Anaerobic and aerobic cleavage of the steroid core ring structure by Steroidobacter denitrificans.

The aerobic degradation of steroids by bacteria has been studied in some detail. In contrast, only little is known about the anaerobic steroid catabolism. Steroidobacter denitrificans can utilize testosterone under both oxic and anoxic conditions. By conducting metabolomic investigations, we demonstrated that S. denitrificans adopts the 9,10-seco-pathway to degrade testosterone under oxic conditions. This pathway depends on the use of oxygenases for oxygenolytic ring fission. Conversely, the detected degradation intermediates under anoxic conditions suggest a novel, oxygenase-independent testosterone catabolic pathway, the 2,3-seco-pathway, which differs significantly from the aerobic route. In this anaerobic pathway, testosterone is first transformed to 1-dehydrotestosterone, which is then reduced to produce 1-testosterone followed by water addition to the C-1/C-2 double bond of 1-testosterone. Subsequently, the C-1 hydroxyl group is oxidized to produce 17-hydroxy-androstan-1,3-dione. The A-ring of this compound is cleaved by hydrolysis as evidenced by H2(18)O-incorporation experiments. Regardless of the growth conditions, testosterone is initially transformed to 1-dehydrotestosterone. This intermediate is a divergence point at which the downstream degradation pathway is governed by oxygen availability. Our results shed light into the previously unknown cleavage of the sterane ring structure without oxygen. We show that, under anoxic conditions, the microbial cleavage of steroidal core ring system begins at the A-ring.

the aerobic steroid catabolic pathway are in progress. Overall, during the aerobic degradation of testosterone, three reactions are catalyzed by oxygenases. In addition to testosterone, cholesterol and phytosterols are also degraded through the common 9,10-seco-pathway, with C19 androgens as the intermediates ( 12 ).
By contrast, information on the biochemical and molecular details of anaerobic steroid degradation is very limited. Steroids, especially sterols, may remain in anoxic sediments over hundreds of millions of years ( 13,14 ), indicating that steroids are not degraded facilely by anaerobes. Obviously, anaerobes must use a novel, oxygenase-independent catabolic strategy to degrade steroids in the absence of oxygen. Denitrifying bacteria are facultative aerobes that can use various aromatic compounds and terpenoids as sole sources of carbon and energy; thus, they play a crucial role in carbon cycling in the environment. In the last decade, a few denitrifying bacteria that can anaerobically mineralize steroids were isolated and characterized (15)(16)(17)(18). Among them, S. denitrifi cans DSMZ18526 has an unusual ability to degrade testosterone under both oxic and anoxic conditions. The Blast results showed that S. denitrifi cans strains are widely distributed in diverse oxic and anoxic ecosystems, e.g., agriculture soil, bioremediated soil, anoxic sediment, activated sludge, and anoxic sludge (supplementary Fig. I).
Recently, the initial reactions involved in the anaerobic metabolism of cholesterol and testosterone were reported (19)(20)(21), albeit the ring cleavage details of the anaerobic pathways are yet to be unraveled. In this study, we adopted a 13 C metabolomic approach to investigate the anaerobic degradation of testosterone using S. denitrifi cans as a model organism. The aerobic testosterone degradation by the same model organism was also studied for comparison. The results obtained shed light into the previously unknown cleavage of the sterane ring structure without oxygen. To our knowledge, this is the fi rst study showing that under anoxic conditions, the microbial cleavage of steroidal core ring system begins at the A-ring.

Aerobic growth of S. denitrifi cans with testosterone
The S. denitrifi cans was grown in phosphate-buffered shakefl ask cultures (500 ml in 2 l Erlenmeyer fl asks) containing 4 mM testosterone. The cultures were incubated at 28°C in an orbital

Anaerobic in vivo transformation of [2,3,4C-13 C] testosterone
S. denitrifi cans was grown with 2 mM of unlabeled testosterone in a 250 ml glass bottle. After the unlabeled testosterone was completely consumed, 10 ml of the anoxic culture was transferred into a 12 ml glass bottle sealed with a rubber stopper. The S. denitrifi cans cells were subsequently fed with 2 mM testosterone (unlabeled testosterone and [2,3,4C- 13 C]testosterone were mixed in a 1:1 molar ratio) under denitrifying conditions. The samples (1 ml) were withdrawn after 10 min, 6 h, and 12 h of incubation at 28°C. After the second sampling (6 h), 0.5 mM of mercaptopropionic acid [an inhibitor of acyl-CoA dehydrogenase ( 22 )] was immediately added to the anoxic culture. The culture samples were immediately extracted three times with the same volume of ethyl acetate to recover testosterone-derived intermediates. The ethyl acetate fractions were combined, the solvent was evaporated, and the residue was redissolved in 100 l of methanol. The testosterone-derived intermediates were identifi ed using UPLC-HRMS.

Anaerobic growth of S. denitrifi cans with unlabeled testosterone
The S. denitrifi cans was grown with 4 mM of unlabeled testosterone at 28°C in anoxic fed-batch cultures (2 l) according to published procedures ( 20 ). The amounts of residual testosterone in the anoxic culture were monitored using HPLC. After the consumption of 2 mM testosterone, 0.5 mM mercaptopropionic acid was added to the cultures, and incubation was continued for an additional 12 h. The cultures were subsequently extracted three times with the same volume of ethyl acetate to recover residual testosterone and its derivatives from the aqueous phase. Separation of ethyl acetate extracts was performed using silica gel chromatography, TLC, and HPLC. The structures of HPLC-purifi ed intermediates were determined using NMR spectroscopy.  by using the 2,6-dimethylphenol photometric method as described elsewhere ( 21 ).

Silica gel chromatography
A 385 ml silica gel column (55 × 3 cm; SiliaFlash P60; Silicycle) was equilibrated with 2 bed volumes of dichloromethane/ethyl acetate/ethanol (14:4:1, v/v). The ethyl acetate extract (approximately 350 mg dissolved in 3 ml ethyl acetate) was loaded to the column and eluted with the same solvent system at a fl ow rate of 2 ml min Ϫ 1 . The eluate was collected in 5 ml fractions, and a 0.5 ml sample was withdrawn from each fraction. The solvent was evaporated until dry, and the residue was redissolved in 10 l of methanol. The samples after silica gel chromatography (SGC) were analyzed using TLC. The fractions that contained the same compounds were pooled and evaporated to dryness, and 200 l of methanol was used to redissolve the residue. Further purifi cation of testosterone-derived intermediates was performed using TLC.

Thin layer chromatography
The steroid standards and products were separated on silica gel aluminum TLC plates (Silica gel 60 F 254 , thickness, 0.2 mm, 20 × 20 cm; Merck). The following developing solvent system was used: dichloromethane/ethyl acetate/methanol (14:4:1, v/v). The steroid compounds were visualized under UV light at 254 nm or by spraying the TLC plates with 30% (v/v) H 2 SO 4 .

High-performance liquid chromatography
A reversed-phase Hitachi high-performance liquid chromatography (HPLC) system was used for the fi nal separation. The separation was achieved on an analytical RP-C 18 column [Luna 18 ( 2 ), 5 m, 150 × 4.6 mm; Phenomenex] with a fl ow rate of 0.5 ml min Ϫ 1 . The separation was performed isocratically at room temperature with 50% (v/v) methanol as an eluent. The steroid products were detected in the range of 200-300 nm using a photodiode array detector. In addition, HPLC was used for the quantifi cation of steroids present in the S. denitrifi cans cultures. The quantity of steroids (testosterone, 3,17-DHSA, and 2,3-SAOA) was calculated from their respective peak areas using a standard curve of individual standards. The R 2 values for the standard curves were greater than 0.98. Data are averages of three measurements.

O-Incorporation experiments
The denitrifying growth of S. denitrifi cans with testosterone and the preparation of cell extracts were performed as previously described ( 21 ). To determine the origins of the oxygen atoms at C-1 and/or C-3 of 17-hydroxy-androstan-1,3-dione and 17-hydroxy-1-oxo-2,3-seco-androstan-3-oic acid (2,3-SAOA), two in vitro assays were performed. The two reaction mixtures (3 ml for each assay) were prepared anaerobically and were incubated at 30°C for 16 h with shaking. The steroid products were extracted from the assays using ethyl acetate, and the extracts were analyzed using UPLC-APCI-mass spectrometry.
Control assay. The  shaker (180 rpm). After the consumption of 2 mM testosterone, 1 mM of 3-chlorocatechol [a meta -cleavage inhibitor ( 23 )] was added to the cultures, and incubation continued for an additional 12 h. The cultures were extracted using ethyl acetate, and testosterone-derived intermediates present in the extract were analyzed using UPLC-HRMS. Separation of ethyl acetate extracts was performed using silica gel chromatography, TLC, and HPLC.

S. denitrifi cans grown under various concentrations of oxygen
A S. denitrificans culture (500 ml) was first anaerobically grown on 2 mM testosterone. After testosterone and its derivatives were completely consumed, 50 ml of the preculture was mixed with 450 ml of fresh phosphate-buffered medium (pH 7.0) containing 2.2 mM testosterone, 10 mM NH 4 Cl (the nitrogen source), and 10 mM NaNO 3 (the potential electron acceptor). The resulting cultures (100 ml) were transferred to fi ve 1 l glass bottles sealed with rubber stoppers and were incubated under various concentrations of oxygen [headspace (900 ml); 0, 2.5, 5, 10, and 20% (v/v)]. The culture containing 20% oxygen in headspace was prepared in air. The remaining four cultures were prepared in an anaerobic chamber containing 95% nitrogen and 5% hydrogen gas. Oxygen gas was injected into the headspace after passing through a 0.22 m membrane fi lter (Millipore). The fed-batch cultures were incubated at 28°C with shaking (180 rpm). Samples (3 ml) were retrieved every 4 h to measure the growth of bacterial cells (measured as total proteins), the residual amount of nitrate and testosterone, and the production of ring cleavage intermediates (3,17-DHSA and 2,3-SAOA). NaNO 3 was added continuously to 10 mM when the nitrate added initially was consumed. After the consumption of 1 mM testosterone, 0.5 mM mercaptopropionic acid and 1 mM 3-chlorocatechol were added to the cultures, and incubation was continued for an additional 12 h.

Measurement of protein content and nitrate
The protein content in the culture samples and in cell extracts was determined using a BCA protein assay according to manufacturer's instructions, with BSA as the standard. Nitrate was determined achieved with a linear gradient of Solvent B from 10% to 99% in 8 min. In APCI -MS analysis, the temperature of the ion source was maintained at 100°C. Nitrogen desolvation gas was set at a fl ow rate of 500 l h -1 and the probe was heated to 400°C. Nitrogen was used as the APCI nebulizer gas. The corona current was maintained at 20 A , and the electron multiplier voltage was set to1700 eV. The parent scan was in the range of m / z 50-500. The predicted elemental composition of individual intermediates was calculated using MassLynx Mass Spectrometry Software (Waters).

UPLC-ESI-HRMS
The ethyl acetate extractable samples or TLC-purifi ed testosterone-derived intermediates were also analyzed using UPLC-ESI-HRMS. The separation conditions for UPLC were the same as those for UPLC-APCI-HRMS. Mass spectral data were collected in +ESI mode in separate runs on a Waters HDMS-QTOF synapt mass spectrometer operated in a scan mode from m / z 50 to 500. The capillary voltage was set at 3000 V; the source and desolvation temperatures were 100°C and 250°C, respectively. The cone gas fl ow rate was 50 l h Ϫ 1 . 18 O-Labeled water-treated assay. A total of 1.5 ml of 18 O-labeled water (97 atom %, Aldrich) was added to 1.5 ml of 100 mM Tris-HCl buffer (pH 7) containing soluble proteins of S. denitrifi cans (15 mg) and 1 mM mercaptopropionic acid. The fi nal 18 O-water content was approximately 48.5%. The reaction was started by adding 4.5 mM of 1-testosterone to the anoxic assay. The 2-propanol content was also 6.67%.

UPLC-APCI-HRMS
The ethyl acetate extractable samples or purifi ed steroid intermediates were analyzed using UPLC-MS with UPLC coupled to an atmospheric pressure chemical ionization (APCI) highresolution mass spectrometry (HRMS). Mass spectral data were obtained using a Waters HDMS-QTOF synapt mass spectrometer (Waters) equipped with a standard APCI source operating in the positive ion mode. Separation was achieved on a reversedphase C 18

Phylogenetic analysis of S. denitrifi cans strains
Detection of the phylogenetic relationship of S. denitrifi cans strains was conducted using Clustal W and MEGA 5.0 ( 24 ). Sixtythree 16S rRNA gene sequences of S. denitrifi cans were retrieved from the GenBank database of the National Center for Biotechnology Information (NCBI).

Anaerobic testosterone catabolism by S. denitrifi cans
The metabolites of the anaerobic testosterone catabolism were extracted using ethyl acetate from in vivo transformation assays. These testosterone-derived intermediates were identifi ed using UPLC-HRMS. The steroid substrate was composed of [2,3,4C-13 C]testosterone and unlabeled testosterone (mixed in 1:1 molar ratio). Therefore, pairs of molecular adduct ions (with the m / z difference of 3) were observed in the mass spectra of testosterone-derived intermediates ( Fig. 3 ). The mass spectra of their 17-keto structures are shown in Fig. 4 . At the beginning of the assay, only testosterone was detected ( Fig. 3A 1). After 6 h of anaerobic incubation, 1-dehydrotestoterone, 1-testosterone, and their 17-keto derivatives appeared ( Fig. 3A 2). These steroid compounds were identifi ed by comparison with authentic steroid standards by UPLC-HRMS. After a further incubation (6 h) with 0.5 mM mercaptopropionic acid (an acyl-CoA dehydrogenase inhibitor), a few new intermediates were present ( Fig. 3A 3). This phenomenon suggests that ␤ -oxidation may play a role in the degradation of downstream intermediates. The testosterone-derived intermediates detected in the anaerobic S. denitrifi cans cultures are summarized in supplementary  Table I.
A testosterone-derived intermediate (compound X, Fig. 3A 3) with 17 carbons was identifi ed using UPLC-HRMS. Its ESI-mass spectrum ( Fig. 7 ) indicated that this compound is labeled with two 13 C. We assumed that the C-1 and C-2 of intermediate X was removed because i ) testosterone, the steroid substrate, was labeled with three 13 C at C-2/C-3/C-4, and ii ) in the case of 2,3-SAOA, the single bond between C-2 and C-3 was broken. So far, we cannot produce a suffi cient amount of compound X for NMR analysis. Therefore, the exact structure of the intermediate X remains unclear.

Modes of respiration and testosterone catabolism by S. denitrifi cans grown under various oxygen concentrations
To test i ) whether S. denitrifi cans adopts the oxygenindependent 2,3-seco-pathway under microaerobic conditions and ii ) whether the coexistence of the 2,3-seco-and 9,10-seco-pathways is possible, we grew S. denitrifi cans under various oxygen concentrations (0, 2.5, 5, 10, and 20%, v/v). The volume ratio of the S. denitrifi cans cultures (100 ml) to its headspace (900 ml) was 1 to 9; thus, the oxygen concentration in the headspace was not heavily changed during To determine the origins of the oxygen atoms at C-1 and/or C-3 of 17-hydroxy-androstan-1,3-dione and 2,3-SAOA, we conducted two in vitro transformation assays using 1-testosterone as the substrate, as follows: i ) an 18 Olabeled water-treated assay contained approximately 48.5% 18 O-labeled water (mole/mole) in the anoxic reaction mixture; and ii ) a control assay was incubated under anoxic conditions without the addition of 18 O-labeled water. Compared with the 17-hydroxy-androstan-1,3-dione purifi ed from the control assay ( Fig. 6A ), an additional 18 Oisotopic molecular ion ([M+2+H] + , m/z 307.2162) was observed in the APCI-mass spectrum of the 1,3-dioxo product purified from the 18 O-labeled water-treated assay ( Fig. 6B ). The APCI-mass spectrum of 2,3-SAOA purifi ed from the 18 O-labeled water-treated assay showed three dominant protonated molecular ions ([M+H] + , m/z 323.2227, 325.2273, and 327.2320; Fig. 6D ). Their elemental composition was calculated as C 19 Fig. 6C ). These data indicated that under anoxic conditions, after the activation of the A-ring through a hydration reaction, the cleavage of the steroidal core ring system begins with the A-ring by a hydrolysis reaction. Moreover, the presence of 10 mM KCN [an inhibitor generally inactivates members of the xanthine oxidase family ( 25,26 )] considerably absence of oxygen and under 2.5% oxygen ( Fig. 8B ). 2,3-SAOA concentration in the strictly anaerobic culture was 7 ± 1 g ml Ϫ 1 . It is worth mentioning that the two ring cleavage intermediates apparently never coexisted in any tested bacterial cultures. Our data showed that S. denitrifi cans adopts only one testosterone catabolic pathway at any time, depending on oxygen availability.

DISCUSSION
Under oxic conditions, the degradation of testosterone by S. denitrifi cans appeared to follow the 9,10-seco-pathway reported previously ( 27 ). The microbial oxygenolytic cleavage of cycloalkane and aromatic rings under oxic conditions is widely distributed in nature, for example, aerobic catabolism of cyclohexanol and phenol by bacteria ( 28 ). The insertion of hydroxyl groups into the organic substrates by oxygenases is a common catabolic strategy that enables the microbial cells to overcome the inherent inertness of these compounds ( 29,30 ).
The metabolomic data presented in this study show that anaerobic degradation of testosterone by S. denitrifi cans occurs through a catabolic route that differs fundamentally from the aerobic degradation pathway. The crucial differences appear in the mechanisms adopted for the cleavage of the core ring structure of testosterone ( Fig. 1 ). In the aerobic pathway, the cleavage starts with the B-ring, whereas in the anaerobic pathway, the A-ring is opened fi rst. Another signifi cant difference nicely refl ects the infl uence of oxygen availability on the actual ring cleavage mechanism. In contrast to the oxygenase-catalyzed oxygenolytic ring fi ssion under oxic conditions, the opening of the A-ring under anoxic conditions occurs through the oxygen-independent hydrolytic mechanism. The 18 O-incorporation experiments bacterial growth. The bacterial growth was fastest in the presence 20% oxygen in the headspace, whereas the slowest growth was observed under 2.5% oxygen ( Fig. 8A ). S. denitrificans consumed nitrate as the terminal electron acceptor only under strictly anaerobic conditions ( Fig. 8A ). A tiny amount of nitrate was consumed in the other four cultures. The production of ring cleavage intermediates (3,17-DHSA and 2,3-SAOA) by S. denitrifi cans cells in fi ve cultures was quantifi ed using HPLC at 12 h after the addition of 3-chlorocatechol and mercaptopropionic acid. The HPLC detection limits of 2,3-SAOA and 3,17-DHSA concentrations in the bacterial cultures were 100 ng ml Ϫ 1 and 25 ng ml   corroborate the proposed hydrolytic ring cleavage mechanism.
It is well established that the microbial cleavage of the rings of cycloalkane and aromatic compounds under anoxic conditions usually proceeds through hydrolytic mechanism ( 31-34 ). Dangel et al. ( 31,32 ) showed that the hydrolytic ring cleavage substrate in anaerobic cyclohexanol degradation is 1,3-cyclohexanedione. Interestingly, in the anaerobic testosterone catabolism, the hydrolytic ring cleavage substrate, 17-hydroxy-androstan-1,3-dione, has a 1,3-dioxo structure in its A-ring ( Fig. 9 ). For white columns, the amount of 3,17-DHSA produced by S. denitrifi cans grown under 20% oxygen concentration (aerobic treatment) was set at 100%, and those of other four treatments are shown relative to that of the aerobic treatment. For black columns, the amount of 2,3-SAOA produced by S. denitrifi cans grown under strictly anaerobic conditions was set at 100%, and those of other four treatments are shown relative to that of the anaerobic treatment.
Recently, 2-cyclohexenone hydratase catalyzing the addition of water to the C = C bond of ␣ , ␤ -unsaturated carbonyl compounds was purifi ed and characterized from the cyclohexanol-degrading bacterium Alicycliphilus denitrificans ( 35 ). This heterotrimeric enzyme (MhyADH) contains molybdopterin, FAD, and [2Fe-2S] clusters and belongs to the xanthine oxidase family. Our data showed that KCN effectively inhibited the hydration reaction of 1-testosterone, suggesting that this reaction may be catalyzed by a similar molybdopterin-binding enzyme. So far, the ring cleavage enzyme involved in anaerobic cyclohexanol catabolism, 1,3-cyclohexanedione hydrolase, was only partly characterized. 1,2-cyclohexanedione serves as a competitive inhibitor for this enzyme ( 32 ). However, the addition of 1,2-cyclohexanedione (up to 1 mM) to the in vivo or in vitro assays did not inhibit the hydrolytic A-ring cleavage of steroid substrates (data not shown).
A crucial fi nding of our studies is that testosterone is transformed to 1-dehydrotestosterone regardless of the growth conditions. Subsequently, the catabolism proceeds through divergent pathways depending on the availability of oxygen ( Fig. 1 ). Therefore, 1-dehydrotestosterone can be considered a common intermediate or a divergence point for testosterone degradation. According to our current data, S. denitrifi cans adopts only one catabolic pathway (either the 2,3-seco-or 9,10-secopathway) to degrade testosterone, depending on oxygen tension. This mode of catabolism might have an ecological signifi cance. It is more energetically effi cient to start the degradation of a substrate through some common intermediate(s) with the same enzyme(s), regardless of the prevailing conditions. Subsequently, the last common intermediate can be channeled into the relevant pathways. This will help denitrifying bacteria to readily switch their catabolic enzyme inventory between the oxic and anoxic mode and consequently increase their metabolic competence. This hypothesis is supported by the wide distribution of 16S rRNA gene sequences of S. denitrifi cans strains in oxic and anoxic environments. Similar (but not as extreme) cases were reported in the literature for a number of facultative anaerobes, such as Thauera aromatica ( 36 ) and Azoar cus evansii ( 37 ). These bacteria can use benzoate and phenylacetate under oxic and anoxic conditions ( 34,(36)(37)(38). In both cases, the substrate is initially transformed to a common intermediate benzoyl-CoA or phenylacetyl-CoA by the same or isoenyzmes. Then, depending on the availability of oxygen, the common substrate is driven into the relevant pathway. Under oxic conditions, the recently disclosed epoxybenzoyl-CoA or epoxyphenylacetyl-CoA pathway is used; while in the absence of oxygen, the anaerobic benzoyl-CoA is preferred.