in the ambient water. With these organisms bioconcentration is the only possible
mechanism. With other aquatic organisms current evidence suggests that bioconcentration is the
dominant mechanism of bioaccumulation.
THE MECHANISM OF BIOCONCENTRATION AND BIOMAGNIFICATION
Baughman and Paris (1981) have reviewed the available information on bioconcentration
and concluded that bioconcentration of lipophilic compounds by microorganisms occurs as a
result of partitioning between water and the microorganism. Various authors, e.g., Sodergren
(1968), Kerr and Vass (1973), and Harding and Vass (1978), have shown that lipophilic
compounds adsorb onto the outer surface and then diffuse internally within the cell. Mortimer
and Connell (1993) investigated bioconcentration by juvenile crabs and suggested that passive
diffusion from water is the process involved with these organisms. With fish, a number of
authors have shown that gills are the site of uptake and partitioning of lipophilic compounds, as
reviewed by Connell (1988). The evidence indicates that the route of uptake and loss of
lipophilic compounds with aquatic organisms is generally through the oxygen uptake route.
The gills are the primary site; then partitioning with the circulatory fluid occurs, resulting in
deposition of the lipophilic compound in biota lipid (Hawker and Connell 1986). Subsequently,
the compound can be metabolised, generally to more oxygenated and water-soluble forms, and
excreted as indicated in Figure 1. The common mechanism of uptake and loss through the
oxygen pathway suggests that all aquatic organisms can be treated similarly as regards the
water/organism partition process. This indicates that biota lipid/water partitioning is effectively
the dominant factor in the physical process resulting in the uptake and accumulation of lipophilic
compounds from water by aquatic organisms (Mackay 1982).
i
Biomagm'fication is a more complex process but can be diagrammatically represented in
a simplified form as shown in Figure 1. In this process lipophilic contaminants in food can be
seen as partitioning with circulatory fluid and being deposited in lipid tissues by routes similar to
the bioconcentration process.
INFLUENCE OF SOME BIOLOGICAL CHARACTERISTICS
ON BIOCONCENTRATION
The biotic lipid phase is the dominant concentrating phase for lipophilic compounds and
the aqueous and other phases are of little significance with aquatic organisms (Mackay 1982).
Thus the lipid content of an organism is a major factor influencing the amount of chemical
bioconcentrated from water. The concentration of bioconcentrated chemical would be expected
to be in direct proportion to the lipid content. Organisms and groups of organisms exhibiting
low lipid content exhibit correspondingly low bioconcentration capacity.
138
-------
Equilibrium. Exchange
wi-
dth Water
Water
Water Uptake
Excretion
Figure 1. Routes of uptake and clearance of persistent lipophilic chemicals by fish.
Apart from the physical process outlined above, the persistence of a chemical in an
organism is a major factor influencing bioconcentration as previously discussed. The uptake of
lipophilic compounds by biota usually results in the induction of mixed function oxidase (MFO)
enzyme systems in the exposed biota. The MFO systems stimulate oxidation of lipophilic
compounds, leading to their removal from the organism. Not all organisms have the same
capacity to respond to exposure to lipophilic compounds in this way. In fact aquatic organisms
can show considerable variation in MFO activity between species (e.g. Connell and Miller 1981).
This may relate to particular metabolic characteristics of the species or to factors such as
previous exposure patterns to lipophilic compounds. Such differences in MFO activity may lead
to different bioaccumulation characteristics for different species.
The habitat of an aquatic organism, or group of organisms, has an effect on
bioaccumulation characteristics (Cullen and Connell 1992). Of particular importance are aquatic
infauna, which are fauna which reside in bottom sediments in aquatic areas, including aquatic
worms and various other organisms. Bioaccumulation with this group can be best understood by
a three-phase model. This indicates that there are two partition processes involved in
bioaccumulation within the sedimentary system. Compounds are released from the sediment by
the sediment/interstitial water partitioning system and then there is uptake and concentration by
the interstitial water/infauna partition process. The interstitial water/infauna process is
bioconcentration, which parallels the water/fish process and would be expected to exhibit similar
characteristics.
139
-------
QUANTITATIVE STRUCTURE-ACTIVITY RELATIONSHIPS
FOR BIOCONCENTRATION
I
Bioconcentration is characterised by the bioconcentration factor (KB), which is the ratio
of the concentration of the compound in biota (CB) to the concentration in water (Cw). The
process is complex but largely governed by the biota lipid/water partition process, and as a result
it has been found that the most useful characteristic for the prediction of the bioconcentration
factor is the n-octanol/water partition coefficient (KoW). Despite the simplicity of the
octanol/water system as compared with the bioconcentration process, the Kow value provides a
reasonably good estimate of the bioconcentration capacity of lipophilic compounds within
certain limits. The success of the system is largely dependent on the similarity of octanol to
biota lipid since water is common to both systems. Other parameters, including aqueous
solubility and Randic Indices, have also been used to develop bioconcentration quantitative
structure activity relationships (QSARs), but usually these are less successful.
I
Compounds which have log KoW values less than 2 are usually bio concentrated more than
would be expected from the KQW values. This is a result of nonlipoid tissue becoming
increasingly important due to the decreasing lipophilicity and increasing water solubility of these
compounds. On the other hand compounds with log KoW greater than 6.5 bioconcentrate to a
lesser extent than would be expected from the KOW values (Connell and Hawker 1988). With
these compounds the biota lipid and octanol differ in their solubility properties, resulting in a
reduction in the solubility of these compounds in lipid and a reduced bioconcentration factor.
Other properties, such as the size of the molecule, may also be factors causing reduced
bioconcentration.
A QSAR for bioconcentration can be derived as follows. If octanol is a perfect surrogate
for biota lipid, then
KB =KoWYL
and
log KOW = 1 log KOW + log YL
where YL is the fraction of lipid present in the biota.
i
The empirical results obtained from bioconcentration experiments are usually expressed
in the following general form:
log KB = alogKow+=£
and
where a and b are empirical constants.
140
-------
Based on the outline above, constant a is an empirical constant expressing the
nonlinearity of the relationship and indicates how well octanol represents the biota lipid, and
b = log YL. This indicates that under perfect conditions YL is always less than unity, which
means that constant b should always be negative. Some of the empirical relationships, generally
derived from experiments done in aquaria, are shown in Table 1.
Table 1. Characteristics of some relationships between log KB and log KQW for various biota.1
Biota
Microorganisms
Daphnids
Polychaetes and
oiigochaetes
Fish
Fish
Fish
Fish
Molluscs
Constant
a
0.91
0.90
0.99
0.94
0.98
1.00
0.95
0.84
b
-0.36
-1.32
-0.60
-1.00
-1.36
-1.32
1.06
-1.23
Lipid equivalent
to constant b
(% wet weight)
44
4.8
25
10
4.3
4.8
8.8
5.9
Actual lipid
(% wet weight)
n.a.
n.a.
n.a.
1-16
1-16
1-16
1-16
1.2-1.8
Range of
log Kow
3-7
2-8
4-8
3-6
1.5-6.5
0.5-6.0
2-6
3.5-8
1 Source: Cornell (1990). Note: Log KB = a log KOW + b. n.a., Data not available.
There is a large volume of data available on fishes which is reflected by the data
tabulated in Table 1. Mackay (1982), Davies and Dobbs (1984), Connell and Hawker (1988),
Connell and Schuurmann (1988), and Schuurmann and Klein (1988) have collected and collated
sets of data and in some cases evaluated the accuracy and application of this material to the
bioconcentration relationships. These relationships in most cases were derived from the use of
chlorohydrocarbons and polyaromatic hydrocarbons. Kenaga and Goering (1980) reported that
the lipid content offish ranges from 1% to about 16% depending on a variety of factors, which is
in accord with the lipid equivalent to constant b shown in Table 1. The results suggest that with
effectively nonbiodegradable compounds, principally the chlorohydrocarbons and polyaromatic
hydrocarbons, octanol provides a reasonable representation of biota lipid for compounds with log
KOW values between 2 and 6.5.
141
-------
LETHAL AND SUBLETHAL RESPONSES TO BIOCONCENTRATED COMPOUNDS
It is important to address the biological effects of compounds which bioconcentrate in
aquatic biota. This area has not received a great deal of attention in bioconcentration research.
The standard method for assessing the lethal toxicity of chemicals to aquatic organisms involves
exposure of the organism to aqueous solutions of the toxic material, and periodic observations of
effects. The results are usually expressed as the LC50. However, the toxic effect occurs within
the tissues; consequently, the chemical must be taken up from the aqueous medium and retained
by the organism before the toxic effect can occur. With persistent lipophilic compounds, uptake
and retention occur by the bioconcentration process (Connell, 1988). Accordingly, it would be
expected that there would be systematic relationships between exposure to an aqueous solution
of a chemical, bioconcentration, concentration of the chemical in the biotic tissue, and toxicity.
The concurrent investigation of these factors would be expected to present a clearer evaluation of
the toxic effects of a chemical.
Nonspecific, narcotic, or nonreactive toxicity occurs with chemicals which do not exert a
toxic effect through a biochemical process resulting in the breaking or formation of covalent
bonds. The critical concentration hypothesis is based on chemicals exhibiting this toxicity
mechanism, and the observation that bioconcentrated chemicals seem to reach a specific and
fixed internal concentration, on a molar basis, in lipid tissues at which lethal toxicity occurs.
Generally the lethal effect is expected to occur at a consistent tissue level, with all chemicals
which exhibit nonspecific toxicity, regardless of the time taken to reach that level under different
conditions of exposure.
Some critical concentrations are shown in Table 2. These exhibit a reasonable level of
consistency, although there is considerable variability with individual sets of data. From this
limited data-set, the crab would appear to be more susceptible to internal concentrations of
toxicant than would fish. However, considerably more work is required on this matter.
i
Table 2. Observed critical concentrations for aquatic organisms with chlorohydrocarbons.
Organisms
Critical concentrations
(mol kg1)
Crab (Portiimis pelagicus)
Fishes (many species)
Fishes (many species)
Fishes (many species)
3.24 (1/w)
3.24(w/w)
2.0-2.5 (w/w)
5.2-12.2 (w/w)
Data from Mortimer & Connell (1994).
142
-------
A knowledge of the critical internal concentration is of considerable value. It can be used
to estimate the LC5o and corresponding exposure times. The following equation is generally
applicable to bioconcentration:
CB =
where CB is the biotic concentration, Cw the water concentration, ki and k2 the uptake and
clearance rates respectively, and t the exposure period. From this it can be shown that
LC50 = (Ccrit.YL)/[k1/k2(l-e-k2t)]
For any persistent lipophilic chemical, a knowledge of the KOW value allows calculation
of ki and k2. If the lipid content of the organism and its critical concentration are known, then it
may be possible to calculate the LCjo at different exposure times. However, this method requires
further development and confirmation.
CONCLUSIONS
Bioaccumulation occurs with persistent lipophilic organic compounds and can result from
the uptake and retention of contaminants in food and water. However, the pathway from water
seems to be dominant in most situations. It is likely that the QSARs established with fish can be
extended to other aquatic organisms. However, caution must be exercised in applying these
relationships since they are most applicable with the chlorohydrocarbons and deviations may
occur with other substances. Also, these relationships depend on a lack of biodegradation within
the organism during the bioaccumulation process. The least biodegradable compounds are the
chlorohydrocarbons, but biodegradation may occur to different extents with different organisms
since these may have differing capacities to carry out this process. As a result, different
organisms may possibly exhibit different capacities to bioaccumulate various compounds.
REFERENCES
Baughman, G.L., and D.F. Paris. 1981. Microbial bioconcentration of organic pollutants from
aquatic systems - a critical review. CRC Critical Reviews in Microbiology. 205-227.
Connell, D.W. 1988. Bioaccumulation behaviour of persistent organic chemicals aquatic
organisms. Reviews of Environmental Contamination and Toxicology 101:117-154.
Connell, D.W. 1990. Bioaccumulation of Xenobiotic Compounds. CRC Press, Inc.
Boca Raton, Florida.
Connell, D.W., and D.W. Hawker. 1988. Use of polynomial expressions to describe the
bioconcentration of hydrophobic chemicals by fish. Ecotoxicology and Environmental
Safety 16:242-257.
Connell, D.W., and G.J. Miller. 1981. Petroleum hydrocarbons in aquatic ecosystems -
behaviour and effects of sub-lethal concentrations. CRC Critical Reviews of
Environmental Control 11:37-104.
143
-------
Connell, D.W., and G. Schiiurmann. 1988. Evaluation of various molecular parameters as
predictors of bioconcentration in fish. Ecotoxicology and Environmental Safety 15:
324-335.
Cullen, M.C., and D.W. Connell. 1992. Bioaccumulation of chlorohydrocarbons pesticides by
fish in the natural environment. Chemosphere 25:1579-1587.
Davies, R.P., and A.J. Dobbs. 1984. The prediction of bioconcentration in fish. Water Research
18:1253-1262.
Harding, G.C.H., and W.P. Vass. 1978. Uptake from seawater of DDT by marine planktonic
Crustacea. Journal of the Fisheries Research Board of Canada 36:247-254.
i
Hawker, D.W., and D.W. Connell. 1986. Bioconcentration of lipophilic compounds by aquatic
organisms. Ecotoxicology and Environmental Safety 11:184-197.
Kenaga, E.F., and C.A. Goering. 1980. Relationship between water solubility, soil sorption,
octanol water partitioning and bioconcentration of chemicals in biota. In: Aquatic
Toxicology. J.C. Easton, P.R. Parrish, and A.C. Hendricks (Eds.). American Society for
Testing and Materials, Vol. 707.
Kerr, S.R., and W.P. Vass. 1973. Pesticide residues in aquatic invertebrates. In: Environmental
Pollution by Pesticides. C.A. Edwards (Ed.). Plenum Press, London, United Kingdom.
pp. 134-180.
Mackay, D. 1982. Correlation of bioconcentration factors. Environmental Science &
Technology 16:274-278.
Moriarty, F. 1975. Exposure and residues. In: Organochlorine Insecticides: Persistent Organic
Pollutants. F. Moriarty (Ed.). Academic Press, London, UK. pp. 29-72.
Mortimer, M.R., and D.W. Connell. 1993. Bioconcentration factors and kinetic chlorobenzenes
in a juvenile crab (Portunus pelagicus (L)}. Australian Journal of Marine Freshwater
Research 44:565-576.
Mortimer, M.R., and D.W. Connell. 1994. Critical internal and aquatic lethal concentrations of
chlorobenzenes with the crab, (Portunus pelagicus). Ecotoxicology and Environmental
Safety 28:298-312.
Schuurmann, G., and W. Klein. 1988. Advances in bioconcentration prediction. Chemosphere
17:1551-1559.
SOdergren, A. 1968. Uptake and accumulation of DDT by Chlorella sp. Oikos. 19:126-134.
144
-------
A PHYSIOLOGICAL MODEL TO PREDICT XENOBIOTIC
CONCENTRATIONS IN FISHES
Rong Yang1, David J. Randall1, Colin Brauner1, John F. Neuman3, and Robert V. Thurston3
ABSTRACT
A physiological model was developed to estimate fish body toxicant load based on
information regarding the chemical exposure regime, fish body weight, lipid content, and oxygen
uptake. The general model was tested in which an oxygen database (OXYREF) was used to
predict fish toxicant body burden. Based on the quantitative analysis, it was shown that the
model was reliable and accurate in estimating fish body burden of a number of non-metabolized
aquatic toxicants. This modified model possesses some functional reality which enables more
realistic predictions, making it useful in the practice of aquatic environmental risk assessment.
INTRODUCTION
Interest in physiologically-based toxicokinetic (PBTK) models has increased in the past
years (Nichols et al. 1993) because of a growing need to predict the time-course of chemical
concentration in specific tissues. The major advantage of the PBTK model is that it defines an
organism, be it terrestrial or aquatic, in terms of its anatomy, physiology and biochemistry.
Consequently, the PBTK model can be parameterized independently of exposure information
and provides a basis for the comparison of kinetic data among fish species. This type of model,
which leads to improved understanding of the uptake and deposition of chemicals in different
animal tissues (Andersen et al. 1987, Reitz et al. 1990), is also associated with certain
restrictions and disadvantages.
Firstly, the schematic representation of a PBTK model is normally fairly complicated,
requiring a large number of physiological inputs which may not be easily accessible or even
available. This has caused problems, for example, in the development of a PBTK model for
channel catfish (Ictaluruspunctatus) (Nichols et al. 1993), since only very limited physiological
information is available for this fish in the literature (Burggren and Cameron 1980), resulting in
the absence of estimates for several important parameters such as the cardiac output and
respiratory volume. Thus, this model can only be applied following the collection "of
physiological information for this fish, vital for accurate estimation of toxicant loading using
this model.
'Department of Zoology, University of British Columbia, Vancouver, Canada.
Present address: Hong Kong, Environmental Protection Department, Hong Kong
fisheries Bioassay Laboratory, Montana State University, Bozeman, USA.
145
-------
Secondly, the kind of physiological measurements needed, if not available in the
literature, are often extremely difficult and time consuming to obtain. Considering the channel
catfish PBTK model as an example, some of the physiological parameters required are gill
surface area, gill epithelial thickness, effective respiratory volume, blood/water partition
coefficient (Nichols et al. 1993). Such data are available in the literature for a relatively small
number offish species. In addition, measurement of these physiological parameters can be quite
variable due to the differences in the experimental set-up, temperature and animal preparation,
making the data difficult to compare and extrapolate and, subsequently, to standardize.
Moreover, these models are usually developed using a single species that is easy to culture and
handle and assumed to be of importance as a model system for kinetic studies. The same model
parameters, however, when used for a different species, may not provide an accurate prediction
due to the lack of an interspecies extrapolative ability among fish species.
Thirdly, the dependence of toxicant concentration prediction on anatomical
measurements could lead to errors because fishes, at different activity levels, may have different
toxicant accumulation and elimination behaviours, not reflected in the model parameterization.
For example, environmental or physiological factors that influence ventilation volume also
influence the uptake of chemicals. According to Spacie and Hamelink (1982), ventilation
volume (Rv) increases with metabolic rate (Q), which is a function offish body weight (F):
Q = Fa. As the weight exponent is approximately 0.8g, the metabolic rate per gram (Q/F)
decreases with the increase offish body weight. Consequently, a 10-g fish consumes about 60%
as much oxygen per gram as does a 1-g fish. A similar weight relationship should exist for ki as
well. Murphy and Murphy (1971) measured the values of 0.75 and 0.77 for the uptake of DDT
by mosquito fish at 5 and 20 °C, respectively. These values were similar to the corresponding
exponents for oxygen consumption (0.71 and 0.68), but lower than the exponent for gill surface
area per gram (0.89), indicating that uptake is more a function of the fish gill's physiological
rather than anatomical properties.
Thus, the results obtained using these PBTK models involving the measurement
of several anatomical and physiological parameters should be accepted with caution and,
in general, are only applicable after verification, which negates the value of the model as a
predictive tool. The question is posed, therefore, as to whether there exists any suitable
physiological parameter that is not only easy to obtain, but is also functionally realistic.
All of the above-mentioned physiological parameters such as gill surface area, gill epithelial
thickness, gill blood flow, diffusion distance and coefficient, etc., presumably affecting toxicant
transfer, are all related to gas exchange across fish gills. As a result, fish oxygen consumption
is the end-result of the optimization of all of the above parameters. In other words, fish oxygen
uptake is facilitated in a similar manner as toxicant transfer across fish gills.
Yang et al. (1998) reported that oxygen uptake can be used as an indicator for toxicant
transfer regardless offish species and the physico-chemical properties of the non-metabolised
chemicals in question. In addition, OXYREF (Thurston and Gehrke 1993) has been shown to be
applicable to the proposed model since fish oxygen consumption is not altered by toxicant
exposure (Yang and Randall 1997), meaning that the physiological component critical for this
I
146
-------
physiological model can be obtained with ease. Given the argument that fish gills are the main
site for toxicant transfer (Randall et al. 1998) and the toxicant uptake/depuration kinetic
processes are characterized by fish metabolic rate, it becomes a matter of predicting the rate
constants such that the overall toxicant body load can then be estimated at any time using the
appropriate modified rate constant model.
MATERIALS AND METHODS
Compartment Models
A compartmental model is a simplified mathematical description of a chemical's
behaviour in an animal, where the body is represented as a system of compartments. This type
of model has been predominantly applied in aquatic toxicology in that, despite their limitations,
they allow an essential description of the kinetics of xenobiotics in aquatic animals. The
simplest and the most commonly used is a one-compartment, first-order kinetics (1CFOK) model
which has been argued by some researchers to be the standard for pharmacokinetic analysis in
aquatic toxicology (Spacie and Hamelink 1982). The 1CFOK model is actually a rate constant-
based model with the basic relationships expressed as follows:
for uptake:
and for elimination:
= C
to '
-k2t
(1)
(2)
where: Cf = concentration of the chemical in the fish at time t; Cw = water concentration;
Cto = concentration in the fish at time zero; k:, k2 = uptake and elimination rate constants,
respectively.
The uptake of chemicals from the diet can be described with a similar model.
Bruggeman et al. (1981) used the following equation to describe bioaccumulation after dietary
exposure:
Cf(t) = E • f/k2 • Cfd(l-exp (-k2t) ) (3)
where Cfd = concentration of chemical in food; f = feeding rate (food weight • fish weight"1
• time"1); E = absorption efficiency for ingested chemical;
Constant dietary exposure results in increasing fish concentrations until a plateau level is
reached when the clearance rate equals the uptake rate. The ratio between the concentration in
the fish and in the food at steady state is given by the biomagnification factor Km:
km=Cf(t)/Qa=E-f/k2
(4)
More recently, Gobas (1993) has suggested a comprehensive model which combines all
the possible factors that may affect the chemical concentration within a fish body in an overall
147
-------
flux equation, describing the net flux of chemical into the fish as the sum of all of the uptake and
loss fluxes:
i
dCf/dt = k1-Cw + kD-Cfd-(k2 + kE + kM + kG)-CF (5)
where ki, k2 = uptake and elimination rate constants via gills; kD = uptake rate constant from
food; kB = elimination rate constant by faecal egestion; kM = metabolic transformation rate
constant of the chemical; ko= elimination rate constant by gastrointestinal egestion.
The application of these relationships requires three assumptions: (1) the system
operates by first-order kinetics, indicative of passive diffusion of the toxicant into the organism;
(2) a steady-state can be reached; and (3) the body can be treated as one well-mixed
compartment, with the rate of distribution of the toxicant within the organism exceeding the
rate of exchange with the surroundings.
Although the model should be applied with some appropriate restrictions, the 1CFOK
model approach can provide a useful simple approximation. In order to estimate steady-state
chemical concentrations in fish, values are required for all the rate constants (k's) for different
chemicals and fish species.
Special Considerations for the Model Development
Using Fish Oxygen Uptake as an Indicator of Toxicant Transfer Kinetic Processes.
It has been clearly demonstrated by Yang et al. (1998) that fish oxygen consumption is
significantly correlated with both fish toxicant uptake and depuration processes, regardless of
fish species and size. These results indicate that for fish the main route for toxicant entry is
across the gills in exchange with the water, Randall et al. (1998) have shown that toxicant
uptake from food plays only a minor role in determining the total toxicant body burden in
water-breathing animals.
Application of the Oxygen Database and Conditions. The results from the study by Yang
et al. (1998) make it possible to calculate the rate of toxicant transfer using fish oxygen uptake
rate. The oxygen data base, OXYREF, compiled from thousands of fish oxygen consumption
rate values in the literature measured in the absence of toxicant exposure, provides such
information. Moreover, it has been reported that fish oxygen consumption pattern is not affected
by sublethal exposure of tetrachlorobenzene (TeCB) and tetrachloroquaiacol (TeCG) (Yang and
Randall 1997), making the prediction of toxicant transfer kinetic processes even more efficient
and accurate since oxygen consumption rate for different fishes at different activity levels may
be directly retrieved from the database. There are limitations to the utilization of OXYREF
since oxygen uptake can be used as an indicator of toxicant transfer only under normoxic
conditions and for compounds that are not readily metabolized. Altered fish ventilation rate or
diffusion capacity without changes in whole animal metabolic rate, which may occur during
hypoxia or hyperoxia (Randall 1990), will change the rate of toxicant movement across fish gills
without changes in oxygen uptake (McKim et al. 1985). In other words, toxicant absorption may
be altered under these conditions with non-significant changes in fish oxygen consumption.
148
-------
Under such circumstances, fish toxicant transfer could be either underestimated (hypoxia) or
overestimated (hyperoxia), resulting in a greater margin of error in the body burden prediction
Model Development and Testing
Based on all the findings discussed above, the most commonly used one-compartment,
first order kinetic (1CFOK) model (Spacie and Hamelink 1982) was chosen as the basis of this
work. The basic relationships for toxicant uptake and depuration were expressed in Equations
1 and 2.
The uptake and depuration rate constants in Equation 1 were calculated according to fish
oxygen consumption rate by using the correlations determined by Yang (1997):
k, = 1.054 MO2- 2.923, r2=0.87
k2 = 0.0099 MO2 - 2.2975, r2 = 0.87
(6)
(7)
A test of the model was conducted to look at the feasibility and accuracy of predicting
chemical concentration in fishes by incorporating the relationships (Equation 6 and 7)
established between fish O2 uptake and kinetic rate constants (ki and k2) into the 1CFOK model.
In order to test the validity of this modified model, experiments were chosen from the
literature and the experimental data for toxicant load were compared with those derived through
the modeling based on relevant information (e.g., size of the fish being used, experimental
conditions including toxicant exposure concentration, temperature) reported in that literature and
used to derive oxygen uptake and, therefore, uptake and elimination rate constants. Chemicals
selected for the model are reported in Table 1.
Table 1. Chemicals and log KOW values used in test model.
Xenobiotic
Dichlorodiphenyltrichloroethane(DDT)
Sodiumdodecylbenzenesulfonate {LAS)
Ethylenediaminetetraacetic acid (EDTA)
Tetradecylheptaethoxykte (AE)
Di-2-ethylhexylphthalate (DEHP)
Tetrachlorobenzene (TeCB)
Tetrachloroguaiacol (TeCG)
Log KOW
3.99
2.24
2.56
3.78
4.35
4.97
4.41
Reference
Bishop and Maki 1980
Bishop and Maki 1980
Bishop and Maki 1980
Bishop and Maki 1980
Tarretal. 1990
Yang and Randall 1995
Yang and Randall 1995
149
-------
RESULTS AND DISCUSSIONS
Model Application
The result of the model application is shown in Figure i. The predicted uptake data of
DDT at both test concentrations agreed with the measured ones quite well up to an exposure
time of 120 hours. The experimental data of AE, however, seemed to level off after 24 hours
while the predicted values kept increasing. As for the depuration data, model values of LAS and
DEHP concentrations fitted well with the test ones at two water concentrations, and so did the
data for high speed TeCG depuration.
Assuming the model is perfect, and the recorded values are accurate, the predicted values
will be exactly the same as the observed ones. As a result, the linear regression derived between
predicted and measured data should have a slope of 1.0. However, if the values predicted using
the present simplified model are plotted against those obtained from the experiments, the
regression line has a slope of 1.288 (Figure 2), meaning there exist some discrepancies.
Consequently, the predictive power of the simplified model can be judged by the degree the
perfect match regression line overlaps with the 95% confidence intervals of the real regression.
It has turned out that the perfect regression line falls within the 95% confidence intervals for
92% of all the predictions being conducted (67 out of 73 data points). The six points that do not
overlap with the perfect match line are all from one single test with the same chemical, i.e.,
DEHP (log 1^=4.35), using rainbow trout at 12 ° C. .It is with these six points that the model
tends to overestimate, or perhaps the measured values are lower than what they should be.
One possibility is that, in these uptake experiments, the chemical exposure concentration
may have declined over time due to the initial rapid uptake rate by the fish, and as a result
measured values are underestimated. Another possible cause of the discrepancies are errors in
the choice of oxygen consumption values taken from OXYREF. Veith et al. (1979) found that,
between 5 to 15 °C, the bioconcentration of Aroclor® increased much more for fathead minnows
(Pimephales protnelas) and green sunfish (Lepomis cyanellus), which are warm water fish, than
it did for rainbow trout (Salmo gairdneri), a cold water species. It was also reported that
increases in oxygen uptake per gram were greater for carp (Cyprinus carpio) than for brook trout
(Salvelinusfontinalifi) (Beamish 1978). It seems unlikely, however, that the derived oxygen
uptake taken from OXYREF was the cause of the discrepancies observed in the DEHP trout
study because an oxygen uptake error would result in over- or under-estirnation by the model
over the entire range of data points. In the DEHP trout study there was initial agreement
between the model and measured data, the discrepancies only appeared in the later data points.
Considering Equation 1, time (t) is indefinite at equilibrium and the time component
drops out of the equation, leaving Cf= Cw(k!/k2). Thus at a given toxicant concentration in the
water the body burden at equilibrium will be determined by the ratio of the uptake and
depuration kinetics. It can be seen from Equations 6 and 7, that k2 changes more rapidly with
oxygen uptake than does ki, e.g., toxicant load at equilibrium will decrease with increasing
150
-------
400
DEHP
TeCG
1 2 4 8 163248 64 96
Time (h)
TeCB
80
0 0.5 1 2 4 8 16 21
Time (day)
TeCB
20--
| _60-- ••..„._.
I §40 • •
0 0.5 1 2 4 8 16 21
Time (day)
LAS
1 2 4.5 10
Time (day)
LAS
21
6 12 24 48 120
Time (h)
3 8 24 48 120
Time (h)
12 24 48 120
Time (h)
•Experimental Data
-Model Data
Figure 1. Comparison between experimental and predicted values for chemicals of different
DDT = dichlorodiphenyltrichloroethane, LAS = sodium dodecylbenzenesulfonate,
EDTA = ethylenediaminetetraacetic acid, AE = tetradecylheptaethoxylate, DEHP =
di-2-ethylhexyl phthalate, TeCB = tetrachlorobenzene, TeCG = tetrachloroguaiacol.
151
-------
levels of activity (Figure 3). Therefore, if an animal loaded with toxicant at rest becomes active,
the toxicant load, in the fish will fall. Most fish continually change activity level and this means
that an equilibrium between toxicant levels in the water and in the fish will seldom be reached in
the natural environment. In addition, increased activity will tend to reduce toxicant body burden
in chronically exposed fish.
400
03
(D
I
1
• »-i
13
&
300 -
200 -
100 -
0 -
-100
95% Confidence limits
Regression between predicted and
observed values (slope = 1.288)
Perfect match
(slope-1)
-100
100
200
300
400
Observed Values
Figure 2. Quantitative analysis of the predictive power of the simplified model. Each single
data point corresponds to the relevant observed and model-predicted values which
constitute the dashed and solid lines, respectively, in the graphs of Figure 1.
152
-------
2000
• - K1VS.MO2
O- K2VS.MO2
Cf (steady state) vs. MO2
2000
1500
(365.28)
1000
500
II
3 W
9*
TO «?*
(313.28)
£ x
en
S
2
H
g
2 .
•|
0 .
. .0 /
.-** (1.17)
tf
* *
^. • *
— *
^..-- (0.67)
(0.18) ,.-•""
o-'"
o- a,
« n
a'3
*+
eZ
3 S
sv s^
""^^
eg
»
0
240
260
280
300
320
340
360
Fish Oxygen Consumption Rate (mg/kg/hr)
Figure 3. Changes of toxicant transfer rate constants (k! and k2) and fish toxicant body burden
at equilibrium (Cf) with the increase offish oxygen uptake (MO2), assuming constant
toxicant concentration in the ambient water (Cw) being 1 unit.
Model Evaluation
Even though the model test has generated very positive results for non-metabolised
hydrophobic compounds under normoxic condition, there are limitations, in theory, to the use
of the model being developed. These inherent limitations, however, are also applicable to other
PBTK models.
Whether the purpose is to predict environmental residues or to acquire insight into toxic
mechanisms, bioconcentration must be viewed as a dynamic process with competing rates of
uptake and depuration. Factors affecting either one of these two kinetic processes will induce
changes to the overall chemical accumulation within the body of the fish. The relationships
between fish toxicant uptake/depuration and fish oxygen consumption were developed using
153
-------
target compounds resistant to bio-transformation, assuming fish gills are the sites where toxicant
transfer is limited. Toxicant loss from the fish can be achieved through routes other than
breathing for metabolized chemicals (Southworth et al. 1980,1981), as discussed earlier. As a
result, the model reported here is not adequate for predicting tissue residues of substances that
are biotransformed once they have entered the fish. From an environmental standpoint,
however, persistent chemicals (i.e., chemicals that are not biotransformed) are often more toxic
and generally have a high priority in bio-residue monitoring. These chemicals that resist
biotransformation can have a persistent and adverse impact on the aquatic ecosystem, so the
prediction of their levels in fish is important. The model reported here is the means to achieve
this goal.
• i
In any model of the uptake and elimination of xenobiotics in fish, one of the most
important components is the concentration of the chemical that can be absorbed from the water
via the gills offish. Environmental factors that affect the concentration of a chemical in true
solution (C,,,) will affect both the initial uptake and steady-state concentration in fish tissues
(Spacie and Hamelink 1982). In particular for very hydrophobic chemicals, the concentration of
absorbable or bioavailable chemical is often only a fraction of the total chemical concentration
in the water. This fraction is normally referred to as the Bioavailable Solute Fraction (BSF), or
simply bioavailability (Landrum et al 1985, McCarthy and Jimenez 1985). The major factors
affecting bioavailability of organics in natural waters is the presence of adsorbents (Gobas 1993),
namely sediments, suspended solids and dissolved polyelectrolytes such as humic acids. For
groups of pesticides (Wershaw and Goldberg 1972, Kenaga and Goring 1980), polynuclear
aromatic hydrocarbons (PAH) (Means et al. 1980), and industrial organics (Lopez-Avila and
Hites 1980), sorption to the organic fraction of sediments generally increases with the water-
octanol partition coefficient (KoW). Similarly, dissolved organic matter, including humic
substances, binds chlorinated hydrocarbons, phthalates and PAH (Josephson 1982, Sullivan et al.
1982), reducing their bioavailability. The presence of humic substances retards the uptake of
benzo(a)pyrene (log KQW = 6.06) from water by bluegills (Lepomis macrochirus), but does not
significantly affect the accumulation of anthracene (log K<,w = 4.45) (McCarthy and Jimenez
1985).
i
All of these effects should not alter the uptake rate constant kt since the kinetic process is
independent of the actual toxicant concentration in the ambient water. However, they do
sometimes reduce the "true" chemical concentration (Cw) which is one of the important
components of the model presented here. Moreover, the actual value of C^ may be unknown
since it is difficult, analytically, to distinguish between the free and adsorbed fractions in natural
water samples, meaning the validity of this model is, to a large extent, dependent on the accurate
determination of toxicant bioavailabilty in the first place, which adds some restrictions to its
practicability. The traditional PBTK model, on the other hand, can be parameterized
independently of exposure information and the prediction is accomplished partially by using a
number of in vitro equilibrium partitioning coefficients between water/blood and/or tissue/blood
of different organic compounds (Nichols et al. 1993). These equilibrium coefficients, however,
are also influenced by bioavailabiltiy in the water, so the problem is obscured rather than absent
in the traditional PBTK models.
154
-------
Despite all the aforementioned restrictions and preconditions associated with this newly
developed physiological model, its main advantage over the other compartmental or
physiological models lies in the fact that the prediction is based on actual physiological
processes, and fish oxygen consumption rate is easier to measure with acceptable accuracy than
many physiological parameters used in other models (Hayton and Barren 1990). Furthermore,
oxygen consumption rate is available through OXYREF (Thurston and Gehrke 1993) provided
the fish body weight and activity level is known.
The rationale behind this study was to combine the advantages of both compartmental
and physiological models by incorporating physiological processes into a compartmental model,
using an easily accessible physiological parameter correlated with the rate of toxicant transfer,
i.e., fish oxygen consumption rate, to predict chemical concentration in fish with acceptable
accuracy. The model developed in this study possesses some functional reality which enables
more realistic predictions, and is convenient and practical for application in biomonitoring and
toxicant risk assessment.
ACKNOWLEDGEMENTS
This study was supported by the U.S. Environmental Protection Agency grant
(CR-816369) and the British Columbia Science Council Great Award. Tony Farrell and
Chris Kennedy from Simon Fraser University, and James Hill and Jim McKinley from British
Columbia Research Inc., are gratefully acknowledged for their technical assistance and
comments.
REFERENCES
Andersen, M.E., HJ. Clewell, M.L. Gargas, F.A. Smith, and R.H. Reitz. 1987. Physiologically
based pharmacokinetics and the risk assessment process for methylene chloride.
Toxicology and Applied Pharmacology 87:185-205.
Beamish, F.W.H. 1978. Swimming capacity. In: Fish Physiology. Vol.7. W.S. Hoar and D.J.
Randall (Eds.). Academic Press, New York, NY. pp. 101-187.
Bishop W.E., and A.W. Maki. 1980. A critical comparison of two bioconcentration test
methods. Aquatic Toxicology, ASTM STP 707, J.G. Eaton et al. (Eds.), pp. 61-77.
Brauner, C.J., D.J. Randall, J.F. Neuman, and R.V. Thurston. 1994. The effect of exposure to
1,2,4,5-tetrachlorobenzene and the relationship between toxicant and oxygen uptake in
rainbow trout (Oncorhynchus mykiss). Environmental Toxicology and Chemistry
13:1813-1820.
155
-------
Burggren, W.W., and J.N. Cameron. 1980. Anaerobic metabolism, gas exchange, and acid-base
balance during hypoxic exposure in channel catfish, Ictalurus punctatus. Journal of
Experimental Zoology 213:405-416.
Gobas, F.A.P.C. 1993. A model for predicting the bioaccumulation of hydrophobic organic
chemicals in aquatic food-webs: application to Lake Ontario. Ecological Modelling
69:1-17.
Hayton, W.L., and M.G. Barron. 1990. Rate-limiting barriers to xenobiotic uptake by the gill.
Environmental Toxicology and Chemistry 9:151-157.
Josephson, J. 1982. Humic substances. Environmental Science and Technology 16:20-24.
!
Kenaga, E.E., and C.A.I. Goring. 1980. Relationship between water solubility, soil sorption,
octanol-water partitioning, and concentration of chemicals in biota. In: J.G. Eaton,
P.R. Parrish, and A.C. Hendricks (Eds.), Aquatic Toxicology, ASTM ATP 707. American
Society for Testing and Materials, pp. 78-115.
Landrum, P.P., Reinhold, M.D., Nihart, S.R., and Eadie, B.J. 1985. Predicting the
bioavailability of organic xenobiotics to Pontoporeia hoyi in the presence of humic and
fulvic materials and natural dissolved organic carbon. Environmental Toxicology and
Chemistry 4:459-467.
Lopez-Avila V., and R.A. Hites. 1980. Organic compounds in an industrial wastewater and
their transport into sediments. Environmental Science and Technology 14:1382-1390.
Means, J.C., S.G. Wood, J.J. Hassett, and W.L. Beanwart. 1980. Sorption of polynuclear
aromatic hydrocarbons by sediments and soils. Environmental Science and Technology
14:1524-1528.
McCarthy, J.F., and Jimenez, B.D. 1985. Reduction in bioavailability to bluegills of polycyclic
aromatic hydrocarbons bound to dissolved humic materials. Environmental Toxicology
and Chemistry 4:511-521.
McKim, J., Schmieder, P., and G. Veith. 1985. Absorption dynamics of organic chemical
transport across trout gills as related to octanol-water partition coefficient. Toxicology
and Applied Pharmacology 77:1-10.
Murphy, P.O., and J.V. Murphy. 1971. Correlation between respiration and direct uptake of
DDT in the mosquito fish Gambusia affinis. Bulletin of Environmental Contamination
and Toxicology 6:581-588.
Nichols, J.W., J.M. McKim, G.J. Lien, A.D. Hoffman, S.L. Bertelsen, and C.A. Gallinat. 1993.
Physiologically-based toxicokinetic modelling of three waterborne chloroethanes in
channel catfish, Ictalurus punctatus. Aquatic Toxicology 27:83-112.
156
-------
Randall, D.J. 1990. Control and co-ordination of gas exchange in water breathers. In:
Advances in Comparative and Environmental Physiology. Vol 6. R.G. Boutilier (Ed.).
Springer-Verlag, Berlin, Heidelberg, pp. 253-278.
Randall, D.J., D.W. Cormell, R. Yang, and S.S. Wu. 1998. Toxicant body burden in fish is
determined by transfer across the gills, not by feeding. Chemosphere 37:1263-1270.
Reitz, R.H., P.S. McCroskey, C.N. Park, M.E. Andersen, and M.L. Gargas. 1990. Development
of a physiologically based pharmacokinetic model for risk assessment with 1,4-dioxane.
Toxicology and Applied Pharmacology 105:37-54.
Southworth, G.R., C.C. Keffer, and J.J. Beauchamp. 1980. Potential and realized
bioconcentration. A comparison of observed and predicted bioconcentration of
amaarenes in the fathead minnow (Pimephales promelas). Environmental Science &
Technology 14: 1529-1531.
Southworth, G.R., C.C. Keffer, and J.J. Beauchamp. 1981. The accumulation and disposition
of benz(a)acridine in the fathead minnow, Pimephales promelas. Archives of
Environmental Contamination and Toxicology 10:561-570.
Spacie, A., and J.L. Hamelink. 1982. Alternative models for describing the bioconcentration
of organics in fish. Environmental Toxicology and Chemistry 1:309-320.
Sullivan, K.F., E.L. Atlas, and C.S. Giam. 1982. Absorption of phthalic acid esters from
seawater. Environmental Science and Technology 16:428-432.
Tarr, B.D., M.G. Barren, and W.L. Hayton. 1990. Effect of body size on the uptake and
bioconcentration of di-2-ethylhexyl phthalate in rainbow trout. Environmental
Toxicology and Chemistry 9:989-995.
Thurston, R.V., and P.C. Gehrke. 1993. Respiratory oxygen requirements of fishes: description
of OXYREF, a datafile based on test results reported in the published literature. In:
Proceedings of the Second International Symposium on Fish Physiology, Toxicology, and
Water Quality Management. Sacramento, California, USA. September 1990. R.C.
Russo and R. V. Thurston (Eds). United States Environmental Protection Agency,
Environmental Research Laboratory, Athens, Georgia, USA. EPA/600/R-93/157.
pp 95-99.
Veith, G.D., D.L. DeFoe, and B.V. Bergstedt. 1979. Measuring and estimating the
bioconcentration factor of chemicals in fish. Journal of the Fisheries Research Board
of Canada 36:1040-1048.
157
-------
Wershaw, R.L., and M.C. Goldberg. 1972. Interaction of organic pesticides with natural organic
polyelectrolytes. In: Fate of Organic Pesticides in Aquatic Environment, Advances in
Chemistry Series EL R.F. Gould (Ed,). American Chemistry Society, Washington D. C.
pp. 149-158.
Yang, R. 1997. A physiological model to predict xenobiotic concentrations in fishes. Ph.D.
Thesis, Department of Zoology, The University of British Columbia, Vancouver, British
Columbia, Canada.
Yang, R., and D.J. Randall. 1995. Relationships between xenobiotic uptake/depuration and
oxygen consumption by fishes: a physiological model. Proceedings of the Fourth
International Symposium on Fish Physiology, Toxicology, and Water Quality
Management, September 1995, Bozeman, Montana, USA. R.V. Thurston (Ed.). U.S.
Environmental Protection Agency, Environmental Research Laboratory, Athens, Georgia,
USA. EPA/600/R-97/098. pp 1-12.
Yang, R., and D.J Randall. 1997. Oxygen consumption of juvenile rainbow trout
(Oncornhynchus mykiss) exposed to sublethal concentrations of 1,2,4,5-
tetrachlorobenzene and tetrachloroguaiacol. Bulletin of Environmental Contamination
and Toxicology 59:479-485.
Yang, R., D.J. Randall, C. Brauner, R.V. Thurston, and J. Neuman. 1998. Relationship between
toxicant transfer kinetic processes and fish oxygen consumption. Environmental
Toxicology and Chemistry (submitted).
158
-------
MONITORING OF TRACE METALS IN THE AQUATIC ENVIRONMENT
BY ARTFICIAL MUSSELS
R. S. S. Wu1'2, T. C. Lau1, and W. K. M. Fung2
ABSTRACT
An "artificial mussel", a novel chemical device consisting of a polymer ligand suspended
in a semi-permeable container, has been developed for monitoring levels of heavy metals in
seawater. Laboratory studies have shown that this device, constructed from Chelex 100
enclosed in a dialysis tubing, was able to accumulate Cd, Pb and Zn at low ambient
concentrations. Uptake of metals by artificial mussels (AM-I) exhibited a good dose-response
relationship to environmental metal concentration and exposure time. Results of field
transplantation of both AM-I and the green mussel (Perna viridis) showed that metal
concentration factors of the AM-I were significantly higher than those of the green mussels
under natural conditions. Coefficients of variations of metal uptake were also much smaller with
the Chelex®-containing AM-I. Uptake and release of metals by AM-I responded to changes in
temperature and salinity in a predictable manner. Laboratory study also showed that AM-I
mainly uptake the bioavailable fractions of the metals. The overall results indicate that AM-I
meets most (if not all) of the requirements of a good biological indicator, and at the same time,
overcomes the major shortcomings of most biological indicators.
INTRODUCTION
Monitoring of metals is required to detect spatial and temporal changes in coastal waters
to protect living resources and public health. Traditionally, monitoring of metals may involve
quantification and comparison of metal concentrations in water, sediment, or biota. Because
of the difficulties and limitations in water and sediment monitoring (see Table 1), biological
indicators have been widely used in metal monitoring in the last two decades. Certain animals
and plants (e.g. seaweeds, mussels, oysters and barnacles) have been commonly employed to
monitor heavy metals in the marine environment (e.g. Phillips 1977, 1978, 1979, 1980, 1985,
Goldberg et al. 1978, 1983, Bryan et al. 1985, Langston 1986, Ho 1990, Phillips and Rainbow
1988, Claisse 1989, Lauenstein et al. 1990). In particular, mussels have been widely used
to monitor and compare metals in coastal marine areas on a global basis (for example, the
"Mussel Watch" programme) (Goldberg 1975, Goldberg et al. 1978, Lauenstein et al. 1990).
Department of Biology and Chemistry, City University of Hong Kong, Kowloon, Hong Kong
2Centre for Coastal Pollution and Conservation, City University of Hong Kong, Kowloon, Hong Kong
159
-------
Although the use of bioindicators has distinct advantages over the use of water and
sediment in metal monitoring, the dependence of metal bioaccumulation on many physical and
biological factors restricts the use of animals and plants to monitor metal pollution (see Table 1).
Different species may have different accumulation strategies for different metals, and
environmental effects on the uptake and depuration of the metals in the biological indicators are
not well understood. Furthermore, limited natural distribution of indicator species prevents
comparison between monitoring results in different geographic areas and under different
hydrographic conditions. More importantly, most biological indicators are not able to survive
in heavily polluted environments.
Table 1. Summary of problems with monitoring metals in water, sediment, and bioindicators.1
Water
(a) Low concentration makes analysis difficult (pre-concentration is required; contamination is important)
(b) Large temporal variations (c) No time integration (d) No differentiation of soluble and paniculate
fractions of metals (e) No information on bioavailability
Sediment
(a) Effects of particle size, organic content and redox conditions on metal levels are significant and difficult
to account for (b) Unavailabilities of metals are difficult to determine
Bioindicator
(a) Metal contents in tissues are significantly affected by: physical factors (e.g. salinity, temperature, depth
etc.) and biological factors (e.g. reproductive and nutritional conditions, size, sex, life cycles, seasons, gut
content, well-being of the animal, etc.) (b) Population may be intolerant of (or affected by) prevailing
environmental conditions (c) Different species (even closely related species) may have different
accumulation strategies for different metals; inter-specific comparison may be invalid (d) Kinetics of uptake
and depuration are not well understood (e) Differential distribution of metals in different tissues
(f) Limits of natural distribution often prevent direct comparison between indicators in different geographic
area apd hydrographic regime
1 Compiled from Phillips 1977, 1980, Mayer and Fink 1980, Florence 1982, Waldichuk 1985, Rainbow 1990,
Rainbow and Phillips 1993, Robinson et al. 1993.
To overcome the above mentioned difficulties in the use of biological indicators, we
have developed an 'artificial mussel' (AM-I) with an aim to provide a tool for monitoring heavy
metals in aquatic environments. The AM-I is basically a chemical device that consists of a
polymer ligand suspended in artificial sea water in a semipermeable container. The polymer
ligand has multidentate functional groups and can uptake heavy metal ions by forming stable and
insoluble complexes with them. The semipermeable barrier serves to control the rate of uptake
and release of metal ions. Among the various polymer ligands that we have tested (Wu and Lau,
1996), Chelex®100 is particularly suitable for our purpose since it has very strong affinities for
heavy metal ions (Biorad 1994). Our first design of the chemical device consists of Chelex®
enclosed in a dialysis tubing (designated as AM-I). The suitability of using AM-I for monitoring
metal pollution in field conditions has been investigated. The design of AM-I was later modified
to make it more durable and to withstand rough field conditions (designated as AM-II).
160
-------
MATERIALS AND METHODS
Preparation of AM-I
Chelex® 100 (50-100 mesh) was obtained from Biorad and is termed Chelex® throughout
this paper. Dialysis tubings (pore size 12,000 Dalton) were obtained from Medicell International
Ltd. Chelex® (200 mg) was put into a visking tubing containing 10 ml artificial seawater
(35 %o)> followed by sealing both ends of the tubing. A series of experiments was then carried
out to evaluate the suitability of using AM-I as a chemical device for monitoring metal pollution
in the aquatic environment.
Uptake of Gd, Pb and Zn by AM-I
Laboratory experiments were carried out to determine the "dose-response" relationship
between metal uptake by AM-I with ambient metal concentration. Sixty AM-I were placed in
60 L of each of the following metal solutions in 35%o clean seawater: 0.1 ppm, 1 ppm and
5 ppm Cd2+; 0.1 ppm, 0.5 ppm and 1 ppm Pb2+; and 1 ppm, 5 ppm and 10 ppm Zn2+. Metal
solutions were prepared by dissolving the appropriate amount of Cd(NO3)2.4H2O, Pb(NO3)2,
ZnSO4.H2O and commercial sea salt (Marinemix) in deionized water. Water circulation was
maintained by aerating the tank throughout the experimental period. Uptake of metals by AM-I
was determined over a 21-day period. Five AM-I were sampled at time interval. From this, time
required to reach equilibrium for each metal at different concentration was determined.
Release of Accumulated Metals from AM-I
The release of accumulated Cd, Pb and Zn from AM-I in response to a lowering of
ambient metal concentrations was studied over 23 days. Thirty AM-I were put in 60 L of clean
seawater containing 1 ppm Pb, 5 ppm Cd and 10 ppm Zn for 12 days. The AM-I were then
transferred into 60 L clean seawater after rinsing three times with triple distilled water. Three
AM-I were sampled for metal determination at Days 0, 8, 18 and 23 after transfer.
Field Transplantation
The objective of this experiment was to determine the uptake of metals by AM-I under
field conditions, and compare the results with that of the green mussel (Perna viridis). The
AM-I were individually placed inside 30-ml perforated PVC vials and were secured to the inside
of a plastic cage suspended at 5 m below chart datum at a "polluted" site (Victoria Harbour) and
161
-------
at a "control" site (Kat O). Plastic cages containing 30 green mussels (mean shell length 70 mm)
were set up side by side to the AM-I cages. Ten AM-I and 10 mussels were collected from each
site after 1 month.
Sample Analysis
The content of each individual AM-I was emptied into a sintered glass filter and rinsed
three times with triple distilled water, followed by eluting three times with 5 ml 6M HNO3
(AR grade). The elutriant was made up to a known value with triple distilled water and the
metal levels (Cd, Cr, Cu and Pb) in the elutriant were determined either by inductively-coupled
plasma emission spectrometry (ICPES) (Perkin-Elmer Plasma 1000) or flame atomic absorption
spectrophotometry (Shimadzu 650IS). The detection limit was ~ 0.1 mg/L. Concentration of
metals in the polymer was expressed in terms of mg/g of the polymer.
The mussels were dissected with a plastic knife and their byssus threads removed.
The soft tissue was rinsed with deionized double distilled water. The soft part of each individual
mussel was dried in the oven and digested in a block digestor (Techne DG-1) with temperature
controller (Techne TC-DIA), using 30 % hydrogen peroxide and 70 % nitric acid (1:1 v/v).
Metal concentration in the soft tissue of the samples were then determined by inductively-
coupled plasma atomic emission spectroscopy or atomic absorption spectroscopy.
RESULTS
Sea Water
Levels of Cd, Pb and Zn in the seawater used in the present experiments are negligible
and close to the detection limit.
Effects of Ambient Metal Concentration on Metal Uptake by AM-I
Uptake of Cd, Pb and Zn by AM-I showed progressive increase from Day 0 to 18,
and no significant increase was found after 18 days exposure (Mann-Whitney Utest, p O.05).
The accumulation of metals was directly response to both ambient concentration and exposure
time. Uptake of Cd by AM-I under different ambient concentrations is shown in Figure 1.
Accumulation tended to level-off within 21 days, and levelling-off was faster at low
concentrations. Concentration factors in AM-I after 21 days exposure to various concentrations
of different metals (defined as the metal concentration in Chelex® relative to the exposed
concentration) ranged form 1400 to 4800 (Figure 2).
162
-------
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Figure 1. Uptake of Cd by AM-I under different ambient concentrations.
10 ppm
Figure 2. Concentration factors of Cd, Pb, and Zn in AM-I on Day 21.
163
-------
Release of Accumulated Metals from AM-I
Upon returning to clean seawater, levels of all three metals in AM-I showed a slow but
steady decline over a 23-day period, and were significantly lower than their respective values at
Day 0 (Mann-Whitney Utest, p> 0.05). Release of accumulated Cd, Pb, and Zn from AM-I is
shown in Figure 3.
8 10 12 14 16 18 20 22 24
Days
Figure 3. Release of accumulated Cd, Pb, and Zn from AM-1
upon returning to metal-free seawater.
Field Transplantation
All AM-I transplanted to Victoria Harbour were ruptured and no data or comparisons can
be obtained. Concentration of metals accumulated by AM-I and P. viridis at Kat O as well as
the coefficient of variation (CV) are shown in Table 2, Compared with mussels transplanted at
the same site for the same period of time, AM-I accumulated significantly higher levels of Cd,
Cr, Cu, and Pb, but lower level of Zn under natural field conditions (ANOVA, p > 0.05).
Coefficients of variations of metal uptake by AM-I were also smaller than mussels for Cd, Cr,
and Pb.
164
-------
Table 2. Levels of metals ()J.g g"1) in AM-I and Perna viridis transplanted for 1 month at
a clean site (Kat O) in Hong Kong. (Coefficients of variations are shown in brackets.)
AM-I
Perna viridis
Cd
11.9(0.14)
0.18(0.39)
Cr
0.88(0.27)
0.55 (0.29)
Cu
3.70 (0.52)
1.85(0.27)
Pb
52.5(0.17)
0.96 (0.44)
Zn
11.0(1.16)
25.2(0.19)
DISCUSSION
To be useful in metal monitoring, the following characteristics are desirable for such a
chemical device: (1) should be able to accumulate most (if not all) metals at environmental
concentrations; (2) should have high concentration factors; (3) both uptake and release should
be responsive to environmental concentrations; (4) both uptake and release should not be
drastically affected by short-term fluctuations in environmental concentration; (5) equilibrium
should be reached in 2 to 4 weeks.
The present results indicate that AM-I can meet most, if not all, of the above criteria,
at least for the metals tested. AM-I is able to accumulate Cd, Pb, and Zn at ambient
concentrations. Uptake of metals by AM-I showed a direct response to both the exposed
concentration and exposure time. It is noteworthy that the concentration factors of AM-I were
within the same order of magnitude as those in mussels (Martincic et a/., 1992). Moreover, time
for equilibrium of adsorption of AM-I is about 2 weeks and could, therefore, provide a time
integration of the average metal concentration in the environment. The results of field
transplantation indicated that AM-I can accumulate metals from marine waters under natural
conditions. Accumulation of Cd, Cr, Cu? and Pb by AM-I was more efficient than that of the
mussel P. viridis, whereas the accumulation ability of AM-I and P. viridis for Zn was within the
same order of magnitude.
The fragile nature of the semipermeable membrane used in AM-I, however, cannot
withstand rough field conditions. Therefore, we have designed a more robust 'artificial mussel'
(AM II) that utilizes a semipermeable polyacrylamide gel, rather than a membrane, as a means
of controlling metal uptake. Further experiments indicated that AM-II with Chelex® enclosed
inside was also able to accumulate Cd, Cr, Cu, Pb, and Zn at low ambient concentrations.
Similar to AM-I, uptake of metals was directly responsive to the exposed concentration and
exposure time. More importantly, AM-II was able to withstand the harsh environmental
conditions of field transplantation. Compared with P. viridis transplanted at the same site for
the same period of time, AM-II accumulated significantly higher levels of Cd, Cr, and Pb.
Metal uptake and release by bio-indicators are affected by environmental factors such as
water temperature and salinity fluctuations, often in an unpredictable way. Currently, we are
studying the effects of temperature and salinity on the uptake and release of Cd, Cr, Cu, Pb,
165
-------
and Zn in seawater by our chemical devices. Preliminary results showed that both the uptake of
metals by AM-I and AM-II increased by a factor of less than two when the temperature was
increased from 20°C to 30°C. On the other hand, the uptake was decreased by about 50% when
the salinity was changed from 20%0 to 30%0-
We have also conducted experiments to investigate the forms of metals taken up by AM-I
in solutions of Cd, Cu and Pb, using anodic stripping voltammetry (ASV). It was found that the
water inside AM-I contains mainly the free metal ion and ASV labile organic and inorganic
complexes dissociated by Chelex®, indicating that these fractions are preferentially taken up by
AM-I. These results suggest than AM-I will uptake mainly the bioavailable fraction of metals,
the fraction which is of ecotoxicological significance. Selection criteria for metal biomonitors
suggested earlier by other workers (Phillips 1977,1980; Phillips and Rainbow 1988; Rainbow
and Phillips 1993) are summarized in Table 3.
Table 3. Evaluation of Chelex® against selection criteria of biological indicators suggested by
Rainbow and Phillips (1993)
Selection criteria of biological indicators
Chelex®
Accumulate to high levels and tolerate pollutants
Do not regulate the pollutant
Same simple correlation between environmental concentration and
tissue concentration under all conditions
Respond only to the fraction that is of ecotoxicological relevance
Tolerant of variations in physico-chemical parameters in the
Environment
Sessile/sedentary, representative of the area
Abundant, easy to identify and sample
Wide distribution, to enable comparison over a wide geographic area
Yes
Yes
Generally true
Yes
Yes
Yes
Yes
Yes
It can be seen that AM-I met most, if not all, of the requirements of a good bioindicator.
Furthermore, our artificial mussels have the following additional advantages:
(1) They are much more robust than bioindicators. Their deployment is therefore not
limited by geographical/hydrographic boundaries, thus allowing direct comparison of monitoring
results worldwide.
(2) Metal uptake chemical devices involve simple and predictable processes. Effects
such as concentration, temperature, and salinity can be readily modelled using simple physical
laws. This has major advantage over biological indicators, of which metal uptake involve
complicated and variable biological processes.
166
-------
ACKNOWLEDGEMENTS
We thank Dr. P.H. Ko, Dr. W.S. Lee, Ms. M. Chan, V. Poon, and J. Wong for their
technical support to this project.
REFERENCES
Biorad. 1994. Guide to Ion Exchange, 3-5 pp.
Bryan, G. W., W. J. Langston, L. G. Hommerstone, and G. R. Burt. 1985. A guide to the
assessment of heavy metal contamination in estuaries using biological indicators.
Occasional Publication of Marine Association, UK Marine Biology Association,
Plymouth, UK. 92 pp.
Claisse, D. 1989. Chemical contamination of French coasts. The results of a ten years mussel
watch. Marine Pollution Bulletin 20:523-528.
Goldberg, E. D. 1975. The mussel watch - a first step in global marine monitoring. Marine
Pollution Bulletin 6:111.
Goldberg, E. D., V. T. Bowen, J. W. Farrington, G. Harvey, J. H. Martin, P. L. Parker,
R. W. Risebfough, E. Robertson, E. Schneider, and E. Gamble. 1978. The Mussel
Watch. Environmental Conservation 5:101-125.
Goldberg, E. D., M. Koide, V. Hodge, A. R. Flegal, and J. Martin. 1983. US Mussel Watch:
1977-1978 results on trace metals and radionuclides. Estuarine Coastal Shelf Science
16:69-93.
Ho, Y. B. 1990. Ulva lactuca as bioindicator of metal contamination in intertidal waters in
Hong Kong. Hydrobiologia 203:73-81.
Langston, W. J. 1986. Metals in sediment and benthic organisms in the Mersey Estuary.
Estuarine Coastal Shelf Science 23:239-261.
Lauenstein, G. G., A. Robertson, and T. P. O'Connor. 1990. Comparison of trace metal data in
mussels and oysters form a mussel watch programme. Marine Pollution Bulletin
21:440-447.
Martincic, D., Z. Kwokal, Z. Peharec, D. Margus, and M. Branica. 1992. Distribution of Zn, Pb,
Cd and Cu between seawater and transplanted mussels (Mytilus galloprovicialis).
Science of the Total Environment 119:211-230.
167
-------
Phillips, D. J. H. 1977. The use of biological indicator organisms to monitor trace metals in
marine a estuarine environments-a review. Environmental Pollution 13:281-317.
Phillips, D. J. H. 1978. The common mussel Mytilus edulis as an indicator of trace metals in
Scandinavian waters. 2. Lead, iron and manganese. Marine Biology 46: 47-156.
Phillips, D. J. H. 1979. The rock oyster Saccostrea glomerata as an indicator of trace metals in
Hong Kong. Marine Biology 53:353-360.
Phillips, D. J. H. 1980. Quantitative Aquatic Biological Indicators: Their Use to Monitor Trace
Metal and Organochlorine Pollution. Applied Science Publishers, London, 488 pp.
Phillips, D. J. H. 1985. Organochlorines and trace metals in green-lipped mussels Perna viridis
from Hong Kong waters: a test of indicator ability. Marine Ecology Progress Series
21:251-258.
Phillips, D. J. H., and P. S. Rainbow. 1988. Barnacles and mussels as biomonitors of trace
elements: a comparative study. Marine Ecology Progress Series 49:83-93.
Rainbow, P. S., and D. J. H. Phillips. 1993. Cosmopolitan biomonitors of trace metals. Marine
Pollution Bulletin 26:593-601.
Wu, R. S. S., and T. C. Lau. 1996. Polymer-Ligands: a Novel Chemical Device for Monitoring
Heavy Metals in Aquatic Environment. Marine Pollution Bulletin 32:391-396.
168
-------
A COMPARISON OF MUSSELS AND SPMDS AS MONITORS
FOR TRACE ORGANIC POLLUTANTS IN HONG KONG WATERS
Bruce J. Richardson, Gene J. Zheng, and Edmund S.C. Tse1
ABSTRACT
Mussels and other bivalve shellfish have been widely used as sentinel organisms
(or bioindicators) because of their ability to accumulate conservative contaminants from the
water column. Indeed, their practicality in this regard has been recognized in many countries,
including the USA, Hong Kong, and Australia, where "Musselwatch" programs have been
instigated as part of routine monitoring procedures. However, the usefulness of bivalves in
monitoring such contaminants as organochlorine pesticides, petroleum hydrocarbons (PHCs),
and polycyclic aromatic hydrocarbons (PAHs) can be limited by several factors. For example,
certain biogenic hydrocarbons, which are naturally present in the organisms, may interfere with
chemical analyses, leading to the necessity for extensive cleanup prior to quantitation by gas
chromatography. Furthermore, natural variability in bivalve populations, driven by such factors
as growth and reproduction, can also compromise uptake and sequestration of contaminants, and
may provide complications in the interpretation of monitoring data. Semi-permeable membrane
devices (SPMDs) have been recently developed to overcome such problems. The devices consist
of a length of polyethylene tubing, in which is sealed a small quantity of highly refined lipid
(e.g., triolein). Lipophilic compounds partition into, and are concentrated by the lipid, and thus
the SPMDs act as a "mussel surrogate". This paper describes a recent experiment in which
SPMDs and mussels (Perna viridis) were deployed side-by-side at five sites in Hong Kong
coastal waters. The results allow a comparison to be made between the effectiveness of SPMDs
and mussels as monitoring tools for the estimation of bioavailable trace organic contaminants.
INTRODUCTION
Maintenance and evaluation of water quality criteria are often the basis for marine and
estuarine monitoring programmes which measure the concentrations of toxicants in waters
(see Martin and Richardson 1991). However, despite the known toxicity of petroleum
hydrocarbon compounds (including the polycyclic aromatic hydrocarbons, or PAHs) and
organochlorine compounds (including PCBs), direct measurement of these substances in
seawater has proved to be difficult, due to extremely low ambient concentrations and the
concomitant need for large sample volumes (Murray et al, 1991). One of the more successful
'Department of Biology and Chemistry, City University of Hong Kong, Kowloon, Hong Kong.
169
-------
approaches to monitoring hydrocarbon contaminants has utilised sentinel organisms such as
mussels and oysters (the so-called "Musselwatch"; see Phillips 1980). This technique has been
most widely used by regulatory agencies in the USA (e.g. with the mussel Mytilus
californianus); southern Australia (utilizing Mytilus edulis); and to some extent in Hong Kong
(using Perna viridis). However, the technique suffers from expensive and time-consuming
chemical separation and cleanup of samples in order to remove interfering biogenic
hydrocarbons and lipids. The other limitation of the "Musselwatch" technique has been the
translation of mussel organic contaminant concentrations to water quality standards, thus greatly
limiting the usefulness of the technique for regulatory purposes (Martin and Richardson 1991).
Many researchers have proposed that passive sampling with abiotic devices may allow a
more accurate estimation of pollutant levels in the environment. The abiotic systems which have
so far been proposed are diverse, and include bees wax coated paper (Hettche 1971), glass slides
coated with natural animal fat (Schramm and Hutzinger 1988), tristearin coated glass fibres
(Mueller 1997) and textiles (Horstman and McLachlan 1994). However, in retrospect, a
significant step in such developments occurred in the 1980s, when Sodergren (1987) proposed
the use of hexane filled dialysis tubes. A recent advance in this line of research is the
development (initiated in the USA) of a device which has shown perhaps the greatest promise in
overcoming difficulties in seawater monitoring of hydrocarbons (Huckins et al. 1990). This
sampling tool uses membrane diffusion chemistry and a purified lipid matrix. An inexpensive
passive sampler, known as a semi-permeable membrane device (or SPMD) is the result. The
sampler is constructed of a length (e.g., 4 cm by 1 m) of thin polyethylene membrane tubing
containing up to 10 g of a highly refined lipid, triolein. The development of SPMDs in the USA
has concentrated mainly on using the samplers in freshwater and temperate seawater conditions.
For instance, the Marine Pollution Laboratories of the California Department of Fish and Game,
and the University of California have tested SPMDs in San Francisco Bay for the evaluation of'
PCS congener and dioxin-dibenzofuran contamination (Prest et al. 1992). As a result of such
trials, SPMDs have achieved worldwide application as passive samplers for both hydrophobic
aquatic and atmospheric pollutants.
The successful use of SPMDs requires an understanding of the uptake kinetics of
pollutants in the device's sampling phase, e.g., triolein. The uptake of pollutants in SPMDs
is limited by both the rates of diffusion of pollutants to the membrane, known as water side
resistance, and the rate of diffusion of pollutants through the membrane (i.e., membrane
resistance). Research has shown that, for the uptake of pollutants with KoW values >104 in
SPMDs, water side resistance is usually the limiting factor (Hofelte and Shea 1997). Many
important toxicants, such as PAHs, PCBs, dioxins, or organochlorine pesticides have KoW values
of 104 or greater, and thus the kinetic is greatly dependent on water side resistance. A problem
is that this resistance seems to be an unpredictable parameter, as it is also influenced by a suite
of uncontrollable factors relating to the flow rate and turbulence of the water that is being
monitored.
Despite these limitations, SPMDs have been used in a variety of situations, and for a
variety of purposes. For instance, and of relevance to this paper, SPMDs were deployed in an
east-west direction along the Geelong Arm of Port Phillip Bay, Victoria, Australia (Prest et al.
170
-------
1995) alongside transplanted, cultured local mussels (Mytilus edulis). The aims of this
experiment were: (a) to establish the utility of SPMDs as monitors of organochiorines, including
PCBs and pesticides; and (b) to utilize a known gradient of contamination within the Geelong
Arm as a comparative test of the SPMDs vs. mussels as monitoring tools. Knowledge of the
organochlorine gradient along the Geelong Arm was established as part of Victorian State
Government monitoring exercises, and has been reported in the literature by Murray et al. (1991)
and Phillips et al. (1992).
The experiment illustrated that both SPMDs and mussels provide similar information on
the relative levels of organochlorine contamination in Corio Bay. Nonetheless, the profiles of
PCBs in the mussels and those in the SPMDs differed from each other: for example, SPMDs
suggested that lower chlorinated PCBs dominated in the water column, whilst mussel data
implied essentially the reverse. These differences were attributed at the time to contaminant
solubility, partitioning of contaminants on suspended particulates, and possible
biotransformation (Prest et al. 1995). However, further explanations may be found in the
limitations of SPMDs as true "integrators" of contaminant concentrations (i.e. their ability to
reflect "average concentrations" of contaminants during the exposure period).
This paper describes a similar experiment that has recently been undertaken in Hong
Kong involving the use of SPMDs. The experiment is notable in that SPMDs have not
previously been effectively tested, or compared with local biomonitors, in tropical or sub-tropical
waters. The objectives of this experiment were to: (1) Deploy and retrieve local green-lipped
mussels (Perna viridis) and SPMDs at five sites in Hong Kong coastal waters; (2) Analyse
mussels and SPMDs for organochlorines and petroleum-related hydrocarbons; and (3) Compare
the utility of SPMDs vs. mussels as monitoring tools for the assessment of biologically-available
hydrocarbon contaminants in coastal waters.
MATERIALS AND METHODS
Green-lipped mussels were collected from a clean site (Kat O) and transplanted to four
other sites in Hong Kong (Figure 1). The sites were selected to provide a range of contaminant
concentrations (i.e. effectively a gradient) based upon the previous analyses of sediments for
petroleum related compounds and organochlorines at 22 stations in Hong Kong coastal waters
(see Zheng and Richardson 1999, Richardson and Zheng 1999). Semi-permeable membrane
devices, obtained from Dr Harry Prest (Long Marine Laboratory, Santa Cruz, California, USA)
were deployed in specially constructed frames at the five sites. Approximately 50 mussels of
similar size were placed in a stainless steel mesh container, and deployed alongside the SPMDs
at each station. Both mussels and SPMDs were exposed at the test sites for a period of 30-days.
Following retrieval, mussels and SPMDs were analyzed using previously established protocols
(Murray et al. 1991, Prest et al. 1995, Zheng and Richardson 1999, Richardson and Zheng 1999,
and Xu et al. In Press).
171
-------
RESULTS AND DISCUSSION
Mussels transplanted to the Kwun Tong site (Figure 1) did not survive the 30-day test
period. As noted by Zheng and Richardson (1999), this site is highly hypoxic and is thought to
be a virtual "biological desert". In addition, it contains extremely high concentrations of
petroleum-related hydrocarbons and organochlorines in its sediments (Zheng and Richardson
1999, Richardson and Zheng 1999). At the other sampling sites, all mussels survived, and no
SPMDs were lost at any site during the deployment.
Figure 1. Map of Hong Kong showing the sampling sites.
172
-------
Organochlorine pesticide and total petroleum hydrocarbon concentrations in mussels and
SPMDs are reported in Tables 1 and 2. The results indicate that SPMDs are potent accumulators
of water-borne organic contaminants. The levels reached, and the resulting ease of detectability
of individual contaminants in SPMDs most often exceeded that in mussels (Table 1).
Table 1. Concentrations of organochlorine pesticides in mussels (ng/g dry weight)
and SPMDs (ng/g lipid)
Tolo Harbour
Tsim Sha Tsui
Sai Wan Ho
KatO
Kwun Tong
Pesticide
p,p'-DDT
p,p'-DDE
Total HCH
Heptachlor
Aldrin
Chlordane
Dieldrin
Kepone
Total
Pesticides
Mussels
1.8
10.1
24.1
8.2
nd
nd
3.5
nd
47.7
SPMDs
29.4
21.9
126
nd
44.1
nd
46.8
37.4
305
Mussels
nd1
9.8
10.9
1.9
nd
nd
nd
10.8
33.4
SPMDs
15.7
22.0
79.8
27.5
39
14.2
30.8
50.5
280
Mussels
7.2
10.5
11.6
nd
5.2
nd
5.5
nd
40.0
SPMDs
37.8
31.9
83.3
15.8
38.9
20.8
51.4
39.3
319
Mussels
9.4
7.5
12.0
nd
8.0
1.2
7.1
nd
45.2
SPMDs
29.4
23.6
33.7
38.5
15.8
nd
39.4
33.7
214
Mussels
na2
na
na
na
na
na
na
na
na
SPMDs
nd
nd
70.9
49.3
nd
nd
47.5
49.3
217
!nd = not detected; 2na = data not available as mussels died
Table 2. Total petroleum hydrocarbon concentrations in mussels and SPMDs.
Location
Mussels
(mg/g dry weight)
SPMDs
(mg/g lipid)
Tolo Harbour
Tsim Sha Tsui
Sai Wan Ho
KatO
Kwun Tong
1.21
0.82
0.58
0.66
na1
6.82
5.75
2.46
1.71
7.24
*na = data not available as mussels died
173
-------
SPMDs generally reflected the mussel data in ranking organic contamination at the test
sites; normalized values of these contaminants for both mussels and SPMDs are reported in
Table 2. This was especially true for petroleum hydrocarbons. However, data for total
chlorinated hydrocarbons were less convincing, and may be related to a number of factors,
including localized influences on the bioavailability of individual contaminants (e.g. due to
adsorption to particulate matter), and stress on the mussels at sites where little water exchange
takes place, thus limiting the ability of the organisms to accumulate contaminants (e.g. in
Victoria Harbor). A third possible reason is that there are relatively high levels of chlorinated
pesticides at the Kat O site, which was the source of the "clean" mussels for the experiment.
These initial levels (see Figure 2; in particular for DDT) may have resulted from pesticide
sources outside of Hong Kong. As the mussels would have originally contained relatively high
levels of chlorinated contaminants prior to transplantation, this may have compromised results at
other sites due to depuration vs. accumulation effects.
The SPMDs showed a much greater range of organic contaminants than did the mussels.
This is illustrated by the chromatograms shown in Figure 3 which compare chlorinated
compounds in mussels and SPMDs from the same site. This phenomenon was also reported by
Prest et al. (1995) in their comparative study of SPMDs and blue mussels (Mytilus edulis) in
Corio Bay, Victoria, Australia.
Based upon the results of our experiment in Hong Kong, and of those from previous trials
in Australia, SPMDs show considerable potential as indicators of organic contaminants, although
they have certain limitations. We recommend the use of SPMDs as follows:
1. In conjunction with mussels, which may better reflect contaminants that are truly
bioavailable and not necessarily subject to biotransformations, and which are retained for
extended periods in the tissues of living organisms.
2. As tools to identify the range of contaminants present. This potential is clearly
shown in the chromatograms of organochlorines in SPMDs from this experiment, which indicate
an enormous range of compounds in comparison with the chromatograms of mussel tissues
(Figure 3). Similar data were also reported by Prest et al. (1995) for organochlorine compounds
in Corio Bay, Australia. Combined with appropriate toxicity testing procedures for the
compounds accumulated (which could potentially be extracted and separated during analysis,
and thence subject to ecotoxicological investigation) SPMDs offer a fascinating potential,
deserving further research and development.
3. Where mussels cannot live because of adverse conditions, such as Kwun Tong in this
experiment. Many such circumstances exist world wide, and the use of SPMDs in these
places has potential as an alternative to bivalves, or indeed other living organisms, as indicators
of organic contaminant load.
174
-------
Figure 2. Normalized values of total petroleum hydrocarbons (top); and typical organochlorines,
HCH (centre) and DDT (bottom) in SPMDs (grey bars) and mussels (dark bars).
TH = Tolo Harbour; TST = Tsim Sha Tsui; SWH = Sai Wan Ho; andKO = KatO
(see also Figure 1).
175
-------
,,f«
Figure 3. Electron capture detector gas chromatograms of organochlorine compounds
in mussels (top) and SPMDs (bottom) at the Tsim Sha Tsui test site.
ACKNOWLEDGMENTS
The authors thank City University of Hong Kong for the award of a Strategic Research
Grant to support this work.
176
-------
REFERENCES
Hettche, H.O. 1971. Pflanzenwachse als sammler fuer polycycklische aromaten in der Luft von
Wohngebieten. Staub Reinhalt Luft 31:72-76. (In German)
Hofelt, CS, and D. Shea. 1997. Accumulation of organochlorine pesticides and PCBs by
setnipermeable membrane devices and Mytilus edulis in New Bedford harbor.
Environmental Science and Technology 31:154-159.
Horstmann, M., and M.S. McLachlan. 1994. Textiles as potential passive collectors of
atmospheric PCDD/F. Conference Presentation, Society of Environmental Toxicology
and Chemistry - Europe. Brussels, Belgium, 1994.
Huckins, J.N., M.W. Tubergen, and G.K. Manuweera. 1990. Semipermeable membrane
devices containing model lipid: a new approach to monitoring the bioavailability of
lipophilic contaminants and estimating their bioconcentration factor. Chemosphere
20:533-552.
Martin, M., and B.J. Richardson. 1991. Long term contaminant biomonitoring: views from
northern and southern hemisphere perspectives. Marine Pollution Bulletin 22:533-537.
Milller, J.F. 1997. Occurrence and Distribution Processes of Semi-volatile Organic Chemicals
in the Atmosphere and Leaves. PhD Thesis, Griffith University, Brisbane, Australia.
Murray, A.P., B.J. Richardson, and C.F. Gibbs. 1991. Bioconcentration factors for petroleum
hydrocarbons, PAHs and biogenic hydrocarbons in the blue mussel, Mytilus edulis.
Marine Pollution Bulletin 22:595-603.
Phillips, D.J.H. 1980. Quantitative Aquatic Biological Indicators. Their Use to Monitor Trace
Metal and Organochlorine Pollution. Applied Science Publishers Ltd. London.
Phillips, D.J.H., B.J. Richardson, A.P. Murray, and J.G. Fabris. 1992. Trace metals,
organochlorines and hydrocarbons in Port Phillip Bay, Victoria: a historical review.
Marine Pollution Bulletin 25:200-217.
Prest, H.F., W.M. Jarman, S.A. Burns, T, Weismueller, M. Martin, and J.N. Huckins. 1992.
Passive water sampling via semipermeable membrane devices (SPMDs) in concert with
bivalves in the Sacramento/San Joaquin River Delta. Chemosphere 25:1811-1823.
Prest, H., B.J. Richardson, L. Jacobson, J. Vedder, and M. Martin. 1995. Monitoring
organochlorines with semi-permeable membrane devices (SPMDs) and mussels (Mytilus
edulis} in Corio Bay, Victoria, Australia. Marine Pollution Bulletin 30:543-554.
Richardson, B.J. and G.J. Zheng. 1999. Chlorinated hydrocarbon contaminants in Hong Kong
surficial sediments. Chemosphere 39:61-71.
177
-------
Schramm, K-W and O. Hutzinger. 1988. Accumulation of semi-volatile lipophilic
micropollutants in thin butter layers. Chemosphere 19:1729-1742.
Sodergren, A. 1987. Solvent-filled dialysis membranes simulate uptake of pollutants by aquatic
organisms. Environmental Science and Technology 21:85-87.
Zheng, G.J. and B.J. Richardson. 1999. Petroleum hydrocarbons and polycyclic aromatic
hydrocarbons (PAHs) in Hong Kong marine sediments. Chemosphere 38:2625-2632.
Xu, L., G.J. Zheng, P.K.S. Lam, and B.J. Richardson. In Press. Relationship between tissue
concentrations of polycyclic aromatic hydrocarbons and DNA adducts in green-lipped
mussels (Perna viridis). Ecotoxicology.
178
-------
USE OF SEMIPERMEABLE MEMBRANE DEVICES
TO MEASURE XENOBIOTICS IN LITHUANIAN RIVERS
A. Cetkauskaite1, D. Sabaliunas2, J. Ellington3,1. Sabaliuniene1, and A. Sodergren2
ABSTRACT
Here we report the screening of bioavailable hydrophobic toxicants in surface waters of
Lithuania in 1995-1996 using semipermeable membrane devices (SPMDs). The major
chemicals identified at four field study locations in 1995 were polycyclic aromatic hydrocarbons
(PAHs), polychlorinated biphenyls (PCBs), hexachlorocyclohexane (HCH) isomers, alkyl
hydrocarbons, and phthalates. The dialysates from those SPMDs were toxic to luminescent
bacteria Vibrio flscheri, and decreased bioluminescence 60-85% when compared to control.
Calculated 30-minute EC50 values ranged from 0.4 to 2.4 u.g per ml of SPMD triolein, and the
highest toxicity values were observed in samples from an industrial water well in the territory of
the Kedainiai chemical plant, a creek in the vicinity of the Jonava chemical plant, and the Neris
River downstream from both the Vilnius wastewater treatment plant and a paper mill.
Dialysates of SPMD samples collected in May 1996 were tested in the same bioassay; the
sample from the "relatively clean" Ula River exhibited three times more toxicity than the sample
from the Vilnia River which received effluent from several industrial enterprises and municipal
wastewater. The whole dialysates were subjected to bioassay-directed fractionation on silica gel
columns: Toxicity testing of each fraction revealed that most of the toxicity was contained in
Fraction 10 (eluted with 100% acetone) which contained oleic acid as the major component
(based on GC/FID, GC/ECD and GC/FTIR/MS analyses). Spectral analysis showed that other
toxic fractions contained a number of halogenated compounds and PAHs. Field samples
collected in July 1996 were fractionated by size exclusion chromatography (SEC) and analysed
by GC/ECD on GC/MS. Fraction 1 contained lipids and polyethylene oligomers, Fraction 2
contained all of the PCBs, organochlorine pesticides, and PAHs, and Fraction 3 contained
sulphur. All SPMD samples collected in July 1996 effectively inhibited the bioluminescence of
V. fischeri (Microtox™) with the calculated 30-minute EC50 values in the range of 0.052-0.86 jxg
of the SPMD triolein per milliliter of the bacterial suspension. In general, SPMDs were shown
to be a useful way to screen for hydrophobic toxicants in water. However, sample clean-up
procedures to remove oleic acid and sulfur, followed by sample fractionation, may be required
prior to toxicity testing for the estimation of the true toxic potential of the accumulated different
pollutants.
1 Faculty of Nature Sciences, Vilnius University, Eiurlionio Str. 21, 2009 Vilnius, Lithuania
2 Department of Ecology, Lund University, S 223 62, Sweden
3 National Exposure Research Laboratory, US Environmental Protection Agency, Athens, Georgia 30605, USA
179
-------
INTRODUCTION
River water quality data included in the annual reports of the Environmental Protection
Ministry (EPM) of Lithuania are based mainly on conventional general parameters such as
COD, BOD, total nitrogen, and total phosphorus, and are usually the average of monthly data
(EPDLR 1993,1996,1997a, 1997b). Data on xenobiotic chemicals are scarce and their
contribution to toxicity of aquatic organisms may not be reflected in the general parameters.
Gas chromafpgraphy (GC) analysis of individual hazardous chemicals (such as HCH and its
isomers, sym-triazines, and other chlororganic pesticides) in surface waters is performed by the
laboratories of the Joint Research Center of the Lithuanian Environmental Protection Ministry
(EPM), and usually do not show any negligible pollution in the sensitivity range of ng/L - ug/L).
For example, gas chromatography/electron capture detector GC/ECD and GC/flame ionization
detector (GC/FID) analysis of special organic pollutants in Lithuanian river water was reported
yearly from 1992 in the annual issues of the EPM River Water Quality Chronicle (EPDLR 1993,
1997a). The classes of organic pollutants analyzed in river water samples included p-p'-DDT,
its metabolites, and HCH isomers from year 1992, sim-triazine herbicides from year 1993,
and chlorinated phenols and benzoic acid derivatives from year 1994. According to the data
presented in the River Water Quality Chronicles no chlorinated phenols or benzoic acid
derivatives were detected consistently in a majority of the samples from Lithuanian rivers
from 1994 and later. One exception was the Sidabra River at the Latvia border where
3,4-dichlorobenzoic acid was found at a concentration of 0.003 mg/L (EPDLR 1997a, 1997b).
No sim-triazine herbicides and P-HCH were routinely detected in different sampling places
from 1993 and later. Only prometryn and simazine were detected in the Nemunas River
(concentration 0.6 and 0.575 jag/L) upstream from Druskininkai (i.e. close to the border of
Belarus) in 1993. Similar situations of clean river water having no DDT or its metabolites and
HCH isomers were reported in the River Water Quality Chronicles of 1992. One exception was
a finding of 2 )j.g/L of a-HCH in Neris River sediments downstream from the city of Jonava in
1996 (EPDLR 1993,1997a).
The following observations resulted from reviewing these official EPM documents:
(1) all these specific analyses were performed in river water, except a few in sediments in 1996;
(2) the possible industrial source of pollution of bottom sediments in rivers downstream from the
larger cities (Vilnius, Alytus, and Kaunas) were not investigated; (3) all results are presented in
River Water Quality Chronicles as annual average of appropriate measurements; (4) many
individual organic compounds, such as PAHs, were not routinely investigated (EPDLR 1993,
1996,1997a, 1997b). Accordingly, herein we report independent sampling and analysis of
organic pollutants collected in Lithuanian rivers during 1995-1996 using semipermeable
membrane devices (SPMDs) for field sampling. Chemical identification of compounds was
performed by GC/FID, GC/ECD, and GC/mass spectrometry (G-C/MS).
Since the development of the SPMDs in the late 1980s, they have been successfully
applied under a variety of experimental and environmental settings (Sodergren 1987, Huckins
et al 1990, Johnson 1990, Prest el al. 1995, Huckins et al. 1996, Sabaliunas and Sodergren
i
180
-------
1997, Sabaliunas et al. In Press, Sabaliunas et al. In Prep.). SPMDs containing triolein are
known as potent preconcentrators of bioavailable organic contaminants in the aquatic
environment (Huckins et al. 1990, 1996). These SPMDs consist of a thin film of synthetic lipid
triolein enclosed in thin-walled flat polyethylene tubing. Uptake of chemicals by the SPMDs
containing triolein are known as potent preconcentrators of bioavailable organic contaminants in
the aquatic SPMDs is based on the process of passive partitioning of the chemical between the
aqueous phase and the lipid. Triolein is a major component of the lipid pool of aquatic
organisms such as fishes and molluscs (Chiou 1985, Ewald and Larsson 1995). The greatest
accumulation is observed for organic pollutants having octanol-water partition coefficient (KoW)
with a log value > 4.0. The SPMD water concentration factor for chemicals in flow-trough
systems can vary from 700 for beta-benzenehexachloride (log K<,w = 3.8) or 1,200 for lindane
(log KoW = 3.9), up to 24,100 for hexachlorobenzene (Kow= 6.2) (Huckins et al 1996). Low
density polyethylene is described usually as "nonporous", although in free volume there exist
transient cavities having a diameter of less than 10 A (Hwang and Kammermeyer 1975). The
molecular size-exclusion limit for nonporous polymers (including polyethylene membranes) is
similar to that estimated for biological membranes (Opperhuizen et al. 1985, Hwang and
Kammermeyer 1975). One of the main advantages of SPMDs is the time-integrated monitoring
of hydrophobic pollutants in aquatic ecosystems, i.e., rapid and effective concentration of
different hydrophobic compounds from water during various periods (Huckins et al. 1996).
According to data obtained by Huckins et al., triolein SPMDs have different equilibration
rates (Rs) for different chlorinated organic compounds (Ibid). For example; the equilibration
time for lindane is 5 days in flow-trough experiments, while equilibration time for
hexachlorobenzene is 3 hours. Analytical recovery of chlororganic compounds from SPMDs
averaged about 80%. It was also shown that accumulation of chlororganic compounds from
sediments into SPMDs is a more complicated process than that from water, and the
concentration factor in triolein (for 2,2',5,5'-tetrachlorobenzene and other organic chemicals)
obtained after 14 and 28 days of exposure was 26,600 and 38,000 (Ibid). It also has been shown
that the SPMD technique can be readily integrated with standard bioassays to measure toxic and
mutagenic effects of accumulated pollutants (Sabaliunas and Sodergren 1997, Cleveland et al.
1997). Such an approach allows for rapid and low-cost screening of bioavailable hydrophobic
chemicals in the aquatic environment. Depending on the results of bioassays, the SPMD
dialysates may be analyzed to identify the compounds responsible for the observed effects.
The goal of the present study was to review the potential use of SPMDs in screening
chemical pollutants, and evaluating their effects in rivers of Lithuania using the bioluminescence
inhibition in the marine bacteria V. fischeri assay (Microtox™ test) to assess the toxicity of the
SPMD dialysates. In an attempt to identify chemicals contributing most to the observed effects,
the samples were fractionated, and each fraction was re-tested in the Microtox™ assay.
Chromatography/spectral techniques were used to make tentative identification of the main
components of the fractions.
181
-------
MATERIALS AND METHODS
Field Study Locations and Natural Conditions
Selected locations for the study in 1995 and 1996 are shown in Figure 1. In October-
November 1995 four sites were selected in water bodies in industrial or urban areas. They
were (1) the River Obelis at the rain-water drainage outlet of the Kedainiai chemical plant,
(2) an industrial water well adjacent to this plant, (3) the Neris River downstream from the city
of Vilnius and the paper pulp mill at Grigiskes, and (4) a creek in the vicinity of the Jonava
chemical plant (Sabaliunas and SOdergren 1997).
Latvia
Russia
{Kaliningrad)
Poland
Figure 1. Major rivers of Lithuania and SPMD sampling sites.
year 1995 October-November field sampling places; exposure lasted for 21 days.
year 1996, May field sampling places, exposure lasted for 4 weeks.
year 1996, July field sampling places; exposure lasted for 21 days
182
-------
In May 1996 two additional sites were selected in relatively small rivers with comparable
flow rates: (1) the Vilnia River in the capital city of Vilnius and which receives pre-treated
wastewater from up to seven treatment plants (Huckins et.al. 1990); and (2) the Ula River in
the Dzukija National Park, a river that is regarded as a clean natural river with limited inflow of
waste waters from some rural areas (Sabaliunas et al. In Press). In July, 1996 SPMDs were
deployed in four additional sites: (1) the Neris River upstream from the city of Vilnius, (2) the
Neris River downstream from Vilnius, (3) the Kulpe River, and (4) an underground water
monitoring well at the Zigmantiskes pesticide storage site. In the upper Neris River, where the
river bottom is rocky, the SPMDs were exposed to a column of clear running water. In the lower
Neris River and the Kulpe River the protective shields of the SPMDs sank into the sediment
slurry, so that SPMDs were exposed in the river sediments. The difference between the
sediments of the Neris and Kulpe Rivers was that the former sediments were in standing muddy
water without any vegetation (characteristic of anoxic sediments), while the Kulpe sediments
were exposed to running and relatively clear water with very abundant aquatic vegetation.
The water temperature at all three locations was 20-21°C (Sabaliunas et al. In Prep.). At the
pesticide storage site in Zigmantiskes, the SPMDs were lowered into an underground water
monitoring well located approximately 20 meters from the concrete wall of the underground
chamber in which pesticides were stored. The SPMDs were exposed in the water column 0.5 m
from the well bottom, 2.5 m below the water surface and, 16.5 m from the ground surface. The
underground water temperature was 8°C, and separate samples were taken for routine analysis of
COD, BOD, DO, and pH (Sabaliunas et al. Submitted).
Test Chemicals
Triolein, (95 % purity) was obtained from the Sigma Chemical Co., St. Louis, Missouri.
Cyclohexane, hexane, methylene chloride, acetone, methanol, and other solvents used were all
analytical grade, obtained from the Merck Chemical Co. or from Sigma. Silica gel columns
were obtained from J.T. Baker (6 ml, 1 g), and were pre-rinsed with hexane. A size exclusion
column was obtained from Phenomenex, Torrance, California. It contained Phenogel (350 mm
x 21.2 mm diameter x 5 A). Standard mixtures of PCBs, PAHs, and organochlorine pesticides
were used for identification of sample components by GC/ECD or GC/FID.
Preparation of SPMDs for Field Exposure
The SPMDs for deployment in Lithuania were prepared at the Department of Ecology,
Lund University, Sweden. The first 14 of these were prepared in 1995 in the same way as
described by Huckins et al. (1990). Flat polyethylene tubing (Brentwood Plastics, Inc. Missouri,
USA) 2.7 cm wide and with a thickness of 72-75 jim was cut into 50-cm segments; these were
extracted with cyclohexane for 24 hours to remove potential contaminants. The segments were
filled with 0.5 ml (0.455 g) of triolein and configured to form a thin film. The effective length
of the SPMDs (distance between two thermosealed ends of the segment) was 45 cm. Both ends
of the SPMDs were clamped together with Spectra For® closures so that the tubing formed a
twisted loop. They were then placed in amber-glass jars capped with teflon-lined screw caps and
transported to Lithuania. Time between the preparation of SPMDs and deployment was 7 days
183
-------
and most of that time (except during the transportation) the jars were stored at 4°C. Glass rods
(6 cm in length) were tightened and fixed at the ends of the SPMDs, and used as anchors to keep
the SPMDs in a vertical position in the water (Sabaliunas and Sodergren 1997).
In May 1996 SPMDs were again prepared at Lund University in a similar way as in 1995.
The length of the SPMDs between thermosealed ends was 45 cm and this resulted in the
membrane area to lipid volume ration of 520 cm2/ml. Freshly prepared SPMDs were stored in
clean metal containers at -20°C, and they were transported from Sweden to the study locations in
Lithuania at approximately 4°C (Sabaliunas et al. In Press). Six SPMDs were deployed at each
site, placed in metal protective shields and submerged in the running river water. Field blanks
(control samples) consisted of six SPMDs that accompanied the other SPMDs to the study
locations. Exposure time at each sampling site was 4 weeks. At the end of exposure, the
SPMDs were transported back to Lund University and were processed further in the laboratory
(Sabaliunas et al. In Press).
In July 1996 additional SPMDs of standard configuration (90 cm long, 2.5 cm wide,
membrane thickness 75-80 jj.m, containing 1 ml of 95 % purity triolein) were prepared at Lund
University as described earlier by Sabaliunas and Sodergren (1997), stored in sealed acetone-
rinsed stainless steel containers at 4°C, and transported by air to Lithuania and immediately
taken to and deployed at four field study sites, three in rivers and one in a monitoring well at the
Zigmantis'kes pesticide storage site. At each site eight SPMDs were deployed, placed in a
protective shield made of stainless steel. At all three locations in the rivers, the SPMDs were
lowered to the bottom of the river at a distance of 2-4 meters from the river bank (Sabaliunas
et al. Submitted). Field blanks consisted of eight SPMDs that accompanied the other SPMDs
to the study locations (two field blank SPMDs at each location). In addition, eight SPMDs
from the same lot were left at the laboratory at Lund University as laboratory blanks (Ibid.).
Extraction of Compounds from SPMDs
i
SPMDs collected in field studies in October-November 1995 after 21 days of exposure
were each placed in a separate glass vial and returned to the laboratory at Lund University for
processing. They were first rinsed with tap water and air dried, and after checking for damage
and holes, the SPMDs were dialysed in cyclohexane for 48 hours at 13 °C. This low temperature
was chosen to reduce the carry-over of lipids and polyethylene waxes. The dialysates were
evaporated to 3 ml under a gentle stream of nitrogen gas. The dialysate was divided into two
equal parts; one part was used in bioassays and the other saved for identification of accumulated
compounds. The latter part was purified and cleaned up using a simple acetonitrile/cyclohexane
extraction procedure in which the compounds were extracted from cyclohexane with 3 ml of
acetonitrile saturated with cyclohexane. The cyclohexane fraction was then discarded while the
recovered acetonitrile was diluted with distilled water at a ratio 1:9. The resulting solution was
evaporated to 1 ml under a stream of nitrogen gas and treated for approximately 12 hours with
sodium sulfate and copper. Aliquots of the extract were evaporated under the stream of nitrogen
to the volume of 50 jul and analyzed by GC/MS. Gravimetric analysis of the samples collected
184
-------
in 1995 showed that carry-over of lipids and co-dialysed impurities during the dialysis procedure
were less than 1.7% (7.5 mg) of the total lipids amount in the SPMDs. Carry-over through the
acetonitrile/cyclohexane extraction procedure was negligible and could not be quantified.
In May 1996 at the end of exposure for 4 weeks, the SPMDs were transported back to
Lund University where they were processed in the laboratory as follows. Each SPMD was
initially rinsed with running tap water, and debris was removed manually. Each was then
examined for cuts and leaks and air-dried after a quick rinse with acetone. All six SPMDs from
each site and the field blank were cut into five or six pieces and extracted with 500 ml of hexane
for 48 hours, with solvent replacement after 24 hours. The volume of the two hexane extracts
combined was reduced to approximately 5 ml using rotary evaporation and nitrogen blow-down.
The extracts were then passed through kilned (400°C) glass columns with glass fiber filters
(70 nm) at the bottom and 2 g of anhydrous Na2SO4 on top. The columns had been pre-washed
with solvent before addition of the extracts. After addition of the solvents, several solvent rinses
were used to recover all target toxicants. The volume of the extracts was reduced to 2.5 ml using
nitrogen blow-down, and they were transferred into pre-extracted 2.5-cm wide polyethylene
tubes with one end sealed. The tubes were placed upright in 250-ml graduated cylinders.
Hexane (250 ml) was added to each cylinder, and the extracts in the tubes were dialysed for
48 hours, with solvent replacement after 24 hours. The dialysates were evaporated to a volume
of 2 ml and again filtered through Na2SO4 with a glass fiber filter. The filtrates were divided
into two equal parts which were solvent-exchanged to 1 ml of acetone and 1 ml of methylene
chloride, respectively. The acetone part was tested in the Microtox™ assay while the methylene
chloride part was sealed in glass ampules and shipped to the Ecosystems Research Division,
U.S. Environmental Protection Agency (U.S. EPA), Athens, Georgia, USA for fractionation
and chemical analysis.
SPMDs collected in field studies of July 1996 after 21 days' exposure were processed in
the laboratory exactly in the same way as in May 1996, and the final volume of each extract was
brought to 3 ml, half of which was sealed in glass ampules and shipped for fractionation with
size exclusion chromatography to the J.M. Long Marine Laboratory, University of California,
Santa Cruz, USA.
Adsorbtion Chromatography
Extracts of SPMD field samples collected in May 1996 from the Vilnia and Ula Rivers
were fractionated by adsorbtion chromatography on silica gel by the following procedure.
A sample volume of 800 jj.1 was concentrated to approximately 200 ul by nitrogen blow-down.
Hexane (1 ml) was added and the concentration procedure was repeated two times. Silica gel
columns (Ig, 6 ml), (J.T. Baker Co.) were pre-rinsed with 10 ml hexane; elution continued
until the last portion of hexane reached the surface of the silica gel. The concentrate (200 jul)
was then added to the silica gel column, and the hexane level was lowered to the top of the
column. The fractionation on the silica gel column was performed with 3 ml volumes of
hexane:methylene chloride:acetone:methanol in the following proportions: 3:0:0:0 / 2.85:0.15:0:0
/ 1.2:1.8:0:0 / 0:3:0:0 / 0:0:3:0 / and 0:0:0:3, and twelve 1.5 ml fractions collected (Table 1).
185
-------
Table 1. Toxicity of dialysates of SPMDs collected in May 1996 and their silica gel
chromatography fractions in the Microtox™ bioassay (Sabaliunas et al. In Press)
Sample or
Fraction
number
Field blank
Overall11
Fr. 1
Fr, 2
Fr.3
Fr. 4
Fr. 5
Fr. 6
Fr. 7
Fr. 8
Fr. 9
Fr. 10
Fr. 11
Fr. 12
Solvents
used for
Fractionation
Hexane
Hexane
5% methylene chloride: 95% hexane
5% methylene chloride: 95% hexane
60% methylene chloride: 40% hexane
6Q°7o methylene chloride: 40% hexane
Methylene chloride
Methylene chloride
Acetone
Acetone
Methanol
Methanol
5-minute ECso (mg/L)a
Ola River
(95% CI)
I
34.7(21.9-55^0)
0.062 (0.057-0.067)
n.t. (134)c
n.t. (134)
72.7(40.7-152)
110(-)
197 (143-272)
n.t. (134)
n.t. (134)
213 (150-303)
n.t. (134)
0.666 (0.064-0.068)
52.1 (45.4-59.8)
79.8 (77.6-82.1)
"
Vilnia River
(95% CI)
34.7(21.9-55.0)
0.182(1.15-0.224)
n.t. (134)
12.0(11.2-12.8)
123 (104-145)
50.6 ( - )
170(116-249)
n.t. (134)
n.t. (134)
n.t. (134)
193 (171-218)
0.136(0.115-0.162)
161 (142-181)
at. (134)
"Concentration values are based on the amount of SPMD triolein; "^on-fractionated; ''Non-toxic (highest
concentration tested).
Since the void volume of the column was 1.5 ml, the first fraction collected served as a
silica gel column blank in the Microtox™ assay. Each fraction (1.5 ml) was concentrated to
500 jj.1 by nitrogen blow-down and subsequently analyzed on capillary GCs equipped with either
an BCD, FID, or by Fourier transform-infrared spectroscopy/mass spectrometry (FT-ER/MS).
After the final GC analyses, the remainder of the silica gel fraction was quantitatively solvent-
exchanged into acetone, sealed in glass ampules and shipped to Lund University to be retested
in the Microtox™ assay.
Size Exclusion Chromatography
Extracts qf SPMDs field samples, collected in July 1996 were fractionated by size
exclusion chromatography (SEC). Equipment included: (a) a Hewlett-Packard (HP) 1050
quaternary pump (Hewlett-Packard Co., Palo Alto, California) for the mobile phase delivery to
the column; (b) the Phenomenex (Torrance, California) Phenogel 350 mm x 21.2 mm diameter
x 5 A column; (c) an HP 1050 automatic liquid sampler for the injection of samples and
standards; (d) a HP 1050 UV-visible detector for effluent monitoring 254 nm; and (e) an ISCO
Foxy 200 (Nebraska, California) fraction collector that was initiated by the HP 1050 automatic
liquid sampler contact controls. The dichlormethane mobile phase was delivered at 5 ml/minute
by quaternary pumps, injection volume of samples and standards was 300 ul Fraction 1
was taken between 10 and 15 minutes after injection, Fraction 2 between 15 and 23.5 minutes,
and Fraction 3 between 23.5 and 28.5 minutes. Lipids and polyethylene oligomers were
primarily found in Fraction 1. All PCBs, organochlorine pesticides and PAHs were contained in
Fraction 2; sulfur eluted in Fraction 3 (Sabaliunas et al. In Prep.).
186
-------
Gas Chromatographic Analysis of Pesticides Accumulated in SPMDs
The SPMD samples collected in 1995 were analyzed by GC/MS using an HP 5890
Series II GC connected to VG 70E-HF double focusing (EB) medium resolution MS. A fused
silica capillary 30-m column was used (DB-5 from J&W Scientific, Inc.), and temperature was
programmed from 50 to 300 °C. Quantification ion-based reconstructed ion chromatograms and
single ion monitoring (SIM) were used for the detection of target compounds. NIST /EPA/ NIH
Mass Spectral Database was used for the compound identification and standards, when
available, for confirmation and quantification (method of external standard). The list of target
compounds was based on earlier data on water pollution in the study areas (Ellington et al,
1994a, 1994b, 1996, Sabaliunas and Sodergren 1997). Extracts of SPMD field samples,
collected in May 1996 and fractionated by adsorbtion chromatography on silica gel, were
analyzed by GC/ECD and GC/FID detectors. The SPMD extracts were injected on an HP 5890
Series II GC equipped with a DB-5 capillary column (30m length, 0.32 mm ID, and 0.25
micrometer film), a Nickel 63 electron capture and a flame ionization detector. Helium was
used as the carrier gas at 25 cm/sec and nitrogen was the makeup gas. The GC oven was held at
35°C for 4 minutes and then programmed at 9°C per minute, and held there for 10 minutes.
Splitless injections (1 jo.1) were made at an inlet temperature of 250°C.
Gas Chromatography/Fourier Transform-Infrared Spectroscopy/Mass Spectrometry
Analyses of samples of May 1996 were performed on a HP 5890 Series II GC interfaced
with a 5965B infrared detector and a 5971 Series mass selective detector. For injections, the
capillary columns and oven temperature program were the same as for ECD and FID analysis.
The FT-IR lightpipe and transfer line were held at 290°C and the MS transfer line was
maintained at 285°C. Before any samples were analyzed, acceptable system performance was
verified by injection of a "standard" mixture of more than 70 organics covering a broad range of
functional groups, volatilities, polarities, and molecular weights, all at 20 mg/L in methylene
chloride (Sabaliunas et al. In Press). Analysis of this standard prior to sample injection also
allowed the estimation of sample component concentrations based on detector response relative
to selected components of the standard. Under acceptable system performance, an injection
of 3 ul of this standard afforded easy identification of all test mix compounds with a signal-to-
noise-ratio of approximately 20 for both IR and MS.
Extracts of SPMDs field samples, collected in July 1996 were fractionated by SEC.
All fractions were analyzed by GC/ECD for the screening of halogenated compounds.
The second fraction obtained from SEC was analyzed also by GC/MS for possible identification
of the compounds (Sabaliunas et al. Submitted). For the screening of halogenated compounds,
aliquots of Fraction 2 of the SPMD samples were injected on a Varian 3500 GC/ECD.
With a 30-meter fused silica capillary column (DB-5 from J&W Scientific, Inc.) and temperature
programmed from 110 to 240°C. Standard mixtures of PCBs and organochlorine pesticides
were used for tentative identification of the sample components. In addition to the GC/ECD,
Fraction 2 was also screened by GC/MS using an HP 5890 Series II GC connected to a
187
-------
HP 5989A MS engine. The samples were injected on a DB-5 fused silica capillary column
(110-240°C), and helium was used as the carrier gas. Standard PAH mixtures and a library
search were used for tentative identification of compounds (Ibid,).
Bioassays
Toxicity of SPMD dialysates, obtained from samples collected in October-November
1995, was analyzed in the test of inhibition of bioluminescence in V.fischeri Strain 430,
obtained from the culture depository of the Department of Biochemistry and Biophysics, Vilnius
University. Luminescence intensity was measured by aBioOrbit 1250 (LKB-Wallac Co.) single-
sample luminometer. Centrifuged fresh bacterial suspension was diluted with 50 mM phosphate
buffer (pH 7.2) containing 2.5 % of sodium chloride to an optical density of 0.025-0.03 (590 nm)
to obtain bioluminescent light intensity in the range 6000-8000 mV per ml. A 980 jjl sample of
bacterial suspension was mixed with 20 ^1 of test solution. The bioluminescence intensity was
measured after 1, 15, 30 and 60 minutes of exposure. Thirty-minute ECso estimates were
calculated using standard linear regression analysis and least square statistics (ISO/CD 1994).
Mutagenic activity of the samples was evaluated using the Salmonella histidine reversion assay
(Maron and Ames 1983). The strains used were TA 97a, TA 98, TA 100 and TA 102, without
metabolic activation according to procedures described by Maron and Ames (1983).
From samples collected in May and July 1996, both whole SPMD extracts and SPMD
extract fractions were tested in the Microtox™ bioassay, a commercial toxicity test from
Azur Environmental, Carlsbad, California. It is based on the inhibition of luminescence in the
marine bacte.ria V.fischeri as a result of toxic action of chemicals (Bulich 1984). Samples
were tested in a medium containing 2% of sodium chloride and about 107 cells of bacteria
reconstituted from the lyophilized reagent. The luminescence intensity was measured with the
Microbics Model 500 Analyzer at 5- and 15-minute exposure times. Each test was run in
duplicate using four sample concentrations and a negative control. Concentration of the carrier
solvent (acetone) in the test medium, including controls, was 1%. The calculated 30-minute
ECso estimate denote sample concentration expressed on the basis of the original SPMD triolein
that reduces the light output by 50%. The toxicity equivalents were used for graphical
representation of the data and they were calculated as 1/EC50.
RESULTS
Gas Chromatography Analysis
Field samples collected in 1995 and analyzed by GC/ECD contained different polycyclic
aromatic hydrocarbons, alkyl hydrocarbons, polychlorinated biphenyls. HCH isomers were
obtained in one of the SPDM samples from the Neris River downstream from Vilnius.
A detailed list of chemicals extracted from the SPMDs, with individual chemicals confirmed by
analytical standards, is reported in Table 2. Major chemicals identified in all SPMD samples
from different locations were PAHs. These included phenanthrene, fluoranthene, and pyrene.
Chrysene was present in SPMD samples from the Neris River downstream from Vilnius and
188
-------
from the creek in the vicinity of the Jonava chemical plant (Sabaliunas and Sodergren 1997).
High levels of fluorene and alkylated naphthalenes were detected in the SPMD samples from
the territory of the Kedainiai chemical plant. As this was a pilot study, exact quantification of
substances was not a primary goal. However, the levels of PAHs in the samples were estimated
based on calibration curves of the standards and the available data. The highest PAH levels
estimated were in SPMD samples from the Kedainiai chemical plant (0.5-12 |ng/ml SPMD
triolein), the lowest were in the SPMD samples from the River Obelis (< 0.2 |j.g/ml SPMD
triolein). The SPMD samples from the Neris River contained PCBs (0.02-0.08 ng/ml SPMD
triolein) and HCHs (< 0.03 ng/ml SPMD triolein). Alkyl hydrocarbons and phthalates were
detected in all the samples including the field blanks of SPMDs; composition of these
compounds differed significantly and the levels were lower in blank SPMDs than in the field
blanks that indicated the real presence of these compounds in the field samples (Sabaliunas and
Sodergren 1997).
Table 2. Chemicals identified from the field study SPMDs in 1995 (Sabaliunas and Sodergren
1997)
Chemical
Field River Kedainiai Neris Jonava Confirmed
blank Obelis" chemical River0 chemical by
plantb plantd
standard
Polycyclic aromatic hydrocarbons:
Naphthalene and alkyl derivatives
Fluorene
Phenanthrene
Anthracene
Pyrene
Fluoranthene
Benz(a)anthracene
Chrysene
Other PAHs
Other aromatic hydrocarbons
Polychlorinated biphenyls
Pesticides
a-HCH
Y-HCH (lindane)
Alkyl hydrocarbons
Phthalates
Sulphur compounds, including S8
4-
aAt the rainwater drainage outlet of the Kedainiai chemical plant; blndustrial water well on the territory of the
Kedainiai chemical plant; "Downstream from the city of Vilnius; dCreek in the vicinity of the Jonava chemical plant.
Field SPMD samples collected in May of 1996 were concentrated, and part of the extract
after testing in the Microtox™ bioassay was sent to the Ecosystems Research Division, U.S.
EPA, Athens, Georgia, for fractionation and chemical analysis in the attempt to identify
189
-------
chemicals responsible for the observed toxic effects (Sabaliunas et al. In Press). The overall
non-fractionated Ula River, Vilnia River, and blank SPMD dialysates were screened by capillary
gas chromatography using GC/ECD, FED, and FT-IR/MS. The GC/FID chromatograms of the
Vilnia and Ola River samples were essentially identical (Figure 2).
FED r
500000 •
400000 •
300
Oleic acic
200000
100000
0
30
-100 000 '
-200000
-300000 •
-400
Oleic acic
espon
35 n
se
in
Vilnia
L ,
5
Ola
0
1
i
n*
'V
Jt
*
JL
IE
*-UlLc
ifr;
000
x
uuu
-500 000
ECD response
25000'
20000'
15000
10 000
5000'
min.
0
16
Vilnia
__J^_^_
468
IS 20
LJjjJ
r"^ rr" ' -
10 12 14
Ola
-10 000
Figure 2. GC/FID (A) and GC/ECD (B) chromatograms of SPMD samples collected
May 1996 from the Vilnia and Ula Rivers (from Sabaliunas and SQdergren 1997).
The notable difference was that the Ula River extract contained three times as much oleic
acid as the Vilnia River extract. The other major peaks in the chromatograms were identified as
a homologous series of straight chain alkyl hydrocarbons by IR/MS and coinjection of a mixture,
which contained the C-10 through C-19 straight chain hydrocarbons. The chromatograms
190
-------
contained a continuum of unresolved peaks in the C-13 to C-19 retention time window.
The GC/ECD chromatograms from the Vilnia River differed from those from the Ula River in
both the number and intensity of peaks, especially in the retention times bracketed by the C-13 to
C-19 window. One peak, with the same retention time and almost the same area in all three
chromatograms, was used as a relative reference peak. The Ula River extract contained only
three additional peaks of larger area than the reference peak, while the Vilnia River extract
contained 15 peaks of equal or greater area. The response to the GC/ECD was an indication of
the possible presence of compounds, which contained halogens and heteroatoms such as sulfur,
oxygen, and nitrogen, and at much higher levels in the Vilnia River extract.
The complexity of the SPMD chromatograms precluded identification of minor
components by GC/FT-IR/MS. To enhance chromatographic resolution and to facilitate
identification of individual peaks by GC/FT-IR/MS, the extracts were separated according to
functional groups by use of 1 gram silica gel columns. The individual silica gel fractions were
re-tested in the Microtox™ bioassay.
To determine the functional group elution sequence from the silica gel columns, a
mixture that contained nonadecane (C-19), iodohexadecane (1C-16), 2,6-dimethlnaphthalene
(2,6-DMN), p,p'-DDT, chrysene (Chry), 4-chlorophenol (4-C1P), and palmitic acid (PA) was
eluted from a 1-gram column of silicic acid using successive 3-ml portions of the solvent
compositions described in the Materials and Methods Section (Adsorbtion Chromatography).
Fractions (1.5 ml) were collected, concentrated to 0.5 ml, and analyzed by GC/FID or GC/ECD.
It was found that C-19 and IC-16 eluted mainly in the 100% hexane fractions, and 2,6-DMN in
the solution of 5% methylene chloride and 95% hexane, while 4-chlorophenol and palmitic acid
eluted with the 100% acetone fractions.
Gas Chromatography Analysis of Silica Gel Fractions.
The 12 silica gel fractions were analyzed by GC/ECD and GC/FT-IR/MS. In both the
Vilnia and Ula River GC/ECD chromatograms, Fractions 2, 3, and 4 contained the majority of
the GC/ECD sensitive compounds. The Vilnia River chromatograms were the most responsive
to GC/ECD both in number of peaks and peak areas. Fraction 1 was the void volume. The MS
and IR spectra for peaks in Fraction 2 of both extracts were typical for alkyl and branched chain
alkyl hydrocarbons; the "hump" in the C-13 to C-19 region was still present. Fraction 3 in both
extracts contained the remainder of the alkylhydrocarbons and the C-l and C-2 naphthalenes.
Fractions 4 and 5 from the Vilnia River sample contained pyrene and fluoranthene, but neither
compound was detected in the Ula River fractions. However, phenanthrene, anthracene,
9H-fluorene, methyl 9H-fluorene, biphenyl, alkyl-substituted biphenyls, and alkyl-substituted
naphthalenes and phenanthrenes were detected in Fractions 4 and 5 in both the Vilnia River and
Ula River extracts. A compound identified as benzene-1,1' -sulfonylbis was identified by
MS and IR in Fractions 8 and 9 for both extracts.
The results of the chemical analysis of Fractions 2 to 5 showed a higher content of
halogenated organic compounds and PAHs in the Vilnia River sample compared to the Ula River
191
-------
sample. These results are in line with the Microtox™ toxicity data reported in the bioassays
section below. The Ula River and Vilnia River fraction 10 contained the same suite of
compounds based on retention times and their MS and IR spectra, except the major component
identified as oleic acid was at a level three higher level in the Ula River extract.
Field samples collected in July 1996 were fractionated by size exclusion chromatography
into three fractions (Fl, F2 and F3) and analyzed by GC/ECD; Fraction 2 was also analyzed by
GC/MS. Lipids and polyethylene oligomers were primarily found in Fraction 1. Only one large
peak and almost no others were observed in Fraction 1, which is probably due to the general
incapability of high-molecular weight compounds such as lipids to absorb light at 254 nm
wavelength in the UV region of the spectrum (Sabaliunas et al. Submitted). Fraction 2 contained
PCBs, organochlorine pesticides and PAHs. Sulfur eluted in Fraction 3 (Sabaliunas et al.
In Prep.). The profiles of the SEC chromatograms were essentially the same for all the samples
except for those from the lower Neris River and the ZigmantiSkes pesticide storage site. In
Fraction 3 of these, a large peak was evident, indicative of elemental sulfur (S8). This peak was
absent in the upper Neris and Kulpe River SPMD dialysates (Figure 3). Fraction 2 of the SEC
chromatograms contained an unresolved continuum of peaks, the total area of which was similar
for all the samples, except that from the upper Neris River where it was less than half of the
average total peak area of the other three samples.
A detailed chemical analysis was not the major goal of this study, and GC/ECD and
GC/MS techniques were used only to screen the mixtures for major compound classes present in
the samples. In general, the GC/ECD and GC/MS chromatographic patterns of Fraction 2 of all
the samples were very complex and contained continuums of unresolved peaks. Again, the
number of peaks and total peak areas were similar for the lower Neris River and Kulpe River site
samples, and significantly smaller for the upper Neris River sample. The ZigmantiSkes pesticide
storage site SPMD GC/ECD chromatogram contained a relatively higher number and more
abundant peaks than all the other samples (Sabaliunas et al. Submitted). All samples contained
a number of PCBs, PAHs (mainly two to four rings, most often alkylated), some organochlorine
pesticides (mainly hexachlorocyclohexanes and cyclodienes, especially in the Zigmantiskes
pesticide storage site sample) as well as homologous series of alkyl hydrocarbons. The
complexity of the chromatograms precluded the identification of the vast majority of the peaks
(Ibid).
Bioassays
The tests of concentrated dialysates from SPMD samples collected in field studies of
1995 were highly toxic to luminescent bacteria. A marked inhibition of bioluminescence
intensity was observed during the first minute of exposure (Figure 4). After 15 minutes of
exposure the decrease in the bioluminescence intensity reached 60-85% compared to control,
followed by some stabilization. Similar inhibition dynamics were observed for diluted
dialysates. Calculated 30-minute EC50 estimates ranged from 0.4 to 2.4 jig/ml of SPMD
triolein, and the highest toxicities were observed for: (1) the industrial water well on the territory
of the Kedainiai chemical plant: (2) the creek in the vicinity of the Jonava chemical plant:
192
-------
(3) the Neris River downstream from the Vilnius paper mill. The toxicity of reference toxicant
3,5-dichlorophenol (30-minute EC50 = 0.0064 ± 0.6 mg/L) in the V.fischeri test was high
(Sabaliunas and Sodergren 1997). Test samples from the chemical plants were most toxic to
Salmonella typhimurium strains causing the poor growth of bacteria in the background layer
and decrease in the rate of spontaneous revertants for strains TA 9a and TA 100 at doses of
6 to 8 mg per plate, although they exhibited no mutagenic activity in this experiment (Sabaliunas
and Sodergren 1997).
Upper Neris
12.040
— '
Sulfur
10
4
Fraction I
Is ' ' ' '
— >•
^ Fraction
— 1 1 1 —
20
2
1 1 1 ri
25 mi
Fraction 3
Lower Neris
10
IS
j—
20
Figure 3. Size exclusion chromatography fractions of the SPMD samples collected
July 1996 from the upper and lower Neris River (from Sabaliunas et al. In Prep.).
193
-------
Light intensify, %
100
90
80
70
60
50
'40
30
20
10
0
10 20 30 40
Time, min.
50
60
Figure 4. Inhibition of bioluminescent light in Vibrio fisher i by SPMD concentrates obtained
in 1995 field sampling (control = 100%) (from Sabaliunas and Sodergren 1997).
(a) River Obelis at rainwater drainage outlet at Kedainiai chemical plant
(b) Continuous flow study (laboratory experiment), exposure 28 days
(c) River Neris downstream from Vilnius
(d) Creek at the Jonava chemical plant
(e) Industrial water well at the Kedainiai chemical plant
Both the Ola and Vilnia River samples collected in May 1996 exhibited high toxicity in
the Microtox™ assay with the calculated 30-minute EC50 estimates in the range of micrograms
of the SPMD triolein per milliliter of the bacterial suspension (Table 2). the toxicity of the
Ola River sample was three times higher than that of the Vilnia River sample. Such results were
rather unexpected since, as mentioned above, the Ola River is regarded as one of the cleaner
rivers in Lithuania, and it is often used as a reference site in various environmental monitoring
studies. One must note, however, that data on the chemical pollution of the Ola river is virtually
non-existent, and the above assumption is based more on its geographical location away from
major industrial enterprises. The Ola River is located in the forest region, with possibly only
some village sewage pollution as far as 10 km upstream from the sampling area. The results
of the Microtox™ toxicity testing of whole and silica gel fractions for both the Vilnia River
and Ola River SPMD samples collected in May 1996 clearly indicated that almost all of the
toxicity of both samples were contained in Fraction 10 (100% acetone) (Table 1) which was
194
-------
about two times more toxic for the Ula River sample than for the Vilnia River sample. There
is no doubt that the latter fraction predetermined the overall higher toxicity of the Ula River
sample (Table 1). If we exclude Fraction 10 and analyze the remainder of the fractions, it is seen
that the sum of the relative toxicities (defined as 1/ED50) was higher for the Vilnia River sample.
This is largely due to the pronounced difference between the two samples in the toxicity of
Fraction 2 (100% hexane) (Sabaliunas et al. In Press).
The SPMD samples collected in July 1996 were also tested by Microtox™ bioassays.
All the samples effectively inhibited the bioluminescence of V. fischeri with the calculated
30-minute EC50 estimates in the range of 0.052-0.86 ng/ml of the SPMD triolein (or mg/L)
of the bacterial suspension. Some toxicity was also observed in the field blank, and it may be
attributed to toxic impurities in the solvents that were used in large volumes for the SPMD
extraction and dialysis (Figure 5).
20
Kulpe
Pesticide
site
Lower
Neris
Upper
Neris
Add-backs
Bulk sample
Fraction 3
Fraction 2
Fraction 1
Figure 5. Toxicity of SPMD samples collected July 1996 and their size exclusion
chromatography fractions in Microtox™ bioassay (from Sabaliunas et al. In Prep.)
The lower Neris River SPMD dialysate exhibited the highest toxicity (30-minute EC50)
195
-------
in the test (0.052 mg/L), followed by that from the Kulpe River (0.19 mg/L), the Zigmantiskes
pesticide storage site (0.36 mg/L) and the upper Neris River (0.86 mg/L) bulk samples. Testing
of the SEC fractions showed that Fraction 2 (moderate molecular weight compounds) was the
only fraction that was slightly toxic in all the samples. This is in line with the observed
abundance of peaks in this fraction in the SEC chromatograms. However, for the lower Neris
River and the Zigmantis'kes pesticide storage site samples, Fraction 3 was the most toxic
fraction. These were the only samples in which Fraction 3 contained a significant amount of
elemental sulfur. A slight toxicity in Fraction 3 was also observed in the Kulpe River sample.
The toxicity of this fraction correlates well with the presence of the sulfur peak (Sabaliunas et al.
Submitted). Toxicity in Fraction 1 (high molecular weight compounds) was observed only in the
lower Neris and Kulpe River samples. For the latter sample, this fraction was even more toxic
than Fraction 2 (Figure 5).
In terms of relative toxicity for all the samples, excluding the field blank, the sum of
toxicity equivalents (expressed as l/ECs0) of the three fractions was smaller than the number of
toxicity equivalents of the whole sample (Figure 5). To confirm this observation, the three
fractions were combined and re-tested in Microtox™. For samples from the Kulpe River and the
Zigmantiskes pesticide storage site, the toxicity of the "add-backs" (combined three fractions
from SEC) was almost identical to that of the whole samples, whereas for samples from both the
lower and upper Neris River, the "add-backs" were less toxic than the non-fractionated samples
but still the relative toxicity was significantly larger than the sum of toxicity equivalents of the
three fractions. Thus, even though some loss of toxicants during the fractionation cannot be
completely ruled out, these results clearly point to some synergistic toxic action of the complex
mixtures of the SPMD dialysates.
DISCUSSION
The major results of our field studies of river water quality using SPMDs and bioassay-
directed toxicity analysis were: (1) identification of individual organic chemicals: (2) new
problems revealed by bioassays of triolein extracts from the SPMD samples.
The 1995 field studies showed that the major chemicals identified in all samples of
SPMDs from different locations were all PAHs. These PAHs were present also in SPMD
samples from the May 1996 field studies on the Ola and Vilnia Rivers and July 1996 field
studies on the Kulpe and Neris Rivers. The SPMD data add to the general information of PAHs
detected in the surface waters of Lithuanian rivers, as today the laboratories of Joint Research
Center of the Environmental Protection Ministry of Lithuania perform analyses of PAHs by
HPLC (fluorescent detection) techniques on air samples only (EPDLR 1096,1997a, 1997b).
Water samples from the Neris River downstream from Vilnius contained in addition to PAHs,
PCBs (1995 and July 1996) and HCH isomers (1995). The SPMD field studies effectively
concentrated and enabled the identification of organic compounds that have not here-to-for been
reported in official EPM reports (Sabaliunas and Sddergren 1997, Sabaliunas et al. In Press,
Submitted). Similarly, research on organic pollutants was performed in the surface waters of the
196
-------
Nemunas River basin and Kursiu Bay in 1990-1993. This research was performed jointly by
scientists from the USA and the Joint Research Centre of the Environmental Protection Ministry
(Ellington et al. 1994,1996) and reported the presence of chemicals not listed in the Lithuanian
Rivers Water Quality Chronicles (EPDLR 1993, 1996, 1997a) or the Annual Reports of the
Lithuanian EPM (EPDLR 1997b). In the joint USA/Lithuanian survey, solid phase extracts of
organic compounds from water samples were analyzed by GC/FID and GC/ECD. A more
information-rich analysis of the samples was performed by GC-FT-IR-MS (Ellington et al.
1996). In water samples taken from the Neris River downstream from Vilnius (Grigiskes),
some of the chemicals/classes of chemicals found were: (1) alkyl hydrocarbons, (2) aromatic
hydrocarbons, (3) 2,4-dichlorobenzoic acid, (4) dialkylthiophene, (5) 2-(methylthio)-
benzothiazole, (6) phenantrene carboxylic acids, (7) phthalates (diizooctyl-, burylbenzyl-,
di-n-octyl-, di-n-butylphtalates, (9) silicon organics; (8) tetrachlorinated biphenyl carboxyacids
(Ellington et al. 1996). Results from the SPMD field studies in Lithuania reported here have
enlarged the list of organic chemicals identified earlier in the 1990-1993 river water and
sediment studies (Ellington etal. 1994a, 1996, Sabaliunas and Sodergren 1997, Sabaliunas etal.
In Press, Submitted).
The results of silica gel fractionation and gas chromatography analysis of SPMD samples
obtained in May 1996 showed that the Ula River sample had overall three times higher toxicity
as compared to that of the Vilnia River, and that it was due to the highly toxic Fraction 10 which
contained oleic acid (Sabaliunas et al. In Press). Toxicity of fatty acids has been demonstrated
in a number of tests, including Microtox™ (Hakanson et al. 1991). The toxicity of unsaturated
fatty acids is attributable to their membrane disturbing properties: their incorporation into the
lipophilic phase of the double layer of the cell membrane, resulting in increase in membrane
fluidity followed by increased membrane permeability and leakage of cell components (Ewald
and Sundin 1993). It is very likely that by disrupting the membrane structure, oleic acid has also
increased the susceptibility of the cell to some minor, not-identified, components of the toxic
fraction. It has been reported that unsaturated fatty acids, including oleic acid, were responsible
for 87-97% of the toxicity of effluent water from forest product and pharmaceutical industries in
V. fischeri (Svenson et al. 1996). One must note, however, that the possibility of at least some
of the oleic acid being sequestered by SPMDs from the water cannot be completely ruled out.
The oleic acid present in the SPMD samples collected in May 1996 in the Ula and Vilnia Rivers
was probably the hydrolysis product of the SPMD triolein or, even more likely, methyl oleate
which constitutes most of the 5% impurities of commercial triolein used in the SPMDs, and that
readily diffuses to the exterior surface of the SPMD membrane and can be degraded by
microorganisms of periphytic community or abiotically to produce oleic acid (Sabaliunas et al.
In Press). A number of factors, such as the water temperature during the SPMD exposure,
abundance and esterolytic activites of periphytic organisms, oxygen content, light intensity, and
others can affect (bio)degradation. As the concentration of oleic acid increases on the exterior
membrane surface compared to the interior surface, a concentration gradient is established and a
net increase in oleic acid in the SPMD will occur over time. Oleic acid was not detected in the
field blank as the blank SPMDs were stored under conditions not amenable to hydrolysis.
197
-------
If oleic acid is a major contributor to the toxicity of the samples, it may have important
implications for the use of SPMDs in screening the toxicity of hydrophobic pollutants in the
aquatic environment. It may be necessary to refine the sample clean-up procedures or
fractionation (as in this analysis) to eliminate oleic acid prior to toxicity testing in V. fischeri.
The employment of a set of alternative bioassays insensitive to unsaturated fatty acids,
expecially micro-biotests for rapid detection of ecotoxic effects of complex liquid samples, is
of little significance. This is because the main mechanism of action by fatty acids in microbial
biotests overlaps the activities of substituted phenols, organic acids, or other ionogenic organic
compounds by uncoupling of oxidative phosphorylation and permeabilization/fluidization of
membranes. The effect of polar organic compounds acting on membranes (such as fatty acids)
can influence the results of genotoxicity tests, since the physical state of the membranes control
the expression of stress and other genes (Vigh et al. 1998). Similarly, we have demonstrated
(unpublished data) that in anaerobic environments such as some sediments, SPMDs tend to
accumulate elemental hydrophobic sulfur which is mainly of natural origin. Since sulfur is
highly toxic to V. fischeri (Jacobs et al. 1992), it is also necessary to pre-treat such samples for
sulfur removal before testing.
So, in our study the bioassay-directed fractionation using adsorption or SEC applied to
the SPMD samples prior to GC analysis showed that accumulated organic compounds can be
separated into fractions according to their physico-chemical properties, Eind that those fractions
exhibit cumulative toxicity in bioassays (Sabaliunas and Sodergren 1997, Sabaliunas et al.
In Press, Submitted).
CONCLUSIONS
1. During 1995-1996 sampling and analytical chromatography techniques such as
GC/FID, GC/ECD and GC/MS were used for the screening and identification of individual
organic xenobiotics. Chemicals identified in river water from different sampling sites in
Lithuania were: (l) PAHs, (2) PCBs, (3) HCH isomers, (4) alkyl hydrocarbons, (5) phthalates,
and (6) fatty acids. At the sampling places we studied, identification of these compounds had
not previously been reported.
2. Parallel to chemical analysis, all SPMDs field samples extracts were tested in
different bioassays. These were the test of bioluminescence inhibition in V. fischeri
(Microtox™), and the test of mutagenic activity in S. typhimurium. These toxicity tests were
used even prior to chemical analysis in attempt to determine most toxic samples or fractions
(i.e. bioassay-directed fractionation was performed). These fractions were then selectively
analyzed by the more expensive chemical analytical techniques.
3. For an estimation of the true toxic potential of the accumulated different pollutants,
however, sample fractionation or sample clean-up procedures may be required prior to toxicity
testing to remove oleic acid or sulfur.
198
-------
ACKNOWLEDGEMENTS
The authors thank Harry Prest at the University of California, Santa Cruz, for performing
the size exclusion fractionation of the SPMD samples and for his valuable advice and help in
setting the study. We are grateful to Liutauras Stoskus at Vilnius University for his assistance
during deployment of the SPMD's.
REFERENCES
Bulich, A.A. 1984. Microtox - A bacterial toxicity test with several environmental
applications. In: Toxicity Screening Procedures Using Bacterial Systems. D. Liu and
B.J. Dutka (Eds.). Marcel Dekker, Inc., New York, pp 55-64.
Chiou, C.T. 1985. Partition coefficients of organic compounds in lipid-water systems and
correlations with fish bioconcentration factors. Environmental Science and Technology
19:57-62.
Cleveland, L., E.E. Little, J.D. Petty, B.T. Johnson, J.A. Lebo, E. Ora, J. Dionne, and
A. Crockett. 1997. Toxicological and chemical screening of Antarctica sediments:
whole sediment toxicity tests, Microtox™, Mutatox™ and semipermeable membrane
devices (SPMDs). Marine Pollution Bulletin 34:194-202.
Ellington J. J., and T.W. Collette. 1994. Analysis of Fish Tissue from Lithuania by Gas
Chromatography/Fourier Transform-Infrared Spectroscopy/Mass Spectrometry. Internal
Report, ERL-Ath 1994, Environmental Research Laboratory, U.S. Environmental
Protection Agency, Athens, Georgia, 14 pp.
Ellington J. J., J. Sukyte, N. Striupkuviene, and J. Neuman. 1994a. Chemicals identified at two
pesticides storage sites in Lithuania, 1990-1993. In: Environmental Studies in the
Nemunas River Basin, Lithuania. Technical Report 94-1, Fisheries Bioassay Laboratory,
Montana State University, Bozeman, Montana, USA. pp. 39-58.
Ellington J. J., R.V. Thurston, J. Sukyte, and K. Kvietkus. 1994b. Hazardous chemicals in
surface water of Lithuania. In: Environmental Studies in the Nemunas River Basin,
Lithuania. Technical Report 94-1, Fisheries Bioassay Laboratory, Montana State
University, Bozeman, Montana, USA. pp. 1-18.
Ellington J. J., R. V. Thurston, J. Sukyte, and K. Kvietkus. 1996. Hazardous chemicals in waters
of Lithuania. Trends in Analytical Chemistry 15:215-224.
EPDLR (Environmental Protection Department of Lithuania Republic). 1993. Lithuanian
Rivers Water Quality Chronicle of Year 1993. Vilnius, 3.6 pp. (In Lithuanian).
199
-------
EPDLR (Environmental Protection Ministry of Lithuania Republic). 1996. Environmental
Protection in Lithuania. Year 1995 Annual Report. Vilnius, 133 pp. (In Lithuanian).
EPDLR (Environmental Protection Ministry of Lithuania Republic). 1997a. Lithuanian Rivers
Water Quality Chronicle of Year 1996. Vilnius, 64 pp (In Lithuanian).
EPDLR (Environmental Protection Ministry of Lithuania Republic). 1997b. Environmental
Protection in Lithuania. Annual Report. Vilnius, 132 pp. (In Lithuanian).
Ewald, G., and P. Larsson. 1995. Partitioning of 14C-labelled 2,2',4,4'-tetrachlorobiphenyl
between water and fish lipids. Environmental Toxicology and Chemistry 13:1577-1580.
Ewald, G., and P. Sundin. 1993. ATP leakage from ELD cells after exposure to stearic,
monochlorostearic, dischlorostearic, and oleic acids. Pharmacology and Toxicology
73:159-162.
Hakanson, H., P. Sundin, B. Brunstrom, L. Dencker, M. Engwall, G. Ewald, M. Gilek, H. Holm,
S. Hpnkasalo, J. Idestam-Almquist, P. Jonsson, P. Kautsky, G. Lundberg, A. Lund-
Kvernheim, K. Martinsen, L. Norrgren, M. Personen, M. Rundgren, M. Stalberg, M.
Tarkpea, and C. Wesen. 1991. In vivo and in vitro toxicity of fractionated fish lipids,
with particular regard to their content of chlorinated organic compounds. Pharmacology
and Toxicology 69:459-471.
Huckins, J.N., J.D. Petty, J.A. Lebo, C.E. Orazio, H.F. Prest, D.E. Tillitt, G.S. Ellis,
B.T. Johnson, and G.K. Manuweera. 1996. Semipermeable membrane devices (SPMDs)
for the concentration and assessment of bioavailable organic contaminants in aquatic
environments, pp. 625-655. In: Techniques in Aquatic Toxicology, G.K. Ostrander (Ed.).
CRC-Lewis Publishers, Boca Raton, Florida, USA.
Huckins, J.N., M.W. Tubergen, and G.K. Manuweera. 1990. Semipermeable membrane
devices containing model lipid: a new approach to monitoring the bioavailability of
lipophilic contaminants and estimating their bioconcentration potential. Chemosphere
20:533-552.
I
Hwang S.T., and K.Kammermeyer (Eds.). 1975. Membranes in Separations. Robert E. Krieg
Publishing. Malabar, Florida, USA. 559pp.
ISO/CD (International Standard Organization/Current Document). 1994. Water Quality -
Determination of the Inhibitory Effect of Water Samples on the Light Emission of Vibrio
fischeri (Luminescent bacteria test). Document No. 11348.
Jacobs, M.W., J.J. Delfino, and G. Bitton. 1992. The toxicity of sulfur to Microtox™ from
acetonitrile extracts of contaminated sediments. Environmental Toxicology and
Chemistry 11:1137-1143.
200
-------
Johnson G.D. 1990. Hexane-filled dialysis bags for monitoring organic contaminants in water.
Environmental Science and Technology 25:1897-1903.
Maron D., and B.N. Ames. 1983. Revised methods for the salmonella mutagenicity test.
Mutation Research 13:173-215.
Opperhuizen, A., E.W. van der Velde, F.A. Gobas, A.K. Liem, J. van der Steen, and
O. Hutzinger. 1985. Relationship between bioconcentration in fish and steric factors
of hydrophobic chemicals. Chemosphere 14:1871-1896.
Prest, H.F., B.J. Richardson, L.A. Jacobson, J. Vedder, and M. Martin. 1995. Monitoring
organochlorines with semipermeable membrane devices and mussels (Mytilus edulis)
in Corio Bay, Victoria, Australia. Marine Pollution Bulletin 30:543-554.
Sabaliunas D., J. J. Ellington, and I. Sabaliuniene. (In Press). Screening bioavailable
hydrophobic toxicants in surface waters with semipermeable membrane devices: role
of inherent oleic acid in toxic evaluations. Ecotoxicology and Environmental Safety.
Sabaliunas D., J. Lazutka, and I. Sabalixlniene. (Submitted). Toxicity and genotoxicity of
aquatic hydrophobic pollutants sampled with semipermeable membrane devices.
Environmental Pollution.
Sabaliunas D., and A. Sodergren. 1997. Use of semi-permeable membrane devices to monitor
pollutants in water and assess their effects: a laboratory test and field verification.
Environmental Pollution 96:195-205.
Sodergren, A. 1987. Solvent-filled dialysis membranes simulate uptake of pollutants by aquatic
organisms. Environmental Science & Technology 21:855-859.
Svenson, A., H. Norin, and P.A. Hynning. 1996. Toxicity-directed fractionation of effluents
using the bioluminescence of Vibrio fischeri and gas chromatography-mass spectroscopy
identification of toxic components. Environmental Toxicology and Water Quality
11:277-284.
Vigh L., B. Maresca, and J.L. Harwood. 1998. Does the membrane's physical state control the
expression of heat shock and other genes? Trends in Biochemical Sciences 23:369-377.
201
-------
-------
HYPOXIA IN COASTAL WATERS: PRESSING PROBLEMS WORLDWIDE
AND THEIR SCIENTIFIC CHALLENGES
R.S.S. Wu
ABSTRACT
Large scale hypoxia has become a major problem in coastal waters world wide, and is
certain to be exacerbated in the coming years. Progressive eutrophication and decreases in
dissolved oxygen have been observed in many coastal waters and embayments, and many coastal
areas are suffering from permanent or periodic anoxia/hypoxia. Large scale hypoxia/anoxia has
already led to major changes in structure and function of coastal systems and the mass mortality
of fishes and benthos. Both the severity and frequency of large scale hypoxia in many coastal
areas appears to be increasing. Most benthic fauna have adapted to hypoxia, and can tolerate it
(at least at -2.8 mg/L O2) for prolong periods without apparent adverse effects, while hypoxic
tolerance by fishes appears to be much more variable. While much is known about the
biochemical and physiological responses to hypoxia, only scanty information exist on the effects
of hypoxia on the ecology of natural populations and communities. Limited studies have shown
that hypoxia may change the structure of benthic and fish communities and alter trophic
relationships, which may have serious ecological and fisheries implications. Recovery from
hypoxia of benthic communities in temperate regions may take several years, while recovery
may be quicker in the sub-tropics. Developing cost effective pollution control technologies will
be one of the major challenges in the years to come. Deep sea disposal and wetland treatment
appears to be cost effective. To this end, the carrying capacity and "no observable effect
concentration" of coastal waters and wetland systems to organic waste and nutrients needs to be
determined. The development of specific biomarkers may provide a more cost-effective
technology for frequent monitoring of dissolved oxygen over large areas. Very few attempts
have been made to relate hypoxia to Darwinian fitness traits, or to relate biochemical/
physiological responses to significant ecological effects. Future research in this area is required
for the better prediction of the ecological risks of hypoxia.
INTRODUCTION
Hypoxia is generally defined as dissolved oxygen levels below 2 ml/L C>2 (equivalent
to 2.8 mg/L O2 or 47.8 mm Hg). Hypoxia may be a natural phenomenon caused by the
formation of a thermocline, halocline, or pynocline in a water body. Paleological evidence
indicates the occurrence of hypoxia in pre-historical ages (Tyson and Pearson 199 la, 1991b).
In fact, all marine sediments are anoxic a few mm below sediment surface. Under normal
circumstances, bioturbation is the major mechanism for oxygen replenishment in sediment
(Schaffner et al. 1992). Anthropogenic activities contributing to organic pollution and nutrient
'Centre for Coastal Pollution and Conservation, City University of Hong Kong, Hong Kong, SAR, China.
203
-------
enrichment/eutrophication have led to large scale hypoxia over large areas. Some marine
systems (e.g., the bottom waters of the Black Sea) are permanently anoxic, while hypoxia or
periodic hypoxia are common in many coastal systems (Diaz and Rosenberg 1995).
Large scale hypoxia has become a major problem in coastal waters world wide, and this
problem is likely to be exacerbated in coming years for the following reasons. First, some
65% of existing cities of >2.5 million people are presently located along ocean coasts. By the
year 2000, 70% of the world's population (some 4.2 billion) will live within 60 km of coast lines
(UNEP 1991). The volume of wastewater generated by such a huge population is enormous.
It should be noted that wastewater treatment (especially nutrient removal) is expensive.
For example, the cost of secondary treatment (which removes approximately 90% BOD and
30-40% of N and P) is some three to four times more expensive than primary treatment.
Due to the high capital and recurrent costs incurred, it is highly unlikely that construction for
sewage treatment facilities can match population and GNP growths, at least in developing
countries. This implies that an enormous amount of untreated wastewater and nutrients will be
discharged into many coastal waters in the coming years.
Second, large scale clearing of land vegetation and deforestation are occurring at an
alarming rate in many part of the world, and there is no good evidence to suggest that levels of
these activities will decrease in the coming years (World Resources Institute 1992). Runoff of
nutrients from crop and farm lands are typically an order of magnitude greater than that from
pristine forest, and such a large nutrient input from terrestrial runoff has already greatly altered
the geological cycle (Gabric and Bell 1993).
Third, intensive farming practices in the last few decades have added a significant load
of nutrients into coastal waters. For example, marine fish farming activities have increased
dramatically in the last two decades, and production of marine culture fish is expected to increase
further to 1 million ton year"' by 2025 (New and Csavas 1995). It should be noted that some
80% of N input into the fish culture systems is lost to the environment, mainly through feed
wastage and fish excreta (Gowen and Bradbury 1987, Handy and Poxton 1993), and the total
nutrient input from increased mariculture activities is significant.
The problem of organic and nutrient enrichment is further augmented by the fact that
many of the nutrient and organic inputs are derived from non-point sources. Atmospheric fall-
out of N is also significant (constituting some 10-50% of the total anthropogenic N input) and
is expected to increase further in the coming years (Paerl 1993). This means that control of
discharges is much more difficult.
GLOBAL SITUATION AND TREND
Since 1970, progressive eutrophication and decreases in dissolved oxygen have been
observed in the Baltic Sea, North Sea, Adriatic Sea, along the Chinese coast, and in the Japan
Sea (GESAMP 1990). For example, P and N in the coastal waters of Germany has increased
by 1.5 and 1.7 times respectively over the last 25 years (GESAMP 1990); N and P in the Black
Sea had increased 5 and 20 times respectively from the 1960s to the 1980s (Gomoiu 1992);
204
-------
levels of P and N in the Baltic Seas have increased some 1.5 to 4.5 times from 1970 to 1990
(HELCOM 1996); P and N in the Dutch Seas increased 4 times and 2 times respectively from the
1930s to 1980s (GESAMP 1990); and input of N and P into coastal waters of Queensland,
Australia has increased by 3 to 5 times in the last 65 years (Moss et al. 1992, Bell and Elmetri
1995). It is therefore not surprising that decreases in dissolved oxygen have been reported in
many coastal waters of the USA, Norway, UK, Sweden, Denmark, the Black Sea, and the
Adriatic Sea in the last 30 to 80 years (Mirza and Gray 1981, Justic et al. 1987, Weigelt 1990,
Diaz and Rosenberg 1995). Dissolved oxygen levels < 4 mg/L were found in about 40%
of 130 estuarine and coastal areas throughout USA (Whitledge 1985), and many coastal areas
and embayments in Europe, Scandinavia, North America, and China are suffering from
permanent or periodic anoxia/hypoxia (Wu 1982, Kuo and Neilson 1987, Diaz and Rosenberg
1995). Indeed, large scale hypoxia/anoxia has already led to major changes in structure and
function of coastal systems. Mass mortality of fishes and benthos have also commonly been
reported in coastal areas all over the world, and both the severity and frequency of large scale
hypoxia in many coastal areas appears to be increasing (GESAMP 1990, Diaz and Rosenberg
1995).
BIOLOGICAL AND ECOLOGICAL RESPONSES TO HYPOXIA
Since hypoxia occurs commonly in bottom waters, most benthic fauna have adapted,
and can tolerate hypoxia (at least at -2.8 mg/L O2) for prolonged periods without apparent
adverse effects (Diaz and Rosenberg 1995). However, oxygen concentration becomes critical to
benthos at around 2 mg/L, and acute effects generally occur within a very narrow range (between
1 to 2 mg /L ) (Rosenberg et al 1992, Diaz and Rosenberg 1995). Hypoxic tolerances of fishes,
however, appear to be much more variable. For example, LT50 values at 1 mg/L O2 of the red
snapper (Chrysophrys major) is only 60 minutes, while the sea bass (Lates calcarifer) and the
green grouper (Epinephelus tauvind) can tolerate the same level of hypoxia for more than
8 hours without any apparent behaviorial changes (Wu 1990). Active species and predators
appear to be less tolerant than sessile species (Wu 1982, Rosenberg et al. 1992, Diaz and
Rosenberg 1995).
Many studies have been carried out to investigate the biochemical, physiological and
organismal responses to hypoxia. In general, several strategies have evolved to deal with
hypoxia. Avoidance of hypoxic waters has been reported in fishes and crustaceans under both
laboratory and field conditions (Pihl et al. 1991, Diaz et al: 1992). For example, the shrimp
Metapenaeus ensis has been shown to avoid hypoxic waters and move towards normoxic water
(Wu et al. In Prep.). Animals may tolerate or adapt to hypoxia by reducing their activities and
metabolism, so as to reduce their energy expenditure and hence oxygen requirements (Hill et al.
1991, Dalla-Via et al 1994). Facing falling oxygen concentrations, animals may enhance
oxygen uptake and delivery of oxygen to essential tissues. This may be achieved through
increase in red blood cell production, increase in hemoglobin-oxygen affinity, and increase in
ventilation and oxygen transport. If hypoxia persists for prolonged periods, animals may resort
to enhanced supply of oxygen from anaerobic sources (Holeton and Randall 1967a, 1967b,
Buggren and Randall 1978, Woo and Wu 1984, Dunn and Hochachka 1986, Chew and Ip 1992,
Randall et al. 1992).
205
-------
While much is known about the biochemical and physiological responses to hypoxia,
limited studies have been carried out to determine the effects of hypoxia on natural populations
and communities. Hypoxia may eliminate sensitive species, thereby limiting their natural
distribution and changing the structure (species composition, dominance, and diversity) of
benthic and fish communities (Rosenberg et al. 1992, Diaz and Rosenberg 1995). Persistent
hypoxia over large areas has been shown to alter trophic relationships (Wu 1982, Kimor 1992a,
Breitburg 1992), which may lead to major ecological consequences. Pihl (1994) related the shift
from demersal to pelagic fish to hypoxia in bottom waters. Such a shift may be caused by
elimination of sensitive demersal species or their natural prey items, or a combination of both.
Occurrence of hypoxia appears to favor the dominance of small size prey with a shortened life
cycle (opportunistic species) as well as small size fish (Harper et al. 1991, Diaz and Rosenberg
1995).
1 ; ' , . : • . ; . : - . , ., "', . -.'.
The above changes may have long term fisheries implications. A decrease in dominance
of predators, reflecting fundamental changes in trophic structure of a benthic community has
been demonstrated along a gradient of decreasing oxygen levels (Wu 1982). Various studies
have reported the emergence of benthos from their burrows during hypoxia (Baden et al. 1990,
Vismann 1990, Tyson and Pearson 1991b, Nilsson and Rosenberg 1994). Weigelt and Rumohr
(1986) reported mass catches of moribound polychaetes and migration of fishes into nearshore
waters during oxygen depletion. Such behavioral changes may make animals much more
vulnerable to their natural predators. Indeed, Breitburg et al. (1994) demonstrated an increase
in predation on the goby Gobiosoma bosc by the sea nettle and a corresponding decrease in
predation by the stripped seabass and adult gobies. Likewise, Pihl et al. (1992) demonstrated a
general shift of natural diet of demersal fish from crustaceans to polychaetes and echinoderms in
a hypoxic coastal system. Effects of such dietary change on the structure and function of
ecosystems and fisheries production, however, remained unknown.
RECOVERY
Recovery of natural communities from hypoxia is poorly known, although such
information is obviously important from an environmental management point of view.
Limited studies in Europe and Scandinavia have shown that recovery of benthic communities
after hypoxia caused by algal bloom may take some 2 to 8 years (Diaz and Rosenberg 1995).
By 18 months after a major hypoxic event, macrobenthic communites in Gullmarsfjord, Sweden,
had not been re-established (Josefson and Widbom 1988). Likewise, little or no recovery was
found after 2 years in the Kattegat arm of the North Sea after benthic defaunation caused by
hypoxia (Rosenberg et al. 1992), and the nematode community in Swedish waters had not
returned to its original state a year after complete defaunation caused by hypoxia (Austen and
Wibdom 1991). Studies carried out in the sub-tropics, however, provide evidence that recovery
may be much quicker and may only take a few months to a year, although it is not sure whether
the community has fully been restored to the original state (Wu 1982, Gamenick et al. 1996,
Wu and Shin 1998, Lu and Wu 1998). Diaz and Rosenberg (1995), however, are of the view that
no large marine ecosystem system has so far been known to recover after development of
persistent hypoxia or anoxia.
206
-------
CHALLENGES
Development of Cost Effective Pollution Control Technologies
There is little argument that the best way to prevent or minimize the occurrence of
hypoxia is by source reduction of nutrients and organic inputs. To this end, substantial changes
in current land use and farming practices are required. Source reduction of nutrient and waste is,
however, very expensive. For example, reduction of N by 50% from a population of 85 million
people in the Baltic region is estimated at US $20 billion (HELCOM 1996). While there is no
shortage of pollution control technology for removing organic waste and nutrients from
wastewater, the main problems lie with cost-effectiveness, especially considering that the
volume of wastewater to be treated is enormous. No significant adverse effects have so far been
demonstrated for deep sea disposal, so such disposal methods should remain a practical and cost
effective option. The same, however, cannot be said for disposing wastewater in coastal waters
with limited carrying capacity. The trigger level of the Baltic Sea from oligotropy to eutrophy
is estimated at somewhere between 0.05-0.25 g of P and 1-3.5 g of N per m2 per year
(Vollenweider 1992). The trigger level of algal blooms in Hong Kong coastal waters is
estimated at 1 mg/L N (Wu et al. 1999). Although the concept of no observable effect
concentration (NOEC) has been used extensively in ecotoxicology, there are few studies that
apply such a concept to the determination of carrying capacity, and to the NOEC of coastal
waters to organic waste and nutrients.
Wetlands are an effective means for removing organic waste and nutrients from
wastewater. It has been shown that natural wetland can remove some 70-90% BOD, 94% of P
and 77% of N from wastewater (Patruno and Russel 1994, Thomas et al. 1995, Hiley 1995).
The major limitation of this treatment method is the long retention time and low loading rate,
and the long term viability of using wetlands for treating large volumes of wastewater has yet to
be firmly established. Nevertheless, the use of wetlands for nutrient and organic waste removal
may be an attractive option for developing countries where construction of treatment facilities
may not be affordable.
Development of Cost Effective Monitoring Techniques
Regular monitoring of dissolved oxygen levels is essential to provide an early warning
and to protect the health of coastal systems. Currently, this is done by regular measurement of
dissolved oxygen in water (using oxygen electrodes or chemical methods such as Wrinkler
titration). Since daily variations of oxygen levels are typically large, and hypoxic events are
unpredictable over a large area, it is not cost effective to monitor oxygen frequently over a large
region and over a long period. Recently, biomarkers have been developed to provide time-
integrated information on the levels of contaminants in the aquatic environment, mainly trace
metals and xenobiotics such as PAHs, organophosphates, and organochlorines (Peakall 1992,
Depledge 1993). The use of biomarkers as an early warning system in pollution monitoring not
only alleviates the difficulties of frequent and large scale sampling, but also provides information
on biological effects. Although there have been many studies of the responses offish to hypoxia
(e.g. Holton and Randall 1967a, 1967b, Burggren and Randall 1978, Wu and Woo 1985,
207
-------
Dunn and Hpchachka 1986, Val et al 1995), only limited attempts have been made to identify
possible biomarkers for hypoxia in aquatic systems (Foerlin et al. 1996). Field transplantation
experiments carried out by Wu and Lam (1997) indicated that activities of glucose-6-phosphate
dehydrogenase and lactate dehydrogenase (LDH) in adductor muscles of the green lipped mussel
(Perna virvidis) showed a significant, negative correlation with dissolved oxygen levels at the
transplantion sites.
To be useful in environmental monitoring, biomarkers should be specific, reasonably
long lasting, and easy to determine (Peakall 1992). Although many physiological and
biochemical changes in organisms can be associated with hypoxia, most of these changes are
non-specific. For example, the activities of LDH in many fishes not only increase during
hypoxia, but also increase with a fall in temperature (Pierce and Crawford 1997). Thus,
measurement of LDH alone would not be a reliable biomarker for hypoxia, unless all other
factors are constant. Although much is known about the responses of fishes to hypoxia, it is not
clear which response is specific to hypoxia. Recent studies showed certain biochemical changes
(e.g. Epiotein and Hypoxic Inducible Factor 1) in mammalian tissue culture may be specific to
hypoxia (Wang and Semenza 1993, Jelkmann 1994), and the use of these biochemical changes
as specific hypoxic biomarkers in practical monitoring should be explored.
Relating Biochemical and Physiological Responses to Ecological Effects
While much is known about the tolerance of marine animals to acute hypoxia, as well
as biochemical and physiological responses, surprisingly, the effects of hypoxia (especially long
term hypoxia) on growth and reproduction of marine animals are poorly understood. A recent
study by Keckeis et al. (1996) showed reduced hatching success and deformation in nace
(Chondrostoma nasus) when exposed to hypoxia. Thus far, very few attempts have been made
to relate hypoxia to Darwinian fitness traits (e.g. growth and reproduction), or to relate observed
biochemical/physiological responses of marine animals to population effects and significant
ecological consequences. The paucity of data, therefore, does not permit us to extrapolate many
of the observed biochemical and physiological responses to population consequences in
ecological risk assessment. Research contributing to such information gaps would be extremely
useful in environmental management.
REFERENCES
Austen, M.C., and B. Wibdom. 1991. Changes in and slow recovery of~a meiobenthic nematode
assemblage following a hypoxic period in the Gullmar fjord basin, Sweden. Marine
Biology 111:139-145.
Baden, S.P., L.O. Loo, L. Phil, and R. Rosenberg. 1990. Effects of eutrophication on benthic
communities including fish: Swedish west coast. Ambio 19:113-122.
Bell, P.R.F., and I. Elmetri. 1995. Ecological indicators of large-scale eutrophication in the
Great Barrier Reef lagoon. Ambio 24: 208-215.
208
-------
Breitburg, D. L. 1992. Episodic hypoxia in Chesapeake Bay: interacting effects of recruitment,
behaviour, and physical disturbance. Ecological Monographs 62:525-546.
Breitburg, D.L., N. Steinberg, S. DuBeau, C. Cooksey, and E.D. Houde. 1994. Effects of low
dissolved oxygen on predation on estaurine fish larvae. Marine Ecology Progress Series
104:235-246.
Burggren, W.W., and D.J. Randall. 1978. The pattern of carbon dioxide excretion in the
rainbow trout Salmo gairdneri. Journal of Experimental Biology 72:17-24.
Chew, S.F., and Y.K. Ip. 1992. Biochemical adaptations of the mudskipper Boleophthalmus
boddaerti to a lack of oxygen. Marine Biology 112:567-571.
Dalla-Via, J., G. VnD. Thillart, O. Cattani, and A.D. Zwaan. 1994. Influence of long term
hypoxia exposure on the energy metabolism ofSolea solea. 2. Intermediary metabolism
in blood, liver and muscle. Marine Ecology Progress Series 111: 17-27.
Depledge, M.H. 1993. Ecotoxicology: a science or a management tool? Ambio 22:51-52.
Diaz, R.J., R.J. Neubauer, L.C. Schaffner, L. Pihl, and S.P. Baden. 1992. Continuous
monitoring of dissolved oxygen in an estuary experiencing periodic hypoxia and the
effect of hypoxia on macrobenthos and fish. The Science of the Total Environment.
Supplement 1992: 1055-1068.
Diaz, R.J., and R. Rosenberg. 1995. Marine benthic hypoxia: a review of its ecological effects
and the behavioural responses of benthic macrofauna. Oceanography and Marine
Biology: An Annual Review 33:245-303.
Dunn, J.F., and P.W. Hochachka. 1986. Turnover rates of glucose and lactate in rainbow
trout during acute hypoxia. Canadian Journal of Zoology 65:1144-1148.
Foerlin, L., S. Phil-Baden, S. Eriksson, A. Granmo, E. Lindesjoeoe, K. Magnusson, R. Ekelund,
A. Esselin, and J. Sturve. 1996. Effects of contaminants in roundnose grenadier
(Coryphaenoides rupestris) and Norway lobster (Nephrops norvegicus) and contaminant
levels in mussels (Mytilus edulis) in the Skagerrak and Kattegat compared to the Faroe
Island. Journal of Sea Research 35:209-222.
Gabric, A.J., and P.R.F. Bell. 1993. Review of the effects of non-point nutrient loading on
coastal ecosystems. Australian Journal of Marine and Freshwater Research 44:261-283.
Gamenick, L, A. Jahn, K. Vopel, and O. Giere. 1996. Hypoxia and sulphide as structuring
factors in a macrozobenthic community in the Baltic Sea shore: Colonization studies and
tolerance experiments. Marine Ecology Programming Series 144:73-85.
209
-------
GESAMP (Joint Group of Experts on the Scientific Aspect of Marine Pollution). 1990.
The State of the Marine Environment UNEP Regional Seas Reports and Studies.
No.ll5,UNEP.
Gomoiu, M. 1992. Marine eutrophication syndrome in the north-western part of the Black Sea.
In: Marine Coastal Eutrophication. R.A. Vollenweider, R. Marchetti and R. Vicviani
(Eds.). Elsevier, Amsterdam. 683-696 pp.
Gowen, R.J., and N.B. Bradbury. 1987. The ecological impact of salmonid farming in coastal
waters: a review. Oceanography. Marine Biology Annual Review 25:263-575.
Handy, R.D., and M.G. Poxton. 1993. Nitrogen pollution in mariculture: toxicity and excretion
of nitrogenous compounds by marine fish. Review of Fisheries Biology 3:205-241.
Harper, D.E., L.D. McKinney, J.M. Nance, and R.B. Salzer. 1991. Recovery responses of two
benthic assemblages following an acute hypoxic event on the Texas continental shelf,
northwestern Gulf of Mexico. In: Modern and Ancient Continental Shelf Anoxia.
R. V. Tyson and T.H. Pearson (Eds.). London: Geological Society. Special Publication
58. pp. 49-64.
HELCOM (Helsinki Commission). 1996. Third Periodic Assessment on the State of the Marine
Environment of the Baltic Sea, 1989-93 Background Document. Baltic Marine
Environment Protection Commission, Baltic Sea Environment Proceedings No. 64B.
252pp.
Hiley, P.D. 1995. The reality of sewage treatment using wetlands. Water Science and
Technology 32:329-338.
Hill, A.D., A.C. Taylor, and R.H.C. Strang. 1991. Physiological and metabolic responses of the
shore crab Carcinus meenas (L.) during environmental anoxia and subsequent recovery.
Journal of Experimental Marine Biology and Ecology 150:31-50.
Holeton, G.F., and D.J. Randall. 1967a. Changes in blood pressure in the rainbow trout during
hypoxia. Journal of Experimental Biology 46:297-305.
Holeton, G.F., and D.J. Randall. 1967b. The effect of hypoxia upon the partial pressure of
gasses in blood and water afferent and efferent to the gills of rainbow trout. Journal of
Experimental Biology 46: 317-327.
Jelkmann, W. 1994. Biology of erythropoietin. Clinical Investigator 72:S3-S 10.
Josefson, A.B., and B. Widbom. 1988. Differential response of benthic macrofauna and
meiofauna to hypoxia in the Gullmar Fjord basin. Marine Biology 100:31-40.
210
-------
Justic, D., T. Legovic, and Sandrini Rottini. 1987. Trends in oxygen content 1911-1984 and
occurrence of benthic mortality in the northern Adriatic Sea. Estuarine and Coastal
Shelf-Science 25:435-445.
Keckeis, H., E. Bauer-Nemeschkal, and E. Kamler. 1996. Effects of reduced oxygen level on
the mortality and hatching rate of Chondrostoma nasus embryos. Journal of Fish Biology
3:430-440.
Kimor, B. 1992a. Impact of eutrophication on phytoplankton composition. In: Marine Coastal
Eutrophication. R.A. Vollenweider, R. Marchetti, andR. Vicviani (Eds.). Elsevier,
Amsterdam, pp. 871-878.
Kimor, B. 1992b. The impact of eutrophication on phytoplankton composition in coastal marine
ecosystems. The Science of the Total Environment. Supplement 1992: 871-878.
Kuo, A. Y., and B. J. Neilson. 1987. Hypoxia and salinity in Virginia estuaries. Estuaries
10:277-283.
Lu, L., and R.S.S. Wu. 1998. Benthic recolonization on organic polluted sediment.
Environmental Pollution 101:241 -251.
Mirza, F.B., and J. S. Gray. 1981. The fauna of benthic sediments from the organically enriched
Oslofjord, Norway. Journal of Experimental Marine Biology and Ecology 54:181-207.
Moss, A.J., G.E. Rayment, N. Reilley, and E.K. Best. 1992. A preliminary assessment of
sediment and nutrient export from Queensland coastal catchment. Report of Queensland,
Department of Environment and Heritage and Department of Primary Industry, Brisbane,
Australia. 33 pp.
New, M.B., and I. Csavas. 1995. Aquafeed in Asia: a regional review. Farm Made Aquafeed.
Food and Agricultural Organization of the United Nations, Technical Paper No. 343.
24 pp.
Nilsson, H., and R. Rosenberg. 1994. Hypoxic response of two marine benthic communities.
Marine Ecology Progress Series 115:209-217.
Paerl, H.W. 1993. Emerging role of atmospheric nitrogen deposition in coastal eutrophication:
biogeochemical and trophic perspectives. Canadian Journal of Fisheries and Aquatic
Science 50:2254-2269.
Patruno, J., and J. Russel. 1994. Natural wetland polishing effluent discharging to Wooloweyah
Lagoon. Water Science Technology 29:4.
Peakall, D. 1992. Animal Biomarkers as Pollution Indicators. Chapman and Hall, London.
211
-------
Pierce, V.A., and D.L. Crawford. 1997. Phylogenetic analysis of glycolytic enzyme expression.
Science 276:256-259.
Pihl, L. 1994. Changes in the diet of demersal fish due to eutrophication-induced hypoxia in the
Kattegat, Sweden. Canadian Journal of Fisheries and Aquatic Science 51:321-336.
Pihl, L., S.P. Baden, and R. J. Diaz. 1991. Effects of periodic hypoxia on distribution of
demersal fish and crustaceans. Marine Biology 108:349-360.
Pihl, L., S. P. Baden, R. J. Diaz, and L.C. Schaffner. 1992. Hypoxia-induced structural changes
in the diet of bottom-feeding fish and Crustacea. Marine Biology 112:349-361.
Randall, D.J. D. J. McKenzie, G. Abrami, G.P. Boniolotti, F. Natiello, P. Bronzi, L. Bolis,
and E. Agradi. 1992. Effects of diet on responses to hypoxia in sturgeon (Acipemer
naccarif). Journal of Experimental Biology 170:113-125.
Rosenberg, R., L.O. Loo, and P. Moeller. 1992. Hypoxia, salinity and temperature as
structuring factors for marine benthic communities in a eutrophic area. Netherlands
Journal of Sea Research 30:121-129.
Schaffner, L.C., P. Jonsson, R.J. Diaz, R. Rosenberg, and P. Gapcynski. 1992. Benthic
communities and bioturbation history of estuarine and coastal systems: effects of
hypoxia and anoxia. In: Marine Coastal Eutrophication. R.S. Vollenweider,
R. Marchetti, and R. Viviani (Eds.). The Science of the Total Environment.
Supplement 1992, pp. 1001-1016.
Thomas, P.R., P. Glover, and T. Kalaroopan. 1995. An evaluation of pollutant removal from
secondary treated sewage effluent using a constructed wetland system. Water Science
and Technology 32:87-93.
Tyson, R.V., and T.H. Pearson. 1991a. Modern and ancient continental shelf anoxia. London:
Geological Society, Special Publication 58.
Tyson, R.V., and T.H. Pearson. 1991b. Modern and ancient continental shelf anoxia: an
overview. In: Modern and Ancient Continental Shelf Anoxia. R. V. Tyson and
T.H. Pearson (Eds.). London: Geological Society, Special Publication 58 pp. 1-24.
UNEP (United Nations Environmental Program). 1991. The State of the World Environment.
Nairobi, 48 pp.
Val, A.L., J. Lessard, and DJ. Randall. 1995. Effects of hypoxia on rainbow trout
(Oncorhynchus mykiss): Intraerythrocytic phosphates. Journal of Experimental Biology
198:305-310.
212
-------
Vismann, B. 1990. Sulfide detoxification and tolerance in Nereis (Hediste) diversicolor and
Nereis (Neanthes) virens (Annelida: Polychaeta). Marine Ecology Progress Series
59:229-238.
Vollenweider, R.A. 1992. Coastal marine eutrophication: principles and control. In: Marine
Coastal Eutrophication. R.A. Vollenweider, R. Marchetti, and R. Vicviani (Eds.).
Elsevier, Amsterdam, pp. 1-20.
Wang, G.L., and G.L. Semenza. 1993. General involvement of hypoxia-inducible factor 1 in
transcriptional response to hypoxia. Proceeding of the National Academy of Sciences of
the United States of America 90:4304-4308.
Weigelt, M. 1990. Oxygen conditions in the deep water of Kiel Bay and the impact of
inflowing salt-rich water from the Kattegatt. Meeresforschung 33:1-22.
Weigelt, M., and H. Rumohr. 1986. Effects of wide range oxygen depletion on benthic fauna
and demersal fish in Kiel Bay 1981-1983. Meeresforschung 31:124-136.
Whitledge, T.E. 1985. National Review of Oxygen Depletion and Eutrophication in Estuarine
and Coastal Water. Informal Report BNL 37144, National Oceanic and Atmospheric
Administration, Rockville, Maryland, USA.
Woo, N. Y.S., and R.S.S. Wu. 1984. Changes in body composition in the red grouper
(Epinephelus akaard) and the black sea bream (Mylio macrocephalus) during hypoxic
exposure. Comparative Biochemistry and Physiology 77A:475-482.
World Resources Institute. 1992. World Resources 1992-93. Oxford University Press.
Oxford. 369 pp.
Wu, R.S.S. 1982. Periodic defaunation and recovery in a subtropical epibenthic community, in
relation to organic pollution . Journal of Experimental Marine Biology and Ecology
64:253-269.
Wu, R.S.S. 1990. Environmental tolerance of some marine fish: applications in mariculture
management. In: Proceedings of an International Symposium on Fish Physiology, Fish
Toxicology, and Fisheries Management, Guangzhou, PRC, September 1988, R.C. Ryan
(Ed.). U.S. Environmental Protection Agency, Environmental Research Laboratory,
Athens, Georgia, USA. EPA/600/9-90/011, pp. 197-207.
Wu, R.S.S., and P.K.S. Lam. 1997. Lactate dehydrogenase and glucose-6-phosphate
dehydrogenase in the green mussel Perna virvidis: biomarkers for hypoxia in the marine
environment. Water Research 31:2797-2801.
Wu, R.S.S., and P.K.S. Shin. 1998. Sediment characteristics and colonization of soft-bottom
benthos: a field manipulation experiment. Marine Biology. 128:475-487.
213
-------
Wu, R.S.S, P.K.S., Shin, D.W. Mackay, M. Mollowney, and D. Johnson. 1999. Management
of marine fish farming in the sub-tropical environment: a modeling approach. Aquaculture.
174:279-298.
Wu, R.S.S., K.L. Wan, and P.K.S. Lam. (In Preparation). Behavioral and cardiac responses of
the shrimp Metapenaeus ensis to hypoxia.
Wu, R.S.S., and N.Y.S. Woo. 1985. Respiratory responses and tolerance to hypoxia in two
marine teleosts, Epinephelus akaara (Temminck & Schlegal) and Mylio macrocephalus
(Basilewsky). Hydrobiologia 119:209-217.
214
-------
TOWARDS PREDICITNG METAL ECOTOXICOLOGY: APPLYING
COORDINATION THEORY, SURFACE CHEMISTRY, AND
SIMULATION MODELS
George W. BaileyWd Z. Z. Zhang2
ABSTRACT
Predicting metal toxicity in fishes requires relating metal properties and speciation, ligand
properties, environmental parameters, the resultant metal-ligand complex and the alteration in
the original ligand properties with regard to a specific physiological and/or anatomical function.
We will examine metal and ligand properties in terms of Lewis acid-base and coordination
theory, hard-soft acid-base principles, complexation, and electronegativity as a way to determine
the relative affinities of a specific metal for a specific ligand. The validity for using the above
principles to relate metal and ligand properties is demonstrated using a National Institute of
Science and Technology database for metal stability constants (NIST 1998). The speciation of a
metal or metalloid at a given metal or metalloid concentration, pH, Eh, ionic strength, alkalinity,
and temperature was calculated using MINTEQA2, a geochemical equilibrium model.
Speciation data are presented for aluminum, cadmium, calcium, chromium, copper, selenium,
and zinc as a function of pH.
INTRODUCTION
Metals are released into the aquatic environment from a variety of sources and through
different pathways. Sources include direct discharge of industrial effluents and urban runoff into
surface waters, introduction of acid mine drainage and tailing washings, nonpoint source runoff
(surface and subsurface routes) of land-applied metal wastes and metal-bearing pesticides, and
rainout of metal particulates from flue gases and automobile exhausts. Metals may enter into the
aquatic environment as different species and phases including the free aqua metal ion, an ion
pair, hydrolysis/polymerized species, sorbed onto mineral, organic matter and microbial surfaces,
complexed with soluble organic matter as a soluble oxy-anion, and as a metal particulate of
variable chemical composition and particle size. The complexity of metal speciation reactions,
sorption and multi-phase transport can be seen in Figure 1. The concentration (really the
thermodynamic activity) of the toxic species is determined by the resultant of all interactions
shown in Figure 1.
Ecosystems Research Division, National Exposure Research Laboratory, U.S. Environmental Protection Agency,
Athens, Georgia, USA
2National Research Council, c/o U.S. Environmental Protection Agency, Athens, Georgia, USA
215
-------
Metal-Pore Water and
Water Column Equilibria
z-(n-y)
MJ Uptake
and Transport
Within the Fish
Reaction in
Fish Cell
K
Adsorption
ces
K
Chemical Processes
+Lj =
Solid Phases
•Minerals
-Silicates
-Oxides
—Carbonates
-Sulfides
•Humic Acid
•Microorganisms
-Bacteria
-Fungi
-Algae
Transport Pathways
Metal-Pore Water
Parucle-Surface Water— *Advection
Metal-Bottom Sediment
LEGEND
Complexation
-HT = Hydrolysis
+A~ = Precipitation
-A" = Dissolution
-me" = Oxidation
+me = Reduction
* = treat as [M(H20)nf+
species, process-wise
It = rate constants
K = S K- equilibrium constants
M = metal
, , j
L = S L ligands
A = anion, e.g.,S2", SO*", CO3"
so^po2;
LJ = N,O,P,S ligands
+M+ = same equilibrium
components as pore water
Z =• valence
m,n,x,y = stochiometric coefficients
Figure 1. Metal behavior in the aquatic environment.
216
-------
Predicting metal ecotoxicity depends on defining metal properties and speciation, ligand
properties, environmental parameters, the resultant metal-ligand complex and the alteration in
the original ligand properties with regard to a specific physiological and/or anatomical function.
The formation of the metal-ligand complex and the resultant toxicological effect of this complex
must be understood at the molecular scale and related to observations made at the macroscopic
scale.
Metal speciation is the key factor in determining the effect of metals in the aquatic
environment as different species of the same metal may vary widely in toxicity. For example,
copper hydroxy-carbonate (azurite), Cu3(OH)2CO3, is much less toxic than copper hydroxide,
Cu(OH)2, or the cupric ion, Cu(H2O) 6 2+ (Stine and Brown 1996). The difference in toxicity
may be due to a difference in water solubility, since the free aqua cupric ion is much more
soluble than either of the copper salts. Metal speciation in solution is determined by such factors
as metal concentration, anion type and concentration, pH, Eh, ionic strength, alkalinity, and
temperature. Sorption of a metal ion species to mineral, microbial and organic matter surfaces
(humic and fulvic acids) (Figure 1) may lower the activity of that species in solution (therefore,
its bioavailability) depending on the partition coefficient or apparent stability constant for that
metal species and for each surface type.
Limited success has been attained in predicting toxicity based on atomic properties or
indices. Kaiser (1980) had some success in relating toxicity to various aquatic organisms by
empirical equations involving several basic atomic properties. This relationship is
pT = ao + ai log (AN/AIP) + a2 AE0
Where: pT is the negative log of metal ion concentration (M) with certain toxicity;
AN is the atomic number of the metal;
AIP is the difference in the ion's ionization potential (eV) and the ionization
potential of the next oxidation state of that ion;
AF-o is the absolute value of the electrochemical potential between the ion and the
first reduced state;
Coefficient ao, ai, and a2 depend on the group of metals, biota, and type of toxic
effect being determined.
The terms AEo and AIP are related to outer orbital electronic properties of atoms and AN is
related fairly closely to ionic size. Inclusion of AN in the regression equation allows successful
predictions for ions with similar ionization and electrochemical potentials but different radii.
Remember that a given correlation relationship appears to have little transferability beyond the
set of data from which it was generated.
217
-------
To predict metal ecotoxicology, we must understand the complex biochemistry of metal
interactions at the specific cellular site or mode of action and relate the property of the metal to
the ligand character at the site of action. Toxicants exert an adverse metabolic efficacy through a
chain of events that begins with a specific target or cellular site of action. This target molecule
with which the metal toxicant binds or reacts may be a protein, lipid, or a nucleic acid within the
cell. Symptoms resulting from this exposure may relate directly to a specific molecular event or
may be complicated by secondary effects. The identification of the primary site of action
requires an understanding of the coupled biochemistry of the site of action and the observed
physiological or anatomical manifestation of this reaction (Merian and Haerdi 1992). The site of
action may involve ligands present on structural proteins, hormones, receptor proteins, enzymes,
proteins that make up Ion channels that regulate the flow of ions across membranes, transport
proteins, antibodies, and nucleic acids.
Although the toxic reaction from the metal occurs at the molecular level, the effect is
generally observed at the system level. Effects may be seen in the cardiovascular system,
respiratory system, nervous system, immune system, hepatic system, renal system, and in
reproduction and development. For example, lead affects the reproductive, cardiovascular,
immune, and nervous systems, cadmium affects the reproductive, cardiovascular, and renal
systems, and mercury affects the reproductive, nervous, and renal systems (Newman and
Mclntosh 1991). As we shall see later, these metals have certain similar atomic properties.
The purpose of this paper is to: (a) examine metal and ligand properties in terms of Lewis
acid-base and coordination theory, hard-soft acid-base (HSAB) principles, molecular orbital
perturbation theory, complexation and electronegativity as a way to determine the relative
affinities of a specific metal for a specific ligand; (b) evaluate these concepts using stability
constants from the National Institute of Science and Technology Critical Metal Constants
Database (NlST 1998); (c) demonstrate that the speciation of a given metal can be calculated
using a geochemical equilibrium model.
METAL AND LIGAND PROPERTIES
A theory or a set of principles are needed to interpret the affinity of a suite of metals to a
sorbent or set of ligands. This will be done in terms of acid-base chemical relationships.
Acid-Base Chemistry
The meaning of the term "generalized acid-base reactions" is derived from the definition
of acids and bases by Lewis (1923). G.N. Lewis laid a foundation that may be applied to this
problem when he proposed a definition of acids and bases. A Lewis acid is a substance that can
accept an electron pair from a base while a Lewis base is a substance that can donate an electron
pair. There have been concerted efforts over the ensuing seven decades to classify Lewis acids
and bases into categories based upon their observed behavior. The base has the ability to donate
218
-------
partially a pair of electrons to form a coordinate or dative covalent bond, while an acid has the
ability to accept at least one pair of electrons from a base, A base, therefore, is an atom, ion, or
molecule that has at least one pair of valence electrons that has not already been shared in a
covalent bond. An acid is a ion, atom, or molecule where at least one atom has a vacant orbital
in which a pair of electrons can be accommodated. A Lewis base is also a ligand and in organic
chemistry terminology a nucleophile or an electron donor, while a Lewis acid is an electrophile
or an electron acceptor. All metal atoms or ions are Lewis acids. Lewis acids normally are
coordinated to several Lewis bases simultaneously. The resulting complex may be either
electrically charged, e.g., hydration of a metal ion, [M(H2O)n]z, where M is the metal ion of
valence z and having a coordination number of n, and water is a Lewis base or an electrically
neutral molecule like Fe2O3. Anions like COO", OH", and -SH are Lewis bases and may be
important constituents of proteins. Other types of acid-base complexes are charge transfer
complexes and free atoms and radicals acting as Lewis acids forming complexes with a variety
of bases. These complexes with free radicals have yet to be isolated, but they exert a great
impact on the reactivity of the radicals. Jensen (1980) explains this in greater detail.
Complex Formation
A metal ion in aqueous solution is not present as a bare ion, but is complexed with water
(hydration) or complexed through coordination with either inorganic or organic ligands. These
complexes may be soluble (generally, organic complexes are water-soluble) or insoluble.
Precipitation occurs when the solubility product of the metal ion and the ligand is exceeded.
Complexation occurs through formation of coordinate covalent bonds, electrostatic bonds, or a
combination of both.
Metal chelates are metal complexes where there are two or more ligands capable of
forming a five- or six-member ring containing the metal ion. Chelate formation enhances the
stability of the complex system compared to the stability of a similar system that contains no
rings, or a lesser number of rings. Chelate formation may be very important in binding metals to
ligands and may help explain the anomalies in the data, particularly with variable pH. Chelate
formation may increase the stability constant of a metal by orders of magnitude compared to that
found for that same metal binding to either one of the ligands separately.
Metal cations act as electron acceptors, i.e.., Lewis acids, and ligands with anionic or polar
molecules act as electron donors, i.e., Lewis bases. Therefore, formation of metal complexes is a
branch of acid-base chemistry.
Complexes with charge due to direct cation-to-metal bonds are "organometallic species",
while "metallorganic" refers to metal ions bound in coordinate bonds that employ a variety of
organic ligands (e.g., O, N, S, P, As, etc.) that utilize non-carbon donors. Unsaturated carbon-
carbon bonds can form stable TC complexes with certain metals.
219
-------
Approaches for Predicting Metal-Ligand Reactions
Three approaches have been developed that may be used to predict chemical reactions a
priori, and these include: (1) hard-soft acid-base principles (HSAB) (Pearson 1963, 1967a,
1967b); (2) perturbed molecular orbital theory (Klopman 1968, 1974); and (3) a dual softness
parameter approach (Misono et al. 1967, Misono and Saito 1970). Such approaches can be used
to characterize the chemical nature of the ligands present in proteins, DMA, etc. within the
biological system. Perturbed molecular orbital theory provides a theoretical basis for the HSAB
concept set forth by Pearson (1963). The development of a general molecular orbital
perturbation theory (Klopman 1968, 1974) has permitted quantum mechanical calculations to
help elucidate roles played by specific orbitals of molecules and atoms engaged in a chemical
reaction. This generalized perturbation molecular orbital theory addresses the problem of what
happens to the energy when a metal and a ligand interact, and also addresses the process of
electron transfer during bond formation. Each will be discussed separately. Our focus will be on
Pearson's Hard-Soft Acid-Base Principles. The other two approaches generate the same results
by considering electronegativity or size to predict softness or the extent of covalent formation.
Hard-Soft Acid-Base Principles: The principle of HSAB is based upon two assumptions:
(1) that if a bond exists between two atoms, one will be and acid and the other play a role as a
base; and (2) that electrons hold the bonded atoms together. A typical acid-base reaction can be
written
A + :B = A:B
Solvation of the acid and base results in the metal cation bonded to the basic end of a water
molecule - A(OH2), and the anionic ligand bound to the acidic end - (HOH)m :B.
The strengths associated with the acid or base at the ends of a bond arise from two factors,
intrinsic strength (S), and their softness parameter (5). The strength of the A:B bond, which
roughly is proportional to its formation constant K, is defined as:
LogK= SASB + 6A5B
Softness arises from the electron polarizability or mobility of a species. If the electrons are
easily mobilized, the species is soft; conversely if the electrons are firmly held the species is
hard. Another perspective on a soft base is one that is highly polarizable, easily oxidized, or both
valence electrons are loosely held. This translates into a low density of charge on the base. It is
the relative rather than absolute strengths that are important in complex formation in solution and
involves replacing one ligand (usually water) by another ligand. Furthermore, the bonds formed
by species at the extremities of softness would be purely covalent (Sa = Sb = 0), and of hardness
would be purely ionic (5A = 83 = 0).
Metal reactions with ligands, whether with water molecules or ligands present in the
Structure of cells in biological tissues, are interactions of generalized acids and bases. What is
220
-------
needed are rules to predict the stability of the metal-ligand reaction. Pearson (1963,1967a,
1967b, 1968a, 1968b, 1973) suggested the "Principle of Hard and Soft Acids and Bases" which
has a direct application to complex formation of metals with ligands, whether they are ligands
present in biological tissue, or competing ligands present in humic acids or minerals. The HSAB
Principle can be stated: hard acids prefer to coordinate, bind, or react with hard bases, and soft
acids prefer to coordinate, bind, or react with soft bases, and there is no cross-over. The kinetic
analogue of the HSAB Principle is that hard acids react faster with hard bases and that soft acids
react faster with soft bases. The preferred stereochemical arrangement of bonds around a
chelated metal group may affect the ability of the base to function as a multi-donor ligand (which
results in a net entropy increase). A complex formation may, therefore, be favored despite a
poorer matching of donor acceptor atoms so as to lead to a violation of the HSAB rules.
A corollary to the HSAB Principle stated above is that given by Arhland et al, (1958)
namely that hard acids coordinate best to the lightest atom of a family of elements. Soft acids
coordinate best to one of the heaviest atoms of the same family. Hard acids bind best to the least
polarizable (hardest) atom of the family while soft acids bind to a more polarizable atom of a
family of elements (Table 1). However, soft acids do not form their most stable complexes with
the most polarizable atoms, the reason being that some very soft atoms are very weak bases
towards all acids.
A hard Lewis acid is one that has small size, high charge and electropositivity, low
polarizability, few and not easily excitable outer electrons on donor atoms, and forms
ionic/electrostatic bonds; the properties of a soft-acid are the converse. Strictly speaking it is the
acceptor atom of the acid that has these properties. A soft Lewis base is one that is large in size,
whose electrons are easily distorted, polarized, easily accessible or removed, and forms covalent
7t-type bonds. A hard base is the converse, i.e., one that holds on to its electrons more tightly.
Lewis acid-base reactions present problems in understanding and predicting the nature of
chemical reactivity - the nature of the products formed and the kinetics with which a reaction
occurs. Both the products formed and the kinetics are, however, closely related and are a
function of the internal nature/structure of the reagents - the Lewis acid and Lewis base, as well
as the external (environmental) conditions under which the reaction occurs.
A number of representative Lewis acids which are classified as being either hard, soft, or
borderline are listed in Table 1. Note that many of the divalent transition metals are classified as
borderline between hard and soft acids. Whether they act as hard or soft acids will be influenced
by the solution environment. Water is a very hard solvent both as an acid and as a base. It
strongly solvates small anions and small, highly charged cations. In most cases hard acids and
hard bases are held together by ionic or polar bonds while most soft acid-bases are held together
by largely covalent bonds. Also important in HSAB reactions are n bonding centers. Most hard
acids are 7t bond acceptors, i.e., they have an empty outer orbital that can accept re electrons from
the base. Most soft acids are n bond donors. They have filled orbitals that can donate n
electrons. Hard bases have filled outer orbitals that can donate via a n bond to a hard acid. Soft
acids have certain empty outer orbitals that can accept n electrons from a soft base.
221
-------
Table 1. Metal classification schemes and trends in ligand preferences.
a. Pearson Hard and Soft Acids
Hard acids
2+
if, Na+, K+, Be2+, Mg2+, Ca
2*, A13+, La3*, Cr3+, Fe3+, Co3+
Inert gas configuration
Borderline
Mn2+, Fe2+, Co2+, Ni2+,
Cu2+, Zn2+, Pb2% Bi3*
1 to 9 outer shell electrons;
not spherically symmetric
Soft acids
Cu+, Ag+, Au+, TI+, Hg+,
Hg2+,Pd2+,Pt2+,Tl3+
10 to 12 outer shell electrons;
highly polarizable
b. Klopman Orbital Electronegativity of Cations and Softness Character
Low softness Medium softness
AI3+ As > Sb
O»S>Se>Te
F »C1 > Br > I
Soft acid-type cations
Strongest complexes with heavier ligands in 2nd, 3rd,
and 4 row; mainly strong covalent bonding;
enthalpy changes dominate in energy of reaction
N «P > As > Sb
O «S - Se =Te
F «C1 < Br < I
e. Irving-Wiliams Order
For divalent transition-metal cations, the following well-established sequence of complex stability is based on
crystal-field stabilization energies, which depend on the number of d orbital electrons. This order is independent of
ligand type.
Mn < Fe < Co < Ni < Cu > Zn
222
-------
Stability and Conditional Stability Constants
A methodology has been developed for estimating the stability constant for a given
metal with a ligand of known chemical functionality (Zhang and Bailey, Submitted). We will
use the database from this methodology to demonstrate the validity of the above approaches to
correlate metal and ligand properties to predict relative binding affinities. The values of stability
constants, their standard deviations, and a number of data entries of metal complexes with
selected functional groups obtained from the National Institute of Science and Technology
(NIST) database are reported in Table 2. Based on this table, we make the following
observations:
1. For divalent metals, the conditional stability constants for each organic ligand follow
the extended Irving-Williams series (Williams 1971), i.e., they increase from Ca, Mg, Mn, Fe,
Co, Ni to Cu and decrease with Zn.
2. The formation of a five- to six-member ring increases the stability of the complex.
For this reason, we only considered dicarboxylic acids and diamines separated by either one or
two carbons, and aromatic carboxylic and phenolic groups in the adjacent positions.
3. The selectivity of metal-ligand binding obeys the HSAB principles. Trivalent metals,
such as Fe and Cr, are hard Lewis acids and are preferred by carboxylic sites, whereas heavy
divalent metals, such as Cd, Hg, and Pb, are soft Lewis acids and are preferred by soft sulfur-
containing ligands. Most of the transitional metals are either borderline or soft Lewis acids,
therefore, the stability and conditional stability constants for N- and S-bearing ligands are
generally higher than for oxygen-bearing ligands. This would be true for nucleic acids and for
sulfur-bearing proteins - cysteine and methionine.
4. The conditional stability constants for each functional group increase with pH (data
not shown). However, the conditional stability constants of the carboxylic acid group reached a
plateau at pH 5 to pH 6. Therefore, carboxylic acid sites are important for metal binding at
pH < 6.0. At higher pH levels, diketone, amino acid, diamine, sulfhydryl, and catechol groups
become more important. There is also independent experimental evidence showing that metals
are bound to nitrogen ligands. It is known that Cu forms many complexes with nitrogen ligands
(Nicholls 1974), therefore, it is expected that Cu is bound by nitrogen-bearing functional groups
in proteins. Indeed, Keefer et al. (1984) extracted and fractionated organic matter from sludge-
amended soil and found that Cu was most often bound by hydrophilic bases (amino acids).
5. For most of the metals, the conditional stability constants for an aliphatic dicarboxylic
acid is greater than for an aromatic dicarboxylic acid (e.g., phthalic acid). Between phthalic acid
and salicylic acid, the former is preferred at lower pH levels, whereas the latter is favored at
higher pH levels. However, the catechol site forms the most stable complex at higher pH levels
among the three groups.
6. Single amine groups and single carboxylic acid groups are not important for metal
binding. Similarly, alcoholic OH groups and S bound to two carbons (RrS-R2) do not contribute
significantly to metal binding (data not shown).
223
-------
Table 2. Values of stability constants (log K), their standard deviation (s.d.), and number
of data entries (N), compiled from NIST database.
Amtno Acid
logK
s.d.
N
DIamtnc
log K
s.d.
N
Dtcarboxvllc Acid
logK
sxl.
N
L/HL
9.25
0.45
35
9.75
ass
43
5,32
0.95
34
Tctracarboxvlic Acid
log K 5.60
s.d,
N
Phlhalic Acid
s.d.
N
Salicylic Aclti
logK
s.d.
N
Catcchol
logK
s.d.
N
Dlkclone
logK
s,d
N
R-SH
logK
' s.d.
N
0.60
4
4.66
0.48
9
11.85
2.09
18
12.77
0.84
14
8.61
0.28
2
9.94
0.67
10
HL/H2L
2.36
0.91
34
6.81
1.05
43
2.65
0.63
33
4.42
0.55
4
2.53
0.35
8
2.18
0.82
18
8.73
1.16
14
-
-
6.86
1.97
10
*»
1.76
0.45
8
0.37
1
2.00
0.61
13
4.23
0.71
4
-
-
5.33
0.33
2
6.47
0.75
6
3,35
-
1
3.12
1.06
3
Cai+
1.51
0.44
11
0.87
1.07
2
1.88
0.64
17
5.40
0.30
4
1.60
—
1
4.08
0.86
2
5.18
0.47
6
2.32
-
1
3.20
1.47
3
Cr" Mnz+
4.47 2.72
0.24 0.54
3 18
5.48 . 2.54
0.34
"l "" 5
3.89 1.89
0.04 . 0.64
2 10
5.32
- : 0,56
' *
2.48 2.68
0.53'
1 4
8.15 4.86
0.77 0.96
2 ' " 4
8.06
0.67
9
5.96 3.9
-
1 1
1
6.15
2.75
. -- 3
Fe2+
3.72
0,68
17
4.26
1
2.10
0.50
4
4.41
--
1
_
--
--
6.23
0.39
2
8.26
-
1
5.Q7
-
1
9,16
3.62
2
Co2+
4.40
0.66
28
5.16
0.63
nl"1
14
2.40
0.48
19
4.55
--
1
2.74
0.49
4
5.77
1.37
4
9.02
0.76
11
4.74
0.48
2
8.84
2.70
8
224
-------
Table 2. (Continued)
Amino Acid
logK
s.d.
N
Diamine
logK
s.d.
N
NP
5.50
0.75
29
6.37
1.15
34
Dicarboxylic Acid
logK 2.50
s.d.
N
0.50
22
Tetracarboxylic Acid
logK 6.21
s.d.
N
Phthalic Acid
logK
s.d.
N
Salicylic Acid
logK
s.d.
N
Caiechol
logK
s.d.
N
Diketone
logK
s.d.
N
R-SH
logK
s.d.
N
1.37
4
2.03
0.21
9
6.32
1.23
5
9.35
0.74
11
4.98
1.02
2
10.68
2.75
7
Cuz+
8.17
0.62
32
9.49
0.98
43
4.29
0.99
31
7.06
1.23
4
2.98
0.39
9
9.04
1.57
12
13.88
0.96
13
7.85
0.21
2
-
-
—
Cr* Fe*"
8.84 9.44
2.39
1 11
..
-- .
..
8.26 7.18
0.42
1 10
11.38
0.66
3
-
..
-
9.41 15.09
2.40
1 13
19.45
2.25
— 9
10.6
_
1
-
„
—
Ag+
3.62
0.65
11
5.31
0.59 ,
5
1.86
0.76
4
..
-
-
-
-
-
.
--
--
-
-
-
-
-
-
12.4
1
Znz+
4.68
0.61
33
5.27
0.65
19
2.58
0.43
25
6.54
0.82
4
2.86
0.59
3
6.48
0.46
2
9.94
0.88
11
4.35
0.49
2
9.71
3.09
8
Cd"
3.85
0.48
23
5.52
1.18
14
2.41
0.28
18
5.10
0.39
4-
2.37
0.29
2
5.13
0.53
2
8.44
1.08
9
: 3.48
0.04
2
10.70
3.04
7
Hg»
7.08
1.93
5
13.20
1.56
2
-
-
14.12
0.28
3
-
—
-
--
-
19.90
—
1
11.93
—
1
16.45
2.21
3
PbH
4.77
0.56
16
5.52
0.36
6
3.07
0.33
8
7.39
0.35
3
2.78
—
1
5.60
--
1
13.55
0.43
4
4.57
-
1
12.52
2.93
5
AIW
3.72
1.20
• 6
-
—
-
, 5.25
1.16
6
8.41
0.77
3
2.94
—
1
12.89
0.59
3
16.23
1.20
10
-
-
-
-
-
„
225
-------
MODELING METAL SPECIATION
To realize the goal of predicting metal ecotoxicity we must first predict metal speciation in
solution and at the site of action. Aquatic and biochemical systems comprise large numbers of
components and Interactions. Therefore, equilibrium computations of these systems can involve
a large array of equations that incorporate the principles of mass action and mass balance to
model such reactions as hydrolysis, complexation, dissolution/precipitation, oxidation/reduction,
ton exchange, and partition/adsorption. A number of computational codes have been formulated
to solve these arrays of equations using a thermodynamic data base and an "equilibrium constant
approach" that uses the mass action principle that relates the activities of free metal, free ligand
and the metal ligand complex. The mass action relationships are linked with mass balance
equations, resulting in a set of linear equations. Generally, the assumption is made that the
solution is in equilibrium with the thermodynamically-predicted solid phases and that only the
most stable phase or phases can occur, and that the solution is always in equilibrium with them.
The equilibrium speciation of each component is obtained by solving the set of linear
equations. In the solution of chemical equilibrium models, it is assumed that equilibrium exists
for all reactions in the systems of interest. This assumption is appropriate, because a majority of
these interactions are known to occur rapidly. Because of the rapidity of these reactions,
predictions of equilibrium for aqueous and adsorbed phases are generally reliable, except in the
case where reactions are controlled by diffusion. Such reactions really are transport-controlled
rather than chemically controlled. Caution must be exercised when predictions involving solid-
phase precipitation/dissolution and redox transformations are involved, because these depend on
the knowledge of chemical kinetics.
Several chemical equilibrium models have been developed including WATEQ (Truesdell
and Jones 1974), GEOCHEM (Mattigod and Sposito 1979, Sposito and Mattigod 1980),
MINEQL (Westall et al, 1976), MINTEQ (Felmy et al 1984), MINTEQA2 (Allison et al.
1991), GEOCHEM-PC (Parker, et al. 1995), GMIN (Felmy 1995), SOILCHEM (Sposito and
Coves 1995), ALCHEM (Schecher and Driscoll 1995), C-SALT (Smith et al. 1995), and CHESS
(Santore and Driscoll 1995). For a more in-depth discussion of the subject of chemical
equilibrium models the reader is referred to a publication by Loeppert et al. (1995).
What metal cations and or anions are important in fish behavioral toxicology? Henry and
Atchison (1991) summarized the literature up through 1990 and found that the following
metals/metalloids are important: (a) aluminum, (b) arsenic, (c) cadmium, (d) chromium, (e)
copper, (f) selenium, and (g) zinc. MINTEQA2 is used to assess the effect of pH on speciation
of the five metals and the metalloid, selenium; arsenic speciation can not be simulated.
The input parameters for a simulation run included metal/metalloid type and
concentration: Al (3.4 umol L"1); Cd, Cu, Cr, Se, and Zn (0.47 umol I/1); pH (4.0 to 8.0);
partial pressure of CO2 (10~3'5 alms); ionic strength (1.0 x 10~3 M Ca (NO3)2); and temperature
(15 C). Precipitation of solids was allowed and activities were calculated by the Davies
equation.
226
-------
The MINTEQA2 speciation simulation results are reported in Table 3. The pH of the
system certainly has a great effect on speciation of each element and the relative distribution of
each species. We will look at the speciation of each element separately below.
Aluminum. For aluminum the presence of the [A1(H2O)6] species decreased from 96%
at pH 4.0 to 12% at pH 5.5, while [Al (OH)(H2O)]2 increased from 4% to 55% over the same
pH range. Diaspore started to precipitate out at pH 5.0 and represented 86% of the speciation at
pH 7.0. Highly charged aluminum polymers can form and also play a role in aluminum
speciation. These polymers may have problems passing through membranes.
Cadmium. In the presence of chloride the major species is CdQ2 (aq) and the distribution
is 56% regardless of pH. Independent simulations with nitrate as the balancing anion indicated
that [Cd(H2O)6]2+ was the overriding species. The other two species are [CdCl(H2O)5]+ and
[CdCl3(H20)5]-.
Copper. At pH below 6.5 [Cu(H2O)6)]2+ is the predominant or equally dominant species
with the ion pair [CdCl(H2O)5]+. This changes at pH 7.0 and above; the predominant species is
Cu(OH)2 (aq) with [CuCl (H2O)s]" being a minor species.
2-
Chromium. Below pH 7.0 [Cr2O7] is the predominant species and [CrO4] " is an
important one, while the order is reversed at pH 7.5 and above.
Calcium. Below pH 8.0 the species is the aqua ion [Ca(H2O)8]2+, while above pH 8.0
calcium occurs as calcite,
Selenium. Up to pH 6.0 in the presence of zinc, selenium occurs as ZnSeC>4, while above
pH 6.0 the predominant species is
i2-
Zinc. In the presence of the selenate anion ZnSeO4 is the major species increasing up to
more than 60% at pH 8.0. In the absence of selenate, the aquo cation [Zn(H2O)6]2+ is the
predominant species up to pH 7.0.
ACKNOWLEDGEMENTS
This research was conducted while Z. Z. Zhang was a Research Associate of the National
Research Council. This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies, and is approved for presentation
and publication. Mention of trade names or commercial products does not constitute
endorsement or recommendation by the U. S. Environmental Protection Agency.
227
-------
Table 3. Metal speciation in aqueous solution
pH4.0 pH4.5
Metal Species
Al Al*
A1(OH)2*
Cd CdCl2
-------
Table3. (Continued)
pH 6.0
Species
A1(OH)3M
A1(OH)2
A1(OH)2*
CdCl2(aq)
CdCV
CdCL,2'
CdSeO4
Cu2*
cucr
CuCI2(aq)
Cr2072'
HCr04-
Cr042-
Ca2+
ZnSeO4
Se042+
ZnSeO4
ZnCf
Zn2*
ZnCl2(aq)
ZnCl3"
ZnCl42-
Percent
50
46
4
56
21
21
1
43
43
14
83
13
5
100
70
30
44
17
16
12
8
3
pH6.S
Species
Al(OH)3(aq)
A1(OH)2*
A1(OH)4'
CdCl^
CdCl3-
CdCf
CdSeO4
CuCf
Cu2t
CuCl2(aq)
Cu(OH)2(aq)
Cr2072-
CrO42'
HCr04-
Ca2+
Se042'
ZnSeO4
ZnSeO4
ZnCf
Zna+
ZnCl2(aq)
ZnCV
ZnCl/'
Percent
76
22
2
56
22
20
2
41
39
14
6
75
14
12
100
73
27
54
14
12
10
8
3
pH7.0
Species
AlCOH)^
A1(OH)2+
A1(OH)4-
CdCl2(aq)
CdCl3-
CdCf
CdSeO4
Cu(OH)2(a
CdCl3-
cdcr
CdSeO4
Cu(OH)2(aq)
CuCf
Cu2+
CuCl2(aq)
CuOH"
Cr2072'
CrO42'
HCrO/
Ca2"
SeO42'
ZnSeO4
ZnCl*
Zn2+
ZnCl2(W!)
ZnCV
ZnCl42'
ZnOHCl
Percent
80
18
2.
55
22
20
2
84
6
6
2
2
74
20
6
100
99
60
10
9
7
6
5
2
pHS.O
Species
AlCOHte,)
A1(OH)4-
CdCl2(aq)
CdCV
CdCl+
CdSeO4
Cu(OH)2(aq)
Cr042'
Cr2072'
HCrO/
CaCOs
Ca2+
Se042'
ZnSeO4
ZnOHClCaq)
ZnCl+
Zn2+
ZnClv
ZnCV
ZnCl42-
Percent
59
41
54
22
19
5
97
94
4
3
95
5
100
63
12
7
6
5
4
2
229
-------
REFERENCES
Allison, J.D., D.S. Brown, and K.J. Novo-Gradac. 1991. MINTEQA2/PRODEFA2, a
geochemical assessment model for environmental systems: Version 3.0 User's Manual.
U.S. Environmental Protection Agency, Athens, Georgia. EPA/600/3-91/021.
Arhland, S., J. Chart, and N.R. Davies. 1958. The relative affinities of ligand atoms for acceptor
molecules and ions. Quarterly Review of the Chemical Society XII(3):265-276.
Felmy, A.R., D.C. Girvin, and E.A. Jenne. 1984. MINTEQ - A Computer Program for
Calculating Aqueous Goechemical Equilibria. U.S. Environmental Protection Agency,
Athens, Georgia. EPA/600/3-84-032.
Felmy, A.R. 1995. GMIN, a computerized chemical equilibrium program using a strained
minimization of the Gibbs free energy: Summary report. In: R.H. Loeppert, A.P. Schwab,
and S. Goldberg, (Eds). Chemical Equilibrium and Reaction Models. Soil Science Society
of America, Madison, Wisconson. Special Publication Number 42. pp 377-407.
Henry, M.G., and G.J. Atchison. 1991. Metal effects on fish behavior - Advances in
determining the ecological significance of responses. In: M.C. Newman, and A.W.
Mclntosh, (Eds.). Metal Ecotoxicology - Concepts and Applications. Lewis Publishers,
Chelsea, Michigan, pp 131-169.
Jensen, W. B. 1980. The Lewis Acid-Base Concepts: An Overview. John Wiley & Sons,
New York. 351 pp.
Kaiser, K.L.E. 1980. Correlation and prediction of metal toxicity to aquatic biota. Canadian
Journal of Fisheries and Aquatic Sciences 37:211-218.
Keefer, R.F., E.E. Codling and R.N. Singh. 1984. Fractionation of metal-organic components
extracted from a sludge-amended soil. Journal of the Soil Science Society of America
48:1054-1059.
Klopman, G. 1968. Chemical reactivity and the concept of charge-and-frontier-controlled
reactions. Journal of the American Chemical Society 90:223-234.
Klopman, G. 1974. Reactivity and Reaction Paths. Wiley Interscience, New York, pp 55-72.
Lewis, G.N. 1923. Valence and the structure of atoms and molecules. In: The Chemical
Company Catalog, New York pp 141-142.
Loeppert, R.H., A.P. Schwab, and S. Goldberg (Eds). 1995. Chemical Equilibrium and
Reaction Models. Soil Science Society of America, Madison, Wisconsin. Special
Publication Number 42. 422pp.
230
-------
Mattigod, S.V., and G. Sposito. 1979. Chemical modeling of trace metal equilibria in
contaminated soil solution using the computer, GEOCHEM. In: Jenne, E. A. (Ed)
Chemical Modeling in Aqueous Systems. American Chemical Society Symposium Series
92, Washington, D.C. pp 837-855.
Merian, E., and W. Haerdi (Eds). 1992. Metal Compounds in Environment and Life.
Middelsex, United Kingdom. Science Technology Letters. 514pp.
Misono, M., and Y. Saito. 1970. Evaluation of softness from the stability constants of metal-ion
complexes. Bulletin of the Chemical Society of Japan 43:3680-3684.
Misono, M., E. Ochiai, Y. Saito, and Y. Toneda. 1967. A new dual parameter scale for the
strength of Lewis acids and bases with evaluation of their softness. Journal of Inorganic
and Nuclear Chemistry. 29:2685-2691.
Newman, M.C., and A.L. Mclntosh (Eds). 1991. Metal Ecotoxicology: Concepts and
Applications. Lewis Publishers, Chelsea, Michigan. 397pp.
Nicholls, D. 1974. Complexes and First Row Transition Elements. Macmillan, London, pp
215.
NIST (National Institute of Science and Technology). 1998. Standard Reference Database No.
46 Critically Selected Stability Constants of Metal Complexes, Version 5.0. Gaithersburg,
Maryland.
Parker, D.R., W.A. Norvell, and R.L.Chaney. 1995. GEOCHEM-PC - A chemical speciation
program for IBM and compatible personal computers. In: R.H. Loeppert, A.P. Schwab, and
S. Goldberg, (Eds.). Chemical Equilibrium and Reaction Models. Soil Science Society of
America, Madison, Wisconsin. Special Publication Number 42. pp 253-270.
Pearson, R.G. 1963. Hard and soft acids and bases. Journal of the American Chemical Society
85:3533-3539.
Pearson, R.G. 1967a. Hard and soft acids and bases. Chemistry of Britain. 3:103-107.
Pearson, R.G. 1967b. Application of the principle of hard and soft acids and bases to organic
chemistry. Journal of the American Chemical Society 89:1827-1836.
Pearson, R.G. 1968a. Hard and soft acids and bases. Part I, fundamental principles. Journal of
Chemical Education 45:581-587.
Pearson, R.G. 1968b. Hard and soft acids and bases. Part II, underlying theories. Journal of
Chemical Education 45:643-648.
Pearson, R.G. (Ed.) 1973. Hard and Soft Acids and Bases. Dowden, Hutchinson, and Ross.
Stroudsburg, Pennsylvania.
231
-------
Santore, R.C., and C.T. Driscoll. 1995. The CHESS model for calculating chemical equilibria
in soils and solutions. In: R.H. Loeppert, A.P. Schwab, and S. Goldberg, (Eds.), Chemical
Equilibrium and Reaction Models. Soil Science Society of America, Madison, Wisconsin.
Special Publication Number 42. pp 357-376.
Schecher, W.D. and C.T. Driscoll. 1995. ALCHEM: A chemical equilibrium model to assess
the acid-base chemistry and speciation of aluminum in dilute solutions. In: R.H. Loeppert,
A.P. Schwab, and S. Goldberg, (Eds). Chemical Equilibrium and Reaction Models. Soil
Science Society of America, Madison, Wisconsin. Special Publication Number 42. pp 325-
356.
Smith, G.R, K.K. Tanji, R.G. Burau, and JJ. Jurinak. 1995. C-SALT - A chemical equilibrium
model for multicomponent solutions. In: R.H. Loeppert, A.P. Schwab, and S. Goldberg,
(Eds). Chemical Equilibrium and Reaction Models. Soil Science Society of America,
Madison, Wisconsin. Special Publication Number 42. pp 289-324.
Sposito, G., and J. Coves. 1995. SOILCHEM on the Macintosh. In: R.H. Loeppert, A.P.
Schwab, and S. Goldberg, (Eds). Chemical Equilibrium and Reaction Models. Soil
Science Society of America, Madison, Wisconsin. Special Publication Number 42. pp
271-288.
Sposito, G., and S.V. Mattigod. 1980. GEOCHEM. A computer program for the calculation of
chemical equilibria in soil solutions and other natural water systems. Kearney Foundation
for Soil Science. University of California - Riverside, California.
Stine, K.E., and T'.M. Brown. 1996. Principles of Toxicity. CRC Press, Boca Raton, USA.
259pp.
Truesdell, A.H., and B.E. Jones. 1974. WATEQ, a computer program for calculating chemical
equilibria in natural water. U.S. Geological Survey Journal of Research. 2:321-326.
Westell, J.C., J.L. Zachary, and F.M.M. Morel. 1976. MINEQL. Technical Note 18.
Department of Civil Engineering, Massachusetts Institute of Technology, Cambridge,
Massachusetts.
, ). t, ' • • • • , i f , ~ \ .
Williams, D.R. 1971. The Metals of Life: The Solution Chemistry of Metal Ions in Biological
Systems. Van Nostrand Reinhold Co., New York
Zhang, Z. Z., and G. W. Bailey. (Submitted). Modeling of metal binding on humic substances
using the NIST database. (Submitted to Environmental Science & Technology).
232
-------
DEVELOPMENT OF METHODOLOGIES TO EVALUATE
THE HEALTH OF RIPARIAN AND WETLAND AREAS
Paul L. Hansen1, William H. Thompson1, Robert C. Ehrhart1,
Dan K. Hinckley2, William Haglan3, and Karen Rice4
ABSTRACT
Since 1988, we have been developing various assessments to address a wide range of
riparian and wetland questions. Throughout this process, we have worked with numerous
personnel from the U.S. Department of Interior (USDI) Bureau of Land Management (BLM)
and the Fish and Wildlife Service (F&WS). Out of this collaborative effort, the following
assessments for riparian and wetland areas have been developed: (1) lotic inventory; (2) lotic
health evaluation (derived from the lotic inventory); (3) lotic health assessment (stand-alone);
(4) river health assessment (stand-alone); (5) lentic inventory; (6) lentic health evaluation
(derived from the lentic inventory); and (7) lentic health assessment (stand-alone). Each of the
assessments also includes a discussion on the codes or instructions used with each form.
BACKGROUND
In 1986, work began at The University of Montana on developing a statewide riparian
and wetland vegetation-based ecological site classification for Montana. This resulted in the
document Classification and Management of Montana's Riparian and Wetland Sites (Hansen
et al. 1995). While developing this statewide classification, The University of Montana was
asked by the U.S. Department of Interior (USDI) Bureau of Land Management (BLM) in the
spring of 1988 to develop and conduct a large-scale inventory and assessment for the Upper
Missouri National Wild and Scenic River corridor in central Montana. The major goal of the
work was to develop a sampling protocol that would allow the BLM to address some basic
questions about the location, extent, and health of the various plant communities along 253 km
(157 mi.) of the Missouri River and its tributaries. In addition, some basic soil and physical site
information was collected.
Since 1988, The University of Montana has continuously worked with the BLM Montana
State Office, and later the BLM Upper Snake River District in eastern Idaho, to develop various
assessment protocols to address a wide range of management questions. In addition, personnel at
the USDI Fish and Wildlife Service at Charles M. Russell National Wildlife Refuge in central
'Riparian and Wetland Research Program, School of Forestry, The University of Montana, Missoula, Montana,
USA. 2Montana State Office, USDI Bureau of Land Management, Billings, Montana, USA.
3Charles M. Russell National Wildlife Refuge, USDI Fish Wildlife Service, Lewistown, Montana, USA.
''Upper Snake River District, USDI Bureau of Land Management, Idaho Falls, Idaho, USA.
233
-------
Montana provided invaluable field input and critical review. Out of this collaborative effort,
we have developed the following assessments for riparian and wetland areas: (1) lotic inventory;
(2) lotic health evaluation (derived from the lotic inventory); (3) lotic health assessment (stand-
alone); (4) river health assessment (stand-alone); (5) lentic inventory; (6) lentic health evaluation
(derived from the lentic inventory); and (7) lentic health assessment (stand-alone). Each of the
forms also includes a discussion pn the codes or instructions used with each form. In addition,
we utilize the Pfankuch channel assessment (Pfankuch 1975), the BLM's lotic proper
functioning condition (PFC) checklist (USDIBLM 1998), and the BLM's lentic proper
functioning condition (PFC) checklist (Prichard et al. 1994).
INTRODUCTION
Public and private land managers in the United States are being asked to improve or
maintain riparian and wetland habitat and water quality. Those who live and work on the land
can usually tell which sites support diverse, vigorous plant and animal communities, which sites
have lost their capacity to retain spring season waters long into the summer dry season, and
which sites are biologically depauperate. While it may be easy for an astute observer to see that
a site has been degraded by human use, it is often difficult to quantify such changes. Presented
here are methods for rapidly assessing riparian and wetland health. These methods provide an
indexed site rating useful for setting management priorities and stratifying segments for remedial
or more rigorous analytical attention. These methods are intended to serve as a first
approximation, or "coarse filter," by which to identify segments in need of closer attention so
that the manager can more efficiently concentrate effort.
,•"' ,',",'' i ,, ' • , ' • , i •" : n • , • • ,i! H
Three questions that are generally asked about a riparian or wetland site are: (1) what
is the potential of the site (e.g., climax or potential natural community)?; (2) what plant
cpmmunities currently occupy the site?; and (3) what is the overall health (condition) of the site?
For riparian and wetland sites in Montana, the first two questions can be answered using the
Classification and Management of Montana's Riparian and Wetland Sites (Hansen et al. 1995).
Other regions of North America may have similar publications to aid in addressing these two
questions. The assessments outlined in this paper address the third question: what is the site's
overall health (condition)? These methods provide an indexed site rating useful for setting
management priorities and stratifying riparian and wetland sites for remedial or more rigorous
analytical attention.
We use the term "health" to mean the ability of a riparian or wetland area to perform
certain functions such as adequate vegetation, landform, or woody debris present to dissipate
stream and wave energy associated with high water levels, thereby reducing erosion and
improving water quality; filter sediment, capture streambed load, and aid floodplain
development; improve flood-water retention and ground-water recharge; develop root masses
that stabilize streambanks and shorelines against stream cutting and wave action; develop diverse
ponding and channel characteristics to provide the habitat and the water depth, duration, and
temperature necessary for fish production, waterfowl breeding, and others uses; and support
greater biodiversity.
234
-------
In some cases management steps may have already been taken to remedy a functionally
degraded riparian or wetland areas. In many such cases, however, it is unclear how the results of
those changes can be assessed. How, for example, can we stratify sites on a large management
unit among those functioning well, those functioning with slight impairment, those having lost
much of their functional capacity, and those so severely impaired that restoration would be too
costly and difficult?
Flowing Water (Lotic) Wetlands vs. Still Water (Lentic) Wetlands
Cowardin et al. (1979) point out that no single, correct definition for wetlands exists,
primarily due to the nearly unlimited variation in hydrology, soil, and vegetative types.
Wetlands are lands transitional between aquatic (water) and terrestrial (upland) ecosystems.
Windell et al. (1986) state that "wetlands are part of a continuous landscape that grades from
wet to dry. In many cases, it is not easy to determine precisely where they begin and where
they end."
In the semi-arid and arid interior western North America, a useful distinction has been
made between wetland types based on association with different aquatic ecosystems. Several
authors have used the terms "lotic" and "lentic" to separate wetlands associated with flowing
water (lotic) from those associated with still water (lentic). The following definitions represent
a synthesis and refinement of terminology from Shaw and Fredine (1956), Stewart and Kantrud
(1972), Boldt et al. (1978), Cowardin et al. (1979), American Fisheries Society (1980), Johnson
and Carothers (1980), Cooperrider et al. (1986), Windell et al (1986), Environmental
Laboratory (1987), Kovalchik (1987), Federal Interagency Committee for Wetland Delineation
(1989), Mitsch and Gosselink (1993), and Kent (1994).
Lotic wetlands are associated with rivers, streams, and drainageways. Such wetlands,
also referred to as riparian wetlands, contain a defined channel and floodplain. The channel is
an open conduit that periodically or continuously carries flowing water and dissolved and
suspended material. Beaver ponds, seeps, springs, and wet meadows on the floodplain of,
or associated with, a river or stream are part of the lotic wetland.
Lentic wetlands are associated with still water systems. These wetlands occur in
basins and lack a defined channel and floodplain. Included are permanent (i.e., perennial) or
intermittent bodies of water such as lakes, reservoirs, potholes, marshes, ponds, and stockponds.
Other examples include fens, bogs, wet meadows, and seeps not associated with a defined
channel.
Functional vs. Jurisdictional Wetland Criteria
Defining wetlands has become more difficult as greater economic stakes have increased
the potential for conflict between politics and science. A universally accepted wetland definition
satisfactory to all users has not yet been developed because the definition depends on the
objectives and the field of interest. However, scientists generally agree that wetlands are
235
-------
characterized by one or more of the following features: (1) wetland hydrology, the driving force
creating all wetlands; (2) hydric soils, an indicator of the absence of oxygen; and (3) hydrophytic
vegetation, an indicator reflecting wetland site conditions. The problem is how to define and
obtain consensus on thresholds for these three criteria and various combinations of the three
criteria.
In the United States jurisdictional wetlands are those wet areas that are protected
by law through Section 404 of the Clean Water Act and the Swampbuster Provision of the
Food Security Act (Mitsch and Gosselink 1993). The U.S. Army Corps of Engineers and the
U.S. Environmental Protection Agency jointly define wetlands for purposes of Section 404
of the Clean Water Act as:
Those areas that are inundated or saturated by surface or ground water at a
frequency and duration sufficient to support, and that under normal circumstances
do support, a prevalence of vegetation typically adapted for life in saturated soil
conditions. Wetlands generally include swamps, marshes, bogs, and similar
areas, (33 CFR 328.3 (b), 1984).
Currently, jurisdictional wetlands in the United States are those that meet the criteria
defined in the 1987 Corps of Engineers Wetlands Delineation Manual (Environmental
Laboratory 1987). These are not inclusive of all wetlands included in the classification of
Cowardin
-------
METHODS
Since 1988, we have been continually developing various assessment protocols to address
the health of riparian and wetland sites. The first assessment that was developed was the lotic
inventory. At the beginning of the process, we evaluated a wide range of inventory procedures.
Most were developed for upland sites and not always applicable. Therefore, in many instances,
we had to develop new procedures for use in a riparian or wetland site. Through a series of
workshops with a large number of natural resource professionals, a lotic inventory assessment
began to take shape. We utilized the Delphi approach or expert opinion approach for developing
the assessment. The Delphi approach is designed to bring together the experts in the field of
study and develop a consensus on a topic.
In the beginning, the protocols evolved at a rapid rate as the field personnel provided
invaluable feedback. As the years have progressed, the assessment and the codes and
instructions for the form has evolved to where today it contains over 800 data base fields
comprising detailed information on vegetation, physical site, and hydrology data.
As work progressed on the lotic inventory assessment, the public and many natural
resource professionals began asking questions about the "health" of a riparian or wetland zone.
In early 1992, we began the process of developing a methodology that would address the health
issue. Once again, through a series of workshops with natural resource professionals, a lotic
health assessment was developed.
To date the collaborative effort has resulted in the development of the following
assessments for riparian and wetland areas: (1) lotic inventory; (2) lotic health evaluation (derived
from the lotic inventory); (3) lotic health assessment (stand-alone); (4) river health assessment
(stand-alone); (5) lentic inventory; (6) lentic health evaluation (derived from the lentic
inventory); and (7) lentic health assessment (stand-alone). Each of the assessments also includes
a discussion on the codes or instructions used with each form.
RESULTS AND DISCUSSION
The health of a site may be defined as the ability of that system to perform certain
wetland functions. A site's health rating may also reflect management considerations.
For example, although noxious weeds such as Centaurea maculosa (spotted knapweed) or
Euphorbia esula (leafy spurge) may help to trap sediment and provide soil-binding properties,
other functions (i.e., productivity and wildlife habitat) will be impaired and their presence should
be a management concern.
No single factor or characteristic of a wetland site can provide a complete picture
of either site health or the direction of trend. For example, the lotic assessment is based on
consideration of channel and riparian vegetation factors. It relies extensively on vegetative
characteristics as integrators of factors operating on the landscape. Because they are more
visible than soil or hydrologic characteristics, plants may provide early indications of riparian
health as well as successional trend. These are reflected not only in the types of plants present,
237
-------
but also by the effectiveness with which the vegetation carries out its wetland functions of
stabilizing the soil, trapping sediments, and providing wildlife habitat. Furthermore, the
utilization of certain types of vegetation by animals may indicate the current condition of the
wetland and may indicate trend toward or away from potential natural community (PNC).
In addition to vegetation factors, an analysis of site health and its susceptibility to
degradation must consider physical factors (soils and hydrology) for both ecological and
management reasons. Changes in soil or hydrologic conditions obviously affect functioning of
a wetland ecosystem. Moreover, changes in physical characteristics are often (but not always)
more difficult to remedy than vegetative changes. For example, extensive incisement (down-
cutting) of a stream channel may lower the water table and thus change site potential from a
Fraxinus pennsylvanica'Prunus virginiana habitat type to an Artemisia cana/Agropyron smithii
habitat type. Sites experiencing significant hydrologic, edaphic (soil), or climatic changes will
likely also have a change in plant community potential.
The assessments attempt to balance the need for a simple, quick index of health against
the reality of an infinite range of wetland situations. Although this approach will not always
work perfectly, we believe that in most cases it will yield a usefully accurate index of riparian
health. Some more rigorous methods to determine status of a stream's channel morphology are
Dunne and Leopold (1978), Pfankuch (1975), and Rosgen (1996). These relate their ratings to
degree of channel degradation, but do not integrate other riparian functions into the rating. Other
methods are available for determining condition from perspectives that also include vegetation;
most notably the user guide on proper functioning condition (PFC) published by the BLM
(1998).
Potential Uses
The rapid lotic health (stand-alone) assessment procedure promulgated by the BLM
(1998) has been tested in Montana, surrounding states, and western Canada since 1992.
Currently, over 5,000 people have been trained using the assessment. Some potential uses for
this health rating include: (i) stratifying streams or wetlands by degree of ecologic dysfunction;
(2) identifying ecologic problems; and (3) when repeated over time, monitoring to detect
functional change. A less direct, but also important, value of an environmental assessment of
this kind is its educational potential. By getting land managers to focus on individual riparian
functions and ecologic processes, they may come to a better understanding how the parts work
together and are affected by human activities.
Once land managers have determined health of the reach of stream in question, they next
need to determine the appropriate course of action, if any. If the stream reach rated Proper
Functioning Condition (Healthy), then no action may be needed. If the stream reach rated
Functional At Risk (Healthy, but with problems) or Nonfunctional (Unhealthy), the manager
needs to determine what remedy is appropriate. The assessment form is divided into two
categories: vegetation and physical site factors. The land manager should review the assessment
to see which category rated low. This will indicate the prime area of focus. The publication by
Hansen et at. (1995), Classification and Management of Montana's Riparian and Wetland Sites
offers assistance in this area. Suppose, for example, a stream reach was rated at 54%, and
238
-------
a review of the health assessment form revealed major problems in five areas: (1) altered
streambanks; (2) lateral cutting of the streambank; (3) cover of undesirable herbaceous species;
(4) utilization of trees and shrubs; and (5) tree and shrub regeneration. These major problems are
determined by comparing the actual value against the possible value for each factor. This tells
the manager that the banks are eroding because high use is impacting the banks and reducing
woody species cover. If potential for the site is woody species as determined from the habitat
types or community types recorded on the lotic inventory form, and there are low values for both
utilization and regeneration of woody species, the manager may accelerate the restoration
process by planting woody species to help stabilize the streambanks. The appropriate woody
species and methods for planting them can be found in Hansen et al. (1995) or some other
appropriate publication. If livestock are causing the problem of tearing down stream banks or
overbrowsing them, changes in grazing regime are needed before planting to prevent new plants
from being browsed. Management change can include measures designed to discourage
livestock from spending long periods along the streambanks.
Types of Assessments
Through the years we have developed a variety of assessments for both lotic systems
and lentic systems. The following is a listing and brief description of the assessment protocols.
Lotic Inventory — A comprehensive inventory of a stream segment and its associated
riparian area, including detailed vegetation data, physical site data, some wildlife data, trend
commentary, and photographs. The inventory form contains over 800 data base fields. The
vegetation data collected includes species identification and canopy cover estimations, as well
as age class breakdowns for each tree and shrub species. Physical site data include channel
morphology and condition, substrate composition, disturbance degree and kind, amount and
cause of bare ground, and commentary. Wildlife data include details of activity by beaver
(Castor canadensis) and observations of fishery, amphibian, and reptile data. Currently, this
approach has been used on over 4,828 km of streams and rivers in western North America.
Lotic Health Evaluation -- An evaluation of riparian functional health derived from data
collected in the Lotic Inventory form. An array of vegetation (biotic) and physical site (abiotic)
items are weighted and rated for calculation of a health evaluation index score. The items
include information on hydric soils, hydrophytic vegetation, and wetland hydrology.
Lotic Health Assessment (stand-alone), (Thompson et al. 1998) - A rapid assessment
of lotic site functional health based on a similar set of factors as the Lotic Health Evaluation,
but derived from on-site estimation instead of from the detailed Lotic Inventory form. This
assessment has been taught to over 5,000 land owners/managers in Montana, Idaho, North
Dakota, Colorado, Utah, and the four western Canadian Provinces of Alberta, Saskatchewan,
Manitoba, and British Columbia.
River Health Assessment (stand-alone) — A rapid assessment of river functional health
based on a set of factors similar to the Lotic Health Assessment, but with some differences to
take into account differences of a river system vs. a stream system.
239
-------
Lentic Inventory — A comprehensive inventory of a lentic site and its associated
functional wetland area, including detailed vegetation data, physical site data, some wildlife
data, trend commentary, and photographs. The inventory form contains over 800 data base
fields. The vegetation data collected includes species identification and canopy cover
estimations, as well as age class breakdowns for each tree and shrub species. Physical site data
include shoreline morphology and condition, substrate composition, disturbance degree and kind,
amount and cau§e of bare ground, and commentary. Wildlife data include observations of
fishery, amphibian, and reptile data.
Lentic Health Evaluation, — An evaluation of wetland functional health derived from
data collected in the Lentic Inventory form. An array of vegetation and physical site items are
weighted and rated for calculation of a health evaluation index score. The items include
information on hydric soils, hydrophytic vegetation, and wetland hydrology.
Lentic Health Assessment (stand-alone) — A rapid assessment of lentic site functional
health based on a similar set of factors as the Lentic Health Evaluation, but derived from on-site
estimation instead of from the detailed site inventory.
The various assessments in terms of type of data collected (vegetation vs. physical data),
level of effort required, and estimated distance each day that can be traveled by an evaluator are
listed in Table 1.
Table 1. Type of data collected, level of effort required, and estimated kilometers/day
Assessment
Detailed
yeg. data
Detailed
physical data
Level of effort
required
Potential Km/day
by evaluator
Lotic Wetlands
Lotic Inventory
Lotic Health Evaluation
(derived from lotic
inventory form)
Lotic Health Assessment
(stand-alone)
River Health Assessment
(stand-alone)
Yes
No
No
No
Yes
No
No
No
High
High
Moderate
Moderate
Low
Low
Moderate
Moderate
Lentic Wetlands
Lentic Inventory
Lentic Health Evaluation
(derived from lentic
inventory form)
Lentic Health Assessment
(stand-alone)
Yes
No
No
Yes
No
No
High
Moderate
Low
Low
Moderate
240
-------
The current assessment protocols can be obtained at the web site www.rwrp.umt.edu.
The forms and their codes and instructions are available for downloading using the free program
called Adobe Acrobat®. The files are PDF (Portable Document Format) files.
Assessment Limitations
These assessments are not designed for in-depth or comprehensive analysis of ecologic
processes. Such analysis may be warranted on certain sites, and these can be done after this
evaluation has identified areas of concern.
These assessments attempt to balance the need for a simple, quick index of health against
the reality of an infinite range of situations. There are some visible changes to site health for
which we have no simple way to measure. An obvious and commonly encountered example is
excess entrained sediment. This may indicate serious degradation, but we leave it out of the
assessment due to difficulty in knowing how much is normal. Instead, we address on-site causes
of sediment production: bare ground, banks with poor root mass protection, and human-caused
structural damage to the banks. Another potentially serious degrading factor for which we have
no simple measurement yet is de watering of the system by irrigation diversion/pumping and by
upper drainage retention dams. Although these approaches will not always work perfectly, we
believe that in most cases they will yield a usefully accurate index of riparian or wetland health.
No single factor or characteristic of a riparian site can provide a complete picture of
either site health or the direction in which it might be heading. Because of inherent dynamics of
such systems, riparian sites often contain a mix of indicators. Moreover, characteristics that in
traditional evaluations of ecological sites have been considered negative may not be so in
riparian sites. For example, bare soil, which often reflects overgrazing or erosion on upland
sites, may be only a reflection of normal riparian activity, such as recent sediment deposits
resulting from spring runoff or a high water event. The ratings in the evaluation form have been
weighted to take such situations into consideration. Thus, only human-caused bare ground is
rated negatively, although naturally occurring bare ground can be considered an indicator of
susceptibility to impacts such as erosion and weed invasion.
A single evaluation provides a rating at only one point in time. Due to the range of
variation possible on a riparian or wetland site, a single evaluation cannot define absolute status
of site health or reliably indicate trend (whether the site is improving, degrading, or stable).
To measure trend, health assessments should be repeated in subsequent years during the same
time of year. Evaluation should be conducted when most plants can be identified in the field
and when hydrologic conditions are most nearly normal (e.g., not during peak spring runoff or
immediately after a major storm).
Each assessment has its strengths and weaknesses. Our overall goal has been to provide
land managers with a variety of "tools" from which to choose in their "toolbox". Determination
of which tool is the best for each job can only be answered after the specific goals and objectives
are determined for the project (e.g., a "needs assessment").
241
-------
ACKNOWLEDGEMENTS
Financial support for this work was provided by the USDIBLM and F&WS. In addition,
we thank the numerous USDI BLM and F&WS personnel for their valuable critique through the
development of the assessment protocols. In particular, we thank Vito Ciliberti, Tim Bozorth,
Joe Frazier, Buck Damone, Jim Rosco, Brian Hockett, Jim Sparks, Jeff Gardetto, and Dan
Bricco. We also thank Greg Hale, Barry Adams, and Lome Fitch of the Alberta Riparian Habitat
Management Program (Cows & Fish) for their assistance 'during the review stage.
REFERENCES
American Fisheries Society (Western Division). 1980. Position paper on management and
protection of western riparian stream ecosystems. American Fisheries Society, Bethesda,
Maryland. 24 pp.
Boldt, C.D., D.W. Uresk, and K.E. Severson. 1978. Riparian woodlands in jeopardy on
Northern High Plains. In: Strategies for Protection and Management of Floodplain
Wetlands and other Riparian Ecosystems. R.R. Johnson and J.F. McCormick (Technical
Coordinators). USDA Forest Service General Technical Report WO-12. Washington,
D.C. pp. 184-189.
Cooperrider, A.Y., R.J. Boyd, and H.R. Stuart. 1986. Inventory and Monitoring of Wildlife
Habitat. USDI Bureau of Land Management, Denver Service Center, Denver Colorado.
858 pp.
1 „"! "';,' •, • : • • i "'"ii" ' ' ; ! ' • •'' '"
Cowardin, L.M., V. Carter, F.C. Golet, and E.T. LaRoe. 1979. Classification of Wetlands and
Deep Water Habitats of the United States. USDI Fish and Wildlife Service, Office of
Biological Services, Washington, D.C. Publication Number FWS/OBS-79/31. 107 pp.
Dunne, T. and L.B. Leopold. 1978. Water in Environmental Planning. W.H. Freeman &
Company, San Francisco, California. 818 pp.
Environmental Laboratory. 1987. Corps of Engineers Wetlands Delineation Manual. Technical
Report Y-87-1. U.S. Army Engineer Waterways Experiment Station, Vicksburg,
Mississippi. 100pp.
Federal Interagency Committee for Wetland Delineation. 1989. Federal Manual for Identifying
and Delineating Jurisdictional Wetlands. U.S. Army Corps of Engineers, U.S.
Environmental Protection Agency, USDI Fish and Wildlife Service, and USDA Soil
Conservation Service Cooperative Technical Publication, Washington, DC. 76 pp.
242
-------
Hansen, P.L., R.D. Pfister, K. Boggs, B.J. Cook, J. Joy, and D.K, Hinckley. 1995. Classification
and Management of Montana's Riparian and Wetland Sites. Miscellaneous Publication
No 54. Montana Forest and Conservation Experiment Station, School of Forestry,
University of Montana, Missoula, Montana. 646 pp.
Johnson, R.R., and S.W. Carothers. 1980. Riparian Habitats and Recreation: Interrelationships
and Impacts in the Rocky Mountain Region. Produced under agreement 53-82 FT-0-125
of the Eisenhower Consortium for Western Environmental Forestry Research, Fort
Collins, Colorado. 109pp.
Kent,D.M. 1994. Applied Wetlands Science and Technology. D.M. Kent (Ed.). CRC Press,
Inc., Lewis Publishers, Boca Raton, Florida. 436 pp.
Kovalchik, B.L. 1987. Riparian zone associations: Deschutes, Ochoco, Fremont, and Winema
National Forests. USDA Forest Service Region 6 Ecology Technical Paper 279-87.
Pacific Northwest Region, Portland, Oregon. 171 pp.
Mitsch, W.J., and J.G. Gosselink. 1993. Wetlands. Second Edition. Van Nostrand Remhold,
Publishers, New York, New York. 722pp.
Pfankuch, D.J. 1975. Stream Reach Inventory and Channel Stability Evaluation. USDA Forest
Service, RI-75-002. Government Printing Office #696-260/200, Washington, DC. 26 pp.
Prichard, D., C. Bridges, S. Leonard, R. Krapf, and W. Hagenbuck. 1994. Riparian Area
Management: Process for Assessing Proper Functioning Condition for Lentic Riparian-
Wetland Areas. Technical Reference 1737-11. USDI Bureau of Land Management.
Denver Service Center. 39 pp.
Rosgen, D.L. 1996. Applied River Morphology. Wildland Hydrology, Pagosa Springs,
Colorado. 246pp.
Shaw, S.P., and C.G. Fredine. 1956. Wetlands of the United States: Their Extent and their
Value for Waterfowl and Other Wildlife. USDI Fish and Wildlife Service, Circular 39.
Washington, D.C. 67pp.
Stewart, R.E., and H.A. Kantrud. 1972. Classification of Natural Ponds and Lakes in the
Glaciated Prairie Region. USDI Fish and Wildlife Service, Research Publication 92.
57pp.
Thompson, W.H., R.C. Ehrhart, P.L. Hansen, T.G. Parker, and W.C. Haglan. 1998. Assessing
health of a riparian site. In: Proceedings of the American Water Resources Association
Specialty Conference, Rangeland Management and Water Resources. D.F. Potts (Ed.)
American Water Resources Association, Herndon, Virginia. TPS-98-1, pp 3-12.
243
-------
USDIBLM (U.S. Department of Interior Bureau of Land Management). 1998. Riparian Area
Management: A User Guide to Assessing Proper Functioning Condition and the
Supporting Science for Lotic Areas, Technical Reference 1737-15. Bureau of Land
Management National applied Resource Sciences Center, Denver, Colorado. 125 pp.
Windell, J.T.,B.E. Willard, DJ. Cooper, S.Q. Foster, C.F. Knud-Hansen, L.P. Rink, and G.N.
Kiladis. 1986. An Ecological Characterization of Rocky Mountain Montana and
Subalpine Wetlands. USDI Fish and Wildlife Service Biological Report 86(11).
National Ecology Center, Division of Wildlife and Contaminant Research, Fish and
Wildlife Service, U.S. Department of the Interior, Washington, DC. 298 pp.
244
-------
TRACE METALS IN SURFACE SEDIMENTS FROM THE BAJA CALIFORNIA-
CALIFORNIA BORDER AREA AND THE SEA OF CORTEZ
J. Vinicio Macias-Zamora1
ABSTRACT
The Southern California Bight and the Sea of Cortez (or Gulf of California) are very rich
and complex systems. Circulation patterns have not been very well established although more is
known for the former area and some general models of circulation have been proposed. The
circulation pattern for the Sea of Cortez however, still remains to be understood although again,
good progress has been made on this subject. The centers of population surrounding the
Southern California Bight are larger than those on the Sea of Cortez, and as such, larger
anthropogenic impacts are to be expected and have been recorded. The sharing of resources
between the United States and Mexico has resulted in a large monitoring project coordinated
between scientists in both countries. Large efforts have been devoted to intercalibration
exercises to insure comparability. Fluxes of pollutants at the border are controversial and remain
to be elucidated over longer periods of time. Recurrent, north to south patterns of pollutant flow
have been found in Mexican waters. Simultaneously, water discharged from the Tijuana River
on the U.S. coast has been largely responsible for frequent beach closures and has been a source
of irritation for local communities. A Pilot Project for the Southern California Bight will attempt
to resolve some of the questions which we are facing. The insured comparability of data will
permit valid comparisons across the border, allowing the interpretation of sediment and
pollutants flow.
INTRODUCTION
Marine pollution has been in the minds of people in general and scientist in particular for
several decades now. Recently however, a large effort has been devoted by nations with the
signing, by more than 100 countries, of the Global Programme of Action for the Protection of the
Marine Environment from Land-Base Activities. Among the stated purposes of the plan, one of
them is to develop or strengthen national as well as regional and sub-regional programs to reduce
the degradation of the marine environment. This program, for our region, reinforces the fact that
many national studies have stopped at the border. This is clearly not the case for pollutants. If
we plan to understand what is happening to these materials in the region, combined efforts have
to be undertaken.
Institute de Investigaciones Oceanologicas, Universidad Autonoma de Baja California
245
-------
Several monitoring studies have already been conducted on both sides of the border.
the purpose of these studies has been, among others, to establish the effects of local point source
discharges of trace metals on the biota and, in general, on the quality of water and sediments.
These studies, however, have been carried out independently in both countries. U.S. scientists in
thie, San Diego, California area have been monitoring the impact, if any, of the Point Loma
wastewater plant on local beaches and its effect on biological resources. Almost simultaneously,
scientists on the Mexican side have attempted to establish baseline data for the area, as well as
conducting some studies to determine possible impacts of the local Punta Banderas wastewater
treatment plant. Similar programs have been carried out for organic pollutants such as
polyaromatic hydrocarbons (PAHs) and polychlorinated biphenyls (PCBs). The Mexican efforts
however, have been more limited, both in area and scope. Additionally, these efforts have met
with the inconvenience of different methodology, differences in the quality controls of the
measurements, different timing, and even different collection, preservation, and sample plan
design. As a result, the numbers generated are difficult to compare (see Table 1 for examples).
In particular, in the case of trace metals, they are normally present due to naturally
occurring geochemical cycles. Normal accumulation of trace metals also occurs in places where
organic matter content is highland where grain size is small (less energy, more sedimentation).
Researchers have devised procedures to normalize against the organic matter content effect
and/or grain size effects The challenge is, then, to distinguish these background levels from
those enriched by anthropogenic activities. One procedure that has been extensively used is to
normalize with respect to a metal. Many authors have included such a normalization procedure.
However, the metals have to be present in such a large quantity that anthropogenic effects can
not significantly affect their concentrations. Typically, for example, aluminum, iron, lithium, as
well as others, have been selected for this purpose, but unfortunately, different authors use
different metals. In our case, the choice of metal for normalization has been traditionally
decided based on the method used to digest the samples. We have not used hydrofluoric acid to
digest the sediment samples, as a consequence the only choice has been iron content of the
sediment. The percent recovery for aluminum, under these conditions, is poor.
Variations observed in values reported in Table 1, although they might be real, e.g., the
concentrations for manganese, might also be partially attributable to differences in sample
treatment, digestion procedure and so on. A similar situation has also been observed for some
ofganic pollutants that have been investigated on both sides of the border. In previous studies,
PAHs have been determined in Mexican sediments along the border (Macias-Zamora 1996, and
Macfas-Zamora et al. 1997) and have been compared to those reported in the San Diego area of
the U.S. Comparisons again should be made with some reservation because of the differences in
methods of sample collection, extraction, and analysis.
One of the most critical points is the perception by the public, that one country is
polluting the beaches, the waters, and affecting also the biota in the other country. The public
could be misinformed, or maybe mislead about the problem, but more often than not, at least at
this geographic location, pollutants circulate in both directions with differences in both quality
and quantity, and the extent of this has yet to be determined.
246
-------
Table 1. Metal concentrations reported in different studies which varied in methodology,
collection at different times, and at different locations. (Concentrations in ng/g)
Study
Gutierrez-Galindo, etal. (1994)
Villaescusa-Celaya, etal. (1997)
Thompson, et al. (1993)
Romero-Vargas (1995)
Cu
2.6
7.6
5.3
11.8
Zn
7.3
40
29.1
46
Cd
0.13
0.10
0.26
--
Cr
11.9
68
17.0
42.6
Mn
34.7
428
99.0
510
Ni
—
21
8.0
11.8
Ag :
0.14
0.02
0.10
—
THE SOUTHERN CALIFORNIA BIGHT
Description
The Southern California Bight, also know as the Bight of the Californias, is an area
caused by a pronounced change in the geographic shape of the Pacific coast, making the general
circulation pattern of the ocean a very complex one (Figure 1).
31 -
121 120
w
119
118
117
116
115
Figure 1. The Southern California Bight. The figure shows the narrow platform were sampling
took place in 1998 for the Bight of the Californias pilot project. The change in direction
of the coast produces complex nearshore circulation.
247
-------
As one of the consequences, residence times of pollutants in the Bight tend to be longer
than elsewhere along the coast, making this area an inappropriate place to discharge pollutants.
This is particularly true for those pollutants known as persistent organic pollutants (POPs) and
trace metals, both" of which are expected to remain part of the biogeochernical cycles for long
periods of time. The circulation pattern close to the beach is not well known. The waves have
been reported to reach the coast frequently with an angle that contains a net component towards
the south. Evidence for this behavior has been found in the size distribution of sediments along
the beach on the Mexican side of the international border. Another very important feature of this
area is the large human population living along the coast. This population has been estimated to
be around 20 million people with approximately 1.5 million living on the Mexican side of the
Bight.
Most of the wastewater discharges to the Bight occur from four large wastewater
facilities. tKese are the Hyperion Treatment Plant, the Joint Water Pollution Control Plant,
Wastewater treatment Plants 1 and 2 in Orange County, and the Point Loma wastewater
treatment Plant in San Diego. Additionally, there are another 15 smaller facilities all of which
discharge to the Bight (Raco Rands 1997). South of the international border, there are two point
sources that are of concern. The first one is the Tijuana River, which drains a basin located
mainly on the Mexican side, but the river mouth discharges are located on the U.S. side and,
second, the wastewater treatment facility at Punta Banderas, located a few miles from the
international border. There is a third wastewater treatment facility under construction in San
Ysidro, California called the Binational Wastewater Treatment Plant. This will process most of
the discharge from the Tijuana River and will release the treated water (primary treatment) out
into the Ocean, very close to the international border.
Research Program
In 1995 Mexico joined more than a hundred other countries in an international effort
designed to protect the marine environment from pollutants originating on land. The program
calls for international cooperation, especially on shared water bodies such as the Southern
California Bight. Through the coordination of the Commission for Environmental Cooperation,
Mexico and the United States have started one of two pilot projects designed to improve
collaboration and optimization of resources in search of solutions to areas perceived to be in
danger. Those areas are especially important and it is where regional cooperation is fundamental
to protect valuable resources on both sides of the border.
11 • i ' i
To achieve some of the objectives for this work, in particular, to permit full comparison
of data between scientists on both sides of the border, a large effort has so far been devoted to
intercalibration exercises. The intercalibration exercises have included the use of the same
samples by all involved laboratories, and a set of quality assurance procedures and achievable
quality controls. Each set of controls was designed for a particular group of chemicals and
analyses, for example, all measured chemicals in sediments, microbiology tests for total and
fecal coliforms by the most probable number (MPN) technique, and intercalibration for the
Response of all sensors of each conductiviry/temperature/density (CTD) electronic probe under
similar conditions.
248
-------
The Mexican efforts are focussing locally, to answer questions mostly related to
geographical aspects of pollutant distribution and the understanding of pollution history in the
region. The coastal region shared by California in the U.S. and Baja California in Mexico is
located within the previously described well-defined oceanographic system, historically known
as the Southern California Bight (SCB). On the Mexican side, the SCB area extends from the
U.S. border to Cabo Colonet, just north of San Quintin. The Mexican effort, however, extends
only to Todos Santos Bay in coincidence with other border programs such as the Mexico-U.S.
Border XXI Program. Additionally, because most of the population activities are related to
coastal or nearshore activities, the program as a whole has been designed to focus the sampling
effort only from the coast, from a depth of as little as 6 m to as much as 200 m.
Although the levels of concentrations for contaminants at which action should be
implemented is still a controversial area of research, we are using, only as a matter of
comparison, those values already used in the previous survey in SCB waters conducted in 1994.
The effect range low (ERL) and the effects range median (ERM) approach described by Long
and Morgan (1990), MacDonald (1994), and MacDonald et al (1996) will be used as an
approximation to determine the areal extent of the contamination. An extra situation that is
coinciding with these efforts is the development of an "El nino" condition. This phenomenon
has recently brought to the area more than usual precipitation. This, in turn, has produced larger
than average sediment and particle loads to the area. It is expected that the presence of large
amounts of new material might affect the otherwise average concentration of pollutants, although
this remains to be seen.
The area to be studied is shown in Figure 2, and the information that we hope to obtain to
answer some of our questions can be summarized as follows: (1) The concentration levels of
both linear and PAHs in surface sediments on the coastal shelf; (2) The concentration levels for
pesticides, in particular DDT and its derivatives, as well as Dieldrin; (3) The concentration and
distribution of PCBs by specific congeners, 41 of which will be studied; (4) The concentration
levels of trace metals; (5) The concentration and the distribution of linear alkyl benzenes in the
sediments; (6) The measurement of concentrations of those chemicals listed in (1) - (5) on at
least three to five sediment cores; (7) The water quality parameters temperature, salinity,
dissolved oxygen, pH (and chlorophylls at some stations) at more than 60 stations divided in
transects in agreement with the sampling design used in the U.S. monitoring plan. Nutrients and
chlorophyll measurements will be taken in at least 10 % of the sample sites; (8) The
measurement of bacteriological variables (total coliforms, fecal coliforms, and enterococcus)
using the multiple fermentation tube (MFT) technique known as the most probable number
(MPN).
The selection of sampling sites has been based on a randomly stratified design (Stevens
1997). In close coordination with U.S. scientists, we have selected only three strata (Figure 3)
based on two reasons. First, there are fewer features along the Mexican coast than along the U.S.
coast. For example, there is only one port, one finished marina, and only two other marinas are
under construction. There is also, only one bay (Todos Santos Bay), which makes selection of
strata simpler. The first stratum selected represents the northern area extending from the border
249
-------
OCZAK)
mtuitu
Figure 2. The water quality component of the Mexican plan consisted of transects perpendicular
to the coast and starting from known continental water and drainage inputs.
along the coast to just south of Rosarito Bay. This area is expected, because of previous studies,
to be the most polluted (see for example: Sanudo-Wilhelmy and Flegal 1991, Macias-Zamora
1996, Macias-Zamora et al 1997, Villaescusa-Celaya et al. 1997). These studies have been
Carried out to answer different sets of questions and each has been conducted with different
methodologies and degrees of quality assurance/quality control, making it difficult to make
comparisons across the border and even between studies in the same area.
250
-------
33
N
32 -
III., ..
U.S.A.
MEXICO
ENSENADA
Todos Santos Bay
W
117
116
Figure 3. Sampling program for the Mexican waters in the Bight of the Californias pilot project.
The figure shows the three strata selected for the project.
The second stratum, south of the first, was selected to include Todos Santos Bay. Both
this and the first stratum were designed to include 35 sampling sites assigned by the same
statistical procedure as the sites in U.S. waters (Stevens 1997). The third stratum is expected to
represent a reference site. It is also expected to be the least polluted of the three due to a scarce
population distribution in the area. Given the narrow platform and the expectation of rocky
bottoms, a lesser number of samples (25 samples) was originally assigned to this area. A total of
72 samples were actually collected. Some were eliminated because of rocky bottom; a few were
eliminated because they were located at depths exceeding 200 meters, and so on.
251
-------
With the present design, we expect for the first time to be able to compare the magnitude
of pollution among these areas, as well as calculate the areal extent affected by certain chemicals
or combination of chemicals. At the same time, and given that a large effort has been devoted to
intercalibratipn of techniques used on both sides, a valid comparison will be established, also for
the first time, between Mexican and U.S. sites. In addition to the geographic perspective, we
have decided to take a look at the historical aspects of pollution. In this respect, four core
sediments were collected using a box corer to obtain undisturbed sediments. The length on each
core is between 30 to 50 cm. If well preserved, and no significant bioturbation is found, these
cores will offer a first look at the pollution history and its rate of change versus time on the
Mexican northwest coastal region.
A specific list of variables for which samples were collected and which are now being
analyzed is reported in Table 2. Additionally, organic carbon content and grain size analysis will
be performed on these samples.
,j,
Table 2. Chemical variables being measured in Mexican waters under the joint U.S. Mexican
study program.
PAH's - 24 family members
PCB's - 41 specific congeners
Organics (n-hydrocarbons)
Pesticides - 6 DDT and related derivatives
LAB's - 16 members of the family
Inorganic (Trace metals, in sediments)
Antimpny
Arsenic
Cadmium
Cobalt
Copper
Iron
Lead
Mercury
Nickel
Selenium
Silver
Zinc
Benthic Infauna
Frequently, the concentrations of contaminants on sediments have an impact on the
distribution and types of benthic organisms (see for example Clarke and Ainsworth 1993,
Warwick and Clarke 1993 and references therein). More often than not, however, environmental
studies do not include a benthic component. The researchers in southern California have made a
part of their monitoring research the inclusion of the structure of benthonic communities. Our
sampling procedure also included the collection of a full set of 72 samples for analysis of the
benthic composition and those are in the process of separation and cleaning. At this time,
however, there has not been a calibration or standardization between our technicians and the
scientists working on the U.S. side. This lack of standardization could limit the validity of our
results with respect to the intercomparability. After considerable discussion it was decided that
one of the parameters to be measured in these samples is the biomass of the six main groups into
which these organisms are classified. Although at this time no group or individual has indicated
the willingness to participate in collecting and analyzing this part of the project, our chemistry
group will be responsible for collecting and preserving the benthic organisms for later
252
-------
identification, and for which additional funding will be pursued. At each of those sites where
sediments were collected for chemical measurement, an additional sample was collected.
Sieving, collection, and preservation of these benthic organisms was made according to the
established procedures.
Water Chemistry
CTD measurements will be carried out at approximately 100 stations (Figure 3) starting
at the border and extending to Punta Banda in Todos Santos Bay. The parameters to be included
are pH, temperature, salinity, and dissolved oxygen. At present, the Mexican participants have
received support for the October 1999 field sampling program, a date selected as one
representing the dry season conditions. Two other dates will be sampled in the rainy season on
the U.S. side; however, we might be able to collect samples only on one other occasion during
the rainy season and after one important event. The coordination for these samplings, because
they are dependent on the rain pattern, will not necessarily occur simultaneously, and also, no
specific date can be set. Discrete samples will be collected for calibration purposes on dissolved
oxygen, salinity and chlorophyll. This last parameter will be done by fluorometer (in situ),
spectrophotometer, and by high-performance liquid chromatography (HPLC) techniques.
Coastal Microbiology
One of the most sensitive issues is the one due to diminishing of water quality by the
presence of bacteria on frequently used beaches. The complex circulation pattern makes the
situation worse. The Mexican group will evaluate, by the Multiple Fermentation Tube (MFT)
technique and using the MPN counting, three parameters in about 30 locations during 5 weeks,
one sampling per week, and on two occasions including, rainy and dry seasons. The parameters
included are: Total coliforms, fecal coliforms and enteroccocus. Extensive intercalibration
exercises were also conducted and the Mexican group performed well on all exercises to date.
THE GULF OF CALIFORNIA
The Gulf of California (Figure 4), known also as the Sea of Cortez, is probably one of the
most studied geographic locations in the world. There is an excellent book of references that
lists most of the works done in the area (Schwarztlose et al. 1992). However, the chemistry of
sediments is one of the most neglected topics. Much effort has gone into understanding the
circulation, the biology, and the general hydrographic parameters. A good deal of effort has also
been placed on geochemical studies. In particular, due to its rich nature and very high rates of
biological productivity, the decomposition of the large amount of organic material produced has
resulted in a very large volume of water with little or no oxygen, creating suboxic or possibly
anoxic condition. This enormous feature of the Gulf has also resulted in, among other effects,
the preservation of undisturbed sediments. One of the characteristics of these sediments is the
alternating dark and light colored sequence. Although the explanation of why these so called
253
-------
"varved" sediments present this alternancy in color is still a matter of debate (see for example
Baumgartner et al. 199la, 1991b and references therein), in any case it has allowed for a
particularly useful marking of events in geological sense.
The northern part of the Gulf, close to what used to be an almost permanent source of
fresh water, the Colorado River, has seen very large changes in the frequency of discharges from
that river. Dewatering of the Colorado River, principally within the U.S. before discharging into
the Gulf, has produced significant alterations in the physiognomy of the lower Colorado River
delta. Vegetation has dramatically diminished, wetland and riparian environments have almost
disappeared, and the flow of nutrients and sediments to the northern portion have all but stopped.
In the Gulf itself, the lack of frequent inputs of nutrients and sediments from the Colorado River
have altered the ecology of the northern portion of the Gulf. In particular, it has been frequently
reported that fish species such as the drum totoaba (Totoaba macdonaldi) and the "vaquita"
[Phocoena sinus), both of which are endangered, as well as the shrimp fisheries, have been
altered as a result of the lack of input of both fresh water and nutrients. Saline intrusion of Gulf
water represents an additional threat to the agricultural areas in the Mexicali Valley, through
Which the Colorado River flows before entering the Gulf.
N
34
32
30
28
26
24
22
20
i * ' ' i ' ' ' r '
USA
Mexico
Angel de la Guardla Island
Tiburon Island
Pacific Ocean
118
W
116
114
112
110
108
106 104
Figure 4. The Gulf of California. The thin arrows indicate the main water circulation pattern.
The tidal forcing of currents occurs through the channel closer to Mexico's mainland and
the exit of the water occurs through the channel between Angel de la Guarda Island and
Baja California. The curved arrow indicates the main pathway for sediment transport.
254
-------
At the same time, the lack of input of sediments due to extensive up-river damming has
forced the reworking of sediments deposited on the delta. An indication that the lack of
sediment input from the Colorado River has not had an effect on sedimentation patterns inside
the Gulf was advanced by Baba et al. (1991). He concluded that although the recent discharges
of Colorado River water to the Gulf have been reduced to only about 17500th of those prior to
damming of the River, there appears to be a mechanism by which, sedimentation rates (measured
in a sediment core collected in the northern part of the Gulf) have not changed significantly.
The mostly counter-clockwise and highly structured surface circulation (anticyclonic,
during winter) on the northern part of the Gulf has been producing a net transport of sediments
toward the Baja California coast and away from the original delta (Soto-Mardones et al. 1999).
The confirmation that the transport of sediments or re-working of the delta sediments has been
occurring was presented recently by Carriquiry and Sanchez (1999). In their work, the proposed
mechanism for a net transport of sediments from the delta into the upper Gulf consisted of two
main flows. One, carrying sediments from the ocean and along the Sonoran Coast into the delta
system, and the other transporting sediments from the delta system, along the Baja California
coast into the Gulf of California. This counter clockwise transport appears to be responsible of
slowly reshaping the sedimentary deposits in the Colorado River Delta.
These relict sediments have not been studied with respect to their trace metal content. In
fact, most trace metal studies in the region have been focused on organisms, and in particular, on
the mussels Mytilus calif ornianus or M. edulis and Modiolus capax (see, for example, Gutierrez-
Galindo et al. 1990 and Gutierrez-Galindo et al. 1994). However, given that there is little
industrial activity in most of the region, most results reflect the difference in quality among local
sources.
One of the questions still under scrutiny relates to the circulation of water and sediments
within the Gulf. There are two large islands located in the northern part of the Gulf (Angel de la
Guardia and Tiburon), and these play a role in the circulation pattern of the water to their north.
The water circulation pattern within this northern region is generally (frequently) counter-
clockwise, and sediments circulating with the water within this region tend to stay within this
region, although some of the sediments may be exported. Because of continuous evaporation of
water confined within this northern region there is an increase in its salinity and buoyancy, and
this water mass has been reported well to the south of the islands. It is expected that sediments
might also be transported outside the region of the islands. Although a definitive answer has yet
to be provided, some indication of the export of these sediments has been reported by Da Costa
Gomez-Bueno and Valle-Diaz (1989). In this work, the trace metal content in the mussel
M. capax was investigated. Organisms were collected in many sites along the Baja California
coast. The results showed a significant gradient from north to south for the trace metals
aluminum, manganese, and zinc from Punta Estrella to Bahia de Los Angeles, this last place
located outside the restricted circulation area and south of the Island. In particular, the first two
elements are both well known for its terrigenous origin and they are presumed to be originating
in the sediments from the Colorado River delta.
255
-------
Other increases in trace metal concentrations (Cu, Zn, Al, and Mn) were attributable to
local sources, for example, to mining at Santa Rosalia. Finally, cadmium concentrations were
found to be high in the region near Bahia de los Angeles, but this increase is believed to be
associated to up-welling events within the Gulf
REFERENCES:
Baumgartner, T.R., V. Ferreira-Bartrina, J. Cowen, and A. Soutar. 199 la. Reconstruction of a
20th century varve chronology from the Central Gulf of California. In: The Gulf and
Peninsular Province of the Californias. J.P. Dauphin, and B.R.T. Simoneit (Eds.).
American Association of Petroleum Geologists Memoir 47.
Baumgartner, T.R., V. Ferreira-Bartrina, and P. Moreno-Hentz. 1991b. Varve formation in the
Central Gulf of California: a reconsideration of the origin of the dark laminae from the
20th century varve record. In: The Gulf and Peninsular Province of the Californias.
J.P. Dauphin and B.R.T. Simoneit (Eds.) American Association of Petroleum Geologists
Memoir 47.
"; , , •• ' ':- '••• •.• V"vr •• ,;
Carriquiry, J.D., and A. Sanchez. 1999. Sedimentation in the Colorado River Delta and Upper
Gulf of California after nearly a century of discharge loss. Marine Geology. (In press).
Clarke, K.R. and M. Ainsworth. 1993. A method of linking multivariate community structure to
environmental variables. Marine Ecology Progress Series 92:205-219.
DaCosta Gomez-Bueno, C.A., and N.A. Valle-Diaz. 1989. Disponibilidad biologica de metales
traza en el mejillon Modiolus capax del Mar de Cortez. Tesis. Facultad de Ciencias
Marinas, Ensenada, Baja California. 77pp.
Gutierrez-Galindo, E.A., G. Flores-Munoz, G. Olguin-Espinoza, and J.A. Villaescusa-Celaya.
1990. Bioavailability of trace metals in clams and mussels of the agricultural valley of
Mexicali and upper Gulf of California. Ciencias Marinas 16:1-28
Gutierrez-Galindo, E.A., G. Flores-Munoz, J.A. Villaescusa-Celaya, and A. Arreola-Chimal.
1994. Spatial and temporal variations of arsenic and selenium in a biomonitor (Modiolus
capax) from the Gulf of California, Mexico. Marine Pollution Bulletin 28:330-333.
•.: "', ":; ' i, ':. • ' ' '- • • ' ,i i
Long, E., and L. Morgan. 1990. The Potential for Biological Effects of Sediment-Sorbed
Contaminants in the National Status and Trends Program. Technical Memorandum
NOSOMA 52. National Oceanic and Atmospheric Administration, Seattle, Washington.
MacDonald, D. 1994. Approach to the Assessment of Sediment Quality in Florida Coastal
Waters. Volume 1. Development and Evaluation of Sediment Quality Guidelines.
MacDonald Environmental Sciences, Ltd. Ladysmith, British Columbia, Canada.
256
-------
MacDonald, D., R. Carr, F. Calder, and E. Long. 1996. Development and evaluation of
sediment quality guidelines for Florida coastal waters. Ecotoxicology 5:253-278.
Macias-Zamora, J.V., J. A. Villaescusa-Celaya, E. A. Gutierrez-Galindo, and G. Florez-Mufioz.
1997. Pollution studies on Pacific coastal waters of northern Baja California, Mexico:
In: Proceedings of the Fourth International Symposium on Fish Physiology, Toxicology
and Water Quality, Bozeman, Montana, September 1995. R.V. Thurston (Ed.). U.S.
Environmental Protection Agency, Ecosystems Research Division, Athens, Georgia,
USA. EPA/600/R-97/098. pp. 179-188.
Macias-Zamora, J.V. 1996. Distribution of hydrocarbons in recent sediments off the coast of
Baja California. Environmental Pollution 92:45-53.
Raco-Rands, V. 1997. Characteristics of Effluents from Small Municipal Wastewater
Treatment Facilities in 1995. Southern California Coastal Water Research Project
Authority. Annual Report 1996. Stephen B. Weisberg (Ed.). Santa Ana, California.
Romero-Vargas-Marquez, I.P. 1995. Metales pesados y su fraccion quimica en sedimentos de la
Bahia Todos Santos, Baja, California, Mexico. M.Sc. Thesis. Universidad Autonoma de
Baja California. Facultad de Ciencias Marinas. 86 pp.
Safiudo-Wilhelmy, S.A., and A.R. Flegal. 1991. Trace elements distribution in coastal waters
along the U.S.-Mexican boundary: relative contributions of natural processes vs.
athropogenic inputs. Marine Chemistry 33:371-392.
Schwartzlose, R.A., D. Alvarez-Millan, and P. Brueggeman. 1992. Golfo de California.
Bibliografia de las Ciencias Marinas. Universidad Autonoma de Baja California.
Ensenada, Baja California, pp.425.
Stevens, D.L., Jr. 1997. Variable density grid-based sampling designs for continuous spatial
populations. Environmetrics 8:167-195.
Thompson, B., D. Tsukada, and D. O'Donohue. 1993. 1990 Reference Site Survey. Technical
Report #269. Southern California Coastal Water Research Project Authority.
Westminster, California. 105 pp.
Villaescusa-Celaya, J.A., E.A. Gutierrez-Galindo, and G. Flores-Mufioz. 1997. Metales pesados
en fracciones geoquimicas de sedimentos de la region fronteriza de Baja California,
Mexico, y California, EUA. Ciencias Marinas 23:43-70.
Warwick, R.M., and K.R. Clarke. 1993. Comparing the severity of disturbance: a meta-analysis
of marine macrobenthic community data. Marine Ecology Progress Series 29:221-231.
257
-------
-------
RISK MANAGEMENT OF MERCURY POLLUTION IN CHINA
Lin Yuhuan1 and Chen Jianhua1
ABSTRACT
Mercury is used in a wide variety of industrial and civil applications in China, including
the chemical industry, chlor-alkali and polyvinylchloride (15%); metallugical industry, gold mines
and non-ferrous refining (35%); electronic and electrical industry, household batteries and electric
lighting (38%); medical instrument, thermostats, and dental uses (5%); and military and laboratory
applications (4%). From the 1980's to the 1990's, the amount of mercury used increased as much
as 3.5-fold in electronic and battery productions, and 30-80% in the gold mines. Environmental
scientists have warned of pollution by mercury. In this paper a strategic decision for the
management of pollution is discussed, and a proposal for management is suggested based on an
investigation of mercury application in industry and energy production.
INTRODUCTION
In China, the emission of mercury from anthropogenic processes is mainly from industrial
processes, energy production, and metal mining production, including the chlor-alkali industry,
battery and electric lighting industry, municipal waste, waste solid incineration, and waste disposal
sites. All this consumption and emission contributes to mercury pollution in the air, so the
concentration of mercury has increased 10-100 fold higher than the background level of the
northern hemisphere, especially during the winter heating season. The water and soil pollution
of mercury occurring in mining areas and the industrial regions of China is a serious problem.
The quality of drinking water in some area is threatening, and the risk of mercury pollution is
threatening civilian health and sustainable development of the economy. The increasing rate of
mercury use in China (Figure 1) is much higher than the world rate, especially in the metallurgical
industry, gold mining, and production of household batteries and electric lights (Xi 1997).
Research Centre for Eco-Environmental Sciences, Chinese Academy of Sciences, 100085 Beijing, China
259
-------
600
500
400
300
200
100
82
84
86
88
90
92
94
96
Year
Figure 1. Diagram of mercury consumed in China from 1983-1995.
INVENTORY OF ANTHROPOGENIC MERCURY EMISSIONS
Environmental releases of mercury can be natural in origin (e.g. geological deposits,
volcanic activities) or can occur from anthropogenic sources. In order to understand the
anthropogenic sources, an inventory of the amount of mercury used in China in 1995 is reported in
Table 1. The industries that process mercury the most are the electrical and electronic industry, the
metallurgical industry, and the chemical industry; these three industries alone account for over
85% of the total amount used in 1995. An inventory of anthropogenic mercury emissions in China
in 1995 demonstrate that energy production is another significant source (Table 2). In industrial
activities, the most significant were metal mining and production such as gold mining, mercury
mining, and non-ferrous metal mining and refining. Chlor-alkali production, household battery
production, and electric light production also release mercury into the environment, and discarded
household batteries have been important sources of mercury pollution since the 1990's.
" I ' ''"if •:•','. ' ' . ' , Til ' 1 . I .' . I .,.„•.•
Pollution in mining areas in China from gold and mercury mines is also a serious problem.
A major portion of regional emission of mercury to water bodies comes from small-scale gold
mining. Concentrations of mercury in wastewater and amounts of mercury released to river waters
are shown in Table 3. The pollution of mercury from the chemical industry is also an old problem
in China; the discharge of mercury by this industry is a source of water pollution (Table 4),
especially in adjacent coastal areas, such as Bohai Bay and Jinzhou Bay in northern China.
260
-------
Table 1. Mercury use in China in 1995. (From Lin 1998)
Industry
Chemistry
Metallurgy
Electronics and electrical
industry
Medical instrument
Production
Chlor-alkaliandPVC
Mercury, gold mining,
and non-ferrous refining
Batteries, electric lighting, and
electronic equipment
Percent of
total used
15
35
38
5
manufacturing and
medicine
Laboratory instruments and
military applications
Other uses
Painting, etc
4
3
Table 2. Emission of mercury from anthropogenic sources in China in 1995. (From Lin 1998)
Source
Percent of total emissions
Small scale gold mining
Energy production, coal and fossil fuel combustion
Chlor-alkali and PVC production
Household batteries disposal
Small-scale mercury mining
National gold mining
National mercury mining
Electric lighting
Non-ferrous metal refining
Other sources (Tooth fillings, thermometer,
laboratory instruments, military, paint)
38.8
27.9
9.3
7.75
4.65
3.1
3.1
2.33
1.6
1.6
Table 3. Mercury in wastewater of gold mining in China in 1995.
Sample
1
2
3
Average
Discharge mouth
(mg/L)
0.291
0.301
0.301
0.298
Amount
(g/ton of ore)
1.16
1.21
1.20
1.19
Drainage
(mg/L)
0.0116
0.0105
0.0109
0.0110
Background water
(mg/TL)
<0.0001
261
-------
Table 4. Mercury in wastewater of chlor-alkaline industry in China in 1997.
Sample
Volume (ktons) Concentration (mg/L) Amount (tons)
Discharge to river
Discharge to reservoir
Wastewater of alkaline
Wastewater of PVC
10,800
1,810
77.6
216
0.0948
0.219
0.552
1.36
1.01
0.396
0.0428
0.293
RISK OF MERCURY POLLUTION
Because a large amount of mercury has been discharged into the environment from
anthropogenic sources, and the concentration of mercury in water, air and soil has thereby been
considerably enhanced, there is a risk that pollution will occur in food webs, involving surface
water, sediments, drinking water (tap water and well water), soil, and rice and vegetable
E reduction. For example, the concentration of mercury in sediments and water in the cities of
eying and f ainjing were found to be higher than elsewhere, exceeding 1 ppm. The concentration
of methylmercury is also high (Table 5), as is the biota accumulation in fishes.
Table 5. Mercury in water and sediments of lakes and rivers in Beijing and f ainjing in 1997.
Sampling location
Total mercury
Methyl mercury
Kunming Lake
Shuizhuizi Lake
Beihai Lake sediment
Yuyantan Lake
Jinyun River
(mg/L)
0.599
0.360
4.68
0.451
0.101
S.D.(%)
14.0
19.4
6.96
10.9
11.3
(u.g/L)
0.11
0.35
1.25
0.37
0.017
S.D.(%)
9.09
5.71
16.8
13.5
18.2
Pollution of drinking water in mercury mining areas is a serious problem. The
concentration of mercury in drinking water and surface water is much higher than the national
standard level (Table 6). The concentration of mercury in food sources, such as rice and
vegetables, is also much higher than the national standard (Table 7).
Similar pollution has also occurred at gold mining areas. The concentration of mercury
in the food produced near refining plants, and fishes in rivers near mining areas, are high.
Epidemiological investigations have shown that concentrations of mercury in urine samples
(Table 8) are higher than the normal value (20ug/L), the highest concentration of mercury in urine
being 50-fold higher than normal. The youngest boy measured was only 2 years old.
262
-------
Table 6. Mercury in drinking water and surface water in mercury mining areas in China
in September 1997. (National standard 0.0001 mg/L).
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Sample
Drinking water
Drinking water
Drinking water
Drinking water
Well water
Well water
Well water
Well water
Well water
Well water
Background
Leakage water from tailing dam
Waste water from refining plant
Waste water from precipitating tank
Waste water from concentrating plant
Sampling date
4
4
4
5
5
5
5
6
6
6
4
6
6
6
6
Cone. (mg/L)
0.0003
0.001 1
0.0002
0.0008
0.0004
0.0002
0.0002
0.0001
0.0003
0.0027
0.0001
0.0002
0.0770
0.0773
0.0004
Table 7. Mercury in rice from mercury mining areas in China in 1997.
Field types
Rice field
Sloping field
Background
Numbers of
samples
30
40
20
Concentration range
(mg/kg)
8.98-228
5.57-58.0
0.222-0.321
Average
78.7
12.9
0.266
Table 8. Mercury in urine of 185 villagers in gold mining areas of China in 1995
(Abnormal is greater than 20 ug/L).
Sample group
Adult men
Adult women
Girls under 1 6
Boys under 16
Total
Number of
persons in
group
131
21
19
14
185
Number of Range of
persons with urine Hg
urine Hg (l-ig/L)
greater than
20fig/L
125
17
14
12
168
3.6-290
4.0-418
4.7-195
8.3 - 540
3.6 - 540
Average
urine Hg
(ug/L)
66.5
64.1
38.1
87.0
65.7
Percentage
of persons
with
abnormal
urine Hg
95
81
71
86
91
263
-------
..., , RISK MANAGEMENT OF MERCURY POLLUTION
In general, the need to do something about the mercury pollution situation in China is
urgent, and decisions and policy are needed. In order to reduce the risk of mercury pollution, an
economic and environmental analysis has been conduced. The procedure for risk management for
mercury pollution is shown in Figure 2. The priority of action for reducing risk has been decided
as follows:
Risk Reduction
Risk Management
Inventory of Industry Mercury
used
Reduced amount
of Mercury used
Predication of
Mercury
Inventory of Mercury
emission from Sources
The Regional Situation of
Mercury pollution
Regional Analysis economic
losses
Standards for controlling and
measuring
Decision of
management
Figure 2. Procedure of risk management for mercury pollution
264
-------
Reduction of Mercury Use
A survey of mercury consumption showed that in an inventory of industries using mercury,
small scale gold mining was first., and small-scale mercury mining also played an important role in
the emission of mercury from anthropogenic processes. The economic effect and production rate
of these two mining industries is very low, but the extent of pollution is serious. As a consequence,
an official order forbidding small-scale gold and mercury mining was announced in 1997 by the
national government. Most small-scale mining was closed, and 80% of the amount of mercury
used in this industry has now been reduced. Also, the mercury cell had been replaced in the
chlor-alkaline industry in the last decade, and the remaining two cell factories will be improved
before the year 2010. In this area the amount of mercury consumption has so far been reduced
by 50%.
Regulation of Mercury Use
The regulation of mercury use by industry is very important, and legislation and
registration of mercury use in industry is made by the government. For example, the degree
of limitation for use of mercury in household batteries has been decided by the National
Environmental Protection Agency, and the content of mercury in batteries will be reduced.
Emission of Mercury
Cooperation is needed to reduce the emission of mercury from the metallurgical industry.
For example, cooperation by the non-ferrous smelter industry is needed to provide the technical
support for removing mercury in exhaust gases.
Study of Mercury Air Pollution and Its Transportation
It will also be necessary to measure air pollution of mercury from the combustion of coal
and incineration of solid waste. An investigation of mercury in coal and transportation of air
mercury will need to be conducted.
SUMMARY
In China, the amount of mercury used in some industrial applications has recently
increased, and the emission of mercury from anthropogenic processes has been enhanced due to
low technological level and poor regulation. The risk of mercury pollution is threatening civilian
health. The inventory of mercury emission has shown the probable anthropogenic sources of this
pollution. These are the metallugical industry, small-scale gold and mercury mining, the chemical
industry, chlor-alkaline and PVC production, and energy production. To reduce the amount of
mercury used in industry, a strategic decision for management of pollution will be necessary.
To eliminate small-scale gold and mercury mining, an official order forbidding these was
265
-------
announced in 1997. The technology of the chloro-alkaline industry has improved since 1990.
The application of mercury in household batteries will be reduced by the year 2010. The
investigation of air mercury pollution from the combustion of coal will be studied and cooperation
in this study is expected.
1 ' , REFERENCES' , ' ' [
Lin, Y.H. 1998. Final Report of Mercury Pollution Protection, Volume 12, pp 2-10.
Xi,X. 1997. Battery. 27:1
266
-------
REMEDIAL STRATEGIES TO REDUCE IMPACTS FROM METAL MINE WASTES
ON A WESTERN UNITED STATES WATERSHED
Dennis R. Neuman1
ABSTRACT
Metal mining in the western United States has had a relatively short history of
approximately 150 years. Ecological damage from this metal mining legacy includes soil and
water contamination caused by tens of thousands of abandoned mine sites. Many of these sites
contain uncontrolled waste materials bearing contaminants of environmental consequence such
as arsenic, copper, cadmium, lead, zinc, and mercury. The U.S. Bureau of Mines estimated that
19,000 km of streams and 7,300 ha of lakes are affected by metal wastes in the United States
(Evangelou and Zhang 1995). In addition to environmental degradation, exposure to these
metals is a threat to human health. Costs of reclamation of abandoned sites is estimated to be
between $33 billion and $72 billion for all minelands in the United States (Lyon et al. 1993),
The U.S. Environmental Protection Agency has designated several Superfund sites in the Butte
and Anaconda area in southwestern Montana. A large landscape has been impacted by over a
century of world class metal mining, milling, and smelting. This paper describes several
remedial actions that are underway or planned for this 10,000 km watershed. Mitigation
strategies are selected on their efficacy to reduce risks to human and environmental health.
Different technologies are being implemented depending on land use and landscape position, and
these include removals, capping, in-situ treatment, subaqueous disposal, channel reconstruction,
waste repositories, and land reclamation.
INTRODUCTION
Metal mining in the western United States began with the great gold rushes in the 1850s
and 1860s and continues today with large operating metal mines in several states. This industry
developed into significant components of the regional, national, and world economies. Impacts
of this historical and current mining have not been without environmental consequences.
Damage from the metal mining legacy includes soils and water contamination caused by tens of
thousands of abandoned mine sites. Many of these sites contain uncontrolled waste materials
bearing environmental contaminants such as copper (Cu), lead (Pb), arsenic (As), cadmium (Cd),
zinc (Zn), and mercury (Hg). The presence of the mineral pyrite, in particular, exacerbates the
situation by producing sulfuric acid, which can mobilize and release these elements into the
surrounding ecosystem. Results include acid drainage, phytotoxic soils, contaminated surface
and groundwater, fish mortality, and, in some extreme cases, mortality to grazing cattle. In
addition to environmental degradation, exposure to these metals is a threat to human health.
'Reclamation Research Unit, Montana State University, Bozeman, Montana, USA.
267
-------
Costs of reclamation of abandoned hardrock mine sites is estimated to be between $33 billion
and $72 billion (Lyon et al. 1993). Some of the largest metal mine sites are Iron Mountain in
California, Bunker Hill in Idaho, Summitville in Colorado, and the Butte/Anaconda area in
Montana.
Environmental Impacts
Sulfide bearing minerals are commonly associated with metal ore deposits. Acid
production occurs when some sulflde minerals, particularly FeS2, react with oxygen and water in
the presence of bacteria such as Thiobacillwferrooxidans. Several complex chemical steps are
involved (Eyangelou and Zhang 1995) which can be summarized as follows:
FeS2 + 15/4 O2+ 7/2 H2O s? Fe(OH)3 + 2H2SO4
The resulting acidic solution mobilizes most metals and they may be released to environmental
receptors including receiving streams, groundwater, and both terrestrial and aquatic organisms.
Mineral weathering is a natural geologic process and in undisturbed settings the oxidation and
release of metals is slow, and insults to ecological receptors are minimal. However,
anthropogenetic activities of mining, milling, and smelting greatly accelerate this process by
exposing vast volumes of wastes over relatively short time periods. This acid generation rate and
associated release rates of the metals can be in great excess to what the receiving environment
Can equilibrate and the effects on the surrounding ecosystems can be significant.
Upper Clark Fork River Basin
Mining for gold began in the Butte, Montana area in the 1860s. By the 1880s, Butte
became the center of a world class copper industry. Underground mines as well as smelters
dominated the Butte landscape and the population swelled to nearly 100,000 people by 1920. In
the mid 1950s open pit mining began. To the west, giant mills and smelters in Anaconda refined
the copper ore. At the height of operations, the Anaconda Smelter produced 11 million kg (25
million pounds) of copper monthly from the mining operations in Butte. The smelter complex
ceased operations in 1983. New mining continues in Butte today, but the ore is shipped to other
facilities for processing. The legacy of this huge industrial complex is waste piles of acid
producing materials in Butte, and a lifeless stream, Silver Bow Creek, which received much of
the industrial waste materials from Butte. In Anaconda, there are large smelter slag piles,
tailings impoundments containing wastes of low pH and high metal concentrations, contaminated
soils, surface and groundwaters. Fluvially deposited tailings along Silver Bow Creek and farther
dpwnstream along the Clark Fork River have contaminated the riparian corridor, and impacted
agricultural lands. The upper Clark Fork River watershed drains 10,000km2 in southwestern
Montana (Figure 1). Silver Bow Creek begins in Butte and extends some 40 km to the Warm
Springs Ponds. The Clark Fork River begins at the outfall from the Ponds and extends
approximately 200 km to Missoula.
268
-------
N
Clark Fgjjc River Site
Anacon
Anaconda Smelter Site
Silver Bow Creek/Butte Area Site
10 0 10 20 Miles
City
Streams
'/\/ Highways
Superfund Sites
Anaconda Smelter Site
Clark Fork River Site
Silver Bow Cr./Butte Area
MONTANA
[Area of Superfund Sites
Figure 1. Clark Fork River basin Superfund sites.
269
-------
Major contaminants found in wastes from the ore processing are As, Cd, Cu, Pb, Hg, and
Zn. Media contaminated are soils, groundwaters, surface waters, terrestrial and aquatic
feceptors. Concentration ranges of these elements in contaminated soils within the drainage
basin of the upper Clark Fork River are provided in Table 1. As a comparison, both regional
background soil concentrations and earth crust concentrations for these elements are also shown
in Table 1. Some mineral processing wastes, such as smelter flue dusts, contain even greater
concentrations of these elements. The acidity of these wastes and contaminated soils ranges
from pH below 2 to near neutral.
Table 1. Elemental concentrations in contaminated soils within the Clark Fork River basin.
Characteristic
Arsenic
Cadmium
Copper
Lead
Mercury
Zinc
Contaminated soils1
(mg/kg)
19 to 3,140
2.6 to 108
260 to 11,200
83 to 6,500
—
19 to 22,000
Regional background2
(mg/kg)
5.6 to 15. 5
0.5 to 1.5
17.2 to 29.1
18.1 to 70.4
—
56 to 78
Mean U.S. soils3
(mg/kg)
6.7
0.73
24
20
0.09
50
1989; 2PTI1996; 3Kabata-Pendias and Pendias 1984.
Comprehensive Environmental Response, Compensation, and Liability Act
In 1980 the U.S. Congress passed the Comprehensive Environmental Response,
Compensation, and Liability Act (CERCLA) which directs the cleanup of uncontrolled
hazardous waste sites in the United States. Some mine wastes are covered under this so-called
"Superfund" law. Of the approximately 1100 designated Superfund sites, 66 hardrock mine sites
were listed on the National Priorities List (NPL) as of August 1996. Hardrock mine sites are
considered among the largest and most expensive Superfund sites to remediate. In 1993, the
US. Environmental Protection Agency (U.S. EPA) estimated that reclamation of the 52 hardrock
niine sites then listed on the NPL would cost a minimum of $15 billion. Several areas in the
upper Clark Fork River basin were placed on the NPL including the metals-impacted areas
designated as Silver Bow Creek/ Butte Area Site, Anaconda Smelter Site and the Milltown
Reservoir Site (Figure 1). Within each of these sites approximately 14 smaller "operable units"
have been designated for efficient management. The U.S. EPA began the process of Remedial
Investigations and Feasibility Studies at these sites in the early 1980s. These activities include
site characterization to define the nature and extent of contamination, human and environmental
risk assessments, and evaluation of potential cleanup alternatives. An administrative vehicle
called a Record of Decision is then prepared that defines remediation of the site. Cleanup
alternatives must be protective of human health and the environment, and they must meet all
relevant legal requirements. Remediation must be effective and permanent, and it must reduce
toxicity, mobility, or volume through treatment. The cleanup also needs to be cost effective and
be implementable. Lastly, the State of Montana and locally affected communities must accept
the remediation alternatives.
270
-------
The CERCLA site's various "operable units" in Montana's Clark Fork River Basin are on
different schedules within the Superfund process. Some sites and "units" have been
characterized, human and ecological risk assessments have been completed, and remedial
alternatives have been selected. Cleanup of these sites has begun. Alternatives chosen have
been quite varied and depend on several factors, including human and/or ecological risk, current
land use, landscape position, proximity to residential homes, the types of contaminated media,
and the metals of concern. In general, the alternatives for wastes and contaminated soils can be
described as soil removals, repositories for treated wastes and some soils, in-situ chemical
treatment to neutralize acidity and immobilize metals, and constructed soil covers. These
alternatives all have land reclamation as a component to the cleanup. Cleanup strategies for
groundwaters have varied from in-situ chemical treatment for arsenic contamination to chemical
precipitation of the metals followed by subaqueous disposal in large ponds. Other groundwater
remedial approaches include natural attenuation and monitoring.
REMEDIAL STRATEGIES FOR ACID METALLIFEROUS MINE SITES
Butte
The city of Butte was built concurrently with the development of many underground
mines and several smelters. Historically, wastes were indiscriminately discharged resulting in
contaminated soils and receiving streams, and numerous acid metalliferous waste piles.
Remedial alternatives for this urban area have concentrated on risk reduction for humans. Butte's
setting is urban and the landscape position is upland. Major contamination problems and
remedial strategies for Butte are summarized in Table 2.
Table 2. Remedial strategies for Butte.
Major contamination problems
Remedial strategies
Acid metalliferous waste piles
Contaminated residential soils
Railroad embankments constructed of mine wastes
Contaminated storm water runoff
Underground mine flooding
Berkley Pit
Removal of waste piles and off-site disposal, followed
by soil cover and revegetation
Removal of residential soils and off-site disposal, followed
by clean soil cap and grass sod
Removal of abandoned railroad embankments
Storm water controls, soil caps, revegetation, detention
ponds, and experimental wetland treatment cells
Treatment at municipal water treatment plant and
Monitoring
Treatment with yet to be selected technology
271
-------
Contamination problems in Butte are acid producing metalliferous waste piles,
contaminated storm water runoff, and residential and recreational soil contamination. Remedial
strategies include consolidation of waste piles and disposal in active mining areas. Remaining
areas are covered with a veneer of lime rock to provide a chemical barrier to the low pH mine
wastes, and then 45 cm of uncontaminated cover soil are spread over the limerock.
These areas are then revegetated using native and adapted drought-tolerant species. Since many
residential areas in Butte are in relatively steep terrain, contaminated storm water runoff is being
channeled over the newly revegetated areas and into culverts for transport off the steep slopes.
Detention ponds have been built into the system to store excess water and to limit erosion. Some
residential soils have been removed because ofelevated levels of lead. Railroad beds that are no
longer in use will eventually be removed. When the underground mines were in production,
large pumps were used to remove water from their workings. These pumps were shut down and
\Vater began to fill the shafts and tunnels. Most of this mine water is currently routed to the vast
open pit known as the Berkley Pit. Water rising in the Berkley Pit will require treatment prior to
discharge to receiving waters. Treatment technologies are currently under development.
Silver Bow Creek
j
Silver Bow Creek flows from Butte approximately 40 km to the Warm Springs Ponds
(Figure 1). Large floods in the early century deposited mine tailings along its banks resulting in
a metals-contaminated riparian zone. The creek bed also contains metals and the surface water at
times exceeds State of Montana water quality standards. Remedial strategies (Table 3) include
removal of saturated tailings and placement in on-site or off-site repositories. These wastes may
be treated with neutralizing amendments if required. Other tailings and phytotoxic soils which
are not in the saturated zone will be treated in place, using neutralizing amendments to control
pH and to immobilize metals. Amended materials will then be seeded and planted with
vegetation species that can establish, grow, and reproduce in the newly altered environment.
Table 3. Remedial strategies for Silver Bow Creek
Major contamination problems
Remedial strategies
Fluvially deposited acid
metalliferous tailings
and phytotoxic soils
Contaminated streambed
Contaminated surface water
Removal of saturated tailings, treatment and on-site and/or off-
site repositories. In-situ treatment of unsaturated tailings with
neutralizing amendments and revegetation
Removal, stream channel reconstruction, imported soils, pool and
riffle habitat construction, bank stabilization, and revegetation
Natural attenuation for surface waters.
These phytostabilization techniques for Silver Bow Creek wastes were developed in a
series laboratory/greenhouse bench studies and field trials (RRU and Schafer & Associates
1993). Some portions of the Creek will be reconstructed to provide aquatic habitat, and these
reconstructed banks will be stabilized with vegetation, mainly willow (Salix spp.) and other
272
-------
riparian species. This work was initiated in 1998 on Silver Bow Creek. First water quality data
show improvement of concentration of As, Cu, and Zn in the surface waters (Table 4). In
addition some wildlife, including muskrat (Ondatra zibethicus) and geese (Branta canadensis),
have moved into the newly reclaimed areas along the Creek.
Table 4. Initial post-remedial water quality for Silver Bow Creek.
Parameter
Montana water Pre-remedial concentrations
quality standards1 mean and (range)2
Post-remedial mean
concentrations3
Arsenic (|J.g/L)
Copper ((J.g/L)
Zinc (ug/L)
18
18 (acute)
12 (chronic)
120 (acute)
110 (chronic)
18
(11 to 39)
252
(130 to 550)
1060
(900 to 1,600)
9
36
415
'Aquatic Life Standards for Surface Waters, Montana Department of Environmental Quality 1995;
/. 1995; 3Pantino 1998
Warm Springs Ponds
The Warm Springs Ponds were initially constructed to treat water entering from upper
Silver Bow and Warm Springs Creeks (Figure 1) prior to discharge downstream. The pond
system was inefficient, and leaks and breeches occurred that severely impacted the downstream
ecology. As part of the remedial cleanup, the earthen dikes surrounding the ponds were
improved and raised. Waters entering the pond system are now treated with calcium oxide (CaO)
to precipitate dissolved metals. Retention time of treated waters is monitored as well as the
quality of the discharge from the system entering the Clark Fork River. Precipitated metals are
essentially immobilized in reducing conditions, and are thus less available to the environment.
This can be considered subaqueous disposal of metals. State and federal agencies conduct
extensive monitoring of the wildlife indigenous to the Ponds.
Anaconda Smelter
The Anaconda Copper Smelter complex operated for nearly a century producing Cu,
AsOs, EkSCvfe Zn, and other metals. Waste products remaining from this milling and smelter
operation include massive amounts of smelter slag and metal enriched wastes. These cover
approximately 2000 ha of contaminated, denuded surface soils which also house tailing ponds
at low pH. Both ground and surface waters have also been impacted. The affected area is
approximately 6,000 ha including urban and rural settings, as well as agricultural lands.
Remedial strategies (U.S. EPA 1998) have been selected to reduce human and ecological risk
from metals and arsenic. Some residential yards within the town of Anaconda, and perhaps
other, smaller communities and rural residential areas within the impacted area, will have soils
273
-------
removed to reduce human exposure to arsenic. Clean fill and sod will replace the contaminated
soils in these residential areas. Some of the most toxic industrial wastes were a beryllium-laden
waste arid arsenic-laden flue dusts. These materials have been treated and encapsulated in
repositories near the smelter. Remaining major problems and remedial approaches are listed in
TableS.
Table 5. Remedial strategies for the Anaconda Smelter site.
Major contamination problems
Remedial strategies
Smelter slag piles
Smelter tailings ponds
Phytotoxic soils
Contaminated ground
and surface waters
No remedial action due to economic value
Combinations of soil covers and in-situ treatment of tailings
wastes followed by revegetation
Land reclamation techniques
Removal of wastes adjacent to impacted surface waters and
revegetation; natural attenuation of ground and surface waters
Smelter slag, which is chiefly a silica material, is currently being used for a variety of
purposes including the base material for roofing products, and as an abrasive to remove paint
from military ships and planes. Since the slag is ofeconomic value, no remedial action was
developed during the Feasibility Study. The massive tailings ponds are scheduled for
combinations of constructed soil covers and/or in-situ treatment with lime and organic matter
followed by seeding with adapted metal- and acid-tolerant plant species. Lands that are
contaminated from aerial emissions are scheduled for reclamation using a variety of in-situ
techniques from light tillage to intensive use of chemical and biological amendments. These
phytostabilization methods were developed specifically for the wastes and contaminated soils in
the vicinity of Anaconda (RRU 1997).
The U.S. EPA (1998) determined that cleanup of the contaminated ground water beneath
the tailings ponds is technically impractical. Monitoring of the quality of the groundwater will
continue and it is expected that as other site cleanups are implemented natural attenuation will
eventually reduce concentrations of metals to acceptable levels.
Clark Fork River
The last major component of the drainage basin is the Clark Fork River itself. Site
characterization was recently completed (ARCO 1998). Large floods deposited metal containing
tailings along the river's banks. In addition, river water was used extensively for flood irrigation
of terraces above the historic fioodplain. As a result, the 100- year floodplain is contaminated
with phytotoxic soils and nearly 7000 ha of agricultural lands contain phytotoxic soils. In
addition, the river's bed sediments are also contaminated with metals. During dry climatic
274
-------
conditions, metal salts form on the surface of some tailings in response to evaporation and
capillary rise. These are very soluble and can be washed into the river during time of intense
thunderstorms. In 1989, such an effect took place and was well documented. The pH of the
river water dropped and metal concentration rose several orders of magnitude in a short period of
time. At least 5,000 fishes, many of them trout, were killed, and concentrations of metals in their
internal organs were extremely elevated. Land reclamation efforts to eliminate such ecological
events were reported by Munshower et al. in 1997.
Residential areas adjacent to irrigated agricultural fields have elevated soil concentrations
of As, Cu, and Zn. Some residential homes have surface soil arsenic concentrations above what
the U.S. EPA recognized as a threat to human health. These soils are scheduled for removal, and
clean soil will be spread followed by sod or seeding of the yards.
Strategies for the Clark Fork River are currently being studied in the Feasibility Study. It
seems likely that removal of the most contaminated tailings areas will be conducted. Other areas
with fluvially deposited tailings may be treated with amendments, tilled and seeded with plant
species that would be compatible with post-land uses. Bank stabilization may be conducted
using combinations of engineering designs as well as vegetative methods. Agricultural lands
may also be treated with amendments to control acid production and with organic matter to
improve the nutrient status. Selected plants compatible with land use practices will then be
seeded. The river's contaminated sediment is a serious problem with no clearly identified
solution. Massive removal of the sediment or building new channels seems unlikely because
these techniques may compromise the geomorphic stability of the current river.
CONCLUSIONS
There are tens of thousands of abandoned hardrock mines and mills in the western United
States. Uncontrolled releases of acid drainage from sulfide weathering at these sites have
impacted both aquatic and terrestrial habitats in many watersheds. The U.S. Bureau of Mines
estimates that 19,000 km of streams and some 7,300 ha of lakes are impacted by metal mine
wastes in the United States (Evangelou and Zhang 1995). The largest sites, and those which
represent risks to humans, have been placed on the U.S. EPA's "Superfund" list. Remedial
strategies being developed and implemented at sites within the Clark Fork River Basin have
direct applicability to other large and small impacted sites in the western United States.
REFERENCES
ARCO (Atlantic Richfield Company). 1998. Final Draft Remedial Investigation Report, Clark
Fork River Operable Unit, Milltown Reservoir Sediments NPL Site. Atlantic Richfield
Company, Anaconda, Montana.
Evangelou, V.P., and Y.L. Zhang. 1995. A review: pyrite oxidation mechanisms and acid mine
drainage prevention. Critical Reviews in Environmental Sciences and Technology 25:
141-199.
275
-------
Kabata-Pendias, A., and H.Pendias. 1984. Trace Elements in Soils and Plants. CRC Press,
Boca Raton, Florida. 315 pp.
Lambing, J.H., M.I Hornberger, E.V.Axtmann, and D.A. Pope. 1994. Water quality, bed-
sediment, and biological data (October 1992 through September 1993) and statistical
summaries of water quality data (March 1985 through September 1993) for streams in the
Upper Clark Fork River Basin, Montana. U.S. Geological Survey. Open File Report 94-
375, Denver, Colorado.
Lyon, J.S., T.J. Hillard, and T.N. Bethell (Eds.) 1993. Burden of Guilt. Mineral Policy Center,
Washington, D.C. 68pp.
MDEQ (Montana Department of Environmental Quality). 1995. Montana Numeric Water
Quality Standards, Circular WQB-7. Water Quality Division, Helena, Montana.
Munshower, RF., D.R. Neuman, S.R. Jennings, and G.R. Phillips. 1997. Effects of land
reclamation techniques on runoff water quality from the Clark Fork River floodplain,
Montana. In: Proceedings of the Fourth International Symposium on Fish Physiology,
Toxicology and Water Quality, Bozeman, Montana, USA. September 1995. R.V.
Thurston (Ed). U.S. Environmental Protection Agency, Ecosystems Research Division,
Athens, Georgia, USA. EPA/600/R-97/098. pp. 199-207.
Pantino, J. 1998. Surface water monitoring at remediated Lower Area One of Silver Bow
Creek. Data courtesy of Atlantic Richfield Company, Anaconda, Montana.
PTI Environmental Services. 1996. Final Remedial Investigation Report for Anaconda Regional
Water and Waste Operable Unit. Prepared for Atlantic Richfield Company, Anaconda,
Montana.
RRU (Reclamation Research Unit), Schafer & Associates, and CH2M Hill. 1989. Final
Summary Report for Streambank Tailings and Revegetation Studies (STARS). Volumes
I and It Bench-scale soil column and greenhouse treatability studies and tailings ranking
system. Montana State University, Bozeman, Montana.
RRU (Reclamation Research Unit), and Schafer & Associates 1993. Final Report for
Streambank Tailings and Revegetation Studies (STARS). Volumes I through IV. Field
Monitoring and Evaluation. Montana State University, Bozeman, Montana.
RRU (Reclamation Research Unit). 1997. Anaconda Revegetation Treatability Studies (ARTS),
Phase IV: Monitoring and Evaluation. Volumes I and II. Montana State University,
Bozeman, Montana.
U.S. EPA (U.S. Environmental Protection Agency). 1998. Record of Decision, Anaconda
Regional Water, Waste, and Soils Operable Unit. Anaconda Smelter National Priorities
List Site, Anaconda, Montana. U.S. EPA, Region VIII, Montana Office, Helena,
Montana.
276
-------
WATERSHED MODELING - WHERE ARE WE HEADING?
Yongqin David Chen1
ABSTRACT
In this paper a review of the history of watershed hydrologic and water quality modeling
is presented, leading to analysis and discussion of a number of critical issues often concerned
and debated by modelers. The purpose of this discussion is twofold: (1) to analyze the role of
the Geographic Information System (GIS), remote sensing, and graphical user interface (GUI)
- the "bells and whistles" for modeling - in the development of a new generation of watershed
models, and (2) to address the importance of algorithm and computation - the underlying model
"engine" - in watershed modeling, thus proposing a research agenda to outline the most
important areas for model improvement now and in the foreseeable future.
INTRODUCTION
In the past several decades, with the advent and rapid progress of computer technologies,
numerical simulation models have increasingly become important and effective tools for
tackling a wide range of environmental and resources management issues. Among these many
types of models, watershed hydrologic models simulate the dynamic behavior of significant flow
and storage processes, generating water balance information (quantity and associated hydraulic
characteristics, source and pathway, residence time, etc.). Historically, most early hydrologic
models were designed for estimating water quantities in engineering applications such as flood
forecasting, urban storm water management, and many other water resources planning activities
such as reservoir design and water supply.
More recently, water quality components have been developed and incorporated into
some water balance models since the late 1970s as the importance of non-point source pollution
was gradually recognized. As the water quality objectives have been expanded from human
health protection to include ecosystem health in recent years, research efforts have been made to
introduce new state variables of habitat and ecological characteristics into watershed models for
conducting multi-stressor analysis of whole ecosystem effects (Chen et al. 1993).
'Department of Geography, The Chinese University of Hong Kong, Shatin, New Territories, Hong Kong
277
-------
"' i)« i ;t I' SI i" • ''IIIIH
r t. "i!!!!"1!:1: T'1;
Watershed hydrologic models as a base, together with their applications in various areas
as shown in Figure 1, have made tremendous contributions to numerous aspects of water
resources management. As the era of watershed planning and management of the 1990s
continues into the 21st century, the demands for more sophisticated and practical watershed
models will certainly continue to increase (e.g., see white papers on the needs for hydrologic
modeling by various federal agencies in Burton 1993).
Watershed Hydrologic Model
Figure 1. Watershed model and applications.
The "Information Age" has significantly influenced the way we conduct scientific
research and engineering practices in the modern society, including our capabilities of
environmental simulation and forecasting. For the development and application of watershed
models, two technological advancements have made fundamental and extensive contributions in
recent years. One is the computer technology which has been able to meet essentially all of the
modeling demands for computing power and data storage capacities in these days. In fact, the
upgrading speed of both personal computers and workstations has already gone beyond the needs
and expectations of watershed modelers for many years. The other is the geoinformation
technology mainly consisted of the Geographic Information System (GIS) and remote sensing
that have made numerous spatial data available, and the processing and analysis of these data
possible for watershed modeling. Actually, since the early 1990s, most modeling activities have
focused on the use of computer and geoinformation technologies to develop Windows -based
graphical user interface (GUI) and integration of watershed models with GIS and remote sensing
data. In the meantime, it seems that efforts for the development and enhancement of model
algorithms arid computational techniques have been comparatively weakened. The objectives of
this article is first to review the history of watershed modeling, and then based on this review to
examine the role of computer interfaces and GIS techniques, as well as the importance of
algorithms and computation in the current practice and future of watershed modeling. The
historical review and analysis of current modeling activities will help watershed modelers better
understand where we have come from and where we are heading as the 21st Century arrives.
278
-------
HISTORY OF WATERSHED MODELING: A BRIEF SUMMARY
The Stanford Watershed Model (SWM), first created in the mid-1960s, has been
commonly considered as the first watershed hydrologic model (Crawford and Linsley 1966).
Over the past few decades, countless (at least hundreds of) watershed hydrologic and water
quality models have been developed around the world. Numbers of reports, book chapters, and
articles that offer comprehensive reviews of these models have periodically appeared in the past
two decades. Some authors focused only on hydrologic models while others were concerned
about both runoff and water quality simulations. In chronological order, reviews of hydrologic
models include: (i) Fleming (1975) who devoted a good portion of his pioneering textbook to
summarize the structures, similarities and differences, and applications of 19 hydrologic models
developed in the preceding 10 years; (ii) Renard et al. (1982) who conducted an extensive survey
of hydrologic models developed in the United States through computer and manual literature
searches and collection of questionnaires, resulting in a summary of 75 watershed models; and
(iii) Troendle (1985) who reviewed the development of variable source models that can provide
better interpretations of the source, pathway, and residence time, solutes fate, and energy
„ disposition within a forested watershed.
Reviews of watershed models with hydrologic and water quality components include:
(i) Donigian (1981) who discussed the watershed hydrologic processes affecting water quality,
and offered an overview on the computer models developed in the 1970s; (ii) Beasley and
Thomas (1989) who published a report on the selection, evaluation, modification, and
application of five hydrologic and water quality models to simulate the effects of agricultural
and silvicultural practices on the quantity, quality, and utilization of surface and subsurface
waters; (iii) El-Kadi (1989) who reviewed 29 watershed models and discussed their applicability
to conjunctive use management; (iv) the U.S. Environmental Protection Agency (U.S. EPA)
who developed two major documents that provide practical guidance for hydrologists and
watershed managers to select and apply appropriate models for specific management issues and
modeling objectives (Donigian and Huber 1991, U.S. EPA 1992); (v) DeVries and Hromadka
(1993) who described a number of commonly used watershed models and also provided the
reader with some model selection guidance; and (vi) the book edited by Singh (1995) which is
so far the most comprehensive and detailed compilation of representative watershed hydrologic
models developed in many countries around the world. Although personal bias was unavoidable
as admitted by the editor, Singh (1995) did a good job in selecting the 26 popular models and
inviting the original author(s) of each model to contribute a book chapter which describes their
models in details. Basic information of the 26 models (see Table 1) should provide the readers a
general idea about the availability and history of models developed in various countries over the
past few decades.
279
-------
Table 1. Twenty-six representative watershed models included in Singh (1995)
Country
USA
(14)
Canada
(2)
UK
(3)
Denmark
(2)
Sweden
0)
Australia
(2)
China (1)
Japan (1)
Agency/
Institute
USEPA, USGS
USEPA
USGS
USNWS
USACOE
USACOE
USDA-ARS
USDA-ARS
USDA-ARS
USDA-ARS
USDA-ARS
USDA-ARS
USDA-ARS
University of British
Columbia
National Hydrology
Research Institute
University of Newcastle
Upon Tyne
Institute of Hydrology
University of Lancaster
Danish Hydraulic Institute
Danish Hydraulic Institute
Swedish Meteorological
&
Hydrological Institute
Monash University
University of Melbourne
Hohai University
National Research Center
for Disaster Prevention
Model Name
Hydrologic Simulation Program-Fortran
(HSPF1
Storm Water Management Model (SWMM)
Precipitation-Runoff Modeling System
(PRMS)
National Weather Service
River Forecast System (NWSRFS)
HEC-1 Flood Hydrolograph Package
Streamflow Synthesis and
Reservoir Regulation (SSARR)
Snowmelt Runoff Model (SRM)
Kinematic Runoff and Erosion Model
(KINEROS, called KINGEN first developed in
1977)
Simulation for Water Resources
in Rural Basins (SWRRB)
Erosion Productivity Impact Calculator
(EPIO
Agricultural Non-Point Source Model
(AGNPS)
Simulation of Production and
Utilization of Rangeland (SPUR)
Chemicals, Runoff and Erosion from
Agricultural Management System (CREAMS)
Groundwater Loading Effects of
Agricultural Management System (GLEAMS)
UBC Watershed Model
Simple Lumped Reservoir Parametric
(SLURP)
SHE/SHESED
Institute of Hydrology Distributed Model
(WDM)
TOPMODEL
A Generalized River Modeling Package
(MIKE 11)
MIKE SHE
HBV
RORB
THALES/TAPES-C
The Xinanjiang Model
The Tank Model
First
Developed
1980
1971
1983
1976
1968
1956
1975
1990
1990
1984
1987
1987
1980
1963
1978
1986
1987
1979
1989
1986
1975
1975
1992
1973
1974
Current
Update
1996
1995
"*
""
1991
1990
1992
On-
going
1995
1990
1996
1991
1987
1993
"
'
"
1994
"
-
1992
-
-
280
-------
To summarize the above reviews, the history of watershed modeling can roughly be
divided into three stages. First, the advent of mainframe computers (even though slow and
clumsy) allowed hydrologists, for the first time, to implement hydraulic computations and
conceptual water balance algorithms on a digital platform during the mid and late 1960s.
This first generation of models mainly includes the most famous SWM, HEC-1 and SWMM.
These classical and long-lasting models laid down the theoretical and technical basis for
constructing conceptual hydrologic models which have become the major tools for water
resources planning and management. They have also been expanded into more advanced
systems or significantly improved. For example, SWM was used as the base for the development
and later numerous enhancements of the first comprehensive watershed model called HSPF. For
nearly three decades, the U.S. EPA, joined in recent years by the U.S. Geological Survey, has
provided enormous funding for the development and technical support of HSPF, leading to the
establishment of a large and powerful modeling system which is supported by data manager
ANNIE, meteorologic data analyzer METCMP, model calibration expert system HSPEXP, and
statistical analysis program SWSTAT. Another example is the HEC model series which has
been the focus of modeling activities for the past three decades at the Hydrologic Engineering
Center of the U.S. Corps of Engineers located in Davis, California.
In the second stage, as modeling techniques became more sophisticated and with the
spread and rapid advancement of personal computers, numerous watershed modeling systems
were developed throughout the 1980s around the world. Renowned examples, just to name a
few, include HSPF, CREAMS and GLEAMS, AGNPS, and ANSWERS in the United States,
and SHE and TOPMODEL in Europe. The emphasis of the second stage was development and
implementation of simulation algorithms, which produced a large of computer codes that still
constitute the core of today's modeling technology.
The third stage, beginning in the early 1990s, has been signified by the increasing
emphasis on the development of computer interfaces and application of GIS techniques. And in
fact, GUI development and GIS integration have dominated in most modeling activities (Huber
1995). However, the pros and cons of this trend, as well as the relationship between the internal
algorithms and external features of models, have not yet been fully addressed. The rapid
proliferation of watershed models and numerous application studies have attributed to many
theoretical and technological advances, mainly including those given below.
(1) The advent and development of watershed modeling as we know it, would never have
been possible without computers. The spread of personal computers and the availability of
workstations and supercomputers, as well as the ever escalating computing power and data
storage capacity of all machines, have dramatically reduced the amount of effort and computer
time for designing and operating complex models. Nowadays the use of computers has entered
or even dominated every aspect of hydrologic research and practice, while only a small number
of (research) hydrologists started touching the clumsy mainframe computers three or four
281
-------
decades ago. One can easily see this fundamental change by simply comparing the contents
related to hydrologic computations and model simulations in the two versions of the Handbook
of Hydrology, i.e., Chow (1964) and Maidment (1993a). As noted above, the recent
development is that the growth of GIS technology and the expanding availability of digital data
for various landscape features have significantly enhanced the capabilities of many watershed
rttodels to use, process, and display spatial information (e.g., see overview by Maidment 1993b).
(2) The theoretical basis for evaluating basin responses to storm events has been altered
and enhanced" by progress in the understanding of runoff generation mechanisms in both urban
and wildland watersheds. One of the most momentous innovations in modern hydrology is the
evolution from Horton and Sherman's over-simplified infiltration excess theory (Linsley et at.
1982) that was widely accepted since the 1930s to Hewlett's variable source area concept
(Hewlett 1969). Hewlett's concept was originally derived from the experimental watershed
studies at Coweeta Hydrologic Laboratory, Franklin, North Carolina, USA in the 1960s and later
verified and enhanced by many other field and modeling studies mainly in the United States and
Europe (e.g., see the most recent review by Bonell 1993; the benchmark text edited by Kirby
1978; and the updated assembly of numerous studies in hillslope hydrology edited by Anderson
and Burt 1990).
(3) The modernization of hydrometeorologic and hydrogeochemical measurement
techniques has also significantly contributed to the development and application of hydrologic
models. Increased quantity and quality of on-site measurement data for model calibration and
testing have resulted from modern technologies such as the automation of data acquisition and
transmission, and the use of chemical tracers. More importantly, as distributed parameter
models have become increasingly popular in recent years, remote sensing and image processing
technologies have made it possible to generate a tremendous amount of inexpensive information
quickly with High spatial and temporal resolution. In the United States, for example, the Next
Generation Weather Radar System (NEXRAD), which consisted of more than 120 radar stations
fully set up by 1996, can provide rainfall estimates for time intervals as small as 5 minutes and
spatial resolution as small as 1 km2. Critical watershed characteristics for spatial disaggregation
and model parameterization such as updated and detailed land use and land cover data are now
available from satellite imageries (e.g., see Engman and Gurney 1991).
(4) A number of leading scientists continue to advance environmental modeling as a
discipline through philosophical thought and debates on (i) modeling objectives and
assumptions, (ii) model design criteria and methods, (iii) scale issues focusing on the
representation of spatial complexities of watershed features and climatic inputs, (iv) model
calibration and validation, and (v) sensitivity and uncertainty analysis in the development and
application of numerous "physically-based" models. See representative essays by Beck (1983,
1988), Beven (1989), Bloschl and Sivapalan (1995), Jakeman and Hornberger (1993), and
Wheater e/a/. (1993).
282
-------
COMPUTER AND GEOINFORMATION TECHNOLOGIES
FOR WATERSHED MODELING
A number of reasons have caused the dominance of user interfaces in watershed
modeling activities in recent years. Firstly, rapid advances in computer technology (hardware
and software) have offered tremendous opportunities and potential for modelers to upgrade the
traditional DOS-based models to be associated with various user-friendly Windows®-based
interfaces. This fundamental change has made model operation much easier and efficient.
Secondly, development of GIS and proliferation of spatial data have opened a new field for
hydrologists to derive, to process, and to use spatial characteristics of watersheds in
unprecedentedly great details for modeling exercises and to display and analyze simulation
outputs in impressive and dynamic graphical forms. As model input providers, GIS and remote
sensing technologies have not only dramatically increased the efficiency and data accuracy of
watershed characterization, but also enabled modelers to generate and use much new
information (e.g., spatially continuous rainfall and soil moisture data from radar, digital
elevation model data, etc.) which was not previously available for modeling. Thirdly, as shown
in Table 1, most watershed models were developed in the 1970s and 1980s before the current
generation of computers and geoinformation technologies emerged and rapidly advanced since
the early 1990s. This is obviously a major reason that the needs and market for GUI-oriented
and GIS-based work in the watershed modeling field have emerged and increasingly expanded in
recent years. Lastly, although the user friendliness offered by GUIs and the representation of
spatial features provided by GIS have certainly brought in tremendous benefits, the fact that
watershed modeling has been driven, often exclusively, by computer and GIS technologies
should be cautiously and critically examined.
There are three types of interfaces for watershed modeling. (1) Windows®-based
environment as model input developer, data editor, and output display was the focus of
watershed model software engineering in the early 1990s when a number of models were
upgraded from DOS to Windows® environment. Examples include SWMM, HEC-1, AGNPS,
and SWRRB. The only contribution of these efforts has been a significant increase in the user
friendliness and data management efficiency of the models. (2) Loose linkage to GIS has been
the eye-catching part of numerous modeling research and application projects that have been
widely reported in a voluminous literature. Basically, derivation and processing of spatial
features by GIS are only loosely linked with model operation in these studies. In other words,
the use of GIS to characterize and represent watersheds is pretty much an independent
component of a modeling project. (3) Integration with GIS for developing modeling systems
that can better represent and analyze the spatial characteristics of simulated basins and
hydrologic processes has been a new trend for the last few years. The Center for Research in
Water Resources (CRWR) led by D. R. Maidment at the University of Texas at Austin, together
with its partners in industry, government, and university, has been making many pioneering
efforts in this field. This group of researchers strongly believe that the synthesis of GIS and
283
-------
hydrology will be of tremendous benefits to the simulation of water flow and its constituents
over the land surface and in the subsurface environment, and achieving this synthesis requires
nuinerous efforts to expand GIS capabilities such as introducing explicit time variation in its
data structures and to design new hydrologic modeling frameworks such as distributed models
(Maidment 1993b). The CRWR group has developed and periodically upgraded a GIS
Hydrology program (a set of computer models, databases, atlases) for research and teaching
purposes (see http.7/www.ce. utexas.edu/prof/maidment/ for details and abundant information of
the most recent development in spatial hydrology). Other exciting developments by CRWR
partners include the Arc View-based BASINS by the U.S. EPA Office of Water, Watershed
Modeling System (WMS) and Groundwater Modeling System (GMS) by the Engineering
Computer Graphics Laboratory of Brigham Young University, banish Hydraulic Institute's
models (so far including Mile-11, Mike Basin, Mike-SHE, and Mouse) integrated with Arc View,
and a hydrologic module built in Arc View of Environmental Systems Research Institute (ESRI)
in Redlands, California. Most recently, Research Systems, Inc., a clata analysis and visualization
software company based in Boulder, Colorado, has developed a full-featured topographic and
hydrologic analysis package called RiverTools (see http://www.rsinc.com/rivertools/index.cfm
for details).
Hydrologists and modelers believe that models are much more than just visually
attractive computer interfaces. Given the prevalence of GUI-oriented and GIS-based modeling
activities in recent years, then, questions such as "what are the contributions of interfaces to
watershed modeling?" and "what is the relationship between the development of interfaces and
the enhancement of underlying model codes?" must be addressed. The answers to the first
question include: (1) revolution in digital representation of watershed characteristics and
meteorological inputs both in terms of accuracy and resolution; (2) automated processes for
watershed characterization, delineation, and segmentation that were impossible tasks before;
(3) great improvements in data access, storage, transformation, derivation, exploration, and
visualization; (4) unprecedented user-friendliness in model operation and calibration, analyses of
model performance, sensitivity and uncertainty, scenario generation and evaluation and
interpretation of model output for decision support; and (5) rapid development in integrated
modeling systems with many new features and spatial analysis capabilities, as well as physically-
based and distributed-parameter models. Based on this analysis, the importance of GIS and
interfaces in watershed modeling must not be under-estimated while a parallel requirement
- improvements in the underlying model "engine" as called by Huber (1995) - has always
existed but may have been neglected to some extent in recent years. To summarize in an
allegorical way, a watershed model as compared to an automobile consists of the most important
core — the underlying algorithm and computer code as the "engine" - and other parts among
which interfaces can substantially improve the model performance and reliability. Therefore,
interfaces are more than just "bells and whistles", and work like automatic transmission, power
lock and power steering that make automobiles much safer, comfortable, and easier to drive.
284
-------
ALGORITHMS AND COMPUTATION - THE MODEL "ENGINE"
There is no doubt that improvement of the most important underlying model "engine" has
been and will continue to be the key to producing more and more reliable tools for managing our
environment and water resources. While enormous progress has been made in developing and
refining interfaces, greater efforts are now needed to focus on model formulation - the
conceptualization of hydrologic and water quality processes, algorithms and computational
techniques, including both new developments and enhancements of existing codes. In fact, the
two types of tasks are complementary and simultaneous advances in both areas will certainly be
of utmost importance for the future of watershed modeling. As discussed previously, computer
and geoinformation technologies have made fundamental contributions to model interfaces and
GIS integration. These technologies, as well as their resulting software systems and spatial
analysis tools, have also fostered numerous opportunities for the incorporation of new research
findings into better models. On the other hand, developments in the model "engine", e.g., from
spatially-lumped to distributed-parameter formulations and from simple linear to complicated
non-linear systems, will continue to generate demands for more powerful computing and spatial
analysis tools.
If we all agree that greater emphasis should be placed on the underlying core of process
algorithms and computational techniques in the future, identification and analysis of important
areas for model improvement are surely needed for guiding us in the right direction. A few
experts have given some thoughts and written on this topic (e.g., Donigian et al. 1995; Huber
1995). The state-of-the-art of watershed modeling depends upon many factors such as
management concerns, environmental assessment techniques and criteria, scientific knowledge
of hydrologic and water quality processes, and availability of technology and data. Since all of
these factors are dynamic by their nature, the future of watershed modeling will continue to face
many new challenges and opportunities. With direct support of new developments in computer
and GIS technologies or in advances to be made independently, major areas with needs for
research and modeling creativity include:
(1) Simulation of hydrologic and water quality effects of human activities will continue
to be the focus of model development and enhancement in the future. In some cases,
representation of these effects can be made by adjusting parameter values, while development of
new algorithms or refinement of existing formulation may be needed under other circumstances.
In order to simulate the environmental benefits of best management practices such as
constructed wetlands, riparian buffers, and porous pavements, etc., with greater confidence, we
will need to decide on the technical approach and conduct research for making the models more
accurate and robust. Due to the complexity of the issue and the lack of field data, a vital task of
enhancing the physical basis in conceptual/empirical hydrologic and water quality algorithms
will continue to be difficult. As illustrated by Huber (1995), a good example is the attempt to
replace the existing build-up/wash-off concept in urban non-point source water quality models
with physically-based algorithms for describing the dependence on shear stress and rainfall
energy.
285
-------
(2) The need of expanding model capabilities for additional state variables has arisen as
a result of the augmentation of water quality objectives from human health protection to include
ecosystem protection. In order to use models for making multi-stressor analysis of whole-
ecosystem effects, efforts are being made to design and introduce biological, habitat, and
ecosystem variables in watershed models (Donigian et al. 1995).
(3) Remote sensing has demonstrated tremendous potential for providing unprecedented
hydrologic and water quality data for modeling studies. GIS is a powerful and versatile tool for
processing remote sensing data and building a linkage between models and data storage. The
emergence and proliferation of digital spatial data have required modelers to develop new
algorithms for making use of these data. Recently, there are many examples of this type of
model development efforts reported in the literature. Just to mention a few, "hot topics" for
research include algorithms for processing high resolution (spatial and temporal) precipitation
data from weather radar (e.g., NEXRAD in the United States) for rainfall-runoff modeling,
DEM-derived aspect and canyon topographic shading characteristics of mountainous land
surface for snowmelt and stream temperature simulation (e.g., see Chen et al. 1998a, 1998b).
(4) Similar to other fields in environmental modeling, development or enhancement of
computational techniques for taking advantage of more and more powerful computing resources
has been an important area for research. Generally, a major improvement is that faster machines
with more memory allow reduced time step while numerical stability can be maintained (Huber
1995). Nowadays computing power is normally no longer a constraint for the implementation of
numerical solutions of differential and other equations in the context of watershed modeling.
SUMMARY AND CONCLUSIONS
Over the past several decades, computers and information technologies have brought
tremendous advances in every aspect of hydrologic research and engineering practices. The
history of watershed hydrologic and water quality modeling can be roughly divided into three
stages to reflect the development of science and technology. The most recent stage has been
signified by the emphasis on Windows®-based user interface and GIS linkage or integration in
modeling activities. This focus was directly caused by the revolutionary advances in computer
technology, together with the advent and rapid progress in geoinformation technology, and more
recently has been justified by the software products and research outputs of enormous efforts in
the past few years. Interfaces and GIS have made many important contributions to watershed
modeling, and they play important roles. However, there is much more to watershed modeling
than just an attractive interface and a seamless integration with GIS. Improvement and
sophistication of the underlying model "engine" should be the utmost important field for
research and engineering invention now and in the years to come. Specific areas with needs for
this work have been identified and addressed as an attempt to propose a research agenda for
developing better models.
286
iiti
• .:,:.„; i
-------
REFERENCES
Anderson, M.G., and T.P. Burt (Eds.)- 1990. Process Studies in Hillslope Hydrology.
John Wiley & Sons, Chichester, England, 539 pp.
Beasley, D.B., and D.L. Thomas. 1989. Application of Water Quality Models for Agricultural
and Forested Watersheds. Southern Cooperative Series Bulletin No. 338, Coastal Plain
Experiment Station, University of Georgia, Tifton, Georgia, 116 pp.
Beck, M.B. 1983. A procedure for modeling. In: Modeling of Water Quality: Streams, Lakes,
and Reservoirs. John Wiley and Sons, Chichester, England, pp 11-41.
Beck,M.B. 1988. Water quality modeling: a review of the analysis of uncertainty. Water
Resources Research 23:1393-1442.
Beven,K. 1989. Changing ideas in hydrology-the case of physically-based models. Journal of
'Hydrology 105:157-172.
Bloschl, G., and M. Sivapalan. 1995. Scale issues in hydrological modelling: a review.
Hydrologic Processes 9:251 -290.
Bonell, M.. 1993. Progress in the understanding of runoff generation dynamics in forests.
'journal of Hydrology 150:217-275.
Burton, J.S. (Ed.). 1993. Proceedings of the Federal Interagency Workshop on Hydrologic
'Modelling demands for the 90s. Water Resources Investigation Report 93-4018. U.S.
Geological Survey, Reston, Virginia, USA.
Chen, Y.D., R.F. Carsel, S.C. McCutcheon, and W.L. Nutter. 1998a. Stream temperature
simulation of forested riparian areas: 1. watershed-scale model development. Journal of
Environmental Engineering 124:304-315.
Chen, Y.D., S.C. McCutcheon, D.J. Norton, and W.L. Nutter. 1998b. Stream temperature
simulation of forested riparian areas: 2. model application. Journal of Environmental
Engineering 124:316-328.
Chen, Y.D., S.C. McCutcheon, T.C. Rasmussen, W.L. Nutter, and R.F. Carsel. 1993. Integrating
water quality modeling with ecological risk assessment for nonpoint source pollution
control: a conceptual framework. Water Science and Technology 28:431-440.
Chow, V.T., (Ed.). 1964. Handbook of Applied Hydrology. McGraw-Hill, New York, NY USA.
287
-------
Crawford, N.H., and R.K. Linsley. 1966. Digital Simulation in Hydrology: Stanford Watershed
Mo
-------
Maidment, D.R. (Ed.). 1993a. Handbook of Hydrology. McGraw-Hill, New York.
Maidment, D.R.. 1993b. GIS and hydrologic modeling. In: M.F. Goodchild, B.O. Parks, and
L.T. Steyaert (Eds). Environmental Modeling with GIS. Oxford University Press, New
York, pp. 147-167.
Renard, K.G., W.J. Rawls, and M.M. Fogel. 1982. Currently available models. In: Hydrologic
Modeling of Small Watersheds. C.T. Haan, H.P. Johnson, and D.L. Brakensick (Eds.).
Monograph No. 5, American Society of Agricultural Engineering, St. Louis, Missouri, pp
507-522.
Singh, V.P., (Ed.). 1995. Computer Models of Watershed Hydrology. Water Resources
Publications, Highlands Ranch, Colorado, USA. 1130 pp.
Troendle, C.A.. 1985. Variable source area models. In: M.G. Anderson,and T.P. Hurt (Eds.).
Hydrological Forecasting. John Wiley & Sons, Chichester, England, pp. 347-403.
U.S. EPA (U.S. Environmental Protection Agency). 1992. Compendium of Watershed-Scale
Models for TMDL Development (EPA/841 -R-92-002). Office of Water, U. S.
Environmental Protection Agency, Washington, DC, USA.
Wheater, H.S., A.J. Jakeman, and KJ. Seven. 1993. Progress and directions in rainfall-runoff
modelling. In: Jakeman, A.J., M.B. Beck, and M.J. McAleer (Eds.). Modelling Change
in Environmental Systems. John Wiley & Sons, Chichester, England, pp. 101-132.
289
-------
-------
STRATEGY FOR DEVELOPING NUTRIENT AND SEDIMENT LOAD
ALLOCATIONS FOR WATER QUALITY PROTECTION
Rosemarie C. Russo1 and Robert F. Carousel1
ABSTRACT
The issue of non-point source pollution is receiving increasing attention within the
United States, and U.S. Environmental Protection Agency researchers in Athens, Georgia, are
undertaking an environmental regulatory assessment support research effort directed toward total
maximum daily load (TMDL) determinations. The TMDL program's strategy is intended to
facilitate the research and development necessary for providing fully validated operational
requirements, proven operational concepts, and transition of mature technologies to support
successful development and production of TMDL support tools. These tools will be used for
regulatory needs in the short term and ultimately incorporated into an open architecture, object-
oriented, modeling framework over the long term. The research strategy to accomplish these
goals is described and discussed.
INTRODUCTION
Many rivers, and estuaries into which they flow, have environmental problems due to
an over-abundance of nutrients (nitrogen, phosphorus) from both point sources (municipal
sewage, animal feedlots, etc.) and non-point sources (agricultural runoff, land use changes, etc.).
Of particular concern are oxygen-demanding wastes and nutrients, and the resulting impacts on
dissolved oxygen and trophic status-habitat in river and estuary systems undergoing rapid
growth, industrial redevelopment and expansion, and agricultural decentralization and
modernization. Nutrient over-enrichment from human activities is one of the major stresses
impacting coastal ecosystems, causing reduced photic zone depth, loss of habitat, decrease in
dissolved oxygen and impacts on living resources. Excess nutrients lead to excessive growth of
macrophytes or phytoplankton, increased algal production, and increased availability of organic
carbon. This algal over-production may sink to the bottom and decay, consuming a significant
amount (hypoxia) or all (anoxia) of the available oxygen in these bottom waters.
National Exposure Research Laboratory, US Environmental Protection Agency, Athens, Georgia 30605, USA
291
-------
Eutrophication is an important problem for coastal ecosystems such as the Gulf of
Mexico (U.S. EPA and USGS 1999, NOAA 1998); the Gulf region, of all regions in the United
States, has the highest number of wastewater treatment plants and the most land use devoted to
agriculture. Similarly, eutrophication is a problem in the Baltic Sea (Lochmuller and Malin
1998).
In order to be able to develop environmentally protective and economically feasible water
quality management activities for such water bodies, it is important to be able to model the
streams to assess waste load allocations and the efficacy of various improvement options. In the
United States, The Clean Water Act requires states to develop and implement Total Maximum
Daily Loads (TMDLs) to attain water quality standards. TMDLs for both sediments and
nutrients are important environmental issues. Addressing these issues requires updated models
with state-of-the-art science, and models must be field-tested and verified. User training and
assistance are also required.
State water quality reports indicate that over-enrichment of waters by nutrients (nitrogen
and phosphorus) is the biggest overall source of impairment of the nation's rivers and streams,
lakes and reservoirs, and estuaries (U.S. EPA 1998a). States have reported that 40 percent of
surveyed rivers, 51 percent of surveyed lakes, and 57 percent of surveyed estuaries were
impaired by nutrient enrichment. Agriculture is the most widespread source of these
impairments, followed by municipal sewage treatment plants, urban runoff, and storm sewers.
Table 1 summarizes the leading causes of water quality problems in the United States.
Table 1. Five leading causes of water quality impairment (U.S. EPA 1998a)
Rank
1
2
3
4
5
Rivers
Agriculture
Municipal point sources
Hydrologic modification
Habitat modification
j: ••;: •
Resource extraction
Lakes
Agriculture
Unspecified non-point sources
Atmospheric deposition
Urban runofFStorm sewers
Municipal point sources
Estuaries
Industrial discharges
Urban runoff? Storm sewers
Municipal point sources
.Upstream sources
Agriculture
Air deposition of nitrogen is another important source of nutrient loading to water bodies.
More than 23 million tons of nitrogen are emitted to the atmosphere each year in the United
States (U.S. EPA 1998a). About half of the nitrogen compounds emitted from fossil-fuel-
burning plants, vehicles, and other sources in the United States are deposited on U.S. watersheds
(Ibid.). Storm water runoff from cities, construction sites, and agricultural fields is a major cause
of sedimentation problems in water bodies. Sediment movement causes scour around bridge
abutments and piers, stream channeling, and habitat degradation due to deposition of sediment.
292
-------
Under Section §303(d) of the Clean Water Act, each state within the U.S. must produce
and provide the U.S. EPA with a list of waters where water quality standards are not being
attained, to prioritize the development of TMDLs for the water bodies that will result in
attainment of standards, and to develop and implement the TMDLs. In the event a state fails to
develop the list or to develop TMDLs, the U.S. EPA is obligated to do so. Development of
TMDLs typically requires the use of one or more environmental fate and transport models.
Often these models are sophisticated and new users may require assistance (often in the form of
hotline support) and/or training. Special efforts are needed because 16,000 TMDLs must be
developed over the next 8-15 years in order to attain Jthe goals of The Clean Water Act and to
meet court-ordered deadlines. Nearly all of these TMDLs are for contaminants for which major
contributions come from non-point sources. The TMDL research program represents an initial
effort to design, build, and demonstrate the building blocks for TMDL environmental analysis
systems.
RESEARCH PLAN - OVERVIEW
The TMDL Research Plan (U.S. EPA 1999a) encompasses strategies to address
operational requirements, operational concepts, and technology demonstrations to support
successful development, production, and foundation of future designs for regulatory support
exposure assessment systems. The plan is an attempt to implement the risk paradigm (Figure 1)
in terms of evaluating human health, ecological systems and environmental protection.
This approach to developing methodologies for sediment and nutrient TMDLs falls
within larger strategies developed by the U.S. EPA Office of Research and Development (ORD).
These strategies are the "Strategic Plan for the Office of Research and Development" (U.S. EPA
1996,1997); the "Ecological Research Strategy for the Office of Research and Development"
(U.S. EPA 1998b); "NERL Research Strategy" (U.S. EPA 1998c); and the Research Strategy for
Nitrogen (U.S. EPA 1999b).
Research is planned to address major knowledge gaps that have been identified.
These knowledge gaps are:
1. How to extrapolate the stream concentration-discharge dynamic relationship from
hillslope to watershed to river basin scales, including how to utilize existing and remote-sensed
data optimally to simulate multiscale, hydrochemical ecosystems.
2. What are the sensitive hydrologic, geochemical, landscape modifications, and habitat
changes that mediate exposure/response of watershed/regional ecosystems? How do we quantify
the interplay between these changes to the equations of motion for flow and solute transport,
including non-linear controls and responses?
3. How to characterize stressors for exposure assessment. This will include land use
changes~what exists at present, compare to the past, where is it being done on the watershed,
intensity, and projected future land use. Other inputs include multimedia inputs of chemicals
(e.g., air deposition, farming), point sources (chemical and non-chemical), and
climate/meteorological change.
293
-------
Ecological Risk Assessment
PROBLEM FORMULATION
A
N
A
L
Y
S
I
S
Characterization
of
Exposure
Characterization
of
Ecological
Effects
RISK CHARACTERIZATION
Discussion Between the
Risk Assessor and Risk Manager
(Results)
Risk Management
Figure 1. Framework for ecological risk assessment (From U.S. EPA 1992)
294
-------
4. How to develop watershed/regional models that can be used to account for the
interaction of (1) and (2) under the influences of (3) for exposure/response prediction.
The research questions we are attempting to address are:
1. What are multi-component concentration-storage or concentration-discharge
relationships at the watershed or river-reach scale for interactions of hydrochemical systems, and
how to separate or distinguish among essential spatial and temporal components influencing
these relationships?
2. What are the geochemical kinetic processes that influence these exchanges between
soil organic material and relatively important TMDL gases (carbon, nitrogen)? How are these
processes modified by natural and human-induced changes in earth/ecological systems? How do
geochemical processes mediate the exposure and response of earth/ecological systems to climate
change and other stressors?
3. What are the feedbacks between geochemical cycles and changes in habitat or
hydrology in terms of spatial-temporal processes and perturbation? How do biogeochemical
transformations influence exposures to ecological stressors (e.g., mercury)?
4. What geochemical processes control nitrogen- and phosphorous-cycling? How do
land uses or management regimes modify exposures to nonpoint stressors?
5. How do biogeochemical processes mediate the exposure and response of regional
ecosystems to natural and anthropogenic stressors, including land use and climate change,
multimedia pollutant inputs, etc.? Processes include carbon- and nitrogen-cycling, storage and
release, organic pollutant transportation/fate, metal speciation/fate, and plant/soil-pollutant
interactions.
6. Under what conditions are chemical stressors more (or less) important than physical
and land use stressors?
a. How will anticipated global climate changes exacerbate the structural and functional
responses of ecosystems that are already subject to chemical and land use stressors?
What constitutes a reasonable framework for evaluating watershed ecosystem responses
to climate change?
b. How does the relative importance of chemical, water quality, and land use stressors
change when predicting and evaluating the ecological and agricultural sustainabilities
of multi-use watersheds and landscapes?
c. What natural systems are most vulnerable to climate and land-use change, and what
effects on sustainability should be expected?
d. How fast is the biosphere being changed through natural and human-induced causes?
295
-------
I
e. Under what conditions is dietary and foodweb exposure the most important route of
exposure for aquatic wildlife? How do changing prey availability and ontogenetic
changes in prey preferences affect expected patterns of dietary exposure?
f. How do the spatial and temporal dynamics of habitat availability influence the dietary
exposure of highly mobile or migratory wildlife to contaminants?
7. What are the interactions between biogechemical cycles and habitat development,
succession, and sustainability, including soils formation and regional water budgets and
hydrology?
8. ^Tiat are the appropriate tools for integrating the interactions of chemicals, chemical
processes, nitrogen/carbon/phosphorus geochemistry, land use changes, and alternative land
management?
The TMDL research program will employ several tools to link the identified needs and
the research to address those needs to the technologies that could potentially satisfy these needs.
The initial tool for analysis should be multivariable analysis techniques that are widely accepted
throughout the defense, automotive, and electronic industries as an aid in the decision-making
process. These analytical techniques are designed to prioritize potential solutions for established
needs, and is especially well suited as a vehicle for including the "customer" voice in the overall
process.
Modeling and simulation has several distinct levels or building blocks, all of which
contribute to the TMDL decision evolution. The three levels are Constructive and Simulation
Modeling, Interactive Digital Simulation, and Visual Simulation. Each level utilizes and
expands upon the previous underlying levels. Also, at each level the potential solutions can
become more focused and refined. Even though many of the decision tools already exist, this
process will continue over the next several years, but will eventually provide the linkage from
initial requirements to validated, cost-effective solutions.
'.'"'. ' •' , •' . :" .•)•••'..
Constructive and Simulation Modeling builds on the analysis results, providing more
quantifiable and focused estimates of potential needs and solutions. Constructive and Simulation
Modeling can evaluate various levels of characteristics and can complement the "quick look"
direction provided by analysis. The TMDL Program will use a central core of several existing,
accepted, and complementary models for the Constructive and Simulation Modeling effort and
augment this core with additional simulations, as required. This hierarchy of models needs to be
identified, and the databases need to be collated, guided by the initial analysis prioritization.
" ' " , ' , ........ •' '• ! "' ' •
The second level of modeling and simulation, Interactive Digital Simulation, will
capitalize on the results of Constructive and Simulation Modeling by combining Geographic
Information Systems (GIS) technology with satellite imagery and the Internet. Activities will
maximize exploitation of computer analysis technology, will build on the refinements provided
by Constructive and Simulation Modeling, and it is relatively inexpensive to evaluate several
potential solutions. For this step of modeling, the field of potential technology solutions will
have been narrowed and refined as requirements and technologies evolve.
296
-------
Visual Simulation is the final simulation tool and will evaluate a few options within a
visualization network. This will combine TMDL configuration benchmark modeling cases,
along with digital simulation, into the visual simulation. This arena is an architecture that allows
free play within the modeling area. This step is interactive with the other existing and potential
systems and will evaluate the true advantage that this program brings to regulatory support.
We believe that this approach will provide a useful and effective means of viewing
the relationship between sources of environmental stress and eventual health risks to both the
human population as well as the environment itself. Key to this approach is the organization of
information within a more holistic systems framework. Instead of regulating individual
discharges to individual environmental media, the approach will assess entire industrial sectors
and the environments in which they function as an integrated whole to provide tools that can be
used to minimize the collective risk of all pollutant and risk generating activities.
As an example of this process we will use the U.S. EPA Hydrological Simulation
Program - FORTRAN (HSPF) (Bicknell et al 1993) that can simulate the hydrologic and
associated water quality processes on pervious and impervious land surfaces, and in streams and
well-mixed impoundments. HSPF was initially released in 1980, and has undergone subsequent
revisions, modifications, and enhancements; these include the release of a PC version of the
model in 1988, improved sediment and nutrient interaction capability, provision for atmospheric
deposition and inclusion of a forest nitrogen module, and many others. The fundamental
equations used to simulate the hydrologic water budget on pervious land segments have
remained essentially unchanged. Although the HSPF algorithms are robust, and the model has
been successfully applied to a wide range of situations, many of the water quality modules
contained in its architecture are simplistic in their representation of impoundments (e.g., large
lakes, estuaries) or are in need of algorithmic enhancement (e.g., in stream routing/sediment
interactions). The simplistic nature of these water quality modules limits HSPF's ability to
perform complete TMDL assessments for either sediment or nitrogen/phosphorus which are high
priority pollutants for TMDLs. Model upgrades are consequently underway to provide the water
quality components necessary for simulating eutrophication, including riverine (i.e., in-stream)
routing/interactions, receiving water bodies such as lakes/impoundments, and end-points such
as estuaries and their associated biological and chemical interactions. The following
processes/components are necessary to conduct TMDL water quality assessments and are
planned to be included in the modeling system:
1. Heat exchange processes that need to be considered include incident shortwave
radiation, longwave back radiation, conduction-convection, evaporation and precipitation;
thermal enrichment from heat absorbing surfaces; and shading.
2. Transfer across media boundaries, depletion by sediment and benthal oxygen
demand, depletion by nitrification, depletion by carbonaceous oxygen demand, depletion by
biotic respiration, and influx due to photosynthesis.
3. Interactions between the water column and the benthos are also critical processes
for nutrient cycling in many waterbodies.
297
-------
rsi'ii1!1"1
',,,1,1'vj . ^
-------
The enhanced HSPF model will be tested using a set of carefully designed benchmark
problems designed to include all of the modified features in HSPF. This task also includes
comparing the modified version of HSPF to the original version (Version 11). Where
practicable, comparisons will also be made with other hydrologic simulation models to provide
additional testing.
The TMDL Program is designed to create the building blocks for affordable, successful
development of TMDL modeling systems. Ultimately the models will be integrated with a
parallel effort to develop an open architecture multimedia modeling toolbox, representing the
next-generation modeling systems.
ACKNOWLEDGEMENT
This paper has been reviewed in accordance with the U.S. Environmental Protection
Agency's peer and administrative review policies and approved for presentation and publication.
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
REFERENCES
Bicknell, B.R., J.C. Imhoff, J.L. Kittle, A.S. Donigian, and R.C. Johanson. 1993. Hydrological
Simulation Program - FORTRAN (HSPF): Users Manual for Release 10. EPA-600/
R-93/174, U.S. EPA, Athens, Georgia, 30605.
Lochmuller, C.H., and J.M. Malin (Eds.). 1998. Proceedings of the U.S.-Baltic Workshop on
Environmental Chemistry, Palanga, Lithuania, June 1997. Critical Reviews in Analytical
Chemistry, 28:170 pp.
NOAA (National Oceanic and Atmospheric Administration). 1998. Committee on Environment
and Natural Resources Hypoxia Work Group for the Mississippi River/Gulf of Mexico
Watershed Nutrient Task Force. Gulf of Mexico Hypoxia Assessment Plan. Coastal
Ocean Program Office, National Oceanic and Atmospheric Administration, Silver Spring,
Maryland.
U.S. EPA (United States Environmental Protection Agency). 1992. Framework for Ecological
Risk Assessment. Risk Assessment Forum, Office of Research and Development,
Washington, D.C. Publication No. EPA/630/R-92-001.
U.S. EPA (United States Environmental Protection Agency). 1996. Strategic Plan for the
Office of Research and Development. Office of Research and Development, U.S.
Environmental Protection Agency, Washington, D.C. EPA/600/R-96/059.
U.S. EPA (United States Environmental Protection Agency). 1997. 1997 Update to ORD's
Strategic Plan. Office of Research and Development, U.S. Environmental Protection
Agency, Washington, D.C. EPA/600/R-97/015.
299
-------
U.S. EPA (United States Environmental Protection Agency). 1998a. Clean Water Action Plan:
Restoring and Protecting America's Waters. Office of Water, U.S. Environmental
Protection Agency, Washington, D.C.
•
U.S, EPA (United States Environmental Protection Agency). 1998b. Ecological Research
Strategy. Office of Research and Development, U.S. Environmental Protection Agency,
Washington, D.C.
U.S. EPA (United States Environmental Protection Agency). 1998c. NERL Research Strategy.
National Exposure Research Laboratory, Office of Research and Development, U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina.
, ' i
U.S. EPA (United States Environmental Protection Agency). 1999a. TMDL Research Plan
(Draft). Office of Research and Development, U.S. Environmental Protection Agency,
Athens, Georgia.
i
ilh ifl i
US. EPA (United States Environmental Protection Agency). 1999b. Nitrogen Research
Strategy (Draft). Office of Research and Development, U.S. Environmental Protection
Agency, Washington, D.C.
•;. , " " . . • •; i i • '
U.S. EPA (United States Environmental Protection Agency) and USGS (United States
Geological Survey). 1999. The Ecological Condition of Estuaries in the Gulf of Mexico.
EPA/620-R-98-004.
*U.S. GOVERNMENT PSINTING OFFICE: 2000-550-101-20025
300
-------
-------
United States
Environmental Protection Agency
Center for
Environmental Research Information
Cincinnati, OH 45268
Please make all necessary changes on the below label,
detach or copy, and return to the address in the upper
left-hand corner.
If you do not wish to receive these reports CHECK HEREO ;
detach, or copy this cover, and return to the address in the
upper left-hand corner.
Official Business
Penalty for Private Use
$300
EPA/600/R-00/015
------- |