UnilH-d 3t^*
Environmental protection
/, . Genetic Diversity as an Indicator
of Ecosystem Condition and Sustainability
"^ f t
... • >
,
BIODIVERSITY
*•
[
.
Utility for Regional Assessments of Stream
Condition In the Eastern United States
* *
.-*i
-------
Condition and Sustainability
EPA/600/R-03/056
September 2002
Genetic Diversity as an Indicator of
Ecosystem Condition and Sustainability
Utility for Regional Assessments of Stream
Condition in the Eastern United States
Mark J. Bagley
Susan E. Franson
Suzanne A. Christ
Eric R. Waits
Greg P. Toth
U.S. Environmental Protection Agency
National Exposure Research Laboratory
Ecological Exposure Research Division
Molecular Ecology Research Branch
Cincinnati, OH 45268
Recycled/Recyclable
Printed with vegetable-based ink on
paper that contains a minimum of
50% post-consumer fiber content
processed chlorine free.
-------
Genetic Diversity as an
Ill I I
Notice
The research described in this report has been funded wholly or in part by the U.S. Environmental Protection
Agency. This document has been prepared at the EPA National Exposure Research Laboratory (Ecological
Exposure Research Division, Cincinnati, Ohio) based on data collected by EPA and Sobran, Inc. (Contract 68-C6-
0019) and with publication assistance provided by CSC, Inc. (Contract 68-WO-1032).
Mention of trade names or commercial products does not constitute endorsement or recommendation of use.
The correct citation for this document is:
Bagley, M.J., S.E. Franson, S.A. Christ, E.R. Waits, G.P. Toth. 2002. Genetic Diversity as an Indicator of
Ecosystem Condition and Sustainability: Utility for Regional Assessments of Stream Condition in the Eastern
United States. U.S. Environmental Protection Agency, Cincinnati, OH.
-------
Condition and Sustainability
Preface
In 1995, the U.S. Environmental Protection Agency's National Exposure Research Laboratory (NERL)
enhanced its ecological research and development efforts to support the delivery of the "next generation"
of biological indicators. These new ecological condition and stressor diagnostic indicators would follow
the risk paradigm organization of the U.S. EPA's Office of Research and Development (ORD). ORD's
Ecological Research Strategy (US EPA, 1998) called for the "development of a set of indicators for estu-
arine, stream, and lake systems that can be interpreted relative to status and changes in fundamental eco-
logical and hydrologic processes that influence and constrain the integrity and Sustainability of these sys-
tems." Genetic diversity indicator research was initiated in NERL to address this objective.
In 1998, ORD's National Center for Environmental Research (NCER) requested proposals from the aca-
demic research community for Science to Achieve Results (STAR) grants focused on "Ecological
Indicators". Research sponsored by these grants emphasized genetic diversity and landscape ecology, both
of which can be interpreted at a number of geographic scales, a requirement for the next generation of eco-
logical indicators. In May 2000, NCER sponsored a review of genetic diversity science including both
ORD scientists and STAR grant recipients during which a roundtable discussion delineated contributions
as unique to either academic or federal laboratories. Large, regional-scale evaluations of genetic diversi-
ty within species, deemed beyond the scope of academic laboratories, were seen to be appropriate for fed-
eral laboratories and easily incorporated into existing environmental monitoring studies such as the
Environmental Monitoring and Assessment Program (EMAP). These views were mirrored by recom-
mendations in the July 1999 NSF Task Force on the Environment document "Environmental Science and
Engineering for the 21st Century" which called for genetic diversity research within federal laboratories.
Current advances in molecular biological science and technology have converged with classical genetic
research, large-scale field biological monitoring, and remote sensing capabilities to provide unprecedent-
ed opportunities for multifaceted studies of species population structure and dynamics. These types of
studies are critical to understanding the integrity and Sustainability of ecosystems.
This report chronicles significant strides made in the development of an ecological indicator based on
genetic diversity that is suitable for environmental monitoring studies at a range of geographic scales.
Implementation of regional studies of genetic diversity required development of protocols for inclusion of
genetic sampling in field studies and large-scale laboratory throughput of these samples. Robust statisti-
cal methods facilitated the meaningful interpretation of genetic diversity DNA fingerprinting data in the
context of other environmental data collected concurrently. Documentation of these procedures for meas-
uring genetic diversity is presented herein, along with the background and rationale for employing genet-
ic diversity as an ecological indicator. Case studies are presented which demonstrate the application of
genetic diversity in two field-monitoring efforts. Finally, recommendations are given for genetic diversi-
ty study design and technology transfer based on field and laboratory experience with large-scale studies.
-------
Genetic Diversity as an
Ill I I
Summary
Genetic diversity is a fundamental component of biodiversity and is as critical to sustainability of our nat-
ural resources as are diversity of species and ecosystems. It encompasses all of the genetically determined
differences that occur between individuals of a species. Virtually all species are composed of populations
that exist somewhat independently of each other, and thus genetic diversity exists both within and among
populations. Levels of genetic diversity in any one population are determined primarily by four forces:
(1) mutation, the ultimate source of all genetic diversity; (2) migration, the exchange of individuals
between populations; (3) natural selection, the removal of "unfit" individuals from the population; and (4)
genetic drift, random changes in gene frequency each generation due to limited numbers of breeding
adults. The natural history of a species and the structure and dynamics of populations provide the arena
in which these forces interact to drive evolutionary adaptation of populations to their environments. Thus,
natural and anthropogenic environmental changes lead to changes in genetic diversity, both within and
among populations, and genetic diversity measurement can provide insights into the consequences of envi-
ronmental changes.
Genetic diversity can be measured by examining common morphological or morphometric traits. Such
observable characteristics often result from the interaction of many genes, the expression of which is influ-
enced by environmental factors. Assessments of molecular markers based directly on DNA have simple
inheritance patterns and are not influenced by environmental factors. This simplification of the genetic
system allows precise estimates of genetic diversity for any one marker and, by assessing many markers,
can provide more precise estimates of overall levels of genetic diversity within and among populations of
a species. Mathematical tools have been developed that allow diagnosis of the relative strengths of the
four genetic forces and, indirectly, properties of populations, such as population size, breeding structure,
and dispersal abilities.
Measurement of genetic diversity with molecular markers can add value to assessments of ecological con-
dition derived from other ecological indicators, such as landscape and species assemblage indicators.
Population parameters can be effectively estimated with molecular markers and used to characterize the
geographic structure and connectivity of populations critical to interpreting data for ecological assess-
ments. Genetic diversity also serves as an independent indicator of environmental condition.
Environmental stressors typically reduce genetic diversity, primarily through the forces of selection and
genetic drift, so that a recent reduction in genetic diversity is indicative of deteriorating environmental
condition. As an indicator of ecological condition, genetic diversity integrates the genetic effects of mul-
tiple sources and is cumulative over time. In addition, it is a naturally 'scalable' indicator, as the geo-
graphic structuring of genetic diversity at the population, watershed, and regional levels is easily inferred.
The importance of genetic diversity to long-term sustainability is widely accepted, although the efficacy
of molecular measures of genetic diversity for diagnosing extinction risk remains unclear and needs fur-
ther investigation. Standing levels of genetic diversity in populations contribute to long-term sustainabil-
ity in several ways. First, the ability of populations to adapt to changing environments is directly depend-
ent on the amount of genetic diversity they possess. Second, small populations that lose genetic diversity
may experience fitness reductions and increased extinction risk. Finally, populations that are adapted to
IV
-------
Condition and Sustainability
local conditions may become less fit if individuals from other areas that are adapted to different conditions
are allowed to interbreed with them; the ensuing reduction in genetic diversity between populations influ-
ences long-term sustainability of the species.
This report documents research undertaken to determine if the theoretical promise of genetic diversity as
an ecological indicator is realized in real-world applications. Results of two case studies confirm that
genetic diversity is a useful indicator of environmental condition. The first case study incorporated the
genetic diversity indicator in a larger Regional Environmental Monitoring and Assessment Program study
of the Eastern Cornbelt Plains Ecoregion, done in collaboration with US EPA Region 5 and Ohio EPA.
Genetic diversity of a small cyprinid minnow, the central stoneroller (Campostoma anomalum), was meas-
ured at 91 sites in nine watersheds using the RAPD (random amplified polymorphic DNA) fingerprinting
technique. Although the RAPD technique was chosen primarily for ease of technology transfer, experi-
ences with the technique suggested that it might not be robust to the normal variations in equipment and
technical skills that exist among different laboratories. Nonetheless, the genetic diversity data obtained
proved highly informative. Although sample sizes varied and were sometimes small (3-10 individuals per
site), large differences in genetic diversity within sites and among sites were detected. Significant differ-
ences in the average levels of genetic diversity within populations were observed among major river
drainage basins, leading to the conclusion that populations of stonerollers are highly differentiated within
the Eastern Cornbelt Plains Ecoregion and there is geographical structuring of these populations within
and among watersheds. Genetic diversity is related to environmental condition, particularly impacts from
urbanization, channelization, and impaired riparian zones. Expected relationships between genetic diver-
sity and existing ecological indicators such as the IBl and QHEI were seen, although the small degree of
correlation suggests that the genetic diversity indicator provides supporting and not highly redundant
information for environmental condition assessments.
The second case study examined the genetic diversity indicator applied to populations of the creek chub
(Semotilus atromaculatus) in a small region of western Pennsylvania and West Virginia underlain by coal-
bearing geology and for which the history of coal mining operations is known. Samples of between 9 and
28 creek chubs were collected from 10 sites within 4 watersheds. Two molecular methods were used: the
amplified fragment length polymorphism (AFLP) fingerprinting technique was used to assess diversity in
the nuclear genome, while a portion of the mitochondrial genome was assessed using DNA sequencing.
Mitochondrial DNA differences showed a strong spatial component. The nuclear DNA also differentiat-
ed the populations although the genetic structure was not as strong as that seen in the mitochondrial DNA.
Environmental factors (derived from principal components analysis of 25 key environmental measure-
ments) accounted for about half of the differences in mitochondrial DNA diversity and virtually all of the
differences in nuclear DNA (AFLP) diversity.
These two case studies clearly demonstrate that genetic diversity can serve as an indicator of environ-
mental condition. They also provided the practical experience upon which recommendations for future
implementation are based.
At present, genetic diversity indicators will be used most effectively if they are incorporated into multi-
indicator assessments at large, regional scales. Typically, reduced genetic diversity in particular popula-
-------
Genetic Diversity as an
Ill I I
tions is inferred from assessments of spatially separated populations, although it can be detected more eas-
ily through temporal monitoring. Thus, incorporation of genetic diversity indicators into monitoring pro-
grams at intensively studied index sites will be useful. Regional genetic diversity data can then serve as
baseline data for future monitoring of temporal patterns in genetic diversity at different spatial scales.
A number of different molecular technologies can be used for genetic diversity analysis, including
allozyme, DNA fingerprint, microsatellite DNA, and mitochondrial DNA fragment or sequence analysis.
At present, it appears that the most cost-effective strategy is to incorporate microsatellite markers into
existing or planned ecological assessments. It may be beneficial to supplement microsatellite analysis
with mitochondrial DNA analysis since mitochondrial DNA can yield complementary information. This
approach is the most technologically challenging of all the genetic diversity assessment options. Thus, the
recommendation is that a "three-laboratory approach" be used to obtain and interpret genetic diversity
data. The lead lab would be the regional field lab, which would be responsible for design of the assess-
ment, field collections, and preparation of DNA samples. A marker development laboratory would design
molecular markers specific for the target species identified by the regional lab. A genetic analysis labora-
tory would use the molecular markers and DNA samples obtained from the other labs to perform the genet-
ic diversity assessment and, together with the regional laboratory, derive the ecological assessment.
VI
-------
Condition and Sustainability
Table of Contents
Preface iii
Summary iv
Table of Contents vii
1 Introduction 1
1.1 What is Genetic Diversity? 1
2 Rationale for a Genetic Diversity Indicator 4
2.1 Relevance of genetic diversity to assessment of ecological condition 5
2.2 Relevance of genetic diversity to assessment of population sustainability 11
2.3 Measurement of Genetic Diversity 13
3 Case Studies 16
3.1 Genetics of Central Stonerollers in The Eastern Cornbelt Plains Ecoregion 16
3.1.1 Background 16
3.1.2 Key findings and their implications 19
3.2 Genetics of Creek Chubs in a Mining-Impacted Region 32
3.2.1 Background 32
3.2.2 Key Findings and their Implications 35
3.3 Generalizations from the Case Studies 40
4 Considerations when Implementing a Genetic Diversity Assessment 43
4.1 Sampling Design 43
4.2 Species Choice 45
4.3 Which Genetic Marker? 46
4.4 Sample Size Considerations 53
4.5 Personnel Training and Specialized Equipment 55
4.6 Information Management 55
4.7 Costs 56
4.8 Summary Recommendations 59
Glossary 62
References 65
Acknowledgements 72
Appendix 1: Laboratory and analytical procedures for RAPD analysis 73
Appendix 2: Laboratory and Analytical Procedures for AFLP Analysis 76
Appendix 3: Laboratory and Analytical Procedures for DNA Sequence Analysis 77
VII
-------
Genetic Diversity as an Indicator of Ecasjfijjll
ill
viii
-------
Condition and Sustainability
1. Introduction
Historically, the U.S. Environmental Protection Agency and other governmental agencies charged with
monitoring and safeguarding the quality of aquatic environments tended to concentrate their efforts on
assessment of the immediate toxicological effects of discharges from known pollution sources. Gradually,
as the more flagrant pollution sources were identified and removed, it became clear that much of the
remaining pollution in our waterways could not be ascribed to specific point sources but was generated at
low levels from many nondescript sources. In addition, there was concern that aquatic ecosystem health
was not a simple function of water quality and could not be fully measured based on the results of a series
of short, standardized laboratory tests. Following the publication of an influential report over a decade ago
(U.S. Environmental Protection Agency, 1988), the EPA began to shift its ecological research to focus
attention on ecosystem responses to cumulative, regional, and long-term anthropogenic disturbances. The
EPA's Environmental Monitoring and Assessment Program (EMAP) was born of this movement, and was
given the mission to evaluate the status and trends in the condition of the Nation's ecological resources
(Messer,etal, 1991).
A number of ecological indicators have since been developed that have, in one fashion or another, aided
in measuring the overall health of aquatic ecosystems. These include indicators based on types of assem-
blages of species present and landscape or land cover indices. The application of these indicators has
clearly expanded our toolbox for monitoring the status of aquatic ecosystems. Nonetheless, important
information about the current status and predicted future status of the Nation's aquatic ecosystems still
remain to be addressed. For example, current indicators tell us little about the biological independence or
connectivity of different geographic areas. How important to ecosystem health are the natural migration
barriers and corridors that separate and connect different populations? If populations of a species from
one region are extirpated, will individuals that are similarly adapted to that environment replace them?
How much of the evolutionary history of a species will be lost if the habitat of one distinct population is
destroyed to make room for a new development?
In addition, we know little about the many hidden effects of anthropogenic stressors on species and ecosys-
tems that do not manifest as gross deformities, dead or dying fish, or extinct species. Do these stressors
evoke heritable changes in organisms that may affect the Sustainability of populations many generations
into the future? If past or current stressors have altered the genetic characteristics of populations, have
these changes affected their ability to withstand the additional stresses of future human population growth,
global warming, and associated environmental modifications? Are some populations of a species geneti-
cally "pre-programmed" to withstand severe environmental modifications while others are not?
These are important questions that must be answered to truly understand the condition and Sustainability
of the Nation's ecological resources. Clearly, they cannot be answered using simple streamside assays or
laboratory tests. To answer these questions, we must come to understand the dynamics of populations and
how those dynamics will be altered in the face of anticipated environmental changes. To understand the
long-term effects of environmental changes on populations, we must learn how environmental modifica-
tions impact genetic diversity within and among populations and how changes in genetic diversity ulti-
mately influence the ability of populations to withstand further environmental change.
1.1 What is Genetic Diversity?
The term 'genetic diversity' has been defined in various ways to suit different purposes. Here, we restrict
our interests to intraspecific genetic diversity, which represents the range of heritable differences of a trait
-------
Genetic Diversity as an
Ill I I
or set of traits among individuals within a species and includes diversity among individuals within popu-
lations as well as variation among different populations (Figure 1-1). The term is essentially synonymous
with 'genetic variation' or 'genetic variance'. As defined, we consider genetic diversity a trait of popula-
tions and species, not of individuals. Examples include variation in the genetic determinants of eye color,
growth rate, and disease resistance within and among populations. Similar variation can be assessed in
specific stretches of DNA using molecular markers. Genetic diversity is one of the three components of
biodiversity, along with species diversity and ecosystem diversity. In a sense, it is the fundamental orga-
nizational component of biodiversity, since species diversity is built from genetic diversity, and ecosystem
diversity derives from species diversity. Thus, any goal to monitor or maintain biodiversity is incomplete
and superficial if it considers only species and ecosystems.
Genetic diversity is shaped by past population processes and affects the sustainability of species and pop-
ulations in the future. Physical, chemical, and biological environments are constantly changing over short
and long time scales. To keep up with changing environments, species must adapt or face extinction. The
ability of species to adapt to altered environments is directly related to the amount of genetic diversity
available to natural selection. For two populations that are under similar environmental pressures to adapt,
the population with the greater genetic diversity for fitness traits is expected to adapt more quickly. That
the maintenance of genetic diversity is key to the long-term survival of most species is a central paradigm
of the field of conservation genetics (Soule, 1987).
The main forces that determine current levels of genetic diversity within species are: mutation, migration,
selection, and genetic drift (Figure 1-2). Mutation is the ultimate source of all genetic diversity, but is usu-
ally a relatively weak force in comparison to the other three. Most new mutations are either neutral or
harmful with respect to fitness. A very small proportion of mutations is expected to increase fitness.
Migration or gene flow represents the movement of breeding individuals between populations. It is usu-
ally a very strong force on genetic diversity, although it can be weak for species with low dispersal abili-
ties. Disturbance of a species' migration regime may have substantial effects on fitness. In general, migra-
tion has the effect of increasing genetic diversity within populations and homogenizing genetic differences
among populations.
Natural selection, along with genetic drift, is one of the main forces that cause separate populations to
become differentiated. It is the primary force that increases the average fitness of populations. Selection
can have variable effects on genetic diversity, depending on the form of selection that is operating. In
cases where an environmental change has caused a different set of genes or genotypes to be optimal, selec-
tion is expected to decrease genetic diversity at selected loci, at least until the population mean has shift-
ed to the new genetic optimum.
Genetic drift represents the random change in gene frequencies in each generation due to a finite number
of breeders producing each new generation. It is the statistical equivalent of sampling error. Without the
input of new mutations, all genetic diversity within species would eventually be lost due to the effects of
genetic drift. Genetic drift tends to be a very strong force on genetic diversity in small populations, even
for loci that are under intense selection. As a consequence, adaptation is more difficult in small popula-
tions while random fixation of deleterious alleles that would normally be removed by selection is more
likely. The effect of genetic drift in large populations is weaker but may still be a dominant force for loci
that are not under selection. Genetic drift has the effect of decreasing genetic diversity within populations
and increasing genetic diversity among populations.
-------
r
Diversity as an Indicator of Ecosystem Condition and Sustainability
o.
(D p_o ^
°=0=!
ro .^ o
at?
" «g o
§ = is
- o
±^ 1- Q. O Q.
£ n
-------
Genetic Diversity as an Indicator of Ecosystem Condition and Sustain
I I I
Migration (+)
Genetic drift (-)
Local adaptation (-)
Within
Populations
Selection (+/-)
Mutation / recombination (+)
Genetic drift (-)
Migration (-)
Genetic drift (+)
Local adaptation (+).
Among
Populations
o
Species
Figure 1-2. Forces that influence overall (species-level) genetic diversity, as well as components of genetic diversity
among and within populations. Forces that are annotated with a + generally increase genetic diversity; those anno-
tated with a - generally decrease diversity. Selection may either increase or decrease genetic diversity, depending
on specific circumstances.
2 Rationale for a Genetic Diversity Indicator
It is now a widely accepted practice to integrate genetic management ideas into resource management
planning (e.g, Hessel, 1992; Moritz, 1994). The primary goal of genetic management is to ensure that suf-
ficient genetic diversity is retained in order to maintain short-term fitness and long-term sustainability and
to ensure that local adaptations are not lost due to intermixing of previously distinct populations. An addi-
tional goal is to identify and protect those populations that represent highly distinct or important evolu-
tionary lineages, because if such populations become extinct, an important component of the genetic diver-
sity of the species would be lost. The idea that genetic monitoring can be used as an indicator of envi-
ronmental condition is less prevalent in resource management plans (salmon management in the Pacific
Northwest being a notable exception), primarily because most resource agencies still consider genetic
techniques to be novel, expensive, or technically challenging. This has created the somewhat paradoxical
situation in which resource managers pay close heed to conservation genetic principles, but have not
-------
ITTT
jjjU''&f.Ecosystem Condition and Sustainability
adopted policies to measure the success of man-
agement actions in maintaining genetic diversity.
Nonetheless, an enormous database of genetic
diversity studies exists in the scientific literature
and can provide guidance in the development and
evaluation of genetic diversity indicators.
of the of
a for
is
to the
of
I • •
2.1 Relevance of genetic diversity to
assessment of ecological condition
The genetic diversity of populations responds to
environmental heterogeneity via alterations in the
relative strengths of the four opposing genetic forces: mutation, migration, selection, and genetic drift.
The balance and cumulative history of these forces determines actual levels of genetic diversity at any one
time. The accuracy and sensitivity of measurements of genetic diversity has steadily increased with
advances in molecular marker technologies. An enhanced ability to describe patterns of genetic diversity
in space and time has, in turn, made it easier to diagnose the genetic mechanisms that produce these pat-
terns (Carvalho, 1998).
By far, the easiest genetic forces to measure in natural populations are equilibrium levels of migration
(gene flow) and genetic drift. Gene flow provides a direct measure of the evolutionary connectivity of
populations while genetic drift provides a measure of effective population size. Populations that have low
connectivity with others become genetically differentiated and more unique. These populations are often
assigned higher conservation 'value' because a large portion of the genetic diversity of the species would
be lost if they were extirpated (Moritz, 1995). Populations with low effective population size may be more
vulnerable (see section 2.2) and therefore may be targeted for immediate conservation efforts. The use of
genetic markers to measure connectivity and effective population size and to assign conservation values
and priorities is commonplace, as evidenced by numerous articles in scientific journals such as Molecular
Ecology, Evolution, Conservation Biology, and Conservation Genetics. Thus, the concept that genetic
diversity is a useful indicator of population structure and conservation status is well established. However,
in these studies, evaluations of genetic diversity are almost always focused narrowly on the status of par-
ticular populations residing in specific habitats; inferences to overall ecological condition based on genet-
ic diversity are not usually made.
For the purpose of performing environmental assessments, one of the most important potential uses of a
genetic diversity indicator is simply to characterize the natural population structure of component popula-
tions within ecosystems. The effective population size and connectivity of populations are fundamental
attributes of populations that can be diagnosed only with genetic markers. Because populations are the
basic biological units that respond to changes in the environment, delimitation of population boundaries is
critical to effective ecosystem monitoring and management.
Without this basic knowledge of population structure, it is impossible to judge the extent or range over
which local or regional environmental impacts will affect ecosystems. In addition, concordance of popu-
lation genetic structure across several species may indicate biogeographic boundaries not recognized
through analysis of species distributions alone (Moritz and Faith, 1998). By defining population bound-
aries, genetic diversity indicators provide fundamental data that enhance the value and interpretation of
other ecological assessment data, such as those obtained from landscape and species assemblage indica-
-------
Genetic Diversity as an Indicator of Ecosystem Condition and Sustainabilit
I I I
tors.
In addition to delineating these long-term, evolutionary characteristics of populations, a useful indicator
of ecological condition should be responsive to environmental change (US EPA, 1998). Genetic diversi-
ty responds to environmental change, primarily through the actions of genetic drift and selection, but also
through mutation and gene flow. The indicator
'remembers' population effects. In other words,
once genetic diversity is reduced it will remain low
until mutation or gene flow replenishes it (Figure
2-1). Because the effects of selection and drift are
cumulative across generations, there is often a lag
time before significant changes in genetic diversity
are revealed (Bickham et a/., 2000). Thus, a genet-
ic diversity indicator is expected to be useful pri-
marily for multigenerational exposures; the rele-
vant timescale for indicator response is years or
decades, not months. Exceptions occur when pop-
ulations are severely bottlenecked or severe selec-
tion pressures are imposed on populations, induc-
ing rapid and severe changes in genetic diversity.
Decreases in genetic diversity in response to experimental bottlenecks of laboratory populations are well
known (Hartl and Clark, 1997). In general, these studies show that declines in heterozygosity at marker
loci follow theoretical predictions regarding the loss of genetic diversity in relation to effective population
size. "Natural" population bottleneck experiments, such as those that occurred when elephant seals were
hunted to near extinction and when European starlings were introduced to North America, also show a
remarkable correspondence with genetic theory (Cabe, 1998; Hoelzel, 1999).
Genetic diversity indicators provide
fundamental data that enhance the
value and interpretation of other
ecological assessment data, such
as those obtained from landscape
and species assemblage indicators.
1.2
e
8 0.8
31 0.6
2
oj 0.4
o>
1 0.2
Ne
•2
•5
•20
•100
•5-100
°>
Generations
Figure 2-1. Predicted loss of heterozygosity over time from genetic drift. Heterozygosity (a measure of genetic
diversity within populations) decreases more rapidly in populations with smaller effective population sizes (Ne). For
the series labeled "5-100", the population started at an effective size of Ne=5 and remained there for five genera-
tions. Starting at generation 6 the effective size grew to Ne=100. Note that information about the past bottleneck
continues to be reflected by heterozygosity after the population has rebounded.
Numerous field studies have demonstrated a correlative relationship between levels of genetic diversity
and single or mixed contaminant exposures. Examples of changes in the genetic structure of plant popu-
lations exposed to mine wastes are well known, and were recently reviewed by Shaw (1998). Several
examples of correlations between genetic diversity and contaminant exposures for freshwater and marine
species are listed in Table 2-1. In most of these cases, the frequency of a particular allele or genotype shift-
-------
irrr
', Jii-^yf^
mtUEiCOsystem Condition and Sustainability
...
a
is
I I
ed dramatically in exposed populations relative to
frequencies in reference populations. Often, a sim-
ilar shift was observed for the same marker locus
(e.g., the allozyme locus GPI-2) in different
species, suggesting that selection was operating at
that particular marker (Gillespie and Guttman,
1999). In addition, genetic diversity of exposed
populations was often lower than for reference
populations, suggesting that the effective sizes of
exposed populations were reduced thereby induc-
ing strong genetic drift effects. In at least one case,
a significant difference in genetic diversity
between reference and exposed populations was
observed when indicators based on species assem-
blages suggested homogeneity (Roark and Brown, 1996).
Many factors can contribute to differences in allele frequencies so demonstration of population differences
does not, by itself, convincingly demonstrate that the specific contaminants directly impacted genetic
diversity. However, a causal link between susceptibility and resistance to specific contaminants and genet-
ic diversity has been demonstrated in several laboratory and mesocosm experiments (Table 2-1). For
example, Gillespie and Guttman (1989) found that central stonerollers with allozyme genotypes that were
common in contaminated environments had greater survival following laboratory exposure to copper than
those with other genotypes. While significant effects on genetic diversity have been observed for a large
number of organisms and experimental exposures, other studies have failed to identify significant genetic
changes, even in cases where correlations were observed in natural populations (e.g., Diamond et al.,
1989). Although a number of potential reasons can explain the lack of genetic response in these studies,
they demonstrate that genetic diversity is as yet an imperfect indicator of exposure.
Besides chemical inputs, other anthropogenic stressors that affect survival or selection would be expected
to impact genetic diversity as well. In practice, few examples of genetic diversity assessments relative to
recent environmental alterations have been reported. An exception is the rather large number of studies
that have documented the strong effects on genetic diversity of interspecific and intraspecific hybridiza-
tion due to species introductions (e.g., Campton and Johnston, 1985; Gyllensten et al, 1985; Leary et al,
1995; Williams et al., 1996; Ayres et al., 1999). The effect of hybridization on genetic diversity is imme-
diate and, if the hybridizing taxa are genetically distinct, easy to detect with genetic markers. The effect
of habitat fragmentation on genetic diversity also has been investigated. Increased genetic differentiation
of populations in relation to habitat shrinkage has been reported for woodpeckers (Haig et al., 1996),
lizards (Schneider et al, 1998), mint plants (Morden and Loeffler, 1999), and beetles (Knutsen et al.,
2000). The construction of a highway was reported to have increased genetic differentiation between vole
populations on each side of the road (Gerlach and Musolf, 2000). Smith et al. (1983) reported results of
a survey of genetic diversity in mosquitofish that suggested genetic differentiation due to water impound-
ments. If population size within fragments is decreased, fragmentation should result in lower genetic
diversity within populations as well, but this was either not observed or not reported in these studies.
Although there is much interest in measuring the ecological effects of environmental mutagens and car-
cinogens, studies of the heritable effects of these exposures on natural populations are extremely rare.
However, measures of genomic mutation rates and inferences to possible changes in mutation rates, both
globally and in specific environments, are extremely important. Population genetics theory suggests that
even a very small increase in the mutation rate may decrease the fitness of many populations enough to
-------
Genetic Diversity as an
p I1 I1
Table 2-1. Natural populations and laboratory or mesocosm studies of freshwater and marine taxa that indicate a
correlation between genetic diversity and exposure to contaminants (modified and extended from Gillespie and
Guttman, 1999).
Known
Stressors
Mercury, other
metals
Acidity, [Al]
Arsenate
Pesticides
PAH
(fluoranthene)
Radionuclides
Overall water
quality,
complex
effluents
Molecular Marker
Allozyme
Allozyme
____A21ozyjn^___
Allozyme
Allozyme
^^^AM5?XH!£^^-
RAPD, Allozyme
RAPD
Allozyme
CYP1A Sequence
MtDNA/RAPD
MtDNA
RAPD
RAPD
Allozyme
Taxa Natural Population References
(NP)Or
Laboratory /
Mesocosm (L/M)
Study
Gastropods
Shrimp
Fish
Fish
Shrimp
Gastropods
Insect
Fish
Amphipod
Shrimp
Gastropod
Mollusc
Fish
Fish
Fish
Amphipod
Fish
Fish
Fish
Bivalve
Fish
Amphipod
Fish
Fish
Fish
Fish
Fish
Fish
Copepod
Fish
Crayfish
Mussel/Barnacle
Fish
NP
NP
NP
NP
NP
NP
NP
L/M
L/M
L/M
L/M
L/M
L/M
L/M
L/M
L/M
L/M
L/M
L/M
NP
NP
NP
L/M
L/M
T/M
NP
L/M
L/M
L/M
L/M
NP
NP
L/M
NP
NP
NP
NP
NP
NP
NP
NP
NP
NP
Nevoe/a/., 1984
Bentone/a/., 1994
Nevoe/a/., 1984
Reagleretal., 1993
Keklake/a/., 1994
Bentone/a/., 1994
Roark and Brown, 1996
Diamonds a/., 1989
Heaglere/a/., 1993
Mulvey etal, 1995
Tatarae/a/., 1999
Ben-Shlomo andNevo, 1988
Lavie and Nevo, 1986b
Benton and Guttman, 1992a,b
Chagnon and Guttman, 1989
Schluetere/a/., 1995, 1997, 2000
Duane/a/.,2000a
Ben-Shlomo andNevo, 1988
Lavie and Nevo, 1982, 1986a,b
Moragaetal., 2002
Larno etal., 2001
Koppe/a/., 1992
Koppetal., 1992
Duane/a/.,2000a
Newman et al., 1989
Hughes et al, 1991
Brown Sullivan and Lydy, 1999
Tanguy etal., 1999
Schlueter, et al, 2000
Duane/a/.,2000b
Larno et al., 2001
Theodorakis and Shugart, 1997
Theodorakis and Shugart, 1998
Gillespie and Guttman, 1989, 1993
Fore et al., 1995a,b
Heithaus and Laushman, 1997
Roy etal., 1996
Murdoch and Hebert, 1994
Street and Montagna, 1996
Nadige/a/., 1998
Krone etal., 1999
Maet al, 2000
Roark et al, 2001.
..
-------
Condition and Sustainability
drive them to extinction (Lynch, 1996). Measurement of mutations with highly mutable molecular mark-
ers (e.g., microsatellites, minisatellites) can reveal differences in mutation rates at both the individual and
population levels. At the population level, an increase in the mutation rate is expected to result in a large
frequency of very rare alleles. In practice, it may be more straightforward to measure the accumulation of
new mutations directly, by assessing genotypes of parents and progeny, or by assessing the frequency of
novel genotypes in gametes of reproductive individuals. For example, analysis of multi-locus DNA fin-
gerprints of herring gulls and their offspring from an industrialized urban harbor indicated higher muta-
tion rates than for gulls originating from areas with lower anthropogenic inputs (Yauk and Quinn, 1996).
A similar analysis of microsatellite loci in barn swallows near Chernobyl, demonstrated two-to-tenfold
higher mutation rates near the radioactively contaminated area (Ellegren et a/., 1997).
The scientific literature reviewed in this section shows a clear relationship between genetic diversity and
ecological stressors that is consistent with genetic theory. In general, different stressors should have pre-
dictable consequences for measurements of genetic diversity. This is not to say that genetic diversity is
likely to be useful to diagnose specific stressors, but that an established cause and effect relationship exists.
Thus, loss of habitat that causes a decrease in the rate of migration between populations (fragmentation)
is expected to increase the component of genetic diversity among populations, which may be measured as
a decrease in the average similarity between populations or an increase in estimates of one of Wright's
(1978) F-statistics, FST (Figure 2-2). If fragmentation results in small breeding population sizes, then a
second predicted outcome would be reduced within-population genetic diversity or heterozygosity (H),
and reduced numbers of segregating alleles per locus (Na). Similarly, significant degradation of habitat
due to contaminant exposure, siltation, changes in flow regime, or other modifiers that greatly reduce pop-
ulation size are expected to decrease genetic diversity within populations (reduced H, Na) and greatly
reduce genetic diversity of genetic markers that are linked to loci affecting adaptation to the altered habi-
tat. The predicted effect of introduced species depends on whether they can successfully hybridize with
native populations. Thus, introduced species are expected to have an effect similar to habitat degradation
(decreased H, Na) due to competition or predation but will increase genetic diversity through hybridiza-
tion. Exposure to environmental mutagens is predicted to increase the average number of alleles at a mark-
er locus (Na) but, since each new mutation will be rare in the population, should have little effect on het-
erozygosity.
One of the principle advantages of genetic diversity measures is that they are highly scalable indices. For
aquatic populations, the basic units of measure are populations that reside in different stretches of streams
or rivers that are nested within watersheds that are nested within larger regions. Measures of genetic diver-
sity, such as Wright's F statistics or Nei's genetic diversity statistics (reviewed in Nei, 1987) are especial-
ly appropriate for analyzing this hierarchical structure. For example, these population statistics can be
used to assess the proportion of all genetic diversity that resides within populations, within local water-
sheds, within larger river drainages, and within regions. In fact, the only limitation to the number of hier-
archical levels of analysis that can be assessed are biological, as the range a species can occupy ultimate-
ly limits the regional application of genetic diversity indicators.
Another important advantage of a genetic diversity indicator is that it naturally lends itself to analysis of
temporal trends. In fact, temporal monitoring of genetic diversity, either regionally or at index sites, is the
preferred application of the indicator (see section 3.1.1). Temporal monitoring provides the most direct
and most effective measure of actual changes in the level or structure of genetic diversity. Analytical tech-
niques have been developed to estimate the effective number of breeding adults from temporal changes in
genetic diversity (Waples, 1990; Jorde and Ryman, 1996).
-------
Genetic Diversity as an
TTT
STRESSOR
EFFECT
MEASURE
Habitat Loss /
Fragmentation
Habitat
Degradation
Exotics /
Admixtures
Dispersal
"ST
Breeders
* H, N
Adaptation
Gene Flow/
Hybridization
Environmental
Mutagens
Mutations
*H, Na at
linked loci
Figure 2-2. Population-level effects of different stressors and their predicted effects on measurements of genetic
diversity. FST is a measure of population subdivision. Heterozygosity (H) and the number of segregating alleles
(Na) are measures of genetic diversity within populations.
Some characteristics of genetic diversity measures that support its use as an indicator of ecological condi-
tion are summarized in Table 2-2.
Table 2-2. Rationale for indicator of ecological condition based on genetic diversity
• Population-level measure rather than individual-level measure
• Measure of cumulative impact of multiple stressors on populations
• Integrative measure at multiple geographic and temporal scales
• Can be implemented as a nondestructive measure (e.g., DMA can be taken from fin
tissue or scales)
• May identify problems within species before species assemblage indicators
become significant
• Well-defined relationships between indicator and the size and connectivity
of populations
• Highly complementary to species assemblage and landscape indicators
• Useful indicator of population trends through temporal monitoring
• Quantitative measure of population and community "conservation worth"
10
-------
iror
ijjOffijcosystem Condition and Sustainability
2.2 Relevance of genetic diversity to assessment of population sustainability
Genetic diversity is both an input and an output of ecosystems. As shown in the previous section, its role
as an output tells us something about the processes occurring within ecosystems. As an input, genetic
diversity tells us something about the future sustainability of ecosystems. The importance of genetic diver-
sity for sustainability of ecosystems is due to its role in determining population fitness in the short term
and population sustainability in the long term.
Due to rapidly increasing human pressures, current rates of environmental modification may be greater
than those most modern species have experienced. For populations with little genetic variation, we expect
that the rate of adaptation to altered environments will be slow. If the rate of environmental change is too
great, then limitations on genetic diversity combined with high demographic costs of selection will not
allow populations to replace their numbers each generation and the populations will slowly slide towards
extinction (Lande and Shannon, 1996; Lynch, 1996; Lande, 1999). Rapid anthropogenic changes such as
global warming are expected to greatly increase the importance of genetic diversity to population persist-
ence in the foreseeable future (Lande, 1999).
is
the
of
i I I
Besides loss of adaptive potential, low genetic
diversity within populations can have an immediate
effect on overall fitness. Experimental evidence
suggests that most, if not all, diploid species harbor
recessive lethal alleles and partially recessive sub-
lethal alleles (Lande, 1999). If genetic diversity is
reduced rapidly, strong genetic drift will result and
many individuals will be homozygous for the
(sub)lethal alleles, allowing them to be fully
expressed. If the population size becomes very
small, sublethal alleles may actually become fixed
(all individuals homozygous for the deleterious
allele), causing a permanent decrease in fitness
and, consequently, increasing extinction risk. The
effect is often called "inbreeding depression" since it is associated with inbred populations in laboratory
experiments. The effects of inbreeding depression on population survival can be severe; most experi-
mental populations of animals that are purposefully inbred will eventually become extinct due to contin-
uous erosion of fitness (Soule, 1980; Frankham, 1995). On the other hand, a more gradual decline in
genetic diversity may allow (sub)lethal alleles to be purged by selection without greatly harming fitness.
A correspondence between loss of genetic diversity at marker loci and reduced fitness has been difficult
to document in wild populations, primarily because it is difficult to quantify population fitness in natural
populations. Nonetheless, evidence for a relationship between particular fitness components and het-
erozygosity has been shown in a number of species (reviewed in Allendorf and Leary, 1986). More recent
studies also suggest that a relationship between genetic diversity and fitness holds for a number of natural
populations (Table 2-3). In only one case, however, was a direct relationship between genetic diversity
and population extinction observed.
In addition to genetic diversity within populations, maintenance of genetic diversity among populations is
important to the long-term sustainability of many species. Populations of many species are adapted to
unique local conditions. In the long run, continued evolution of the species depends on these unique pop-
ulations and a major conservation effort is dedicated to seeking out and protecting these 'evolutionarily sig-
nificant' lineages. Removal of barriers to dispersal, either through habitat modification or translocation of
11
-------
Genetic Diversity as an
Ill I I
Table 2-3. Examples of genetic diversity studies of natural populations that demonstrated an association between
genetic diversity and population fitness.
Endpoint
Local population extinction
Lifetime breeding success
Colony growth and survival
Fertility, hatching rate
Mortality, growth, fecundity,
developmental stability
Developmental stability
Male reproductive success
Male reproductive success
Birth weight, neonatal survival
Molecular marker
Allozyme,
micro satellite
Micro satellite
Allozyme
Micro satellite
Allozyme
mtDNA,
microsatellites
Allozyme
Allozyme
Micro satellite
Taxon
Butterfly
Red deer
Ant
Prarie chicken
Topminnow
Elephant Seal
Butterflies
(2 species)
Gastropod
Seal
Reference
Saccheriefa/., 1998
Slate efa/., 2000
Cole and Wiemsasz, 1999
Westemeir etal, 1998; Bouzat
etal, 1998
Quattro and Vrijenhoek, 1989
Hoelzel etal., 2002
Carter and Watt, 1988
Rolan-Alvarez etal, 1995
Coltmanefa/., 1998
individuals, allows the transfer of locally maladaptive genes into populations, thereby reducing population
fitness. This effect is sometimes called "outbreeding depression". It is believed that the widespread prac-
tice of rearing fish stocks to supplement wild populations may have inadvertently reduced fitness and
caused local extinctions of many fish populations (Nehlsen etal., 1991). In addition, previously isolated
but reproductively compatible species have been allowed to come into contact, creating "hybrid swarms"
(Leary et a/., 1995). In the worst cases, native species diversity may be overrun by genes from hybridiz-
ing introduced species, effectively eliminating native species through dilution.
Long-term sustainability is affected by the ability of populations to adapt to changing environments.
Gilpin and Soule (1986) summarized the theoretical relationship between genetic diversity, adaptation,
demography, and population extinction in the "extinction vortex" model. Under normal circumstances, a
temporary decrease in population fitness due to a change in the environment is corrected via the effects of
selection acting on genetic diversity. The population is able to "track" the environment as it changes.
However, under circumstances of low genetic diversity where insufficient alleles are available for selec-
tion, adaptation will be slowed and inefficient (Figure 2-3). Instead, the population size will be reduced
due to poor fitness, which reduces genetic diversity and, in turn, reduces fitness. The separate processes
of fitness loss, population loss, and demographic effects feed off each other until the population finally
crashes.
While the concept that adaptation is dependent on genetic diversity is a cornerstone of evolutionary biol-
ogy, there is little empirical evidence that measurements of genetic diversity are correlated with long-term
sustainability or adaptive potential. Difficulty in defining and measuring adaptive potential may partly
explain the present lack of evidence. However, a more fundamental problem regarding inferences made
from measurements of genetic diversity with molecular markers also should be considered. Most molec-
ular markers are selected and analyzed because they show sufficient variation to elucidate patterns of
genetic variation within and among populations. With a few exceptions (e.g., some cases in Table 2-1),
little or nothing is known about the direct influence of these molecular markers on fitness, although the
12
-------
Condition and Sustainability
vast majority is believed to be effectively neutral (unselected). Neutral markers are preferable for sever-
al questions of interest, such as estimation of effective population sizes, levels of connectivity between
populations, and evolutionary distinctness of certain populations. They are problematic, however, when
the desired inference is to levels of genetic diversity at loci that directly influence fitness. These loci are
clearly under the influence of selection and therefore influences from neutral loci may not be accurate.
The issue is still under debate (Vrijenhoek, 1994; Lynch, 1996; Rodriguez-Clark, 1999) but, in general, it
is reasonable to conclude that if genetic diversity is high at neutral molecular markers then genetic diver-
sity at fitness loci also is probably high. If, however, genetic diversity is low at neutral markers then genet-
ic diversity at fitness loci is potentially depressed, but may be sufficient to avoid inbreeding depression
and loss of adaptive potential. A second way to approach the problem of extinction risk is to use genetic
diversity measures as estimates of effective population size. Estimates of the rate at which new genetic
variation in quantitative traits (e.g., fitness) is generated by mutation were used to suggest that populations
should be maintained at a minimum effective population size of 500 individuals for long-term sustain-
ability (Soule, 1980; Franklin, 1980). This number was estimated as the minimum needed to maintain
enough quantitative genetic variation to adapt to future environmental changes. Lande (1995) subse-
quently suggested that the minimum effective size estimate was too small by an order of magnitude. In
addition, at very low effective sizes, populations are at risk of immediate extinction from demographic
effects. Therefore, estimates of genetic diversity at neutral markers should be useful indirect measures of
both short and long- term sustainability because they can be used to estimate effective population sizes.
Some characteristics of genetic diversity that support its use as an indicator of ecological sustainability are
summarized in Table 2-4.
Table 2-4. Rationale for indicator of ecosystem sustainability based on genetic diversity.
• Indicator of long and short-term extinction risk when combined with demographics
• Measure of potential to adapt to modified environments (e.g., susceptibility to novel pathogens)
• Amenable to temporal monitoring, trends analysis
• Can be implemented as a nondestructive measure
2.3 Measurement of Genetic Diversity
A variety of methods have been developed to measure genetic diversity within species. The oldest and
most widely known is to simply measure morphological traits of individuals within populations. For traits
that are determined predominantly by single genes, such as flower color or seed color in Mendel's peas,
this provides a direct and straightforward measure of genetic variation. Fitness however, as well as most
other morphological, physiological, and behavioral traits, is influenced by many genes, by environmental
differences, and by interactions between different alleles, loci, and environments. In order to measure
genetic variation of these traits, it is necessary to first control for environmental variation, a decidedly dif-
ficult task when working with natural populations.
Over the last several decades, a number of molecular markers have been developed to aid in the measure-
ment of genetic diversity (see Avise, 1994). Since the 1970s, innumerable studies have assessed genetic
diversity in the electrophoretic properties of proteins, particularly blood group proteins and enzymes
(allozymes). The advantage of this system of analysis over measurement of morphological characters is
that the marker patterns are almost entirely free of environmental influences and one or, at most, a few
genes determine each pattern. A number of additional techniques emerged in the 1980s and 1990s for
measuring genetic diversity directly at the DNA level. One of the advantages of DNA-based analysis is
13
-------
Genetic Diversity as an
1 I I
Inbreeding depression
Heterozygosity
t Genetic Drift
Inbreeding
+ Adaptation
Environment tracking)
+ Population size
Density
A Replacement
rate
variability
Demographic randomness
Figure 2-3. Simplified diagram of Gilpin and Soule's (1986) Extinction Vortex model. Only the adaptation and
inbreeding vortices are shown. Two other vortices (population size and density) also interact with these vortices to
cause population extinction. Ne: effective population size.
that genetic diversity is measured directly at the most fundamental level, before intracellular modifiers
such as gene transcription, RNA splicing, and protein translation and post-translational processing can
modify the pattern. Another advantage of new DNA-based methods is that essentially all of the genetic
material within organisms is available for analysis. This has allowed researchers to target for analysis par-
ticular types of DNA that are most suited for answering specific questions. For example, mitochondrial
DNA rarely (if ever) recombines, has a relatively high mutation rate, and is usually only transmitted from
the maternal parent to offspring. This makes mitochondrial DNA particularly useful for tracking the
maternal pedigree of populations back through time (e.g., for humans, back to the 'mitochondrial Eve'),
and it makes genetic diversity in mitochondrial DNA particularly susceptible to the effects of small popu-
lation size (genetic bottlenecks). Other types of DNA may be targeted because they have exceptionally
high mutation rates and therefore generate many alleles, because they code for genes with known func-
tions that may respond to particular environmental modifiers, or because they are believed to be selectively
neutral so genetic diversity at these loci more accurately reflects evolutionary relationships between pop-
ulations and species.
14
-------
Condition and Sustainability
The mathematical foundations for measuring and interpreting genetic diversity and population genetic
structure were developed in the early part of the last century. However, progress in understanding genet-
ic processes in populations was limited until the recent explosion of molecular biological methods made
it possible to accurately apply and test genetic theory in the field. Presently, genetic diversity indicators
can be used to assess effective population sizes, past or present population bottlenecks, gene flow and,
under some circumstances, mutation rates and selection. The availability of new types of genetic markers
with different genetic properties, together with inputs from related efforts such as the human genome proj-
ect and DNA forensics, has fueled a new round of mathematical development that is continuing today. As
new methods to measure and interpret genetic diversity are rapidly increasing, the information about past
and present population processes that ultimately can be extracted from analyses of molecular markers is
almost certain to experience phenomenal growth.
15
-------
Genetic Diversity as an Indicator of Ecosystem Condition and Sustain
I I I
3 Case Studies
Two case studies are presented in this section to demonstrate the technical feasibility of incorporating
genetic diversity measurements into an ecological assessment and to help evaluate the environmental rel-
evance of this indicator. Many variables can affect the success of a genetic diversity indicator, ranging
from the quality of the genetic markers employed, to the technical abilities of the people involved, to the
appropriateness of the species surveyed. The two case studies described here utilized different approach-
es with different technical requirements. The first study represents a large regional survey in which a large
number of sites were assessed in order to create a regional genetic diversity profile for a single, moder-
ately sensitive minnow species. The study was based on the RAPD fingerprinting technique and utilized
very few samples per site. The second study included a much smaller number of sites, but evaluated more
fish per site. A moderately tolerant fish was evaluated in this study with both AFLP and mitochondrial
sequence data. Ideally, a genetic diversity assessment will evaluate multiple species over the study region.
In fact, analyses of additional fish species are underway in both study areas; this multi-species assessment
will be presented in a future document. Both case studies demonstrate the utility of integrating genetic
diversity information and other ecological indicator data into a single ecological assessment.
3.1 Genetics of Central Stonerollers in The Eastern Cornbelt Plains Ecoregion
3.1.1 Backgroun d
The Central Stoneroller (Campostoma anomalum). The study organism is a small minnow in the fam-
ily Cyprinidae (Figure 3-1). It is common and relatively abundant in runs or riffles over hard bottom in
clear running streams throughout much of the Eastern and Midwestern USA, exclusive of Southeastern
states. Stonerollers are bottom feeders, feeding primarily on algae and detritus they scrape off of rocks.
Males build nests in gravel pools at the tops of riffles in which females lay between 100 and 500 eggs.
Stonerollers are classified as moderately tolerant, preferring clean water in organically enriched streams
with thick growths of attached algae.
Figure 3-1. The central stoneroller (Campostoma anomalum). (Photo
courtesy of the Ohio Department of Natural Resources)
16
-------
Condition and Sustainability
The Study Area. The Eastern Cornbelt Plains (ECBP) Ecoregion (Omernick, 1995) encompasses most of
central and western Ohio as well as the central and eastern parts of Indiana (Figure 3-2). A finger of the
ecoregion extends into southern Michigan. The region once was composed of tall grass prairie, eastern
deciduous forest, and wetlands. Modern agriculture, industrialization, and urbanization have introduced
both point source (e.g., industrial discharge, waste water treatment plants, combined sewer overflows) and
non-point source (e.g., nutrient enrichment, toxic chemicals, sedimentation) stressors to the network of
streams in this ecoregion. Figure 3-3 provides an aerial view of the typical ECBP landscape. The genet-
ic study concentrated on watersheds in the eastern part of the ECBP, primarily in Ohio. Major watersheds
within this region include the Scioto, Great Miami, and Little Miami, which drain into the Ohio River, and
the Maumee, Sandusky, and Huron watersheds, which drain into Lake Erie.
Integration. This work was undertaken as part of a Regional Environmental Monitoring and Assessment
Program (REMAP) study initiated as a collaborative effort between USEPA Region 5, the Ohio EPA and
USEPA-NERL to answer questions about the overall ecological condition of the ecoregion. Specifically,
the REMAP project's purpose was to assess the stream resource status and to develop biological measure-
ments that would serve as quantitative indicators of the condition of those stream resources, the type and
magnitude of stress placed on the streams, and whether the resource condition changed over time.
Randomly selected sample sites along first through third order (wadeable) streams represented a variety
of land uses and impact sources (Figure 3-4) present in the ecoregion.
Field sampling. Fish were collected by electroshocking (Figure 3-5) using the EMAP protocols for wade-
able streams (USEPA, 1993). Central stonerollers were collected from 91 sites representing 10 large
watersheds in 1995. Sample sizes at each site were small (Table 3-1) because genetic diversity sampling
was not the primary purpose of the REMAP assessment. The caudal peduncle and caudal fin were removed
and placed into a cryovial labeled with the site name, frozen in liquid nitrogen, and shipped back to the
laboratory where the sample was stored at -80°C.
Laboratory methods and data analysis. A brief description of laboratory and analytical procedures is
provided here. We recommend that readers interested in designing a genetic study using this or related
techniques consult the more detailed description in Appendix 1.
DNA was extracted from each sample using standard procedures. For this study, a genetic "fingerprint"
or profile was constructed for each fish using the technique of randomly amplified polymorphic DNA
(RAPD) (Williams et a/., 1990; Welsh and McClelland, 1990). This is a relatively simple technique based
on agarose gel electrophoresis of anonymous PCR (polymerase chain reaction) fragments. A drawback of
the technique is that scoring of gels (detecting and sizing bands) has been found to be less repeatable than
for some other methods (e.g., AFLP, microsatellite analysis). To overcome this problem, bands were
scored semi-automatically using a fluorescence imaging system (Fluorimager 595, Molecular Dynamics)
and dedicated image analysis software (FragmeNT, Molecular Dynamics) (Figure 3-6). Additionally, each
sample was analyzed in triplicate. Bands were classified into size groups (bins) using cluster analysis.
These bins were further refined by discriminant analysis of band size and intensity characteristics. In
excess of 200 such bins were identified; however, 53 bins that were most repeatably scored (based on com-
parison of the three replicates for each individual) were selected for analysis. The genetic profile for each
individual was derived from the composite information from three replicates for these 53 fragment size
bins. These profiles were compared using a "similarity index" (Lynch, 1990; Leonard et a/., 1999).
Average genetic similarities within populations provide an inverse measure of genetic diversity while aver-
age genetic similarity between two populations provides an inverse measure of the "genetic distance"
between them.
17
-------
Genetic Diversity as an Indicator of Ecosystem Condition and Sustain
I I I
V
25 Q 25 SO Milts
ECBP StudySites
RF-1
ECBP
St. Joseph
Maumee
Sandusky
I Huron
Muskingum
Scioto
Great Miami
Little Miami
Wabash
White
Figure 3-2. Map showing sampling locations for the study. Major watersheds are highlighted. Borders between
USGS watersheds (8-digit hydrologic units or HUCs) also are indicated.
18
-------
r
Diversity as an Indicator of Ecosystem Condition and Sustainability
Figure 3-3. Aerial view of typical Eastern Cornbelt Plains landscape.
3.1.2 Key findings and their implications
Populations of stream fish are highly differentiated within the Eastern Cornbelt
Plains Ecoregion. A common statistic used to measure how different or similar populations are with-
in a geographic area is F^ (Section 2.1). Technically, FST estimates the "inbreeding", or increased likeli-
hood that the two copies of a gene within an individual are identical (homozygous), due to population
genetic structure. An estimate of F^ = 0 indicates that there is no evidence for geographic structure from
the samples analyzed and thus they may represent a single population. An estimate of FST =1 indicates
that none of the populations share the same alleles at polymorphic loci. For diploid sexual species, FST
values generally fall between 0 (no genetic structure and 0.3 (large genetic structure), with cases of FST >
0.4 being quite unusual. For stonerollers sampled from these ECBP sites, our estimate of FST was about
0.39 (Table 3-2) so we conclude that stoneroller populations are highly structured within the study region.
A nice feature of the FST statistic is that it can be used to estimate gene flow, or the average number of
migrants between sites each generation. Table 3-2 shows that over the entire study area, an average of just
0.4 individuals move between populations each generation (a generation is about 2-4 years for stonerollers
in this area). Of course, individuals should be much more likely to migrate to nearby streams, rather than
to streams that are more distant. Looking within major watersheds, and for large watersheds within USGS
8-digit hydrologic units (HUCs), we see that the average number of migrants between sites does tend to
be greater in the smaller areas. However, the pattern is not consistent for different watersheds, as some
(e.g., the Scioto) show very little gene flow, on average, while others (e.g., the Wabash) show very high
gene flow.
19
-------
Genetic Diversity as an Indicator of Ecosystem Condition and Sustain
I I I
Figure 3-4. A highly modified suburban stream. Urban run-off, non-native riparian growth,
bank erosion, and combined sewer overflows are common urban and suburban impacts to
streams.
Figure 3-5. Backpack electroshocking for central stonerollers.
20
-------
Condition and Sustainability
Table 3-1. Geographic location of 91 REMAP sites in the Eastern Cornbelt Plains Ecoregion from which central
stonerollers were sampled. The 11-digit USGS hydrological unit code (HUC) for the sub-basin is provided, along
with the REMAP site identification number, stream order classification, ecological impacts identified, Qualitative
Habitat Evaluation Index (QHEI), Index of Biotic Integrity (IBI), and number of stonerollers collected. KEY— RM:
River mile; A: agriculture/livestock; R: riparian degradation, C: channelization; U: urban/suburban/industrial; S: waste
water treatment plant or combined sewer overflow; n.c.: not calculated.
Basin
St. Joseph
Maumee
Sandusky
Huron
Muskingum
Scioto
USGS HUC
04050001 110
04100004010
04100006040
04100007010
04100007010
04100008030
04100008030
04100011 030
04100011 030
04100011 050
04100011 060
04100011 060
04100011 070
04100011 100
04100012010
04100012010
04100012030
04100012040
05040002010
05040003010
05040003010
05040006010
05040006 040
05060001 010
05060001 010
05060001 010
05060001 020
05060001 030
05060001 070
05060001 070
05060001 080
05060001 090
05060001 090
05060001 090
05060001 100
05060001 140
05060001 140
05060001 140
05060001 140
05060001 150
05060001 170
05060001 180
05060001 220
05060002010
Site Name
Pigeon Creek
E. Br. St. Marys R. (Clear Creek)
Beaver Creek
Dry Run
Auglaize River
Eagle Creek
Eagle Creek
Broken Sword Creek
Broken Sword Creek
Trib. to St. James Creek (RM 0.85)
Oak Run
LittleTymochtee Creek
Sycamore Creek
E. Br. Wolf Creek
Marsh Run
Walnut Creek
Trib. to trib. to Huron R.
Chappel Creek
Trib. to Black Fk. Mohican R. (RM 54.45)
Kokosing River
Kokosing River
Otter Fk. Licking R.
Trib. to S. Fk. Licking R.
Taylor Creek
Trib. to Taylor Creek (RM 7.60)
Trib. to Scioto River (RM 227.76)
Rush Creek
Panther Creek
Grassy Run
Trib. to Mill Creek (RM 41 .24)
Eversole Run
Flat Run
Trib. to Olentangy R. (RM 76.95)
Rocky Fork
Shaw Creek
Blacklick Creek
Blacklick Creek
Rocky Fk. Big Walnut
Rocky Fk. Big Walnut
Bunker Run
(none provided)
Lick Run
Hellbranch Run
Yellowbud Creek
Site
ID
399
173
403
268
270
283
284
436
446
280
273
275
448
431
440
443
428
426
435
142
423
145
151
168
172
263
170
277
155
171
123
140
437
438
138
129
130
146
148
144
152
338
126
308
Stream
Order
3
3
2
1
3
3
3
3
3
2
3
3
3
3
3
3
1
2
2
3
2
2
2
2
1
2
3
2
2
3
3
3
1
3
3
3
3
2
2
3
2
2
3
3
Impacts
A
A
A,R,C
A,R,C
A
A
A
A,R,C
A
A,R
A,R,C
A
A
A
A
A
U,R,C
none
A
none
A
R,C
none
A
A
A
A
A
A,C
A
none
none
A,R,C
R
A,C
U,S
U,S
U
A
none
n.c
A
U,S
A
QHEI
80
57
18.5
38
72.5
49.5
49
34
59.5
33
24
52.5
62.5
54.5
51
69
30.5
49
58
42.5
73
26
61
71.5
31
32
67.5
62.5
27
58.5
71.5
79
24.5
53.5
44
63.5
75
67.5
67
68.5
n.c
54
75
80.5
IBI
36
29
26
46
44
34
24
36
32
34
28
20
42
28
34
46
28
36
40
44
53
36
38
36
28
36
37
40
26
40
48
43
30
32
38
48
38
48
36
54
n.c
46
45
56
Sample
Size
10
10
9
10
10
10
10
7
10
9
10
8
10
10
10
10
9
10
9
10
10
10
10
10
3
9
9
10
10
9
10
9
8
10
10
10
6
3
8
10
10
10
18
20
(continues on next page)
21
-------
Genetic Diversity as an
Ill I I
Table 3-1. (Continued)
Basin
Great Miami
Little Miami
Wabash
White
USGS HUC
05060002010
05060002010
05060002010
05060002 030
05060002 030
05060002 040
05060003 020
05060003 040
05060003 040
05060003 050
05060003 050
05060003 060
05060003 090
05080001 010
05080001 020
05080001 030
05080001 030
05080001 040
05080001 150
05080001 170
05080002 030
05080002 040
05080002 040
05080002 040
05080002 040
05080002 050
05080002 060
05080002 060
05080002 070
05080002 080
05080002 080
05080003 080
05090202010
05090202010
05090202010
05090202 020
05090202 020
05090202 040
05090202 050
05090202 080
05090202 080
05090202 1 00
05120101 010
05120101 010
05120104070
05120107010
05120204020
Site Name
Hargus Creek
Scippo Creek
Congo Creek
Opossum Run
Bradford Creek
Hay Run
Sugar Creek
Bull Creek
Hardin Creek
Paint Creek
Sugar Run
Clear Creek
Trib. to Little Creek (RM4.22)
Trib. to S. Fk. Great Miami R. (RM 5.27)
Willow Creek
Brandywine Creek
Blue Jacket Creek
Little Indian Creek
Kings Creek
E. Fk. Buck Creek
Twin Creek
Trib. to Twin Creek (RM 18.29)
Toms Run
Bantas Fork
Toms Run
Millers Creek
Sevenmile Creek
Paint Creek
Fourmile Creek
Indian Creek
Indian Creek
Trib. to Dry Fk. Whitewater R. (RM 6.73)
Little Miami River
Trib. to N. Fk. L. Miami R. (RM 5.60)
Lisbon Fork
Clark Run
Little Beaver Creek
Painters Creek
N. Br. Caesar Creek
E. Fk. Todd Fork
E. Fk. Todd Fork
Turtle Creek
(none provided)
Wabash River
W Br Twelvemile Cr.
Middle Fork Creek
Sixmile Creek
Site
ID
341
342
343
125
294
309
292
302
330
305
306
329
307
160
271
162
163
236
166
217
248
254
255
258
259
324
256
257
1
353
355
358
220
221
223
224
253
320
297
313
314
318
177
178
87
100
35
Stream
Order
3
3
3
2
3
3
3
2
2
3
2
1
2
3
2
2
2
2
3
2
2
2
2
3
2
2
3
3
3
3
3
1
3
2
2
2
2
2
3
3
3
3
3
3
2
1
2
Impacts
u,s
A,S
A
R
A
A
A,R
none
A
A
A
A
none
A
A,R,C
A,R,C
A
A
A
A,R,C
A
A
A
A
none
none
A
A
A,R,C
S
none
none
A
A,R,C
A,R
A
U,S
A
A,R
A
A
S
n.c
A
A
A,C
None
QHEI
58
73.5
82
72.75
84.5
63.5
43
76
71
76.5
62
50
53
76.5
26
30.5
32
75
58
49
59
54
62
75
82
65.5
73.5
74
29
71.5
80.5
68.5
75.5
23.5
54
58
91
71.5
49
60
54.5
58.5
n.c
66
53.5
58
60
IBI
36
54
54
47
55
48
50
54
50
37
48
40
44
52
46
38
20
50
38
40
48
50
n.c
42
42
46
46
42
24
45
40
46
38
28
46
38
38
46
44
50
46
54
n.c.
31
31
40
48
Sample
Size
10
10
10
10
9
10
10
10
10
3
10
7
10
9
10
9
8
10
10
19
10
7
10
10
10
10
9
9
8
7
9
10
10
4
10
10
10
10
7
10
10
9
10
9
10
10
5
22
-------
r
Diversity as an Indicator of Ecosystem Condition and Sustainability
Populations of stream fish are geographically structured among and within water-
sheds. Another way to look at these data is to examine the patterns of relatedness between each of the
sampling sites. One can think of the populations as connected in a phylogenetic tree-like pattern, with the
caveat that, unlike relationships between different species, the patterns of relationships are determined by
both evolutionary descent and ongoing gene flow. Figure 3-7 shows a tree diagram constructed from a
matrix of genetic distance between each of the stoneroller collection sites using the neighbor joining
method. Populations in the same basin are colored similarly. The genetic distance between any two pop-
ulations is represented as the total length of the branch between them. Two things are immediately obvi-
ous from the figure. First, groups of populations in the same watershed tend to cluster together on the tree
(genetically similar). Second, some watersheds have more than one cluster of genetically related popula-
tions while some clusters include more than one watershed. For example, the Scioto basin (sites colored
green in Figure 3-7) is clearly separated into two highly distinct genetic groups. Examination of the geo-
graphic location of sites (Figure 3-2) indicates that the smaller genetic group is composed of populations
located in the extreme northern part of the basin. All are from a single USGS watershed (HUC 50600001),
but other populations from this same watershed are from the second genetic group. In contrast, popula-
tions in the Little Miami (sites shown in blue) and Great Miami (sites shown in purple) watersheds are not
highly distinct from each other.
Another way to visualize genetic relationships is with multidimensional scaling (MDS). In MDS analy-
sis, the relationships between each population and all other populations are represented in two dimensions
in order to more easily observe patterns. In Figure 3-8, the separation between the two genetic groups in
the Scioto basin is evident, as is the lack of strong separation between the Little Miami and Great Miami
Figure 3-6. Interpretation of marker patterns is aided by gel image analysis software
23
-------
Genetic Diversity as an Indicator
1 1 1
basins. However, it also is clear that many of the populations from streams in the Erie drainage are genet-
ically very similar, and are not strongly differentiated from streams in the Wabash basin. Based on this
genetic analysis, we can define at least 5 large genetic groups within the ECBP ecoregion.
The major significance of this analysis is that understanding the genetic structure of stream fish popula-
tions is requisite to drawing conclusions about the status of those species and, by extension, the habitats
that support them. Populations within each of the five genetic groups are likely to have some degree of
genetic interdependence among them, either due to ongoing gene flow or recent separation. These popu-
lations are likely to share common genetic traits (e.g, susceptibilities, behaviors) and have some degree of
demographic connectivity. Populations in different genetic groups, however, are likely to be highly inde-
pendent, demographically, genetically, and evolutionarily. These genetic groups, therefore, are appropri-
ate units for performing ecological analyses if the assessment question concerns the current and future sta-
tus of stoneroller populations. If central stoneroller population structure parallels that of other minnow
species, the geographic boundaries between these genetic groups delineate appropriate assessment units
for minnows in general. It should be noted that the boundaries of these genetic groups do not compare
well to watersheds defined by 8-digit USGS hydrologic unit codes (HUCs), although these are the typical
assessment units in current use.
A useful
method for measuring genetic diversity with DNA fingerprint data is to estimate the average genetic sim-
ilarity of individuals within populations (Sw; Lynch, 1990). This is an inverse measure of genetic diver-
sity, so populations with high genetic similarity are less diverse. Average genetic similarity within these
populations ranged from 58% to 84%. A somewhat surprising result of this analysis was that genetic diver-
sity within populations differed greatly among the five genetic groups (Figure 3-9). Populations in the
Great Miami and Little Miami basins (genetic groups 1 and 2) were slightly more genetically diverse (less
average genetic similarity within populations) than populations in the lower Scioto Basin (genetic group
3) and were much more diverse than populations in the Erie drainage (genetic group 5). These differences
in genetic diversity at geographically broad scales are not likely to reflect different anthropogenic impacts.
Instead, they probably reflect evolutionary differences in the amount of diversity within these groups. The
general trend of decreasing diversity with increasing latitude may reflect residual effects of the Wisconsin
glaciation, which would have wiped out fish populations over most of the study area. Regardless of how
the differences arose, these results suggest that the future sustainability of populations in the Little Miami
and Great Miami basins is much less in question than the sustainability of Erie drainage populations.
Table 3-2. Summary of population genetic structure analysis.
Geographic grouping
Overall
Maumee
Sandusky
Huron
Muskingum
Scioto
HUC 5060001
HUC 5060002
HUC 5060003
Little Miami
Great Miami
HUC 5080001
HUC 5080002
Wabash
Populations
91
6
7
4
5
34
20
7
7
10
19
7
11
4
FST
0.388
0.046
0.033
0.026
0.298
0.325
0.360
0.065
0.096
0.148
0.110
0.077
0.120
0.006
Average number of
migrants per generation
(Nem)
0.4
5.2
7.3
9.4
0.6
0.5
0.4
3.6
2.4
1.4
2.0
3.0
1.8
41.4
Significance
level
< 0.001
0.001
0.16
0.05
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
0.31
24
-------
I~TT
Genetic Diversity as an Indicator of Ecosystem Condition and Sustainability
140. , 306
123
263
172
. 438
342
152
305
178
446
273
280
448
275
436
341
130
338
Figure 3-7. Tree diagram depicting the estimated genetic relationships between populations at each site. This is an
unrooted tree which was made by employing the neighbor-joining algorithm to impose a tree structure on a matrix of
Nei's genetic distances between each site. Site IDs are color coded to match the map in Figure 3.2.
25
-------
Genetic Diversity as an Indicator of Ecosystem Condition and Sustain
I I I
I * $ * $ f
E E E
_:• ^ _:•
d1 O CJI
E E E -e
•3 5 J3 §
X :- ^k
2 ° ° Si
O O LJ *^-
w) w v) Q
OQOaaaQQQQOQQOQQIN — '-'-
OOOOOOOOOOOOOOOOOOOQ
1O O O O O Q O ^" "^J" ''J' ^C* 4O ^D "^O "90 "DO •OT iN CN CN
c^ i— ^~ i— T~ ^^* T"~ CT- o CTI c^> c^ >r^ -TTp ^r^ r^ r~i T~ i— ^~
^r «r ^y *r sr "^y ^3f u^ uft lA u^ ^^ ^b1 in I" u? >un ift irt irt
OOOOQOOQOQOOOOOOODOQ
o
a
26
-------
I~TT
Genetic Diversity as an Indicator of Ecosystem Condition and Sustainability
A
A
AB C
I
1
I
>W
Genetic Group
Figure 3-9. Box and whisker plot of genetic similarity within populations (Sw) for each of the five genetic groups
identified in Figure 3-8. Groups with low average genetic similarity have higher genetic diversity. Groups with differ-
ent letters have significantly different levels of genetic diversity (p < 0.05, Duncan's multiple range test).
Genetic diversity within populations is affected by environmental condition.
Correlations between the average genetic similarity within populations and various environmental meas-
ures are presented in Table 3-3. The environmental measures include the Qualitative Habitat Evaluation
Index (QHEI) and 8 metrics that are used to create the index. Other measures include tissue metabolite
assays for benzo(a)pyrene (BAP)-type compounds (generally associated with combustion by-products)
and naphthalene (NAPH)-type compounds (associated with oil contamination), and EROD assays which
measure enzyme activity induced by planar xenobiotics such as halogenated hydrocarbons and polycyclic
aromatic hydrocarbons (PAHs). The raw correlations were generally small. Among the QHEI metrics.
only the overall QHEI index and the cover and pool metrics were significant, although all correlations with
genetic similarity were negative, as would be expected (larger values of the metrics indicate better envi-
ronmental quality).
27
-------
Genetic Diversity as an
Ill I I
Table 3-3. Spearman (rank) correlations between average genetic similarity within populations (Sw) and various envi-
ronmental measures.
Environmental Variable
Qualitative Habitat Evaluation Index
Substrate QHEI metric
Cover QH El metric
Channel QHEI metric
Riparian QHEI metric
Pool QHEI metric
Riffle QHEI metric
Gradient QHEI metric
Gradient (m/km)
Average sampling depth (cm)
Latitude
Longitude
Stream Order (first-third)
Watershed Area (hectares)
Watershed perimeter (km)
Elevation (meters)
BAP (ug/mg protein)
NAPH (ug/mg protein)
EROD (pmol/min/mg protein)
N1
84
84
84
84
84
84
84
84
84
83
86
86
86
86
86
86
78
78
76
Mean
58.33
12.20
11.53
13.02
5.21
6.16
2.50
7.71
3.02
34.58
40.17
-83.63
2.47
8361
54.62
279.86
0.15
27.77
6.96
Std. Dev
16.85
5.31
3.99
4.27
1.96
2.64
1.95
2.20
2.98
17.96
0.62
0.77
0.63
3844
16.82
40.09
0.11
10.19
7.84
Spearman
Correlation
-0.229*
-0.166
-0.216*
-0.160
-0.022
-0.269*
-0.135
-0.121
-0.101
-0.302**
0.492***
0.268*
0.088
-0.050
-0.050
-0.170
-0.088
0.104
0.070
*p<0.05 **p<0.01 ***p< 0.001
Sample sizes vary because not all environmental measures were available for all sites.
As an exploratory analysis, variables that suggested a strong association with genetic similarity in simple
correlation analyses were included in a linear model to examine their explanatory power in determining
average genetic similarity [Sw was transformed (SW/(1-SW) to approximate normality]. Both quantitative
variables (Table 3-3) and categorical variables (genetic group and the impact factors from Table 3-1) that
were significant at the p < 0.1 level were included in the initial model. The least significant effects were
eliminated in a stepwise fashion until only effects that were significant at the p < 0.05 level were left in
the final model.
The final model (Table 3-4) was highly significant (F8 74 = 10.8, p < 0.0001) and explained about half of
the variation in Sw (R2 = 0.53). Significant effects in the final model included three impact factors: Urban,
Riparian, and Channelization, along with genetic group and average stream depth at the sampling location.
Latitude, though highly significant in the correlations analysis, was not significant in the linear model
because it was strongly associated with genetic group. It is interesting that the riparian and channeliza-
tion impact factors were significant while the QHEI riparian and channel metrics were not. The explana-
tion for this may be a nonlinear response of genetic diversity to habitat degradation, as only highly affect-
ed sites were classified as having impacts to the riparian zone or channelized habitat. Stream depth may
be significant as it is an indicator of stream size, and thus the carrying capacity of the habitat.
28
-------
nnr
Condition and Sustainability
Table 3-4. Final model explaining genetic similarity within populations following stepwise elimination of non-signifi-
cant effects.
Source
Impact factor- Urban
Impact factor - Riparian
Impact factor- Channelization
Major Genetic Group
Depth (covariate)
Degrees of
Freedom
1
1
1
4
1
Mean Square
5.369
4.882
2.452
6.311
1.974
F Value
11.20
10.19
5.12
13.17
4.12
Probability
0.001
0.002
0.027
< 0.001
0.046
We can use these relationships to derive a more precise ecological indicator based on genetic diversity.
Specifically, the deviation of the estimated genetic diversity within each site from its predicted value based
on historical biogeographic factors (genetic group) and habitat carrying capacity (stream depth) indicates
the relative genetic health of the population. Figure 3-10 shows a representation of this type of analysis,
in which the corrected genetic diversity at each site is classified as low, moderate, or high. As with any
ecological indicator, these results should be interpreted cautiously and only in relation to other indicators
of ecological condition.
is
dition, In addition to indicators that might directly affect genetic diversity, other environmental variables
were measured that are not expected to affect genetic diversity directly, but might be similarly affected. If
measures of genetic diversity are redundant with other indicators, then direct measures of genetic diversi-
ty may be unnecessary. Table 3-5 demonstrates that genetic similarity within central stoneroller popula-
tions is significantly correlated with two common indicators of environmental condition, the Index of Well
Being and Index of Biotic Integrity, but the correlation coefficients are small. Genetic similarity also is
correlated with several other environmental measures, however these correlation coefficients also are
small. We interpret these results to signify that the genetic similarity measure for central stonerollers is
related to but not highly redundant with these other indicators of ecological condition.
29
-------
Genetic Diversity as an Indicator of Ecosystem Condition and Sustain
I I I
• High genetic divecisty
Moderate genetk dtvernty
Not measured
Figure 3-10. Map depicting the relative genetic diversity at each site after correcting for genetic group and average
stream depth at the sampling site. Green: highest one third of sites; yellow: middle third of sites; red: lowest third
of sites
30
-------
Condition and Sustainability
Table 3-5. Spearman rank correlations of condition indicators with average genetic similarity.
Variable
Number of fish (per 300 m)
Weight of fish (kg/300 m)
Total fish species
Shannon Diversity Index (on number)
Shannon Diversity Index (on weight)
Index of Well Being
Index of Biotic Integrity
Number of sunfish species
Number of darter species
Number of sucker species
Number of Cyprinid minnows
Number of intolerant spp.
Percent tolerant fish
Percent top carnivores
Percent omnivores
Percent fish with anomalies
Number of sensitive or
moderately sensitive species
Number of sensitive species
Percent simple lithophils
Number simple lithophils
N
83
83
83
83
83
83
83
83
83
83
83
83
83
83
83
83
83
83
83
83
Mean Std.Dev. Spearman
Correlation
1454.00
14.73
17.88
1.93
1.79
8.32
40.71
2.40
3.57
1.98
7.03
0.76
45.79
1.72
18.10
0.28
822.59
4.31
27.63
6.34
1108.00
15.94
6.13
0.35
0.35
1.12
8.58
1.35
1.80
1.20
2.13
1.18
20.81
2.37
13.86
0.65
832.45
3.47
14.98
2.94
-0.235*
-0.243*
-0.310**
-0.025
-0.078
-0.228*
-0.287**
-0.168
-0.163
-0.270*
-0.321**
-0.130
0.137
-0.139
-0.096
0.049
-0.260*
-0.306**
-0.047
-0.302**
:p<0.05
p<0.01
p< 0.001
31
-------
Genetic Diversity as an Indicator of Ecosystem Condition and Sustain
I I I
3.2 Genetics of Creek Chubs in a Mining-Impacted Region
3.2.1 Backgroun d
The Creek Chub (Semotilus atromaculatus). Like the central stoneroller, this organism is a minnow in
the family Cyprinidae (Figure 3-11). It is common in small to moderate size hard bottom streams from
the Great Plains to the Atlantic coast. Creek chubs are omnivorous, eating primarily insects and plant
material. Some adults can reach substantial size (250-300 mm) and become significant fish predators.
Males build and defend nests in gravel beds where there is moderate current. Females lay a clutch of
approximately 25-50 eggs in the nest, which is covered with gravel by the male but not defended. Creek
chubs are fairly tolerant of turbidity and pollution.
Figure 3-11. The creek chub (Semotilus atromaculatus) (Photo courtesy of the Ohio
Department of Natural Resources)
The Study Area. The EPA initiated a regional demonstration project in the mid 1990s to evaluate the
condition of wadeable streams in the Middle Atlantic region of the United States. For this pilot study, we
focused on a smaller region within Pennsylvania and West Virginia underlain by coal-bearing geology and
for which coal mining operations are historically known. Five watersheds, as determined by USGS 8-digit
hydrologic units (HUCs) were represented: Upper Ohio, Connoquenessing, Upper Ohio-Wheeling, Little
Muskigum-Middle Island, and Little Kanawha (Figure 3-12); all are within the Western Allegheny Plateau
ecoregion. While this region is primarily forested, between 14 and 28% of the stream length in each of
the watersheds studied in this pilot has agricultural land cover in the riparian zone (USEPA, 1997).
Integration. This study was performed in conjunction with the EPA's Mid-Atlantic Integrated Assessment
(MAIA) project, under the Environmental Monitoring and Assessment Program (EMAP). EMAP is a
research program charged with developing the tools necessary to monitor and assess the status and trends
of national ecological resources and typically employs a range of ecological indicators in order achieve
this purpose.
32
-------
r
Diversity as an Indicator of Ecosystem Condition and Sustainability
• Sample Silts
Rf-l
States
K ffigi' III I
l)SH3flKI5
(HCGOIQI
y
97-074
9S-I32
97-UI6
WV
¥74*24
98-133
97-023
98-036
» II » 4U
Figure 3-12. Map indicating the location of sample sites within the five watersheds studied. The USGS 8-digit
hydrologic units are indicated. These watersheds are (from North to South): Upper Ohio, Connoquenessing, Upper
Ohio-Wheeling, Little Muskigum-Middle Island, and Little Kanawha. All sites fall within Pennsylvania and West
Virginia.
Field sampling. Fish were collected using the EMAP protocols for wadeable streams (USEPA, 1993).
Sites were chosen probabilistically in order to establish the overall condition of first through third order
streams in the area. Measurements of physical habitat, water and sediment chemistry, and biotic assem-
blage structure were used to develop quantitative indicators of the condition of stream resources and the
types and magnitudes of stresses placed on streams. Samples of between 9-28 creek chubs were collect-
ed in 1997-98 from 10 sites within the five watersheds (Table 3-6).
33
-------
Genetic Diversity as an Indicator of Ecosystem Condition and Sustain
I I I
Laboratory methods and data analysis. A brief description of laboratory and analytical procedures is
provided here. We recommend that readers interested in designing a genetic study using this or related
techniques consult the more detailed description in Appendices 2 and 3.
Total genomic DNA was extracted from the caudal fin of each sample using standard procedures. For this
case study, two analytical techniques were used: a genetic "fingerprint" was constructed for each fish
based on its nuclear DNA characteristics and a portion of its mitochondrial DNA was sequenced. The
DNA fingerprint was produced using the technique of amplified fragment length polymorphism (AFLP;
Vos et al. 1995). This is a relatively simple technique based on gel electrophoresis of anonymous PCR
fragments. Twenty percent of the creek chub samples were analyzed in duplicate. The genetic profile for
each individual was derived using the presence or absence of 109 polymorphic AFLP markers. As in the
Figure 3-13. Manual gel loading of samples in an acrylamide gel for electrophoresis.
first case study, these genetic profiles were compared using a similarity index approach (Lynch, 1990;
Leonard et al., 1999). A 590 base pair (bp) portion of the cytochrome B gene was sequenced in order to
provide a measure of genetic diversity in the mitochondrial genome.
To provide an ecological context for the genetic diversity patterns, a set of 25 environmental measures also
were examined. These environmental factors were analyzed and reduced to six independent "factors" that
explained most of the environmental differences between sites using principal components analysis.
Reducing the number of environmental variables and interdependence among them is very helpful in
understanding the difference between sites and performing statistical analyses. Relationships between
genetic diversity and these environmental factors were explored with multiple regression analysis.
34
-------
r
Diversity as an Indicator of Ecosystem Condition and Sustainability
—_
Figure 3-14. An AFLP fingerprint gel. The first lane and every 8th lane following it contains a 10 bp ladder to aid in
sizing the fragments. Lanes between these ladders contain AFLP fingerprints for different creek chub samples.
Polymorphisms are scored as the presence or absence of individual fragments in each lane.
3.2.2 Key Findings and their Implications
Mitochondrial DNA differences between creek chubs have a strong spatial compo-
nent. Among the DNA samples sequenced, 27 unique types of mitochondrial DNA (mitochondrial hap-
lotypes) were identified. By inspecting the pattern of mutational differences between these different
forms, a network or phylogenetic tree can be constructed that shows how one mitochondrial form has
evolved into another through time. One way to represent this data visually is in the form of a minimum
spanning network (Figure 3-15). In the figure, each colored circle represents one of the mitochondrial
DNA haplotypes identified. The frequency of the mitochondrial type in the overall sample is proportion-
al to the area of the circle, and its spatial distribution is represented by different colors. Lines connecting
circles represent mutations that differentiate different mitochondrial haplotypes; black dots represent hap-
lotypes that are inferred to exist (or have existed) but were not observed. The figure shows that two com-
mon, presumably ancestral haplotypes were identified, along with many related but infrequent haplotypes.
The five populations from the four northern watersheds have mitochondria that are from one group, while
the five populations from the Little Kanawha watershed (HUC 05030203) were predominantly of the other
35
-------
Genetic Diversity as an Indicator of Ecosystem Condition and Sustainabilit
I I I
Table 3-6. Geographic location of 10 sites within the Western Allegheny Plateau ecoregion from which creek chubs
were sampled. The name and 8-digit USGS hydrological unit code (HUC) for each watershed is provided, along with
the individual stream/site name, stream ID, stream order classification, and number of creek chubs used in nuclear
(AFLP) and mitochondrial genome analysis. Colors correspond to those of the watersheds depicted in Figure 3-12.
us
Watershed H
MAIA
« «.»*.
Stream
Order
Sample Size
for AFLP
analysis
Sample Size
for
mitochondrial
analysis
Upper Ohio
Connoquenes
sing
Upper Ohio-
Wheeling
Little
Muskigum-
Middle Island
05030105
05030101
05030106
05030106
Taylor Run
Millers Run
98-134
97-073
Short Run Creek 97-074
Upper Bowman Run 98-132
27
14
14
28
27
14
14
25
05030201 Little Fishing Creek 97-016
05030203
05030203
05030203
05030203
Neal Run
Spruce Creek
Walnut Fork
Oil Creek
Left Fork Steer
Creek
97-008
97-024
98-133
97-023
98-036
13
14
27
13
15
13
13
26
13
13
mitochondrial group. It is interesting to evaluate this geographically structured genetic pattern (or "phy-
logeographic" pattern) in relation to the geologic history of this area. Prior to the Pleistocene glaciations.
streams in the four northern watersheds drained into the Great Lakes/Laurentian basin while streams in the
Little Kanawha watershed drained into the Mississipi River via the ancestral Teays River. Current patterns
of mitochondrial genetic differentiation appear to reflect this historical structure.
A useful way to evaluate the geographic structure of mitochondrial DNA difference is with an analysis of
molecular variance (AMOVA, Excoffier et ctl, 1992), which is analogous to the common analysis of vari-
ance. Using this method, we can determine the proportion of all major variation that is due to differences
among regional groups (the northern four watersheds vs. the southernmost watershed), differences among
populations within watersheds, and differences among individuals within populations. The results of this
analysis indicate that 65% of all mitochondrial variability is associated with the regional differences, while
7% is associated with differences among populations within watersheds. Just 28% of mitochondrial vari-
ability occurs among individuals within populations (Figure 3-16). This information can be used to derive
an estimate ofFST of 0.72, indicating that mitochondrial haplotypes of creek chubs are very highly differ-
entiated in the study area.
36
-------
r
Diversity as an Indicator of Ecosystem Condition and Sustainability
Creek chub populations are strongly differentiated in the nuclear genome, though
the genetic structure is not as strong as for mitochondrial DNA. As was shown for
stoneroller DNA fingerprints in the first case study, the patterns of genetic similarities within and among
populations can be converted into genetic distances and subjected to a standard cluster analysis to reveal
the evolutionary relationships among populations. Figure 3-17 shows the pattern of evolutionary rela-
tionships among these ten populations. The length of the branch between any two populations is roughly
Figure 3-15. Minimum spanning network for 27 mitochondrial DNA haplotypes identified by sequencing a portion the
cytochrome B gene. Colors correspond to those of the watersheds depicted in Figure 3-12. Circle area is proportional
to the number of samples possessing that haplotype (the actual number is indicated if more than one. The letter "V"
indicates a transversion-type mutation. Black nodes represent hypothetical haplotypes not identified in the sample.
proportional to the genetic distance between them. The basic pattern observed for this nuclear DNA analy-
sis is similar to that observed for mitochondrial DNA: populations in the Little Kanawha watershed appear
to be genetically differentiated from the four northern watersheds. In addition, there is some suggestion
from the network diagram of a north-south cline.
An analysis of molecular variance can be constructed for the nuclear AFLP data, similar to the procedure
for mitochondrial data. The results of the nuclear DNA analysis are strikingly different than for the mito-
37
-------
Genetic Diversity as an Indicator of Ecosystem Condition and Sustainabilit
I I I
chondrial DNA. Just 5% of nuclear DNA variation is due to differences between the southern watershed
and the four northern watersheds. While differences among populations within each of these groups
accounts for 8% of the variation, similar to mitochondrial DNA, some 87% of all nuclear DNA variation
exists within populations (Figure 3-18). In contrast to the large mitochondrial FST statistic of 0.72, FST
calculated for nuclear AFLPs is just 0.13. This translates into estimates of effective migration rates
between populations of 0.2 based on mitochondrial DNA, and 1.7 based on nuclear DNA. Although we
cannot be sure of the cause of this discrepancy, it should be noted that the estimate based on mitochondr-
ial DNA tracks female gene flow (since mitochondria are maternally inherited) while nuclear DNA tracks
both male and female gene flow.
Among
Regions
64.8%
Within
Populations
27.8%
Among
Populations
within
Regions
7.4%
Figure 3-16. Distribution of mitochondrial DNA variation, as indicated by an analysis of molecular variance (AMOVA).
For this analysis, the four northernmost watersheds were considered one region while the southernmost watershed
was considered a second region. Each of the variance components was significantly greater than zero (p < 0.0001).
Environmental factors account for about half of the differences in mtDNA diversi-
ty within sites, and virtually all of the differences in nuclear (AFLP) diversity with-
in Sites. Measures of genetic diversity within each population, along with several environmental meas-
ures recorded for each site, are provided in Table 3-7. AFLP diversity was actually measured as average
pairwise similarity of individuals within the population (an inverse measure of genetic diversity), as was
done for RAPD fingerprint data in the first case study. Mitochondrial genetic diversity was measured as
the average number of mutational differences between two mitochondrial haplotypes in the population.
Six independent environmental "factors" (as determined by principal components analysis and varimax
rotation) accounted for about 98% of all the environmental variance in the 25 measurements taken.
Different types of measurements were related and were associated with different factors (Table 3-8).
Factor 1, for example, included many geochemical measurements (aluminum, calcium, chloride, etc)
while factor 4 was associated with spatial scale measurements (size of watershed, stream width and depth).
The relationship between mitochondrial genetic diversity (haplotype diversity) and these six environmen-
tal factors was explored by multiple regression analysis (forward selection model). Factor 2, which was
38
-------
r
Diversity as an Indicator of Ecosystem Condition and Sustainability
0.01
Figure 3-17. An unrooted neighbor joining network based on Nei's genetic distances (see Lynch, 1991), as estimat-
ed from AFLP polymorphisms. Populations in the same basin are colored similarly and correspond to the watersheds
depicted in Figure 3-12.
associated with environmental variables nitrate, total nitrogen, total phosphorous, and organic carbon, was
the only factor that was found to significantly explain mitochondrial diversity within populations. This
factor explained about 50% of the differences in mitochondrial diversity (R2 = 0.51, F = 8.4, p = 0.02).
Three environmental factors were found to significantly affect nuclear (AFLP) diversity within popula-
tions. Latitude and variables that were highly confounded with latitude, which included channel slope,
elevation, silica and zinc concentrations, (Factor 3) explains about 43% of the differences in AFLP diver-
sity (Table 3-9). Nitrogen, phosphorous, and organic carbon, (Factor 2) which was important in explain-
39
-------
Genetic Diversity as an Indicator of Ecosystem Condition and Sustainabilit
I I I
ing mitochondrial diversity, explains an additional 35% of AFLP diversity within sites. Finally, stream pH
and ammonium ions (Factor 5) explain an additional 18% of AFLP diversity. Together these environ-
mental factors explained almost all of the differences in AFLP diversity among sites (R2 = 0.98). Since
latitude and variables highly correlated to latitude accounted for 43% of differences, we conclude that at
least 53% of differences among sites in AFLP diversity are explained by non-geographic environmental
factors. Thus, it appears that both mitochondrial and nuclear genetic diversity of creek chubs in this area
are responsive to stream condition.
It should be noted that we hypothesized that geochemistry would have the strongest effect on creek chub
genetic diversity, since it represented the greatest environmental gradient and should be associated with
mining impacts. However, no association was observed between geochemistry and genetic diversity.
Interestingly, factor 2, which represented nitrogen/nitrate/phosphorus/organic carbon, was significantly
associated with both mitochondrial and nuclear genetic diversity. This may suggest that agricultural inputs
have a greater impact on stream condition in this area than mining inputs.
Among
Regions
64.8%
Within
Populations
27.8%
Among
Populations
within
Regions
7.4%
Figure 3-18. Distribution of nuclear (AFLP) DMA variation, as indicated by an Analysis of Molecular Variance
(AMOVA). For this analysis, the four northernmost watersheds were considered one region while the southernmost
watershed was considered a second region. Each of the variance components was significantly greater than zero
(p< 0.0001).
3.3 Generalizations from the Case Studies
Though quite different in scope, geographic area, and in the nature of the organisms examined, the two
case studies reveal several commonalities. First, the population structure of stream fishes in the Eastern
United States can be quite strong, even for very common, ubiquitous species. This structure is partly a
consequence of natural barriers to migration within stream networks, but the data indicates that it cannot
be entirely explained by watershed boundaries. In fact, defining the USGS watersheds (8-digit hydrolog-
ic units) as the basic assessment unit would appear to be misleading for both species studied. Future eco-
logical assessments that incorporate the population structure of stream fishes when measuring ecological
condition will result in more precise estimates of the condition offish populations.
40
-------
I~TT
Genetic Diversity as an Indicator of Ecosystem Condition and Sustainability
Table 3-7. AFLP and mtDNA diversity (mean ± S.D.) and some key environmental measures for 10 sites sampled.
Colors correspond to the 8-digit HUCs as depicted in Figure 3-12.
Stream
ID
98-134
97-073
97-074
98-132
97-016
97-008
97-024
98-133
97-023
98-036
AFLP
within population
similarity
0.828
±0.025
0.899
±0.007
0.882
±0.006
0.879
±0.005
0.868
±0.008
0.845
± 0.022
0.898
±0.006
0.899
±0.005
0.907
±0.009
0.911
±0.005
mtDNA
pair wise
differences
0.507
± 0.443
0.407
±0.398
1.000
± 0.722
0.581
± 0.484
1.643
±1.078
2.538
±1.458
1.198
±0.814
0.548
± 0.466
0.000
±0.000
0.308
±0.338
Latitude
(decimal
degrees)
41.054
40.347
40.160
39.786
39.606
39.254
39.073
39.054
38.870
38.763
Conductivity
(US)
672
1192
3590
308
213
339
153
162
188
158
Sulfate
(|leq/L)
2849
9222
28569
807
447
616
316
297
233
176
Total
Nitrogen
(Hg/L)
741
624
243
273
195
2011
223
111
260
368
Watershed
Area
(km*)
28
57
49
11
98
3
31
9
70
28
pH
8.31
6.61
8.31
8.01
7.77
7.69
7.73
7.48
7.44
7.50
In addition to strong genetic differences between populations, both assessments also found genetic diver-
sity within populations that varied in amount among populations. Both studies found strong associations
between the amount of this genetic diversity and various measures of environmental condition.
Unfortunately, the types of environmental information collected in the two studies were different so it is
not possible to generalize the types of environmental factors most likely to influence levels of genetic
diversity. However, the correlations that were observed should prove useful in focusing research to diag-
nose particular stressors that may impact fish populations. It is important to note that these inferences were
made for fish populations that appeared outwardly healthy, as the species were numerically abundant at
each of the sites measured. However, the genetic data revealed a scarcity of alleles in some populations,
presumably a consequence of differential environmental quality, that suggest concerns about future sus-
tainability.
41
-------
Genetic Diversity as an
Ill I I
Table 3-8. Environmental factors revealed by principal components analysis and the percent of total environmental
variation attributed to each factor. The mean and range of values for each variable within the factors are presented.
Factor 1
(Geochemistry)
(37.43%)
Conductivity
698 uS
(1 53-3590)
Aluminum
27 ug/l
(5-1 84)
Calcium
2,628 ueq/l
(940-7,394)
Chloride
1 ,044 ueq/l
(85-6,471)
Potassium
80 ueq/l
(39-179)
Magnesium
1 ,366 ueq/l
(341-6,066)
Sodium
3,669 ueq/l
(192-27,665)
Sulfate
4,353 ueq/l
(176-28,569)
Factor 2
(N/P/C)
(24.33%)
Nitrate
15 ueq/l
(1-81)
Total Nitrogen
505 ug/l
(111-2,011)
Total
Phosphorus
63 ug/l
(5-470)
Organic Carbon
80 mg/l
(39-1 79)
Factor 3
(Latitudinal
dines)
(14.14%)
Latitude
39.6°
(38.76-41 .05)
Elevation
274m
(240-350)
Channel Slope
1.0%
(0.10-2.10)
Silica
6 mg/l
(3-13)
Zinc
0.01 mg/l
(0.00-0.04)
Factor 4
(Spatial Scale)
(11.81%)
Watershed Area
38.4 km2
(3.2-98.1)
Stream Width
5.6 m
(2.4-11.1)
Stream Depth
29.9 cm
(13.2-59.4)
Factor 5 Factor 6
(pH-Ammonium) (Substrate
(6.71%) Condition)
(6.35%)
pH Percent
7.69 Substrate
(6.61-8.31) (>16mm)
68%
(20-96)
Ammonium Embeddedness
5.8 ueq/l 55%
(0.0-32.8) (10-86)
Percent Riffle
31%
(5-50)
Table 3-9. Forward stepwise multiple regression of AFLP within population similarity value as dependent variable.
Three variables met the 0.05 significance level for entry into the model.
Environmental
Factor
partial
R2
model
R2
F value Pr>F
PCA Factor 3
(Latitudinal clines) 0.4328
PCA Factor 2
(N/P/C) 0.3489
PCA Factor 5
(pH/Ammonium) 0.1841
0.4328
0.7917
0.9758
6.10
12.06
45.60
0.0387
0.0104
0.0005
42
-------
Condition and Sustainability
4 Considerations when Implementing a Genetic Diversity Assessment
Suggestions that genetic diversity should be used as an indicator of ecological health are not new (e.g.,
Beardmore et al, 1980; Nevo et al, 1988; Fore et al, 1995a, 1995b). However, specific guidance on
how genetic markers should be incorporated into an ecological assessment program is lacking. The
overwhelming majority of reported scientific studies of genetic diversity are at relatively small scales,
incorporating assays of at most a few dozen populations. Aspects of how to scale-up to region-wide
analyses with dense geographic sampling, including data management and quality assurance issues, have
not been considered. In addition, the focus of genetic diversity studies in the scientific literature is usu-
ally on the status of the species under study, not the ecosystem. The US EPA is currently engaged in
several studies designed to evaluate the utility and practicality of implementing genetic diversity indica-
tors as part of ecological assessment or monitoring. Based on our experiences with these studies and
relevant information from the scientific literature, several guidelines are suggested.
4.1 Sampling Design
Two different types of sampling designs have traditionally been employed for ecological assessments:
source-biased studies, in which sites with known exposures are compared to reference sites, and region-
wide studies, in which a large number of sites are sampled according to a defined sampling scheme in
order to create a regional profile. Some studies may have aspects of both designs, in which sites with
known exposures are compared to a relatively large number of "reference" sites within the region. Both
designs are amenable to genetic diversity analysis. Many examples of the source-biased design applied
to genetic diversity exist in the ecotoxicological literature (see Table 2-1). In addition, there are many
examples in the conservation genetics literature of genetic diversity analyses that incorporate regional
scales, although they rarely include large numbers of sample sites within the region. To our knowledge,
no examples yet exist of intensive regional ecological assessments that have utilized a genetic diversity
indicator.
The source-biased design has obvious cost advantages when the assessment question of interest is
whether a known, local exposure has an impact on the genetic diversity of resident populations.
However, considerable care must be exercised when implementing this design. Because the intent of
this design is to measure a recent genetic change, the reference population(s) will ideally be identical to
the test population(s) in all aspects except for the application of the specific exposure, yet independent
of the test population(s) following the exposure. Thus, the populations must have had similar genetic
diversity before the exposure, either because they recently diverged or because they experienced high
gene flow prior to the exposure. In addition, significant gene flow between the populations must have
stopped immediately following the exposure and any genetic differentiation that occurred must have
been due to the exposure and not to other population or environment-related factors. These standards
are likely to be difficult to meet. In practice, genetic diversity is often measured at a number of refer-
ence sites and compared to the exposed site. If genetic diversity at the exposed site is outside the norm
for the reference sites then the exposure is implicated as the cause of the change in genetic diversity.
Here too, there can be difficulties with interpretation. Often, reference populations are chosen to be geo-
graphically distant from the exposed population in order to ensure that they represent "reference condi-
tions". Typically, it is not clear that the reference populations are not each more closely related to each
other than any is to the exposed population and that any genetic diversity differences uncovered did not
predate the exposure.
Regional studies offer much greater ability to characterize patterns of intraspecific genetic diversity and
their possible causes than do source-biased studies. Genetic diversity will naturally vary among popula-
43
-------
Genetic Diversity as an Indicator of Ecosyiiijjf
I 1 1
tions for a variety of reasons, including variation
in the size of populations that different habitats
can support, as well as evolutionary relationships
and patterns of dispersal among populations.
These natural levels of genetic diversity are
indicative of fundamental population data, such as
effective population sizes and population bound-
aries (section 2.1). In addition, any attempt to
determine whether anthropogenic factors have
influenced present levels of genetic diversity must
be able to distinguish historical (evolutionary) pat-
terns from recent change. Since regional assess-
ments allow better characterization of the natural
variation in genetic diversity measures, they can
provide guidance for selecting specific areas for
more intensive study. For example, if genetic
diversity of one population is determined to be
low, it can be compared to evolutionarily similar
populations to determine whether a recent genetic
change is implicated.
ic to
of
is tte
to
af
• I I
The natural synergisms that genetic diversity data lend to landscape-level analyses and species assem-
blage studies suggest that incorporation of genetic diversity sampling into existing regional assessment
programs, including EMAP and Regional EMAP programs, is the preferred strategy to obtain genetic
diversity data at present. Sampling of target species for genetic diversity analyses can easily be accom-
plished within existing EMAP guidelines, and will help reduce costs of sample collection. One of the
advantages of DNA-based analyses is that sufficient DNA can be obtained from a single fish fin-clip to
perform a large number of genetic analyses. Thus, tissue can easily be collected in the course of normal
field identification procedures and, in most cases, specimens can be released back into the environment.
In the long run, the most efficient method to measure anthropogenic changes in genetic diversity is to
measure genetic changes directly through temporal monitoring. For this purpose, regional assessments
provide an excellent means to collect the necessary baseline genetic data for future comparisons. In
addition to regional-scale assessments, intensive temporal analyses of genetic changes at a network of
index sites will be extremely valuable. Index sites typically are assessed with a greater range of diag-
nostic and condition indicators, which will allow
for better integration of the genetic data. Since
index sites are intended for long-term monitoring,
they provide an excellent opportunity to measure
the temporal scale of genetic changes. For both
regional and index site studies, it will be critical
that DNA or tissue is archived for future analyses
as part of the assessment. Given the rapid
changes occurring in molecular methodologies, it
is likely that whatever marker is used to measure
genetic diversity initially will not be the optimal
strategy at later stages of the assessment. The
availability of archived samples will allow future
retrospective analyses to assess genetic changes
using the most appropriate technologies available.
In the the
to
in is to
i i i
44
-------
Condition and Sustainability
Scale-up issues. Several project management issues emerge when the scale of genetic analysis changes
from assessment of genetic structure at a few sites at a single point in time to assessment at regional
scales and/or over time periods that may encompass decades. Clearly, the greater management complex-
ity and cost of this type of project requires greater forethought in design of the genetic survey. The utili-
ty of a pilot study to guide project management cannot be overemphasized. The pilot study serves to
validate the choice of molecular markers and species in the study region and allows for initial assess-
ment of the feasibility of more intensive or large-scale sampling (Baverstock and Moritz, 1996). It is
very likely that the project plan will be redefined based on the results of the pilot study. For example, a
target species may be difficult to sample within the region or a number of molecular markers that were
found to be useful in other areas may lack polymorphism (and thus discriminatory power) within the
region. A pilot study also may determine whether the scale or intensity of sampling for a regional
assessment is appropriate.
Management of a large genetic diversity survey will be logistically simpler if it can be divided into
smaller units that are analyzed individually. For example, the EPA's pilot study of genetic diversity for
central stonerollers throughout the Eastern Cornbelt Plains Ecoregion (section 3.1) was divided into
analysis units that were equivalent to large watersheds or basins. In addition, analyses were completed
for central stonerollers before attempting to assess genetic diversity in additional species. Geographical
and species-stratified analysis has the additional advantage that fairly intensive genetic diversity assess-
ments can be completed for specific basins on a regular basis throughout the life of the project. While
the advantages of such compartmentalization of analyses are clear, it does lead to a potential for bias if
variation in laboratory procedures occurs over time (thus, between sample units). Inclusion of duplicate
samples from prior analyses as a type of positive control against temporal variation in laboratory proce-
dures should control for this type of bias.
Additional scale-up issues involve planning for adequate data management structure and are dealt with
in section 4.6.
4.2 Species Choice
Most genetic diversity studies in the scientific literature are focused on conservation or management of a
particular species, usually one that is threatened, endangered, or of some economic importance. In
choosing species as indicators of environmental health, other issues clearly come into play. Table 4-1
presents "optimal" species attributes for environmental monitoring. Many of these species attributes
stem from consideration of basic population genetics. Genetic diversity of species that are highly sensi-
tive to degraded habitat and that have short generation intervals will respond more quickly and more
dramatically than other species. Species with low dispersal ability will have higher exposure to specific
environments and may provide finer resolution of environmental differences between sites. Asexual
species, including microbes and algae, are not optimal mainly because the distinction between intraspe-
cific and interspecific genetic diversity is blurred. Valid use can be made of such species (e.g., Ford et
a/., 1998) but they become functionally equivalent to species assemblage indicators.
Other species characteristics listed in Table 4-1 help make collection, analysis, and interpretation of the
data simpler. Selection of species that are easily distinguished morphologically will ensure that genetic
diversity is measured at the intraspecific level, and that comparisons are not erroneously made between
different species (in fact, cryptic species complexes are readily diagnosed using molecular markers,
which is one advantage of combining analysis of genetic diversity with species assemblage assess-
ments). Selection of broadly distributed species allows simpler analysis of scale issues. Use of species
that are important to resource managers will allow easier integration of genetic diversity monitoring into
existing monitoring programs. Species that are moderately abundant within the study area are easier to
45
-------
Genetic Diversity as an
Table 4-1. Optimal characteristics of species assessed with a genetic diversity indicator.
Optimal Species Characteristics
• Short generation interval
• Moderate-high exposure to stressor(s)
• Moderate-high sensitivity to stressor(s)
• Low-moderate dispersal ability or highly philopatric
• Sexual reproduction
• Native species
• Broad distribution
• Moderate abundance
• Management importance
• Morphologically distinct
• Known life-history parameters (age structure, sex ratio, etc)
• Availability of comparative genetic and demographic data
• Availability of historical DNA or tissue samples
• Amenable to laboratory culture
collect, although these species are usually not management priorities. The population genetic structure
of species that have not been excessively moved around is generally easier to interpret than the structure
of introduced species. This is because the stocking history of introduced species is generally poorly
known; non-native species could be useful indicators in cases where the history of introductions and
their sources are well documented. Availability of historical data, including the availability of archival
samples (DNA, fins, scales, or whole preserved specimens) is useful for reconstructing changes in
genetic diversity that may have occurred prior to or during known exposures in the past. Finally, the
selection of species that can be cultured in the laboratory will aid in assigning causality to genetic
changes, if such studies are desired in the future.
Ideally, the indicator will be applied to several species, since a multi-species index should better predict
ecosystem status than a single-species index. Genetic responses of individual species are not always
predictable (Gillespie and Guttman, 1999), as different species "see" different aspects of the environ-
ment, and not always what we expect. If several species are selected then additional considerations
become relevant, such as sampling from phylogenetically and ecologically diverse taxa.
4.3 Which Genetic Marker?
Several books have been published in recent years that review the biological and analytical properties of
different molecular markers (e.g., Avise, 1994; Hillis et al, 1996; Caetano-Anolles and Gresshoff
1997). A general consensus is that no genetic marker is best for all applications and in the hands of all
investigators; each provides different insights and requires different levels of investment in equipment
and training. Here, we will summarize the properties of some of the most common genetic markers and
discuss their relative strengths and weaknesses as ecological indicators.
46
-------
r
Diversity as an Indicator of Ecosystem Condition and Sustainability
Allozymes. Allozyme electrophoresis is a simple and time-honored technique in the field of population
genetics. The principle of the technique is that allelic forms of enzyme proteins with different net
charges will have different mobilities when induced to move through a matrix by application of an elec-
tric current (electrophoresis). Alternative forms of the enzyme at the enzyme locus (alleles) migrate dif-
ferent distances through the matrix and are visualized by histochemical staining (Figure 4-1). Numerous
manuals have been developed that outline general equipments needs, procedures, and gel pattern (zymo-
gram) interpretations (see May, 1992; Murphy et al, 1996; and references therein).
Although allozymes are often thought of as "old technology", they have some clear advantages over
other genetic markers as ecological indicators. Few other markers can match allozymes in the simplici-
ty and economy of standard procedures. Allozymes have a much longer history than other genetic mark-
ers and have been analyzed in many more laboratories so the historical database of comparative popula-
tion genetic data is much larger for this marker than for any other. In fact, it is highly likely that
allozyme datasets can be found for any of the common stream fishes in the country. In comparison to
some other markers, allozyme techniques suffer from a limited number of marker loci available for
study and a limited number of alleles per locus. Less than a dozen polymorphic markers are assessed in
typical studies, and most of these markers only segregate for two or three variant alleles. As noted earli-
er, allozyme loci are more likely to be affected by natural selection than most DNA markers, which may
bias estimates of gene flow and genetic diversity. For example, allozyme markers suggested significant
gene flow in oysters along the eastern and Gulf coasts of North America, but both mitochondrial and
nuclear DNA markers indicated a sharp biogeographic boundary between northern and southern popula-
tions in northeastern Florida (reviewed in Avise, 1994).
Figure 4-1. A histochemically stained starch gel showing GPI allozyme loci of rainbow trout. Each vertical lane rep-
resents a different individual.
47
-------
Genetic Diversity as •anAriKiiSlJ
1 1 I
'' - ''-.
Mitochondrial DNA is a circular DNA molecule containing about 16,500 base pairs that is present in
multiple copies in the cells of eukaryotic organisms. Mitochondrial DNA has a number of properties
that lend it to ecological assessments. Mutation rates tend to be higher than for most nuclear DNA
regions, so large numbers of alleles (called haplotypes) are generated. Each individual usually only pos-
sesses one mitochondrial haplotype, which it typically inherits from its maternal parent. In addition,
recombination within the mitochondrial genome appears to be rare or absent. Unlike different allozyme
alleles, mitochondrial DNA sequences (Figure 4-2) can be analyzed to determine patterns of evolution-
ary relationships between different haplotypes. Thus, analysis of mitochondrial DNA sequences adds a
different dimension to the analysis of genetic diversity; one can move beyond asking whether two indi-
viduals are the same or different at a genetic locus to ask how different they are at that locus. This pro-
vides a straightforward method for assessing (maternal) genealogical relationships between individuals
of a population, and between individuals of different populations and different species.
Figure 4-2. Electropherogram of DNA sequence generated by an automated genetic analyzer. Identification of muta-
tional relationships among mitochondrial types (haplotypes) can elucidate patterns of evolutionary relationships
among populations.
Mitochondrial DNA is not without drawbacks, the most important of which is that the lack of recombi-
nation within the molecule causes the entire mitochondrial genome to behave effectively as a single
genetic marker; different mitochondrial genes are linked and therefore not independent. Since the histo-
ry of a mitochondrial lineage is not identical to the population history (most notably, it says nothing
about male contributions), interpretations made strictly from mitochondrial DNA may be erroneous.
The tremendous information content derived from DNA sequencing comes at a cost in terms of equip-
ment, supplies, and technical expertise (Figure 4-3). Constant technological improvements are leading
to rapid reductions in these costs. For example, the development of the polymerase chain reaction
(PCR) and "universal" PCR primers has negated the need to isolate mitochondrial DNA away from
nuclear DNA, or to clone specific fragments prior to sequencing (instead, the target sequence is simply
PCR amplified). Meanwhile, a number of techniques have been developed that reduce the need for
DNA sequencing or the number of required sequencing reactions. A common strategy is to screen for
restriction fragment length polymorphisms (RFLPs). RFLPs provide a coarse indicator of DNA
sequence variability, typically capturing less than one-eighth of the DNA sequence variability in a
region. However, RFLP patterns contain information about relationships between haplotypes and can be
analyzed relatively inexpensively. Other methods are used to "prescreen" mitochondrial DNA, so that
only unique or previously uncharacterized haplotypes are sequenced. These include single strand con-
formation polymorphism analysis (SSCP), denaturing gradient gel electrophoresis (DGGE) and various
commercial strategies. These techniques typically identify 80% to 100% of single-base mutations within
DNA, but usually say little about haplotype relationships. Still, when combined with DNA sequencing,
they can be more informative than RFLP analysis with only slightly more technical difficulty and cost
(Figure 4-3).
48
-------
I~TT
Genetic Diversity as an Indicator of Ecosystem Condition and Sustainability
Information per Marker
RAPD
AFLP
Minisatellite
VNTR
DMA
Sequencing
RFLP
SSCP DGGE
Cost/Equipment/Technical Inputs
Figure 4-3. Relative advantages and disadvantages of different molecular marker strategies in relation to information
per marker (darker colors in the figure being more informative), the number of markers per typical study, and costs per
study in terms of capital outlay and technical expertise. The figure is not based on quantitative data and is presented
for illustration purposes only.
Nuclear DNA Sequencing, RFLPs, and Prescreening Strategies. Strategies that are
available for analysis of mitochondrial DNA also are available for the analysis of nuclear DNA genes.
with similar advantages and limitations. The general strategy is often referred to as single-copy nuclear
DNA (scnDNA) analysis. Typically, non-protein coding intervening sequences (introns) or flanking
regions are targeted for analysis since they are usually more polymorphic than coding sequences. The
primary advantage of scnDNA analysis is that many more genetic markers that are independent are
available for analysis, so it can be highly complementary to mitochondrial DNA analysis. The develop-
ment of gene sequence databases for many organisms, combined with PCR technology, has made avail-
able a number of well-studied genes for population analyses. A number of "universal" PCR primers
have been published (e.g., Palumbi, 1996) to aid in the development of genetic markers for different
species, although a certain amount of primer modification is often required. Because many of the gene
sequences available for analysis are believed to have an impact on fitness, these, like allozymes, have
the potential to be developed as diagnostic indicators of natural selection (and thus population stress).
There are a number of disadvantages with scnDNA markers. In general, levels of polymorphism are low
compared to mitochondrial DNA and some other nuclear DNA markers. In addition, the analysis is rela-
tively intensive, even when using mutation-prescreening techniques, so relatively few scnDNA markers
are generally analyzed per study.
49
-------
Genetic Diversity as an Indicator of Ecosystem Condition and Sustain
I I I
Multi-locus DNA Fingerprints. DNA fingerprinting is a strategy that is in many respects the opposite
of scnDNA analysis. Instead of targeting single, relatively well-characterized genes, DNA fingerprints
target many anonymous chromosomal regions for analysis simultaneously. Typically, little is known
about these regions except that they possess a small region of similarity to specific probes or PCR
primers. Any region that has such homology presents as a distinct DNA fragment or "band" following
gel electrophoresis. For any one individual, the pattern of presence and absence of bands can be highly
complex, resembling a bar code. With the most variable markers, individuals can be uniquely identified
by these band patterns. The most common DNA fingerprinting technique is random amplified polymor-
phic DNA (RAPD) analysis (Figure 4-4a), in which short primers of arbitrary sequence are used to
amplify DNA fragments from 10-50 discrete regions in the genome. A newer method called amplified
fragment length polymorphism (AFLP) analysis (Figure 4-4b) is similar, but depends on amplification of
polymorphic fragments generated by restriction enzymes (RFLP) from arbitrary regions of the genome.
I 1 I I I I
= -.
Figure 4-4. Examples of multilocus DNA fingerprints. Both RAPD (A) and AFLP (B) fragments can be generated
without a priori knowledge of an organism's sequence. As dominant markers, homozygous and heterozygous indi-
viduals cannot be distinguished. The AFLP fingerprint here differs from Figure 3-14 because this is a multiplex
AFLP reaction generated with an automated genetic analyzer.
50
-------
r
Diversity as an Indicator of Ecosystem Condition and Sustainability
There are two principal advantages of multilocus fingerprints for ecological assessments. The first is
that no sequence information about the genome of the organism is needed in order to apply the methods.
Thus, marker development costs are minimal and species can be chosen for analysis based on ecological
or management criteria rather than the amount of sequence information already known. The second
advantage is that genetic differences between individuals can be distinguished with relative economy.
Dozens to hundreds of markers are analyzed in a typical DNA fingerprinting study.
The overriding disadvantage of DNA fingerprints is the poor quality of genetic information from each
individual fingerprint band. Individual fingerprint bands cannot reliably be assigned to independent
genetic loci. More importantly, RAPD and AFLP markers are dominant markers, which means that indi-
viduals that are heterozygous for a marker (i.e, only one chromosome of the pair has the marker) cannot
be distinguished from individuals that are homozygous for the marker (both chromosomes have the
marker), significantly decreasing the genetic information available. As a result of these ambiguities,
comparisons are generally made in terms of overall 'similarities', taking into account the proportion of
bands that are shared between individuals. Another concern with DNA fingerprinting methods is that
sophisticated procedures must be implemented to reliably sort through the complex fingerprint patterns
to identify homologous fingerprint bands from different individuals. Related to this is concern about the
overall reliability of fingerprinting methods, particularly RAPD fingerprints. The RAPD technique is
known to be very fickle, and adherence to exact protocols by different labs, often including use of the
same brand of equipment and reagents, is considered critical to repeatability.
Microsatellite DNA Markers. Microsatellite DNA, also called simple sequence repeats (SSRs) are
regions of repetitive DNA that consist of tandem repeats of a core sequence of two to five base pairs,
such as (CA), (TAGA), and (CAT). Different alleles at a microsatellite locus differ in the number of tan-
dem repeats of the core sequence. These sequences appear to be ubiquitous in the genomes of eukary-
otes, and thousands of potential microsatellite markers could be developed for most species.
Figure 4-5. Flourescently labeled microsatellites detected using an automated genetic analyzer. Use of flourescent-
ly labeled markers allows differentiation between multiple loci (illustrated above by blue, green and yellow labeled
markers) within the same reaction (multiplex PCR) thus reducing cost and increasing throughput.
51
-------
Genetic Diversity as an
Ill I I
Microsatellite DNA markers have some tremendous strengths for ecological assessments. They are sub-
ject to very high mutation rates relative to scnDNA, sometimes producing dozens of alleles (Figure 4-5).
Like allozymes and scnDNA markers, inheritance of microsatellite markers is codominant, so heterozy-
gotes can usually be reliably differentiated from homozygotes. The proportion of individuals that are
heterozygous in a population is much higher than for other nuclear loci, often approaching 100%.
Although heterozygosity at microsatellite loci is somewhat less sensitive to genetic bottlenecks than
mitochondrial DNA diversity, it is much more sensitive than other nuclear DNA markers due to the large
number of rare segregating alleles at these loci. Loss of these rare alleles actually provides a more sen-
sitive measure of population bottlenecks than does heterozygosity (Leberg, 1992). It appears that the
predominant mode of mutation is to an allele one repeat-unit different from the original allele [e.g., from
(CA)12 to (CA)13], thus genealogical information can be captured from allelic relationships of microsatel-
lite loci, although this information is less precise than that captured from DNA sequence analysis of
mitochondrial DNA and nuclear genes. The primary disadvantage of microsatellite DNA markers is
development cost. Technical expertise necessary for microsatellite marker development is greater than
for any of the other markers listed, although, like DNA sequencing, the technical demands are decreas-
ing. Microsatellite markers are developed from non-protein coding DNA regions and, therefore are not
conserved across taxa, so that microsatellite DNA markers developed for one species are often only use-
ful for very similar species. The number of organisms for which microsatellite markers have been
developed is increasing rapidly, so it is possible that microsatellite development will be less costly in the
near future. In addition, the very large number of alleles present at some microsatellite loci requires
large sample sizes be used to estimate allele frequencies accurately.
Recommendations. This report has considered only a subset of the available genetic markers, but these
are the most common and well supported in the scientific literature. The "best" marker for ecological
assessments will vary, depending on the specifics of each situation. Until reliable methods are devel-
oped that allow economical analysis of nuclear DNA sequences from more than just a few genes per
study, RFLP, SSCP, DGGE and sequence analysis of nuclear genes will likely be less informative than
allozyme or microsatellite studies per unit effort. Mitochondrial DNA can be highly informative and
offers insights not available from analysis of nuclear DNA. However, information from mitochondrial
DNA may be misleading if interpreted alone so it is recommended that mitochondrial DNA be assessed
only in conjunction with other markers.
Among the remaining markers, microsatellites undoubtedly offer the best combination of information
per genetic marker and potential for analysis of many genetic markers. Although most microsatellite
studies of natural populations to date have utilized relatively few microsatellite markers, there is now no
technical reason why dozens, even hundreds of microsatellite markers cannot be developed and applied
to genetic analyses. The technical challenges of microsatellite development can be overcome (for exam-
ple, several commercial laboratories will now develop microsatellite markers on contract), however, an
advanced laboratory is still required for microsatellite analysis. Reliable scoring of microsatellite mark-
ers requires the use of labeled PCR primers (either isotope or fluorescence) for visualization so, mini-
mally, laboratories must have the ability to detect these labels. For large-scale, regional analyses utiliz-
ing many microsatellite markers, automated laboratory analysis using commercial genetic analyzers
(automated sequencers) is essential.
The many advantages of allozymes, (economy, standardized methods, large existing database of infor-
mation, homology of loci across species) should not be overlooked, particularly when microsatellite
analysis is infeasible. Targeted analysis of specific allozyme loci (e.g., GPI) also may be useful as diag-
nostic indicators of specific stressors (e.g., heavy metals), and could complement analyses of other
genetic markers. However, the requirements for ultra-cold storage of tissue samples in the lab and in
52
-------
Condition and Sustainability
the field may make allozyme analysis impractical for regional analyses, particularly if the genetic collec-
tions are done as part of a multi-indicator assessment. In addition, if genetic diversity of allozymes is
too low overall in the indicator species (as determined, perhaps, by a small pilot study), or if lethal sam-
pling is not acceptable, then other methods should be explored. AFLPs and RAPDs, in that order,
should be considered if neither allozyme nor microsatellite studies are feasible. Both methods allow
assessment of many different genetic markers, but identification of AFLP markers is believed to be more
repeatable between laboratories. However, AFLP analysis, like microsatellites, requires a more
advanced laboratory able to detect isotopically or fluorescently labeled PCR products. Mitochondrial
DNA analysis may be combined with any of the above nuclear markers to provide semi-independent
genetic information. For example, analysis of evolutionary relationships among mitochondrial haplo-
types may provide information on the historical biogeography of the species. Such information can help
interpret patterns of genetic diversity in nuclear markers. A flow chart to aid in choosing genetic mark-
ers is provided in Figure 4-6.
4.4 Sample Size Considerations
Clearly, statistical power to detect differences in genetic diversity within populations and to detect genet-
ic differentiation among populations will increase with increasing sample size. Two different sample
sizes need to be considered: the number of individuals sampled per population and the number of mark-
ers assessed per individual. For any one marker, the ability to discriminate allele frequency differences
between populations is low so large numbers of individuals need to be sampled from each population.
\n general, as more markers are assessed per individual, fewer individuals need to be sampled from
each population. However, the balance between the number of markers and the number of individuals
varies with the marker chosen. Recommendations for determining appropriate sample sizes are
reviewed in current literature (Baverstock and Moritz, 1996).
In general, genetic markers that allow estimation of molecular relationships (DNA sequences, RFLP, and
microsatellites) should require smaller sample sizes to achieve the same statistical power as markers that
only consider allele frequencies, although this depends on the level and complexity of molecular rela-
tionships. Estimates of allele frequencies for microsatellite markers with high polymorphism (e.g., 10 or
more alleles) can have very high standard errors since few individuals in the sample will possess any
one allele.
For dominant markers (RAPD, AFLP) in which heterozygotes cannot be distinguished from one of the
homozygotes, less information is available per locus and therefore more individuals or more markers
must be sampled for the same statistical power as analysis of codominant loci. Often, logistical prob-
lems limit the number of individuals that can be sampled so statistical power is gained by sampling as
many loci as possible. Simulation studies by Mariette et al. (1999) suggest that at least four times as
many AFLP markers are needed as microsatellite markers to estimate genetic diversity within popula-
tions that are at equilibrium between migration and genetic drift. Comparatively more AFLP or RAPD
markers will be needed to measure genetic diversity of populations that are not at equilibrium.
Mitochondrial DNA, which is only transmitted from female parent to offspring (haploid inheritance),
also has less genetic information than a single codominant marker, based on analysis of haplotype fre-
quencies alone. Reduction in the standard error of estimates of haplotype frequencies can only be
accomplished by increasing the number of individuals sampled.
Although it is difficult to come up with specific numbers, as a rule of thumb, sample sizes of between
50 and 100 samples per site are typically targeted in population genetics studies, although actual sample
sizes are often lower. Analyses that utilize dominant marker systems should aim to assess between 100-
53
-------
Genetic Diversity as an Indicator of Ecosystem Condition and Sustainabilit
I I I
Are suitable microsatellite markers
available, or can they be developed or
purchased?
NO
YES
Is the lab equipped for
fluorescent DMA fragment
detection?
YES
NO
Is there
allozyme
diversity?
I
NO
YES
Consider
microsatellite
study
YES
Is lethal sampling
and -70° field
storage feasible?
NO L
Consider
allozyme
study
NO
Is the lab equipped for
fluorescent DNA
fragment detection?
Consider
RAPD study
Does the lab have a DNA
seauencina capability?
YES
Consider
AFLP study
YES
Consider combination of
mitochondria I and
nuclear genetic markers
Figure 4-6. Flow diagram for deciding the best genetic marker or combination of genetic markers in relation to avail-
able resources.
54
-------
Condition and Sustainability
200 polymorphic markers. Analyses that utilize codominant markers should aim for between 20 and 50
polymorphic loci.
4.5 Personnel Training and Specialized Equipment
Significant technical expertise is required for field sampling, laboratory analysis, and data analysis and
interpretation. The degree of field expertise is equivalent to that needed for species assemblage indica-
tors. Personnel must be able to operate sampling gear such as backpack electroshockers or seines effec-
tively. In addition, they must have the ability to identify target species in the field. In practice, species
discriminatory skills are probably less important than for assemblage indicators because the validity of
the field identification will likely be confirmed by the genetic analysis.
Laboratory analyses require varying levels of skills. DNA extraction and quantitation techniques have
become highly commercialized and many simple kits can be used that yield DNA of suitable quality and
concentration for PCR-based analyses. The primary skill necessary is accurate pipetting. PCR analysis
also is relatively straightforward and usually only requires good pipetting skills. However, a number of
factors can disrupt the PCR process, and troubleshooting problems is a common occurrence in PCR
assays. For this reason it is recommended that M.S.-level person with molecular biology training be on
hand to supervise or consult during the PCR phase. A similar skill-level is recommended for gel prepa-
ration, sample loading and electrophoresis, and for operating automated DNA sequencers of genetic ana-
lyzers. Interpretation of gel patterns to determine the allelic complement at each marker analyzed gener-
ally requires a skilled, M.S.-level molecular biologist. In general, improvements in technology, such as
highly robotic capillary electrophoresis systems for genetic analysis, have decreased the training require-
ments necessary to perform genetic analyses. Development of microsatellite markers requires sophisti-
cated molecular biology skills.
Equipment requirements to perform different laboratory operations are described in Table 4-2. A num-
ber of software packages are available for genetic analysis, and similar analyses can be performed with
general-purpose statistical software such as SAS®. Selection and interpretation of appropriate indices is
best done in consultation with a population geneticist or a statistician who is familiar with genetic data.
4.6 Information Management
A typical genetic diversity study will generate large amounts of data. Critical data sets include the loca-
tions and dates of sampling, the number of individuals of each species that are sampled per site, and the
genotype of each individual at each of the molecular markers analyzed. These data sets must be related
to other databases that may exist, including phenotypic data (age, size, developmental abnormalities)
biomarker data, chemical and physical habitat data, assemblage indicator data, and landscape-level data.
This suggests that use of relational database software will be useful to manage the data, particularly if
this software has already been incorporated to manage linked databases. However, simpler database
management tools, including spreadsheets and simple database software, can be appropriate for less
complex datasets.
Large numbers of tissue samples and DNA extractions will be collected that must be archived for valida-
tion purposes and to aid future retrospective assessments. Minimally, a database is needed that docu-
ments for each sample a unique sample ID, a population ID, the storage location of the tissue sample,
the storage location of the DNA sample, dates of collection, DNA extraction, DNA quantification, as
well as amounts and quality assessment of the archived material. If voucher specimens were collected
along with the genetic samples then the database should include this link as well. Field data collection
55
-------
Genetic Diversity as an Indicator of Ecosystem Condition and Sustain
I I I
forms and forms that document dates of laboratory manipulations and spatial orientations of samples
during those manipulations (DNA extraction, PCR, electrophoresis, marker scoring) will need to be
physically archived.
Metadata requirements include documentation of field collection procedures, as well as detailed labora-
tory methods (see appendices) and data analysis procedures. The procedure for translating genotypic
data for an array of genetic markers into a data file must be explained, including descriptors of each of
the fields in the data file. Similarly, documentation of the rationale and procedures for the statistical
analyses, including software documentation, are needed.
4.7
Costs
Monetary costs of implementation of a genetic diversity indicator are incurred during sampling, labora-
tory analysis, and data analysis. Costs of field sampling for DNA analyses are similar to costs to collect
species assemblage and biomarker data, and will generally be shared with those indicators. McCormick
and Peck (2000) estimate the cost to field a contracted crew of 4 people at $1200 per site visited. They
estimate the cost of field equipment, including a backpack electrofishing unit, to be $3515 per crew,
with a 15% annual maintenance and depreciation rate. Sampling for allozyme analyses requires the use
of special storage containers and a regular supply of dry ice or liquid nitrogen. This may add an addi-
tional $50 per site in supply costs, plus approximately $400 in cold storage equipment. One possible
consequence of the necessity for cold storage is that the crew may not be able to remain at remote sites
Figure 4-7. DNA quantitation is performed using a commercially available fluorescent nucleic acid stain that is
detected with a fluorescent scanner.
56
-------
r
Diversity as an Indicator of Ecosystem Condition and Sustainability
Figure 4-8. Thermal cyclers are used for fragment amplification (RAPD, AFLP, microsatellites) and DMA sequencing
assays.
Figure 4-9. A capillary-based, auto-loading genetic analyzer can perform both fragment analysis (AFLP, microsatel-
lite) and DMA sequencing.
57
-------
Genetic Diversity as an
Ill I I
Table 4-2. List of standard and specialized equipment for different types of genetic marker studies. * optimal but not
necessary.
All laboratories
1. Freezer-samples
2. Freezer- chemicals
3. Refrigerator
4. UV transilluminator
5. Ultrapure water source
6. Pipetters
7. Variable speed centrifuges
8. Microcomputer with statistical genetics analysis software
Standard Equipment - DNA laboratories
1. Separate sample preparation room/area (DNA extraction)
2. PCR room/area
3. Post-PCR room/area with photodocumentation and/or fragment analysis equipment
4. thermal cyclers
5. Agarose gel electrophoresis rigs
6. Microcentrifuges (10,000 RPM)
Specialized equipment- allozymes
1. -80 freezers
2. dry ice or liquid nitrogen and canisters
3. Starch gel electrophoresis rigs
4. Power supplies (to 350 V, 150mA)
5. Incubator oven.
6. Chiller (for cooling starch gels during run)*
Specialized Equipment - microsatellites
Microsatellite Development
1. Hybridization oven
2. Incubator oven
3. Automated DNA sequencer
4. Shaking incubator
5. Fluorescence detection system (see microsatellite screening equipment)
Microsatellite Screening
1. Fluorescence detection system, either
(a) Acrylamide gel electrophoresis rigs, 1000 V power supply, fluorescence scanner, fragment
analysis software, microcomputer
(b) Automated DNA sequencer with fragment analysis software
Specialized Equipment - RAPD
1. Power supplies (to 350 V, 150 mA)
2. Specialized agarose gel electrophoresis rigs for recirculating buffer
3. Chiller unit (to 0° C) for cooling agarose gels during run*
4. Microcomputer with fragment analysis software
Specialized Equipment - AFLP
1. Fluorescence detection system, either
(a) acrylamide gel electrophoresis rigs, 1000 V power supply, fluorescence scanner, or fragment
analysis software, microcomputer
(b) Automated DNA sequencer with fragment analysis software
58
-------
Condition and Sustainability
long before samples must be shipped to the laboratory, which may increase sampling costs.
Laboratory costs are more variable, and depend greatly on both the type of molecular marker assessed
and the technical skills of laboratory personnel. All estimates given here are provided with the caveat
that they are likely to change rapidly. As with field sampling, labor is the greatest cost. Based on
review of the EPA's pilot study that utilized the RAPD fingerprinting technique, it is estimated that a
crew of four laboratory technicians can extract DNA from 95 samples, perform nine RAPD PCR assays
per sample, electrophorese each PCR product individually, and size each of the fingerprint bands in a
period of approximately 9 days. Assuming atypical sample size of 50-100 individuals per site, this indi-
cates that a crew of four will require between one and two weeks to assay a single site. Supply costs,
including enzymes, agarose, plastics and chemicals, are estimated at $9 per sample for nine RAPD
assays. Equipment costs included two fluorimagers and associated computer hardware and software
($70,000 each) ten agarose gel units ($4000), chiller ($3000), and many smaller items (pipetters, stirrers,
centrifuges, incubators, approximately $10,000). We assume an average depreciation of approximately
20%. The fluorimagers could reasonably be replaced with cheaper photodocumentation systems for less
than $20,000.
Guidance on costs of other types of molecular marker studies can be obtained from estimates by com-
mercial laboratories. It is assumed that this is a maximum cost estimate because a profit margin is built
in, but depreciation costs are built into the estimates and these labs may experience higher efficiency
than other labs. Allozyme electrophoresis typically costs between $10 and $40 per sample, with
between 5 and 15 polymorphic loci (and often many more monomorphic loci) assayed. Costs to set up
an allozyme laboratory fall in the range of $10,000 to $20,000. DNA sequencing typically costs $25-
$40 per sample, for a sequence of up to 500 bp. It is often recommended that sequences be generated
for both the upper and lower DNA strands, doubling the cost. This cost does not include the cost of
DNA extraction and PCR amplification of the target locus, approximately $2 per sample. Automated
DNA sequencers vary greatly in price and quality, but useful models can be obtained for between
$70,000 and $130,000. Microsatellites and AFLP fingerprints that are assayed on automated DNA
sequencer cost between $2 and $5 per sample run. For microsatellites, a single automated run may
include between one and 8 distinct PCR reactions, increasing the efficiency of the assay. Costs of DNA
extraction ($2) are not included.
Costs of marker development range from essentially zero for RAPD, AFLP, and allozymes to several
thousand dollars for microsatellites. At least one commercial laboratory will guarantee production often
usable microsatellite markers for any species for a cost of $20,000.
4.8 Summary Recommendations
Among the myriad different ecological indicators available for environmental assessments, the optimal
niche of a genetic diversity indicator is to map patterns of population structure and to identify cumula-
tive genetic changes in populations through spatial and/or temporal comparisons. Like other indicators
of ecological condition, its value will be greatly enhanced if it is interpreted as part of a multi-indicator
assessment. By its nature, genetic diversity is a generalized indicator of long-term changes in popula-
tions; it will be difficult to assign causation to any specific stressor. There are exceptions to this rule.
For example, assays for specific allozyme genotypes may be developed as diagnostic indicators of par-
ticular classes of stressors (e.g., heavy metals) and, as more is learned about functional consequences of
nucleotide variation at specific genes, DNA-level diagnostic indicators also may be developed.
Presently, however, a genetic diversity indicator will be most useful as an integrative indicator of genetic
effects imposed by multiple stressors. This suggests that the genetic diversity indicator will prove most
59
-------
Genetic Diversity as an
useful as one component of a multi-indicator regional or index-site assessment.
Ill I I
Although the point can be debated, of the currently available technology, microsatellite analysis, perhaps
combined with mitochondrial DNA analysis, is likely to provide the most useful information per unit
effort for both regional genetic diversity assessments and temporal genetic diversity monitoring. The
US EPA's Molecular Ecology Research Branch employs this methodology for most current and planned
assessments. Unfortunately, this strategy also requires relatively advanced instrumentation and techni-
cal expertise compared to other strategies. The EPA's ECBP pilot genetic diversity study (section 3.1)
was conceived and implemented to use RAPD fingerprinting precisely because of concerns about the
ability to transfer genetic diversity indicator technologies to end-users. Because RAPD fingerprinting is
technically simple, it was assumed that it would have the greatest prospects for technology transfer.
However, during the course of the pilot study it became clear that the trade-off that comes with ease of
implementation is that of extreme sensitivity to minor variations in laboratory techniques. Similar con-
cerns are echoed in the scientific literature (see Perez et al., 1997). Concerns about repeatability
between different laboratories appear to be great enough to negate any perceived technology transfer
advantages. Current work with AFLP fingerprints indicates much less concern with repeatability, but
technology transfer is considerably more difficult than for RAPD analysis. Transfer of allozyme tech-
nology should not be difficult, but the needs for lethal sampling and methods to transport samples at
ultra-cold temperatures limit its general application.
While complete protocols for development and analysis of microsatellite markers may be difficult to
transfer to environmental labs, certain aspects of the analysis are relatively straightforward. As men-
tioned, methods for extracting DNA from animal tissues have been commercialized and are now sold by
several vendors as kits. The quality and quantity of DNA extracted using commercial kits is typically
very high. In addition, improvements in thermal cycler technology, together with packaging of PCR
reagents as standard assay kits have greatly eased technology transfer of PCR. Several vendors now sell
test kits for genotyping domestic animals and humans with fluorescent microsatellite DNA markers.
This suggests that regional field labs could accomplish major parts of the genetic diversity analysis
while one or two "core" molecular biology labs handle other aspects. A model for this approach is
shown in Figure 4-10. The regional lab, perhaps with some advice from a genetic analysis lab, would
design the assessment, which presumably will be performed in conjunction with other ecological indica-
tors. A marker development laboratory can then be employed to develop a panel of markers that are
appropriate to the assessment. The regional lab can then collect samples and prepare DNA for analysis.
The actual PCR reactions can be performed by the regional lab or by the genetic analysis lab. The
genetic analysis lab performs the genetic analysis and derives the genetic diversity interpretation. The
regional and genetic analysis labs can then assess the ecological significance of the genetic diversity
indicator data.
Under this scenario, the marker development laboratory could be a commercial laboratory or a laborato-
ry internal to the EPA. The EPA currently has the required expertise for such a laboratory in the
Molecular Ecology Research Branch of NERL. A laboratory internal to the EPA, possibly the same lab
that is used for marker development, or a laboratory under contract to the EPA could perform the genetic
analysis.
60
-------
I~TT
Genetic Diversity as an Indicator of Ecosystem Condition and Sustainability
Regional Laboratory
Marker Development Laboratory
Design Assessment
Field Sampling
DNA extraction
(PCR?)
• Develop and test
microsatellites, other
markers
• (PCR)
• Marker screening
• Genetic Diversity
assessment
Ecological
interpretation
Genetic Analysis Laboratory
Figure 4-10. A model showing how three different labs, the regional field lab, a genetic analysis lab, and a marker
development lab could interact to apply a genetic diversity indicator such as microsatellites or mitochondrial DNA
sequences to an environmental assessment.
61
-------
Genetic Diversity as an
Glossary
Ill I I
Adaptation: (noun) A genetically determined trait that enhances the ability of an organism to survive
and reproduce in its environment; (verb) the evolutionary process by which a population undergoes pro-
gressive genetic modification to increase its ability to survive and reproduce in a given environment.
Allele: one of several alternative forms of a gene that differs from other forms by a mutation in the
DNA sequence.
Codominance: the condition where both alleles at a locus are expressed and influence the observed
properties of the heterozygote.
Diploid: An organism that normally has two sets of chromosomes, and thus two copies of each gene.
Dominance: the condition in which one allele (the dominant allele) masks the expression of the other
(the recessive allele), so that heterozygotes cannot be distinguished from individuals that are homozy-
gous for the dominant allele.
Effective population size: The size of an "ideal population" (with a mathematically simplified breeding
structure) that loses genetic diversity at a rate equivalent to the population under study. The effective
population size is typically smaller than the census size because of differences in the breeding success of
individuals, skewed sex ratios, variation in population size, and other factors. The effective population
size provides a measure of how fast genetic diversity is being lost from the population under study.
Electrophoresis: The separation of DNA, RNA, or proteins by differential migration through a matrix
in the presence of electric current. Electrophoretic mobility is determined by differences among mole-
cules in size, charge, or shape.
Fitness: The genetic contribution by an individual's descendants to future generations of a population.
Individuals with greater fitness have genotypes that are better matched to the environment than those
with lower fitness.
Gene flow: The exchange of genes between generation via migration of individuals that will eventually
breed.
Genetic diversity: Variation among individuals for some heritable trait. Intraspecific genetic diversity
may be partitioned into at least two components: genetic diversity within populations and genetic diver-
sity among populations.
Genetic drift: Stochastic change in allele frequency due to random variations in the contributions of
breeding adults to the next generation. Since one consequence of genetic drift is that one or more alleles
will not be passed to the next generation, genetic drift results in a decrease in genetic diversity.
Genetic marker: an easily discerned attribute or probe that indicates the genotype of an individual at a
locus. Examples include molecular markers (usually DNA-based), cytological markers, and morpholog-
ical markers
Genome: The complete set of genetic information contained within an individual
62
-------
Condition and Sustainability
Genotype: The genetic make-up of an organism, typically with respect to one or a few genes of interest,
as distinguished from its appearance or phenotype.
Haplotype: A mitochondrial DNA genotype. One of several forms of the circular mitochondrial DNA
molecule that differs from other forms due to mutations in the DNA sequence.
Heterozygote: For diploid organisms, an individual whose cells contain two forms (alleles) of a gene,
one derived from each parent.
Heterozygosity: The proportion of individuals in a population that are heterozygous (possess more than
one allele) at a locus
Inbreeding: Mating among related individuals.
Inbreeding depression: The deterioration in a population's fitness and other traits due to a reduction in
genetic diversity within populations.
Loci: Plural of locus.
Locus: A site on a chromosome where a gene or other measurable variation resides. The DNA at this
chromosomal site may or may not have any functional significance to the organism.
Mitochondria: self-replicating organelles found in the cytoplasm of all eukaryotic cells that produce
energy via oxidative phosphorylation.
Molecular marker: An easily discerned measure of the genotype of an individual at a locus based on
molecular biological methods. Molecular markers are usually based on attributes of DNA, but may be
based on RNA or protein.
Monomorphic: State in which virtually all individuals have the same genotype at a locus. It is a prop-
erty of a locus in a population.
Neutral marker: A genetic marker derived from a locus that has a negligible effect on the ability of the
organism to survive and reproduce.
Outbreeding depression: The deterioration in a population's fitness or other traits due to immigration of
individuals from other populations that disrupts local adaptations.
Phenotype: The observable structural and functional characteristics of an organism that result from the
interaction of the genotype with the environment; the outward appearance of the organism.
Philoptatry: The tendency of species or groups to remain in or habitually return to their native regions
or territories. Organisms that are highly philopatric tend to stay in a defined geographic location.
Phylogeography: Reconstruction of biogeographic relationships, usually among populations within a
species, by identifying the phylogenetic relationship between a genetic trait or set of traits in relation to
geography.
63
-------
Genetic Diversity as an
Ill I I
Polymerase chain reaction (PCR): A process for the exponential amplification of a specific region of
DNA using DNA primers that flank the region of interest and a DNA polymerase to catalyze the reac-
tion.
Polymorphic: State in which individuals in a population have more than one genotype at a locus. It is a
property of a locus in a population.
Polymorphism: One of several forms of a genetic characteristic at a locus in a population. Also, the
presence in a population of two or more relatively common forms of a gene, chromosome, or genetically
determined trait.
Population: A group of conspecific organisms that occupy a more or less well defined geographic
region and exhibit reproductive continuity from generation to generation; it is generally presumed that
ecological and reproductive interactions are more frequent among these individuals than between them
and the members of other populations of the same species.
Recessive: An allele whose properties are not observed because they are masked by the expression of a
dominant allele. Thus, it is only expressed in homozygotes.
Trait: Any aspect of the appearance, behavior, development, biochemistry, or other feature of an organ-
ism.
64
-------
Condition and Sustainability
References
Allendorf, F.W., and Leary, R.F. (1986) Heterozygosity and fitness in natural populations of
animals. In Soule (ed.) Conservation Biology: The Science of Scarcity and Diversity.
Sinauer. Sunderland, MA. Pp. 57-76.
Avise, J.C. (1994) Molecular Markers, Natural History and Evolution. Chapman and Hall,
New York.
Ayres, D.R., Garcia-Rossi, D., Davis, H.G., and Strong, D.R. (1999) Extent and degree of
hybridization between exotic (Spartina altemiflord) and native (S. foliosd) cordgrass (Poaceae)
in California, USA determined by random amplified polymorphic DNA (RAPDs).
Molecular Ecology, 8, 1179-1186.
Baverstock, P.R., and Moritz, C. (1996). Project design. In: Hillis, D. M., Moritz, C. Mable,
B.K. (eds) Molecular Systematics. Second Edition. Sinauer Associates, Sunderland MA.
Pp. 17-27.
Beardmore, J.A., Barker, C.J., Battaglia, B., Berry, R.J., Crosby Longwell, A., Payne, J.F.,
and Rosenfield, A. (1980) The use of genetical approaches to monitoring biological effects
of pollution. Rapports et Proces-Verbaux des Reunions, 179, 299-305.
Ben-Shlomo, R. and Nevo, E. (1988) Isozyme polymorphism as monitoring of marine environments:
the interactive effect of cadmium and mercury pollution on the shrimp, Palaemon elegans.
Marine Pollution Bulletin, 19, 314-317.
Benton, M.J. and Guttman, S.I. (1992a) Allozyme genotype and differential resistance to mercury
pollution in the caddisfly, Nectopsyche albida \. Single-locus genotypes. Canadian Journal
of Fisheries and Aquatic Sciences, 49, 142-146.
Benton, M.J. and Guttman, S.I. (1992b) Allozyme genotype and differential resistance to mercury
pollution in the caddisfly, Nectopsyche albida II. Multilocus genotypes. Canadian Journal
of Fisheries and Aquatic Sciences, 49, 147-149.
Benton, M.J., Diamond, S.A., and Guttman, S.I. (1994) A genetic and morphometric comparison
oftfelisoma trivolvis and Gambusia holbrooki from clean and contaminated habitats.
Ecotoxicology and Environmental Safety, 29, 20-37.
Bickham, J.W., Sandhu, S., Hebert, P.D.N., Chikhi, L., and Athwal, R. (2000) Effects of chemical
contaminants on genetic diversity in natural populations: implications for biomonitoring and
ecotoxicology. Mutation Research, 463, 33-51.
Bouzat, J.L., Cheng, H.H., Lewin, H.A., Westemeier, R.L., Brawn, J.D., and Paige, K.N. (1998)
Genetic evaluation of a genetic bottleneck in the greater prarie chicken. Conservation Biology,
12, 836-843.
Brown Sullivan, K., and Lydy, M.J. (1999) Differences in survival functions of mosquitofish
(Gambusia affinis) and sand shiner (Notropis ludibundus) genotypes exposed to pesticides.
Environmental Toxicology and Chemistry, 18, 906-911.
Cabe, P.R. (1998) The effects of founding bottlenecks on genetic variation in the European
starling (Sturnus vulgaris) in North America. Heredity, 80, 519-525.
Caetano-Anolles, G and Gresshoff, P.M. (eds) (1997) DNA Markers: Protocols, Applications,
and Overviews. Wiley-Liss, Inc., New York.
Campton, D.E., and Johnston, J.M. (1985) Electrophoretic evidence for a genetic admixture of
native and nonnative rainbow trout in Yakima River, Washington. Transactions of the
American Fisheries Society, 114, 782-793.
Carter, PA., and Watt, W.B. (1988) Adaptation at specific loci. V. Metabolically adjacent enzyme
loci may have very distinct experiences of selective pressures. Genetics,\\9, 913-924.
Carvalho, GR. (1988) Molecular ecology: origins and approaches. In: Carvalho, GR. (ed.) Advances in
Molecular Ecology, IOS, Amsterdam. Pp. 1-24.
65
-------
Genetic Diversity as an
Ill I I
Chagnon, N.L. and Guttman, S.I. (1989) Differential survivorship of allozyme genotypes in
mosquitofish populations exposed to copper or cadmium. Environmental Toxicology
and Chemistry, 8, 319-326.
Cole, B.J., and Wiernasz, B.C. (1999) The selective advantage of low relatedness. Science,
285, 891-893.
Coltman, D.W., Bowen, W.D., Wright, J.M. (1998) Birth weight and neonatal survival of harbour
seal pups are positively correlated with genetic variation measured by microsatellites.
Proceedings of the Royal Society of London B, 265, 489-495.
Diamond, S.A., Newman, M.C., Mulvey, M., Dixon, P. M., and Martinson, D. (1989) Allozyme
genotype and time to death of mosquitofish, Gambusia affinis (Baird and Girard), during
acute exposure to inorganic mercury. Environmental Toxicology and Chemistry, 8, 613-622.
Duan, Y., Guttman, S.I., Oris, J.T., and Bailer, J.A. (2000a) Genotype and toxicity relationships among
Hyalella azteca: \. Acute exposure to metals or low pH. Environmental Toxicology and
Chemistry, 19, 1414-1421.
Duan, Y., Guttman, S.I., Oris, J.T., Huang, X., and Burton, GA. (2000b) Genotype and toxicity
relationships among Hyalella azteca: II. Acute exposure to fluoranthene-contaminated sediment.
Environmental Toxicology and Chemistry, 19, 1422-1426.
Ellegren, H., Lindgren, G, Primmer, C.R., and Moller, A.P. (1997) Fitness loss and germline
mutations in barn swallows breeding in Chernobyl. Nature, 389, 593-596.
Excoffier, L., Smouse, P.E., and Quattro, J.M. (1992) Analysis of molecular variance inferred from
metric distances among DNA haplotypes: Applications to human mitochondrial DNA
restriction data. Genetics, 131, 479-491.
Ford, T., Sorci, J., Ika, R., and Shine, J. (1998) Interactions between metals and microbial
communities in New Bedford Harbor, Massachusetts. Environmental Health Perspectives,
106 (Supplement 4),1033-1039
Fore, S.A., Guttman, S.I, Bailer, A.J., Altfeter, D.J., and Counts, B.V. (1995a) Potential use of
genetic diversity as an indicator of water quality. I. Pimephales notatus. Ecotoxicology
and Environmental Safety, 30, 24-35.
Fore, S.A., Guttman, S.I, Bailer, A.J., Altfeter, D.J., and Counts, B.V. (1995b) Potential use of
genetic diversity as an indicator of water quality. II. Campostoma anomalum. Ecotoxicology
and Environmental Safety, 30, 36-46.
Frankham, R. (1995) Inbreeding and extinction: A threshold effect. Conservation Biology, 9, 792-799.
Franklin, A.R. (1980) Evolutionary change in small populations. In Soule, M.E., and Wilcox,
B.A. (eds.) Conservation Biology: An Evolutionary-Ecological Perspective. Sinauer,
Sunderland, MA. Pp 135-149.
Gerlach, G. and Musolf, K. (2000) Fragmentation of landscape as a cause for genetic subdivision
in bank voles. Conservation Biology, 14, 1066-1074.
Gillespie, R.B. and Guttman, S.I. (1989) Effects of contaminants on the frequencies of allozymes
in populations of the central stoneroller. Environmental Toxicology and Chemistry, 8, 309-317.
Gillespie, R.B. and Guttman, S.I. (1993) Correlations between water quality and frequencies of
allozyme genotypes in spotfin shiner (Notropis spilopteris} populations. Environmental
Pollution, 81, 147-150.
Gillespie, R.B., and Guttman, S.I. (1999) Chemical-induced changes in the genetic structure of
populations: effects on allozymes. In: Forbes, V.E. (ed.) Genetics and Ecotoxicology. Taylor
and Francis, Philadelphia. Pp. 55-77.
Gilpin, M.E. and Soule, M.E. (1986) Minimum viable populations: processes of species extinction. In
Soule (ed.) Conservation Biology: The Science of Scarcity and Diversity. Sinauer. Sunderland,
MA. Pp 19-34.
66
-------
Condition and Sustainability
Gyllensten, U.R., Leary, R.F., Allendorf, F.W., and Wilson, A.C. (1985) Introgression between two
cutthroat trout species with substantial karyotypic, nuclear, and mitochondrial genomic
divergence. Genetics, 111, 905-915.
Haig, S.M., Bowman, R. and Mullins, T.D. (1996) Population structure of red-cockaded woodpeckers
in south Florida; RAPDs revisited. Molecular Ecology, 5, 725-734.
Hartl, D.L. and Clark, A.G. (1997) Principles of Population Genetics. 3rd edition. Sinauer,
Sunderland, MA.
Hessel, D.L. (1992) Genetic resource program strategic plan. U.S. Forest Service, Timber
Management. Washington, B.C.
Heagler, M.G, Newman, M.C. , Mulvey, M., and Dixon, P.M (1993) Allozyme genotype in
mosquitofish, Gambusia holbrooki, during mercury exposure. Temporal stability, concentration
effects and field verification. Environmental Toxicology and Chemistry, 12, 385-395.
Heithaus, M.R., and Laushman, R.H. (1997) Genetic variation and conservation of stream fishes:
influence of ecology, life history, and water quality. Canadian Journal of Fisheries and
Aquatic Sciences, 54, 1822-1836.
Hillis, D. M., Moritz, C. Mable, B.K. (1996) Molecular Systematics. Second Edition.
Sinauer Associates, Sunderland MA.
Hoelzel, A.R. (1999) Impact of population bottlenecks on genetic variation and the importance of
life-history; A case study of the northern elephant seal. Biological Journal of the Linnean
Society, 68, 23-39.
Hoelzel, A. R., Fleischer, R. C., Campagna, C., Le Boeuf, B. J., and Alvord, G (2002). Impact of
a population bottleneck on symmetry and genetic diversity in the northern elephant seal.
J. Evol. Biol. 15, 567-575.
Hughes, J.M., Harrison, D.A., and Arthur, J.M. (1991) Genetic variation at the PGI locus in the
mosquitofish Gambusia affinis (Poeclidae) and possible effect on susceptibility to an
insecticide. Biological Journal of the Linnean Society, 44, 153-167.
Jorde, P.E., and Ryman, N. (1996) Demographic genetics of brown trout (Salmo trutta) and estimation
of effective population size from temporal change of allele frequencies. Genetics, 143,
1369-1381.
Keklak, M.M., Newman, M.C., and Mulvey, M. (1994) Enhanced uranium tolerance of an exposed
population of eastern mosquitofish (Gambusia holbrooki Girard 1859). Archives of
Environmental Contamination and Toxicology, 27, 20-24.
Knutsen, H., Rukke, B.A., Jorde, P.E., and Ims, R.A. (2000) Genetic differentiation among
populations of the beetle Bolitophagus reticulatus (Coleoptera : Tenebrionidae) in a
fragmented and a continuous landscape. Heredity, 84, 667-676.
Kopp, R.L., Guttman, S.I., and Wissing, T.E. (1992) Genetic indicators of environmental stress in
central mudminnow (Umbra limi) populations exposed to acid deposition in the
Adirondack Mountains. Environmental Toxicology and Chemistry, 11, 665-676.
Krane, D.E., Sternberg, B.C., and Burton, GA. (1999) Randomly amplified polymorphic DNA
profile-based measures of genetic diversity in crayfish correlated with environmental impacts.
Environmental Toxicology and Chemistry, 18, 504-508.
Lande, R. (1995) Mutation and conservation. Conservation Biology, 9, 782-791.
Lande, R. (1999) Extinction risks from anthropogenic, ecological, and genetic factors. In L.F.
Landweber and A.P. Dobson (eds.) Genetics and the Extinction of Species. Princeton University
Press. Princeton, NJ. Pp 1-22.
Lande, R. and Shannon, S. (1996) The role of genetic variability in adaptation and population
persistence in a changing environment. Evolution, 50, 434-437.
Lavie, B., and Nevo, E. (1982) Heavy metal selection f phosphoglucose isomerase allozymes in
marine gastropods. Marine Biology, 71, 17-22.
67
-------
Genetic Diversity as an
Ill I I
Lavie, B. and Nevo, E. (1986a) Genetic selection of homozygote allozyme genotypes in marine
gastropods exposed to cadmium pollution. The Science of the Total Environment, 57, 91-98.
Lavie, B. and Nevo, E. (1986b) The interactive effects of cadmium and mercury pollution on
allozyme polymorphisms in the marine gastropod Cerithium scabridum. Marine Pollution
Bulletin, 17, 21-23.
Larno, V., Laroche, J., Launey, S., Flammarion, P., and Devaux, A. (2001). Responses of
Chub (Leuciscus cephalus) populations to chemical stress, assessed by genetic markers,
DNA damage and Cytochrome P4501A induction. Ecotoxicology. 10, 145-158.
Leary, R.F., Allendorf, F.W., and Sage, O.K. (1995) Hybridization and introgression between introduced
and native fish. In (Schramm, H.L. Jr. and Piper, R.G.) Uses and Effects of Cultured Fishes in
Aquatic Ecosystems. American Fisheries Society Symposium 15, Bethesda, MD.
Leberg, P.L. (1992) Effects of population bottlenecks on genetic diversity as measured by
allozyme electrophoresis. Evolution, 46, 447-494.
Leonard, A.C., Franson, S.E., Hertzberg, VS., Smith, M.K., and Toth, G.P. (1999) Hypothesis
testing with the similarity index. Molecular Ecology, 8, 2105-2114.
Lynch, M. (1990) The similarity index and DNA fingerprinting. Molecular Biology and Evolution,
75 478-484.
Lynch, M. (1991) Analysis of population genetic structure by DNA fingerprinting. In Burke, T, Dolf,
G., Jeffreys, A.J., and Wolff, R. (eds.) DNA Fingerprinting: Approaches and Applications.
Birkhauser Verlag, Boston.
Lynch, M. (1996) A quantitative-genetic perspective on conservation issues. In Avise, J.C. (ed.),
Conservation Genetics: Case Histories from Nature, Chapman Hall, New York. Pp 471-501.
Ma, X. L., Cowles, D. L., and Carter, R. L. (2000). Effect of pollution on genetic diversity in the
bay mussel Mytilus galloprovincialis and the acorn barnacle Balanus glandula. Marine
Environmental Research. 50, 559-563.
Mariette, S., Lecorre, V. and Kremer, A. (1999) Sampling within the genome for measuring
within-population diversity: trade-offs between markers. In Molecular Tools for Biodiversity.
Compendium of the Research Project: Development, optimization and validation of molecular
tools for assessment of biodiversity in forest trees (EM Gillet,ed.) European Union DGXII
Biotechnology FW IV Research Programme.
May, B. (1992) Starch gel electrophoresis of allozymes. In: Hoelzel, A.R. (ed.) Molecular Genetic
Analysis of Populations: A Practical Approach. 2nd Ed. Oxford University Press. New York.
Pp. 1-28.
McCormick, F.H., and Peck, D.V. (2000) Application of the indicator evaluation guidelines to a
multimetric indicator of ecological condition based on stream fish assemblages. Chapter 4 in:
U.S. Environmental Protection Agency. Evaluation Guidelines for Ecological Indicators. Office
of Research and Development. EPA/620/R-99/005. U.S. Environmental Protection Agency,
Washington, D.C.
Messer, J.J., Linthurst, R.A., and Overton, W.S. (1991) An EPA program for monitoring ecological
status and trends. Environmental Monitoring and Assessment, 17, 67-78.
Moraga, D., Mdelgi-Lasram, E., Romdhane, M. S., EL Abed, A., Boutet, I., Tanguy, A., and Auffret,
M. (2002). Genetic responses to metal contamination in two clams: Ruditapes decussates
and Ruditapesphilippinarum. Marine Environmental Research. 54, 521-525.
Morden, C.W. and Loeffler, W. (1999) Fragmentation and genetic differentiation among subpopulations
of the endangered Hawaiian mint Haplostachys haplostachya (Lamiaceae). Molecular Ecology,
8, 617-625.
Moritz, C. (1994) Defining 'evolutionarily significant units' for conservation. Trends in Ecology
and Evolution, 9, 373-375.
68
-------
Condition and Sustainability
Moritz, C. (1995) Uses of molecular phytogenies for conservation. Philosophical Transactions of
the Royal Society of London, Series B, 349, 113-118.
Moritz, C. and Faith, D.P. (1998) Comparative phylogeography and the identification of genetically
divergent areas for conservation. Molecular Ecology, 7, 419-429.
Mulvey, M., Newman, M.C., Chazal, A., Kelak, M.M., Heagler, M.G., and Hales, L.S. Jr. (1995) Genetic
and demographic responses of mosquitofish (Gambusia holbrooki Girard 1859) populations
stressed by mercury. Environmental Toxicology and Chemistry, 14, 1411-1418.
Murdoch, M.H., and Hebert, P.D.N. (1994) Mitochondrial DNA diversity of brown bullhead from
contaminated and relatively pristine sites in the Great Lakes. Environmental Toxicology and
Chemistry, 13, 1281-1289.
Murphy, R.W., Sites, J.W. Jr., Buth, D.G, and Haufler, C.H. (1996). Proteins: Isozyme electrophoresis.
In: Hillis, D. M., Moritz, C. Mable, B.K. (eds) Molecular Systematics. Second Edition. Sinauer
Associates, Sunderland MA. Pp. 51-120.
Nadig, S.G, Lee, K.L., and Adams, S.M. (1998) Evaluating alterations of genetic diversity in sunfish
populations exposed to contaminants using RAPD assay. Aquatic Toxicology, 43, 163-178.
Nehlsen, W., Williams, J.E., and Lichatowich, J.A. (1991) Pacific salmon at the crossroads: Stocks at
risk from California, Oregon, Idaho, and Washington. Fisheries, 16, 4-21.
Nei, M. (1972) Genetic distance between populations. American Naturalist, 106, 283-292.
Nei, M. (1987) Molecular Evolutionary Genetics. Columbia University Press, New York, NY, USA.
Nevo, E., Ben-Shlomo, R. and Lavie, B. (1984) Mercury selection of allozymes in marine organisms:
Predictions and verification in nature. Proceedings of the National Academy of Sciences USA,
81, 1258-1259.
Nevo, E., Lavie, B., and Ben-Shlomo,R. (1988) Population genetic structure of marine organisms
as detectors and monitors of marine pollution. FAO/UNEP Scientific Reports, 24, 37-49.
Newman, M.C., Diamond, S.A., Mulvey, M., and Dixon, P. (1989) Allozyme genotype and time
to death of mosquitofish, Gambusia afftnis (Baird and Girard), during acute toxicant exposure:
a comparison of arsenate and inorganic mercury. Aquatic Toxicology, 15, 141-156.
Omernick, J.M. (1995) Ecoregions: A framework for environmental management. In Davis, W. and
Simon, T. (eds.) Biological Assessment and Criteria: Tools for Water Resource Planning and
Decision Making. Lewis Publishers, Boca Raton, FL. Pp 49-62.
Palumbi, S.R. (1996) Nucleic Acids II: The polymerase chain reaction. In: Hillis, D. M., Moritz,
C. Mable, B.K. (eds) Molecular Systematics. Second Edition. Sinauer Associates,
Sunderland MA. Pp. 205-247.
Perez, T., Albornoz, J, and Dominguez, A. (1998) An evaluation of RAPD fragment reproducibility
and nature. Molecular Ecology, 1, 1347-1357.
Quattro, J.M. and Vrijenhoek, R.C. (1989) Fitness differences among remnant populations of the
endangered Sonoran topminnow. Science, 245, 976-978.
Roark, S., and Brown, K. (1996) Effects of metal contamination from mine tailings on allozyme
distributions of populations of Great Plains fishes. Environmental Toxicology and Chemistry,
15, 921-927.
Roark, S.A., Andrews, J. F., and Guttman, S. I. (2001). Population genetic structure of the
western mosquitofish, Gambusia affinis, in a highly channelized portion of the San Antonio river
in San Antonio, TX. Ecotoxicology, 10, 223-227.
Rodriguez-Clark, K.M. (1999) Genetic theory and evidence supporting current practices in captive
breeding for conservation. In L.F. Landweber and A.P. Dobson (eds.) Genetics and the
Extinction of Species. Princeton University Press. Princeton, NJ. Pp 47-74.
Rolan-Alverez, E., Zapata, C., and Alvarez, G (1995) Multilocus heterozygosity and sexual selection
in a natural population of the marine snail Littorina mariae (Gastropoda: Prosobranchia).
Heredity, 75, 17.25.
69
-------
Genetic Diversity as an
Ill I I
Roy, N.K., Konkle, B., Wirgin, I. (1996) Characterization of CYP1A1 gene regulatory elements in
cancer-prone Atlantic tomcod. Pharmacogenetics, 6, 273-277.
Saccheri, I., Kuussaari, M., Kankare, M., Vikman, P., Fortelius, W., Hanki, I. (1998) Inbreeding
and extinction in a butterfly metapopulation. Nature, 392, 491-494.
Schlueter, M.A, Guttman, S.I., Oris, J.T., and Bailer, A.J. (1995) Survival of copper-exposed
juvenile fathead minnows (Pimephales promelas) differs among allozyme genotypes.
Environmental Toxicology and Chemistry, 14, 1727-1734.
Schlueter, M.A, Guttman, S.I., Oris, J.T., and Bailer, A.J. (1997) Differential survival of fathead
minnows, Pimephales promelas, as affected by copper exposure, prior population stress,
and allozyme genotypes. Environmental Toxicology and Chemistry, 16, 939-947.
Schlueter, M.A, Guttman, S.I., Duan, Y, Oris, J.T., Huang, X., and Burton, G A. (2000) Effects of
acute exposure to fluoranthene-contaminated sediment on the survival and genetic variability
of fathead minnows (Pimephales promelas). Environmental Toxicology and Chemistry, 19,
1011-1018.
Schneider, C.J., Cunningham, M., and Moritz, C. (1998) Comparative phylogeography and the
history of endemic vertebrates in the Wet Tropics rainforests of Australia. Molecular Ecology,
75 487-498.
Shaw, A.J. (1998) The evolution of heavy metal tolerance in plants: adaptation, limits, and costs.
In. Forbes, V.E. (ed.) Genetics and Ecotoxicology. Taylor and Francis, Philadelphia. Pp. 9-30.
Slate, J., Kruuk, L.E.B., Marshall, T.C., Pemberton, J.M., Clutton-Brock, T.H. (2000)
Inbreeding depression influences lifetime breeding success in a wild population of red deer
(Cervus elaphus), Proceedings of the Royal Society Biological Sciences, SeriesB, 267,
1657-1662.
Smith, M.W., Smith, M.H, and Chesser, R.K. (1983) Biochemical genetics of mosquitofish. I.
Environmental correlates, a temporal and spatial heterogeneity of allele frequencies within a
drainage. Copeia, 1, 182-193.
Soule, M.E. (1980) Thresholds for survival: Maintaining fitness and evolutionary potential. In M.E.
Soule and B.A. Wilcox (eds.), "Conservation Biology" An Evolutionary-Ecological Perspective.
Sinauer, Sunderland, MA. Pp. 151-169.
Soule, M.E. (1987) Viable Populations for Conservation. Cambridge University Press, Cambridge, U.K.
Street, G.T. and Montagna, PA. (1996) Loss of genetic diversity in harpacticoida near offshore
platforms. Marine Biology, 126, 271-282.
Tanguy, A., Fernandez-Castro, N., Marhic, A., and Moraga, D. (1999) Effects of an organic
pollutant (tributytin) on genetic structure in the Pacific Oyster Crassostrea gigas.
Marine Pollution Bulletin, 38, 550-559.
Tatara, C.P., Mulvey, M., and Newman, M.C. (1999) Genetic and demographic responses of
mosquitofish (Gambusia holbrooM) populations exposed to mercury for multiple generations.
Environmental Toxicology and Chemistry, 18, 2840-2845.
Theodorakis, C.W., and Shugart, L.R., (1997) Genetic Ecotoxicology :II. Population genetic structure
in radionuclide-contaminated mosquitofish {Gambusia affinis). Ecotoxicology, 6, 335-354.
Theodorakis, C.W., and Shugart, L.R., (1998) Natural selection in contaminated environments: A
case study using RAPD genotypes. In: Forbes, V.E. (ed.) Genetics and Ecotoxicology. Taylor
and Francis, Philadelphia. Pp. 123-149.
U.S. Environmental Protection Agency (1988) Future Risk: Research Strategies for The 1990s.
Science Advisory Board. SAB-EC-88-040. US Environmental Protection Agency.
Washington, B.C.
70
-------
Condition and Sustainability
U.S. Environmental Protection Agency: Klemm, D.J., and Lazorchak, J.M. (1993) Environmental
Monitoring and Assessment Program. Surface Waters and Region 3 Regional Environmental
Monitoring and Assessment Program. 1993 Pilot Field Operation and Methods Manual.
Streams, USEPA, ORD, NERL-Cincinnati, Washington, D.C.
U.S. Environmental Protection Agency (1997) An Ecological Assessment of the United States
Mid-Atlantic Region. EPA/600/R-97/130. U.S. Environmental Protection Agency,
Washington, D.C.
U.S. Environmental Protection Agency (1998) Ecological Research Strategy. Office of Research
and Development. EPA/600/R-98/086. U.S. Environmental Protection Agency,
Washington, D.C.
Vos , P., Rogers, R., Bleeker, M. Reijans, M., van der Lee, T., Homes, M., Firjter, A., Pot, J.,
Peleman, J., Kuiper, M., and Zabeau, M. (1995) AFLP: a new technique for DNA fingerprinting.
Nucleic Acids Research, 23, 4407-4414.
Vrijenhoek, R.C. (1994) Genetic diversity and fitness in small populations. In Loeschcke, V., Tomiuk, J.,
and Jain, S.K. (eds) Conservation Genetics. Birkhauser Verlag Basel. Switzerland. Pp.37-53.
Waples, R.S. (1990) Conservation of Pacific salmon. III. Estimating effective population size.
Journal of Heredity, 81, 267-276.
Welsh, J. and McClelland, M. (1990) Fingerprinting genomes using PCR with arbitrary primers.
Nucleic Acids Research, 18, 7213-7218.
Westemeier, R.L, Brawn, J.D., Simpson, S.A., Esker, T.L, Jansen, R.W., Walk, J.W., Kerchner,
E.L. (1998) Tracking the long-term decline and recovery of an isolated population.
Science, 282, 1695-1698.
Williams, J.G., Kubelik, A.R, Livak, K.J., Rafalski, J.A., and Tingey, S.V. (1990) DNA
polymorphisms amplified by arbitrary primers are useful as genetic markers. Nucleic Acids
Research, 18,6531-6535.
Williams, R.N., Shiozawa, O.K., Carter, J.E., and Leary, R.F. (1996) Genetic detection of
putative hybridization between native and introduced rainbow trout populations of the
upper Snake River. Transactions of the American Fisheries Society, 125, 387-401.
Wright, S. (1978) Evolution and the Genetics of Populations. Vol. 4: Variability Within and
Among Natural Populations. University of Chicago Press. Chicago.
Yauk, C.L., and Quinn, J.S. (1996) Multilocus DNA fingerprinting reveals high rate of heritable
genetic mutation in herring gulls nesting in an industrialized urban site. Proceedings of
the National Academy of Sciences, USA, 93, 12137-12141.
71
-------
Genetic Diversity as an
Ill I I
Acknowledgements
A number of individuals and organizations were instrumental in collecting data and contributing to this
document. Tom Wessendarp and Tammy Goyke (USEPA) provided valuable contributions as part of the
genetic diversity team. Field crews from DynCorp, including Dan Williams, Scott Jacobs, Joseph
Loucek, Steve Harmon, Jason Jannot and Dennis McMullen helped to collect fish samples. The Ohio
EPA providing environmental data on the Eastern Cornbelt Plains study as part of a joint Ohio EPA and
USEPA Region 5 REMAP assessment. Environmental data for the Mid-Atlantic Region was provided by
field crews participating in the EMAP MAIA project. Laboratory analysis and data management was
provided by Rich Silbiger, Manju Garg, Ana Braam, Jodie Walker, Malika Humphries, Jared Smith, Paul
Weaver, Richard Converse, Sandra Millward, Shawn Mills and Tony Leonard of PAI/SoBran. Susan
Cormier, Bhagya Subramanian, and Annette Roth (USEPA) provided additional laboratory and data
support. Graphics support was provided by Keith Adams of Computer Sciences Corporation. Frank
McCormick, Jim Lazorchak and Kate Smith provided helpful discussions and support of these projects.
72
-------
Condition and Sustainability
Appendix 1: Laboratory and analytical procedures for RAPD analysis
DNA Extraction. For DNA purification, a portion of the frozen tail section (approximately 25 mg) was
homogenized in a 1.5 ml microfuge tube containing 100|il PBSET (standard PBS, lOOmM EDTA, 0.1%
Triton X). 400|il of PBSET, 21 \A Proteinase K (20mg ml -1) and 30|il of 20% SDS was added to the
homogenate. The tube was mixed by gentle inversion and incubated overnight at 65 °C. Following
digestion, the preparation was centrifuged at 12,000 rpm for 10 minutes to remove particulate and undi-
gested material. The supernatant was removed to a clean microfuge tube, incubated for 10 minutes at
91 °C to inactivate the Proteinase K, then allowed to cool to room temperature. 10 |il of RNAse(A) (20
mg ml •') was added to the supernatant and incubated for 30 minutes at 42 °C. 1 ml of ProCipitate was
added to the tube, which was then mixed by gentle inversion on a rotator for 5 minutes and centrifuged
at 14,000 rpm for 15 minutes at room temperature. To facilitate the separation of the DNA-containing
supernatant and the semi-solid, protein-containing pellet, a dab of light PhaseLock gel (5'-3', Inc.) was
added to the inside of the microfuge cap and the sample was centrifuged for an additional 15 minutes at
14,000 rpm at room temperature. The supernatant was carefully transferred to a 2.0 ml microfuge tube,
after which an equal volume of chloroform-isoamyl alcohol (24:1) was addedto the supernatant and
mixed by gentle inversion on a rotator for 15 minutes. The aqueous fraction was transferred to a new
1.5 ml microfuge tube and the DNA was precipitated by adding 1 |il of glycogen (20 mg ml'1) and an
equal volume of ice cold 100% isopropanol, mixing well by inverting, and incubating at -20 °C for 2
hours. The precipitated DNA was pelleted by centrifuging for 15 minutes at 15,000 rpm at 0 °C. The
supernatant was removed and the DNA washed with ice cold 70% ethanol. The DNA was again pelleted
by centrifuging for 15 minutes at 15,000 rpm at 0 °C. The supernatant was removed and the DNA
washed again with ice cold 95% ethanol. The DNA was pelleted one last time by centrifuging for 15
minutes at 15,000 rpm at 0 °C. The supernatant was removed and the DNA pellet was allowed to air
dry for 10 minutes at room temperature. After drying, the pellet was resuspended in 200 |il of IX PCR
buffer (lOmM Tris-HCl, 50mM Kcl, pH 8.3), warmed for 15 minutes in a wet block at 65 °C, then incu-
bated overnight at 37 °C in ensure complete resuspension. The DNA stock concentration was estimated
by comparing the fluorescence of 1000-fold dilutions of the stock with known standards both prepared
with PicoGreen fluorescent dye (Molecular Probes, Inc.) and scanned using a Fluorlmager 595
(Molecular Dynamics) with 488 nm excitation and 530df30 bandpass filter. Stock DNA was diluted to a
working concentration of 1.25 ng jil'1 with IX PCR buffer (lOmM Tris-HCl, 50mM Kcl, pH 8.3).
PCR. RAPD analysis used was modified slightly from the original descriptions (Williams et a/., 1990;
Welsh and McClelland, 1990). A total reaction volume of 25 |il was prepared by combining a 5 |il
aliquot of the PCR master mix (2mM Tris- HC1 (pH 8.5), 10 mM KC1, 3mM MgCl2, 1 pmole RAPD
primer, 2.0 units of Native Tag polymerase (Gibco BRL), 0.05% acetamide, ImM dNTP mix) with a 20
|ll aliquot of diluted DNA (1.25 ng (il'1) in a 0.2 ml PCR tube (individual or strip) and mixing well. All
PCR prep was done on ice. All media preparation and sample aliquoting was done using aerosol resist-
ant pipet tips. Samples were amplified by 34 cycles of 45 seconds at 94 °C, 1 minute at 41 °C, 1 minute
at 72 °C on a RoboCycler with Hot Top (Stratagene). Three decamer oligonucleotides (primers) were
used to generate RAPD profiles; OPF-04 (5'-GGTGATCAGG-3'), OPL-02 (5'-TGGGCGTCAA-3'),
OPL-05 (5'-ACGCAGGCAC-3'). Triplicate PCR reactions were performed, 95 samples plus 1 negative
control per replicate.
Gel Electrophoresis. A 15% Ficoll 400/0.25% bromophenol blue solution was added to the completed
RAPD products, along with 400 and 900 bp fragments and buffered salt solution; the additional DNA
fragments provided in-lane standards to aid in sizing. Samples were arranged on a 20-lane agarose gel
(1.65% w v-' containing IX TBE [89mM Tris Base, 89mM Boric Acid, lOmM Na2 EDTA-H20]) such
that each sample lane was placed adjacent to a molecular weight marker lane (500/100 bp ladder,
73
-------
Genetic Diversity as an
Ill I I
GenSura). Samples were electrophoresed at 7 °C, 6.8 V cm'1, for 4.5 hours with continuous buffer (0.6X
TBE) recirculation. The gels were subsequently stained in 1000ml of IX TBE with 4 |jg ml"1 of ethidi-
um bromide for 30 minutes with constant agitation and then destained in water for 15 minutes. The gels
were scanned using a Fluorlmager 595 (Molecular Dynamics) with 514 nm excitation and 590df30
bandpass filter, producing a visual image of the PCR products.
Image Analysis. Bands were declared using FragmeNT Analysis software (Molecular Dynamics) and
then further edited visually to eliminate both false bands and those bands below a level of optical densi-
ty that could be reliably scored. Sample band molecular weights were calculated by a point-to-point
logarithmic interpolation method within FragmeNT using the standard lane nearest to the unknown.
Marker Identification. Despite the large number of molecular weight ladders, sizing errors inevitably
occurred so fragment size alone is not a reliable indicator of band homology. Bands were classified
into size groups (bins) using cluster analysis. These bins were further refined by discriminant analysis
of band size and intensity characteristics. In excess of 200 such bins were identified; however, 53 bins
that were most repeatably scored (based on comparison of the three replicates for each individual) were
selected for analysis. The number of bins created ranged from 39 for primer F04 to 42 for primer L05.
Bins in which the average band intensity was low (less than 0.5% of total lane intensity) were excluded
from the analysis, as were bins in which the repeatability of scoring across replicates was poor. Once the
bins were created, each bin was assessed for each individual to determine whether the bins were "filled"
or empty. If a band was observed within the size bin for two of the three replicates scored then the bin
was considered filled for that particular individual.
Similarity Analysis. RAPD fingerprints were compared between each pair of individuals using Lynch's
(1990) similarity index (S), which is defined for two individuals, x and y, as Sxy=2Nxy/(Nx + Ny), where
Njy is the number of bins that are filled for both individuals, and Nx and Ny are the total number of bins
filled for individuals x and y, respectively. The average similarity of individuals within a population
(S w) provides an inverse measure of the genetic diversity within that population. Comparisons of aver-
age similarities among_individuals in the same populations to average similarities among individuals in
different populations (S^) provide a measure of population substructure. Hypothesis tests for significant
differences in Sw's were constructed and tested according to the method of Leonard et al. (1999).
Analogs of Nei's (1972) genetic distance estimate D' between each pair of populations were calculated
following the method of Lynch (1991). Population substructure was estimated using the estimator of
Wright's (1978) FST for DNA fingerprint data described by Lynch (1991). Estimates of the average
number of migrants among populations were made using the approximate equilibrium relationship
FST= !/(!+ 4Nem), where Ne is the effective size and m is the rate at which migrants are exchanged
between populations per generation. Nem, therefore, represents the average (effective) number of
migrants exchanged between populations each generation. Dendrograms relating genetic relationships
among populations were assessed using unweighted pair group means analysis (UPGMA).
Quality Assurance. Several quality assurance measures were employed throughout this study.
Incorporation of three replicates of each RAPD reaction into the design helped to ensure that RAPD fin-
gerprints were highly repeatable. Each set of 95 fish to be processed was assigned a blind sample num-
ber based on the tube assignment of a 96-well PCR plate (8 rows [A-H] x 12 columns [1-12]). This
sample number was carried through all steps of DNA extraction and quantitation, PCR, gel electrophore-
sis and scanning, FragmeNT analysis, and subsequent data handling. A SAS database was maintained to
record the dates and any comments of all aforementioned steps. To avoid any PCR block position bias,
the 95 samples and negative control within one PCR replicate were randomly assigned a position in the
96-well PCR block using a SAS-based random number generator. To avoid any gel lane position bias,
74
-------
Condition and Sustainability
all samples and the negative controls for one PCR replicate of 3 primers (3 primers x 96 DNA
samples/negative control = 288 PCR reactions per replicate) were randomly assigned a lane position on
a gel (288 PCR reactions / 12 reactions per gel = 24 gels). A pre-printed form listing the load order of
samples and specific gel numbers was used at the time of gel loading and any deviations in the assigned
order were noted and changes were made to the database. Following electrophoresis and gel scanning,
gel images were reviewed to verify that the samples were actually loaded in the lanes as assigned. The
appearance of bands in a gel lane designated as a negative control was used to indicate contamination
during the PCR preparation; any questionable replicate was discarded and repeated. Any PCR reactions
that failed to amplify any bands were noted and an additional replicate PCR was performed. Any gels
that produced a poor visual image due to problems associated with electrophoresis were run a second
time with the balance of the PCR reaction. SAS-based programs were employed to compare the original
gel load order assignments with those actually reported as an output from the FragmeNT Analysis; dis-
crepancies were flagged and reconciled where possible.
75
-------
Genetic Diversity as an
Appendix 2: Laboratory and Analytical Procedures for AFLP Analysis
Ill I I
DNA Extraction. DNA was extracted from a section of tail fin (25-50mg) using a commercial kit
(DNeasy, Qiagen) and quantitated using a fluorescently labeled nucleic acid stain (PicoGreen, Molecular
Probes).
AFLP digestion/ligation/amplification. 200 ng of genomic DNA was restriction digested and ligated
overnight at room temperature in a single 15 yd reaction as follows: IX T4 ligase buffer (50mM Tris
Hcl, lOmM MgCl2, lOmM dithiothreitol, ImM ATP), 50mM NaCl, 75 ng BSA, 5 units each Eco RI and
Mse I, 100 Weiss units T4 DNA ligase, 30 pmol Eco RI adapter and 300 pmol Mse I adapter. The
digestion-ligation products were diluted 10-fold with TLE [lOmM Tris-HCL, pH 7.6, O.lmM EDTA].
Preamplification reactions were performed using 5 (il of the diluted digestion-ligation product in a 20 (il
reaction with standard EcoRI (Eco+G) and Msel (Mse+C) adapter primers (30ng each), 20mM Tris
HCL, 50mM KC1, 1.5mM MgCl2, 0.2mM each dNTP, and 0.5 units DNATaq polymerase. The follow-
ing thermocycling profile was used for preamplification: 72°C for 2 min, 24 cycles of 94°C for 30 sec,
65°C for 30 sec, 72°C for 1 min, and ending with a final extension at 72°C for 30 min. The preamplifi-
cation products were diluted 10-fold with TLE (lOmM Tris-HCL, pH 7.6, O.lmM EDTA). Selective
amplification reactions were performed using a 5 (il of the diluted pre-amplification product in a 20 (il
reaction with 20mM Tris HCL, 50mM KC1, 1.5mM MgC12, 0.2mM each dNTP, 0.5 units DNA Taq
polymerase, 5 ng MseC+1 primer, 15 ng EcoG+2 primer with a 5' fluorescein attachment. Ten
EcoG+2/MseC+l primer combinations were used for selective amplification PCR conditions for the
selective amplification were as follows: 12 cycles of 94°C for 30 sec, 65°C for 30 sec (dropping 0.7°C
per cycle), 72°C for 1 min, followed by 24 cycles of 94°C for 30 sec, 56°C for 30 sec, 72°C for 1 min; a
final extension at 72°C for 30 min completed the PCR.
Gel Electrophoresis. Selective amplification products were diluted 1:1 with sequencing load dye
(deionized formamide, 0.5 mg/ml xylene cyanole); 28 samples and 5 lanes of a 10 bp size ladder were
electrophoresed on 5% denaturing polyacrylamide gels in 0.5X TBE buffer at 25W for 1.3 hours. Gels
were scanned using a Molecular Dynamics 595 Fluorlmager (488nm excitation, 530df30 filter) to visu-
alize the fluorescein-tagged amplification products.
Fragment Analysis. Each lane was assigned a score based on visual quality (0 = no amplification, 1 =
very few and/or light bands, 2 = only very intense bands are able to scored, 3 = dark bands are scorable,
4 = most bands are distinct over a range of intensity); lanes with a score <3 were not analyzed and those
samples were subjected to repeat amplification and electrophoresis. Images were evaluated using Cross
Checker (v 2.91) and a binary table indicating presence or absence for each sample/primer combo/mark-
er was generated. A second Cross Checker evaluation was made by another analyst and the resulting
binary scores were compared with the original interpretation. Replicates (DNA extraction through selec-
tive amplification) were done on 20% of the samples; binary scores from the replicate samples were
used to determine reliability of a particular AFLP marker. Those markers whose binary scores were in
disagreement, either between analysts or replicate samples were eliminated from the final analysis.
Within and among population similarities were calculated; pairwise t-tests of the within population simi-
larities were performed. A genetic distance matrix (Nei, 1987) based on the among- population similari-
ties was constructed to test for evidence of population subdivision. A molecular analysis of variance
(AMOVA, Excoffier et a/., 1992) was performed to partition genetic diversity within and among popula-
tions.
76
-------
Condition and Sustainability
Appendix 3: Laboratory and Analytical Procedures for DNA Sequence Analysis
Mitochondrial DNA amplification. Genomic DNA was amplifed with mitochondrial DNA primers for
part of the cytochrome B region. Reactions were performed using 1.5 yd of diluted genomic DNA (25
ng/(il) in a 20 (il reaction with 50 (il each of cytochrome B primers GLU (5'-TGA CTT GAA GAA
CCA CCG TT-3') and THR (5'ATC TTC GGA TTA CAA GAC CGA); 20mM Tris HCL, 50mM KC1,
1.5mM MgCl2, 0.2mM each dNTP, and 0.5 units DNA Taq polymerase (Gibco BRL). The following
thermocycling profile was used for preamplification: 94°C for 5 min, 28 cycles of 94°C for 40 sec, 52°C
for 30 sec, 72°C for 1 min, and ending with a final extension at 72°C for 5 min. The reaction products
were diluted 1:1 with buffered load dye containing bromophenol blue and were electorphoresed in 1%
agarose gels containing ethidium bromide in 0.6X TBE buffer at 250V for 2.5 hours. The band was
visualized under UV light, excised from the gel, and purified using QIAquick kit (Qiagen). The result-
ing amplification product was quantitated using using a fluorescently labeled nucleic acid stain
(PicoGreen, Molecular Probes) and diluted to a concentration of 2.5ng/ (il.
Mitochondrial DNA sequencing. Cycle sequencing reactions were performed in 20 (il using 1 (il of
template (2.5 ng/ (il purified cytochrome B reaction), 2 (il ABI Ready Reaction mix (ABI BigDye
Terminator [v.l] kit), 60mM Tris, 1.5mM MgCl2, and 0.8 M GLU primer. Positive control reactions
were similarly prepared using M13 primer and pGEM template. The following thermocycling profile
was used for cycle sequencing: 25 cycles of 94°C for 10 sec, 50°C for 5 sec, 60°C for 4 min. Excess
terminator dyes were eliminated by isopropanol precipitation of the product. The sample was resus-
pended in deionized formamide and capillary electrophoresed (Applied Biosystems 3100 Genetic
Analyzer; 50cm array, POP-6 polymer).
Sequence analysis. Sequences were reviewed in Sequence Analysis (v3.7, Applied Biosystems), edited
and aligned by hand. Genetic structure was analyzed by AMOVA and minimum spanning networks
using the software Arlequin (v2.0).
77
-------
Genetic Diversity as an Indicator of Ecasjfijjll
ill
78
-------
------- |