PE86-1C9485
EPA/60Q/1_85/013
June 1985
AN ISOTOPIC STUDY OF THE INHALATION TOXICOLOGY OF OXIDANTS
John M0 Hayes and Jeffrey Santrock
Indiana University
Department of Chemistry
Bloomington, IN 47405
Cooperative Agreements CR807322 and CR811261
HEALTH EFFECTS RESEARCH LABORATORY
OFFICE r~ RESEARCH AND DEVELOPMENT
U.S. ENVlrx IMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
TECHNICAL REPORT DATA
/Please read /ntinicnons un the reverse before completing)
1. REPORT NO. 2.
EPA/600/1-85/013
4. TITLE ANDSUBTITLE
AN ISOTOPIC STUDY OF THE
^INHALATION TOXICOLOGY OF OXIC&NTS
7. AUTHOR(S)
Dr. John M. Hayes, Dr. Jeffrey Santrock
^,
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Department of Chemistry
Indiana University
Bloomington, IN 47405
12. SPONSORING AGENCY NAME AND ADDRESS
Office of Research & Development
Health Effects Research Laboratory
USEPA
Research Triangle Park, NC 27711
3. RECIPIENT'S ACCESSION NO.
rr, ': 1 -; '. G'57/iS
5-R5KReT?9&
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NC
1O. PROGRAM ELEMENT NO.
C9GA1A
11. CONTRACT/GRANT NO.
CR807322 and CR311261
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA-600/11
is. SUPPLEMENTARY NOTES
16. ABSTRACT
The purpose of these studies was to develop novel methods to investigate the biological
fate of inhaled ozone and other oxygen-containing pollutants in animal and human tis-
sues using the heavy isotope of oxygen, oxygen-18 (180). Methods were developed which
facilitated the conversion of tissue oxygen to CO2 and the subsequent trapping of the
(X>2 so that it could be subjected to isotope-ratio mass spectranetry. The ratios of
the various masses of evolved COj were used to calculate the ±JO content of the orig-
inal .tissues, thus enabling the detection of isotopic enrichments as small as 0.4%.
The above procedures were performed by modification of a commercially available ele-
•».antal analyzer to include effluent columns and trapping devices, development of oxygen
isotopic standards, and by derivation of mathematical models for correction of blank
and memory effects originating during sample pyrolysis. These techniques were applied
with success to the determination of the biological fate of inhaled ozone, and to the !
measurement of tissue oxidation induced by a model peroxidation initiator, carbon I
tetrachloride.
17. KEY WORDS AND DOCUMENT ANALYSIS
». DESCRIPTORS
18. DISTRIBUTION STATEMENT
Release to Public
b. IDENTIFIERS/OPEN ENDED TERMS
19. SECURITY CLASS (This Report)
Unclassified
20. SECURITY CLASS (This page/
Unclassified
c. COSATI Field/Croup
21. NO.OR.PAGES
22.PHICE
CPA Ftrm 2220-1 (R»«. 4-77) PREVIOUS EDITION is OBSOLETE
-------
PE86-1C9485
EPA/60Q/1_85/013
June 1985
AN ISOTOPIC STUDY OF THE INHALATION TOXICOLOGY OF OXIDANTS
John M0 Hayes and Jeffrey Santrock
Indiana University
Department of Chemistry
Bloomington, IN 47405
Cooperative Agreements CR807322 and CR811261
HEALTH EFFECTS RESEARCH LABORATORY
OFFICE r~ RESEARCH AND DEVELOPMENT
U.S. ENVlrx IMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
NOTICE
This document has been reviewed in accordance with
U.S. Environmental Protection Agency policy and
approved for publication. Mention of trade names
or commercial products does not constitute endorse-
ment or recommendation for use.
11
-------
This is to certify that the thesis submitted by Jeffrey Santrock
has been passed by the Ph.D. Advisory Committee as satisfactory in
partial fulfillment of the requirements for the Ph.D. degree.
Chairman
CResearch Advisor)
QlxruKxtxix
iii
-------
TABLE OF CONTENTS
TABLE OF CONTENTS v
.ILLUSTRATIONS AND FIGURES vill
ACKNOWLEDGMENTS xi
CHAPTER 1 INHALATION TOXICOLOGY OF OXIDANTS "... 1
Inhalation Toxicology ., 3
Respirtory Transport and Absorption of Inhaled
Oxidants , 6
Airway Structure 6
Epithelial Cell Population 9
Flow Dynamics . . . . 11
4!
Chemical Mechanisms 12
f
Moleculat Targets 13
Ozonolysis 14
Aatoxidation ' 17
Mechanism of Tissue Damage 23
Isotopic Studies 24
Physiological Applications of Oxygen-18 .... 25
References for Chapter 1 28
Chapter 2 DETERMINATION OF OXYGEN-18 IN ORGANIC SUBSTANCES . 33
Oxygen Isotopic Standards 35
Pyrolysis of Benzole Acid 36
Isotopic Standards 46
Mathematical Model 46
Chemical Processes 46
Isotopic Mass Balance 52
Correction Parameters 57
Correction of Analytical Results 65
Oxygen Isotopic Analysis 66
Yield of Carbon Dioxide 69
Correction Parameters 69
Evaluation of the Correction Parameters .... 78
Standards 85
-------
Experimental Section 89
Isotopic Standards 89
Oxygen Isotopic Analysis 93
References for Chapter 2 98
Chapter 3 MASS SPECTROMETRIC DETERMINATION OF CARBON-13,
OXYGEN-17, AND OXYGEN-18 IN CARBON DIOXIDE ..... 101
Development of Isotope Ratio Equations 103
Isotopic Distribution 103
Relationships Between 17R and iaR 106
Theoretical Considerations . 106
*'R and *'R in Natural Samples 110
Calculation of Isotope Ratios Ill
Determination of Isotope Abundances 112
Mixed Samples Enriched in ^0 113
17R and »»R in Natural Samples 117
Conclusions 123
Experimental Section 124
Preparation of Carbon Dioxide for Isotopic
Analysis 124
Mass Spectrometric Measurement 124
References for Chapter 3 127
CHAPTER 4 DEPOSITION, DISPOSITION, AND CLEARANCE OF
OZONE-DERIVED OXYGEN IN LUNG 12.9
Detection and Quantitation of the Isotopic Label . 131
Deposition in the Respiratory System 135
Comparative Aspects of Ozone Uptake 135
Distribution Within the Respiratory System . . 138
Deposition and Clearance in the Lung 143
Deposition in the Lung 143
Clearance from the Lung 152
Experimental Section 156
Apparatus 156
-------
Preparation and Purfication of ''O, 162
Aminal Exposures 163
Tissue Sampling and Preparation 163
Elemental and Isotopic Analysis 164
References for Chapter 4 165
CHAPTER 5 DETECTION AND QUANTITATION OF AUTOXIDATION
IN VIVO; A STUDY OF HEPATIC AUTOXIDATION INDUCED
BY CARBON TETRACHLORIDE 167
Lipid Peroxidation 168
Peroxidation of Linolenic Acid ........ 168
Measurement of Lipid Peroxidation In Vivo . . . 171
Carbon Tetrachi.oride Hepatotoxicity 172
Microsomal Cytochrome P-450 173
Carbon Tetrachloride-Induced Autoxidation . . . 174
Conclusions 183
Experimental Section 184
Protocol 184
Tissue Sampling and Preparation 184
Elemental and Isotopic Analysis 185
References for Chapter 5 186
SUMMARY AND CONCLUSIONS 189
vii
-------
ILLUSTRATIONS. AND FIGURES
Figure
1-1 Schematic representation of the mammalian respiratory
system ....... .................. 8
1-2 Reaction of ozone with unsaturated fatty acids ..... 16
1-3 Reaction of nitrogen dioxide with unsaturrated fatty
acids ....................... ... 19
1-4 Autoxidation in the cell ................ 21
2-1 Yield and isotopic composition of carbon dioxide
produced by pyrolysis of benzoic acid .......... 39
2-2- Equilibrium constant and equilibrium isotopic
fractionation factor for the reaction C02 + C *r± 2CO . . 45
2-3 Production of C02 from organic material ....... . . 53
i :
2-4 Results Predicted for replicate analyses of organic
samples ............. .-...' ........ 59
2-5 Flow-chart for determination of correction parameters . . 62
2-6 Illustartion of the method for refinement of values for
and n ....................... 68
2-7 Effect of IjOj temperature on yield of carbon dioxide . . 71
2-8 Determination of HQ and JiFQ ............... 75
2-9 Determination of a and b ............. ... 77
2-10 Sample matrix effects .................. 84
2-11 Calibration of working standards ......... ... 91
2-12 Diagram of modifications to elemental analyzer ..... 95
3-1 Schematic representation of isotope ratio mass
spectrometer ...................... 108
3-2 Relative error in »»R as a function of oxygen-18
abundance ........................ 119
3-3 Relative error in "F as a function of oxygen-18
abundance ........................ 121
4-1 Distribution of ozone-derived oxygen in the respitatory
system ......................... 142
viii
-------
4-2 Uptake of ozone-derived oxygen in the lungs 146
4-3 Molecular distribution of ozone-derived oxygen in the
lungs 148
4-4 Clearance of ozone-derived oxygen from the lungs .... 155
4-5 Apparatus for preparation and purification of 1§03 . . . 158
4-6 Exposure system 161
5-1 Products formed during peroxidation of linolenic
acid 170
5-2 Incorporation of Oa-derived oxygen into liver 181
Table
1-1 Isotopes of oxygen 26
2-1 Effect of pyrolysis time and temperature on yield and
isotopic composition of carbon dioxide 37
2-2 Recovery and isotopic composition of carbon dioxide
heated to 550°C in quatrz tubes 42
2-3 Analysis of benzoic acid 47
2-4 Analysis of heteraromatic carboxylic acids 48
2-5 Iterative determination of correction parameters .... 73
2-6 Comparison of isotopic analysis of benzoic acids by
pyrolytic decarboxylation and by the Schutze-
Unterzaucher procedure 81
2-7 Comparison of isotopic analysis of heteroaromatic
carboxylic acids by pyrolytic decarboxylation and by
Schutze-Unterzaurher procedure 86
2-8 Analysis of water standards and sucrose samples 88
3-1 Isotopic molecular species of carbon dixide 104
3-2 Analysis of CO, with oxygen derived from exchange with
mixed water 115
3-3 Analysis of CO, with oxygen derived from exchange with
distilled water 122
4-1 Contributions to the variance in »»F 134
4-2 Deposition of ozone-derived oxygen in the respiratory
system 136
ix
-------
4-3 Oxygen pools in nasopharynx, trahea, lungs, and blood . . 137
4-4 Uptake of ozone-derived oxygen in the respiratory
system 140
4-5 Deposition of ozone-derived oxygen in the lungs 144
4-6 Mouse-lung oxygen poolr> 151
4-7 Clearance of ozone-derived oxygen in mouse lung . . . . . 153
5-1 Summary of experimental protocol, 175
5-2 Abundance of oxygen-18 in liver 177
5-3 Relat've enrichment of oxygen-18 in liver 178
5-4 Oxygen pools in liver 179
-------
ACKNOWLEDGEMENTS
My sincere appreciation goes to John Hayes, his commitment ani
guidance made this thesis possible.
Dr. Edward Baer provides much advice concerning the chemistry
and preparation of ozone.
The technical support in our lab is outstanding. Steve Studley
had a posative influence on aiy research in many, innumerable ways.
A special thanks for the conscientious analysis of tissue samples, and
tissue samples, and tissue samples ... by David Andrews, David Archer,
and Martha J. Lars en is also well deserved.
Many of the specialized pieces of equipment required in these
4 :
studies were designed and constructed by the technical staff in the
department. The glass blower, Don Fowler/ spent many hours building
rebuilding apparatus. Each of the members of the machine shop, John
Dorsett, Kenny Bastin, Dick Martin, Bill Miller, Larry Sexton, and Don
Swafford, contributed to this work. Several devices had their origin
in the electronics: Bob Ensman, Steve Williamson, and Andy Quails.
This work was supported by a cooperative agreement with the
Unites States Environmental Protection Agency. The physiological
studies were conducted at the laboratories of the Inhalation
Toxicology Division of the EPA. Their support and collaboration made
this work possible. A special thanks must go to Dr. Gary E. Hatch for
his efforts. He initiated this project, provided the necessary
administartive support from within the EPA, and had an equal and ever
present hand in these studies.
-------
CHAPTER 1
INHALATION TOXICOLOGY OF OXIDANTS
-------
2
Oxidant gases, specifically ozone (Oa) and nitrogen dioxide (NOj),
are the most important toxic components of photochemical smog. Toxicity
arises from oxidation of components of biological tissue. Upon
inhalation, the chemice.1 lesion occurs in the respiratory system, at the
interface between inspired air and blood in the lung. Damage to tissue
can occur by two processes: direct reaction of the inhaled oxidant with
molecular components of the cell and free radical-induced oxidation
(autoxidation) of cellular components. Although autoxidation is a
secondary process, initiated by free radicals produced in the initial
4!
chemical insult, it is thought to account for the majority of damage to
tissue. Moreover, a complex series 'of biochemical, physiological, and
anatomical events in the lung ensues, and it is difficult to distinguish
dysfunction resulting from oxidation from the response of the organism.
A major emphasis of current research is the establishment of
relationships between effects in' laboratory animals and in humans. A
realistic assessment of the risk to humans must involve, not only an
understanding of the chemical mechanisms of oxidant injury to tissue,
but an understanding of the chemical and physiological factors which
influence transport, absorption, and reactivity of oxidants in the
respiratory system. A unified approach to the study of the inhalation
toxicology of oxidants utilizing cxygen-18 as a physiological tracer of
inhaled oxidants will be presented. This approach relies on techniques,
developed in this work, for the quantitation of excess oxygen-18 in
complex organic material (physiological samples). Two questions of
immediate concern are addressed directly:
1. What is the retained dose of a specific oxidant, both in the
lung and at specific sites within the respiratory system?
-------
3
2. What are the target molecules for inhaled oxidants in the lung
and the chemical mechanisms of oxidant injury in tissue?
This approach can be used to determine the influence of various
biochemical and physiological factors on the retained dose and oxidant
injury in the lung, and to provide a sound basis for extrapolation of
dose-response relationships frcj laboratory animals to humans. Although
the primary focus of this research has been study of the inhalation
toxicology of photochemical oxidants, the experimental protocol
developed in the course of these investigations is generally applicable
to the study of any agent or pathological condition that causes aberrant
oxidation of the molecular components of tissue.
Inhalation Toxicology
The physiological function of the lung is to provide a medium for
gas exchange between inspired air and the blood, while maintaining the
integrity of the organism. Any process which impairs the ability of the
lung to mediate exchange of respiratory gases is detrimental. In the
case of inhaled oxidants, pulmonary distress results mainly from tissue
damage in the respiratory regions of the lung.
An important conclusion, derived from cytologic and biochemical
observations of lung tissue, is that the peak concentration of an
oxidant in the atmosphere, as opposed to exposure time or the time-
weighted-average concentration, correlates most closely with changes in
lung structure and biochemistry that are associated with toxicity (1).
Although it is difficult to establish a distinction between an acute and
a chronic dose of a particular oxidant or, more precisely, the
concentration which elicits an acute or a chronic response, the oxidant
levels typically found in highly polluted urban areas (ozone < 1 ppm and
-------
4
nitrogen dioxide < 0.5 ppm) are well within the chronic range, except
for a small group of high-risk individuals (2). Exposure to high
concentrations of oxidants for prolonged periods of time results in
extensive destruction of epithelial tissue in the lung and can be lethal
(3). Symptoms of acute toxicity include chest pains, stvere respiratory
distress, and accumulation of edemata in the lung (4). If the exposure
is not lethal, prolonged structural and biochemical changes are observed
(5). Since chronic toxicity is most prevalent, however less obvious, in
the general population, the effects of low levels of inhaled oxidants
are currently of greatest interest.
Chronic effects encompass a complex sequence of events. Exposure
to sub-lethal concentrations of ozone or nitrogen dioxide results in an
increase in breathing frequency with decreases in minute volume, total
ventilation, and oxygen consumption (6). This rapid, shallow breathing
is accompanied by an increase in pulmonary resistance and a decrease in
pulmonary compliance (6). Respiratory function is impaired and
stressed.
Susceptibility to infection is also increased. Increased risk of
respiratory infection is a result of decreased phagocytyzing capacity of
the alveolar macrophage (7). This effect has been observed in mice
exposed to concentrations of ozone as low as 0.08 ppm (7). Furthermore,
the immune response of T and B lymphocytes (8) and the phagocytic rate
of polymorphonuclear leukocytes (9) are suppressed in humans exposed to
low concentrations of ozone. These effects persist for at least two
weeks after the exposure (1-9).
Decreases in the activities of enzymes which are particularly
susceptible to attack by oxidants, such as membrane proteins and
-------
5
sulfhydryl-containing enzymes, are observed initially (10-14). As
regenerative processes begin, biosynthesis increases, as do the
activities of typical microsomal and mitochondrial marker enzymes
(10,11). A certain degree of adaptation to oxidant injury occurs upon
repeated exposure (12,115), but prolonged changes in lung structure
(16,17), cell population (18), enzyme activity (13,14), and pulmonary
function (19) occur. The degree to which the organism is affected is,
in part, a function of the dose of oxidant to tissue. Tissue damage is
greatest in the centriacinar* region of the lung (20-22) and, therefore,
exchange of respiratory gases is greatly impaired.
Attempts to assess the dose of inhaled oxidants to specific regions
within the respiratory system have relied on mathematical models which
account for factors controlling the transport and absorption of reactive
gases in the respiratory system (23). Though predictions of these
models (24) are generally consistent with observations that the major
sites of damage are the respiratory bronchioles and the alveolar ducts
(20), direct experimental verification of the predicted dose is lacking.
Since, in principle, such models can be used to extrapolate doseresponse
relationships from laboratory animals to humans, it is imperative that
direct measurements of the retained dose in tissue be made and that the
mechanisms of respiratory transport and absorption of inhaled oxidants
be understood.
•The centriacinar region collectively refers to the structural units in
the lung where exchange of gases between inspired air and blood occurs.
See pages 6-9.
-------
6
Respiratory Transport and Absorption
The transport and absorption of gases such as nitrogen and oxygen
in the respiratory system is well understood (25). In contrast,
theories which describe the behavior of low concentrations of highly
reactive gases in the respiratory system are not well developed, even
though this issue is of central impor-ance to understanding the toxic
effects of inhaled pollutants. The dose of a reactive gas to various
regions of the respiratory system and the resulting tissue damage depend
on airway structure, the cellular morphology of the lung, and dynamics
of gas flow in the respiratory system.
Airway Structure. Conceptually, the respiratory system can be
divided into three functional zones: a conducting zone, a transitory
zone, and a respiratory zone (26,27). The conducting zone comprises the
nasal pharynx, the trachea, the bronchi, and the bronchioles, and
functions to distribute inspired gases into the lower lung. The
transitory zone forms the link between the conducting zone and the
respiratory zone, and consists primarily of terminal and respiratory
bronchioles. The respiratory zone, consisting of alveolar ducts and the
alveoli, is the site of exchange of gases between inspired air and the
blood. A schematic representation of the lung is shown in Figure 1-1.
The airways in the respiratory system form a series of bifurcated
tubes terminating in small sacs in the periphery of the lung. Each
airway is approximately cylindrical with flattened ends in the dimension
perpendicular to the plane of the bifurcation (28). Systematic
measurements of the size and branching angles of airways have shown that
the lung exhibits a regular, asymmetric branching pattern (25-30). If
the trachea is designated as the "rero-th generation" in the sequence of
-------
Figure 1-1. Schematic representation of the mammalian lungs •
The conducting zone consists of the the trachea, bronchi (br), and
bronchioles (bl). The transitory zone consists of terminal bronchioles
(tbr) and respiratory bronchioles (rbl). The respiratory zone consists
of the alveolar ducts (ad) and alveoli (ad). Each bifurcation of the
airway is assigned a generation number N, where the trachea is
designated as the "zero-th generation" (26,27). Human lungs typically
contain 23 generations between the trachea and the alveoli.
-------
-------
9
bifurcations, the mean length and radius of each airway decreases
exponentially with generation number, whereas total lung volume and
total surface area increase exponentially with generation number (24).
Although variations in the branching pattern, the size, and the geo-
metric shape of the conducting airways, thought to be a consequence of
evolutionary adaptation, are observed in different species (29), lung
structure is similar in mammals. Body mass is the single most important
factor which determines morphometric and physiological respiratory
parameters in mammals. Linear relationships between body mass and total
lung volume, total lung weight, compliance, resistance, diffusing
capacity, oxygen consumption, airway diameter, and alveolar surface area
have been established (25). The systematic morphometry of the mammalian
lung has been utilized in the development of a single convection-
diffusion model to describe the mass transport of inspired gases in the
lungs of mammalians (23).
Cell Population. Cell types with highly specialized functions are
found in epithelial tissue of the respiratory system. The epithelium of
the conducting airways contains ciliated and nonciliated cells, and
receptor cells (25). The nonciliated cells synthesize and secrete onto
the surface of the epithelium an aqueous fluid which becomes the mucous
layer. Mucus is cleared through the epiglotus into the gastrointestinal
tract by ciliary action. This is a major mechanism by which particles
and aerosols are removed from the lung (30). The function of the
receptor cells is not entirely clear. It is suspected that they serve
as neuronal receptors for irritants.
Transition to a different population occurs in the alveolar ducts.
The alveolar epithelium is composed of three cell types: Type I, Type
-------
10
II, and Type III pneumocytes (25). The majority of the alveolar
epithelium consists of type I cells. The function of these cells is to
form the boundary between the alveolar lumen and the capillary endo-
thelium while contributing minimally to the barrier for diffusion of
gases between blood and inspired air. Type I cells extend laterally, up
to 50 urn, from a single nucleus to form a thin cytoplasmic layer (0.1
vat) populated with relatively few organellas (25). Type II cells are
cuboidal with no lateral cytoplasmic extensions. They have a high
concentration of mitochondria, a well-developed endoplasmic reticulum, a
Golgi complex, and numerous multivesicular bodies (25). These cells
account for most of the metabolic activity of the lung, their main
activity being the synthesis and secretion of a phospholipid surfactant
onto the surface of the alveolar epithelium. Type III cells are very
rare and their importance to lung function is unknown. It has bean
proposed that their characteristic brush microvilli extending into the
alveolar lumen serve as neuronal receptors, in a fashion similar to the
receptor cells found in the bronchi.
A consequence of the low metabolic activity and large surface area
of the Type I alveolar epithelial cells is a high susceptibility to
damage by inhaled oxidants (11). Cell damage progresses from severe
cytoplasmic swelling to a complete disintegration of Type I cells, thus
exposing the basement membrane of the intersitiuro (20). Early stages of
tissue repair proceed by a proliferation of Type II cells, which
apparently give rise to the normal population of Type I cells (17).
Prolonged exposure to oxidants can result in a permanent increase in the
number of type II cells (16,17), together with fibrous intrusions into
the alveolar lumen, thickening of the respiratory bronchiolar and
-------
11
proximal alveolar walls, and partial or complete occlusion of the
alveolar ducts (17,21).
Flow Dynamics. In cases of chronic exposure, localization of
injury primarily to the centriacinar region is thought to be, in part, a
consequence of the flow characteristics of gases in the lung. The
driving force for the transport of gases in the respiratory system is a
pressure gradient created by expansion (inhalation) and contraction
(exhalation) of the pleural cavity. Convective flow predominates in the
conducting airways. A transition from convection to diffusion occurs as
the flow front penetrates the lower portions of the lung. Although the
average diameter of lung airways decrease exponentially as a function of
generation number, the total lung volume increases so that the velocity
of the flow front decreases as inspired, air expands into the lung. Air
in the alveolar ducts and in the individual alveoli is stagnant. Gases
diffuse along a concentration gradient between the airway lumen and the
capillary lumen (25). Therefore, mass transport of gases in the lung
has both convective and diffusive components.
Transport processes almost certainly play an important role in the
expression of toxicity. The dose of trace reactive gases to the airway
epithelium is determined by the specific modes of transport which per-
sist in different regions of the respiratory system. If fully developed
laminar flow existed in the conducting airways, then radial diffusion
would be the only mechanism for transport of gases to the airway wall.
Turbulent flow is common in the trachea and in the first several
bronchial generations (31). The asymmetric branching structure creates
nonlaminor flow profiles which are initiated at the bifurcations and
propagated along the daughter branches (32,33). Increased transport of
-------
12
reactive gases to the airway wall would be expected where turbulent flow
occurs and at the branching points, where non-laminar velocity fields
predominate. As velocity decreases, the residence time of a pulse of
inspired air in a single generation increases. Radial diffusion becomes
more important. The transitory and respiratory regions in the lower
lung are also expected to receive a high dose.
Although some protection is provided by tha mucus and surfactant
layers, a large dose may penetrate the airway coating, with a maximum
local deposition to tissue at the bifurcations and in the respiratory
bronchi, alveolar ducts, and the proximal alveoli. This view is
consistent with cytologic observations that cell damage accompanying
inhalation of oxidants is greatest at the branching points and in the
centriacinar region. Not only is quantitation of the dose to the total
lung essential, it is important to assess non-uniform distribution of
inhaled oxidants in the respiratory system, as the local dose to
specific epithelial sites may be disproportionately high.
Chemical Mechanisms
Studies of the reactions of ozone and of nitrogen dioxide with
organic molecules in vitro have provided a substantial framework from
which the present theories of damage to tissue have been developed.
Although it has been shown that ozone and nitrogen dioxide can react
with a number of compounds found in the cell, toxicity is thought to be
the result of destruction of the bronchiolar and alveolar epithelium,
associated with autoxidation of unsaturated fatty acids in the cell
membrane (34,35). There are two essential features to this scheme.
Initially, ozone or nitrogen dioxide reacts directly with carbon-carbon
double bonds in the fatty acid found in membrane lipids and possibly in
-------
13
certain aromatic residues in membrane proteins. These reactions, in
turn, generate free radicals which initiate autoxidation in the cell
membrane. Autoxidation is, however, not limited to membrane lipids.
Uncontrolled free-radical reactions in the cell may involve other
oxidizable compounds, causing changes in membrane fluidity and membrane-
bound protein structure, thus disrupting the normal function of the
membrane. In fact, autoxidation may produce a chemical amplification of
initial oxidative damage caused by the inhaled toxicant, and, in that
way, account for most of the cell damage.
Molecular Sinks for Ozone and Nitrogen Dioxide. Studies of the
reactions of ozone and of nitrogen dioxide with organic solutes and
macromolecules jn vitro have indicated molecular components of the cell
which are susceptible to direct oxidation by inhaled pollutants. In
aqueous solution, ozone was found to add across the carbon-carbon double
bound in tryptophan and histidine, to oxidize the sulfhydryl group in
cytosine to sulfone, and to cause hydroxylation of the aromatic ring in
tyrosine and phenylalanine (36). These amino acid residues were also
found to be susceptible to oxidation in proteins; however, the extent of
oxidation was a function of the accessibility of a particular residue to
the solution (36,37). Cross-linking of sulfhydryl groups in proteins
and glutathione was also observed, and only a portion of the original
biological activity of the oxidized compounds could be restored by
enzymatic reduction (36-38). Aqueous solutions of NADH and NADPH were
oxidized by ozone; however, both oxidized species were biologically
active as coenzymes in the reduced form (39).
Unsaturated fatty acids are particularly susceptible to oxidation
by ozone and nitrogen dioxide. Oxidation of methyl esters of unsat-
-------
14
urated fatty acids occurs in both thin films and aqueous dispersions
(40,41). The major products of ozonolysis of lipid in the presence of
molecular oxygen indicate that lipid peroxidation had occurred.
Further, addition of phenolic: antioxidants (a-tocopherol, butylated
hydroxyanisole, and butylated hydroxytoluene) decrease the rate of
oxidation (40). Products of lipid peroxidation have also been detected
in tissue from animals exposed to both ozone and nitrogen dioxide
(42-45). The mode of initiation and the extent of autoxidation in vivo
is, at present, unknown.
Direct Oxidation. The mechanism by which ozone reacts with
carbon-carbon double bonds was first proposed by Criegee (46). In this
scheme, Figure 1-2, ozone reacts directly with the carbon-carbon double
bond to form an initial adduct, a 1,2,3-trioxylane JT-complex (reaction
1). At room temperature it is unstable.>and rapidly decomposes to an
aldehyde and a zwitterion (reaction 2). In a non-participating solvent,
a non-polar solvent lacking active hydrogen, a resonance form of the
zwitterion (reaction 3) can react with the aldehyde to form a 1,2,4-
trioxylane ring system (reaction 4). However, in a participating
solvent (the cell membrane) the zwitterion can react with any molecule
containing active hydrogen, yielding a peroxide (reaction 5). Destruc-
tion of the lipid bilayer of the membrane by ozone alone could have
disastrous consequences for the cell; however, the peroxide products of
ozonolysis, including the 1,2,4-trioxylane compound, can initiate ion
autoxidation in the cell producing a chemical amplification of oxidative
damage.
Nitrogen dioxide reacts with unsaturated lipids by addition to tJ-.e
carbon-carbon double bond or by abstraction an a-hydrogen (47). The
-------
15
IS
Figure 1-2. Mechanism for the reaction of ozone with
unsaturated fatty acids.
(1) reaction of ozone with carbon-carbon double bond to form a
1,2,3-trioxylane
(2) decomposition of 1,2,3-trioxylane to an aldehyde and a Criegee
zwitterion
(3) resonance form of Criegee zwitterion
(4) reaction of Criegee zwitterion and aldehyde to form
1,2,4-trioxylane (molozonide)
(5) reaction of 1,2,4-trioxylane with molecular components of
tissue containing active hydrogen to form a peroxide
-------
16
=+ 0
(A)
-c-c-
o-o-
C+
I
A
•?-?-
9-0
-f
o
(I)
(2)
(3)
(4)
— C-O-C-R
(5)
0-OH
C-S-R
-------
17
important steps in both mechanisms are shown in Figure 1-3. The addi-
tion reaction (Scheme I) predominates at extremely high concentrations
of nitrogen dioxide (>1%), whereas hydrogen abstraction (Scheme II) is
the predominant mode of reaction at lower concentrations. Addition of
nitrogen dioxide to the carbon-carbon double bond produces a carbon-
centered radical, reaction 1. In the presence of molecular oxygen, a
carbon-centered radical reacts rapidly to form a peroxy radical
(reaction 2). A peroxy radical can initiate autoxidation in the cell by
hydrogen abstraction from an unsaturated lipid (reaction 3). Hydrogen
abstraction by nitrogen dioxide produces a delocalized carbon-centered
radical (reaction 4). In a similar fashion, this radical reacts with
molecular oxygen to form a peroxy radical (reaction 5) thus initiating
autoxidation.
Autoxidation. Lipid peroxidation in vivo is thought to be an
important component of oxidant injury accompanying exposure to nitrogen
dioxide (42,43) and to ozone (44,45). Presumably, carboneentered free
radicals and organic hydroperoxides produced by the reaction of lipids
with nitrogen dioxide and ozone, respectively, initiate autoxidation. A
general scheme for autoxidation is shown in Figure 1-4. Initiation
occurs by the production of carbon-centered free radicals in the cell,
where an organic hydroperoxide is the precursor for the initiation
process (reactions 1 and 2). At low conversions, lipid peroxidation in
the membrane produces radical species which sustain the autocatalytic
cycle (48-50). Propagation of the chain of free-radical induced oxida-
tions involves the reaction of the carbon-centered radical with
molecular oxygen to form a peroxy radical (reaction 3) followed by
hydrogen abstraction from an organic compound to regenerate a carbon-
-------
18
18
Figure 1-3. Mechanism for the reaction of nitrogen dioxide with
unsaturated fatty acids
Scheme I: Addition mechanism
(1) addition of nitrogen dioxide to carbon-carbon double
bound to form a carbon-centered radical
(2) reaction of molecular oxygen with carbon-centered
radical to form a peroxy radical
Scheme II: Hydrogen abstraction mechanism
(3) abstraction o-hydrogen from unsaturated hydrocarbon
to form a delocalized carbon-centered radical
(4) reaction of molecular oxygen with delocalized
carbon-centered radical to form a peroxy radical
-------
19
CsJ
ro
o
i
o
i
o
o
I
•o-
i
•o-
I
y
o-
I
+
X
o
o
I
o—
I
o
II
II
,0,
+
•o—
I
•o—
I
CVJ
o
O
I
O
H
o
CO
o>
o
CO
-------
20
20
Figure 1-4. Autoxidation in the cell
(1) initiation: ferric-ion catalyzed hemolysis of hydroperoxide
to yield an alkoxy radical and a hydroxide ion
(2) propagation: abstraction of nydrogen from an alkane to yield
a carbon-centered radical
(3) propagation: reaction of carbon-centered radical with
molecular oxygen to yield a peroxy-radical
(4) propagation: abstraction of active hydrogen to yield an
organic-hydroperoxide and a free radical
(5) production of a carbon-centered radical (QH « allylic
hydrogen) will result in a chemical amplification of
autoxidation
(6) reduction of ferrous ion by an organic hydroperoxide to yield
a peroxy radical and a hydrogen ion
RH « organic hydrogen; QH = active organic hydrogen; RCH *
organic alcohol; ROOH = organic peroxide; R- and Q- = carbon
centered radical; ROD * organic peroxy-radical; RO » alkoxy
radical; HO- = hydroxide ion; H* * hydrogen ion; FeJ* and
Fe1* « iron ion redox couple.
-------
21
ROM
ROOM
ROO + H+
-------
22
centered radical and an organic hydroperoxide (reaction 4). Propagation
can also occur by the production of alkoxy and peroxy radicals by the
metal ion-catalyzed oxidation of organic hydroperoxides (reactions 1 and
5).
During the initiation phase, free radicals can be produced by
thermal hernolysis of bonds, one-electron redox reactions, -and highenergy
radiation (49). The rate of hemolysis of organic hydroperoxides is
extremely slow at physiological temperatures. However, soluble
transition-metal ions (51) and transition-metal complexes such as iron
protoporphyrins (52) catalyze the production of alkoxy radicals from
hydroperoxides (reaction 1). AlkoxV radicals are, in general, extremely
reactive so that hydrogen abstraction (even'from saturated alkanes)
(reaction 2) proceeds at the diffusion-limited rate. Peroxy radicals
are, however, relatively stable, and can only abstract hydrogen from
easily oxidizable compounds (53,54). Autoxidation in the cell is thus
accelerated by reduced nucleotides, protein-thiol moieties, and poly-
unsaturated lipids. Termination can occur by the combination of two
radical species to form nonradical products or by the formation of a
stable radical. Nonradica.l products of termination reactions include
malonaldehyde (55), conjugated dienes (56), and volatile hydrocarbons
(57).
Autoxidatio.. in tissue is limited by the concentrations of all
oxidizable compounds and the rates of all initiation and termination
reactions, but is independent of the partial pressure of molecular
oxygen (54). As the concentrations of oxidizable compounds in the cell
are fixed, the various mechanisms that have evolved to protect the cell
against oxidation either limit initiation processes or increase termina-
-------
23
tion reactions (58). Initiation-inhibiting antioxidants act by
directing the decomposition of organic peroxides to nonradical products.
This type of protective mechanism includes the glutathione peroxidase
enzymes (59,60), catalase, and superoxide dismutase (61), and require
NADPH, supplied by the pentose monophosphate shunt, for activity. It is
therefore possible that this type of response can lead to a depletion of
reducing equivalents and inhibit mitochondrial production of adenosine
triphosphate. Chain-breaking antioxidants act by forming a stable
radical. For example, the o,0,7,6-tocopherols (vitamin E) are found in
the cell membrane and act by forming stable phenoxy radicals, thus
preventing chain propagation in the lipid bilayer of the membrane (62).
The reduced phenol can then be regenerated by reaction with ascorbate
(vitamin C) (63). Recent evidence suggests that, at low concentrations
of oxygen, ^-carotene also acts as a chain-breaking antioxidant by the
formation of a biologically stable radical (64).
Mechanism of Toxicity. Autoxidation is an uncontrolled branching
chain reaction with potentially disastrous consequences to the cell.
Studies of the effects of ozone on cultured erythrocytes suggest that
toxicity is due to oxidation of the cellular membrane, and it is
unlikely that ozone penetrates the cell membrane (31,38). The sequence
of events is: i) free radical-induced autoxidation, initiated by
ozonolysis of unsaturated phospholipids in the cell membrane; ii) cross-
linking of membrane proteins, either by direct radical combination, or
by formation of Schiff-base adducts between carbonyl moieties (produced
by the peroxidation of unsaturated lipids) and primary amines; and iii)
the formation of disulfide linkages by the oxidation of protein-
sulfhydryl groups. Disruption of membrane processes, especially in the
-------
24
mitochondria, and utilization of reducing equivalents by the various
antioxidant mechanisms may seriously inhibit cellular production of ATP.
Therefore, cell death may ultimately result from depletion of the
available energy sources used to maintain normal anabolic and catabolic
processes.
Isotopic Studies
At present, the key issues in the study of the toxicity of inhaled
oxidants fall into two categories: dosimetry and mechanisms of oxidant
injury in vivo. The important aspects of dosimetry are quantitation of
the retained dose of an inhaled oxidant to epithelial tissue and
determination of depositional fine structure within the respiratory
system. An understanding of the physicochemical factors controlling
respiratory absorption can be developed by correlating differences in
total and local dose with differences in airway structure in various
mammalian species. In conjunction with dosimetry models, this informa-
tion can be used to obtain an accurate prediction of absorption in the
human lung.
Activity in the area of oxidant mechanisms has involved, primarily,
in vitro studies of the oxidation of organic compounds and the effects
of a specific oxidant insult on cell cultures. Although a substantial
framework for tissue damage by photochemical pollutants has been
proposed, the chemical mechanisms involved in oxidant injury in vivo are
still speculative. While autoxidation in the cell membrane, initiated
by free radicals produced by reaction of nitrogen dioxide or ozone with
unsaturated lipids, has been suggested as the primary mode of oxidation
in vitro and in cell cultures, there is no direct evidence that this
mechanism is important in the toxicity of inhaled oxidants in vivo. The
-------
25
measurement of both direct oxidation and autoxidation in vivo is
therefore important because cell death may be the result of a general
depletion of reducing equivalents and inhibition of mitochondrial ATP
production.
Inhalation of oxidants can result in tissue damage by one, or both,
of two distinct processes: direct reaction with molecular components of
tissue and free radical-induced oxidation of tissue involving molecular
oxygen (02). Direct reaction of nitrogen dioxide or ozone with tissue
components results in the fixation of toxicant-derived atoms in the
tissue, whereas free radical-initiated oxidation leads to the uptake of
Oj-derived oxygen in tissue. This is: an important distinction which
allows these mechanisms to be resolved and studied in vivo by the use of
»* ~~*~~~~"~~
oxygen isotopic tracers. The isotopic label provides a means to
determine the retained dose of a specific oxidant (03, N02, or 02) in
tissue, so that, by the appropriate use of labeled oxidant, questions of
dosimetry and mechanism can be addressed.
Physiological Applications of Oxygen-18. All radioactive isotopes
of oxygen are sufficiently short-lived (Table 1-1) that they are of
limited utility as physiological tracers. Therefore, a stable isotope
of minor abundance, either oxygen-17 and oxygen-18, must be employed.
The natural abundance of both isotopes in biological material represents
an irreducible background, and becomes especially important if only
minute quantities of labeled compound enter the body (as is the case
with inhaled pollutants). Although the use of oxygen-17 would offer
greater sensitivity because of its lower natural abundance, the methods
used to quantify oxygen isotopes in organic material are more suited to
the determination of oxygen-18. Moreover, compounds highly enriched in
-------
26
Table 1-1. Isotopes of Oxygen
natural mass
nuclide abundance
% AU
»0 — 13b
I 4Q 14b
»0 — 15b
**0 99.759 15.99491502°
IT0 0.037 16.99913220C
liO 0.204 17.99915996°
b
i »0 19
a OQ 20b
half-life
s
0.0087
71.0
124
(stable)
(stable)
(stable)
29
14
decay
process6
0+, 17.8 MeV
/J+, 5.14 MeV
0+, 2.76 MeV
0-, 4.82 MeV
0-, 3.81 MeV
approximate abundance in mammalian tissue
nominal mass
Wapstra, A. H.; Grove, N. B. 1971 atomic mass evaluation. I.
Atomic mass scale. Nuclear Data, Sect. A. 9:267-301, 1971.
Handbook of Chemistry and Physics , 53rd edition, Weast, R. C. (ed.),
Chemical Rubber Company Press, 1973, p. B-249.
-------
27
oxygen-17 are rare and expensive, whereas compounds containing greater
than 99% oxygen-18 are readily available and comparatively inexpensive.
Oxygen-18 is, therefore, the best choice for use as a physiological
tracer.
In contrast to a radioactive isotopic tracer where quantitation is
based on the detection of a specific decay product, difference in mass
is the only physical property which can be used to distinguish stable
isotopes of a particular element. Mass spectrometry provides the most
accurate and precise determination of oxygen isotopic abundances.
However, analysis of complex organic mixtures, such as physiological
samples, requires that all oxygen in a sample be converted to a simple,
volatile gas for mass spectrometric measurement. Techniques for
preparation of carbon dioxide from organic material and for the mass
spectrometric analysis of carbon dioxide will be discussed in chapters 2
and 3, respectively.
-------
28
• REFERENCES
1. Melton, C. S. Effects of long-term exposure to low levels of
ozone: A review. Aviat. Space Environ. Ked. 53:105-111, 1982.
2. Epidemiologic Appraisal of Photochemical Oxidants, Air Quality for
Ozone and Other Photochemical Oxidants, Vol. II, Chapter 10, U. S.
Environmental Protection Agency, Office of Research and Develop-
ment, Washington, D. C., EPA-600/S-78-004, 1978.
3. Stokinger, H. E. Ozone toxicology. Arch. Environ. Health 10:
719-731, 1965.
4. Griswold, S. S.; Chambers L. A.; Motley, H. L. Report of case
of exposure to high ozone concentrations for t*o hours. Arch.
Industr. Health Ib:108-118, 1957.
5. . DeLucia, A. J.; Mustafa, M. G.; Cross, C. E.; Plopper, C. G.;
Dungworth, D. L.; Tyler, W. S. Biochemical and morphological
alterations in the lung following ozone exposure. Am. Inst. Chem.
Eng. Symp. Series, 71:93-100, 1975.
4 '
6. Dungworth, D. L.; Cross, C. E.; Gillespie, J. R.; Plopper, C. G.;
The effects of ozone on animals. Ozone Chemistry and Technology ,
Chapter 2, Murphy, J. S. and Orr, J'. R. (ed.s), Franklin Institute
Press, Philadelphia, 1975.
7. Coffin, D. L.; Blonaner, E. J.; Gardner, D. E.; Kolzman, R. Effects
of air pollution on alteration of susceptibility to pulmonary
infection. Proceedings of the 3rd Annual Conference on
Atmospheric Contamination in Confined Spaces, Aerospace Medical
Research Laboratories, Wright-Patterson Air Force Base, Ohio, p.
71, 1968.
8. Savino, A.; Peterson, M. L.; House, D.; Turner, A. G.; Jeffries, H.
E.; Baker, R. The effect of ozone on human cellular and humoral
imnunity: characterization of T and B lymphocytes by rosette
formation. Environ. Res. 15:65-69, 1978.
9. Peterson, M. L.; Harder, S.; Rummo, N.; and House, D. Effect of
ozone on leukocyte function in exposed human subjects. Environ.
Res. 15:485-493, 1978.
10. Ospital, J. J.; Hacker, A. D.; Mustafa, M. G. Biochemical changes
in rat lungs after exposure to nitrogen dioxide. J. Toxicol.
Environ. Health 8:47-58, 1981.
11. Witschi, H.; Cote, M. G. Primary pulmonary response to toxic
agents. Crit. Rev. Toxicol. 5:23-65, 1977.
12. Stokinger, H. E.; Coffin, D. L. Biological effects of urban air
pollution. Air Pollution , 2nd edition, Chapter 1, Stern,
A. C. (ed.), Academic Press, Hew York, 1968.
-------
29
13. Buckley, R. D.; Balchum, 0. J. Enzyme alterations following
nitrogen dioxide exposure. Arch. Environ. Health, 14:687-692,
1967.
14. Werthamer, S.; Penha, P. D.; Amaral, L. Pulmonary lesions induced
by chronic exposure to ozone. Arch. Environ. Health, 29:164-166,
1974.
15. Frager, N. B.; Phalen, R. F.; Kenoyer, J. L. Adaptation to ozone
in reference to mucociliary clearance. Arch. Environ. Health
34:51-57, 1979.
16. Stephens, R. J.;, Freeman, G.; Wright, P. G. Cytologic changes
in dogs' lungs induced by chronic exposure to ozone. Am. J. Path.
73:711-726, 1973.
17. Freeman, G.; Stephens, R. J.; Coffin, D. L.; Stara, F. J. Changes
. in dogs' lungs after long-term exposure to ozone. Arch. Environ.
Health. 26:209-216, 1973.
18. Evans, M. J.;, Cabral, L. C.;, Stephens, R. J.; Freeman, G. Trans-
formation of alveolar type 2 cells to type 1 cells following
exposure to N0a. Exp. Mol. Pathol. 22:142-150, 1975.
19. Bates, D. V. Physiological effects on man of air pollutants. Fed.
Proc., 33:2133-2135, 1974.
20. Stephens, R. J.; Sloan, M. F.; Evans, M. J.; Freeman, G. Early
response of the lung to low levels of ozone. Am. J. Pathol. 74:31-
44, 1974.
21. Hugod, C. Ultrastructural changes of the rabbit lung after a 5 ppm
nitric oxide exposure. Arch. Environ. Health 34:12-17, 1979.
22. Plopper, C. G.j Chow, C. K.; Dungworth, D. L.; Tyler, W. S.
Pulmonary alterations in rats exposed to 0.2 and 0.1 ppm ozone: A
correlated morphological and biochemical study. Arch. Environ.
Health 34:390-395, 1979.
23. Miller, F. J. A mathematical model cf transport and removal of
ozone in mammalian lungs . Ph.D. thesis, North Carolina State
University, Raleigh, North Carolina, 1977.
24. Miller, F. J.; Menzel, D. B.; Coffin, D. L. Similarity between
man and laboratory animals in regional pulmonary deposition of
ozone. Environ. Res. 17:84-101, 1978.
25. Weibel, E. R. Morphological basis of alveolar-capillary gas
exchange. Physiol. Rev. 53:419-495, 1973.
26. Weibel, E. R.; Gomez, D. M. Architecture of the human lung.
Science 137:577-585, 1967.
-------
30
27. Weibel, E. R. Morphometry of the human lung . Springer, Heidel-
berg, 1963.
28. Personal communication from Dr. R. Schreck, Biomedical Science
Department, General Motors Research Center, Warren, Michigan.
29. Horsfield, K.; Gumming, G. Morphology of the bronchial tree in
man. J. Appl. I'hysiol. 24:373-383, 1968.
30. Horsfield, K.; C'.uraming, G. Functional consequences of airway
morphology. J. Appl. Physiol. 24:384-390, 1968.
31. Chan, T. L.; Schreck, R. M.; Lippman, M. Effect of the laryngeal
jet on particle deposition in the human trachea and upper bronchial-
airways. J. Aerosol Sci. 11:447-459, 1980.
32. Schroter, R. C.; Sudlow, M. R. Flow patterns in models of the
hujian bronchial airways. Respir. Physiol. 7:341-355, 1969.
4 '
33. Olson, D. E.; Sudlow, K.; Horsfield, K.; Filley, G. F. Convective
patterns of flow during inspiration'. Arch. Intern. Med. 131:51-57,
1973.
34. Goldstein, B. D.; Balchum, 0. J. Effects of ozone on lipid
peroxidation in the red blood cell. Proc. Soc. Exp. Med. 26:356-
358, 1967.
35. Menzel, D. B.; Roehm, J. N.; Lee, S. D. Vitamine E: the biological
and environmental antioxidant. Agr. Food Chea. 20:481-485, 1972.
36. Mudd, J. B.; Leavitt, R.; Ongun, A.; McManus, T. T. Reaction of
ozone with amino acids and proteins. Atm. Environ. 3:669-682,
1969.
37. Leh, F.; Warr, T. A.; Mudd, J. B. Reaction of ozone with protease
inhibitors from bovine pancrease, egg white, and human serum.
Environ. Res. 16:179-190, 1978.
38. Chan, P. C.; Kindya, R. J.; Kesner, L. Studies on the mechanism of
ozone inactivation of erythrocyte membrane (Na* + K*)-activated
ATPase. J. Biol. Chem. 252:8537-8541, 1977.
39. Menzel, D. B. Oxidation of biologically active reducing substances
by ozone. Arch. Environ. Health 23:149-153, 1971.
40. Roehm, J. N.; Hadley, J. G.; Menzel, D. B. Oxidation of
unsaturated fatty acids by ozone and nitrogen dioxide. Arch.
Environ. Health 23:142-148, 1971.
41. Kesner, L.; Kindya, R. J.; Chan P. C. Inhibition of erythrocyte
membrane (Na* * K*)-activated ATPase by ozone-treated
phospholipids. J. Biol. Chem. 254:2705-2709, 1979.
-------
31
42. Thomas, H. V.; Mueller, P. K.; Lyman, R. L. Lipoperoxidation of
lung lipids in rats exposed to nitrogen dioxide. Science 159:532-
534, 1968.
43. Csallany, A. S.; Ayaz, K. L. Long-term exposure of mice in the
presence and absence of vitamin E. I. Effects on body weights and
lipofuscin pigments. Arch. Environ. Health 33:285-291,1978.
44. Goldstein, B. D.; Lodi, C.; Collinson, C.; Balchum, 0. J. Ozone
and lipid peroxidation. Arch. Environ. Health 18:631-635, 1969.
45. Dumelin, E. E.; Dillard, C. J.; Tappel, A. L. Effect of vitamin E
and ozone on pentane and ethane expired by rats. Arch. Environ.
Health 33:129-135, 1978.
46. Criegee, R. The Course of ozonation of unsaturated compounds.
Rec. Prog. Chem. 18:111-120, 1957.
47. Pryor, W. A.; Lightsey, J. W.; Church, D. F. Reaction of nitrogen
dioxide with alkenes and polyunsaturated fatty acids: addition and
hydrogen abstraction mechanisms. J. Am. Chem. Soc. 104:6685-6692,
1982.
48. Demopoulos, H. B. The basis of free radical pathology. Fed. Proc.
32:1879-1861, 1973.
49. Pryor, W. A. Free radical reactions and their importance in
biochemical systems. Fed. Proc. 32:1862-1869, 1973.
50. Pryor, W. A. Free Radicals. McGraw-Hill, New York, 1966.
51. Kochi, J. K. Mechanism of organic oxidation and reduction by metal
coa.olexes. Science 155:415-424, 1967.
52. Barber, A. A.; Bernheim F. Lipid peroxidation: its measurement,
occurrence, and significance in animal tissues. Advan.
Gerontological Res. 2:355-403, 1967.
53. Mayo, F. R. Free-radical autoxidation of hydrocarbons. Ace. Chem.
Res. 1:193-201, 1968.
54. Walling, C. Limiting rates of hydrocarbon autoxidations. J. Am.
Chem. Soc. 91:7590-7594, 1969.
55. Dahle, L. K.; Hill, E. G.; Holman, R. T. The thiobarbituric acid
reaction and the autoxidations of polyunsaturated fatty acid methyl
esters. Arch. Biochem. Biophys. 98:265-261, 1962.
56. Farmer, E. H.; Koch, H. P.; Sutton, D. A. The course of
autoxidation in polyisoprenes and allied compounds. Part V.
Observations on fish-oil acids. J. Chem. Soc. 1943, 122-125.
-------
32
57. Dumelin, E. E.; Tappel, A. L. Hydrocarbon gasses produced during
in vivo peroxidation of polyunsaturatod fatty acids and
decomposition of preformed hydroperoxides. Lipids 8:194-198, 1973.
58. Burton, G. W.; Ingold, K. U. Autoxidation of biological molecules.
1. The antioxidant activity of vj.tamine E and related chain-
breaking phenolic antioxidants in vitro. J. Am. Chem. Soc.
103: 6472-6477, 1981.
59. Chow, C. &.; Tappel, A. L. An enzymatic protective mechanism
against lipid peroxidation damage to lungs of ozone-exposed dogs,
Lipids 8:518-524, 1972.
60. Chow, C. K.; Tappel, A. L. Activities of pentose shunt and
glycolytic enzymes in lungs of ozone-exposed rats. Arch. Environ.
Health 26:205- 208, 1973.
61. • Fridovich, I. Superoxide radicals and superoxide dismutase. Ace.
Chem. Res. 5:321-326, 1973.
62. Beri, J. G.; Farrell, P. M. Vitamine E. in Vitamine and hormones,
v. 34, pp. 31-74, Academic Press, New York, 1976.
63. Parker, J. E.; Slater, T. F.; Willson, R. L. Direct observation of
a free radical interaction between vitaraine E and Vitamine C.
Nature (London) 278:737-738, 1979.
64. Beri, J. G.; Farrell, P. M. 0-Carotene: an unusual type of lipid
antioxidant. Science 224:569-573, 1984.
-------
33
CHAPTER 2
DETERMINATION OF OXYGEN-18 IN ORGANIC SUBSTANCES
-------
34
Determination of the abundance of l§0 in organic materials requires
conversion of oxygen to a volatile gas such as C02 or Oj for mass
spectrornetric analysis (1). Up to now, several techniques for the
preparation of carbon dioxide from oxygen in organic material have been
described (2-7). Two problems commonly encountered have been i) in-
accuracies linked to fractionation of the oxygen isotopes between two or
more intermediates in the conversion pathway and ii) interferences
arising from oxygen contamination. While these problems have been
successfully overcome in procedures where the sample is pyrolyzed in a
sealed nickel tube (7), such methods require a great deal of skilled
manipulation and, at their present state of development, are limited to
the analysis of compounds containing only carbon, hydrogen, and oxygen.
A general technique for oxygen isotopic analysis of physiological
samples, which utilizes the Schutze-Unterzaucher procedure (8,9) for the
preparation of carbon dioxide, is described in this chapter. In this
procedure, organic oxygen is converted to carbon dioxide in a sequence
of reactions: i) pyrolysis of the organic sample, ii) conversion (by
equilibration with solid carbon at high temperature) of the oxygen-
containing pyrolysis products to carbon monoxide, and iii) oxidation of
the carbon -monoxide by iodine pentoxide to produce carbon dioxide. An
unavoidable blank and memory (2) associated with the pyrolytic
production of CO, and the addition of oxygen from iodine pentoxide
during the production of CO,, complicate the isotopic analysis.
Ultimately, the abundance of oxygen-18 in the initial sample must be
reconstructed from that measured in carbon dioxide contaminated with
oxygen from other sources. Interferences in the isotopic analysis
arising from contamination of sample-oxygen have been addressed by the
-------
35
development of a mathematical model based on the physical processes
occurring in the conversion of organically bound oxygen to carbon
dioxide. The model facilitates accurate and precise determinations of
oxygen-18 in organic material.
Determination of the effects of non-sample-derived oxygen requires
the use of samples with known isot.^ic compositions. Because there is
no standard method for analysis of oxygen isotopes in organic materials,
no organic compounds which could be classified as "oxygen isotopic
standards" have been available. To solve this problem, a procedure for
the quantitative pyrolytic decarboxylation of aromatic carboxylic acids
has been developed. The isotopic composition of the pyroly^ically
produced carbon dioxide can be measured directly, and the isotopic
composition of the organic compound thus determined independently and
accurately. These new standards have allowed accurate quantitation of
contaminants and sensitive tests of the performance of the method.
Oxygen Isotopic Standards
Benzole acid has been used to define optimum conditions for
pyrolysis and to test the fidelity of the isotopic analysis. The major
products of the thermal decomposition of benzoic acid are benzene and
carbon dioxide. At approximately 500*C, decarboxylation of benzoic acid
has been shown (10) to occur predominantly by a unimolecular process
C,H,COOH . C.H. + CO, (2-1)
with small quantities of CO, K,, and biphenyl being produced by minor
side reactions involving radical intermediates. The occurrence of
heterogeneous catalysis cannot be excluded (10), especially when group-
VIII transition metals are present (11). Further, carbon dioxide
-------
36
produced by reaction 2-1 can also yield carbon monoxide if excess carbon
is present (12)
COj + C -, '.•»• 2CP (2-2)
Under the conditions that prevail during pyrolysis, reaction 2-1 is
irreversible, whereas reaction 2-2 is reversible and may approach
equilibrium.
The yield of carbon dioxide is affected by both processes. Kinetic
isotope effects (KIE) associated with reaction 2-1, with both the
forward and reverse processes in reaction 2-2, and an equilibrium
4!
isotope effect (EIE) associated wi.th reaction 2-2 can shift the isotopic
r>
composition of the carbon dioxide from that of the carboxyl group if
products other than COj are formed''from it during pyrolysis.
Pyrolysis _of .Benzoic,Acid. With pyrolysis of benzoic acid, the
yield and isotopic composition of the C02 were affected by pretreatment
of the quartz tubes. If the quartz tubes were used without prior
treatment, yields of carbon dioxide were low and variable (90 ± 5%), and
the isotopic composition of the C02 was irreproducible (± 3%e). Use of
the cleaning procedure described in the experimental section improved
both the yield of carbon dioxide and the consistency of the isotopic •
analysis. Only quartz tubes treated by this procedure were used to
obtain the results summarized here.
Samples of benzoic acid were pyrolyzed at 500, 550, and 600»C for
5, 10, 15, and 20 minutes. The yields and isotopic compositions of
carbon dioxide are given in Table 2-1 a.id their variation with reaction
conditions is graphically summarized in Figure 2-1. A maximum yield of
approximately 97% was attained in 5 rnin at 600eC and at longer times
-------
37
Table 2-1: Effect of Pyrolysis Time and Temperature on Yield and
Isotopic Composition of Carbon Dioxide
time yielda 613C_.. vs PDBa $l'0rri vs SMOWa
min % C°2 C°2
500'C
5 76.1 ± 4.0 -32.71 ± 0.92 +25.72 ± 0.09
10 94.0 ± 0.4 -26.33 +. 0.14 +25.72 ± 0.03
15 95.2 ± 0.4 -25.93 ± 0.05 +25.68 + 0.05
20 94.9 ± 0.1 -25.74 ± 0.02 +25.62 ± 0.04
550'C
5 94.7 ± 0.1 -26.48 ± 0.13 +25.67 t 0.02
10 98.4 ± 1.0 -25.56 ± 0.01 +25.61 t 0.02
15 96.1 ± 0.4 -25.67 ± 0.04 +25.57 ± 0.03
20 94.0 + 0.5 -25.51 ± 0.04 +25.66 ± 0.05
600«C
5 98.8 t 0.2 -25.56 ± 0.02 +25.59 ± 0.02
10 97.7 ± 0.1 -25.62 ± 0.08 +25.61 ± 0.13
15 93.5 i 0.3 -25.64 ± 0.04 +25.30 ± 0.02
20 92.0 t 0.2 -25.79 ± 0.04 +25.17 + 0.05
aMeans and standard errors of means of 3 replicate analyses.
-------
33
38
Figure 2-1. Yield and isotopic composition of carbon dioxide
produced by pyrolysis of benzoic acid. Each point indicates the mean
and standard error of the mean of 3 replicate analyses: (@) 500°C, dD
550»C, and (V) 600eC.
-------
39
100
10 15
time, min
20
-------
40
for lower temperatures. Initial low yields of carbon dioxide were the
result of incomplete decarboxylation; residual benzoic acid was
observed. The systematic decrease in yield of C02 under more rigorous
conditions was apparently associated with the partitioning of carbon and
of oxygen between CO ard CO, (reaction 2-2). In all cases, benzene
(m/z 78} condensed on the inside of the quartz tube as it cooled.
Carbon monoxide (m/z 27.9951) was also present in the liquid-nitrogen-
noncondensible fraction of all samples.
A kinetic isotope effect is associated with decarboxylation. The
decarboxylation of aromatic carboxylic acids in aqueous solution has
been studied extensively (13-16). In general, the mechanism involves
displacement of the carboxyl group by a proton. At low acidity, the
rate-limiting step is protonation of the ring carbon to form a
delocalized carbanion, and carbon isotopic fractionations are small. At
high acidities, protonation is fast and the rate-limiting step is
*• *•
breaking of the C-C bond. Kinetic isotope effects range between 2.5 and
3.5% (13-16). For pyrolysis of benzoic acid at 500 and 550°C, initial
low yields at short pyrolysis times (5 and 10 min, and 5 min,
respectively) were accompanied by depletion of carbon-13 in the CO,
(Fig. 2-1, Table 2-1). A carbon-12/carbon-13 kinetic isotope effect of
approximately 1.5% was calculated from these results. The smaller value
for the KIE is consistent with (i) the high temperature of the reaction
and/or (ii) a concerted mechanism where the rate-limiting step involves
intramolecular displacement of the carboxyl moiety by hydrogen. No
secondary KIE for oxygen was observed.
The magnitude of isotopic fractionation between the carboxyl group
and carbon dioxide depends on the magnitude of the isotope effect and
-------
41
the yield of CO,. If decarboxylation were quantitative, no isotopic
fractionation would occur, and the isotopic composition of the COj would
be identical to that of the carboxyl group. Isotopic fractionations
diminished as decarboxylation approached completion. However,
quantitative yields of CO, were never observed. A maximum yield was
attained in 5 min at 600°C (98.8%), in 10 min at 550°C (98.4%), and in
15 min at 500»C (95.2%). At maximum yield, the carbon isotopic
composition of the CO, was constant (6ls^co2 vs roB * -25.65%»). The
isotopic composition of the oxygen in the CO, was independent of yield
at 500 and 550"C (<Oco vs SHOW « +25.64%0). At 600»C, the oxygen-18
content decreased with yield. These variations in yield and oxygen-
isotopic composition can be attributed to pyrolytic decomposition of CO,
and an accompanying equilibrium isotope effect.
To examine isotopic fractionaticns associated with decomposition of
COj, 30-umole aliquots of CO, were transferred to quartz tubes r^nd
heated to 550°C for intervals of 5 to 30 minutes. The recoveries and
isotopic compositions of the residual carbon dioxide are summarized in
Table 2-2. Although small amounts of CO, were lost, no statistically
significant isotopic shifts were observed (p < 0.05). However, the data
in Table 2-2 suggest that decomposition of CO, under conditions of
pyrolysis resulted in an enrichment of carbon-13 and a depletion of
oxygen-18 in the residual carbon dioxide. These trends differ from
those observed with pyrolysis of benzoic acid (Table 2-1, Fig. 2-1), and
were probably due to a KIE associated with decomposition of C02 to CO
(the forward process of reaction 2-2). Acceleration of reaction 2 in
both directions in the prasence of elemental carbon produced by
decomposition of benzene during pyrolysis of benzoic acid (10) allows
-------
42
Table 2-2: Recovery and Isotopic Composition of Carbon Dioxide
Heated to 550°C in Quartz Tubes
time n recovery3 6lscco2 vs fDn& 51*°C02 vs SM°tjl?a
% %« %o
100 -10.409 ± 0.09 -*-31.33 ± 0.04
10 4 99.8 ± 0.1 -9.96 ± 0.02 +31,28 ± 0.03
20 4 99.5 ± 0.2 -9.94 ± 0.03 +31,28 ± 0.02
30 4 99.4 4 0.1 -9.92 ± 0.06 +31.27 ± 0.04
± standard errors of means of n replicate analyses.
-------
43
this reaction to approach equilibrium. In this case, the isotopic
compositions of the products are controlled by the equilibrium isotope
effect associated with reaction 2-2.
Estimates of isotopic fractionations between carbon dioxide and
carbon monoxide were obtained from calculations of the thennodynamic
equilibrium constant and of the equilibrium oxygen isotopic fractiona-
tion factor for reaction 2-2 (Figure 2-2). The isotopic composition of
the product gas mixture at 550°C was calculated using values of 14A»a
= 1.000735, K = 0.12. The temperature and pressure of the gas inside
the quartz tube during pyrolysis were estimated assuming decarboxylation
was quantitative. This calculation indicated that, at equilibrium, the
composition of gases derived from the carboxyl group would be 94.1% C02
and 5.S% CO, resulting in a shift in the oxygen isotopic composition of
the COj from that of the carboxylic acid moiety of AS^CLg vs
carboxylic acid = -0.30%0. Although the decrease in yield and the shift
in isotopic composition observed during thermal decomposition of benzoic
acid at 600eC (Table 2-1, Fig. 2-1) were significantly greater (p >
0.05) than the predicted values, the model calculations serve only to
estimate isotopic fractionations arising from an EIE. The differences
between the experimental and theoretical results ar? not serious. Since
the rates of the forward and reverse steps in reaction 2-2 are
temperature dependent (17), only at 600eC does this reaction approach
equilibrium and do the isotopic fractionations become significant.
Isotope effects associated with decarboxylation are unimportant when
near-quantitative (maximum) yields of carbon dioxide are obtained. Only
with pyrolysis at 600°C does the oxygen isotopic composition of carbon
dioxide change significantly urom that at maximum yield.
-------
44
44
Figure 2-2. Calculated values for the thermodynaaiic equilibrium
constant (K) and the equilibrium oxygen isotopic fractionation factor
(a) for reaction 2 over the temperature range of 300 to 1500°K.
-------
45
log(Kp)
-------
46
Suitable conditions for isotopic analysis are those which result in
no isotopic fractionation from either incomplete decarboxylation or
production of carbon monoxide.. At 550eC there is a considerable range,
between 10 and 20 minutes, where there is no apparent isotopic
fractionation, which is ideal for isotopic analysis. Any combination of
time and temperature of pyrolysis which gives a maximum, riear-
quantitative yield of carbon dioxide also satisfies the condition of
minimal isotopic fractionation. The isotopic composition of the carbon
dioxide is, therefore, identical to that of the carboxyl group.
Isotopic Standards. Results of isotopic analyses of CO, derived
by replicate pyrolytic decarboxylations (550°C, 10 min) of five benzoic
acids are summarized in Table 2-3. Isotopic analysis of CO, obtained by
pyrolysis (550°C, 10 min) of 2,6-pyridinecarboxylic acid, pyrrole-
2-carboxylic acid, and 2-thiophenecarboxylic acid are shown in Table
2-4.
MATHEMATICAL MODEL
The following description of the chemical processes involved in the
production of carbon dioxide from organic material by the Schutze-
Unterzaucher procedure explains the physical basis for the mathematical
treatment.
Chemical Processes
Pyrolysis of organic material. Depending on the composition and
the structure of the initial sample, pyrolysis products generated in the
first step of the process may include CO, C03, H,0, NO, COS, H,, N,,
CH,, and traces of H,S and C,Ht (18,19). Some generalizations can be
made about the pyrolytic behavior of specific oxygen functionalities.
Carboxyl groups yield their oxygen predominantly in the form of carbon
-------
47
Table 2-3: Analyses of Benzole acid Standards
sample %-yielda 1§FCO x 10'
benzole acid-0 97.0 ± 0.2 2.03548 ± 0.00015
t!
benzole acid-1 96.9 ±,0.2 • 2.04756 ± 0.00009
benzole acid-2 96.5'i 0.2 2.11813 i 0.00015
benzole acid-3 96.3 ± 0.3 2.18844 ± 0.00015
benzole acid-4 96.2 ± 0.3 2.24562 ± 0.00040
± standard errors of means of 5 replicate analyses.
-------
48
Table 2-4: Analysis of Heteroaromatic Carboxylic Acids
?C02
sample %-yielda 1§FCO x 10*
2f6-pyridine-
dicarboxylic acid 99.1 ± 0.4 2.03740 ± 0.01017
pyrrole-2-
carboxylic acid 95.6 ± 0.2 1.97134 ± C 00006
2-thiophene-
carboxylic acid 99.0 ± 0.2 1.97804 ± 0.00005
Means ± standard errors of 5 replicate analyses.
-------
49
dioxide, whereas aldehydes, ketones, and alcohols yield carbon monoxide.
Carbohydrates are the only compounds to produce water in appreciable
amounts. Traces of carbonyl sulfide are present when the sample
contains organic sulfur. The interactions of the various oxygen-
containing pyrolysis products with carbon in the second step are of
central importance in formation of carbon monoxide and in the appearance
of the blank and the memory.
Production of CO from pyrolysis products. A helium carrier gas is
employed to sweep the pyrolysis products through a quartz tube packed
with carbon and maintained at 1060°C. The oxygen-containing compounds
react with the carbon surface to yield carbon monoxide as the
thermodynamically favored product (12). This reaction proceeds through
a surface-oxide intermediate. The reactions of interest are (20):
CO, + C « * CO + Cx(0) fast (2-3)
2ND + 2C * * N, •«• 2CX(0) fast (2-4)
H,0 + C • V H, •*• Cx(0) fast (2-5)
2COS ••• 2C « * CS, + 2CX(0) fast (2-6)
where Cx(0) represents oxygen chemisorbed CO on the surface of the
carbon and x is the number of atoms in the carbon matrix which
participate in the surface-oxygen complex. The surface oxide then
undergoes a unimolecular decomposition
Cx(0) ^ CO + C slow (2-7)
However, the carbon surface is microscopically heterogeneous, containing
two physically distinct reaction sites (21-27): a short-lived inter-
-------
50
mediate that rapidiy decomposes to yield carbon monoxide (reaction 2-3),
and a more stable surface oxide (reaction 2-7) in which oxygen is
strongly adsorbed to the surface (25). The activation energy for
desorption at both sites is inversely proportional to the local surface
coverage (23,26,27). Thus, oxygen accumulates on the carbon surface as
chemisorbed carbon monoxide.
Reactions 2-3 through 2-6 mediate the exchange of oxygen between
the carbon surface and the gas phase. Desorption of CO from the surface
(reaction 2-7) is rate limiting in the formation of carbon monoxide, and
is essentially irreversible (25), whereas reaction 2-3 is readily
reversible (22-24) during the time when the pyrolysis products are in
contact with carbon. Thus, any oxygen entering the system will be
distributed between the surface oxide and gaseous carbon monoxide. As
multiple samples are sent through the reactor, slow desorption of carbon
monoxide effectively generates a sorbed-oxygen pool charged with oxygen
from previous inputs. As subsequent samples are produced, exchange of
oxygen with this pool can buffer the isotopic composition of the carbon
monoxide produced.
The analytical blank. The blank arises from a reaction between
carbon and silica (the quartz tube used to contain the carbon bed). A
direct interaction
SiO, + C •- CO •*• SiO f C (2-8)
is apparently of minimal importance (28). Significant amounts of carbon
monoxide are produced by reaction 2-8 at temperatures in excess of
1200°C, but a large blank is observed at lower temperatures. A process
involving a silicon-hydroxyl moiety as the active oxygen species
-------
51
2SiO(OH) + 2C *• 2CO + 2SiO + H, •*• 2C (2-9)
has a lower overall energy of activation and could explain the blank
observed at lower temperatures. Surface hydroxyl groups are likely to
be produced by reaction of acidic gases and HJf produced in the
pyrolysis reactor, with the walls of the quartz tube (20):"
SiO, + HX *• SiO(OH) + X (2-10)
Due to its lower energy requirements, a mechanism involving reactions
2-9 and 2-10 is a more likely explanation of the blank (20).
Two factors influence the magnitude of the blank: contact between
carbon and the quartz tube in the pyrolysis reactor, and the temperature
of the the reactor in the region of the contact. Each has been utilized
by other investigators in schemes to reduce the blank. The quartz tube
has been lined with platinum foil (2) to separate the carbon from the
quartz, but carbon from the sample is inevitably deposited on the
internal surface of the tube in the region where it undergoes volatil-
ization and initial pyrolysis. Smaller blanks are observed at temper-
atures below 1200*C, but use of a metal catalyst in direct contact with
the carbon is required to ensure a quantitative yield of carbon monoxide
in a dynamic system (20). The metal on the carbon surface acts as a
catalyst for reaction 2-7, increasing the rate of evolution of carbon
monoxide at the lower temperature (26,29).
The presence of low levels of a volatile chlorinated hydrocarbon in
the carrier gas has been found to improve the yield and reduce the
tailing of carbon monoxide (30). The mechanism leading to this improve-
ment is not known. However, acidic gases are adsorbed by carbon (31),
and it seems likely that HC1 produced by pyrolysis of the halocarbons
-------
52
(19) competes successfully for sorption sites on the carbon surface. As
a result, interferences arising from the sorbed-oxygen pool are reduced
and stabilized.
Oxidation of CO to yield CO;. Carbon monoxide produced in the
pyrolysis reactor is oxidized to carbon dioxide as it is swept through a
tube packed with iodine pentoxide (9) and maintained at 120°C. Under
these conditions, hydrocarbons, methane in particular, are not oxidized
by IjOs, nor is oxygen exchanged between either carbon monoxide or
carbon dioxide and the reagent (2). Therefore, no interferences to the
isotopic analysis other than the addition of oxygen from iodine
pentoxide accompany this reaction.
I
Isotopic Mass Balance ;
An isotopic mass balance is required to understand the relationship
between the isotopic composition of carbon dioxide produced by the
Schutze-Unterzaucher procedure and those of the components of the carbon
dioxide-oxygen pool. The mass-balance equations must reflect the
specific reactions that result in the production of carbon dioxide from
an organic material. As noted in Figure 2-3, oxygen in the carbon
dioxide derives from several sources: the sample being analyzed,
previous samples (memory effect), silica in the quartz tube which
contains the carbon packing (blank), and iodine pentoxide.
Mass Balance for CO. The abundance of oxygen-18 in the total
oxygen pool is the weighted sum of the abundances in each component.
The mass-balance equations are
nil'FT = HS^FS + ngi'FQ * Injl'Fj (2-11)
and
-------
53
53
Figure 2-3. Schematic representation of the production of C02 from
i!
an organic sample. The sorbed-oxygen pool is represented by {c (0)}.
r X
f
Oxygen from the present sample, from previous samples (memory), and from
the quartz tube (blank) are incorporated in the carbon monoxide through
the surface-oxide. Carbon monoxide is oxidized to carbon dioxide by
1,0, with the addition of a single atom of oxygen to each molecule of
CO.
-------
54
o
•g
*x
O
O
CVJ
CD
(T
in
J*?
Q.
e
o
O)
13
O
o
CD
6
CO
'en
O
v.
>
CL
o ^O
X o c\j
<-> O X
C.
CJ
O
CO
o
o
-------
55
nT = ng + nQ + In. (2-12)
where l»F is the fractional abundance of oxygen-18 and n is the amount
of oxygen (MIDO! 0). The subscript T designates the total oxygen pool; S
designates oxygen derived from the sample itself; Q is used to designate
oxygen from the blank, thought to be derived largely from the quartz
reactor tube; and j is used to designate oxygen derived ultimately from
a previous sample carrying the index j, that is, a "memory effect" (32).
Only a finite number of samples is considered to be significant in the
sorbed oxygen pool.
An expression for the fractional abundance of oxygen-18 in the
carbon monoxide can be obtained if it is assumed that no isotopic frac-
tionation accompanies the partitioning of oxygen between the gas phase
and the sorbed phase in the reduction reactor. The carbon monoxide is
then a representative sample of the total oxygen pool, lgF = **F , and
\s\J 1
equations 2-11 and 2-12 yield the following relationship
+ Zn »•?.
»Tro- - - - 2 - - - 3 - J - ±- (2-13)
1 + (nQ + Znj)-l/ns
where lt?rn is the fractional abundance of oxygen-18 in the carbon
monoxide oxygen pool. Equation 2-13 can be rearranged to yield the
fractional abundance of oxygen-18 in the sample:
1§FS = I§FCO + + Inj
-------
56
Mass Balanca for CO, . The oxidation of carbon monoxide by iodine
pentoxide results in the combination of two oxygen pools of equal size
nC02l'FC02 " nC01>FCO * n!l'FI <
nC02 * nCO * nl (2-16)
nco - nj. (2-17)
where the subscripts (CO, and I) designate the total carbon-dioxide
oxygen pool and that portion of it derived from iodine pentoxide,
respectively. It follows that, since the size of the oxygen pool is
doubled, isotopic fluctuations in the carbon monoxide are reduced by a
factor of two in the carbon dioxideu'
»'FCO = 2-»Fc'02 - "Fj- (2-18)
Note that ltF]. is the only quantity required for definition of the
relationship between I*FCO and 1§FCO .
Combining equations 2-13 and 2-18 yields a relationship which
relates the isotopic composition of the carbon dioxide to those of its
component cxygen pools:
i »FCO - - : - 3 - = - = - — (2-19)
1 + (n + Zn
Equation 2-19 summarizes the dependence of the analytical result on
experimental parameters, only one of which is the isotopic composition
of the sample. The relationship between *'FS and 1§FCO is complex. In
particular, it will be different for each sample because the isotopic
composition of the sorbed oxygen pool depends on the history of the
-------
57
reactor. Nevertheless, if ail other terms in equation 2-19 are known,
1§FS can be calculated from observed values of IIFCO • In practice, it
is most convenient to use equations 14 and 16 to calculate 1§Fg .
Determination of Correction Parameters
In devising a procedure to determine J*PI, lt?Qt nq* .^^ an
expression for n^ it is helpful to consider the relationship between
1»FCO and l'Fs for each member in a series of analyses of a single
material where each sample differs only in size. Equation 2-19
indicates the effects of each parameter on the observed results. A plot
of 2'1*FCO as a function of l/ns for a hypothetical series of replicate
analyses of a single organic material with a true isotopic composition,
1>FS, is shown in Figure 2-4. Several different curves appear as a
result of changes in the isotopic composition of the sorbedoxygen pool.
If a sample is infinitely large d/ng approaches zero), the effects of
the blank and the memory will be insignificant. Accordingly, all Curves
converge to a common y-intercept at 14FS + 1$Fi' For samples of finite
size, curve A represents the locus of all possible results of the first
analysis in the series, curve B the second, curve C the third, etc..
Each curve is displaced from its predecessor because each sample changes
the isotopic composition of the sorbed oxygen pool, making it more like
that of the sample. A small input affects the sorbed oxygen pool less
strongly than a large one; therefore, as shown in Figure 2-4, the
magnitude of the shift between curves depends on the size of the
previous sample (recall that a small value of 1/n corresponds to a large
sample). Curve Z defines the limit reached when, after repeated
analyses of a single material, the isotopic composition of the sorbed-
oxygen pool differs from that of the input sample only due to
-------
58
58
Figure 2-4. Results predicted" for a plot of 2*18FCQ2 vs. 1/n for
replicate analyses (A, B, C, D, ... s Z) of"an organic compound with a
true isotopic composition of 1*FS. .Equation 2-17 defines the locus of
all possible results for a single analysis. In this example, both the
blank and the initial sorbed-oxygen pool have been assumed to be
depleted in cxygen-18 relative to-the sample, so all analyses fall below
up + itp tne y-intercept. Curve Z defines the limit where memory
D JL
effects have been eliminated by the continuous, repetitive processing of
samples of the same material.
-------
59
.
o
-------
60
contributions from the blank. The condition where the effects of memory
have been suppressed by saturation of the sorbed-oxygen pool with oxygen
from the sample will be referred to as "memory equilibration". In the
construction of Figure 2-4, it was assumed that both the isotopic
composition of the blank and the sorbed-oxygen pool were depleted in
oxygen-18 with respect to the sample.
Values of the correction parameters (i. e., all terms in equation
2-19 except 1§Fj» l'Fco ' anci °S^ can be determined from results of
analyses of materials for which I*FS is known. However, since the terms
used to correct for the blank and the memory are not separable an
iterative procedure is required. In practice, it is useful to
manipulate experimental conditions so that the effects of the blank or
the memory are maximized in a particular experiment, thus improving the
precision with which the correction parameters can be determined.
An outline of the procedure used to determine l*F-ri 1§Fof "O'
an expression for n is shown in Figure 2-5. Initial estimates of
»»FQ, and nQ are obtained from a single experiment in which memory
equilibrium has been established (step I). Subsequently, the effects of
the memory can be quantitated in a separate experiment (step II). These
estimates are then refined by a procedure of successive approximations
until the values for all parameters converge within predetermined limits
(step III). A detailed description of this procedure follows.
Blank and I20,-added Oxygen. Under conditions of memory
equilibration, the difference between »«FS and ''FCO? is controlled by
the isotopic composition and relative size of the blank and by the
isotopic composition of the oxygen derived from the 1,0,. The latter
quantity O'Fj) can be determined from the y-intercept of a plot of
-------
61
61
Figure 2-5. Flow-chart of the procedure for determining, the
correction parameters.
-------
62
2'18FC0., vs l/ns for two
standards where memory
equilibrium has been established
determine:
I'
18
FQ, and nQ
1 fl
F^Q vs sample number
for a series including
isotopically enriched samples
1
n^ and
r
determine :
an expression
for n •
next approximation
using current values
of correction parameters
yes
final values for:
F!» FQ> ng, and n:
-------
63
2-l'FCo2 vs l/ns if i'Fg is knovm:
{y-intercept} «= lt?z + ''Fj (2-20)
The size and the isotopic composition of the blank are determined
from memory-equilibrated curves for two samples of different but knovm
isotopic compositions. Comparison of difference between the observed
and true isotopic compositions (1 , a parameter whose
values are, furthermore, unknown at this point. An iterative procedure
can be adopted to derive a consistent set of correction parameters once
the characteristics of the memory have been derived.
Memory. The amount of a previous sample which is present in the
sorbed oxygen pool at any time is a function of the size of that sample
and of the rate of its displacement by subsequent throughputs. As an
approximation
-------
nj
64
(2-23)
where n* 0 is ^e initial size of the jth previous sample and f(j) is an
exponential function of the index, j. The quantitative characteristics
of f(j) cannot easily be derived from a specific model of the sorbed
oxygen pool since the surface oxide is heterogeneous. In practice, the
exponential function which best describes the wash-out of a particular
sample (see below) is determined empirically (32).
The turnover of oxygen in the sorbed oxygen pool can be determined
by examining the response of the system to isotopic perturbations (32).
In this approach, a series of samples of known isotopic composition
(1»F1) is interrupted by an input of material highly enriched in
,*
oxygen-18 (1'FJ). If the sizes of the samples are constant, then the
amount of oxygen from the enriched input appearing in the k-th sample
after that input is given by
(ns
lip _
rco
- "F
(2-24)
where l*FCo is calculated using equation 2-13. Since f(j) is initially
unknown, values for n. cannot be determined and, in the first
approximation, are set equal to zero in equations 2-13 and 2-24. The
values of n^ are then fit to an exponential function of the index (j).
Further approximations are obtained by using equation 2-23 to calculate
values for n* in equations 2-13 and 2-24 aM recalculating values for n^-
Tt is also possible that more than one enriched sample appears in
the series. In this case, equation 2-24 must be modified to account for
the introduction of multiple isotopic perturbations at different times.
The most recent enriched sample causes the greatest shift in ''FCQ , so
-------
65
its effects alone are considered explicitly. Components from all
previous enriched samples are grouped in a single summed term which
treats these effects as a small contribution to the observed shift
nk *
'•FCO -
?. - 1§F,
(2-25)
where H. « 1 when j corresponds to an enriched sample prior to the most
recent and Hj * 0 for all other samples, independent of their isotopic
composition. As in the previous example, initial estimates of n are
obtained by setting n* • 0 and only isotopic perturbations caused by the
most recent enriched sample are considered in the calculation of n^.
This procedure is only an expedient adopted during study of the sorbed-
oxygen pool. Once the functional form of f(j) has been specified, the
effects of all previous enriched samples enter into the calculation.
To complete the procedure, the first approximations of ''FT* ru/
**FQ, and n. are used to calculate values for all unspecified terms.
The procedure is repeated until convergent values for all correction
parameters are obtained.
Adjustment of Analytical Results
Application of the correction procedure can be visualized most
simply in terms of equations 2-14, 2-18, and 2-23. Since 1'FI is known,
**FCO can be determined from equation 2-18. The characteristics of the
blank (HQ and I§FQ) and of the memory (nj) have been determined directly
by the procedure outlined above. Solution of equation 2-14 for i*Fs
requires, in addition, values for ns and lIFj. The size of the sample
(n,) is determined from the elemental analysis, and values for **F* are
•J J
given by the results of previous analyses in the series. Equation 2-23
-------
66
is used to determine values for n., and liFj is, in each case, set equal
to 1§Fg for the j-th previous sample.
In the correction of a series of analytical results, equations 2-14
and 2-16 are used to calculate values of *•*?$ based on the final set of
correction parameters and the observed values of 1*FC02' Tne initial
isotopic composition of the rorbed oxygen pool is assumed to be that of
atmospheric 02 (1'FAIR » 2.04435 x. 10-') because the introduction of air
into the system is unavoidable when the sampling head is loaded. Values
for i§Fs are calculated sequentially. In this way, values for n$ and
J>FS for all previous samples are available and provide the infurination
necessary to correct for the effects of the memory. A limit of 30 is
placed on the index j. Software allowing rapid completion of these
calculations has been developed.
Refinement of correction parameters. The inclusion of three
samples of a working standard in each sequence of analyses allows
refinement of parameters initially determined using the more elaborate
procedure outlined above. It is of particular interest to check the
values assigned to 1§F]. and HQ. The procedure is graphically summarized
in Figure 2-6. The value assigned to nq is adjusted until the value
calculated for lfFs is independent of n
-------
67
67
Figure 2-6. A schematic illustration of the method for refinement
t!
of values for I§FI and nq. The latter parameter is first adjusted to
minimize the sum of the residuals noted here as &F. The value of ltFIis
then adjusted until the calculated value'of J*FS matches the known
value.
-------
68
I
H
U_
CO
+
CO
CO
-»-•
/
/
*-
-------
69
tests, the benzole acid samples listed in Table 2-3, the three hetero-
aromatic carboxylic acids (pyrrole-2-carboxylic acid, 2-thiophene
carboxylic acid, and 2,6-pyridine dicarboxylic acid)/ and three primary
standard water samples, SHOW, NBS--1, and KBS-1A, were used to determine
the accuracy and precision of this technique.
Yield of Carbon Dioxide
Kanonetric techniques were employed to determine the yield of
carbon dioxide from the Schutze-Unterzaucher procedure for samples of
from 3 to 60 wnoles oxygen. Quantitative yields of carbon dioxide were
obtained from samples containing less than 50 ymoles oxygen at I20S-
reactor temperatures above 125°C (see Figure 2-7). Incomplete yields of
carbon dioxide were obtained at lower temperatures even though the
thermal conductivity detector just downstream from the pyrolysis reactor
indicated that quantitative yields of CO were being produced. It. is
evident, therefore, that oxidation of CO to C0a by iodine pentoxide can
be slow and can lead to non-quantitative yields of carbon dioxide at the
flow rate required by the analyzer (25 mL/min.). The isotopic
composition of the C0a, however, did not vary even when the oxidation of
CO was incomplete. This result demonstrates that isotope effects in the
oxidation of carbon monoxide by iodine pentoxide are minimal. During
routine analyses, the temperature of the iodine pentoxide was maintained
between 120 and 140°C to insure quantitative yields.
Correction Parameters
Blank, 1,0^. Values of l'Fn, nn, and 1§FT were determined by five
— y W •>•
replicate analyses each of samples of benzoic acid designated 0 and 1 in
Table 2-3 (step I). The sample sizes in each set of analyses varied
-------
70
70
4
Figure 2-7. A plot of yield of carbon diojude (/anole CO,) vs
^
oxygen input (ymole 0) for sucrose at various temper&tures of the 1,0,
reactor: 25"C (©), 50'C (0), 75«C (A), 100»C (O), and 125«C (D)«
-------
71
3|OLU/7'
-------
72
from 30 to 2 wool 0. Memory equilibration was obtained by processing
five samples containing 40 nmol 0 of the same material prior to each
set. A plot of 2-1*FCO vs l/ns for these data is shown in Figure 2-8.
The curves were obtained by fitting each set of data to an equation of
the form y * {O'Fg * Ax)/(l + Bx)} + C (that is, the form of equation
2-19).
The isotopic composition of the iodine pentoxide-derivsd oxygen was
determined from the y-intercept of each curve (equation 2-20). The size
of the blank was determined from the difference between the two curves
'
at l/ns = 0.1 (tfDol O)-1 (equation 2-21), and the isotopic composition
of the blank was determined from the slope of each curve at the same
point (equation 2-22). These results are shown in the "initial
estimates" column of Table 2-5.
Memory. The washout of a single contribution to the sorbed-oxygen
pool was quantitated by analysis of a series of 23 samples of glycine
which was interrupted at positions 9 and 16 with glycine highly enriched
in oxygen-18 (step II). Results are summarized Figure 2-9. The
contribution (n.) of the enriched sample to each total oxygen pool was
initially determined using equation 2-25. The values of n, were fit to
an equation of the form f(j) = a(j)b (32). Once initial estimates of a
and b were established, equation 2-23 was used to calculate values for n,
in equation 2-24, and values for a and b were refined.
Values for l*Fj» 1§F0» and nn were then recalculated using these
values of a and b to specify n in equations 2-21 and 2-22 (step III).
The results appear in the "first iteration" column in Table 2-5. Three
iterations were necessary to obtain convergent values for the correction
parameters (Table 2-5). For every recalculation 1§FI, X*FQ» and "o'
-------
73
Table 2-5: Iterative determination of correction parameters
initial
estimates
10' x »»Fi 1.9895
10' x *»FQ 2.0413
Hq 0.71
1 0.072
a 2 0.080
3 0.082
1 -1.80
b 2 -1.72
3 -1.75
iterative determinations
1 2
1.9895 1.9895 1.
1.9766 1.9745 1.
1.00 1.01 1.
0.080 0.081
0.081 0.082 0.
0.081 0.081
-1.56 -1.60
-1.61 -1.60 -1.
-1.61 -1.60
3
9895
9750
02
081
60
Three iterations of the memory-correction parameters (a and b) were
performed for each refinecent of the blank parameters (nn and 1*Ff)).
-------
74
74
Figure 2-8. A plot of 2-F^ vs. 1/ns for five replicate
analy.es of two benzole acids, -FBA.O = 2.03547 x 10- and "F
2.04757 K 10-; uncorrected: BA-0 <•), BA-! («. corrected for'th^
blank and the 1,0,-added oxygen: BA-0 O, BA-1 (O>.
BA-1
-------
2.05
2.04
75
4.02
4.01
2
cvi
4.00
3.99
-i L
O.I
0.2
0.3
0.4
05
-------
76
76
Figure 2-9. A plot of *»FS vs sample number for a series of
glycine (ltFl, = 2.06012 x 10-3) analyses interrupted at positions 9 and
16 with highly enriched material (1§Fj «* 4.489 x 10-3); corrected for
the blank and the IjO,-added oxygen (©); corrected for the blank, the
IjO,-added oxygen, and for the memory (O)» The sample size was constant
at 26 wrool 0. The filled symbols are the raw results corrected only for
the blank and the iodine pentoxide-added oxygen.
-------
77
1 I 1 I
—
-
—
—
• •
^
—
"
—
-
mm^ ^h
\ - ^^^^^S^S.
—
•—
V
—
_J — 1 J 1
IIIIIIII[|IIIP
* o -
• c
• (
• 0
_«...._ ^ '
• (
0
5 -
D—
"
) -
O—
• o -
* c
• o
o
0>
-
—
^ ®
(
«
V
<
1
1
t
'
I ••••
» -
t—
) -
^_
• -
1 1 1 1 1 1 1 1 1 1 1 1 1 I
3 8 !Q 2 ?
U o
CJ
0
r\i
CD
•
^
C\J
O
•
CO
*
CM
O
5
J cJ evi cvi
0
XI
1
©
"o.
E
-------
78
three successive refinements were necessary to adjust the parameters a
and b. The calculated value of the isotopic composition of the iodine
pentoxide-added oxygen remained constant during this calculation,
whereas the other values changed substantially between the first and the
second iterations (Table 2-5). Convergence was rapid once values for
all terms were specified.
The limits for convergence were chosen so that the total error due
to uncertainty in the correction parameters was smaller than the
observed precision of repeated measurements, s* = 2 x 10-' (see be low).
Convergence limits accepted for the various correction parameters were:
nQ , 0.05 wool 0; I>^QI 10-*; I§FI» 10-7; a, 0.002; and b, 0.02. The
<;
standard deviation of calculated values of 1$FS will be less than 2 x
10-' under these circumstances, provided ns is greater than 5 wool 0 and
**F is within a factor of 2 of-all other components. In general* the
mathematical correction is most sensitive to 1'FI, and this value must
be highly accurate under all conditions. Uncertainties in the terns
specifying the contribution of the memory are much less important when
memory effects are small. Thus, uncertainties in the values of a and b
can be as large as 0.5 and 0.1, respectively, when samples within the
natural abundance range are analyzed.
Evaluation of the Correction Procedure
The correction procedure outlined here incorporates some assump-
tions that are essentially untestable. We cannot be certain, for
example, that sample-derived oxygen retained in the sorbed-oxygen pool
does not differ in isotopic composition from that in the initial sample.
Although it would be extremely difficult to show that all such assump-
tions are correct, it is possible to show at least that the correction
-------
79
procedure performs adequately. Three questions arise and will be
discussed in sequence: i) is the functional form of the correction
procedure essentially correct, ii) what levels of accuracy and precision
can be obtained, and iii) are correction parameters derived with one set
of test samples generally applicable?
Validation of the Model. Results summarized in Figures 2-8 and
2-9 show that the correction procedure has accurately adjusted results
obtained during experimental runs designed for evaluation of correction
parameters. The open symbols in both figures are the results of these
calculations. The correction procedures accurately describe the
contribution of the blank over a range of sample size from 2 to 30 umol
0. The means of the calculated values of 1§FS in both sets of data
(open symbols in Figure 2-8), 1'FBA_0 * 2.03512 ± 0.00030 x 10*J and
ltpBA-l = 2-04768 * 0.00028 x 10-' (x ± S.E.), do not differ
significantly (p < 0.05) from the true isotopic compositions of both
materials, as determined by the pyrolytic decarboxylation procedure.
The effect of the blank on the isotopic composition of the carbon
dioxide (the filled symbols) becomes more pronounced, shifting 2«1>FCO
further from the y-intercept, as the size of the sample decreases. A
comparison of the slopes for each curve also reveals that the effect of
the blank is, as expected, greater for benzoic acid-1. However, the
correction procedure effectively eliminates these variations.
The effects of the memory are also adequately described and taken
into account by the mathematical correction. The exponential relation-
ship, a(j) , closely described the wash-out of samples 9 and 16 in the
total oxygen pool, with a correlation coefficient for a least-squares
fit of r» - 0.995 (Figure 2-9). Thus, equation 2-23 can be used to
-------
80
estimate n<, and to provide an accurate indication of the amount of a
particular component in the sorbed-oxygen pool for the memory correc-
tion. The open symbols in Figure 2-9 are the calculated values of lfFs
for samples 10-15 and 17-23. Only the values for samples 10 and 11
differ significantly (p > 0.05) from the mean of samples 1-8, which is
unaffected by the memory. This example is an extreme demonstration of
the memory effect, the great difference in the isotopic compositions of
the two materials resulting in a large shift in the abundance of
oxygen-18 in carbon dioxide produced from samples following each
enriched sample. The effect of the memory will be small and more easily
guantitated when samples of similar isotopic composition are analyzed.
Accuracy and Precision. Accurate results were obtained in
analyses of the benzoic-acid standards (Table 2-6). The correction
parameters were refined using the sample designated 0; the only value
found to differ from those in the original parameter set was that for
1(F0. The mean isotopic abundances of the remaining samples (1 through
4) were compared to the results obtained by the pyrolytic
decarboxylation procedure. A paired t-test indicates that the only
significant difference between calculated and true isotopic compositions
occurs for sample 1 (p > 0.05). This discrepancy, albeit small, is
possibly the result of impurities in this sample, which is unique in
that it was used "as supplied," and was not equilibrated with water and,
consequently, was not recrystallized. If the impurities happened to
yield CO, on pyrolysis, the pyrolytic result could have been affected.
The precision of the analyses in Table 2-6 is typical of this tech-
nique. In general, precision decreases as the sample becomes more
enriched or depleted in oxygen-18 with respect to the blank and 1,0,-
-------
81
Table 2-6: Comparison of Isotopic Ana]yses of Benzole Acids
by Pyrolytic Decarboxylation and by the
Schutze-Untarzaucher Procedure
Pyrolytic a
Decarboxylation
1IFBAX10> .
Schutze- b>c>d
Unterzaucher
1SFBA x 101
Benzoic Acid-0
2.03548 ± 0.00015
2.03544 ± 0.00022
(-0.02)
Benzoic Acid-1
2.04756 ± 0.00009
2.04957 ± 0.00018
(+0.98)
Benzoic Acid-2
2.11705 ± O'.OOOIS
2.11782 ± 0.00087
(+0.36)
Benzoic Acid-3
2.18844 ± 0.00015
2.18976 ± 0.00095
(+0.60)
Benzoic Acid-4
2.24562 ± 0.00040
2.24429 ± 0.00066
(-0.59)
oean ± S.E. of 5 replicate analyses
b
mean ± S.E. of 4 replicate analyses
The correction parameters were optimized for Benzoic Acid-0:
nQ « 1.0 wnol 0, IIFQ = 1.977 x 10-J, »»FT « 1.9740 x 10-»,
a » 0.081, and b « -1.60. All other isotopic abundances were
calculated with these values.
/l«F -1> X
Schutze-Unterzaucher pyrolysis '
1000
-------
82
added oxygen. This occurs because small variations in the contaminant
oxygen pools have, under these circumstances, a more pronounced effect
on the calculated result. For samples in the range of natural abundance
(e. g., benzoic acid samples 1 and 2), the standard deviation of *•? is
typically 2 x 10-' (equivalent to 0.1 %„). The trend toward increasing
standard deviations is evident in the results obtained from the more
highly enriched samples.
General Applicability of Correction Parameters. Sanple matrix
effects have been identified. That is, the chemical form in which the
oxygen is introduced can affect the analytical result. Differences in
the reactivity of the initial oxygen-containing pyrolysis products with
the carbon, which persist even at high temperatures (32), and
interaction of non-oxygen containing gases such as Na, K,, and H,S (18,
20} with the surface oxide are probable causes. Doping the carrier gas
(i. e., with hydrocarbons and halocarbons) minimizes these effects, but
does not eliminate them completely.
An example of these effects is shown in Figure 2-10, which plots
2*"FCO vs. l/ns for a series of analyses of acetanilide, benzoic acid,
and glycine (filled symbols, lower panel). Three different curves,
indicative of different values for Tin, are obtained. These differences
are particularly clear in the upper panel of Figure 2-10, which plots
the calculated (symbols) and the true (lines) isotopic compositions of
each sample. The causes of these differences are not known in detail,
but it is notable that, to obtain the corrected values in the upper
panel, the correction parameters have been optimized for benzoic acid,
which yields COj as its principal pyrolysis product, whereas acetanilide
yields CO and glycine yields a mixture of CO and CO,. The deviations
-------
33
83
Figure 2-10. A plot of 2-1»Fco vs l/ng for benzole acid, glycine,
and acetanilide; corrected for the IaO,-added oxygen: benzole acid (©),
glycine 3D, and acetanilide (A); corrected for the blank, the memory,
and the I20,-added oxygen: benzole acid (O), glycine O, and acet-
anilide (A). The correction parameters were optimized for the benzole
acid.
-------
84
2.07 -
IO
O
& 2.06 h
2.05
406
r 1 1 1 1 1 1 ' 1
A
A
- A
a
: a
O
- , 9 ...
1
•
-
p"
a ;
B:
A I
„ '
.
4.05
10
O
CVJ
4.04
4.03
0.0
O.I
J_
0.2
l/ns
0.3
0.4
0.5
-------
85
for acetanilide and glycine at the smaller sample sizes *ie plausibly
the result of differences in the size of the blank for CO and C02. The
presence of N, apparently also influences the blank.
The effects of nitrogen and sulfur on analytical results are
evident in the analyses of heteroarorcatic carboxylic acids (Table 2-7).
The correction parameters were established using benzole acid, which, of
course, yields no sulfur- or nitrogen-containing pyrolysis products.
The results calculated for the Schutze-Unterzaucher procedure for the H-
and S-containing materials differ significantly from the accurate values
determined by pyrolytic decarboxylation. Sensitivity to heteroelement
content is also observed for the blank in oxygen elemental analysis
using the Schutze-Unterzaucher procedure; the blank is found to be
smaller for C,H,0 compounds than for C,H,N,0 or C,H,0,S compounds
(18-20, 33). The differences seen in Table 2-7 arise from an increase
in the size of the blank due to the presence of nitrogen or sulfur in
the sample. Paired t-tests indicate that all differences are
significant (p > 0.05) if n « 1.0 wool 0, but that the results of the
two procedures do not differ significantly (p < 0.05) if n » 3.0 insol 0.
The mathematical model accurately reflects dependence of the size of the
blank on heteroelement content of the sample. The importance of
choosing a working standard which closely matches the composition and
structure of the sample in order to correctly establish the size of the
blank and obtain the greatest accuracy is thus demonstrated.
Standards
Analysis of Standard Waters and Carbohydrates. The primary
standard for oxygen isotopic abundances is "Standard Mean Ocean Water"
(34). Samples of sucrose intended as supplementary standards for the
-------
66
Table 2-7: Comparison of Isotopic Analyses of Heteroaroaatic Carboxylic
Acids by Pyrolytic Decarboxylation and by the
Schutze-Unterzaucher Procedure
Pyrolytic a Schutze- b>c
Decarboxylation Unterzaucher
up x 10' *«F x 10»
A
d c
n » 1.0 n « 3.0
pyrole-2-
carboxylic acid 1.97134 ± 0.00006 1.97756 ± 0.00018 1.97229" ± 0.00018
(+3.15) (+0.51)
2-thiophene s
carboxylic acid 1.97804 ± O.C0005 1.98241 ± 0.00018 1.97771 ± 0.00020
(+2.21) (-0.17)
benzoic acid-0 2.03548 ± 0.00015 2.03530 i 0.00039. 2.03557 ± 0.00041
(-0.09) (+0.04)
2,6-pyridine
carboxylic acid 2.03740 ± 0.00017 2.03827 1 0.00010 2.03875 ± 0.00012
(+0.43) (+0.66)
mean ± 5.E. of 5 replicate analyses
b
mean ± S.E. of 4 replicate analyses
CA' ^''Schutze-ttiterzauche/^pyrolysis-" « 100°
d
n0» 1.0 Mmol 0, all other correction parameters were optimized for
behzoic acid: 1§F0« 1.975 x 10-», »»FX » 1.98500 x 10-J, a -
0.077, and b » -1.60. All other isotopic abundances were calculated
using these values.
nQ« 3.0 MiBOl 0, all other correction parameters were optimized for
bfchzoic acid-0: l'FQ- 1.975 x 10-', »»Fj» 1.98975 x 10-', a «
0.077 and b « -1.60. All other isotopic abundances were calculated
using these values.
-------
87
analysis of oxygen isotopes in organic materials have been distributed
by I. Friedman (4). Because sucrose should yield water as its
predominant oxygen-containing pyrolysis product (19), direct comparison
of results obtained for water and sucrose standards should be possible.
Results of the analysis of cane and beet sugar (4) and standard
waters (35) are shown in Table 2-8. The correction parameters were
optimized for V-SMOW (34). A paired t-test indicates that the results
calculated for KBS-1 and NBS-1A do not differ significantly (p < 0.05)
from their accepted "0 abundances (34). The calculated values for cane
and beet sugar are both slightly greater than previous measurements (4-
6,36). At least four different laboratories have analyzed these
samples, each reporting a slightly higher oxygen-18 content than the
previous measurement, and each accounting for the discrepancy by virtue
of reduction of blank and memory effects in their respective procedures.
Our measurements extend the trend. The accuracy of our results for NBS-
1 and NBS-1A in the same series of analyses suggests that the effects of
the blank and of the memory have been eliminated in the mathematical
correction and that the isotopic abundances for cane and beet sugar
reported in Table 2-6 are, in fact, correct.
Calibration of Working Standards. Although small variations in
the size of the blank are unimportant if large samples are used, the
effects of these variations can be minimized for samples of all sizes if
working standards yielding pyrolysis products similar to those of the
sample are available. The provision of such standards is always
possible because the accurate isotopic composition of any sample can b«
determined by obtaining multiple analyses and fitting a curve similar to
that shown in Figure 2-4. Such multiple analyses are tedious, but need
-------
88
Table 2-8: Analysis of Water Standards and Sucrose Samples
a,b
Accepted value Schutze- c
Unterzaucher
»»P x 10J »»P x 10*
(S^'O vs SHOW) (&**0 vs SHOW)
°/6o %c
cane sugar 2.0627 - 2.0685 2.06978 ± 0.00073
(+31.3 - +34.1) (+34.7)
beet sugar 2.0489 - 2.0547 2.05373 ± 0.00055
(+24.3 - +27.2) (+26.7)
V-SMOW 2.00044 2.00042 ± 0.00004
(0.00) (-0.01)
KBS-1 1.98499 1.98496 ± 0.00035
(-7.74) (-7.75)
NBS-1A ' 1.9519 1.9526 ± 0.0010
(-24.3) (-24.0)
a
Samples o2 cans and beet sugar were provided by Dr. I. Friedman (4).
The range indicates results obtained by other investigators (4-6,32).
b
The absolute abundance of oxygen-18 in V-SMOW was measured by
Baertschi (30), and the abundances of oxygen-18 in NBS-1 and NBS-1A
relative to V-SMOW were measured by Gonfiantini (31).
Results of the triplicate analysis of samples containing approximately
50 junol 0. The correction parameters were optimized for V-SMOW:
n « 1.2 wnol 0,l'FQ » 1.977 x 10-», 1§FT « 1.9890 x 10-',
a = 0.077, and b B -1.50. All other isotopic abundances are
calculated with these values.
-------
89
not be repeated once a standard allowing adequate determination of
correction parameters has been established. As an example, a plot of
2'1'FCQ vs. l/ns for benzoic acid, glycine, mouse lung hcmogenate, and
mouse blood is shown in Figure 2-11. Samples of benzoic acid were used
to determine 1§Fj and the values of l*Fg for each of these materials
were determined from the y-intercept. These materials can now be used
as working standards in the analysis of similar tissue samples, where
the values of nQ and 1*FI for a particular set of analyses are optimized
by the procedure previously discussed in the section titled "Adjustment
of Analytical Results".
EXPERIMENTAL SECTION
Isotopic Standards '
Pyrolysis of Benzoic Acid. .-Benzoic acid (elemental analysis
standard, Erba, Inc.) was pyrolyzed in sealed, evacuated quartz tubes.
Prior to use, the quarlr tubes (o.d. » 6 mm, length « 15 cm) were heated
in air at 550»C for 12 hours, soaked in concentrated hydrochloric acid
for 12 hours, and again heated in air at 550*C for 12 hours.
Immediately before use, each tube was evacuated, heated to 800*C with a
flaae, and allowed to cool under vacuum. A sample of benzoic acid,
approximately 3.6 mg (30 wnoles), was weighed to the nearest ug with a
Cahn electrebalance. The tube was vented, the sample was added, and the
tube was then evacuated for 60 to 120 sec prior to sealing. The sealed,
evacuated tube was placed in a preheated muffle furnace for the
specified time, removed from the furnace and allowed to cool in air.
In order to investigate reaction 2 under the conditions of this
work, approximately 30-ttnole aliquots of carbon dioxide (99.99%,
-------
90
90
Figure 2-11. A plot of 2-1§Fco vs l/ns for benzole acid (©), acet-
anilide (^), mouse blood 03>, and mouse-lung homcgenate (A). All results
were corrected for the I,0t-added oxygen.
-------
91
-------
92
Matheson) were sealed in quartz tubes cleaned as described above and
placed in a muffle furnace, preheated to 550'C, for times from 0 to 60
minutes. The tubes were removed from the furnace and allowed to cool in
air.
Preparation of Carbon Dioxide. Carbon dioxide from sealed-tube
experiments was purified by cryogenic distillation. The yield of C03
was measured using a gas buret and mercury manometer.
Mass Spectrometric Analysis. Isotopic analysis of carbon dioxide
was performed with a Nuclide 6-60 RMS isotope-ratio mass spectrometer
and isotopic abundances were calculated as described in Chapter 3.
Pyrolysis products other than carbon dioxide were identified using a
Kratos MS-80 high-resolution mass spectrometer.
Theoretical Calculations. The equilibrium constant (K) for •
reaction 2 for temperatures between 300 ar.d 1500'K was calculated from
values of the enthalpies and entropies of formation of CO and CO, (37)
and values of the heat capacities of CO, CO,, and graphite (36). The
equilibrium oxygen isotopic fractionation factor for this reaction was
calculated by means of the Bigeleisen-Mayer formulation (39), with
literature values of the force constants for CO and CO, (40). These
results were used to predict '"he yield of carbon dioxide and the shift
in isotopic composition of the carbon dioxide from that of the carboxyl
group.
Isotopic Standards. Five samples of benzoic acid with different
isotopic compositions were prepared by exchange with water (41). For
each of five isotepically different waters, approximately 4 g of benzoic
acid were added to 100 mL of water. Gaseous HC1 was bubbled through the
mixture until the concentration of HC1 was 5 N. The mixture was sealed
-------
93
under an atmosphere of N, and kept at 80»C fcr 120 hours. The benzole
acid was extracted in ether and the ether was removed under vacuum.
Samples of 2,6-pyridinedicarboxylic acid, pyrrole-2-carboxylic acid, and
2-thiophenecarbcxylic acid (Aldrich) were used as purchased. In order
to determine their oxygen-isotopic compositions, the products were
pyrolytically decarboxylated in sealed, evacuated quartz tubes at 550eC
for 10 minutes. 7he carbon dioxide was purified by cryogenic
distillation and tfe abundance of urygen-18 in the CO, was measured mass
spectrometrically.
Oxygen Isotopic Analysis
Apparatus. A Carlo Erba model 1106 elemental analyzer (Erba
Instruments, Inc.) was modified for the preparation of organic samples
for isotopic analysis (Figure 2-12). These modifications do not
interfere with the normal operation of the elemental analyzer, a
description of which has appeared elsewhere (42,43), and allow the
preparation of CO, for the determination of oxygen-18 or for the
preparation of CO,, N, and H,0 for determination of carbon-13,
nitrogen-15, and deuterium, respectively. Here, only procedures
pertinent to oxygen isotopic analysis will be discussed.
The effluent of the chromatographic column was redirected so that
it passed through an oxidation reactor before entering an auxiliary
vacuum line for collection, purification, and packaging of carbon
dioxide. The oxidation reactor was a pyrex tube, 10 mn o.d. x 40 cm in
length, 30 cm of which were packed with an iodine pentoxide reagent
(44). The remaining 10 cm were filled with dry silica gel. The I,0S
was heated to 135'C. A series of traps and a vacuum system were
constructed, so that carbon dioxide could be trapped from the effluent of
-------
94
94
Figure 2-12. Schematic diagram of modifications to the Carlo Srba
elemental analyzer used for the preparation of carbon dioxide from
organic material for isotopic analysis. Details of the operation of the
elemental analyzer appear elsewhere (42). Valves V2, V3, and V4 were
connected to the high-vacuum system at positions A, B, and C. Carbon
dioxide was trapped' from the affluent of the 1,0, reactor in T3.
-------
95
-g
v
cr
-------
96
the oxidation reactor, purified, and transferred to a sample tube for
mass spectrometric analysis. All valves were stainless steel and
utilized l/8"-low dead volume fittings (Valco Instruments Co.; VI, V2,
and V3 were type HP-6a; V5 was type HP-4a; and V4 was type HP-lOa).
Traps Tl, T2, T3, and T5 were multicoiled, containing 6 loops each, and
were constructed from type 321 stainless steel tubing, 0.125" o.d. x
0.010" wall (Tube Sales). Trap T4 was constructed from 6-mm pyrex
tubing and contained 0.1 g of molecular sieve 5A, 60/80 mesh ^Analabs).
The connections between the valves were made with teflon-lined aluminum
tubing, 1/8" o.d. x 1/16" i.d. (Analabs). Connections to the highvacuum
system were made using one-piece, graphite-impregnated Vespel ferrules
(Supeltex M-2A, Sup«lco). Standard two-piece Swagelok Teflon ferrules
(Crawford Fitting Co.) were used for other connections.
Reagents. High-purity helium, 99.995% (Air Products) was used as
the carrier gas for the elemental analyzer. The carrier gas was doped
with approximately 550 ppm of CC14 and 1500 ppm of C7H,,. These
materials ware mixed with the helium by volatilization from a toroidal
pyrex vessel with a 1-mm orifice, containing a 5% (v/v) solution of
carbon tetrachloride (Fisher, Spectranalyzed) in n-heptane (Fisher,
Spectranalyzed). The vessel was placed in the sample dispenser, just
upstream from the reaction chamber.
Iodine pentoxide was prepared by mixing a solution containing 50 g
of 1,0, (Mallinckrodt Analytical Reagent) in 100 mL of distilled water
with 100 g of 8-12 mesh silica gel (Preiser Scientific, Inc.). This
mixture was allowed to dry at 80°C for 96 hours. Concentrated sulfuric
acid (25 mL) was added to the dried material and the mixture was again
allowed to dry. Finally, the mixture was heated to a temperature of
-------
97
220»C for 4 h in a dry helium stream (44). The active reagent, a ccarse
yellow material, was kept dry and not exposed to light.
Isotopic Analyses. SampJes were weighed into silver-foil cups
(Erba, Inc.) on a Cahn 25 Electrobalance (Cahn Instruments, Inc.) and
placed in the s.impling-head of the instrument. The analyzer was
cperated normally, so that oxygen elemental abundances in .the samples
were also obtained. A dry ice-acetone bath was placed around Tl to
remove residual amounts of I,, a byproduct of the oxidation of CO by
1,0,, which were present in the carrier gas. To prevent aerosol forma-
tion and ensure quantitative recovery, T3 was ixnaersed to half its
length in liquid nitrogen when trapping carbon dioxide from ths analyzer
effluent. Valve V3 was switched to-direct the effluent of the oxidation
reactor through T3 60 sec after a carbon dioxide peak was detected by
thermal conductivity detector Dl (the delay was determined using D2).
Helium was removed and the carbon dioxide was transferred to a sample
tube for mass spectrometric analysis. Isotopic analysis of carbon
dioxide is described in Chapter 3.
-------
98
REFERENCES
1. Mook, W. G.; Grootes, P. M. The measuring procedure and
corrections for the high-precision mass-spectroaetric analysis of
isotopic abundance ratios, especially referring to carbon, oxygen,
and nitrogen. Int. J. Mass Spectrom. Ion Phys. 12:273-298, 1973.
2. Taylor, W.; Chen, I-J. Variables in oxygen-18 isotcpic analysis by
mass spectronetry. Anal. Chem. 42:224-228, 1970.
3. Fong, B. J.; Smith, S. R.; Tanaka, J. Reevaluation of the prepara-
tion of organic compounds for mass spectrometric analysis for
oxygen-18. Anal. Chem. 44:655-659, 1972.
4. Hardcastle, K. G.; Friedman, I. A method for oxygen isotope
analysis of organic material. Geophys. Res. Lett. 1:165-167,
1974.
5. Per hi, A. M.; Letolle, R. R.; Lerman, J. C. Oxygen isotope ratios
of organic matter: Analysis of natural compositions. Proceedings
of the Second International Conference on Stable Isotopes, 716-724,
1976.
6. Thompson, P.; Gray, J. Determination of II0/»»0 ratios in
compounds containing C, H, and O. Int. J. Appl. Hadiat. Isot. 23:
411-415, 1976.
7. Brenninkmeijer, C. A. M.; Hook, W. G. A batch process for direct
conversion of organic oxygen and water to CO, for l»0/140 analysis.
Int. J. Appl. Radiat. Isot. 32:137-141 1981.
8. Unterxaucher, J. The direct micro-determination of oxygen in
organic substances. Analyst 77:584-595, 1952.
9. Schutze, M. Die direh.e bestimmung des Sauerstoffs in Zinkoxyd.
2. Anal. Chem. 118:241-245, 1939.
10. Winter, K.; Barton, D. The thermal decomposition of benzoic acid.
Can. J. Chem. 48:3797-3801, 1970.
11. Maier, W. P.; Roth, R. V.; Thies, I.; Schleyer, P. v. R. Gass
phase decarboxylation of carboxylic acids. Chem. Ber. 115:808-115,
1982.
12. Rhead, T. F. E.; Wheeler, R. V. The effect of temperature and of
pressure on the equilibrium 2CO ~ — CO, * CO. J. Chan. Soc. 99:
1140-1153, 1911.
13. Stevens, K. H.; Pepper, J. M.; Lounsbury, M. The decarboxylation
of anthranilic acid. Can. J. Chem. 30:529-540, 1952.
14. Dunn, G. E.; Dayal, S. K. Mechanism od decarboxylation of
substituted anthranilic acids at high acidity. Can. J. Chem.
48:3349-?353, 1970.
-------
99
15. Dunn, G. E.; Lee, G. K. J. Kinetica and mechanism of the
decarboxylation of pyrrole-2-carboxylic acid in aqueous solution.
Can. J. Chen. 49:1032-1035, 1971.
16. Dunn, G. E.; Gordon, K. J. L.; Thimm, L. Kinetics and mechanism of
decarboxylation of some pyridinecarboxylic acids in aqueous
solution. Can. J. Chen. 50:3017-3027, 1972.
17. Rhead, W. F.; Wheeler, R. V. The rate of reduction of carbon
dioxide by carbon. J. Cheia. Soc. 101:831-845, 1912.
18. Haraldson, L. An investigation of the products of pyrolysis in the
determination of oxygen in sulphur-containing organic substances.
MikrochiB. Acta 1962, 651-670.
19. Belcher, R.; Ingram, G.; Majer, J. R. Direct determination of
organic compounds in the determination of oxygen. Mikrochim. Acta
. 1968, 418-426.
20. Belcher, R.; Ingram, G.; Majer, J. R. Direct determination of
oxygen in organic materials-I. Talanta 16:881-892, 1969.
21. Bonner, F.; Turkevich, J. J. Study of the carbon dioxide-carbon
reaction using C14 as a tracer. J. Am. Chem. Soc. 73:561-564,
1951. '
22. Oring, A. A.; Sterling, E. Oxygen transfer between carbon dioxide
and carbon monoxide in the presence of carbon. J. Phys. Chem. 58:
1044-1047, 1954.
23. Ergun, S. Kinetics of the reaction of carbon dioxide with carbon.
J. Phys. Chen. 60:480-485, 1956.
24. Vastola, F. J.; Hart, P. J.; Walker, P. L. A study of carbon-
oxygen surface complexes using Oi§ as a tracer. Carbon 2:65-71,
1964.
25. Manser, M; Ergun, S. Kinetics of oxygen exchange between CO, and
CO on carbon. Carbon 5:331-337, 1967.
26. Grabke, H. J. Oxygen transfer and carbon gasification in trie reac-
tion of different carbons with CO,. Carbon 10:587-599, 1972.
27. Treoblay, G.; Vastola, F. J.; Walker, P. L. Thermal desorption
analysis of oxygen surface complex on carbon. Carbon 16:35-39,
1978.
28. Gouverneur, P.; Schreuders, M.; Degens, P. N. The direct
determination of oxygen in organic compounds. Anal. Cl^im. Acta 5:
293-312, 1951.
29. Tamai, Y.; Watanabe, H.; Totaita, A. Catalytic gasification of
carbon with steam, carbon dioxide, and hydrogen. Carbon 15:103-
106, 1977.
-------
100
30. Xirsten, W. J. Micro and trace determination of oxygen in organic
compounds. Anal. Chim. Acta 100:279-288, 1978.
31. Freeman, G. B.; Reucroft, P. J. Adsorption of HCK and H,0 vapor
mixtures by activated and impregnated carbons. Carbon 17:313-316,
1979.
32. Hartley, P. E. Corrections for memory effects in analytical
instruments. Anal. Chem. 54:148-150, 1962.
33. Belcher, R.; Ingrain, G. Some observations on the microdetermina-
tion of oxygen in organic materials. Microchem. J. 11:350-357,
1966.
34. Baertschi, P. Absolute "0 content of Standard Mean Ocean Water.
Earth Planet. Sci. Lett. 31:341-344, 1976.
35. Gonfiantini, R. Standards for stable isotope measurements in
natural compounds. Nature 271:534-536, 1978.
36. Wedeking, K. W.; Hayes,^J. M. unpublished results.
37. Rossini, E. D.; Wagman, D. W.; Evans, W. H.; Lavine, S.; Jaffe, I.
Selected values of thermddynamic properties, National Euro of
Standards Circular 500, 1952.
38. Spencer, H. M. Emperical heat capacity equations of gasses and
graphite. Ind. Eng. Chem. 40:2152-2154, 1949.
39. Bigeleisen, J.; Mayer, M. G. Calculation of equilibrium constants
for isotope exchange reactions. J. Chem. Phys. 15:261-267, 1947.
40. Stern, M. J.; Spindel, W.; Monse, E. V. Temperature dependence of
isotope effects. J. Chem. Phys. 48:2908-2919, 1968.
41. Roberts, I.; Urey, H. C. Kinetics of the exchange between benzole
acid and water. J. Am. Chem. Soc. 61:2580-2584, 1939.
42. Pella, E.; Colombo, B. Improved instrumental determination of
oxygen in organic compounds by pyrolysis-gas chromatography. Anal.
Chem. 44:1563-1671, 1972.
43. Pella, E.; Colombo, B. Study of carbon, hydrogen, and nitrogen
determination by ccmbustion-gas chromatography. Mikrochim. Acta
Wien 1973, 697-719.
44. Smiley, W. G. Note on reagent for oxidation of carbon monoxide.
Nucl. Sci. Abstr. 3:391-392, 1949.
-------
101
CHAPTER 3
MASS SPEC7ROMETRIC DETERMINATION OF CARBON-13,
OXYGEN-17, AND OXYGEN-18 IN CARBON DIOXIDE
-------
102
Measurement of the mass spectrum of a volatile molecule can
quantify only the relative abundances of molecular species of
differing mass. Commonly, elemental isotope ratios are not measurable
directly, but must be determined by calculations based on the observed
molecular spectrum (1). In the case of carbon dioxide, for example, ion
currents appear at m/z 44 through m/z 49. With the exception of m/z 44
(11C1»0,), all measurable ion currents represent sums of multiple
isotopic contributions (e. g., »»C1T01«0 + »»C1<01'0 + 13C1T0, at m/z
46). A problem common to all mathematical formulations (2-4) is that a
measurement of two ion-current ratios (45/44 and 46/44) does not provide
sufficient information to obtain unique solutions for three isotope
ratios (i3C/12C, 1?0/140, and "O/^'O).
The oxygen-isotope ratios have, therefore, commonly been treated as
functionally related; that is, an expression for 170/1'0 as a function
of '•0/ltO-has been written and used as the third equation required for
solution of the problem. This approach is based on the fact that proc-
esses (other than admixture of separated isotopes) which have affected
the abundance of 1»0 must also have affected the abundance of 1-I0, and
vice versa (5). Recently, however, it has been shown (i) that no
single, fixed relationship correctly describes the fractionation of J'0
relative to 1»0 in all processes, and (ii) that the relationship
commonly employed by isotopic mass spectrometrists does not correctly
describe even the "average relationship" prevailing in natural systems
(6). Here, we explore the significance of these findings for isotopic
analysis and show that a simple system of equations for accurate
calculation of isotopic abundances can be developed.
-------
103
DEVELOPMENT OF ISOTOPE-RATIO EQUATIONS
Notation. The letter R is consistently used to designate the
ratio of a particular isotopic species to the related isotopic species
of lowest mass
isotopic species of mass n
R * (3-1)
base species related to n
Left superscripts will be consistently used to designate isotopic
species by mass. For n * 13, the base species is 11C; for n * 17 or 18,
*«0; and, for r. » 45 through 49, the base species is 11C1»0,. The
letter F is consistently used^in the sane way to designate fractional
abundances
isotopic species of mass n
F - -r— (3-2)
sum of all species related to n
for n » 13, the denominator is the sum (1SC •*• 11C); ior n • 17 or 18,
(i*0 •+• i'o * 1§0); and, for n« 45 through 49, the denominator is the sum
of all species of carbon dioxide. The symbol a is used to designate the
fractionation factor, in this case, the ratio of isotope ratios
characteristic of the isotope effect associated with a physical or
chemical process
n
"products
o « (3-3)
nR
"reactants
for n = 13, 17, or 18.
Isotopic Distribution. The isotopic species present in carbon
dioxide are shown in Table 3-1. Since the vibrational modes in CO, are
nearly harmonic and coupling between modes is weak (7), the distribution
-------
104
Table 3-1: Isotcpic Molecular Species of Carbon Dioxide
44 45 46 47 48 49
-------
105
a particular isotopic species is accurately described by a simple
probability function <1). The abundances of the various molecular
species are, therefore, related to the isotopic abundances by the
following equations
44p » I2f.ltf.ltf (3-4)
4»p s up.itp.itF 4. 2-1 IF-*- •F- * 'F (3-5)
44p B itf.iif.iif + 2*l2F-**F-l'F + 2«13F>1*F*J'F (3-6)
tif m iif.tif.nf 4- 2- * JF' * *F* * *F * 2- 13F*17F' * *F (3-7)
4*p B iap.np.i»p ^ 2>"F*I'F*1*F (3-8)
4»F . up.i.p.i.p (3_9)
In practice, it is nost convenient to express these relationships in
terms of ratios. Dividing equations 3-5 through 3-9 by equation 3-4
yields
4.R » 1JR * 2-»'R (3-10)
44R « 2-l«R * 2-»JR-»'R •«• »'R» (3-11)
47R « 1JR.1TR1 4. 2-»>R-»«R + 2-17R«>»R (3-12)
««R - >«R» + 2-1>R'»'R'*»R (3-13)
4»R . iJR.liRJ (3-14)
For samples with isotopic abundances near natural levels, ion currents
large enough to allow precise, rapid measurement occur only at masses
44, 45, and 46. Modern instruments employ a triple-collector system to
-------
106
facilitate simultaneous measurement of *'R and 44R (Figure 3-1).
Relationships Between 1J>R and **R
Theoretical Considerations. Fractionations of multiple isotopes
of the same element are proportional to differences in mass (5). A
quantitative relationship between the values of »'R and 1§R for two
oxygen pools (designated by subscripts 1 and 2) linked by a single,
mass-dependent process has been presented by Clayton and coworkers (6)
(3-15)
where a « ln(* 'a)/ln(» §a) . Note, however, that (i) a is not constant
4 !
and (ii) equation 3-15 refers to instantaneous precursor -product
relationships, not to the integrated compositions cf separated or mixed
isotopic pools. Both of these points are significant in the isotope-
analysis problem.
The value of o depends on the process causing the isotopic
fractionation. Further, variations of »*a and »«o are not linked in a
way that leads a to assume a constant value. As a result, relative
fractionations of the oxygen isotopes can depend on the molecular
species involved or on the mechanism of a particular reaction. In
general, 0.50 < 2 < 0.53 (6). For a series of reactions, the value of a
for the overall process is the weighted average of the values for each
step
I{a InO'cO}
a . - (3-16)
where a^ is the exponent and »»ai is the cxygen-18 fractionation
-------
107
107
Figure 3-1. Schematic representation of an isotope ratio aass
spectrometer. Reference and sample gases are alternatively introduced
into the mass spectroneter, and ion-currents at m/z 44, tn/z 45, and m/r
46 are measured for each gas.
-------
108
^^ r^ Oy
~~o o ~o
ro cvj cvj
o o
ro eu
en
8 >
D %-
o>
cr
^^^^_Zk
?
">•
?
-------
109
factor for each step in the series of n reactions. In cases where a
reaction sequence imparts a large net change in the relative abundance
of one heavy isotope and a small net change in the abundance of the
other heavy isotope, the value of a2 is not restricted to the range
defined by the az, and may be very large or very small. Moreover, if a
net enrichment in one isotope accompanies a net depletion in the other,
then as will be negative.
Natural samples are never the products of a discrete process, or
even a single sequence of processes. For polyisotopic elements,
isotopic variations associated with mixing processes are complex. On
the one hand, relationships governing isotopic fractionations associated
with mass-dependent isotope effecSts are correctly cast in terms of
ratios. On the other hand, accurate calculation of the composition of
mixed pools must be based on fractional abundances. Even if a is known
accurately, equation 3-15 will only approximately describe isotopic
relationships associated with the mixing of oxygen pools with similar
isotopic compositions (i.e., natural samples of terrestrial origin).
In general, a will be some average of the ai weighted by the relative
contribution of each component to the total oxygen pool. If one or more
of the oxygen pools involved in a mixture has been selectively enriched
or depleted in a single isotope, serious inaccuracies can result from
application of equation 3-15 (6).
Given these considerations, it is evident that there is nc basis on
which an equation exactly relating 11R and 1(R in all pools of oxygen
might be developed. In particular, the relationship underlying most
oxygen-isotopic analytical calculations:
-------
110
(3-17)
where "unknown" and "standard" refer to a sample of unknown isotopic
composition and to some primary isotopic standard, is inexact and,
because 0.50 represents a minimal value for a, may not be the most
satisfactory approximation.
1TR and **R in natural Samples. Independent measurements
of 61''0 and 6ltO in lunar and terrestrial silicate minerals and samples
of natural waters quite closely define a single relationship (6,8,9)
S^OSMOW" 0.52«6»»OSJ!ow (3-18)
As indicated by use of the d notation, this relationship is based on
measurements of differences between isotopic compositions of unxnown
samples and isotopic standards. As such, it carries no direct
information regarding absolute values of isotope ratios. Calculations
of 17R and 1§R for unknown samples — and, for that matter, accurat e
calculations based on equations 3-10 and 3-11 — require knowledge of
»'R and 1»R for the isotopic standards. Although there have been recent
measurements of the absolute value of I*RSHOW ^and ****¥ are in 9CXX*
agreement, see refs 2,10,11), there have been no direct determinations
°f ItRSMOW
It is, therefore, of great interest that the isotopic compositions
of all natural terrestrial and lunar oxygen pools appear to be related
by equation 3-18, which is closely related to equation 3-15. The 6
terms are identical to the first terms in the series expansions of the
In terms in:
-------
Ill
(3-19)
\ »-R I
SHOW /
Higher-order tern.s are negligible when l'Runknown anr^ l 'Unknown are
close to l1RsMOW zx^ llpSMOW- Jt is evident that equation 3-19 is a
more general form of equation 3-18, which can be recast as
*'R « »«Ra-K (3-20)
where K is a constant characteristic of the relationship between 17R and
l»R in the terrestrial oxygen pool. Because the isotopic compositions
of all samples appear to be^related by this equation, evaluation of K
requires only that both 1-IR a^d ^R-be known accurately in, a single,
representative sample.
There has been only one set of measurements of the absolute
abundances of oxygen-16, oxygen-17, and oxygen-18 in samples of
terrestrial origin. Nier. determined values of the ratios ^C^'O/^'O,
(J)R) and (1101*0 * 11!0,)/l«0, (J«R) for samples of air and of 0, from a
commercial gas cylinder (12). Discordant results were obtained for the
atmospheric sample (perhaps because total air, not purified atmospheric
0,, was admitted to the ion source), but, for the latter analysis, Nier
reports (corrected mean and probable error) *JR « 7.55 t 0.01 x 10-' and
"R » 4.103 ± 0.005 x 10-J, which yield isotope ratios (mean and
standard error of mean) of l'Ro2 • (3.775 ± 0.074) x 10-« and l»Ro2 *
(2.0514 * 0.0037) x 10-». It follows that, for a « 0.52, K « (9.43 ±
0.18) x 10-».
Calculation of Isotope Ratios
Combining equations 3-10 and 3-11 and substituting equation 3-20
-------
112
for i'R yields an expression for »'R in terms of ion-current ratios, K,
and a. Apart from uncertainties in K and a, the equation is exact.
2K'4iR-»«Ra •«• 2-1§R - «'R - 0 (3-21)
Rearrangement of this equation to yield an explicit solution of the form
l'R = f(a, K, 45R, «'R) is not possible except where a is exactly 0.5.
An accurate value for 1(R can, however, be easily and rapidly obtained
using numerical techniques. A value for 11R can be determined from
equation 3-20 or, if a direct solution in terms of ion-current ratios is
required, a derivation similar to that of equation 3-21 yields
-3-i'R» + 2-«'R-"R * (l'R/K)1/Q - «'R » 0 (3-22)
The calculation of J1R follows directly from «'R using equation 3-10.
For clarity in subsequent discussions, we will refer to the
following sequence of calculations as "routine:"
1. »»R is calculated from «»R and ««R (eq. 3-21).
2. "R is calculated from »'R (eq. 3-20).
3. »'R is calculated from * 'R and «*R (eq. 3-10).
DETERMINATION OF ISOTOPE ABUNDANCES
The accuracy of calculated isotope ratios will depend on how well
equation 3-20 actually describes the linked variations of 17R and 1§R
and on the development of representative values for a and K. Here, we
will explore, by means of the application of equations 3-20 - 3-22 to
calculation of results from analytical data, two particularly
interesting cases:
(i) oxygen pools derived by mixing natural-abundance material
-------
113
with material highly enriched in 1§0
(ii) natural-abundance material with a range of oxygen-
isotopic compositions wide enough that correct assignment of the
values of a and K will be important
Mixed Samples Enriched in **0. If a sample has bean selectively
enriched in l»0, inaccuracies will result from use of the routine
sequence of calculations. Because ««R is primarily dependent on 1»R,
solution of equation 3-21 will yield a roughly correct value for »'R.
The value of *'R will, however, be overestimated by equation 3-20.
Insertion of that overestimate in equation 3-10 will lead to an
incorrectly low value for IJR. Further, use of equation 3-20 in the
derivation of equation 3-21 causes the initially calculated value of 1§R
(termed "roughly correct" above) to be low. Because the relative
contribution of the term {2«l'S} to «*R is greater than that of
{2-i»R.iiR + »'Ra} to «'R, the error in »»R will be greater than the
error in 1§R.
Clearly, this cascade of errors can be most appropriately prevented
by "uncoupling" "R and »»R (that is, by not using equation 3-20 to
express the relationship between 17R and »»R). Two approaches are
possible. If 11R is known independently, the sequence of calculations
that we will term "»»C-known" can be employed:
1. 1TR is calculated froa «»R and »»R (eq. 3-10).
2. *»R is calculated from «*R, »»R, and »7R (eq. 3-11). If 1->R is
known independently, or if it is expedient to adopt an estimated value
in order to reduce systematic errors in the analysis of samples for
which neither 1JR nor "R is known, but in which »'0 has been enriched
well above natural abundance, a sequence of calculations that might be
-------
114
termed "1'0-known" can be employed:
1. 1JR is calculated from «'R and "R (eq. 3-10).
2. *'R is calculated from ««R, »>R, and 17R (eq. 3-11).
To explore these problems and possibilities, we have prepared ar.d
analyzed mass spectrometrically a series of carbon dioxide samples
enriched in the heavy isotopes of oxygen. Oxygen in these samples was
derived by exchange with waters prepared by mixing 99% 1*0-water with
distilled tap water. Because only oxygen-isotopic compositions could be
affected by exchange, the true carbon-12/carbon-13 ratio in all samples
was identical.
4'
Results of the mass spectrometric analyses are summarized in Table
r»
3-2, which includes normalized ion-current ratios as we.ll as two
different sets of calculated'isotopic compositions. The columns headed
"R" tabulate apparent isotopic compositions derived from the routine
sequence of calculations. Samples bearing higher numbers have higher
abundances of "0, and it can be seen that, as the oxygen-18 content of
the samples increases, the apparent value of 6tlCpDg does not remain
constant, but, instead, decreases systematically. For samples enriched
in 1§0 by approximately 300%., the error amounts to almost -10%, (see
sample 21: the routine calculation yields -19.11%, even though the
correct result must still be -10.17%,,).
Use of the "l»C-known" sequence of calculations is appropriate in
this case. Columns headed "CK in Table 3-2 tabulate results. Because
routine analysis of the unexchanged material indicated a carbon-isotopic
composition of -10.17%. vs PDB, this value was assigned to all samples
ir. which "0 had been enriched by exchange. Comparison of cxygen-
isotopic compositions reported for each sample in the "R" and "CK"
-------
115
1
t/2
o
M
Li
01
•g
K
tb
I
O
a
01
e>
pi
X
o
0)
•o
-*-l
K
O
3
01
i-H
J3
1
u
a
ca
a
o.
V o
K
•O
I-
re
U)
ro o P- O «» CM rM o in r» mrMcneoco
en *» i*» en fH in o m in »-* en vo ro CM i»"*
inocncoin coinr-i-ir- vo 10 ro in en
m if\ vo p» en o CM *"o in 10 r* en »"4 CM ro
_| ,_, ,-t ,_!,_( rliHfMrMrM
vo CM fr^ CM *v co vo f^ f*s ^y ^j> ^4 r*1 ro ^n co
in enc*>r^coo «*cn*a|'«j|O oo^»oocn
tM inocncoin eo«»f«.i-ir- vovorJinco
f^ ro u^ 10 r** cn ^o CM '*1^ in vo p* CTV ^H CM F^
rHf<>-li-l>-< rHiHCMCMrM
mo**rMO cotnin^eo foin*^orM
vor-vocoo rofMioco-* nrMinoro
^* oo oo en vO c? en en en c? en t*4 ^^ ^D ^?
i-l«H>-fH rovocor-^
en vo O m «* in CM rn *f en en vo o ** vo CM
• ••••• ••'!• *••••
vo COG«*OCO inmC'ivoro cccom>-ico
•H i-i CM m «• «• invovor-co in c?\ o •-( —i
i-4 rH >-*
p^r*t**r*p» r*r*r*p*r* r-r^r-p*"^
fHi-i>-<>-4>-l i-l>-lt~(^tH i-4iHf-4r~4>-l
C3OOC3O O O O O O OOC3OC3
I I I I I I I I I I I I I I I
r- •vvocnvovo vorMCOiMco i~r-<-4inr-
i-( CMVD«HVOO mO«^O»rO p-fMOOCMVO
• ••••• ••••• •••••
O O o »H •—t CM cJror*»fo«* ^ininvovo
I I I I I I I I I I I I I I I I
CO
O
O
o
in
09
Ov
en
cr>
O r- •» m co
p- 10 rs PJ in
<*) m r« o\ \o
^ CO ID in r-t
O -^ m «. 10
O O O O O
vo r~ in co —i
o 10 «» m in
u> in vo IN n
•w o CM vo •-»
r- o\ o r» m eo t-«
O v o \o r>i
eo co o in o
O c\ •» t~ r->
*f m r» oo o
in o fM n n
<») o m m co
CO C7^ O^ O\ (7)
O\ O\ CTv OA C\
in 10 o i*> rM
m o 10 p- >-t
CA o CK o% o
en o ox c\ o
en o en o\ o
en o en co o
o o en CM r»>
«» vo O co O
en o o en o
en o o en o
»
vo r- oo en o r-< IM ro ^- in
-------
116
•o
!*:
1
to
? "a:
>o
*? u
1
CM
CM
r-4
VO
«•>
O
o
o
0
»4
CO
•^
CO
(M
O
VO
CO
CM
O
CO
O
co
rn
o
1
CV
r*
i
o
OV
(M
CM
W
CM
000050 1
CO
fH
VO
CM
CO
CM
O
CO
CM
en
in
CM
in
vo
m
0
1
••
CD
1
VO
CO
«~<
to
m
CM
000078 1
^4
o
CM
•H
ro
••
O'
o
l-(
n
e>v
CM
CO
in
f4
in
O
r»
f-
co
I
3
r-
O
r—
CM
000106 1
*-»
CM
iM
in
CM
p~
OV
CM
CM
CM
f>
CM
CO
C7>
in
«-<
0
i
*-t
t-<
01
1
o
CM
CO
CM
000184 1
CM
• r-
eo p>
in in
r-t C\
co «
CO —I
-4 O
O O
« 1
R Rl
C C
ea D
tn in
K K
m «
• •
T3 "O
•a -a
u
u
i
c
0
Jt
\
u
d CK indicates "'
8
1
u
01
s
a
n
VI
O
rt
in
u
-------
117
columns shows (i) that differences in the abundance of l§0 determined by
•
the routine sequence and the "13C-known" sequence of calculations are,
indeed, small, and (ii) that much larger changes in the calculated
abundance of '- '0 result from uncoupling its abundance from that of 1'0.
To generalize these observations, the difference between the
solution to equation 3-21 and the true value of l*R was determined for
all possible values of ltF ("0 and 1JC were assumed to be at nat ural
abundance). Results are summarized in Figure 3-2. More important in
tracer studies, the relative difference between the calculated end true
values of 1§F was also determined and is shown in Figure 3-3.
"R and **R in Natural paroles. To examine relationships between
*'R and 1>R typical of those found in natural systems, a series of water
f :
samples was produced by fractional distillation. Subsequent exchange of
oxygen between these samples and carbon dioxide yielded a series of
samples of CO, with widely varying "0 content. Because the same stock
of CO, was used to prepare all samples, the carbon-isotcoic compositions
of all samples were identical.
Results of mass spectrcraetric analyses of these samples are
summarized in the first two columns of Table 3-3. The third column,
based on routine calculations with a - 0.52 and K * 0.0094330 (although
the number of significant figures exceeds the precision with which K can
presently be estimated, the value specified exactly reflects the mean
isotopic abundances reported in ref. 14, and was employed in these
calculations), indicates the range of oxygen-isotopic compositions. The
fourth column tabulates carbon-isotopic abundances as determined in the
same set of calculations. Inspection of the latter results shows that
-------
IK
118
Figure 3-2. Relative error in ^R as a function oC oxygen-18
enrichment: I*KZQ " i-12372 x 10~3 **& 17RC02 * 3-73 x 10""•
solutions to equations 14 and 15 are obtained only at natural abundance
(»'RCO * 2.00520 x 10-').
-------
CO
O
h-;
O
119
(O
O
IT)
O
0>
O
O
"O
c
go
i
c
ro
O
cvj
O
X
O
CVJ
•f
CVJ
I
JOJJ3
ro
I
in
i
-------
120
120
Figure 3-3. Relative error in »'F as a function of oxygen-18
4 !
abundance: 1JRco e 1-12372 x 10-' and ''^co? = 3*73 x 10~4- Values of
l*FCO were calculated from equation 18 using solutions to equations 14
and 15. Exact solutions to equations''14 and 15 are obtained only at
natural abundance (I*FCO * 2.000443 x 10-J).
-------
121
I '
OJ
I
U! JOJJ8
ro
I
cp
d
rq
d
OJ
d
0)
o
o
CO
I
c
-------
122
Table 3-3: Carbon Dioxide-Oxygen Derived From Distilled Waters
Scunole
««nb
6»«0,
SHOW'
0.52 0.50 0.52
.009433 .008335 .010090
oc
1
2
3
4
5
6
7
8
9
10
11
12
13
14
0
0
0
0
0
0
0
0
.0
0
0
0
0
1
1
.999857
.996008
.996375
.996499
.996168
.996495
.996547
.996855
.997078 .
.997128
.997114
.997289
.997242
.000050
.000953
1.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
1.
001108
888515
897943
900305
900698
901266
906725
912028
915067
916327
916420
919115
919410
999867
026080
-82
-72
-70
-69
-69
-63
-57
-53
-53
-53
-50
-50
+32
f59
.30
.56
.11
.71
.12
.48
.99
.92
.55
.45
.67
.36
.83
.93
-10.
-10.
-10.
-10.
-10.
-10.
-10.
-10.
-10.
-1C.
-10.
-10.
-9.
-9.
18
14
09
46
13
17
15
06
02
04
12
01
93
89
-10
-10
-10
-10
-10
-10
-10
-10
-10
-10
-10
-10
-9
-9
.28
.23
.19
.55
.22
.26
.23
.13
.09
.11
.19
.08
.88
.81
-10.03
-10.02
-9.98
-10.34
-10.02
-10.07
-10.06
-9.98
-9.94
-9.96
-10.05
-9.94
-10.04
-10.06
1 "Runknovn/ ""standard' "Standard = 0-01188158.
b"Q " ""unknown/'""standard- ""standard » 0.004149577.
C5antple 0 was initial, unexchanged C02.
-------
123
the calculated carbon-isotopic composition is not indeptndent of the
oxygen-18 abundance, as it should be. In fact, regression of 61JC on
BltO indicates a slope of 0.0023%0C/%eO, a result which differs
significantly (p < 0.008) from zero.
Small variations in assigned values of a and/or K have almost no
effect on calculated values of 6»»0, but can substantially affect the
apparent variations in 61JC. As an example, the fifth column in Table
3-3 summarizes apparent carbon-isotopic compositions determined by use
of the routine sequence of calculations with a » 0.50 and K « 0.008335.
This corresponds to the present conventional approach (e. g., ref. 4).
Regression of 61JC on 6l»0 in this case yields a slope of 0.0036 %eC/%<>,
4:
a systematic error even greater than that encountered in the preceding
calculation. The sixth column in Table 3-3 reports apparent carbon-
isotopic compositions determined' by use of the routine sequence of
calculations with a » 0.52 and K = 0.010090. In this case, K was
adjusted to eliminate covariance between 61JC and 61§0.
CONCLUSIONS
We have presented a simple approach to the calculation of isotope
ratios from the mass spectra of carbon dioxide. A rigorous examination
of the effect of assuming a quantitative relationship between »*R and
»'R has shown (i) that determination of isotope ratios by the routine
sequence of calculations for carbon dioxide highly enriched in oxygen-18
will result in errors in the isotopic analysis, and (ii) when a « 0.52
is used in the routine sequence of calculations, a value of K » 0.010090
more accurately describes the mass-dependent fractionations of oxygen
isotopes by distillation and exchange. In the first case, errors in the
isotopic analysis can be eliminated by decoupling »'R and 1§R in the
-------
124
calculation, either with the "*• JC-known", or the "11 0-knovm" sequence of
calculations. In the second case, the value of K which eliminates
covariance between 61JC and 6^0 when a «= 0.52 requires that 1'RSMov«/'"
3.99 x 10-«. It is, however, necessary to reevaluate the oxygen-17
content of SHOW by more direct methods. A systematic study of oxygen-17
and oxygen-18 abundances in many different natural samples and
determination of values of a characteristic of representative processes
in nature are also suggested.
EXPERIMENTAL SECTION
Preparation of Carbon Dioxide for Isotopic Anal/sis. A series of
samples of water selectively enriched in oxygen-18 was prepared by
mixing varying amounts of 99%-H21*0 (Stohler Isotopic Chemicals) with
distilled, deionized water. A second series of water samples with mass-
dependent enrichment of both 1'0 and i$0 was produced by fractional
distillation of tap water. For the exchange of oxygen between water and
carbon dioxide, 200 vL of thoroughly degassed water and 20 lonoles of
carbon dioxide were placed in a sealed, evacuated Pyrex tube (6 mm
o.d.). The tubes were allowed to stand at 25*C for 48 hours to ensure
that exchange of oxygen between water and carbon dioxide was complete
(13). The carbon dioxide was purified by cryogenic distillation.
Mass Spectrometric Measurement. Values of 41R and 4*R used in
equations 3-16 and 3-17 must be numerically equivalent to the ratios of
molecular species in the sample. There are two sets of reasons why
recorded ion-current ratios may differ from absolute molecular ratios.
Specific corrections are required in both cases.
First, there is a set of measurable phenomena such as background
ion currents, amplifier-zero offsets, scattering of major-beam ions into
-------
125
minor-beam collectors, leakage of gas-switching valves, etc. Optimal
procedures for measurement and correction of the effects of these
artifacts have be>en extensively discussed (2-4). In the present work.
only corrections for background ion currents and amplifier-zero levels
have been required, and these have been applied prior to the calculation
of ion-current ratios.
Second, the efficiency of the production of ions varies for the
different isotopic species of carbon dioxide. If, however, this "mass-
discriEination" is constant over time, then this problem can be
eliminated using a reference measurement using carbon dioxide with a
4!
known isotopic composition at a working standard. Reference and sample
gases are alternatively introduced into the mass spectrometer, and ion-
currents at m/z « 44, m/z » 45, and m/r * 46 are measured for each gas
(Figure 3-1). The true molecular-isotope ratios of the sanple are
determined as follows
_ ** sample _
^sample * ~ * ^standard <3-23)
"standard
where °fc refers to the ratio of recorded ion currents, "R refers to the
true ratio of molecular species as defined in equations 3-7 through
3-11, and m and n « 45 or 46. Calibration of the working standard
requires only determination of 4§Rstandard and 4'Rstandard- A primary
reference provides a sample of carbon dioxide with a known isotopic
composition. Values of 4IR and «*R for the primary reference are
calculated from the known »'R, »'R, and i'R using equations 3-7 and 3-8.
The fractional abundance cf a particular isotope (j) is calculated
directly from the isotope ratios
-------
126
(3-24)
where the index i is an integral value indicating all stable isotopes of
a particular element; in the case of carbon i * {12, 13} and in the case
of oxygen i « {16, 17, 18}. For example, »»P = »»R/(1 + -»'R +»»R).
Delta values are calculated from their definition (14)
^sample
— 1 x 1000 (3-25)
Rstandard
Values of delta determined in this way do not incorporate mathematical
approximations.
Mass spectrometric measurements were made using a Finnigan MAT
Delta E isotope ratio mass spectrometer (Finnigan Corp.). Ion currents
at m/z » 44, o/z = 45, and m/z * 46 were measured for the sample and a
working standard. The contribution of instrumental background was
subtracted from the ion current and the molecular ratios 48R and "R for
both gases were determined. The working standard was calibrated using
carbon dioxide produced by phosphorolysis of K-2 and TKL-1 (15).
-------
127
REFERENCES
1. Margrave, J. L.; Polansky, R. B. Relative abundance calculations
for isotcpic molecular species. J. Chem. Ed. 39:789-793, 1962.
2. Craig, H. Isotopic standards for carbon and oxygen and correction
factors for tnass-spectrotnetric analysis of carbon dioxide.
Geochim. Cosmochim- Acta 12:133-149, 1957.
3. Deinis, P. Mass spectronetric correction factors fot the
determination of snail isotopic composition variations of carbon
and oxygen. Int. J. Mass Spectrom. Ion Phys. 4:283-295, 1970.
4. Mook, W. G.; Grootes, P. M. The measuring procedure and
corrections for the high-precision mass-spectrometric analysis of
isotopic abundance ratios, especially referring to carbon, oxygen,
and nitrogen. Int. J. Mass Spectrum. Ion Phys. 12:273-298, 1973.
5. Bigeleisen, J. The effects of isotopic substitution on the rates
of chemical reactions. J. Phys. Chem. 56:823-828, 1952.
6. Matsuhisa, Y.; Goldsmith, J. R.; Clayton, R. N. Mechanisms of
hydrotheriaal crystallization of quartz at 250"C and 15 kbar.
Giochiin. Cosmochim. Acta 42:173-182, 1978.
7. Bigeleisen, J.; Ishida, T.; Lee, M. W. Correlation of the isctopic
chemistry of hydrogen, carbon, and oxygen with molecular forces by
the WIMPER (2) method. J. Chen. Phys. 74:1799-1816, 1981.
8. Clayton, R. N.; Grossman, L.; Mayeda, T. K. A component of
primitive nuclear composition in carbonaceous meteorites. Science
182:485-487, 1973.
9. Clayton, R. N.; Onuma, H.; Mayeda, T. K. A classification of
meteorites based on oxygen isotopes. Earth Planet. Sci. Lett.
30:10-18, 1976.
10. Craig, H. Standard for reporting concentrations of deuterium and
oxygen-18 in natural waters. Science 133:1833-1835, 1961.
11, Baertschi, P. Absolute "0 content of standard mean ocean water.
Earth Planet. Sci. Let. 31:341-344, 1976.
12. Nier, A. 0. A redetermination of the relative abundances of the
isotopes of carbon, nitrogen, oxygen, argon, and potassium. Phys.
Rev. 77:789-793, 1950.
13. O'Neil, J. R.j Epstein, S. A method of oxygen isotope analysis of
milligram quantities of water and some of its applications. J.
Geophys. Res. 71:4955-4961, 1966.
-------
128
14. McKinney, C. R.; McCrea, J. K.; Epstein, 5.; Allen, H. A.; Urey, H.
C. Improvements in mass spectrometers for the measurement of small
differences in isotope abundance ratios. Rev. Sci. Inst. 21:724-
730, 1950.
15. Blattner, P»; Hulston, J. R. Proportional variations of
geochemical 6l*0 scales. An interlaboratory comparison.
Geochim. Cosnochim. Acta 42:59-62, 1978.
-------
129
CHAPTER 4
DEPOSITION, DISTRIBUTION, AND CLEARANCE OF
OZONE-DERIVED OXYGEN IN THE RESPIRATORY SYSTEM
-------
130
In order to provide experimental measurements of the dose of ozone
to the respiratory systeri, the retention and clearance of ozone-derived
oxygen in the respiratory system has been examined using oxygen-18 as a
tracer for inhaled ozone. Extrapolation of health effects from labor-
atory animals to humans begins with a prediction of the uptake of ozone
in the human respiratory system based on mathematical dosimetry models
which describe the transport and absorption of reactive gases in the
lung (1). This approach requires an understanding of the origin of
species differences in the dose of ozone to tissue and an understanding
of the factors controlling the distribution of ozone-derived oxygen in
the respiratory system. Correlation of systematic differences in airway
structure and ventilation patterns (2) with quantitative differences in
local tissue dose in different mammalian species will provide a more
complete understanding of the factors controlling transport and removal
of ozone" in the respiratory system. The first two studies discussed in
this chapter were designed to assess i) the feasibility of utilizing
1(0] to determine species differences in deposition and ii) the ability
to resolve depositional patterns within the respiratory system.
The primary events of injury to the lung must involve interaction
of oxidants with biological molecules. Following the initial insult,
molecules altered by ozonolysis may fail to fulfill their biological
function (3). Products of ozonolysis may initiate autoxidation directly
or may mediate cell damage as chemotactic agents for alveolar
macrophages and polymorphic neutrophile (4-8), both events resulting in
changns in cellular function, and possibly cell death. Mechanisms of
protection against oxidant injury (9,10) and of extrapulmonary effects
(11) have only recently been considered. The dependence of dose on
-------
131
exposure tiae at a single concentration has been examined. The
distribution of czone-derived oxygen in proteins, lipids, and organic
solutes and the clearance of ozone-derived oxygen from the lung also
have been determined.
Detection and Quantitation of the Oxygen-18 Label
Stable isotopes are ubiquitous in nature. Approximately 0.2% of
oxygen in biological material is oxygen-18. Detection and quantitation
of tracer-derived oxygen in tissue therefore requires precise knowledge
of the isotopic composition of i) the initial tissue-oxygen pool and ii)
the tissue-oxygen pool after incorporation of the tracer. An isotopic
4'
mass balance of oxygen in tissue from an animal exposed to labeled ozone
yields an expression which can be used to determine the dose of ozone-
derived oxygen to tissue. For cfzone labeled with oxygen-18, the mole
fraction of ozone-derived oxygen in the tissue oxygen pool is
,E - I§FTI,NA
"TI "FTR ~ 1$FTI,NA
where »»F is the fractional abundance of oxygen-18 and n is the quantity
of oxygen ( e . g . , micromoles 0) in the i-th oxygen pool; the subscripts
designate the oxygen pools {tissue, natural abundance (TI,NA); tissue,
enriched (TI,E); and tracer (TR)}. The numerator in the right-hand side
of equation 1 is the enrichment of oxygen-18 in the tissue-oxygen pool,
the amount in excess of the natural background. The denominator in the
right-hand side of equation 1 quantifies the potential of the tracer to
produce an enrichment in the tissue. This mathematical formalism can be
used in studies employing ozone which is less than 100% oxygen-18.
In this work, "dose" will refer to the total amount of ozonederived
-------
132
oxygen in a tissue sample, whereas "relative enrichment" refers to the
mole fraction of ozone-derived oxygen in the tissue-oxygen pool. Dose
is determined from relative enrichment and the size of the tissueoxygen
pool
dose * (relative enrichment) x (oxygen pool) (4-2)
where oxygen pool refers to the amount of atomic oxygen (moles) in the
tissue. Utilization of oxygen-18 as a physiological tracer (in fact,
utilization of any stable isotope which is normally present in
biological material as a physiological tracer) requires both elemental
and isotopic analyses to quantify dose.
A fundamental limitation on detection and quantitation of an
isotopic tracer is imposed by variations in the natural abundance of
oxygen-18, which include both variations in individual animals with time
and variations among animals in the same population. These variations
appear as unavoidable "noise" on the "stable isotopic signal" in any
biological system. The precision of the measurement of oxygen-18
derived from labeled material in a physiological sample is a function of
i) the precision of the mass spectrometric measurement, ii) isotopic
variations introduced by sample preparation (including conversion of
oxygen to carbon dioxide), and iii) naturally occurring variations in
the abundance of cxygen-18 in the sample material. Mathematically, this
can be expressed as a sum of variances
V " V * V * V <«-3>
where Sj2 is the total analytical variance, SMI is the variance
assoc_ated with the mass spectrometric measurement, SpJ is the variance
-------
133
arising from sample preparation, and s^1 expresses natural variations in
oxygen isotopic abundance as a variance. The level of isotopic
enrichment also affects the precision (enrichments just above natural
abundance can be measured more precisely than large enrichments), but
the exact relationship between enrichment and analytical precision has
not been determined.
Ideally, sample preparation and mass spectrcmetric measurement
procedures should contribute insignificantly to uncertainty in the
isotopic analysis. In practice, this condition is rarely achieved.
Contributions to the variance in *»F from these sources are summarized
in Table 4-1. Each entry represents the variance obtained from five
replicate analyses. The contribution of the mass spectroinetric
measurement (s*) was determined from analysis of identical samples of
carbon dioxide. The variance associated with sample preparation (Sp1)
has two sources: preparation of carbon dioxide from organically bound
oxygen and sampling of an inhomogeneous powder (tissue homogenate). The
first contribution was determined by analysis of carbon dioxide prepared
from crystalline glycine and the second contribution was determined from
analysis of carbon dioxide prepared from samples of a single lung homog-
enate. Variations in the abundance of oxygen-18 between animals (SNI)
were examined by analysis of lung homogenates from different animals.
Temporal variations of oxygen-18 were not considered here. The prepara-
tion of carbon dioxide from organic material is the greatest single
source of uncertainty in the isotopic analysis, accounting for 36% of
the analytical variance. Tissue sampling and variations among animals
account for 34% and 29% of the analytical variance, respecively. The
-------
134
Table 4-1. Contributions to the Variance in l'F
source
a
mass spectrometer
sample preparation
tissue sampling0
d
natural abundance
sj Is1
x 10l» x 101§
6 6
157 e 163
14£e 309
126 e 435
detection limit
ppm
0.10
0.50
0.69
0.82
^replicate analyses of carbon dioxide.
^replicate analyses of glycine.
creplicate analyses of a single lung horaogenate.
replicate analyses of lung homogenates from different animals.
Individual contributions to the variance were calculated sequentially
froa the total variance (Is3).
95% confidence limit for relative enrichment, where sj B 4.35 x
10-» for »»FTT „. and »»F_, r, 1§F-~ «= 0.99, and N - 5.
TI,HA il,t IK
-------
135
mass spectrometric measurement contributes only 1% of the observed
variance. Detection limits corresponding to the cumulative variance at
each step were calculated. A detection limit of < 1 ppm (that is, one
atom of excess oxygen-18 per 10' atoms of oxygen in the tissue-oxygen
pool) is expected if no sources of uncertainty other than those
considered are present. However, differences in physiological states of
individual animals can cause differences in the retention of label in
tissue. Variations in the abundance of oxygen-18 observed in an
apparently homogeneous test population may be significantly greater than
predicted. The detection limit for excess oxygen-18 in a given test
population will be effected by experimental conditions and must be
evaluated in each case.
Deposition in the Respiratory System
Comparative Aspects of Ozone Uptake. This study comprises a
comparison of the uptake of ozone-derived oxygen in the respiratory
system of the rabbit (6.5 kg), the rat (380 g), and the mouse (21.5 g).
Tissue samples included head (nasopharynx), trachea, lungs, and blood.
In two separate exposures, three animals (one of each species) were
placed in an exposure chamber containing 1 ppm of 1>0a for 60 minutes.
Unexposed animals (two of each species) were used to establish the
natural abundance of oxygen-18 in each sample.
Results of isotopic analysis are shown in Table 4-2. Relative
enrichments were calculated using equation 1, where ^F-pj j^ was the
abundance of oxygen-18 in tissue from unexpoSed animals, *'FTI £ was the
abundance of oxygen-18 in tissue from exposed animals, and "FTR* 0.99.
Dose was determined from the relative enrichment and an estimate of the
size of the oxygen pool in each tissue, shown in Table 4-3.
-------
136
Table 4-2. Deposition of Ozone-Derived Oxygen in the Respiratory System
tissue
species
head
mouse
rat
rabbit
xlO>
2
2
2
.0286
.0302
.0330
'•FTI.E
xlO1
2
2
2
.0341
.0339
.0334
enrichment
xlO«
5
3
0
.5
.7
.3
dose
nanomoles
8.0 ± 0.
36 ± 6
0
8
trachea
lung
mouse
rat
rabbit
mouse
rat
rabbit
2
2
2
2
2
2
.0302
.0322
.0296
.0326
.0323
.0335
2
2
2
2
2
2
.0463
.0489
.0373
.0520
.0454
.0438
16
19
7
19
13
10
.4
.9
.7
.6
.2
.4
6.2 ± 1.
Ill ± 32
63 ± 34
8.6 ± 0.
51.7 ± 2.
286 i 22
7
8
9
blood
mouse
rat
rabbit
2
2
2
.0243
.0235
.0220
2
2
2
.0245
.0237
.0225
0
0
0
.2
.2
.5
0
0
0
ax, SE = 0.0003 x 10- J, and N » 2.
enrichments were calculated using the following equation 1, where
0.99,
j ,JA
was the fractional abundance of oxygen-18 in the
unexposed animals, and the l'FTj E was the fractional abundance of
coxygen-18 in the exposed animals, SE • 0.7 x 10-*.
dose is total tissue uptake, uncertainty is propagated SE.
isotopic enrichment was not statistically significant (p < 0.05).
-------
137
Table 4-3. Oxygen Pools in nasopharynx, Trachea, Lungs, and Blood
tissue
species
weight
g
% oxygen
wt/wt
oxygen pool
millimoies
head
mouse
' rat
rabbit
trachea
mouse
rat
rabbit
lung
mouse
rat
rabbit
blood
mouse
rat
0.126 ± 0.012
0.90 ± 0.15
5.89 ± 0.10
0.003 ± 0.001
0.045 ± 0.013'
0.673 ± 0.029"
0.032 ± 0.003
0.289 ± 0.018
2.03 ± 0.14
16.1 ± 0.5
17.3 ± 0.9
19.5 ± 1.3
20.5 ± 0.9
19.8 ± 0.4
19.6 - 0.6
21.9 1 0.7
21.7 i 0.5
21.7 ± 0.7
19.3 * 0.1
19.4 _ 0.1
1.27 i 0.12
9.7 ± 1.7
72 ± 5
0.38 ± 0.13
5.6 ± 1.6
8.2 ± 4.4
0.44 ± 0.04
3.92 ± 0.24
27.5 1 2.1
-------
138
Species differences in the uptake of ozone have their origin in
differences in respiratory rate, lung size, total surface area, and
ventilation. Since these parameters are proportional to body mass (13),
isotopic enrichment and dose also should be a regular function of body
mass. When relative enrichment is considered, the mole fraction of
ozone-derived oxygen in the tissue-oxygen pool decreases with body mass.
The size of the tissue-oxygen pool in each sample increases with body
mass (Table 4-3), so that, with the exception of tissue from the naso-
pharynx, the total amount oi ozone-derived oxygen in each sample
increases with body mass (Table 4-2).
Measurements of the tracheal concentration of ozone in rats, guinea
pigs, rabbits (14), and dogs (15,16) indicate that approximately 50% of
inhaled dose is removed in the head. Nasopharyngeal deposition is
therefore related to minute volume and should be directly proportional
to body mass. In this study, however, an inverse trend was observed.
Measurement of isotopic enrichment in the nasopharynx was complicated by
the sampling procedure. Head samples included portions of extraneous
muscle and skeletal material which was more difficult to remove
reproducibly from the larger samples. Incorporation of small amounts of
label into an increasingly large oxygen pool composed of oxygen from
unexposed tissue (extraneous material) rssulted in tin apparent decrease
in the amount of ozone-derived oxygen retained in the nasopharynx.
Problems of dilution in larger animals have taen circumvented with the
use of different sampling techniques.
Distribution Within _the Respiratory System. The distribution of
ozone-derived oxygen in the respiratory system was determined in
rabbits. One animal was exposed to 1 ppiu 1'0, for 60 minutes and one
-------
139
animal was retained to establish natural abundance. Results of isotopic
analysis of airway epithelium from both animals are shown in Table 4-4.
The relative enrichment in each sample .was calculated using equation 1;
1§FTI NA was tne oxY?511"18 abundance in tissue from the unexposed
animal, l'FTI E *'as the oxygen-18 abundance in tissue from the exposed
animal, and 1'FT, = 0.99 (Table 4-4). The dose to various epithelial
regions was determined using equation 2, where the size of the oxygen
pool was determined from the elemental analysis and the weight of the
sample. Dose is expressed as picomoles of ozone-derived oxygen per
milligram of dry epithelial tissue (Table 4-4).
As expected, the dose is highest in areas of greatest turbulence,
the nasal turbinates and the trachea immediately below the laryngeal jet
<>
(17), and in the conducting airways of the lung (refer to Figure 4-1 for
proper anatomical orientation). A small, but statistically significant
(p > 0.05) isotopic enrichment is observed in the peripheral parenchyma
of the lung. Based on the preceding discussion of variance, the detec-
tion limit is 20 picomoles of ozone-derived oxygen per milligram of
epithelium. Although the dose is greatest in the. nasopharynx and the
conducting airways of the lung, a small amount of ozone does reach the
centriacinar region. Since oxidant injury to respiratory tissue is
responsible for pulmonary toxicity, the ability to measure the dose to
peripheral regions of the lung is necessary in establishing a valid
dose-response relationship for inhaled pollutants. However, the finding
that the dose to the centriacinar region of the lung is extremely small,
even with a high atmospheric concentration of ozone, brings into ques-
tion the ability of ozonolysis alone to account for injury to tissue.
It is therefore likely that other mechanisms of tissue damage occur.
-------
140
Table 4-4. Uptake of Ozone-Derived Oxygen in the Respiratory System
tissue
.
10s
enrichment3 dose
x 10* piccmole/mg
nasal turbinates
nasal septum
nasal pharynx
olfactory mucosa
trachea 1
2
3
4
5
lung 7mm airway
lung parenchyma
2.0402
2.0377
2.0311
2.0344
2.0301
2.0301
2.0301
2.0301
2.0301
2.0262
2.0312
2.1359
2.1179
2.1124
2.0651
2.3518
2.2251
2.1762
2.1554
2.1369
2.1267
2.0365
96.9
81.2
82.3
31.1
325.6
197.4
149.9
126.9
158.7
101.8
5.3
1150
930
930
440
3660
2220
1690
1430
1790
1590
70
aenrichments were calculated using equation 1, where **FTR » 0.99,
14FTj JJA was the fractional abundance of oxygen-18 in the unexposed
animals, and the l*FTI E was the fractional abundance of oxygen-18 in
the exposed animals.
dose is the amount of ozone-derivei oxygen/milligram of dry epithelium.
-------
141
141
Figure 4-1. Distribution of ozone-derived oxygen in the
respiratory system. The dose in a rabbit exposed to 1 ppta 1'0J for 60
minutes is expressed as picoraoles/milligram dry epithelial tissue (refer
to Table 4-2).
-------
142
-------
143
Finally, the demonstrated ability to resolve spatial distributions of
ozone-derived cicygen within the respiratory system by this technique
offers great promise in the validation and definition of mathematical
dosimetry models.
Deposition and Clearance in the Lung
Deposition in the Lung. To determine the effect of exposure time
on retained dose in the lungs, groups of five mice were exposed to 1 ppm
'•0, for times up to 60 minutes. A portion of the lungs from each
animal was retained for analysis. The rema.-'.ning tissue from the five
animals in an exposure group was pooled and chemically fractionated into
an acid precipitable fraction (macromolecules), chloroform extractable
fraction (lipids), and a water:methanol soluble fraction (solutes). The
natural abundance of oxygen-18 in each of these samples was established
using five unexposed animals.
Results of isotopic analysis are shown in Table 4-5. All variables
were analyzed within a linear regression framework. The dependent
variable was »»F and the independent variables were exposure time, the
actual concentration of ozone, and their interaction. There was no
indication that body weight (21.2 ± 0.7 g), lung wet weight (144 ± B
mg), or lung dry weight (30 ± 2 mg) varied significantly (p < 0.05) by
exposure group. Although there was some evidence of lack of fit (p =
0.01), this appears to be due primarily to the anomalously high average
isotopic composition of the group at the 30-minute exposure. Polynomial
fits did not suggest the existence of anything but a linear trend. The
least-squares fit for ozone-derived oxygen in whole lung vs exposure
time is shown in Figure 4-2 and those for each chemical fraction are
shown in Figure 4-3. The uncertainty indicated in Figure 4-3 is the 95%
-------
144
Table 4-5. Deposition of Ozone-Derived Oxygen in the Lungs
exposure
time
min
0
6
17
21
30
32
35 -
44
60
intercept
slope d
r'd
lung
homogenate3
2.033?
± 0.0007
2.0350
± 0.0001
2.0398
± 0.0015
2.0395
± 0.0014
2.0486
± 0.0014
2.0412
t 0.0019
- — --
2.0453
± 0.0017
2.0521
1 0.0024
2.0337
0.0003
0.663
i«F
acid
precipitate"
2.0323
± 0.0003
2.0336
± 0.0012
2.0352
± 0.0013
2.0352
± 0.0011
2.0444
* 0.0017
2.0346
± 0.0019
2.0363
± 0.0014
2.0449
± 0.0020
2.0466
± 0.0022
2.0315
0.0002
0.726
x 10'
chloroform
extract b
2.0353
± 0.0004
2.0351
± 0.0003
2.0388
± 0.0006
2.0448
1 0.0005
2.0569 c
± 0.0007
2.0469
± 0.0007
2.0473
± 0.0003
2.0493
± 0.0004
2.0579
± 0.0006
2.0342
0.0004
0.767
water :methanol
extract53
2.0442
± 0.0004
2.0435
± 0.0006
2.0546
± 0.0007
2.0476
i O.OOOS
2.0527
± 0.0006
2.0505
± 0.0006
2.0503
± 0.0008
2.0570
± 0.0010
2.0554
± 0.0009
2.0451
0.0002
0.669
x ± SE for five animals in a single exposure group.
x ± SE for triplicate analyses of pooled samples.
cexcluded form the linear regression.
values of linear regression
-------
145
145
Figure 4-2. Uptake of ozone-derived oxygen in the lung. A plot of
oxygen-18 abundance in lung tissue vs exposure time for mice exposed to
,*
1 ppn> '"Oj is shown. Each point is the mean and standard error of the
mean for analysis of tissue from five mice in an exposure group. The
regression line with 95% confidence limit for the regression estimate is
also indicated.
-------
146
i i i i i—i i i i i 'i—m—r~r
i i i I I i iiliii\i\ i\i
o
t0
O
ID
O
O
CM
O
CD
q
CM
in
6un|
o
cvj
ro
O
CM
-------
147
147
Figure 4-3. Molecular distribution of ozone-derived oxygen in the
lung. Plots of oxygen-18 abundance in the water:methanol extract, the
acid precipitate, and the chloroform extract vs exposure tine are shown.
Each point is the mean and standard error of the mean for three
replicate analyses of fraction obtained froa peeled lung tissue. The
regression lines and 95% confidence limits for the regression estimates
are indicated.
-------
1435
2.06
2.05
u.
CD
2.04
2.03
2.05
10
Q
* 2.04
u.
s
2.03
I ^ I ' I
water -methanol extract
acid precipitate
10 20 30 40 50 60
exposure time, min
-------
149
confidence limit for the entire population, indicating real variation
among individuals. Since pooled tissue samples were used to obtain the
chemical fractions, uncertainties indicated in Figure 4-3 represent
analytical variation.
The regression explained approximately 73% of the total variation
in the fractional abundance of oxygen-18 in whole lung, 73% in the acid
precipitate, 67% in the chloroform extract, and 77% in the methanolwater
extract. There was evidence of an interaction (p « 0.002) between dose
and exposure time in whole lung, but it was not significant when only
exposed animals were considered. In all cases, regression of 1§F on the
concentration-time product explained less of the total variability than
did exposure time alone. Within a narrow concentration range, exposure
time is the most important factor controlling uptake of ozone-derived
oxygen in the lung.
In the statistical analysis, it was assumed that the variance of
1
-------
150
burying their nose in fur, thus filtering inspired air, while others
regained active throughout the entire exposure.
The dose of ozone-derived oxygen to the lungs and to each fraction
was determined using regression estimates of "F as a function of
exposure time. Equation 1 was used to determine the relative enrichment
in each sample, where the abundance of oxygen-18 in samples from
unexposed animals was used to establish 1'FTj jj^, the regression
estimate of isotope abundance was used to determine 1§FTI E, and 1§FTR *
0.99. Dose was determined from relative enrichment and the size of the
oxygen pool. Estimates of the sizes of the oxygen pools are shown in
Table 4-6. Whole lung retained approximately 100 picomoles per minute
of ozone-derived oxygen. Of this amount, 63% was in the acid precip-
itate, 10% was in the chloroform extract, the 12% was in the methanol-
water extract, and 15% was lost.
These results are somewhat surprising in light of the suggestion
that membrar.e lipids are the primary targets for ozone (18,19). Most of
the retained label appears in the acid precipitate rather than in the
chloroform extract, suggesting that proteins rather than lipids are the
primary targets for attack by ozone. Appearance of the isotopic label
in proteins can result from termination reactions between radicals
produced by ozonolysis of unsaturated lipids and proteins; however,
direct attack on proteins is also a possible explanation. The un-
accounted portion could have been lost by volatilization of labile
molecules or acidification (small organic acids and bicarbonate) and by
exchange of oxygen with water (20). Loss of the isotopic label could
also have occurred during thi exposure and, depending on the products of
ozonolysis, resulted in selective depletion of label from one of the
-------
151
Table 4-6. Lung Oxygen Pool
sample dry weight % of total % oxygen oxygen pool
mg microraoles
whole lung 30 100 21.9 411
chloroform extract 3 10 12.9 24
acid precipitate 20 65 31.1 288
water:methanol extract 7 25 .26.8 117
abody weight * 20.97 ± 0.13 g, lung wet weight = 154 ± 2 mg.
x ± SE, N » 45.
-------
152
chemical fractions. A complete mass balance of inhaled ozone-derived
oxygen would help answer questions of tissue dose and molecular targets.
Further work in this area is therefor.2 needed.
Pulmonary toxicity is more closely related to the concentration of
ozone, as opposed to exposure time or the time-weighted-average concen-
tration (12). Future studies employing different concentrrtions of
ozone will indicate whether tissue-dcse is related to atmospv ^ric
concentration or whether saturation of the ability of the lung to
protect against injury caused by oxidants occurs at higher doses.
Clearance from the Lung. To examine the characteristics of the
clearance of ozone-derived oxygen from the respiratory system, 30 mire
were exposed to 1 ppm itO, for 45 minutes. Groups of five randomly
selected animals were sacrificed at 0, 3, 6, 24, and 48 hours post
exposure. The natural abundance of oxygen-18 was established using six
unexposed animals.
Results of isotopic analysis are shown in Table 4-7. A plot of the
natural logarithm of relative enrichment versus time post exposure is
shown in Figure 4-4. A biexponential decay of ozone-derived oxygen from
the lung is observed. The clearance equation, indicated by the curve in
Figure 4-4, was determined using a curve-stripping procedure (21)
R(t) - 0.83-exp(-0.31t) * 0.17-exp(-0.01t) (4-4)
where R(t) is the fractional retention of ozone-derived oxygen in the
lung and t is time in hours. The first phase accounts for the clearance
of 83% of the isotopic label present in the lung at the end of the
exposure (zero hours) and has a half-life of 2.2 hours. The second
phase accounts for the clearance of 17% of the isotopic label initially
-------
Table 4-7., Clearance of Ozone-Derived Oxygen in Mouse Lungs
153
time
post exposure
h
body weight3 lung wet weight3 ^Flung3
g mg x 103
natural abundance 24.3 ± 0.4
151 ± 7
2.0327 ± 0.0003
23.9':! 0.6
143 1 6
2.0459 ± 0.0030
3
6
24
48
25
25
25
24
.6 ±
.3 1
.9 ±
.7 ±
0
0
1
1
.7
.8
.2
.0
139 ±
139 ±
155 ±
150 ±
5
2 '
15
9
2
2
2
2
.0426
.0369
.0346
.0341
* 0
± 0
1 0
.0036
.0019
.0011
± 0.0005
x i SE for five animals in a single exposure group.
-------
154
154
Figure 4-4. Clearance of ozone-derived oxygen fron the lung. A
plot of the natural logarithm of relative enrichment vs time post
exposure for mice exposed to.l ppra ''O, for 45 minutes is shown. Each
point is the mean and standard error of the mean for analysis of tissue
from five mice. The clearance equation is also indicated in the figure.
-------
155
-------
156
present and has a half live of 70 hours.
The clearance of aerosols from the lung (22) has received much
attention because of the continued health risk posed by particulates
lodged in the centriacinar region. Lung clearance of diesel exhaust
particles is described by a three compartment model, where the three
phases have half-fives of 1 day, 6 days, and 80 days (23). Particles
deposited in the trachobronchial tree and the proximal bronchioles are
transported to the gastrointestinal tract by the mucocillary escalator
(phases 1 and 2, respectively), whereas particles reaching the alveolar
region are removed by endocytosis, adsorption, dissolution, and metab-
olism (phase 3) (24). In contrast, clearance of ozone-derived oxygen
from the lung involves catabolism of ozonolysis products and adsorption
r»
directly into the systemic and lymphatic circulation, with mucocillary
clearance and phagocytosis playing minor roles. Exchange of oxygen
between organic molecules and water (20) also results in removal of the
isotopic label from lung tissue and could be a significant factor in
lung clearance. These data indicate that ozone-derived oxygen is
rapidly cleared from the lung; however, evidence for a two-compartment
model is minimal. Therefore, at this time no assignment of clearance
mechanisms is possible. Unambiguous definition of the phase(s) of
clearance and the mechanism(s) for removal of ozone-derived oxygen from
the lung require a closer examination.
EXPERIMENTAL SECTION
Apparatus. Isotopically enriched ozone was prepared from 99*
oxygen-18 0, in the apparatus shown in Figure 4-5. The ozonizer was an
annular pyrex vessel with a sealed bottom and a cold finger for
condensing liquid oxygen and ozone. Cylindrical copper electrodes were
-------
157
157
Figure 4-5. Apparatus for preparation and purification of Ia0,.
Below the manifold, the apparatus consisted of an annular ozonizer and
two distillation traps for the-synthesis and purification of labeled
ozona. Above the manifold were a 22-L ballast flask for storage of N,
and a 2-L flask for storage of an "0,/H, mixture. See the text for
details of construction and procedure.
-------
158
Q
o
to
S
o
to
5
o 5
-8
-------
159
placed on the inner- and outer-external surfaces of the ozonizer (25).
The distillation traps were pyrex U-tubes. All valves were high-vacuum
valves with Teflon plugs and Kel-F 0-rings, with a maximum clearance of
4 mm (Kontes). The manifold was fitted with a pressure gauge which
measured absolute pressure from 0 to 760 torr (Matheson) and with a
thermal conductivity vacuum gauge (Varian) which operated in the range
from 10-3 to 2 torr (not shown in Figure 4-5). A guard trap filled with
copper turnings heated to 100*C was placed between the vacuum manifold
and the pumping system to prevent ozone from reacting with oil in the
mechanical pump.
Two exposure chambers and a syringe pump for metering isctopically
t!
labeled ozone into the chambers were constructed specifically for
exposure of mice to l*0,. A schematic diagram of the exposure system is
shown in Figure 4-6. A small exposure chamber, which held five mice,
was constructed to minimize internal volume (3 I.), thus minimizing
isotope consumption and expense. A larger exposure chamber (30 L) was
constructed to contain up'to 40 mice. Both chambers consisted of a
series of plenums separated by perforated stainless steel sheets, which
ensured laminar flow of gas over the animals (26). A dispersion element
in the upper plenum ensured a uniform distribution of ozone in the
laminar flow region. The syringe had a capacity of 500 mL (Precision
Sampling). It was constructed of borosilicate glass with a Teflon cap
and plunger. A stepping-motor-drive system displaced the plunger of the
syringe. A variable compression rate from 0.8 to 800 mL/min was
available. The transfer line from the syringe to the exposure chamber
was Teflon tubing.
-------
160
160
Figure 4-6. Exposure system. This system consisted of a 500-mL
stepping-inotor-driven syringe, ah exposure chamber, and an ozone
monitor. See the text for details of construction and procedure.
-------
161
o
/N
0
s!=*
UJ
IT
UJ
§
s
cc
o
-------
162
Preparation and Purification of l'Q,. To synthesize 140,, the
ozonizer was izsaersed in liquid nitrogen and labeled oxygen was admitted
to the system. A potential of up to 25 XV was applied to the
electrodes. The vapor pressure of oxygen at liquid nitrogen temperature
(165 torr) was sufficient to maintain a diffuse discharge at approxi-
mately 15 kv. Ozone that formed in the discharge region immediately
condensed on the walls of the ozonizer and filled the cold finger. At
completion of the reaction, the pressure inside the ozonizer was that of
ozone at liquid nitrogen temperature (5 X 10~J torr) and was
insufficient to maintain a diffuse discharge at full voltage. The
procedure was stopped when this occurred, and a nearly quantitative
yield of ozone was obtained. , Residual oxygen and other possible
contaminants were removed by two cryogenic distillations. The ozone wris
f
transferred to a clean 2-L pyrex flask, diluted with nitrogen from th-^
22-L flask at a concentration of 1% to a pressure of. 740 torr, and used
within 4 hours of preparation.
As ozone is both explosive and toxic, precautions to prevent injury
in case of explosion and to minimize the risk of inhalation were taken.
The danger of explosion is greatest when ozone is a liquid or a solid,
where rapid warming or cooling of the condensed phases could cause a
detonation. Rapid temperature changes should therefore be avoided. The
ozonizer and distillation traps were enclosed by an explosion shield, a
Plexiglas housing with a screen back. The explosion shield was designed
to direct any glass fragments resulting froa detonation of the ozone
away from the experimenter. Two controlled explosions were performed to
ensure that the explosion shield functioned adequately. A ventilation
system was added to remote ozone that might escape during preparation
-------
163
aid transfer.
Animal Exposures. Female CD mice, male Fisher 344 rats, and male
Hew Zealand white rabbits (Charles River) were used in these studies.
Animals were placed in an exposure chamber and allowed to adjust to
their surroundings for 10 minutes prior to beginning the exposure.
Filtered, dry air was supplied to the chamber at & flow rate of one
chamber volume per minute. The syringe-pump was started, and the
measurement of exposure time began when the desired ozone concentration
was established in the laminar flow region of the chamber. The
concentration of ozone inside the chamber was measured using a Bendix
Model 8002 ozone monitor (Bendix). The stepping-rate of the syringe
pump was adjusted to maintain an ozone concentration of 1 i 0.1 ppm "0,
in the laminar flow region. Animals were removed promptly at the end of
an exposure.
Tissue Sampling and Preparation. After removal from the exposure
chamber, animals were individually anesthetized with Halothane and
approximately 3 mL of blood were removed from a heart puncture. The
trachea and lungs were immediately excised. The lungs were homogenized
in distilled water using a Pclytron homogenizer (Brinkmann Instruments).
The head was skinned, muscle and skeleton not part of the nasopharyngeal
region were removed, and the remaining tissue was homogenized in
distilled water. Weights of all samples were obtained prior to homogen-
ization. The tissue homogenates were immediately frozen in liquid
nitrogen and lyophilized to dryness. The samples remained frozen during
lyophilization. Residual water was removed by placing the samples under
high vacuum for 24 hours, or until a pressure of < 10-3 torr was main-
tained when the samples were isolated from vacuum.
-------
164
Using the following procedures, lung tissue was separated into
fractions roughly corresponding to lipids, proteins, and organic solutes
(27). Lung homogenates from animals in each group were pooled and
resuspended in distilled water so that the final mixture was 80% (wt/wt)
water and 20% (wt/wt) tissue. Volumes of solvent in the following
procedure are given for one gran of tissue suspension. Ten drops of
concentrated hydrochloric acid and 10 mL of methanol were added to the
tissue suspension. The mixture was homogenized, 20 mL of chloroform was
then added, and the mixture was again homogenized. The precipitate was
removed by suction filtration through a fritted glass filtering funnel
(ASTM 40-60 C), resuspended in 30 mL of chloroform-methanol (2:1 v/v),
homogenized, and separated by suction filtration. This procedure was
repeated and the precipitated solids were lyophilized to dryness. The
three filtrates were combined, an amount of water equal to 25% of the
total volume of filtrate was added, and the solution was mixed
thoroughly and was centrifuged at 10,000 x g for 5 minutes to separate
the twc layers. The upper layer (methanol-water) was removed by
aspiration and the lower layer (chloroform) was extracted twice with 1/4
volume of oethanol-water (1:1 v/v). The three methanol-water fractions
were pooled. Solvents were removed by evaporation in a stream of dry
nitrogen. The acid precipitate contained proteins and other
mscromolecules, the chloroform extract contained primarily lipids, and
the methanol-water extract contained polar organic solutes.
Elemental and Isotopic Analysis. Both oxygen elemental and
isotopic analyses of each sample were obtained. Procedures for isotopic
analysis were described in Chapters 2 and 3. Working standards (tissue
identical to the samples) were calibrated as described in Chapter 2.
-------
165
REFERENCES
1. Miller, F. J. A mathematical model of transport and removal of
ozone in mammalian lungs. Ph.D. thesis, North Carolina State
University, Raleigh, North Carolina, 1977.
2. Supplement: Comparative Biology of the Lung. Am. Rev. Respir.
Dis. 128, 1983.
3. Janoff, A. Biochemical link between cigarette smoking and
pulmonary emphysema. J. Appl. Physiol. 55: 285-293, 1983.
4. Tate, R. M.; Repine, J. E. Neutrophiles and the adult respiratory
distress syndrome. Aoer. Rev. Respir. Dis. 128:552-559, 1983.
5. Pantone, J. C.; Ward, P. A. Role of oxygen-derived free radicals
and metabolism in leukocycte-dependent inflammatory reaction.
'Amer. J. Pathol. 107:397-418, 1982.
6. Weiss, S. J.; Lo Buglio, A. F. Phagocyte-generated oxygen
metabolites and cellular injury. Lab. Invest. 47:5-18, 1982.
t'
7. Diegelmann, R. F.; Cohen, I. K.; Kaplan, A. M. The role of
macrophages in wound repair: A review. Plastic and Reconstruc.
Surgery 68:107-113, 1931. '
8. Transwell, A. K.; Smith, B;' T. Human fetal lung type II
pneumocytes in monolayer cell culture: The influence of oxidant
stress, cortisol environment, and soluble fibroblast factor.
Pediatric Res. 13:1097-1100, 1979.
9. Chow, C. K.; Tappel, A. L. An enzymatic protective mechanism
against lipid peroxidation damage to lungs of ozone-exposed dogs.
Lipids 8:518-524, 1972.
10. Chow, C. K.; Tappel, A. L. Activities of pentose shunt and
gli'colytic enzymes in lungs of ozone-exposed rats. Arch. Environ.
Health 26:206-208, 1973.
11. Graham, J. A.; Menzel, D. B.; Miller, F. J.; Illing, J. W.;
Gardener, D. E. Influence of ozone on pentobarbitol-induced
sleeping time in mice, rats, and hamsters. Toxicol. Appl.
Pharmacol. 61:64-73, 1981.
12. Melton, C. E. Effects of long-term exposure to low levels of
ozone: A review. Aviat. Space Environ. Med. 53:105-111, 1982.
13. Weibel, E. R. Morphological basis of alveolar-capillary gas
exchange. Physiol. Rev. 53:419-495, 1973.
14. Miller, F. J.; McNeal, C. A.; Kirtz, J. M.; Gardener, D. E.;
Coffin, D. L.; Menzel, D. B. Nasopharyngeal removal of ozone in
rabbits and guinea pigs. Toxicology 14:273-281, 1979.
-------
166
15. Yokoyama, E.; Frank, R. Respiratory uptake of ozone in dogs.
Arch. Environ. Health 25:132-138, 1972.
16. Morrman, w. J.; Chmiel, J. J.; Stara, J. F.; Lewis, T. R.
Comparative decomposition of ozone in the nasopharynx of beagles.
Arch. Environ. Health 26:153-155, 1973.
17. Chan, T. L.; Schreck, R. M.; Lippmann, M. Effect of the
laryngeal jet on particle deposition in the human trachea and upper
bronchial airways. J. Aerosol. Sci. 11:447-459, 1980.
18. Goldstein, B. D.; Balchum, 0. J. Effects of ozone on lipid
peroxidation in the red blood cell. Proc. Soc. Exp. Med. 26:356-
358, 1967.
19. Menzel, D. B.; Roehm, J.; Lee, S. D. Vitamine E: The biological
.and environmental antioxidant. Agr. Food. Chen. 20:431-485, 1972.
20. Murphy, R. C<; Clay, K. L. Synthesis and back exchange of *'0
labeled amino acids for use as internal standards with mass
spectrometry. Biomed. Mass Spectiom. 6:309-314, 1979.
21. Jauquez, J. A. Compartmental Analysis in Biology and Medicine .
Elsevier Publishing Co., New York, pp. 103-104, 1972.
22. Morrow, P. E. Alveolar clearance of aerosols. Arch. Intern. Med.
131:101-108, 1973.
23. Lee, P. S.; Chan, T. C.; Hering, W. E. Long-term clearance of
inhaled diesel exhaust particles in rodents. J. Tox. Environ.
Health 12:801-813, 1983.
24. Green, G. M. Alveolobrcnchiolar transport mechanisms. Arch.
Intern. Med. 131:109-114, 1973.
25. Myers, B. F.; Bartle, E. R.; Erickson, P. R.; Meckstroth, E. A.
Laboratory generator for batch synthesis of pure ozone. Anal.
Chem. 39:415-416, 1967.
26. Laskin, S.; Drew, R. T. An inexpensive portable inhalation
chamber. Am. Ind. Ass. J. 31:6^45-646, 1970.
27. Folch, J., Lee, M., and Sloane-Stanley, G. H. A simple method for
the isolation and purification of total lipids-fron animal tissue.
J. Biol. Chen. 226:497-509, 1957.
-------
167
CHAPTER 5
DETECTION AND QUANTITATION OF AUTOXIDATION IN TISSUE:
A STUDY OF HEPATIC AUTOXIDATION INDUCED BY CARBON TE7RACHLORIDE
-------
168
Evaluation of autoxidation in biological systems has been based on
the measurement of oxidation products of olefinic fatty acids (1-4).
Damage is not limited to lipids, however. Generally not considered but
equally important in the assessment of cellular dysfunction is oxidation
of other cellular components, particularly oxidation of proteins. In
this study, ''Oj was used to determine incorporation of oxygen into
residual organic material during autoxidation in the hepatocytes in
vivo. Determination of isotopic enrichment provides a sensitive
measurement of autoxidation of all cellular components. Conditions
producing hepatic autoxidation were created by interperitoneal injection
of carbon tetrachloride. Metabolism of carbon tetrachloride by the
Bicrososnal cytochrome P-450-mediated monooxygenase system in the liver
produces CC1, and Cl radicals (5,6) which initiate autoxidation (5-7).
Correlation between autoxidation and cytochronie P-450 activity is
demonstrated by modulation of isotopic enrichment in liver tissue.
Lipid Peroxidation
Peroxidation of unsaturated fatty acids results in the formation of
aldehydes, kotones, ethers, esters, epoxides, carboxylic acids, and
hydrocarbons (8). The peroxidation of linolenic acid (18:3A*> iaf**),
for example, can result in the production of propane, 1,3-propion-
dialdehyde (cialonaldehyde), and 9,11-dodecadienic acid during autoxida-
tion of olefinic fatty acids (Figure 5-1).
Peroxidation of Linolenic Acid. . Initiation of lipid peroxidation
(reaction 1) occurs by abstraction of hydrogen from an allylic position
by a free radical (8). Intramolecular rearrangement of the organic
radical (reaction 2) yields a more s'.able conjugated system (9,10),
which reacts with molecular oxygen (reaction 3) to form a peroxy radical
-------
169
169
Figure 5-1. Peroxidation of linolenic acid. Peroxidation of
linolenic acid leads to the formation of hydrocarbons, conjugated
dienes, malonaldehyde, and fluorescent products during autoxidation.
-------
170
(I)
(2)
(3)
(4)
(5)
(6)
(7)
-------
171
(8,11,12). Only one of four possible iscmeric peroxy radicals is shown
in this example (1,8). Intramolecular cyclizaticn of the peroxy radical
(reaction 4) leads to the formation of a five-membered ring peroxide
(1). Propagation of autoxidation proceeds by abstraction of hydrogen
from other cellular components. Formation of an a,7~diperoxy radical
(reaction 5) is followed by formation and subsequent decomposition ot a
diperoxy radical (reaction 6), yielding propane, 1,3-propiondialdehyde
(malonaldehyde), and 9,11-dcdecadienic acid (1). Hydrocarbon production
also .results from decomposition of mono-hydropercxides where ethane and
pentane are the major products (4). Malonaldehyde readily reacts
(reaction 7) with primary amines to form a di-Schiff base adduct (3).
Less than 2% of the malonaldehyde produced during autoxidation of
cellular material remains unbound (1).
Measurement^ of Lipid Peroxidaticn in vivo. The presence of
malonaldehyde has been considered indicative of lipid peroxidation in
biological matrices. Addition of 2-thiobarbituric acid to a solution
containing malonaldehyde produces a water-soluble colored complex, which
is a colorimetric indicator of the concentration of malonaldehyde.
Analysis of lipid peroxidation in tissue requires liberation of bound
malonldehyde, usually by heating in an acidic medium (1). Alterna-
tively, production of conjugated dienes, inferred from an increased
ultra violet absorption, in tissue has been used as evidence of lipid
peroxidation. Absorption occurs between 220 and 250 na, with
-------
172
to fluoresce with an excitation maximum at 360 nm and an emission
maximum at 430-440 nm. Measurement of the fluorescent products in the
chloroform-methanol extract of tissue also has been used as an indica-
tion of lipid peroxidation of biological material (16-18). Most
recently, measurement of volatile hydrocarbons evolved in the breath has
been used as an assay for lipid peroxidation (4).
Determination of the products of peroxidation of unsaturated lipids
as a measure of autcxidation neglects oxidation of all other classes of
compounds. The most general assay for autoxidation would be measurement
of the incorporation of oxygen into organic material. Although the
formation of Schiff-base adducts results in the loss of 0,-derived
oxygen to water, other pathways, not shown on Figure 5-1, lead to the
formation of stable products which retain 0,-derived oxygen (8,19).
Retention of the 0,-derived oxygen in the organic residue is also
expected with oxidation of other cellular constituents.
Carbon Tetrachloride Hepatotoxicity
Hepatotoxicity associated with carbon tctrachloride poisoning is
characterized by the rapid intracellular accumulation of triglycerides.
This dysfunction arises from inhibition of incorporation of lipids into
very low density lipoprotein in plasma (5). Injury to hepatocytes
proceeds from swelling of the rough endoplasmic reticulum, with
decreases in microsossal enzyme activity and protein synthesis, to
mitochondrial swelling and eventual necrosis (6).
Controversy arises, however, as to the nature of the histochemical
lesion. Two distinct mechanisms have *-een proposed. One hypothesis
holds that CC1, and Cl radicals, produced during metabolism of CC1, by
the microsomal cytochrome P-450-mediated monooxygenase system, initiate
-------
173
autoxidative processes that lead to loss of structure and function of
macromolecules, primarily membrane proteins (6). Initial damage is
localized to the rough endoplasmic reticulum, the intracellular site of
microsomal monooxygenases, but spreads throughout the cell as
autoxidation progresses (6). An alternative hypothesis suggests that
irreversible binding of carbon trichloride and chlorine radicals to
cellular components, particularly microsomal and mitochondrial membrane
proteins, is the cause of histological and biochemical dysfunction (20).
In both views, hepatotox:city is related to metabolism of carbon
tetrachloride. The occurrence of lipid peroxidation, as indicated by
the production of conjugated dienes (6,20) and ethane (21,22) during
^!
exposure to carbon tetrachlcride, is undisputed. Autoxidation of other
cellular components has only been inferred.
Microsomal Cytocnrorne P->450. The hepatic microsomal cytcchrome
P-450-mediated monooxygenase system contains at least six distinct
isoenzymes with different, but somewhat overlapping, substrate
specificties (23). Levels of specific isoenzymes can be increased by
repeated administration in vivo of a variety of drugs and toxins which
are normally substrates for these enzymes (24). Induction by
phenobarbital affects four isoenzymatic forms and increases the rate of
carbon tetra- chloride metabolism (25). Phenobarbital induction
increases polysomal P-450 monocxygenase mKNA by 30-fold (26); however,
it is not associated with the Ah (aromatic hydrocarbon) locus (27).
Thus, activities of NADPH-dependent cytochrome c reductase, HADPH-
dependent cytochrome P-450 reductase, epoxide hydroxylase, and
glutathione S-transferase are generally unaffected (28).
A wide variety of compounds has been found to inhibit cytochrome
-------
174
P-450 activity (29). Complete inhibition of carbon tetrachloride
metabolism in vivo can be obtained by administration of piperonal
butoxide. Inhibition occurs by metabolic activation of piperonal
butoxide to generate an intermediate which has a high affinity for Fe»*
contained in the heme group of cytochrome P-450 (29-33). Metabolic
activation requires HADPH and 0,, and a carbanion-cytochrome P-450
complex, which exists in a pH-dependsnt equilibrium, is formed (29,32).
The carbanion-cytochrome P-450 complex is stable in vivo (30), but, as
indicated by the Type I difference spectrum, it does not involve
covalent binding or destruction of cytochrome P-450 (32,33).
Carbon-Tetrachloride-Induced Autoxidaticn. In this study,
oxygen-18 labeled 0, has been used as a tracer for molecular oxygen
incorporated into biomolecules in the liver during autoxidation.
Autoxidation was induced by metabolism of carbon tetrachloride by
microsomal cytochrome P-450-mediated monooxygenases.
Kale Fisher 344 rats were used in this study. Animals were divided
into five groups. A summary of treatment of each group appears in Table
5-1. Animals in Group I were untreated and used to establish the
natural abundance of oxygen-18. Animals in Groups II through V were
ventilated by a mechanical respirator containing a mixture of H, and
"0, (80:20) Cor 60 minutes. Animals in Group II represented an experi-
mental control, and received l*0, as well as I.P. injections of corn
oil, but did not receive CC1,. Animals in Groups III through V received
I.P. injections of a solution of CC1« in corn oil. Hepatic nicroscmal
mixed-function oxidase activity was normal in the animals in Group III,
induced in th* animals in Group IV, and inhibited in the animals in
Group V. A portion of the liver from each animal was retained. Another
-------
175
Table 5-1. Sunrnary of Treatments of Groups
pheno- piperonal
Group »«0, com oila CC14 barbitalc butoxide
II
III
IV
adose » 5 mL/kg body weight.
dose • 1 mL/kg body weight of carbon tetrachloride.
cdose • 20 ng/day of phenobarbital for 3 days prior to the exposure.
dose * 25 lag/kg body weight of piperonal butoxide 60 oinutes prior to
receiving carbon tetrachloride.
-------
176
portion of the liver was fractionated into an acid precipitate, a
chloroform extract, and a nethanol-water extract. Oxygen elemental and
isotopic analyses were performed on ill samples.
Results of isotopic analysis are shown in Table 5-2, and relative
enrichments are s.hown in Table 5-3. Relative enrichments were
calculated using equation 4-1, where I*FXI NA was tne avera9e isotopic
composition of Group I, 1§F _ was the average isotopic composition of
X .L f LJ
the experimental group, and **F__ • 0.99. Estimates of the sizes of the
IR
oxygen pools were derived from the weights of the sample and oxygen
elemental analyses (Table 5-4). The amount of 0,-derived oxygen in
whole tissue and in each fraction was calculated using equation 4-2 from
v!
the relative enrichment and the size of the oxygen pool in the sample.
Results of these calculations'are shown in Figure 5-2. Oxygen derived
from 0, is expressed as nanomoles per gram of liver.
Incorporation of 0,-derived oxygen in animals not receiving carbon
tetrachloride (Group II) arises, in part, from basal activity of hepatic
nonooxygenases and mixed-function oxidases normally involved in metab-
olism, detoxification (34), and phagocytic processes (35). Isotopic
enrichment in control animals may also reflect indirect incorporation of
0,-derived oxygen into organic molecules by exchange of oxygen (36) with
H,"0 produced by oxidative respiration. The background enrichment in
the livers from animals in Group II corresponds to an average incorpora-
tion of 650 nancooles of 0,-derived oxygen per gram of tissue. The acid
precipitate retained 18% of the label, the chloroform extract retained
11% of the label, the methanol-water extract retained 32% of the label,
and 39% of the label was lost in the chemical work-up.
-------
177
Table 5-2. Abundance of Oxygen-18 in Liver
Group N *«F x 10'
liver acid chloroform methanol-water
homogenate3 precipitate3 extract3 extract
I 5 2.0497 2.0416 2.0456 2.0432
± 0.0010 ± 0.0005 ± 0.0009 ± 0.0015
II 5 2.235 2.092 2.291 2.81
1 0.031 ± 0.002 ± 0.027 ± 0.16
III 6 2.404 2.176 3.07 3.34
± 0.049 ± 0.040 ± 0.15 ± 0.12
IV 3 2.618 2.261 4.10 3.60
* 0.098 ± 0.035 ± 0.29 * 0.12
V 2 2.24 2.093 2.40 2.780
1 0.04 ± 0.017 t 0.13 ± 0.075
ax ± SE for N animals.
-------
178
a
Table 5-3. Relative Enrichment in Liver
Group Relative Enrichment x 10*
I
II
III
IV
V
liver
homogenate
0
190 ± 30
360 * 50
580 ± 100
200 ± 40
acid
precipitate
0
50 ± 2
140 ± 40
220 t 40
50 ± 20
chioroiorm
extract
0
250 ± 30
1040 ± 150
2080 ± 300
360 4 130
water imethanol
extract
0
780 ± 160
1310 ± 120
1780 ± 370
750 ± 80
aReiative enrichments were calculated from average isotopic
compositions of the samples from each group, where the isotopic
composition of samples from Group I was used as natural abundance.
Tha uncertainties are propagated standard errors of the means.
-------
Table 5-4. Oxygen Pools in Liver
179
sample
weight
rag
% oxygen
oxygen pool
micromoles
liver
238 ± 18
23 * 1
3400 ± 300
acid
precipitate
168 ±23^
23 i 1
2400 ± 350
CHC1, extract3 2.2 ± 0.2
14 ± 1
19 ± 2
CHjOH-H,0
extract"
2.2 ± 0.3
28 ± 1
38 ± 5
analysis of 0.07 aliquot of total chloroform extract.
3analysis of 0.15 aliquot of total methanol-water extract.
-------
180
180
Figure 5-2. Incorporation of Oj-derived oxygen into liver. The
incorporation of 0,-derived oxygan during free-radical-initiated
autoxidation is shown. The amount of 0,-derived oxygen in whole liver,
and the distribution in the acid.precipitate, the chloroform extract,
and the methanol-water extract of liver are shown. These values are
expressed as nanomoles 0,-derived oxygen/gram liver.
-------
131
UJDJ6-S8|OUJOUDU '9I-U9BAXQ SS90X3
-------
182
A direct correlation between P-450 activity and incorporation of
02-derived oxygen in the liver by autoxidation was observed in Groups
III through V. Autoxidation resulted in an average incorporation of 580
nanomoles Oj-derived oxygen/gram liver in Group III in excess of
background levels (Group II). The molecular distribution of label
within liver was different from that in the control group, in that the
acid precipitate and chloroform extract retained 38% and 36% of the
label, whereas the methanol-water extract retained only 26% of the
label. A two-fold increase in the amount of 0,-derived oxygen was
observed in the livers of animals in which cytochroms P-450 had been
induced with phenobarbital. Autoxidation resulted in an average incor-
poration of 1330 nanomoles 02-derived oxygen/gram liver in Group IV in
excess of background levels (Group II). The distribution of the label
among these fractions was identical to that found in Group III. A
decrease in isotopic enrichment to levels found in Group II occurred in
the livers of animals in which cytochrome P-450 activity had been
inhibited with piperonal butoxide (Group V). All 0,-derived oxygen
incorporated into liver by autoxidation (identified by enrichment above
control levels) was recovered in the chemical fractions. Insignificant
amounts of label were lost during the chemical work-up.
Isotopic enrichments in the chloroform and methanol-water extracts
represent peroxidation products of unsaturated lipids. Upon decomposi-
tion of lipid peroxides, Oj-derived oxygen can be retained in the major
portion of the lipid, thus appearing in the chloroform extract, or can
be retained in a small organic fragment (8) which, because of the polar
oxygen-containing functional group, would appear in the methanol-water
extract.
-------
183
The extent of incorporation of label in the acid precipitate is
surprising. This suggests that macromolecules, proteins in particular,
are also a major target for autoxidation, an observation missed in
previous studies in which only peroxidation products of olefinic lipids
were considered. Incorporation of the isotopic label in the acid
precipitate could result from binding of carbon trichloro-alkcxy
radicals directly to macromolecules (20). In fact, species differences
in hepatotoxicity correlate most closely with the incorporation in
macromolecules of radioactivity from 14CC1« as opposed to lipid
peroxidation measured by absorbance at 243 run (20). Distinction between
these two mechanisms is of minor importance, in that incorporation of
02-derived oxygen in macromolecules by e-ther process most likely
results in loss of structure and function. Isotopic enrichment of the
tissue-oxygen pool is then a good indication of damage resulting from
autoxidation.
CONCLUSIONS
This is the first study in which the incorporation of oxygen into
tissue during autoxidation has been measured directly. This development
is significant for the following reasons: i) this method provides a
sensitive, quantitative assay for autoxidation, ii) these techniques are
applicable to studies of autoxidation both in vitro and in vivo, and
iii) determination of autoxidation is not limited to measurement of the
products of lipid peroxidation, for any class of compounds which retains
oxygen from l»0, can be identified by determination of isotopic
enrichment. Thus, the procedures discussed in this chapter outline a
general approach to the determination of autoxidation and provide a
unifying approach to understanding the relationship between the
-------
184
environmental concentration of a compound and the physicochemical
processes controlling transport, adsorption, and reactivity of the toxic
agent in vivo.
EXPERIMENTAL SECTION
Protocol. Fivu groups of male Fisher 344 rats (Charles River)
were used in this stady; see Table 5-1 for a summary of the experimental
protocol. Group I was untreated and used to determine natural
abundance. Group II was injected I.P. with corn oil at a dose of 5
mL/kg body weight. Groups III, IV, and V were ir.jected with 5 mL/kg
body weight of a corn oil solution of carbon tetrachloride (20%). Group
IV received I.P. injections of phenobarbitol at a dose of 20 mg/day for
3 days prior to the exposure. Group V received an I.P. injection of
f
piperonal butoxide at a dose of 25 mg/kg body weight 60 nir.utes prior to
receiving carbon tetrachloride.
Groups II through V were anesthetized with Nembutal at a dose of 50
mg/kg body weight. The trachea was surgically exposed and cannularized,
and the animal was connected to an artificial respirator (Harvard
Apparatus) containing a mixture of N, and 1§0» (80:20). The animal was
then injected I.P. with Anectine at a dose of 0.06 mi.. Once the
respiratory rate and heart rate had stabilized, the animal received an
I.P. injection of corn oil or the corn oil solution of carbon tetra-
chloride. While on the respirator, some animals received additional
Anectine (0.03 mL) as needed to maintain even and unstressed breathing.
All animals were removed from the respirator 60 minutes after receiving
carbon tetrachloride and allowed to expire.
Tissue Sampling and Preparation. Two one-gram liver samples were
taken. One sample of liver was homogenized in 3 mL of distilled water
-------
185
using a Polytron homogenizer (Brinkmann Instruments). The homogenate
was immediately frozen in liquid nitrogen and lyophilized to dryness.
The sample remained frozen during lyophilization. Residual water was
removed by placing the tissue homogenates under high vacuum overnight.
The other sample of liver was used for chemical separation. One
gram of liver t?as homogenized in 7 mL of chloroform-methanol-water
(2:4:1), with the addition of 2 drops of concentrated HC1. The
homogenate was centrifuged at 20,000 x g and 5"C for. 20 minutes. Two
clear layers separated by a pellicle were visible. The upper layer was
methanol-water and the lower layar chloroform. The pellicle was removed
with forceps and lyophilized to dryness, and the layers were separated.
Elemental and Isotopic Analysis. Samples of the chlorofors.
extract were obtained by evaporation of the solvent in a stream of dry
nitrogen. The residual material was redissolved in 1.00 mL chloroform,
70 ML of this solution (0.070 aliquot) was added to a silver cup, and
the solvent was removed by evaporation in a stream of dry nitrogen. A
similar procedure was used to obtain samples of the methanol-water
extract. Solvent was removed by evaporation in a stream of dry
nitrogen. The residual material was redissolved in 200 uL of water, 30
uL of this solution (0.150 aliquot) was added to a silver cup, and the
solvent was removed by evaporation in a stream of dry nitrogen. The
pellicle was weighed, powdered using a mortar and pestle, and
approximately 2 mg of the powder was placed in a silver cup.
Both oxygen-elemental and isotopic analyses of each sample were
were obtained. Procedures for isotopic analysis were described in
Chapters 2 and 3. Working standards (tissue identical to the samples)
were calibrated as described in Chapter 2.
-------
186
REFERENCES
1. Dahle, L. K.; Hill, E. G.; Holman, R. T. The thiobarbituric acid
reaction and the autoxidation of polyunsaturated fatty acid methyl
esters. Arch. Biochem. Biophys. 98:2:53-251, 1962.
2. Farmer, E. H.; Koch, H. P.; Sutton, D. A. The course of
autoxidation in polyisoprenes and allied compounds. Part V.
Observations on fish-oil acids. J. Chem. Soc. 1943, 122-125.
3. Malshet, V. G.; Tappe.1, A. L. Fluorescent products of lipid
peroxidation: I. Structural requirement for fluorescence in
conjugated Schiff bases. Lipids 8:194-198, 1973.
4. Dumelin,. E. E.; Tappel, A. L. Hydrocarbon gases produced during iji
vitro peroxidation of polyunsaturated fatty acids and decomposition
• of preformed hydroperoxides. Lipids 12:894-900, 3°77.
5. Recknagel, R. 0. Carbon tetrachloride hepatotoxicity. Pharmacol.
Rev. 19: 145-1 , 1967.
6. Recknagel, R. 0.; Glende, E. A. Carbon tetrachloride hepato-
• toxicity: An example of lethal cleavage. Crit. Rev. Toxicol.
2:263-297, 1973.
7. Di Luzzio, N. R. Antioxidants, lipid peroxidation, and chemical-
induced liver injury. Fed. Proc. 32:1875-1881, 1973.
8. Frankel, E. M. Hydroperoxides. Lipids and Their Oxidation.
Schultz, H. W.; Day, E. A.; Sinnhuber, R. 0. (eds), The Avi
Publishing Company, Inc., Westport, Connecticut, 1962, pp. 51-78.
9. Farmer, E. H.; Koch, H. P.; Sutton, D. A. The course of
autoxidation in polyisoprenes and allied compounds. Part VII.
Rearrangement of double bonds during autoxidation. J. Chem. Soc.
1943, 541-547.
10. Farmer, E. H.; Sutton, D. A. The course of autoxidation in
polyisoprenes and allied compounds. Part XI. Double bond movement
during autoxidation of a monoolefin. J. Chem. Soc. 1946, 10-13.
11. Farmer, E. H.; Sutton, D. A. The course of autoxidation in
polyisoprenes and allied compounds. Part IV. The isolation and
constitution of photochsmically formed methyl oleate peroxide. J.
Chem. Soc. 1943, 119-122.
12. Bolland, J. L.; Koch, H. P. The course of autoxidation in
polyisoprenes and allied compounds. Part IX. The primary thermal
oxidation products of ethyl linoleate. J. Chen. Soc. 1945, 445-
447.
13. Bidlack, W. R.; Tappel, A. L. Fluorescent products of
phospholipids during peroxidation. Lipids 8:203-207, 1973.
-------
187
14. Dillard, ('.. J.; Tappel, A. L. Fluorescent products from reaction
of peroxiciizing polyunsaturated fatty acids with phosphatidyl
ethanolamine and phenylalanine. Lipids 8:183-189/ 1973.
15. Reiss, U.; Tappel, A. L. Fluorescent product formation and changes
in structure of DKA reacted with peroxidizing aracnidonic acid.
Lipids 8:199-202, 1973.
16. Fletcher, B. L.; :.'appel, A. L. Fluorescent modification of serum
albumin by lipid peroxidation. Lipids 6:172-175, 1971.
17. Dillard, C. J.; Tappel, A. L. Fluorescent products of lipid
peroxidation of mitochondria and microsomes. Lipids 6:715-721,
1971.
18. Trembly, R.; Tappel, A. L.? Coniglio, J. G.; Grogan,Jr., W. M.;
. Rhamy, R. K. Fluorescent products and polyunsaturated fatty acids
of human testes. Lipids 10:591-596, 1975.
19. Keenen, M. Secondary degeneration products. Lipids and Their
Oxidation. Schultz, H. W.; Day, E. A.; Sinnhuber, R. 0. (eds), The
Avi Publishing Company, Inc., Westport, Connecticut, 1962, pp.
51-78. *!
20. Diaz Gomez, M. I.; de Castro, C. R.; D'Acosta, de Fenos, de
Ferreyra, E. C.; Castro, J. A. Species differences in carbon
tetrachloride-induced hepatotqxicity: The role of CC14 activation
and of lipid peroxidation. Tbxicol. Appl. Pharm,. 34:102-114, 1975.
21. Roster, V.; Albrecht, D.; Kappos, H. Evidence for carbon
tetrachloride- and ethanol-induced lipid peroxidation in vivo
demonstrated by ethane production in mice and rats. Toxicol. Appl.
Pharmacol. 41:639-648, 1977..
22. Burk, R. F.; Lane, J. M. Ethane production and liver necrosis in
rats after administration of drugs and other chemicals. Toxicol.
Appl. Pharmacol. 50:467-478, 1979.
23. Lu, A. Y. H.; West, S. B. Multiplicity of mammalian microsomal
cytochrcmes P-450. Pharmicol. Rev. 31:277-295, 1980.
24. Herbert, D. N. Multiple forms of inducible drug-metabolizing
enzymes: A reasonable mechanism by which any organism car cope
with adversity. Mol. Cell. Biochem. 27:27-46, 1979.
25. Walz, F. G.; Vlas-jk, G. P.; Omiecinski, C. J.; Bresnick, E.;
Thomas, P. E.; Ryan, D. E.; Levin, W. Multiple, immunoidentical
forms of phenobarbitol-induced rat liver cytochrome P-450 are
encoded by different mRNAs. J. Biol. Chem. 257:4023-4026, 1982.
26. Adesnik, M.; Bar-Nun, S.; Maschio, F.; Zunich, M.; Lippman, A.;
Bard, E. Mechanism of induction of cytochrome P-450 by
phenobarbitol. J. Biol. Chem. 256:10340-10345, 1981.
-------
188
27. Pelkonen, 0.; Herbert, D. W. Metabolism of polycyclic aromatic
hydrocarbons: Etiologic role in carcinogenesis, Pharmicol. Rev.
34:189-222, 1982.
28. Herbert, D.-W.; Jensen, N. M. The Ah locus: Genetic regulation of
the metabolism of carcinogens, drugs, and other environmental
chemicals by cytochrome P-450-mediated monooxygenases. Crit.
Rev. Biochem. 6:401-437, 1979.
29. Testa, B.; Jenner, P. Inhibitors of cytochrome P-450s and their
mechanisms of action. Drug Metab. Rev. 12:1-117,1981.
30. Ullrich, V.; Schnabel, K. H. Formation and binding of carbanions
by cytochrome P-450 of liver microsomes. Drug Metab. Dispos.
1:176-183, 1973.
31. . Hodgson, E.; Philpot, R. M. Interaction of methylenedioxy-phenyl
(1,3-benzodioxole) compounds with enzymes and their effects on
mammals. Erug Metab. Rev. 3:231-301, 1974.
32. Franklin, M. R. Inhibitors of mixed function oxidases by
substrates forming reduced cytochrome P-450 metabolic-intermediate
complexes. Phanaacol. Ther. A 2:227-245, 1977.
33. Ullrich, V. The mechanism of cytochrome P-450 action. Microsomes
and Drug Oxidation . Ullrich, V.; Roots, I.; Hildebrandt, A.;
Estabrook, R. W.; Coony, A. H., (eds.), Pergaman Press, Oxford,
1977, pp. 192-201.
34. Coon, M. J. Oxygen activation in the metabolism of lipids, drugs,
and carcinogens. Nutrit. Rev. 36:319-328, 1978.
35. Johnson, R. B. Oxygen metabolism and the microbicidal activity of
oacrophages. Fed. Proc. 37:27-59-27-64, 1978.
36. Murphy, R. C.; Clay, K. L. Synthesis and back exchange of 100
labeled amino acids, for use as internal standards with mass
spectrometry. Biomed. Mass Spectroia. 6:309-314, 1979.
------- |