The mechanisms of PAH-induced carcinogenesis have been characterized for several compounds, including benzo[a]pyrene, initially metabolized by microsomal cytochrome P450 enzymes to benzo[a]pyrene 7,8-epoxide, which is hydrolyzed by epoxide hydrolase to form benzo[a]pyrene 7,8-dihydrodiol and subsequently oxidized into the ultimate carcinogen, benzo[a]pyrene 7,8-diol-9,10-epoxide. The three-step reaction sequence is a common activation pathway for PAH procarcinogens that contain an angular, singly fused benzene ring and a bay region. Additional metabolic pathways may also be important for the metabolic activation of PAHs (7).

PAHs are hydrophobic compounds whose persistence in the environment is due chiefly to low aqueous solubility. They absorb onto soils and other particulates, influencing their bioavailability and biodegradation (19). Whereas low-molecular-weight PAHs, such as naphthalene, phenanthrene, and anthracene, are usually readily degraded in soil and under laboratory conditions, high-molecular-weight PAHs, such as benzo[a]pyrene and benz[a]anthracene, resist extensive microbial degradation (14).

Since the initial studies by Heitkamp et al. (8, 10, 11) on the bacterial degradation of pyrene, there have been numerous reports describing the microbial oxidation of four-ring PAHs (14). However, information in the literature regarding the mechanisms and pathways involved in the degradation of high-molecular-weight PAHs containing five or more rings is less apparent (13). Although reductions in benzo[a]pyrene concentrations were observed after the incubation of the PAHs with pure cultures and with mixed cultures, few investigators have isolated metabolic by-products of benzo[a]pyrene catabolism.

Sphingomonas yanoikuyae, formerly identified as Beijerinckia B-836, oxidized benzo[a]pyrene to form benzo[a]pyrene cis-7,8- and 9,10-dihydrodiols. Ring cleavage metabolites were not identified (6). One study showed that Mycobacterium sp. strain RJGII-135 is capable of transforming benzo[a]pyrene to initial ring oxidation and ring cleavage products (24). The results of that study indicated that initial enzymatic attack occurs at C-4,5, C-7,8, and/or C-9,10 of benzo[a]pyrene. However, only benzo[a]pyrene cis-7,8-dihydrodiol was identified. The formation of benzo[a]pyrene cis-4,5-dihydrodiol was implied by the isolation of the ortho-cleavage ring fission product 4,5-chrysene dicarboxylic acid. Furthermore, it was proposed that meta-cleavage products, either cis-4-(8-hydroxypyren-7-yl)-2-oxobut-3-enoic acid or cis-4-(7-hydroxypyren-8-yl)-2-oxobut-3-enoic acid and either 7,8-dihydropyrene-7-carboxylic acid or 7,8-dihydropyrene-8-carboxylic acid, were formed. These compounds were not rigorously characterized, and the investigators suggested that their meta-fission ring fission products could be formed after initial dioxygenation and subsequent dehydrogenation at C-7,8 and/or C-9,10 (24).

Mycobacterium sp. strain PYR-1 (recently classified as Mycobacterium vanbaalenii PYR-1 [16]) was isolated from sediments taken from a drainage pond in an estuarine area near an oil field which was exposed to petrogenic chemicals in Port Aransas, Tex. (8). M. vanbaalenii PYR-1 is capable of rapidly degrading phenanthrene, fluoranthrene, pyrene, and anthracene (8, 10, 11, 15, 20).

Previously, we reported that M. vanbaalenii PYR-1 (16) grown in a mixture of yeast extract, peptone, and soluble starch oxidized benzo[a]pyrene to 24.7% organic-extractable metabolites (8). Additionally, in sediment-water microcosm experiments, M. vanbaalenii PYR-1 mineralized approximately 36% of the added benzo[a]pyrene (9). Here, we report benzo[a]pyrene oxidation and ring cleavage products formed by M. vanbaalenii PYR-1. The absolute stereochemistry of the benzo[a]pyrene cis- and trans-dihydrodiols was assigned. To our knowledge, this is the first example of the enantiospecific microbial oxidation of benzo[a]pyrene. In addition, this work extends our knowledge of the pathways proposed for the bacterial metabolism of benzo[a]pyrene.

Chemicals. [7-¹⁴C]benzo[a]pyrene (specific activity, 58.4 mCi/mmol) was purchased from ChemSyn Laboratories (Lenexa, Kans.). Unlabeled benzo[a]pyrene (95% pure) was purchased from Sigma-Aldrich (St. Louis, Mo.). Racemic benzo[a]pyrene cis-4,5-dihydrodiol was prepared by the oxidation of benzo[a]pyrene with OsO₄ in pyridine at ambient temperature. Chiral stationary-phase (CSP) high-performance liquid chromatography (HPLC) columns were purchased from Regis Technologies (Morton Grove, Ill.). Bacteriological media and reagents were purchased from BD Biosciences, Difco Laboratories (Detroit, Mich.). Deuterated acetone (99.96% pure), used as the solvent for nuclear magnetic resonance (NMR) analysis, was purchased from Isotec, Inc. (Miamisburg, Ohio). Other solvents were purchased from J. T. Baker, Inc. (Phillipsburg, N.J.).

Growth and biodegradation conditions. Cultures of Mycobacterium vanbaalenii PYR-1 (DSM 7251) were grown and harvested by centrifugation as previously reported (21). Benzo[a]pyrene was dissolved in dimethylformamide and added to the cultures at a final concentration of 0.48 mM. The flasks were incubated for 96 h, and the subsequent extraction and preparation of acidic metabolites were carried out as previously reported (20).

Analytical methods. Benzo[a]pyrene and its metabolites were separated by HPLC with a model 1050 HPLC system (Hewlett-Packard, Palo Alto, Calif.) with a 4.6- by 250-mm C₁₈ Inertsil ODS-3 column (particle size, 5 μm; MetaChem Technologies, Torrance, Calif.) at a flow rate of 1 ml/min. UV absorption spectra were obtained online with a diode array model 1040A detector (Hewlett-Packard) at 254 nm. The compounds were eluted by using a linear gradient of 50 to 95% methanol-water over 30 min. For the collection of larger amounts of metabolites, a model 100A HPLC (Beckman Instruments, Fullerton, Calif.) with an Inertsil ODS-3 10- by 250-mm column (MetaChem) was used. The mobile phase and gradient were the same as those described above, but a 5-ml/min flow rate was used.

Direct exposure probe (DEP) mass spectrometry (MS) was performed with a ThermoFinnigan TSQ 700 triple quadrupole mass spectrometer operated in the electron ionization (EI) mode. The DEP current was ramped to 800 mA at 5 mA/s. Gas chromatography (GC)-MS was also performed with the same instrument with separation on a J&W DB-5ms capillary column (30 m by 0.25 mm, 0.25-μm particle size). HPLC-electrospray ionization MS (ESIMS) was performed with a ThermoFinnigan TSQ 7000 triple-quadrupole mass spectrometer operated in the negative-ion and positive-ion modes. Separation was achieved with a MetaChem Polaris C₁₈ column (250 by 2.0 mm, 5-μm particle size). The mobile phase was a 30-min linear gradient from 5 to 95% acetonitrile with a constant formic acid level of 0.1%.

Derivatization prior to GC-EI MS analysis was performed by silylation with N,O-bis(trimethylsilyl)trifluoroacetamide with 1% trimethylchlorosilane (Regis Technologies). The samples were dissolved in 250 μl of acetonitrile. One hundred microliters of dissolved sample and 150 μl of silylation reagent were mixed and allowed to react for 1 h at 60°C. GC-MS analyses were performed with a column temperature rate of 20°C/min. The injection volume was 0.5 μl, and the total analysis time was 30 min.

NMR spectroscopy was performed with a model AM500 spectrometer (Bruker Instruments, Billerica, Mass.) at 500.13 MHz. Samples were dissolved in 0.5 ml of deuterated acetone. Assignments were made on the basis of homonuclear decoupling and nuclear Overhauser effect (NOE) experiments, integration, and analysis of substituent effects.

Direct resolution of benzo[a]pyrene dihydrodiols by CSP HPLC was conducted with an HPLC system (Waters Corp., Milford, Mass.) consisting of a model 600 controller and a model 996 photodiode array detector. The 4.6- by 250-mm columns (Regis Technologies) used were (i) a column containing (R)-N-(3,5-dinitrobenzoyl)phenylglycine covalently bonded to γ-aminopropylsilanized silica, (ii) a π acceptor type β-Gem 1 CSP column, (iii) a π acceptor type α-Burke 2 CSP column, and (iv) a π electron acceptor and π electron donor Whelk-O1 CSP column. The ratios of the enantiomers were calculated based on comparison of the peak areas of the two enantiomers.

Circular dichroism (CD) spectra of the dihydrodiol metabolites and the enantiomers resolved by CSP HPLC of the synthetically prepared racemic benzo[a]pyrene cis-4,5-dihydrodiol were measured in a quartz cell with a 1-cm path length at room temperature with a Jasco model 500A spectropolarimeter. The concentration of each sample was adjusted to 1.0 absorbance unit (AU) at 273 nm. The CD spectra are expressed as ellipticity (λ) in millidegrees.

The HPLC chromatogram for benzo[a]pyrene and its metabolites is shown in Fig. 1A. Peaks I (retention time, 14.3 min) and II (retention time, 23.1 min) had identical UV absorption spectra with maximum wavelengths at 221, 230, 252, 281, and 291 nm. The DEP-EI mass spectra had base peak molecular ions at m/z 286 [M · ⁺] and significant fragment ions at m/z 268 [M-18]⁺ and m/z 239 [M-18-29]⁺, consistent with a benzo[a]pyrene dihydrodiol.

FIG. 1. Reverse-phase HPLC chromatograms showing benzo[a]pyrene and its metabolites (A) and CSP HPLC separation of benzo[a]pyrene trans-11,12-dihydrodiol enantiomers (B) from peak I (A) and of benzo[a]pyrene cis-11,12-dihydrodiol and cis-4,5 dihydrodiol enantiomers (more ...)

The proton NMR spectrum of compound I consisted of 2 resonances at 5.23 and 5.96 ppm (J = 2.8 Hz) and 10 aromatic resonances. NMR assignments and coupling constants were as follows: 5.23 (H12), 5.96 (H11, J_11,12 = 2.8 Hz), 7.59 (H8, J_8,9 = 7.5 Hz), 7.63 (H2, J_2,3 = 7.3 Hz), 7.66 (H9, J_9,10 = 8.6 Hz), 7.71 (H1, J_1,2 = 7.3 Hz, J_1,3 = 1.5 Hz), 7.73 (H4, J_4,5 = 9.0 Hz), 7.86 (H3), 7.87 (H5), 8.14 (H7, J_7,8 = 8.6 Hz, J_7,9 = 1.3 Hz), 8.47 (H6), and 8.61 (H10). Homonuclear decoupling and NOE experiments showed that the resonances at 5.23 and 5.96 ppm were attached to C-11 and C-12. The chemical shifts were consistent with the carbinol protons of a dihydrodiol, the coupling constant (J_11,12 = 2.8 Hz) was indicative of hydroxyl groups in a quasi-diaxial conformation (30, 31), and the HPLC retention of the peak was consistent with a trans configuration. Peak I was identified as benzo[a]pyrene trans-11,12-dihydrodiol.

The NMR spectrum of peak II (Fig. 1A) indicated that it contained two compounds. The major component had carbinol proton resonances at 5.26 and 5.92 ppm and a coupling constant of 4.5 Hz. The coupling constant was larger than that noted for the carbinol protons of compound I (benzo[a]pyrene trans-11,12-dihydrodiol) and thus indicated that the hydroxyl groups were in a quasi-diequatorial conformation. NMR assignments and coupling constants were as follows: 5.26 (H12), 5.92 (H11, J_11,12 = 4.5 Hz), 7.60 (H8, J_8,9 = 7.5 Hz), 7.67 (H2, H9), 7.73 (H4, J_4,5 = 9.0 Hz), 7.80 (H3), 7.87 (H5), 8.02 (H1, J_1,2 = 7.5 Hz, J_1,3 = 1.5 Hz), 8.15 (H7, J_7,8 = 8.4 Hz), 8.50 (H6), and 8.57 (H10, J_9,10 = 8.6 Hz). On the basis of the NMR experiments and the HPLC data, the larger component of peak II was identified as benzo[a]pyrene cis-11,12-dihydrodiol. The minor compound was identified as benzo[a]pyrene cis-4,5-dihydrodiol, since it had chemical shifts identical to those of the authentic compound (5.09 [H4,5], 7.67 [H8, J_8,9 = 7.5 Hz], 7.69 [H2, J_2,3 = 7.5 Hz], 7.72 [H9, J_9,10 = 8.4 Hz], 7.95 [H3], 7.98 [H1, J_1,2 = 8.2 Hz], 8.09 [H7, J_7,8 = 8.0 Hz, J_7,9 = 1.5 Hz], 8.81 [H12], 8.27 [H6], 8.82 [H11, J_11,12 = 9.0 Hz], and 8.87 [H10]). Its identity was further confirmed by comparing its HPLC retention times to those of the authentic compound by using both reverse-phase and CSP columns.

Peak III (Fig. 1A) eluted at 26.6 min. HPLC-ESIMS was performed, and several peaks were observed in the positive mass spectrum. One peak had a UV spectrum matching that of peak III. The apparent molecular weight of the HPLC-ESIMS peak was 311, indicating that it contained an odd number of nitrogen atoms and was not a benzo[a]pyrene metabolite. The DEP-EI, DEP-MS-MS, and NMR spectra were also obtained and analyzed. These spectra were used to determine that peak III was probably a natural product formed by the Mycobacterium sp.

Peak IV (Fig. 1A) eluted at 29.7 min and had a UV absorbance spectrum with maximum wavelengths at 230, 272, 298, 325, and 340 nm. The GC-EI mass spectrum had a base peak molecular ion at m/z 270 [M ·+] and significant fragment ions at m/z 242 [M-28]⁺, 213 [M-28-29]⁺, 121, 106, and 93.7. The NMR spectrum had 10 aromatic resonances, with one shifted upfield of the rest at 6.69 ppm. NMR assignments and coupling constants were as follows: 6.69 (H8), 8.12 (H2, J_2,3 = 7.5 Hz), 8.13 (H4, J_4,5 = 9.0 Hz), 8.17 (H5), 8.30 (H3) 8.35 (H1, J_1,2 = 7.7 Hz), 8.43 (H7, J_7,8 = 9.5 Hz), 8.50 (H6), and 8.63 (H11, J_11,12 = 9.0 Hz). Homonuclear decoupling and NOE experiments showed that there were no protons at C-9 and C-10. A benzocoumarin-like structure analogous to that formed by M. vanbaalenii PYR-1 from anthracene (20) is consistent with the NMR and mass spectral data for peak IV, and this structure was identified as 10-oxabenzo[def]chrysen-9-one.

The GC-EI and DEP-EI mass spectra showed that peak V (retention time, 32.0 min) was a mixture of two compounds (m/z 270 [M ·+]) with structures similar to that of compound IV, a benzo[a]pyrene phenol (m/z 268 [M · +]), and two compounds containing hydroxy and methoxy groups (m/z 298 [M · +]). The NMR spectrum confirmed that methoxy groups were present, although it was more consistent with the presence of one hydroxymethoxybenzo[a]pyrene compound and one dimethoxybenzo[a]pyrene compound. Many overlapping aromatic resonances were present, so the positions of the methoxy groups could not be determined. The types of compounds that are in peak V are consistent with those formed by M. vanbaalenii PYR-1 from other PAHs (20, 21).

Benzo[a]pyrene eluted at 39.4 min. Its identity was confirmed by comparison of its UV, mass, and NMR spectra to those of authentic benzo[a]pyrene.

Replacement culture experiments. In a separate experiment, cultures of M. vanbaalenii PYR-1 were prepared as described above and incubated with 2 mg of synthetic benzo[a]pyrene cis-4,5-dihydrodiol instead of benzo[a]pyrene. GC-MS analysis of trimethylsilyl-derivatized acidic extracts from 24-, 48-, 72-, and 96-h incubations showed that all four samples produced a compound at 18.4 min that had an apparent molecular weight of 460, the weight of a ring-opened structure (316 Da) with two trimethylsilyl groups (316 plus 144 Da). The mass spectrum consisted of ions at m/z 460(2)[M·+], 445(2)[M-15]⁺, 343(100)[M-117]⁺, 327, 270, 254, 239, 226, 224, 213, 147, 113, and 73. The ion at m/z 147 may indicate adjacent trimethylsilyl groups. The mass spectral data are consistent with those for ditrimethylsilyl-chrysene-4,5-dicarboxylate. The 24- and 48-h acid extract samples yielded an additional compound at 11.7 min that had an apparent molecular weight of 372. The mass spectra consisted of ions at m/z 372(20)[M · ⁺], 357(15)[M-15]⁺, 255(10)[M-117]⁺, 250(7), 239(100), 173(6), and 172(6). The mass spectrum suggests a monoderivatized aromatic structure with an aldehyde and an acid moiety, a probable precursor of 4,5-chrysene-dicarboxylic acid. The 24-, 48-, 72-, and 96-h acid extracts produced chromatographic peaks at 16.3 min that had an apparent molecular weight of 344. Major fragment ions included those at m/z 329 [M-15], 255 [M-89], 227 [M-117], 226 [M-118], 165, and 113. The molecular weight corresponds to a trimethylsilyl ester of chrysene 4-carboxylic acid or chrysene 5-carboxylic acid.

Optical purity and absolute configuration of benzo[a]pyrene cis- and trans-dihydrodiol metabolites. To determine the optical purity and absolute configuration of the dihydrodiols formed by M. vanbaalenii PYR-1, direct resolution of the benzo[a]pyrene dihydrodiols was attempted with four types of CSP HPLC columns. By testing the columns with different solvent systems, it was found that enantiomeric separation was best achieved with the Whelk-O1 column. This CSP HPLC column separated benzo[a]pyrene trans-11,12-dihydrodiol (Fig. 1A, peak I) into two chromatographic peaks that eluted at 4.4 and 5.1 min, respectively (Fig. 1B). Since these two peaks had UV-visible absorption spectra identical to those of benzo[a]pyrene trans-11,12-dihydrodiol, it was evident that the two CSP HPLC-separated compounds were enantiomers. Since the amount of metabolite present was insufficient for derivatization, determination of the absolute configuration by CD analysis was not pursued. However, based on previous chiral resolution and HPLC retention times of a large series of PAH trans- and cis-dihydrodiols (1, 32), we identified the enantiomer eluting at 4.4 min as benzo[a]pyrene-11S,12S-dihydrodiol and the other enantiomer (eluting at 5.1 min) as benzo[a]pyrene-11R,12R-dihydrodiol. Based on the peak areas, the ratio of the two enantiomers was 50% benzo[a]pyrene trans-11S,12S-dihydrodiol and 50% benzo[a]pyrene trans-11R,12R-dihydrodiol (Fig. 1B).

Direct resolution of peak II, containing the benzo[a]pyrene cis-11,12-dihydrodiol and cis-4,5-dihydrodiol enantiomers, by CSP HPLC resulted in the baseline separation of three chromatographic peaks, which eluted at 4.1, 4.8, and 5.2 min (Fig. 1C). These peaks were benzo[a]pyrene cis-11,12-dihydrodiol and the two enantiomeric cis-4,5-dihydrodiols. The benzo[a]pyrene cis-11,12-dihydrodiol that eluted at 4.1 min was optically active, as evidenced by its CD spectral analysis. The metabolite could be either optically pure or a mixture of both enantiomers with one of them predominating. Repeated CSP HPLC separations employing different CSP HPLC columns and various solvent systems to resolve the enantiomers showed only one compound, indicating that the benzo[a]pyrene trans-11,12-dihydrodiol was an optically pure enantiomer. Comparison of the CSP HPLC elution order of structurally similar dihydrodiol enantiomers indicated that the benzo[a]pyrene cis-dihydrodiol had an 11S,12R absolute configuration (1, 32). To determine the optical purity and absolute configuration of the two cis-4,5-dihydrodiol metabolites, the synthetic racemic cis-4,5-dihydrodiol was first directly resolved by CSP HPLC. The chromatographic peaks that eluted at 4.8 and 5.2 min (Fig. 1C) were collected and subjected to UV-visible absorption and CD spectral measurement. The peaks at 4.8 and 5.2 min had UV-visible absorption spectra that were identical to those of the racemic cis-4,5-dihydrodiol (data not shown) and had identical CD spectra (data not shown). Based on previous chiral resolution and HPLC retention times of a series of trans- and cis-dihydrodiols from benzo[a]pyrene, benz[a]anthracene, and chrysene, the enantiomer that eluted at 4.8 min had a 4S,5R absolute configuration, and the other enantiomer had a 4R,5S absolute configuration. The retention times of the standard were the same as those of the metabolite. Based on the peak areas of these enantiomers, the absolute configuration of the cis-4,5-dihydrodiol was 30% 4S,5R and 70% 4R,5S (Fig. 1C). Thus, the metabolite was 40% optically pure.

M. vanbaalenii PYR-1 initiates its attack on benzo[a]pyrene by dioxygenation or monooxygenation at C-4,5, C-9,10, and C-11,12. Figure 2 shows the proposed pathway for benzo[a]pyrene degradation by M. vanbaalenii PYR-1. The formation of benzo[a]pyrene cis-4,5-dihydrodiol has not been confirmed previously, although it was proposed by Schneider et al. (24) based on the identification of the ortho-cleavage ring fission product 4,5-chrysene dicarboxylic acid. We have shown here that M. vanbaalenii PYR-1 produces chrysene 4,5-dicarboxylic acid when it is incubated with benzo[a]pyrene cis-4,5-dihydrodiol. The reaction of dioxygenation in the K region (4,5 positions) to form the cis-dihydrodiol and ortho cleavage to form chrysene 4,5-dicarboxylic acid are consistent with results of our previous studies of phenanthrene, pyrene, and 7,12-dimethylbenz[a]anthracene degradation (10, 20, 21). Other Mycobacterium isolates have been shown to carry out similar reactions (5, 23, 25).

FIG. 2. Proposed pathway for the degradation of benzo[a]pyrene by M. vanbaalenii PYR-1. Compounds in brackets are hypothetical intermediates.

The K region dioxygenase from M. vanbaalenii PYR-1 has been characterized, and the nidA and nidB genes coding for the α and β subunits of the terminal dioxygenase have been cloned, expressed, and sequenced (16). We have also observed that PAH-degrading Mycobacterium spp. isolated from geographically diverse PAH-contaminated sites possess nidA and nidB genes similar to those of M. vanbaalenii PYR-1 (3).

As in anthracene degradation by M. vanbaalenii PYR-1 (20), a benzocoumarin-type metabolite was identified in the culture broths from benzo[a]pyrene incubations. The novel metabolite, 10-oxabenzo[def]chrysen-9-one, may be formed by initial dioxygenation at C-9 and C-10, yielding benzo[a]pyrene cis-9,10-dihydrodiol. Dehydration of the dihydrodiol to the dihydroxy intermediate with subsequent meta cleavage and aromatic-ring closure may lead to the formation of 10-oxabenzo[def]chrysen-9-one. The meta-ring fission product, cis-4-(8-hydroxypyren-7-yl)-2-oxobut-3-enoic acid, was tentatively identified in Mycobacterium sp. strain RJGII-135 (24).

M. vanbaalenii PYR-1 also oxidized benzo[a]pyrene to cis- and trans-11,12-dihydro-11,12-dihydroxybenzo[a]pyrene. The identification of the cis- and trans-dihydrodiols, indicating dioxygenation and monooxygenation reactions, is consistent with results of our studies of naphthalene, anthracene, phenanthrene, pyrene, and 7,12-dimethylbenz[a]anthracene degradation (10, 15, 20, 21). The formation of the benzo[a]pyrene trans-11,12-dihydrodiol was probably due to cytochrome P450 acting to form benzo[a]pyrene 11,12-epoxide, with subsequent hydrolysis by epoxide hydrolase to form benzo[a]pyrene trans-11,12-dihydrodiol. The genes encoding cytochrome P450 and epoxide hydrolase have been detected in M. vanbaalenii PYR-1 (unpublished data). The isolation of the novel bacterial intermediates derived from enzymatic attack at C-11 and C-12 further demonstrates the versatility of M. vanbaalenii PYR-1 in the degradation of high-molecular-weight PAHs and its potential use in the bioremediation of PAH-contaminated sites. Although this metabolic route has not been described previously for the bacterial degradation of benzo[a]pyrene, Lindquist and Warshawsky (18) and Warshawsky et al. (26) demonstrated that the oxidation of benzo[a]pyrene by the green alga Selenastrum capricornutum resulted in the formation of benzo[a]pyrene cis-4,5-, 7,8-, 9,10-, and 11,12-dihydrodiols.

We have also isolated hydroxymethoxy and dimethoxy derivatives of benzo[a]pyrene. Although we have insufficient material to determine the positions of substitution, these data confirm recent results obtained in our laboratory indicating that M. vanbaalenii PYR-1 has a constitutive catechol-O-methyltransferase (unpublished observations).

There are published data on the absolute configuration of the cis-dihydrodiols formed from the bacterial oxidation of naphthalene, anthracene, phenanthrene, benz[a]anthracene, and chrysene (12). The cis-dihydrodiols have an R absolute configuration at the hydroxylated benzylic centers. In this investigation, we demonstrate for the first time the absolute configuration of the benzo[a]pyrene cis- and trans-dihydrodiols formed by the microbial oxidation of benzo[a]pyrene.

M. vanbaalenii PYR-1 oxidized benzo[a]pyrene to 70% benzo[a]pyrene cis-4R,5S-dihydrodiol and 30% benzo[a]pyrene cis-4S,5R-dihydrodiol. The enantiomer-specific cis dihydroxylation of the K region of benzo[a]pyrene is different from that resulting from 7,12-dimethylbenz[a]anthracene, for which the predominant enantiomer was 95% dimethylbenz[a]anthracene 5S,6R-dihydrodiol (21). On the other hand, benzo[a]pyrene cis-11,12-dihydrodiol adopted an 11S,12R conformation with 100% optical purity. The enantiomeric composition of benzo[a]pyrene trans-11,12-dihydrodiol formed by M. vanbaalenii PYR-1 was an equal mixture of 11S,12S and 11R,12R. As described earlier, this metabolite was formed by the epoxidation of benzo[a]pyrene catalyzed by the cytochrome P450 enzymes to form benzo[a]pyrene 11,12-epoxide, with subsequent hydrolysis by epoxide hydrolase. There are two possible enzymatic pathways that explain the formation of nearly equal amounts of the two enantiomers. If the cytochrome P450 of M. vanbaalenii PYR-1 is highly stereoselective, it will form an optically pure benzo[a]pyrene 11,12-epoxide. But if the epoxide hydrolase is not stereoselective in its reaction with the 11,12-epoxide racemic benzo[a]pyrene, trans-11,12-dihydrodiol will be formed. A second pathway would be possible if the stereoselectivity statuses were reversed (i.e., a nonstereoselective cytochrome P450 and a highly stereoselective epoxide hydrolase). This possibility warrants further investigation.

In this study, we have shown that benzo[a]pyrene was metabolized by M. vanbaalenii PYR-1 to a variety of metabolites. Since benzo[a]pyrene is mutagenic, tumorigenic, and widely distributed in the environment, it is important to determine whether this biotransformation has an activation effect or a detoxification effect on tumorigenicity. As shown in Fig. 2, benzo[a]pyrene 11,12-epoxide is the precursor of benzo[a]pyrene trans-11,12-dihydrodiol. Kouri et al. (17) showed that this compound is not tumorigenic in C3H/Cum_f mice. Wislocki et al. (29) reported that it is a weak pulmonary carcinogen in newborn mice and that benzo[a]pyrene exhibited a much higher degree of carcinogenicity. The benzo[a]pyrene cis-dihydrodiols formed by M. vanbaalenii PYR-1 should not be tumorigenic. 10-Oxabenzo[def]chrysen-9-one cannot form a bay region diol epoxide and is therefore unlikely to be carcinogenic. Thus, it is apparent that the biotransformation of benzo[a]pyrene by M. vanbaalenii PYR-1 is a metabolic detoxification process. The ability of M. vanbaalenii PYR-1 to form sterically and optically pure arene dihydrodiols can be exploited in the future for the biosynthetic design of pharmaceutical or industrially important compounds.

We thank John B. Sutherland and Thomas M. Heinze for critical reviews, Diana Mathews for clerical assistance, and Danny Tucker for graphical assistance.