Environmental fate and biodegradability of benzene derivatives as studied in a model aquatic ecosystem.

A model aquatic ecosystem is devised for studying relatively volatile organic compounds and simulating direct discharge of chemical wastes into aquatic ecosystems. Six simple benzene derivatives (aniline, anisole, benzoic acid, chlorobenzene, nitrobenzene, and phthalic anhydride) and other important specialty chemicals: hexachlorobenzene, pentachlorophenol, 2,6-diethylaniline, and 3,5,6-trichloro-2-pyridinol were also chosen for study of environmental behavior and fate in the model aquatic ecosystem. Quantitative relationships of the intrinsic molecular properties of the environmental micropollutants with biological responses are established, e.g., water solubility, partition coefficient, pi constant, sigma constant, ecological magnification, biodegradability index, and comparative detoxication mechanisms, respectively. Water solubility, pi constant, and sigma constant are the most significant factors and control the biological responses of the food chain members. Water solubility and pi constant control the degree of bioaccumulation, and sigma constant limits the metabolism of the xenobiotics via microsomal detoxication enzymes. These highly significant correlations should be useful for predicting environmental fate of organic chemicals.

The discovery of ubiquitous environmental contamination by chlorinated pesticides, polychlorinated biphenyls, and phthalate ester plasticizers has focused attention on the distribution, fate, and possible toxic action of the thousands of synthetic organic chemicals produced commercially.t The accidental entry of these substances into water deserves particular study because of the capacity of many aquatic organisms to biomagnify or- ganic compounds 105to 107-fold, from parts per trillion concentrations in water to parts per million levels in their body lipids. This phenomenon is particularly destructive with pollutants such as DDT, DDE, dieldrin, and PCBs in the Great Lakes as biomagnified in lake trout and coho salmon.
Aromatic compounds make up about 46% of the total synthetic organic chemical production in the United States. The simple aromatics are of great importance as the building blocks of the chemical industry and the benzenoid nucleus is incorporated into thousands of specialty chemicals: pesticides, pharmaceuticals, dyestuffs, and detergents. We chose to study the behavior and fate of six simple benzene derivatives in a April 1975 model aquatic ecosystem with a six element food chain. The compounds studied and their approximate annual U.S. production (1973) are: phthalic anhydride 936 X 106 lb, chlorobenzene 530 X 106 lb, aniline 388 X 106 lb, nitrobenzene 220 X 106 lb, anisole 20 X 106 lb, and benzoic acid 15 X 106 lb (U.S. Tariff Commission). These particular compounds were selected for investigation because of their wide range of physicochemical parameters: reactivity (a constants), lipid/water partitioning (7r constants), and water solubility which could be used to develop quantitative relationships with biomagnification, biodegradability, and comparative detoxication mechanisms. In order to test the predictive properties of these parameters, we have also evaluated several simple benzene derivatives which are important specialty chemicals: hexachlorobenzene and pentachlorophenol which are fungicides and industrial waste pollutants; 2,6-diethylaniline, an intermediate in the production of alachlor herbicide [2-chloro-2',6'-diethyl-N-methoxymethyl) acetanilide], and 3,5,6-trichloro-2pyridinol, an intermediate in the production of chlorpyrifos insecticide, (O,O-diethyl 0-3,5, 6-trichloro-2-pyridyl phosphorothionate). The model aquatic ecosystem behavior of these specialty chemicals provides confirma-tion of the accuracy and applicability of generalizations which can be made from the environmental behavior of the simple benzene derivatives. Radiolabeled DDT and aldrin were also evaluated to provide comparisons of the accumulation and degradation of very water-insoluble organic compounds.

Radioactive Compounds
Radioactive compounds were obtained from various sources as listed in Table 1.
Tritiated anisole was synthesized by the BF3 -H,PO-catalyzed exchange with tritiated water (1). The radiochemical purity of all labeled compounds was evaluated by thinlayer chromatography (TLC) and autoradiography and impurities were removed by recrystallization, redistillation, TLC, GC, and silicic acid column chromatography.

Model Aquatic Ecosystem
A model ecosystem was designed for the evaluation of environmental fate of relatively volatile compounds which are directly discharged into the aquatic system. It consists of a 3-liter round-bottomed flask with three necks to which are fitted (a) a microware condenser to reduce the losses of volatile compounds and to retain a constant water level in the main flask, (b) a removable filter for a constant air flow and for sampling water in the flask, and (c) two traps for collection of the volatile metabolites and/or parent compound, and radioactive CO2, respectively. These traps contain 100 ml of acetonitrile and 100 ml of 0.5N aqueous sodium hydroxide, respectively. This system contains 2 liters of reference standard water that provides satisfactory minerals for the organisms (3). The microenvironment of the food chain members consists of phyto-and zooplanktons, green filamentous algae (Oedogonium cardiacutm), snails (Physa), water flea (Daphnia magna), mosquito larvae (fourth instar), (Culex quinquifasciatus), and mosquito fish (Gambusia affinis). The whole system as shown in Figure 1 was kept in a programmed environmental growth chamber set at 80°F (26.7±2°C) with 12 hr daylight exposure to 750 ft-candles (7500 lux).

Operation of Model Aquatic Ecosystem
The components of the model aquatic ecosystem: 300 daphnia, 200 fourth instar mosquito larvae, 6 snails, strands of alga, and miscellaneous plankton were acclimated in the chamber for 1 day. The water was then treated with 0.01-0.1 ppm of radiolabeled compound to produce the experimental contamination. After 24 hr, 50 mosquito larvae and 100 daphnia were removed for radioassay, 3 fish were added, and after another 24 hr the experiment was terminated. Water and trap samples were taken at zero time for use as blanks and after the first and second days to analyze the distribution of radioactivity. The radioactivity in the water was extracted in diethyl ether and analyzed by TLC on silica gel and autoradiography. The organisms were washed, dried, weighed, homogenized, and extracted three times with acetone. The total radioactivity was measured and the pooled extracts were concentrated and analyzed by TLC and autoradiography. Labeled metabolites and degradation products were scraped from the plates and quantitatively determined by liquid scintillation counting. Wherever possible the identity of the radiolabeled spots was determined by cochromatography with known model compounds. The environmental behavior of the radiolabeled contaminants is expressed in the tables in terms of quantitative distribution of the parent compound and various metabolites and degradation products, and in graphs of the various degradation mechanisms utilized by the various organisms of  the model aquatic ecosystem. The techniques for measurement of radioactivity, partition coefficients, water solubilities, and for the identification of labeled products have been fully described (4,5).

Results and Discussion
The various aquatic contaminants were evaluated by quantitative distribution of the radioactivity in the organisms of the model aquatic ecosystem, in the water, and in the traps. Several of the compounds evaluated were appreciably volatile, and the following percentages were recovered from the traps: anisole, 78.45%o, trap I; chlorobenzene, 95.97%o, trap I; nitrobenzene, 2.22%7o, trap I; and phthalic anhydride, 5.02%o, trap II. More than 99%o of all of the other materials was retained in the water phase and its organisms.

Degradative Pathways
Aniline was rapidly detoxified by methylation, acetylation, hydroxylation and conjugations. Daphnia and snail were able to metabolize aniline completely to polar metabolites as shown in Figure 2. N-Methyl and N,Ndimethylaniline were found in alga and mosquito larva, respectively. The fish was the only species which retained small amounts of aniline with ecological magnification (EM) 6, together with N-methylaniline and N,N-dimethylaniline and almost equal amounts of o-, in-, and p-aminophenols.
Acetanilide was found in fish and water extracts and was further metabolized to p-acetamidophenol then conjugated (6). The quantitative distribution of 14C-aniline equivalents is shown in Table 2. Anisole was stored in various o in substantial amounts, EM for mosquito larva, alga, daphnia, a being 22, 27, 563, 771, and 899, resj This indicated that anisole is a fail ether, even though it can be degrad dealkylation in fish and snail as c from the present experiment. Hydr to o-and p-methoxyphenols occurr species except the daphnia. Conjuga the important detoxification mech alga and snail as shown in Figur quantitative distribution of 3H-anisole equivalents is presented in Table 3.

Environmental Health Perspectives
Benzoic acid was not generally stored or Acid ecologically magnified except in snails and E t daphnia, with EM values of 21 in fish, 102 in alga, 138 in mosquito larva, 1772 in daphnia and 2786 in snail. Conjugation through glycine was an important detoxication mechanism, and substantially higher EM values of hippuric acid were found in daphnia (1317), fish (3973), and snail (4535). Thus these species can detoxify benzoic acid much more effectively by this mechanism than alga and mosquito larva. Hydroxylation to phenolic acids occurred to only limited extent in snail, daphnia, with ies of key traces in mosquito larva and none in the fish following ( Fig. 4). Hydroxybenzoic acids and catechol * .
were found in the water extract. Catechol is rganisms the key compound to enter the f8-ketoadipate the fish, pathway for the dissimilation of aromatic Lnd snail compounds, leading to further complete pectively. degradation of benzoic acid (7). The quanrly stable titative distribution of 14C-benzoic acid equivled by 0-alents is shown in Table 4  further explain its undesirable environmental pollutant properties. Hydroxylation of chlorobenzene to o-and p-chlorophenol and to 4-chlorocatechol was found in mosquito larva and in water extracts. These degradation products accounted for only a very small fraction of the total radioactivity accumulated as compared to that f-rom the parent compound as shown in Figure 5. The quantitative distribution of 4C-chlorobenzene equivalents is shown in Table 5.
Environmental Health Perspectives  Nitrobenzene behaved strangely in the model aquatic ecosystem because of its high polarity and was neither stored nor ecologically magnified. However, it is rather resistant to degradation as shown in Figure  6 and the major portion of the radioactivity was retained as parent compound. Nitrobenzene was also reduced to aniline aerobically in all organisms and subsequently acetylated in fish and water extracts. Hydroxylations of nitrobenzene and aniline to the corresponding nitrophenols were found in mosquito larva and snail. The quantitative distribution of 14C-nitrobenzene equivalents is presented in Table 6.
Phthalic anhydride was almost quantitatively converted to phthalic acid and thus underwent further metabolic reactions. Alga was the only ecosystem component that stored the parent compound to high levels, i.e., EM 3169. However, phthalic acid was only magnified to 200. A substantial amount of 14CO2, 5.02% of total applied radioactivity, was collected over the 3-day period. Thus decarboxylation occurred as a major degradation pathway and was followed by benzoic acid degradation. Conjugation with the acidic proton was the most important degradation pathway (Fig. 7) as reflected by the high BI values: 1.779 in alga, 2.411 in mosquito  Table 7. Hexachlorobenzene was identified as a persistent compound (Fig. 8) that comprised the majority of the total accumulated radioactivity: 84%o in snail, 67% in daphnia, 65 % in mosquito larva, and 64%o in fish. Consequently, HCB was magnified to high levels with EM values: 3969 for alga, 1129 for daphnia, 1166 for fish, 2622 for mosquito larva, and 2672 for snail. Pentachlorophenol was found in alga, mosquito larva, and water extracts as the only identified degradation product. Substantial amount of unknown 1 (Rf 0.72) and several minor degradation products were not identified (Table 8), but presumably were further phenolic degradation products (8).
Pentachlorophenol represented 74%o of the radioactivity accumulated, with an EM value of 296 in fish. The other organisms had lower quantities as shown by the lower EM values: alga, 1.58; mosquito larva, 16; snail, 121; daphnia, 165. Conjugation at the phenolic OH was the most important means of degradation found among the organisms. As the result of this modification, relatively high BI values were found of the snail 1.06, alga 1.48, daphnia, 1.61 mosquito larva 2.80, and fish 0.338. The water extract and snail gave more complicated degradation products than were found in the others. These minor metabolites, with Rf 0.75, 0.34, 0.26, 0.22, and 0.16, have not yet been identified (Table 9). Photolytic degradation of pentachlorophenol has been studied and certain lower chlorinated degradation products identified (9).
2,6-Diethylaniline behaved differently from aniline because the ethyl substituents donate electrons to the ring, thereby changing the Environmental Health Perspectives   electron distribution by increasing the basicity and creating steric hindrance. Thus the compound becomes a less suitable substrate for MFO enzymes. 2,6-Diethylaniline was not readily degraded by hydroxylation, acetyla-tion and methylation in the various components of the ecosystem. However, 2,6-diethylaniline was not stored except in the fish with an EM of 120, and several degradation products (Table 10) which might be due to the presence of microorganisms in the water were isolated. 3,5,6-Trichloro-2-pyridinol was a fairly persistent compound in all the organisms and especially in fish and daphnia where 74-76%o of the total radioactivity was parent compound. However, 3,5,6-trichloro-2-pyridinol, was not accumulated at high levels or magnified through the food chain. Conjugation was not as rapid as with true phenols (10) but was very important in mosquito larva, snail, and alga as 51, 40, and 30% of the accumulated radioactivity respectively (Table 11). 3,5,6-Trichloro-2-pyridinol is one of the significant metabolites of chlorpyrifos and methyl-chlorpyrifos in the model ecosystem (11) and a metabolite of goldfish which is eliminated into the aquatic environment (12). 3,5,6-Trichloro-2-pyridinol is labile to ultraviolet light (13), but no 14CO2 was monitored over the 3-day period in the present study.

Comparative Metabolism
The organisms in the model aquatic ecosystem represent at least five phyla: Chlorophyta, Protozoa, Mollusca, Arthropoda, and various microsomal and conjugative detoxication products present in these organisms. Hydroxylation as a detoxication mechanism occurred in the overall order chlorobenezene > anisole > nitrobenzene > aniline > benzoic acid > hexachlorobenzene. The Culex larva appears to utilize this mechanism most effectively (e.g., chlorobenzene, anisole) and was the only animal to hydroxylate hexachlorobenzene. Hydroxylation of benzenoid compounds with the strongly electron-withdrawing groups Cl and NO2 appeared to take place more uniformly than with the electrondonating CH3O and NH2.
O-Dealkylation of anisole was appreciable only in Gambusia and appears to be of recent evolutionary development as this process is considerably more efficient in mammals than in insects (14).
Reduction of nitrobenzene to aniline was remarkably consistent for all the organisms but was of low total importance presumably because of the aerobic nature of the ecosystem environment.
Methylation of aniline was detected only in alga, mosquito larva, and fish and was of minor importance.
Phenolic conjugation of pentachlorophenol was a highly efficient process in all the organisms but Gambusia, which stored large quantities of free pentachlorophenol. Conjugation of the phenols with electron-withdraw-w I~* 9 FlSH MOSQuITo 5TO IL WIE 0 0 n FM MM= WM UL ing groups Cl (Fig. 5), and N02 (Fig. 6) was relatively and uniformly inefficient compared to conjugation of phenols with electrondonating NH2 (Fig. 2) and CHSO (Fig. 3). 3,5,6-Trichloro-2-pyridinol was also very uniformly subject to conjugation for all the organisms but to a lesser extent in fish.
Carboxyl conjugation was a highly efficient process in all the organisms except Physa (Fig. 4). Hippuric acid, the glycine conjugation product was identified in daphnia, snail, and fish.
Dehydrochlorination of DDT to DDE, a unique detoxication process, was responsible for 42%o of the total radioactivity in Culex, about 15% in Daphnia but much smaller amounts in the other organisms. The very high rate in Culex explains this mosquito's well known natural immunity to DDT (Fig.  11).
Reductive dechlorination of DDT to DDD occurred in a relatively small amount throughout all the organisms (Fig. 11).

Molecular Properties and Environmental Responses
A major goal of this study was to establish the role of intrinsic molecular properties such as water solubility, polarity, and lipid/water partitioning in determining the environmental behavior in such quantitative terms as ecological magnification (concentration of parent compound in organism/concentration in water) and biodegradability index (concentration of polar products/concentration of nonpolar products) (15). If suitable correlations can be established, these molecular properties could be useful in predicting the pollutant potentialities of new or proposed environmental contaminants (16,17).
Water solubility was correlated with eco-Environmental Health Perspectives a9 4 ALADAs I 0O MM logical magnification through a simple two variable linear regression using log values from Table 1 and the graph was plotted as Figure 13. The equation was determined as: y = 3.9950 -0.3891x (1) for n = 11 and r = -0.9228, where y = log EM of fish and z = log water solubility, (ppb). The excellent correlation indicates the fundamental importance of water insolubility in causing organic compounds to partition into the lipid tissue of aquatic organisms (16).
Partition coefficient in octanol/H20 was also correlated with EM. The values used in the two-step linear regression analysis were log partition coefficient (X) as determined experimentally (Table 1). For the simple benzene derivatives, these agreed well with 4 3 2 those of Fujita et al. (18). The graph was plotted as Figure 14 and the regression equation was determined as: y = 0.7285 + 0.6335x (2) for n = 11 and r 0.7879, where y = log EM of fish, and x constant. The correlation is good, but the lower correlation coefficient compared to eq. (1) is probably due to the uncertaintly in the measurement of partition coefficient values.
The constant was correlated with the percentage of the remaining parent compound of four benzene derivatives (aniline, anisole, chlorobenzene, and nitrobenzene). The plot  I.  Table 12 in Figure 15 was determined as: y = 33.8559 + 36.1989x (3) for n = 4 and r = 0.9098, where y = the average of five key organisms (alga, daphnia, fish, mosquito larva, and snail-), and x = a constant. For fish, the top member of the food chain, the correlation showed much betteraseq. (4): y = 41.1086 + 40.7901x (4) for n = 4 and r = 0.9685. It was clearly indicated that the Hammett a constant is one of the important intrinsic molecular properties, not only in controlling the disappearance of the parent compound through hydroxylation but also in limiting the formation of the con-1.0 jugated degradation products. Equations (5) and (6) summarize the data from Table 12 for the correlation of the average of five key organisms and of fishes versus a-constant, respectively ( Fig. 16): y= 54.1393 + 37.2628x (5) for n = 4 and r = -0.9089, where y = average of the total percentage of hydroxylated and conjugated products of five key organisms and x = a constant; and Y= 44.3475 + 46.2330x (6) for n = 4 and r = -0.9908, where y = total percentage of hydroxylated and conjugated products of fish and x= a constant.
Efforts were made to correlate the biodegradability index for the various organisms with various molecular properties. The biodegradability index describes the stability of xenobiotics in living organisms and is strongly influenced by susceptability of the compound to attack by microsomal oxidase enzymes, by aliesterases, by conjugating enzymes, etc. Thus the biodegradability of molecules can be influenced by appropriately chosen substituents or degradaphores. For example, with the simple benzene derivatives, the values for the fish, Gambusia, were -COOH, 2.965; -HN2, 1.784; -NO2, 0.023; and -Cl, 0.014. As might be expected, there Environmental Health Perspectives 0 c:> 0o. O 0 cj a w 0 was lower correlation between BI and v constant (r --0.6763) than EM and v constant. Little correlation was found between BI and of constant (r = -0.5654) and BI and water solubility (r = 0.3617).

Lipid Content of Organisms
Clearly, lipid-partitioning micropollutants are stored mostly in the tissue lipids of organisms. The lipid content of the key organisms of the model aquatic ecosystem was determined by extraction in methanol: alga, 2.32%So; daphnia, 1.28%go; mosquito larva, 1.46%; snail, 1.60 %0; fish, 6.15%o. The mean values of EM of the six benzene derivatives are: fish, 116; mosquito larva, 242; daphnia, 593; snail, 833; and alga, 1407. Regression analysis gave eq. (7): y= 3.0171 -0.1397x (7) for n 5 and r = -0.6618, where y mean value of log EM of the six benzene derivatives and x = lipid content (%s ). The substantial negative correlation suggests that degradation by microsomal oxidase enzymes is more important than passive lipid partitioning in determining bioaccumulation.

Conclusions
The data presented demonstrate that the environmental behavior of a series of simple aromatic compounds, e.g. absorption, bioaccumulation, and biodegradation in a simple food chain system with a wide range of organisms, can be predicted by the basic molecular properties of water solubility, partition coefficient for lipid/water, and reactivity as determined by electron density. Thus, for a large variety of aromatic derivatives it should be possible to estimate potential environmental contamination from such fundamental linear free energy values as Hansch's 7r and Hammett's a in a manner analgous to estimation of distribution and reaction of these compounds in a single chemical or biochemical system. This information and its predictive possibilities should simplify the evaluation of new and potentially toxic substances by emphasizing types of compounds to which major investigational efforts should be directed.
Overall environmental degradative pathways for these simple aromatic compounds are generally qualitatively similar in organisms as diverse as alga, clodoceran, snail, mosquito larva, and fish. However, the quantitative aspects of degradation reflect the generally higher levels of microsomal oxidase enzymes (MFO) present in higher evolutionary forms.