Environmental occurrence, abundance, and potential toxicity of polychlorinated biphenyl congeners: considerations for a congener-specific analysis.

Polychlorinated biphenyls (PCBs) as environmental contaminants often cannot be adequately described by reference to Aroclors or to total PCBs. Although there are 209 possible PCB configurations (congeners), perhaps half that number account for nearly all of the environmental contamination attributable to PCBs. Still fewer congeners are both prevalent and either demonstrably or potentially toxic. If potential toxicity, environmental prevalence, and relative abundance in animal tissues are used as criteria, the number of environmentally threatening PCB congeners reduces to about thirty-six. Twenty-five of these account for 50 to 75% of total PCBs in tissue samples of fish, invertebrates, birds, and mammals. A few PCB congeners that are sterically similar to 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD) are directly toxic. Other PCB congeners, as well as those that are directly toxic, may also be involved in toxicity indirectly by stimulating the production of (inducing) bioactivating enzyme systems. The most consequential of these have the ability to induce aryl hydrocarbon metabolizing mixed-function oxidases (MFOs). A result can be an increased capacity for bioactivation of otherwise nontoxic foreign compounds such as certain polynuclear aromatic hydrocarbons (PAH) to cytotoxic or genotoxic metabolites. The effectiveness of specific PCB congeners as inducers of different types of cytochrome P-450-dependent MFO systems is determined by their stereochemistry. Although MFO induction is not a proximate cause, it is a strong correlate of certain kinds of toxicities. Structural patterns can thus be used to discriminate among PCB congeners on the basis of toxic potential, if not entirely on toxicity per se. Congeners that demonstrate 3-methylcholanthrene-type (3-MC-type) and mixed-type MFO induction have the greatest toxic potential. These congeners most closely resemble 2,3,7,8-TCDD in their structures and in their toxic effects. The larger group of phenobarbital-type (PB-type) inducers have considerably less potential for contributing to toxic effects. Weak inducers and noninducing congeners have the least potential for toxicity. Using the rationale described in this paper, we assigned the most evironmentally threatening PCB congeners to four groups. Congeners assigned to Group 1 are considered most likely to contribute to adverse biological effects attributable to PCBs in an environmental sample. Group 1A contains the three most potent (pure 3-MC-type inducer) congeners, IUPAC numbers 77, 126, and 169. Six congeners, numbers 105, 118, 128, 138, 156, and 170, are assigned to Group 1B. These congeners are mixed-type inducers that have been reported frequently in environmental samples.(ABSTRACT TRUNCATED AT 400 WORDS)

Introduction vironmental contaminants. PCBs are widespread, persistent, and have the potential for harmful biological effects. Polychlorinated biphenyls (PCBs) are among the neu-Current practice in regulatory evaluation of PCBs in entral organic chemicals most commonly of concern as en-vironmental samples involves quantitation as total PCBs or as the total based on Aroclor equivalents (1)(2)(3). Quantitation of environmental samples as equivalents of tech-mixtures is altered (6). The correlation between the chromatogram of an Aroclor standard and that of PCB mixtures from an environmental sample is variable (7)(8)(9)(10)(11), and conventional quantitation methods permit considerable subjectivity on the part of the analyst (12)(13)(14). Additionally, reporting concentration data as total PCBs provides no information about the potential biological significance of the particular mixture of congeners in a sample and may be misleading. For example, it was noted that monochlorobiphenyls constituted as much as onethird of the total PCBs in samples of water, sediment, clams, and fish in an investigation involving contaminated sediments from the upper Hudson River. Such samples, if reported as total PCBs, would be considered potentially more toxic that warranted (15). The most bioaccumulating PCB congeners have five to seven chlorine atoms per molecule. These moderately chlorinated isomer groups (penta-, hexa-, and heptachlorobiphenyls) contain 112 of the 209 possible PCB configurations. They were synthesized in high proportions in many Aroclor formulations and are likely to be prevalent in environmental matrices (8,16). The moderately chlorinated isomer groups also contain most of the mixedfunction oxidase (MFO)-inducing congeners. Two of the three pure 3-MC-type inducers and all but two of the mixed-type inducers are penta-, hexa-, or heptachlorobiphenyls (17)(18)(19)(20)(21)(22)(23)(24). The more highly chlorinated congeners are generally less available to organisms both because they are more tightly bound with soils and sediments and because they usually are present in lower quantities in the environment. Congeners with less chlorination are more readily metabolized and eliminated and so do not tend to bioaccumulate as highly (17,18,25).
Chemical synthesis of all 209 PCB congeners, as well as separation and quantitation as discrete peaks from mixtures, although difficult, is possible using modern capillary column gas chromatographic methods (26,27). However, analysis for all of the congeners is not necessary for routine regulatory evaluations because many PCB congeners have never been reported in environmental samples, are not toxic, or have low bioavailability (15,16,28). An essential step in the development of a regulatory congener-specific analytical protocol is identification of the appropriate PCB congeners for inclusion in an analytical calibration standard. The purpose of this paper is to review current understanding of PCB environmental occurrence, potential toxicity, and prevalence as a rationale for determining the appropriateness of specific PCB congeners for inclusion in an analytical standard intended for application to environmental matrices.
Numbering of congeners in this paper follows the convention proposed by Ballschmiter and Zell (29) and later adopted by the International Union of Pure and Applied Chemists (IUPAC). IUPAC number, structure, and isomer group are given for each congener in Appendix A. The system of numbering and the nomenclature (ortho-, meta-, para-) convention for the biphenyl nucleus, along with the numbering of dioxin and dibenzofuran, are shown in Toxicity, Enzyme Induction, and Structural Patterns of PCBs lbxicity to Aquatic Biota PCBs are not acutely toxic to aquatic biota in the natural environment. The U.S. Environmental Protection Agency water quality criteria document for PCBs (3) noted that problems possibly could exist with the validity of acute toxicity tests because of the low solubility of PCBs in water and because the solubilities of PCBs are less than their acute toxicities. Experiments in which neonate daphnids and fathead minnow fry were exposed to saturated aqueous solutions of pure individual PCB congeners for 48 to 96 hr resulted in no mortalities that were attributable to PCB toxicity (30). The solutions were produced using a generator column and did not employ a carrier, such as acetone, a practice used in earlier work that may have resulted in unrealistically high acute toxicity results for PCBs (31)(32)(33).
These results are in agreement with the work of other researchers (34) who found that water solubility is the primary determinant of acute toxicity of a series of hydrocarbons and chlorohydrocarbons to Daphnia magna and Arternia salina. Abernethy et al. (34) also noted that since there is a trend for larger molecules to be less soluble in octanol, they also may be less soluble in the lipids of organisms. Such large volume molecules (e.g., the PCBs) may partition less easily into sites of toxic action within cells. Kinetic factors also influence acute toxicity, and larger molecules may take longer to establish concentrations necessary to produce toxic manifestations because of their lower diffusivity in water and lipid phases. Tbxic effects of aquatic environmental PCB contamination appear most likely to be sublethal and chronic. Physiological functions that are controlled by steroid hormones may be altered by exposure of organisms to PCBs (35)(36)(37). Growth, molting, and reproduction are primary functions that have been shown to be affected by exposure of aquatic organisms to PCBs in numerous laboratory investigations. The ability of organisms to eliminate foreign organic compounds or endogenous waste products also may be affected. Steroid biosynthesis and the degradation and biotransformation of foreign compounds are metabolic activities in both fish and higher vertebrates that are strongly influenced by terminal oxidase activities of the microsomal cytochrome P-450 systems (referred to also as mixed-function oxidase or MFO systems). Some, although not all, PCB congeners are MFO inducers in fish, mammals, and birds, and to a lesser extent in aquatic invertebrates.
Inhibited growth was reported (38) in juvenile salmon exposed to a tetra-(number 77) and two hexachlorobiphenyl congeners (numbers 153 and 155). Growth of minnows fed food contaminated with a technical PCB formulation, Clophen A50, was stimulated at high PCB doses (39). However, other investigators (40) reported significantly impaired growth and bone development in brook trout fry at 48 days after hatching. Although these and other referenced exposures were conducted at PCB concentrations substantially above those likely to be encountered in the environment, Aroclor 1254 residues in the brook trout from the lowest exposure concentrations were within the range found in some fish from feral populations. The authors (40) speculated that the observed adverse effect on bone development was a realistic indication of a potential toxic impact of Aroclor 1254 in the case of brook trout, and perhaps other fish species. In recent experiments, reproduction was significantly impaired in fathead minnows induced to spawn after exposure to PCB-contaminated natural sediments for 7 weeks. Exposure to pure PCB congeners selected for expected toxic potential did not result in any impairment of reproduction (T. M. Dillon, personal communication).
Molting of fiddler crabs has been reported to be severely impaired in laboratory exposures to Aroclor 1242 and to octachlorodibenzofuran (41). Juvenile crabs, Cancer nwtgister, exposed to suspensions of a harbor sediment containing PCBs, petroleum hydrocarbons, metals, and other contaminants, were deformed and died during molting (42). It has been observed that ovarian development and the molt cycle are closely linked in crustaceans (37). The molting agent, ecdysone, is a steroidal hormone related structurally to androgens and estrogens. A contaminant that causes a dysfunction in steroid biosynthesis and/or biotransformation and degradation could presumably result in impaired molting and reproductive success of some aquatic organisms.
Marine polychaetes, Nereis virens, that were exposed to the procarcinogenic aromatic hydrocarbon benzo[a]pyrene and to Aroclor 1254 were found to have significantly elevated MFO activity (36). Feral Nereis virens were then collected from clean reference and from oil-contaminated sites. The worms from the oil-contaminated sites had MFO activities approximately six times greater than those from the reference site and had only one-sixth the body weight of the reference site worms. Other work (43,44) correlated elevated MFO activities and reproductive failure in flatfish with PCB and oil ingestion, as well as with elevated PCB concentration in sediment at the collection site.
A major difficulty in interpreting the relevance of these findings to PCB contamination in the environment is a lack of clear causal relationships. Aquatic organisms in natural circumstances are almost always exposed to a variety of interacting contaminants. The interactions may be additive, synergistic, or antagonistic. In addition, the data regarding the toxicity of individual PCB congeners to aquatic biota are scant and often contradictory. However, the evidence for PCB congener-specific toxicity to higher vertebrates is much better developed, and it is now known that a large group of halogenated hydrocarbons that are approximately isosteric (including certain of the PCBs) elicit similar biochemical and toxic effects and appear to act by the same or similar mechanisms (45)(46)(47)(48)(49)(50). It is beyond the scope of this report to review the large body of literature reporting research in mammalian and avian toxicology of halogenated aromatic compounds, including the PCBs [for a comprehensive review, see Kimbrough (51)]. However, the stereospecificity required for certain types of MFO induction and that required for many toxic effects is the same, and although MFO induction is not a proximate cause, it is a strong correlate of some kinds of toxicities (52)(53)(54)(55). Contamination of fish and shellfish with PCBs is a concern to the health of human consumers as well as for the protection of avian and terrestrial wildlife. Therefore, the approach taken in this paper relies on the structural specificity of PCB congeners for microsomal enzyme induction as the best characterized indicator of potential toxicity.

Mixed-Function Oxidase Induction by PCBs
The group of microsomal cytochrome P-450-dependent enzyme systems that catalyze oxidative biotransformations of aromatic ring-containing compounds fall in the category of mixed-function oxidases. Different types of cytochrome P-450 systems are frequently characterized by reference to model chemicals that stimulate (induce) or inhibit the production of these enzymes. The MFOs that are induced by PCBs are characterized as being phenobarbital-type (PB-type), 3-methylcholanthrene-type (3-MC-type), or possessing catalyzing properties of both (mixed-type). The different types of inducers are specific for different substrates and have different regioselectivities for catalyzing the insertion of oxygen in the formation of reactive intermediates (56). Bromobenzenes, for ex-ample, were catalyzed by PB-induced rabbits to form 3,4-epoxides, whereas 3-MC-induction resulted in formation of 2,3-epoxides (57). The difference can be consequential in terms of toxicity. In the case of bromobenzenes, the 3,4-epoxide is hepatotoxic and the 2,3-epoxide is not. Dichlorobiphenyls were also metabolized differently by PBinduced animals and by animals induced with (3naphthoflavone (BNF), a chemical having inducing properties in common with 3-methylcholanthrene. With two ortho-chlorine substituents, the dichlorobiphenyls were metabolized predominantly by PB-inducible MFOs, and with no ortho-chlorines, BNF was the effective inducer of metabolizing enzymes (58).
The PB-inducible MFOs catalyze insertion of oxygen into conformationally nonhindered sites of globular or noncoplanar lipophilic molecules, facilitating their conjugation and removal (59). Ordinarily, reactions catalyzed by PB-inducible enzymes in Phase I detoxication go on to conjugation with endogenous substrates such as glutathione, glutamic acid, or sulfate in Phase II (60). Conjugation generally increases the water solubility of lipophilic molecules, making them more easily excreted. In some instances, as in the case of bromobenzene, PBinducible MFOs may have a role in the bioactivation of toxic intermediates. The 3-MC-inducible MFOs function to insert oxygen into conformationally hindered sites of planar molecules such as certain polynuclear aromatic hydrocarbons (PAHs) (59). The results of these reactions may be formation of carcinogenic, mutagenic, or teratogenic biotransformation products from otherwise nontoxic parent compounds. The best characterized example of this is the bioactivation of benzo[a]pyrene (BaP) to the ultimate carcinogens, the BaP-7,8-diol-9,10-expoxides (61). Conformational hindrance of the oxygenated molecule provides stability and tends to inhibit conjugation and detoxication that usually occurs readily in the case of PBinducible enzymic action. Since reactive epoxides are formed as transition products by both PB-and 3-MGinducible enzymes, both types have the potential for producing toxicity through bioactivation. In the case of bromobenzene, toxicity was enhanced in PB-induced animals (58,62,63); however, the potential for contributing to toxicity through bioactivation is considered to be much greater with the 3-MC-inducible enzymic reactions.
The MFOs of fish, and apparently of aquatic invertebrates, are qualitatively similar to the 3-MC-inducible MFOs of vertebrates (37,59,64,65). Phenobarbital-type induction has been reported in mummichog, rainbow trout, and carp in a few investigations; however, most studies have not demonstrated PB-type MFO activity in fish (66). Quantitatively, the detoxifying capability of fish appears to be about one-tenth that of mammals (37). The enzymes aryl hydrocarbon hydroxylase (AHH) and ethoxyresorufin 0-deethylase (EROD) are characteristic in fish as well as in 3-MC-induced mammals.

Molecular Structure and PCB Congener lbxicity
Potency and specificity for MFO induction (and correlatively, for potential toxicity) of individual PCB congeners can be directly related to how closely they approach the molecular spatial configuration and distribution of forces, i.e., are isosteres, of 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD). The dioxin, 2,3,7,8-TCDD (Fig. 2), is generally considered the most potent synthetic environmental toxicant and is regarded as a standard for comparison for other organic toxicants that are more or less isoteric, such as some of the PCBs (48,50,52,55). Dioxins are chlorinated aromatic molecules that form a planar volume in the form of a box or rectangle occupying about 3 by 10 A (67). The cytosolic receptor that binds 2,3,7,8-TCDD is a soluble protein produced by the Ah (aryl hydrocarbon) gene locus. Binding of 2,3,7,8-TCDD or its isosteres to the Ah receptor is facilitated by coplanarity of the phenyl rings within the 3 by 10 A dimensions. Other factors such as polarizability of the lateral chlorines may be important in determining the strength of binding to the receptor (19). Translocation of the inducer-receptor complex to the nuclear Ah locus is thought to initiate the synthesis of AHH, EROD, and related enzymes that may be involved in either biotransformation, conjugation and removal, or the bioactivation of certain planar lipophilic foreign compounds to toxic intermediates (46,47,49). The most toxicologically active PCB congeners are those having chlorine substitution at the para (4 and 4') and at least two meta (3,35, and 5') positions on the biphenyl nucleus, but no ortho (2,2,6, and 6') substitutions ( Fig. la,b) (24). Because the phenyl rings of a biphenyl nucleus are linked by a single carbon:carbon bond, the two rings have relatively unconstrained rotational freedom. Unlike dioxins (Fig. lc) or dibenzofurans (Fig. ld), the phenyl rings of a PCB are not rigidly bound in the same plane. X-ray crystallographic analyses indicate that the preferred conformation for all PCBs, including those without ortho-substituents, is noncoplanar (67). The proportion of molecules of a particular congener assuming a coplanar configuration becomes increasingly small as the energetic cost of confonning increases (67). Chlorines are very bulky atoms, and the substitution of a chlorine at certain positions on the biphenyl nucleus inflicts constraints on rotational freedom. The greatest effect is exerted by substitution of at least two opposing ortho-substituted chlorines on opposite rings (58,68). Increasing the number of chlorine atoms at the ortho points increases steric hindrance to rotation (Fig. 3).

Pure 3-Methylcholanthrene-type Inducers
The PCB congeners that have no ortho substitutions can assume a coplanar configuration (Fig. 4). Because potential toxicity is enhanced by coplanarity, this rule limits the number of PCB congeners that may be expected to be toxicologically most active to four. These congeners are numbers 77, 81, 126, and 169 (Appendix A). With the exception of number 81, the nonortho-coplanar congeners are potent inducers of AHH and EROD in in vitro rat hepatoma cell preparations. The in vitro induction of AHH and EROD is correlated with in vivo demonstrations of mammalian toxicity such as thymic atrophy and inhibition of body weight gain (52)(53)(54)(55)69).
Congeners 77, 126, and 169 are pure 3-MC-type inducers; congener 81 demonstrates mixed-type induction. Tanabe et al. (70) reported the presence of congeners 77, 126 and 169 in tissue samples from a wide range of organisms including marine mammals and humans. Residues of congeners 77, 126 and 169 were present in sufficiently high concentration to conclude that these nonorthocoplanar PCBs, and particularly number 126, pose a greater threat to humans and wildlife than does 2,3,7,8-TCDD itself.
The conclusion of Tanabe et al. (70) was made on the basis of a summation of toxic equivalents, i.e., the product of molar concentration and potency for AHH and EROD induction of chemicals relative to 2,3,7,8-TCDD in in vitro rat hepatoma cell cultures (52)(53)(54)(55). Whether this approach will prove valid remains to be seen. Some results of recent investigations (71,72) have indicated that the PCB mixture Aroclor 1254 is a dioxin antagonist in rat hepatoma H-4-II E cells, in vitro, and in C57BL/6J mice, in vivo, if administered at noneffective (below toxic threshold) doses. The mechanism is thought to be competitive inhibition in which the more abundant and sterically similar but less effective PCB congener molecules outcompete 2,3,7,8-TCDD for the same cytosolic receptor sites. The three nonortho-coplanar PCB congeners are close isosteres of 2,3,7,8-TCDD, and appear to act by the same mechanism. If competitive inhibition by less effective PCB congeners in a mixture reduces the enzymeinducing effectiveness and the toxicity of congeners such as numbers 77, 126, and 169, then summation of toxic equivalents cannot be a viable approach for estimating the toxic significance of these chemicals in the environment.
In reports previously cited (71,72), it was noted that exposure to nontoxic levels of PCBs in the environment may protect humans to some extent from concomitant exposure to much more toxic dioxins and dibenzofurans. Implications regarding other environmental effects are substantial: if this instance of competitive inhibition is a generally operant mechanism in biota, it could provide insight into the ability of most adult organisms, including fish and shellfish, to carry seemingly large body burdens of dioxins, furans, and PCBs without apparent effect.

Mixed-type Inducers
The second group of congeners having enzyme inducing potencies and potential toxicities of high concern are analogs of the four nonortho-coplanar congeners that are still relatively coplanar but have a single ortho-chloro substitution (Fig. 5). These are congeners 105, 114, 118, 123, 156, 157, 167, and 189. This group of congeners has demonstrated mixed PBand 3-MC-type inducing properties. Of these, congener 105 is often present in relatively high concentration in sediments, but analyses of aquatic organism tissues do not reflect this abundance (Table 1). Safe (55) noted that congeners 77 and 105 did not conform to the close relationship between in vitro and in vivo effects characteristic of most of the other nonortho-coplanar and mono-ortho coplanar PCBs and the most potent chlorinated dioxins and dibenzofurans. The presence of adjacent unsubstituted carbons in the biphenyl nucleus of congeners 77 and 105 is thought to facilitate metabolic degradation in mammalian tissues in vivo (55).
Besides the di-ortho coplanar congeners mentioned above, other congeners have been shown to be PB-type inducers (Table 1). Additionally, some congeners have not been demonstrated to be inducers but are classified as theoretical PB-type inducers according to structureactivity rules proposed by Parkinson and co-workers (21,22,24,73,74). The majority of PCBs apparently have no effect on mammalian systems. The 3-MC-type and mixedtype inducer PCB congeners evidently have the highest potential for toxic effect in mammals and birds and possibly in fish and invertebrates as well.

PCB Congeners of Highest Environmental Significance
Priority groupings for specific PCB congeners having highest potential environmental significance are made based upon three factors: a) potential for toxicity; b) frequency of occurrence in environmental samples; c) relative abundance in animal tissues.
Potential for toxicity is inferred by MFO induction. Pure 3-MC-type and mixed-type inducers are considered to be potentially more toxic than PB-type inducers. PB-type inducers are considered potentially more toxic than weak inducers and noninducers. [Classification of PCB congeners according to MFO-inducing capability is based on Safe et al. (18,24), Albro and McKinney (19), Goldstein (20), and Parkinson et al. (21)(22)(23).] Frequency of occurrence in environmental samples is determined from a PCB congener database (Joan U. Clarke, unpublished) developed from information reported in the scientific literature. Environmental samples refer to sediments, water, or organisms collected in the field, as opposed to samples resulting from laboratory exposures to PCBs. At present, 59 literature references included in the database report information on specific PCB congeners in environmental samples. Congeners are considered to occur frequently if reported by at least 20% Several sources were used for information on the relative abundance of specific congeners in animal tissues. A total of 41 single congen.ers and six mixtures of congeners that co-eluted (mixed peaks) were quantitated in fathead minnows exposed for 77 days to natural sediments contaminated with PCBs from an industrial discharge (75). Seventy-two congeners were quantitated in oligochaetes, carp, and ducks from the Detroit River, MI (76). Percent of total PCBs was reported for 43 congeners and 11 mixed peaks in samples of seston, shrimp, plaice, and porpoise from the Dutch Wadden Sea (77). Thirty-six congeners and two mixed peaks were quantitated in caddisfly larvae from the upper Hudson River, NY (25). All 209 congeners were included as instrument calibration standards for analysis of composited human milk samples from Michigan, and percent of total PCBs was reported for each congener (2?). Finally, percent of total PCBs was reported for 55 congeners in human adipose tissue from people living in southwestern Sweden (78). Table 1 presents relative abundances of specific congeners in tissues as percent of total PCBs reported or calculated from these sources. When congeners were expressed as concentrations rather than as percent of total PCBs, percents were calculated from total PCBs as the sum of the reported individual congener and mixed-peak concentrations. Congeners considered relatively abundant comprise at least 2% of total PCBs in one or more reported tissue analyses. Table 1 contains information for all 209 PCB congeners. Those congeners meeting the criteria for potential toxicity, frequency of occurrence, and relative abundance are footnoted in Table 1 and Appendix A and grouped according to priority in Table 2. Based on this evaluation, 36 PCB congeners were assigned to four priority groups. We consider these congeners most appropriate for use in the regulatory evaluation of environmental samples such as water, tissues of fish and other wildlife, and dredged sedi-       ments. The 36 congeners are selected based on environmental relevance, not on analytical considerations. In Table 2, the Group 1 or highest priority congeners, are those that are most likely to contribute to adverse biological effects due to their presence in environmental samples. Group 1 congeners fall into two classifications. Group 1A comprises the three pure 3-MC-type inducers, numbers 77, 126, and 169. Although these congeners have been reported rarely in environmental samples and only in the parts per trillion range, the very high individual toxicities of the three warrant special consideration. All three congeners have been identified as components of technical PCB formulations (79)(80)(81).
Group 1B congeners, numbers 105, 118, 128, 138, 156, and 170, are mixed-type inducers that have been reported frequently in environmental matrices and most are relatively abundant as well. Except for number 105, these congeners occur in nearly all of the samples included in Table 1 and individually represent as much as 16% of total PCBs reported in animal tissues. Collectively, Group 1 congeners make up 24 to 27% of total PCBs in the birds and mammals of Table 1, with a lesser representation (as little as 1.5%) in fish and invertebrates. Congener 105 is the mono-ortho-pentachlorobiphenyl analogue of number 77 and has a higher receptor protein binding affinity and nearly the same AHH and EROD induction potency as does number 77 (24). Because congener 105 has such high potential for toxicity it is included in Group 1B despite its reported lesser environmental occurrence than others in that group.
Group 2 consists of known and predicted PB-type MFO inducers that have numerous reported environmental occurrences, and most are relatively abundant in tissues. These include numbers 87, 99, 101, 153, 180, 183, and 194. Group 2 congeners occur in most of the samples listed in Table 1. Individually they comprise up to 22% of total PCBs in animal tissues. As a group, these congeners contribute 26 to 41% of total PCB in the bird and mammal samples and seven to 25% of total PCBs in fish and invertebrates.
Except for congeners 77 (a tetrachlorobiphenyl) and 194 (an octachlorobiphenyl), all of the congeners in Groups 1 and 2 are members of the penta-, hexa, and heptachlorobiphenyl isomer groups. The less chlorinated PCBs (those having one to four chlorines) are for the most part nontoxic. Many of these congeners may be taken up readily, but are also readily eliminated by organisms and are not bioaccumulated to a great extent. Most of the highly chlorinated congeners (those having seven to ten chlorines) occur in low concentrations in the environment, are tightly bound with soil and sediment, and tend to be less bioavailable. The 16 Group 1 and 2 congeners taken together account for as much as 66% of total PCBs in animal tissues. In particular, congeners 138 and 153 are major components of technical PCB formulations (24) and together account for over 20% of total PCBs in several tissue samples (almost 40% in porpoise).
Congeners assigned to Group 3 are weak or noninducers but occur frequently or represent at least 10% of the total PCBs in tissue samples. Group 3 congeners are most abundant in fish and invertebrate samples of Table   1, collectively contributing from 20 to 48% of the total PCBs. Bird and mammal tissues generally have lower Group 3 abundances, ranging from 13 to 19%. Group 3 congeners individually represent up to 20% of total PCBs in tissues. Group 3 spans the isomer groups from trithrough octachlorobiphenyl.
Group 4 consists of mixed-type inducers that have few reported environmental occurrences. These congeners have low relative abundance in tissue samples (Table 1). Group Table 2. Priority groups of PCB congeners of highest concern as environmental contaminants based on potential toxicity, frequency of occurrence and abundance, and suggested for inclusion in analysis of environmental materials for regulatory purposes. IUPAC no. Group  threatening of these to Groups 1 and 2. A calibration standard for congener-specific analysis of PCB-contaminated environmental materials would ideally include all of the compounds in Groups 1 and 2. The congeners in Group 3 are important in terms of environmental prevalence and relative abundance in animal tissues. Group 4 congeners may be of lesser significance in the environment but are toxicologically active.
An analytical instrument calibration standard for PCB congener-specific analysis by capillary column gas chromatography has been developed in Canada (82). The Canadian standard mixture, CLB-1, contains 51 congeners, including 13 of the 16 congeners in Groups 1 and 2 (Thble 2). Nine of the remaining 20 congeners in Groups 3 and 4 are also included in the Canadian standard. Not included are numbers 126 and 169, two of the three most toxic congeners that make up Group 1A. As yet, the United States has not adopted a selected PCB congener standard mixture. However, most of the congeners in the four priority groups of Table 2 are commercially available and a standardization cocktail could be prepared, but at a high cost. The cost factor coupled with quality assurance considerations suggests that a standard mixture similar to the Canadian CLB-1 should be developed. The specific PCB congeners identified here as being of primary concern as environmental contaminants were selected based on current knowledge. Congener-speciflc PCB analyses are not conducted routinely, and there is not a great wealth of data in the literature concerning concentrations of individual congeners in environmental samples. While this paper was in preparation and then in review, two articles appeared in the scientific literature differing in important respects but also addressing the subjects of environmental occurrence of PCB congeners and the desirability of congener-specific analysis (81,83). Both considered the necessity of using potential toxicity as a criterion for inclusion of PCB congeners in an analysis, and both used MFO induction or the coplanarity of rings as indications of toxic potential. Jones (83) worked from a data set that was largely different from that used here and concluded with a list containing 32 congeners.
With the exception of number 194, Jones included all of the congeners we assigned to priority Groups 1 and 2 (Table 2). Of the remaining 21 congeners in priority groups 3 and 4, Jones listed 10. Duinker, Schultz, and Petrick (81) considered the difficulties in separating mixed peaks as a criterion for inclusion of congeners, but also suggested that toxicity be used as a criterion for analysis of environmental samples. Sixteen of the 36 congeners we have prioritized were identified as being capable of unambiguous analysis by conventional means ( Table 2). The remaining 20 congeners present varying degrees of difficulty in analysis and some may not be analyzable by current methods. Somewhat different patterns of environmental prevalence or relative abundances of congeners in animal tissues may emerge as a result of additional studies. Furthermore, toxicity and bioavailability of PCB congeners are exceedingly complex issues. Potential for toxicity, particularly to nonmammalian species, is probably not fully described by classification of PCB congeners according to type of mammalian microsomal enzyme induction. Much additional research is needed to fully characterize the bioavailability and toxicity of specific congeners or mixtures of congeners in environmental matrices under controlled laboratory conditions, as well as under more variable field conditions. Summary This paper presents an analysis of the potential for adverse environmental effects of specific PCB congeners. Thirty-six congeners are considered most environmentally threatening based on their frequency of occurrence in environmental samples, relative abundance in animal tissues, and potential for toxicity. Ibxic potential is evidenced by type and specificity of mammalian microsomal mixed-function oxidase (MFO) induction. These congeners may be assigned to four priority groups. Group 1 includes the three pure 3-MC-type MFO inducing congeners, along with six mixed-type inducers that have been reported frequently in environmental samples and can be fairly abundant in tissues. The seven Group 2 congeners are PB-type inducers that also have numerous reported environmental occurrences and have high relative abundances, especially in avian and mammalian samples. Group 3 congeners are weak or noninducers that occur frequently or in relatively high concentrations, particularly in fish and invertebrate tissues. Group 4 includes ten mixed-typed inducers that have been reported infrequently in environmental samples and in relatively low concentrations. Group 4 congeners, though scarce in environmental matrices, are considered to be of possible concern because of their potential for toxicity.
Toxicologically relevant evaluations of PCBcontaminated environmental materials can be accomplished better by analyzing samples for specific congeners in the four priority groups as compared to analyzing as total PCBs or as Aroclor equivalents. A more meaningful assessment of a potentially unacceptable adverse ecological impact can be achieved by focusing only on those congeners that are prevalent in the environment, that are preferentially bioaccumulated, or are potentially toxic. This approach would be facilitated by the development of a standard PCB congener mixture that could be used as a reference standard during the analysis of environmental samples.