An Oxygenase-Independent Cholesterol Catabolic Pathway Operates under Oxic Conditions

Cholesterol is one of the most ubiquitous compounds in nature. The 9,10-seco-pathway for the aerobic degradation of cholesterol was established thirty years ago. This pathway is characterized by the extensive use of oxygen and oxygenases for substrate activation and ring fission. The classical pathway was the only catabolic pathway adopted by all studies on cholesterol-degrading bacteria. Sterolibacterium denitrificans can degrade cholesterol regardless of the presence of oxygen. Here, we aerobically grew the model organism with 13C-labeled cholesterol, and substrate consumption and intermediate production were monitored over time. Based on the detected 13C-labeled intermediates, this study proposes an alternative cholesterol catabolic pathway. This alternative pathway differs from the classical 9,10-seco-pathway in numerous important aspects. First, substrate activation proceeds through anaerobic C-25 hydroxylation and subsequent isomerization to form 26-hydroxycholest-4-en-3-one. Second, after the side chain degradation, the resulting androgen intermediate is activated by adding water to the C-1/C-2 double bond. Third, the cleavage of the core ring structure starts at the A-ring via a hydrolytic mechanism. The 18O-incorporation experiments confirmed that water is the sole oxygen donor in this catabolic pathway.


Introduction
Steroids are ubiquitous and structurally diverse in nature. Cholesterol is an essential structural component of animal cell membranes where it acts as a regulator of membrane fluidity and permeability. In addition, cholesterol serves as a crucial precursor for the biosynthesis of steroid hormones, bile acids, and vitamin D. Plants [1,2] and fungi [3,4] also synthesize small quantities of cholesterol. Although eukaryotes are the main producers of steroids, they lack degradation pathways for recycling the carbon content of these compounds. Hence, the degradation of steroids is dominated by bacteria [5]. Because steroids have limited functional groups, they are usually attacked by bacterial oxygenases using molecular oxygen as a co-substrate [6,7].
The ubiquity and abundance of cholesterol renders the biodegradation of the C 27 sterol a crucial issue in biogeochemistry. In previous years, the microbial transformation of steroids has attracted considerable attention because of its potential effects on biotechnological, pharmaceutical, and clinical applications [8,9]. The investigation of cholesterol-degrading microorganisms began 70 years ago. In 1942, Tak observed that several Mycobacterium species could use cholesterol as their sole carbon and energy source [10]. Subsequent studies detected cholesterol-derived intermediates by growing various Gram-positive and Gramnegative bacteria with cholesterol [11]. The use of metabolic inhibitors such as a,a9, -dipyridyl (a,a9-D) enabled the significant accumulation of cholesterol-derived intermediates including androst-4-en-3,17-dione (AD) and androsta-1,4-diene-3,17-dione (ADD) [12][13][14][15].
In the pioneering studies conducted by Sih et al. [16,17], the side-chain degradation of cholesterol by microbial activities was described. Sih et al. [18][19][20] also established the mechanisms of oxygenolytic cleavage of steroidal rings. Kieslich then proposed a complete, oxygenase-dependent catabolic pathway for cholesterol in 1985 [6]. This pathway is characterized by the cleavage of the steroidal core ring between C-9 and C-10 ( Figure 1A) and is called the 9,10-seco-pathway [21]. Following degradation of the aliphatic side-chain, several oxygenases cleave and degrade the core ring system of C 19 steroid substrates. Introducing a hydroxyl group into ADD results in an extremely unstable intermediate, 9a-hydroxyandrosta-1,4-diene-3,17-dione. This compound thus undergoes simultaneous aromatization of the A-ring and cleavage of the Bring (via a non-enzymatic reaction) to form 3-hydroxy-9,10-secoandrosta-1,3,5(10)-triene-9,17-dione. Further cleavage of the ring system proceeds through a hydroxylation at C-4. The aromatic Aring then splits through the well-known meta-cleavage ( Figure 1A). The aerobic testosterone catabolism of Comamonas testosteroni exhibits similar oxygenolytic ring cleavage mechanisms [7,22]. The 9,10-seco-pathway is the only catabolic pathway for the microbial degradation of steroids described to date.
Recent studies have shown that the acquisition and catabolism of host cholesterol is a crucial process for the persistent infection of Mycobacterium tuberculosis in the lungs of chronically infected animals [23,24]. Other studies have reported the purification and characterization of the key enzymes involved in the 9,10-secopathway [steroid C26-hydroxylase (CYP125) for substrate activation and 3-ketosteroid 9a-hydroxylase (KSH) for oxygenolytic core ring cleavage] from M. tuberculosis and its closely related strains [25][26][27][28].
A few reports have suggested the possibility of alternative catabolic pathways for the aerobic degradation of cholesterol [29,30]. For example, the draft genome sequence of Sterolibacterium denitrificans DSMZ 13999 contains no steroid-transforming oxygenases [30]. This indirect evidence prompted us to study the aerobic cholesterol catabolism by the b-proteobacterium S. denitrificans, which is capable of growing aerobically and anaerobically with cholesterol using oxygen and nitrate as the terminal electron acceptors, respectively [31]. In a previous study, the initial steps of anaerobic cholesterol catabolism by S. denitrificans were investigated, and 25-hydroxycholest-4-en-3-one was the last detected intermediate [32] (for its structure, see Figure 1B). Very similar steps for substrate activation were suggested to occur in aerobic cholesterol catabolism by the same organism [33]. Recently, the molybdoenzyme of S. denitrificans that catalyzes catalyzing the anaerobic hydroxylation of the tertiary carbon (C-25) of C 27 steroid substrates was purified and characterized [30].
Here, we adopted a 13 C-metabolomic approach to detect the 13 C-labeled intermediates involved in the aerobic cholesterol catabolism of S. denitrificans. Many detected intermediates are different from those of the classical 9,10-seco-pathway. Based on the 13 C-metabolomics data and the time course data of cholesterol consumption and intermediates production, this study proposes an alternative cholesterol catabolic pathway, that does not require oxygenases for substrate activation and steroidal core ring cleavage ( Figure 1B). The 18 O-incorporation experiments conducted in this study confirm the O 2 -independent mechanisms.

Results
Cholesterol Catabolism by S. denitrificans is not Inhibited by a,a9-D To investigate the effect of a,a9-D on the cholesterol metabolism of S. denitrificans, we added a,a9-D (5 mM) to the culture after 1 mM cholesterol was consumed. Gordonia cholesterolivorans DSMZ 45229 was also tested for comparison. The addition of a,a9-D to the G. cholesterolivorans culture resulted in the accumulation of AD and ADD, indicating an interruption in the cholesterol catabolic pathway ( Figure 2AII). The cholesterol-derived intermediates detected in the G. cholesterolivorans cultures were summarized in Table 1. HPLC analysis showed that two intermediates exhibited the characteristic maximal UV absorption at approximately 280 nm, indicating the presence of a phenolic A-ring (data not shown). These data indicated that G. cholesterolivorans uses the classical 9,10-seco-pathway to degrade cholesterol. On the contrary, a,a9-D did not inhibit the cholesterol degradation by S. denitrificans ( Figure 2AIV).
Steroid C26-hydroxylase activity was detected in aerobically cholesterol-grown G. cholesterolivorans cells, but not in S. denitrificans cells (Table S1). These results suggested that S. denitrificans may adopt an alternative pathway to degrade cholesterol. This alternative pathway does not require monooxygenase-catalyzed hydroxylations at C-9 and C-26 of steroid substrates.
In vivo Transformation of [4C-13 C]Cholesterol by S. denitrificans Cells The S. denitrificans cells were grown with 1 mM [4C-13 C]cholesterol. The time course of substrate consumption and intermediate production is shown in Figure 3A. The strong negative slope for cholest-4-en-3-one indicates that it is the first accumulated intermediate, which drastically decreased after 2 h of incubation. The strong positive slope for 1,17-dioxo-2,3-seco-androstan-3-oic acid (DSAO) indicates that it is the end product. The ADD and androstan-1,3,17-trione behaved like intermediates between cholest-4-en-3-one and DSAO. The 13 C-labeled intermediates present in the ethyl acetate extracts were detected using ultraperformance liquid chromatography -high-resolution mass spectrometry (UPLC-HRMS), and their mass spectra are given in Figures 3B and S1. We detected a series of C 27 ,C 22 acidic metabolites and C 19 androgens. The acidic intermediates are the same as those shown in Table 1. These results indicate that in S. denitrificans cells, after the oxidation of the A-ring of cholesterol to form a 4-en-3-one structure, a series of b-oxidation and retro-aldol reactions degrade the aliphatic side-chain of cholesterol, as in the 9,10-seco-pathway. We also detected certain 13 C-labeled intermediates that do not occur in the 9,10-seco-pathway (Figures 3BIV, 3BV, and S1A). One of them, identified as 1,17-dioxo-2,3-secoandrostan-3-oic acid by mass and NMR analyses, has an open Aring structure.

Structural Elucidation of a Novel Cholesterol-derived Intermediate
To produce sufficient cholesterol-derived intermediates for NMR analysis, we grew four S. denitrificans (500 ml in 2 l Erlenmeyer flasks) cultures with unlabeled cholesterol (2 mM). After the consumption of 1.5 mM cholesterol, the cholesterolderived intermediates were extracted with ethyl acetate. The separation of ethyl acetate extracts involved silica gel chromatography, TLC, and HPLC. Most cholesterol-derived intermediates (cholest-4-en-3-one, cholest-4-en-3-one-26-oic acid, pregn-4-enone-20-carboxylic acid, AD, ADD, androst-1-en-3,17-dione, and 1-hydroxyandrostan-3,17-one) were identified by reference to the TLC, HPLC, UV absorption, and UPLC-HRMS behavior of authentic steroid standards. Compound 1, an unprecedented cholesterol-derived intermediate, was isolated as a white powder. The structural elucidation of this compound relied mainly on mass and NMR spectra (see Figure S4 for original NMR spectra). The ESI-mass spectrum of compound 1 showed a sodium adduct ion of 343.1876 Da ( Figure S3A). In addition, its pseudo-molecular ions [M -2H 2 O+H] + and [M -H 2 O+H] + at m/z 285.1862 and 303.1962, respectively, appeared in the APCI-mass spectrum ( Figure S3B). Its molecular formula is thus deduced as C 19 (Table S2). The connectivity of its C-8 and C-9 was approached by the HMBC spectrum ( Figure  S4D), in which the cross-peaks of H-8/C-9 and H-9/C-8 were observed. Together, the interpretations of distinctive cross-peaks in the COSY and HMBC spectra ( Figure 4A) allowed the assignment of compound 1 as a 2,3-seco-androstane skeleton with one hydroxyl at C-3 and three ketone functionalities at C-1, C-3, and C-17, respectively. Accordingly, compound 1 was characterized as the shown structure, and named as 1,17-dioxo-2,3-seco-androstan-3oic acid ( Figure 4B).

A Hydrolytic Ring Cleavage Mechanism Is Adopted by S. denitrificans to Degrade Cholesterol
In the classical 9,10-seco-pathway, oxygenases catalyze oxygenolytic ring fission using molecular oxygen as the co-substrate. To determine the origin of the oxygen atoms at C-1 and C-3 of 1,17dioxo-2,3-seco-androstan-3-oic acid (DSAO), we conducted three in vitro transformation assays using 1-testosterone (which has two 16   . These data indicate that in the alternative pathway, after the activation of the A-ring through a hydration reaction, the cleavage of the core ring system of cholesterol begins with the Aring by a hydrolysis reaction.

An Alternative Cholesterol Catabolic Pathway is Present in S. denitrificans
In this study, we used 13 C-labeled cholesterol as a tracer to identify the downstream C 27 ,C 19 intermediates (from 26hydroxycholest-4-en-3-one to DSAO, see Figure 1B for their structures) of the aerobic cholesterol catabolic pathway. According to the 13 C-metabolomics and 18 O-incorporation data, we demonstrate that S. denitrificans adopts an oxygenase-independent strategy to degrade cholesterol under oxic conditions. In the proposed pathway, 1,17-dioxo-2,3-seco-androstan-3-oic acid (the product of core ring cleavage) serves as the key intermediate. This compound has never been reported to be involved in aerobic steroid catabolism. Since the 9,10-seco-pathway was established thirty years ago [6], this is the first documented case that clearly demonstrates the existence of an alternative cholesterol catabolic pathway in bacteria. The proposed alternative pathway is significantly different from the 9,10-seco-pathway.
Comparison with the Classical 9,10-seco-pathway In the alternative catabolic pathway, the side-chain degradation also precedes the core ring cleavage ( Figure 1B). However, the alternative pathway differs from the established 9,10-seco-pathway in the mechanisms of substrate activation and core ring cleavage. In the in vivo [4C-13 C]cholesterol biotransformation assay, both 13 C-labeled 25-hydroxycholest-4-en-3-one and 26-hydroxycholest-4-en-3-one could be detected. The NADH-dependent steroid C26-hydroxylase was purified from Rhodococcus jostii [28] and Mycobacterium tuberculosis [34]. However, neither the corresponding gene [30] nor the enzyme activity (this study) of steroid C26hydroxylase could be detected in S. denitrificans cells. Moreover, we confirmed that the terminal hydroxyl group of 26-hydroxycholest-4-en-3-one produced by S. denitrificans originates from water. These data indicate that in the aerobic cholesterol catabolism by S. denitrificans, 26-hydroxycholest-4-en-3-one may be produced from 25-hydroxycholest-4-en-3-one through a novel isomerization reaction. Similar isomerization reactions of hydroxyl groups occur in monoterpene metabolism [35,36]. Researchers have recently isolated and characterized linalool dehydratase-isomerase [37]. This enzyme catalyzes the migration of a hydroxyl group from a tertiary carbon to a primary one. A similar enzyme might catalyze the isomerization of 25-hydroxycholest-4-en-3-one to a 26hydroxyl structure.
After substrate activation by adding a hydroxyl group at C-26, the aliphatic side-chain of the C 27 steroid substrates is degraded through a series of retro-aldol and b-oxidation reactions to form C 24 , C 22 acidic intermediates and C 19 androgens, which were observed in cholesterol-grown S. denitrificans cultures. These sidechain-degrading reactions do not require oxygenases. In contrast to the 9,10-seco-pathway, the proposed catabolic route applies no oxygenases for ring fission ( Figure 1B). The core ring structure opens first at the A-ring through a hydrolytic mechanism. The 18 O-incorporation experiments corroborate the proposed hydrolytic ring cleavage mechanism.
Recently, new pathways for the degradation of aromatic compounds under oxic conditions were unraveled [38,39]. These pathways operate primarily in facultative anaerobes and use a hydrolytic mechanism to open the ring of the substrates. However, these aerobic pathways still employ monooxygenases to introduce hydroxyl groups into the aromatic ring for substrate activation. In contrast, the proposed aerobic cholesterol degradation pathway does not require any oxygenases-catalyzed reactions till the stage at which the steroidal A-ring opens.
Potential Ecological Significance of the Proposed Cholesterol Degradation Pathway S. denitrificans can degrade cholesterol regardless of the presence of oxygen. Several lines of evidence suggest that very similar metabolic strategies may be adopted by S. denitrficans to degrade cholesterol under oxic and anoxic conditions. First, previous proteome analyses revealed no apparent differences in soluble protein patterns of anaerobically and aerobically grown S. denitrificans cells [33]. Second, the steroid-transforming enzymes involved in the initial steps of anaerobic cholesterol metabolism by S. denitrificans are not oxygen-labile in vivo [30,32]. Third, the O 2dependent steroid-transforming enzymes, including steroid C26hydroxylase and 3-ketosteroid 9a-hydroxylase, are not detected in aerobically cholesterol-grown S. denitrificans cells (this study).
It is tempting to speculate that S. denitrificans have developed efficient mechanisms to profit from the available carbon sources regardless of the prevailing redox state. The first adaptive mechanism could be the ability to initiate the degradation of steroid substrates under both oxic and anoxic conditions via similar reactions and intermediates [33]. The second mechanism involves the adoption of oxygenase-independent aerobic catabolic pathways. Both mechanisms would enhance the metabolic competence of these organisms because they can switch quickly between aerobic and anaerobic metabolic modes. Moreover, under micro-aerobic conditions, when the oxygen tension becomes insufficient, the organisms can channel the oxygen flux to the respiratory electron transport chain, and still profit from the steroid substrate through oxygenase-independent catabolic pathways that do not consume molecular oxygen.
Recently, very similar ring cleavage mechanisms were observed in anaerobic testosterone degradation by a c-proteobacterium, Steroidobacter denitrificans DSMZ 18526 [40]. Interestingly, the bacterial strain uses the 9,10-seco-pathway to degrade testosterone when oxygen is available. These data indicated that bacteria adopt the oxygenase-independent 2,3-seco-pathway to degrade steroids not only under anaerobic conditions [40], but also under aerobic conditions (at least in this case). So far, less is known about the enzymes (especially the A-ring-cleavage enzyme) and their corresponding genes involved in the 2,3-seco-pathway. Therefore, in situ 13 C-metabolomics seems to be a feasible approach to investigate the contribution of the 2,3-seco-pathway in the degradation of cholesterol and other steroids in natural environments and engineered systems.

Conclusions
The results of this study demonstrate that microbial degradation of one substrate can proceed via different mechanisms under the same conditions. The cholesterol degradation pathway proposed in this study further underpins the diversity of microbial catabolism of organic compounds. It also broadens our understanding of the strategies that microorganisms use to cope with and adapt to environmental conditions and challenging inert substrates such as steroids.  Table S2. doi:10.1371/journal.pone.0066675.g004

Chemicals and Bacterial Strains
The [4C- 13

Fed-batch Growth of S. denitrificans with Unlabeled Cholesterol
In this study, 0.5% of hydroxypropyl-b-cyclodextrin was always added to the bacterial cultures to improve the solubility of cholesterol in media. S. denitrificans was grown in phosphatebuffered shake-flask cultures (500 ml in 2 l Erlenmeyer flasks) containing 2 mM cholesterol. The culture was incubated at 28uC in an orbital shaker (180 rpm). In 1 l of distilled water, the medium contained the following: 0.77 g cholesterol, 5 g hydroxypropyl-b-cyclodextrin, 1.0 g NH 4 Cl, 0.5 g MgSO 4 ?7 H 2 O, and 0.1 g CaCl 2 ?2H 2 O. After autoclaving, sterile 50 ml KH 2 PO 4 -K 2 HPO 4 buffer solution (1 M, pH 7.0), vitamins (1 ml l 21 ) [42], EDTA-chelated mixture of trace elements (1 ml l 21 ) [43], and selenite and tungstate solution (1 ml l 21 ) [44] were added. The amounts of residual cholesterol in the cultures were monitored using HPLC. After the consumption of 1.5 mM cholesterol, the pH of the cultures was adjusted to pH ,2 using 5M HCl. The acid-treated cultures were extracted 3 times with the same volume of ethyl acetate to recover the residual cholesterol and its Cholesterol Catabolism by S. denitrificans in the Presence of a,a9-D S. denitrificans was grown in two phosphate-buffered shake-flask cultures (50 ml in 250 ml-Erlenmeyer flasks) containing 2 mM cholesterol at 28uC with shaking. After the consumption of 1 mM cholesterol, 5 mM of a,a9-D (an inhibitor of 3-ketosteroid 9ahydroxylase [12][13][14][15]) was added to one of the culture, and the incubation of both cultures continued for 16 h. The pH of the cultures was subsequently adjusted to pH ,2, and ethyl acetate was used to extract cholesterol-derived neutral and acidic intermediates. Cholesterol metabolism by G. cholesterolivorans DSMZ 45229 was also studied using the same procedure for comparison. The four ethyl acetate extracts were analyzed using UPLC-HRMS.
Effect of tert-Butyl Alcohol on Cholesterol Catabolism of S. denitrificans S. denitrificans was aerobically grown in three phosphate-buffered shake-flask cultures (50 ml) containing 2.5 mM cholesterol. After the consumption of 2 mM cholesterol, 2.5% and 5% (v/v) tertbutyl alcohol (an analog of 25-hydroxycholest-4-en-3-one) was individually added to two cultures. 2-Propanol was then added to three cultures to bring the final alcohol concentration to 5% (v/v) in all cultures. The incubation of the three cultures continued further 16 h. The pH of the cultures was adjusted to pH ,2, and ethyl acetate was used to extract cholesterol-derived intermediates. The three ethyl acetate extracts were analyzed using UPLC-HRMS.

Fed-batch Growth of S. denitrificans with [4C-13 C]Cholesterol
A S. denitrificans culture (500 ml) was first grown with 2 mM of unlabeled cholesterol in a 2 l Erlenmeyer flask. After the unlabeled cholesterol was completely consumed, 50 ml of the stock culture was transferred into a sterile 250-ml Erlenmeyer flask. The S. denitrificans cells were subsequently fed with 1 mM [4C-13 C]cholesterol and incubated at 28uC with shaking (180 rpm). Estrone (0.1 mM) which cannot be utilized by S. denitrificans as a carbon and energy source was added as an internal control. Samples (3 ml) were withdrawn every two hours. Culture samples (0.1 ml 63) were centrifuged at 10,0006g for 10 min to harvest the S. denitrificans cells. The protein content in the pellet was determined using bicinchoninic acid (BCA) assay. The residual culture samples (2.7 ml) were acidified to pH ,2, and extracted three times with the same volume of ethyl acetate to recover cholesterol-derived intermediates. The ethyl acetate fractions were combined, the solvent was evaporated, and the residue was re-dissolved in 300 ml of methanol. The [4C-13 C]cholesterol-derived intermediates in 60 ml samples were identified using ultra-performance liquid chromatography -high-resolution mass spectrometry (UPLC-HRMS). The amount of residual cholesterol and cholesterolderived intermediates in the samples (80 ml 63) was determined using HPLC.

C]Testosterone
In another 50 ml in vivo biotransformation assay, S. denitrificans (50 ml culture in a 250-ml Erlenmeyer flask) transferred from the same stock culture was fed with 1 mM [2,3,4C-13 C]testosterone. The samples (3 ml) were withdrawn after 16h incubation. The pH of the culture samples was adjusted to pH ,2 using 5M HCl. The ethyl acetate -extractable samples were analyzed using UPLC-HRMS.

Preparation of Cell Extracts
The S. denitrificans cultures (500 ml in 2 l Erlenmeyer flasks) were grown with 2 mM cholesterol with shaking (180 rpm). Cells were harvested by centrifugation in the exponential growth phase at OD 600 of 0.8,1.0 (optical path 1 cm) and the cell pellet was then stored at 280uC. All steps used for preparation of cell extracts were performed at 4uC. Frozen cells were suspended in twice the volume of 150 mM Tris-HCl buffer (pH 7) containing 0.1 mg of DNase I ml 21 . Cells were broken by passing the cell suspension through a French pressure cell (Thermo Fisher Scientific) twice at 137 MPa. The cell lysate was fractionated using two steps of centrifugation: the first step involved centrifugation for 30 min at 20,0006g to remove the cell debris, unbroken cells and residual cholesterol. The supernatant (crude cell extract) was then centrifuged at 100,0006g for 1.5 h to separate soluble proteins from membrane-bound proteins.

In vitro 18 O-Incorporation Assays for the Production of 1,17-Dioxo-2,3-seco-androstan-3-oic Acid
To determine the origins of the oxygen atoms at C-1 and C-3 of 1,17-dioxo-2,3-seco-androstan-3-oic acid, three in vitro assays were performed in an anaerobic chamber containing 95% N 2 and 5% H 2 (1atm). The three reaction mixtures (3 ml for each assay) were incubated at 30uC for 16 h with shaking. After the acidic treatment, the steroid products were extracted from the assays using ethyl acetate, and the extracts were analyzed using UPLC-APCI-mass spectrometry.
(1) Control assay. A 3-ml reaction mixture containing 50 mM Tris-HCl buffer (pH 7) and soluble proteins (15 mg) of S. denitrificans were sealed in a 10-ml glass bottle with a rubber stopper. The reaction was started by adding 200 ml of 67.5 mM 1testosterone solution (in 2-propanol) to the assay. The final concentration of the steroid substrate in the reaction mixture was 4.5 mM. The final 2-propanol content was 6.67%.
(2) 18O2-Treated assay. 1.8 ml of 18 O 2 gas (99 atom %) was introduced into an anaerobic glass bottle containing 7 ml of headspace (95% N 2 and 5% H 2 , 1atm) and 3 ml reaction mixture containing the same components as the control assay. The final 18 O 2 concentration in the headspace was ,20%.
(3) 18O-Labeled Water-treated assay. A 2.0 ml sample of 18 O-labeled water (97 atom %) was added to 1.0 ml of 150 mM Tris-HCl buffer (pH 7) containing soluble proteins of S. denitrificans (15 mg). The final 18 O-water content was approximately 64.7%. The reaction was started by adding 4.5 mM of 1-testosterone to the anoxic assay. The 2-propanol content was also 6.67%.

Activity Assays for Steroid C26-Hydroxylase (Cyp125)
The Cyp125 activity of G. cholesterolivorans and S. denitrificans was measured by monitoring the product (26-hydroxycholest-4-en-3one) concentration using a Hitachi HPLC module. The reaction mixture (1 ml) contained an air-saturated 100 mM potassium phosphate buffer (pH 7.0), 5 mM NADH, 0.5 mM cholest-4-en-3one, 5% (w/v) 2-hydroxypropyl-b-cyclodextrin, and soluble proteins (20 mg) precipitated at 50% ammonium sulfate saturation. In the anaerobic assays, 2 mM 1,4-dithiothreitol was added to remove residual O 2 present in the reaction mixture (1 ml), which was prepared in an anaerobic chamber containing 95% N 2 and 5% H 2 (1atm). The aerobic and anaerobic assays were incubated at 30uC for 16 h with shaking. The reaction was stopped by the addition of 20 ml of 25% HCl, and the steroids were extracted using ethyl acetate.

In vitro 18 O-Incorporation Assays for the Production of 26-Hydroxycholest-4-en-3-one
To determine the origins of the oxygen atoms at C-26 of 26hydroxycholest-4-en-3-one, three in vitro assays were performed in an anaerobic chamber. In all assays, 20 ml of 100 mM cholest-4en-3-one solution (in 2-propanol) was added to empty glass bottles (3-ml). After complete evaporation of the solvent, the reaction mixture (1 ml) was dispensed anaerobically. The three reaction mixtures (1 ml for each assay) were incubated at 30uC for 16 h with shaking. The ethyl acetate extracts were analyzed using UPLC-APCI-mass spectrometry.
(1) Control Assay. In a 3-ml glass bottle sealed with a rubber stopper, the 1-ml reaction mixture contained 50 mM

Silica Gel Chromatography
A 385 ml silica gel column (5563 cm; SiliaFlashH P60, Silicycle) was equilibrated with 2 bed volumes of n-hexane -ethyl acetate (65:35, v/v). The ethyl acetate extract (approximately 400 mg dissolved in 3 ml ethyl acetate) containing cholesterol-derived intermediates was loaded into the column and eluted with the equilibrium solvent system at a flow rate of 2 ml min 21 . The eluate was collected in 5-ml fractions, and a 0.5 ml sample was taken from each fraction. The solvent was evaporated to dryness, and the residue was re-dissolved in 10 ml of methanol. The samples were analyzed using TLC. The fractions containing the same compounds were pooled and evaporated to dryness, and 200 ml of methanol was used to dissolve the residue. Further purification of cholesterol-derived intermediates was performed using TLC.

High-Performance Liquid Chromatography (HPLC)
A reversed-phase Hitachi HPLC system was used for the final separation. The separation was achieved on an analytical RP-C 18 column (Luna 18(2), 5 mm, 15064.6 mm; Phenomenex) incubated at 35uC. The mobile phase included a mixture of two solvents: A (0.1% aqueous trifluoroacetic acid) and B (methanol containing 0.1% trifluoroacetic acid). The C 27 steroids were eluted at a flow rate of 1.0 ml/min with a gradient from 80%-90% B over 5 min, followed by isocratic elution at 90% B for 10 min, a gradient from 90%-100% B for 5 min, and further isocratic elution for 20 min. The separation of C 19 steroids were performed at a flow rate of 1.0 ml/min with a gradient from 40%-65% B within 50 min. The steroid products were detected in the range of 200-300 nm using a photodiode array detector. The structures of HPLC-purified intermediates were elucidated using NMR spectroscopy and mass spectrometry. In addition, HPLC was used for the quantification of some steroid substrates and intermediates present in the S. denitrificans cultures. The quantification of steroids (cholesterol, cholest-4-en-3-one, ADD, androstan-1,3,17-trione, and 1,17dioxo-2,3-seco-androstan-3-oic acid) was calculated from their respective peak areas using a standard curve of individual standards. The R 2 values for the standard curves were .0.98. Data are averages of three measurements.

Ultra-Performance Liquid Chromatography-Atmospheric Pressure Chemical Ionization-High-Resolution Mass Spectrometry (UPLC-APCI-HRMS)
The ethyl acetate extractable samples or HPLC-purified steroid intermediates were analyzed using UPLC-MS with UPLC coupled to an APCI-mass spectrometer. 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 reversed-phase C 18 column (Acquity UPLCH BEH C18, 1.7 mm, 10062.1 mm; Waters) with a flow rate of 0.3 ml min 21 at 50uC (column oven temperature). The mobile phase comprised a mixture of two solvents: Solvent A (2% (v/v) acetonitrile containing 0.1% formic acid to enable excellent ionization in the APCI) and Solvent B (90% isopropanol containing 0.1% formic acid). Separation was achieved with a linear gradient of Solvent B from 30% to 90% in 12 min. In APCI-MS analysis, the temperature of the ion source was maintained at 100uC. Nitrogen desolvation gas was set at a flow rate of 500 l h -1 and the probe was heated to 400uC. Nitrogen served as the APCI nebulizer gas. The corona current was maintained at 20 mA, and the electron multiplier voltage was set to1700 eV. The parent scan was in the range of 50-500 m/z. The predicted elemental composition of individual intermediates was calculated using MassLynx TM Mass Spectrometry Software (Waters).

Ultra-Performance Liquid Chromatography-Electrospray Ionization-High-Resolution Mass Spectrometry (UPLC-ESI-HRMS)
The ethyl acetate extractable samples and HPLC-purified 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 50-500 m/z. The capillary voltage was set at 3000 V; the source and desolvation temperatures were 100uC and 250uC, respectively. The cone gas flow rate was 50 l h 21 .

NMR Spectroscopy
The 1 H-and 13 C-NMR spectra were recorded at 27uC using a Bruker AV600_GRC 600MHz NMR. Chemical shifts (d) were recorded and shown as ppm values with deuterated methanol (99.8%, 1 H: d = 3.31 ppm; 13 C: d = 49.0 ppm) as the solvent and internal reference. Figure S1 High-resolution mass spectra of other 13