ABSTRACT
Bacterial steroid degradation has been studied mainly with Rhodococcus equi (Nocardia restrictus) and Comamonas testosteroni as representative steroid degradation bacteria for more than 50 years. The primary purpose was to obtain materials for steroid drugs, but recent studies showed that many genera of bacteria (Mycobacterium, Rhodococcus, Pseudomonas, etc.) degrade steroids and that steroid-degrading bacteria are globally distributed and found particularly in wastewater treatment plants, the soil, plant rhizospheres, and the marine environment. The role of bacterial steroid degradation in the environment is, however, yet to be revealed. To uncover the whole steroid degradation process in a representative steroid-degrading bacterium, C. testosteroni, to provide basic information for further studies on the role of bacterial steroid degradation, we elucidated the two indispensable oxidative reactions and hydration before D-ring cleavage in C. testosteroni TA441. In bacterial oxidative steroid degradation, A- and B-rings of steroids are cleaved to produce 2-hydroxyhexa-2,4-dienoic acid and 9,17-dioxo-1,2,3,4,10,19-hexanorandrostan-5-oic acid. The latter compound was revealed to be degraded to the coenzyme A (CoA) ester of 9α-hydroxy-17-oxo-1,2,3,4,5,6,10,19-octanorandrostan-7-oic acid, which is converted to the CoA ester of 9,17-dioxo-1,2,3,4,5,6,10,19-octanorandrostan-7-oic acid by ORF31-encoded hydroxylacyl dehydrogenase (ScdG), followed by conversion to the CoA ester of 9,17-dioxo-1,2,3,4,5,6,10,19-octanorandrost-8(14)-en-7-oic acid by ORF4-encoded acyl-CoA dehydrogenase (ScdK). Then, a water molecule is added by the ORF5-encoded enoyl-CoA hydratase (ScdY), which leads to the cleavage of the D-ring. The conversion by ScdG is presumed to be a reversible reaction. The elucidated pathway in C. testosteroni TA441 is different from the corresponding pathways in Mycobacterium tuberculosis H37Rv.
IMPORTANCE Studies on representative steroid degradation bacteria Rhodococcus equi (Nocardia restrictus) and Comamonas testosteroni were initiated more than 50 years ago primarily to obtain materials for steroid drugs. A recent study showed that steroid-degrading bacteria are globally distributed and found particularly in wastewater treatment plants, the soil, plant rhizospheres, and the marine environment, but the role of bacterial steroid degradation in the environment is yet to be revealed. This study aimed to uncover the whole steroid degradation process in C. testosteroni TA441, in which major enzymes for steroidal A- and B-ring cleavage were elucidated, to provide basic information for further studies on bacterial steroid degradation. C. testosteroni is suitable for exploring the degradation pathway because the involvement of degradation-related genes can be determined by gene disruption. We elucidated the two indispensable oxidative reactions and hydration before D-ring cleavage, which appeared to differ from those present in Mycobacterium tuberculosis H37Rv.
INTRODUCTION
The mechanism by which testosterone is degraded by the representative steroid-degrading bacteria Rhodococcus equi (formerly Nocardia restrictus) and Comamonas testosteroni (formerly Pseudomonas testosteroni) has been intensively studied for the primary purpose of obtaining materials for steroid drugs, and the major intermediate compounds in the A- and B-ring degradation process were identified in approximately 1960 (1–10).
For C. testosteroni, the enzymes catalyzing the early steps of steroid degradation (17β-dehydrogenase, 3α-dehydrogenase, 3-oxo-Δ5-steroid isomerase, Δ1-dehydrogenase, and Δ4-dehydrogenase) were identified in the 1990s (11–25).
C. testosteroni TA441 degrades certain steroids into 2-hydroxyhexa-2,4-dienoic acid and 9,17-dioxo-1,2,3,4,10,19-hexanorandrostan-5-oic acid (compound I) via aromatization of the A-ring and subsequent hydrolysis (26–34). Coenzyme A (CoA) is incorporated into compound I by the ORF18-encoded enzyme (ScdA) (35) (Fig. 1), and the resultant CoA ester of compound I is further degraded via β-oxidation. A double bond is introduced at the C-6 position of the CoA ester of 9-hydroxy-17-oxo-1,2,3,4,10,19-hexanorandrostan-5-oic acid (compound II) by ScdC1C2 to produce the CoA ester of 9-hydroxy-17-oxo-1,2,3,4,10,19-hexanorandrost-6-en-5-oic acid (compound III) (36), followed by the addition of a water molecule at the C-7 position by ScdD (37). Further degradation processes were partially identified in TA441 (compounds IV to VI and scdL1L2 and scdN in Fig. 1) (38, 39). Degradation of C- and D-rings is also reported for Mycobacterium tuberculosis H37Rv (40), but the whole degradation pathway has not been elucidated.
Metabolites and the genes in steroid degradation in Comamonas testosteroni TA441 identified in our previous studies. *, KshA and HsaC are putative orthologs in Mycobacterium, which were used to examine the distribution of bacterial steroid degradation genes in the environment (45). R1/R2 is a hydroxyl moiety when the initial steroid has a hydroxyl moiety at the corresponding position (e.g., cholic acid). These hydroxyl moieties remain at least before the addition of CoA by ScdA. The hydroxyl moiety at R1 is mostly removed before the reaction by ScdC1C2 (35). The hydroxyl moiety at R2 basically remains to directly produce a bracketed compound produced after CoA esters of compound III (35), but a portion is dehydrated because compound III is detected when cholic acid is used as the initial steroid (cf. Fig. 2). The compounds are 9,17-dioxo-1,2,3,4,10,19-hexanorandrostan-5-oic acid (compound I), 9-hydroxy-17-oxo-1,2,3,4,10,19-hexanorandrostan-5-oic acid (compound II), 9-hydroxy-17-oxo-1,2,3,4,10,19-hexanorandrost-6-en-5-oic acid (compound III), 9-oxo-1,2,3,4,5,6,10,19-octanor-13,17-secoandrost-8(14)-ene-7,17-dioic acid (compound IV), 4-methyl-5-oxo-octane-1,8-dioic acid (compound V), and 4-methyl-5-oxo-oct-2-ene-1,8-dioic acid (compound VI). Structures of the bracketed compounds are speculated based on the predicted function of the genes, and the last two are also supported by the molecular weight detected by UHPLC/MS. (Adapted from references 38 and 39 with permission of Elsevier.)
Bacterial steroid degradation has been reported for other genera of bacteria, especially Mycobacterium (41, 42), Rhodococcus (43), and Pseudomonas (44). Recently, steroid-degrading bacteria were shown to be globally distributed; they are found particularly in wastewater treatment plants, the soil, plant rhizospheres, and the marine environment, using the 3-ketosteroid 9α-hydroxylase gene (kshA in Mycobacterium, corresponding to ORF17 in TA441) (Fig. 1) and the 3,4-dihydroxy-9,10-seconandrosta-1,3,5(10)-triene-9,17-dione dioxygenase gene (hsaC in Mycobacterium, corresponding to tesB in TA441) as markers (45). The existence of steroid-degrading bacteria in various environments implies that they might play significant roles in natural circulation. The role of bacterial steroid degradation in the environment is, however, yet to be revealed. Also, bacterial steroid degradation may influence the bacterial flora in intestine, which is abundant with bile acids and is deeply related to human health. To enable exploration of the whole degradation pathway in future studies on bacterial steroid degradation, we hereby present metabolites and the genes involved in the reactions necessary before the cleavage of the steroidal D-ring in C. testosteroni TA441, whose degradation pathway differs from those proposed for M. tuberculosis H37Rv (40).
RESULTS
Isolation and identification of intermediate compounds accumulated by the ORF5-null mutant incubated with cholic acid.Comamonas testosteroni TA441 degrades steroids to 9,17-dioxo-1,2,3,4,10,19-hexanorandrostan-5-oic acid (compound I), and the CoA ester of compound I is presumed to be degraded via a β-oxidation involving scdL1 to ORF33 (Fig. 1). To characterize the detailed degradation pathway of compound I, an ORF5-null mutant was incubated with each one of the cholic acid analogs (cholic acid, chenodeoxycholic acid, deoxycholic acid, and lithocholic acid) and the cultures were analyzed by high-performance liquid chromatography (HPLC) at suitable intervals. The HPLC chromatogram of the culture with cholic acid is shown in Fig. 2, and the others are shown in Fig. S1 in the supplemental material. Major metabolites were the same among all the cultures; new compounds IX to XII were detected with previously identified metabolites, 9,17-dihydroxy-1,2,3,4,10,19-hexanorandrost-6-en-5-oic acid (compound VII; retention [RT], 5.7 min), compound III (RT, 6.9 min), and 9,17-dioxo-1,2,3,4,10,19-hexanorandrosta-6,8(14)-dien-5-oic acid (compound VIII; RT, 7.2 min).
Three-dimensional HPLC analysis of cultures of the ORF5-null mutant incubated in 50% LB plus 50% C-medium with 0.1% (wt/vol) cholic acid for 8 days. Compounds IX to XII were isolated for identification. Compounds III, VII, and VIII were identified in previous studies. The vertical axis indicates wavelength, and the horizontal axis indicates RT; the UV absorbance of each compound is plotted in contours.
From the total 500-ml culture volume of the ORF5-null mutant incubated with 0.1% (wt/vol) cholic acid, compounds X (RT, 5.7 min) and XII (RT, 6.9 min) were extracted under neutral conditions and compounds IX (RT, 5.8 min) and XI (RT, 6.2 min) were extracted under acidic conditions. Fractions containing each of these compounds were purified by HPLC. However, compound XI was readily converted to compound XII during storage in the freezer (−20°C) (Fig. S2). Compounds XI and XII were purified again by HPLC and subjected to further analysis. The amounts of purified compounds IX, X, XI, and XII isolated from the culture and compound XII converted from compound XI were 50 mg, 2.7 mg, 33 mg, 25 mg, and 30 mg, respectively. In addition to these compounds, several other putative metabolites were isolated, but the amount was not sufficient for identification. Compounds IX, X, and XII were directly subjected to nuclear magnetic resonance (NMR) and electron ionization mass spectrometry (EI-MS) analysis. Compound XI was immediately subjected to NMR analysis after purification, and data were successfully obtained. Subsequent EI-MS analysis was not successful, so the sample was methylated and the EI-MS data of the purified methyl ester of XI were obtained. The 1H NMR and 13C NMR data of the isolated compounds are shown in Table S1, and the structures are presented in Fig. 3.
Metabolites isolated from cultures of the ORF5-null mutant incubated with cholic acid. Compounds are 9α-hydroxy-17-oxo-1,2,3,4,5,6,10,19-octanorandrostan-7-oic acid (compound IX), 17β-hydroxy-1,2,3,4,5,6,7,10,19-nonanorandrost-8(14)-en-9-one (compound X), 9,17-dioxo-1,2,3,4,5,6,10,19-octanorandrostan-8(14)-en-7-oic acid (compound XI), and 1,2,3,4,5,6,7,10,19-nonanorandrost-8(14)-ene-9,17-dione (compound XII) (Hajos-Parrish ketone [47]).
The molecular formula of compound IX was determined to be C11H16O4 (found m/z 212.1047 M+, calculated m/z 212.1049). From the assigned NMR data of compound IX, H-8 showed a large vicinal coupling constant (13.1 Hz) with H-14 and a small constant (2.7 Hz) with H-9, and H-9 showed three small vicinal coupling constants (3.0, 2.7, and 2.7 Hz). These results indicated that the hydrogen at C-8 was oriented β-axial and the hydrogen at C-9 was oriented β-equatorial. Compound IX was identified as 9α-hydroxy-17-oxo-1,2,3,4,5,6,10,19-octanorandrostan-7-oic acid.
The molecular formula of compound X was determined to be C10H14O2 (found m/z 166.0996 M+, calculated m/z 166.0994). NMR data of compound X showed the presence of a conjugated ketone and a secondary alcohol. By comparison with the data of Tang et al. (46), compound X was identified as 17β-hydroxy-1,2,3,4,5,6,7,10,19-hexanorandrost-8(14)-en-9-one.
The molecular formula of the methyl ester of compound XI was determined as C12H14O4 (found m/z 222.0914 M+, calculated m/z 222.0892). NMR data of compound XI showed the presence of an isolated ketone (δ 214.2) moiety and an unstable α,β-unsaturated ketone (δ 201.9) moiety with a carboxylic group at the α-position. Compound XI was identified as 9,17-dioxo-1,2,3,4,5,6,10,19-octanorandrost-8(14)-en-7-oic acid.
The molecular formula of compound XII was determined to be C10H12O2 (found m/z 164.0846 M+, calculated m/z 164.0837). All the NMR signals of XII were assigned as presented in Table S1, showing consistency with the decarboxylated compound from compound XI, and compound XII was identified as 1,2,3,4,5,6,7,10,19-nonanorandrost-8(14)-ene-9,17-dione [Hajos-Parrish ketone, 7α-methyl-2,3,7,7α-tetrahydro-1H-indene-1,5(6H)-dione] (47). Since compound XII was produced by decarboxylation of compound XI, the total amount of compound XI in the culture was estimated to be at least approximately 90 mg, suggesting that compound XI is the major metabolite accumulated by the ORF5-null mutant. A homology search indicated that the ORF5-encoded enzyme was enoyl-CoA hydratase. Therefore, the CoA ester of compound XI is presumed to be the dominant metabolite in the culture of the ORF5-null mutant.
Complementation experiments involving ORF5.To determine whether ORF5 was indispensable for the conversion of the compound XI CoA ester, complementation experiments with an ORF5-null mutant were carried out. An ORF5-null mutant carrying the broad-host-range plasmid pMFY42 (negative control) and an ORF5-null mutant carrying a pMFY42-based plasmid with ORF5 (pMFYORF5) were constructed and incubated with androsta-1,4-dien-3,17-dione (ADD) (ADD was used in all the experiments after identification of metabolites because it was necessary to use a steroid compound without a hydroxyl moiety at C-7 for the experiments described in the last section of Results). Ultrahigh-performance liquid chromatography-mass spectrometry (UHPLC/MS) analysis demonstrated the accumulation of compound IX (RT, approximately 0.52 min) and compound XI (RT, approximately 0.57 min) in the culture of the ORF5-null mutant carrying pMFY42 together with a compound detected as a peak of m/z 209 (RT, approximately 0.57 min) (compound XIII), whose CoA ester was likely to be a metabolite produced during the conversion of the compound IX CoA ester to the compound XI CoA ester (see Fig. S3-1; compounds were detected at an RT approximately 0.4 min earlier in this analysis than others). The reason for this is not clear, but a large amount of compounds I, II, and III accumulated in the culture of the ORF5-null mutant carrying pMFY42. Most of compounds I, II, and III were degraded by complementation in the culture of the ORF5-null mutant carrying pMFYORF5 (Fig. S3-2, MS), but the amount of compounds IX, XI, and XIII in the culture of the ORF5-null mutant with pMFYORF5 was almost the same as or more than in the culture of ORF5-null mutant carrying pMFY42 (Fig. S3-2, m/z 211, 209, 207, and 163). Compound V (molecular weight, 202), a metabolite produced after the cleavage of all the rings (Fig. 1), was detected in the culture of the ORF5-null mutant carrying pMFYORF5 but not in the culture of the ORF5-null mutant carrying pMFY42 (Fig. S3-1 and S3-2, m/z 201), suggesting the involvement of the ORF5-encoded enzyme in the conversion of the compound XI CoA ester. Therefore, we constructed an ORF5-null mutant carrying a pMFY42-based plasmid with four copies of ORF5 (pMFYORF5m) and analyzed the culture in the same way. Compound XI was almost undetectable in the culture of the ORF5-null mutant carrying pMFYORF5m, and the amount of compounds IX and XIII was also smaller than the amount in the cultures of the other two ORF5− mutants (Fig. S3-3). The results indicated the involvement of ORF5 in conversion of the compound XI CoA ester.
Involvement of ORF4 and ORF31 in the conversion of compound IX and XIII CoA esters.To identify the genes involved in transforming compound IX and XIII CoA esters, other mutants with gene disruptions were incubated with ADD and the culture was analyzed with UHPLC/MS. Among the mutants in which one of the ORFs from ORF4 to ORF33 was disrupted, the ORF4-null mutant and ORF31-null mutant accumulated larger amounts of compound IX than did other mutants (data for ORF4-null mutant and ORF31-null mutant are shown in Fig. 4, and more detailed data are shown in Fig. S4). The ORF31-null mutant accumulated only compound IX, and the ORF4-null mutant accumulated compound XIII together with compound IX. Therefore, a series of ORF4-null and ORF31-null mutants was constructed for complementation experiments (ORF4-null mutant carrying pMFY42, ORF4-null mutant carrying pMFYORF4, ORF31-null mutant carrying pMFY42, and ORF31-null mutant carrying pMFYORF31; pMFYORF4 and pMFYORF31 are pMFY42-based plasmids with ORF4 and ORF31, respectively) and incubated with ADD. In the culture of the ORF31-null mutant carrying pMFY42, a large amount of compound IX was detected, while compounds XIII and XI were almost undetectable (Fig. 5A, m/z 211, 209, 207, and 163). Upon complementation with pMFYORF31, the amount of compound IX decreased and compound XIII was detected (Fig. 5B, m/z 211 and 209). In the culture of the ORF4-null mutant carrying pMFY42, compounds IX and XIII were detected, while compound XI was undetectable (Fig. 5C, m/z 211, 209, 207, and 163). Upon complementation with pMFYORF4, the amount of compounds IX and XIII decreased and compound XI was detected (Fig. 5D, m/z 211, 209, and 207). Therefore, the ORF31-encoded and ORF4-encoded enzymes are presumed to be involved in conversion of the compound IX CoA ester into the compound XIII CoA ester and the following conversion of the compound XIII CoA ester into the compound XI CoA ester, respectively.
UHPLC/MS analysis of the culture of the ORF5-null mutant (A), ORF4-null mutant (B), and ORF31-null mutant (C). The UHPLC chromatograms and the mass chromatograms are shown with the mass spectra (RT, 0.581 to 0.590 min) (detailed data are in Fig. S4). Peaks appeared in the mass chromatogram with an approximately 0.03-min delay relative to those in the UHPLC chromatogram. Peaks of compounds I (RT, 0.80 to 0.85 min), II (RT, 0.74 to 0.77 min), and III (RT, 0.80 to 0.85 min) are indicated in panel A, MS. In UHPLC chromatograms, the vertical axis indicates wavelength, the horizontal axis indicates RT, and the absorbance is plotted in contour. In mass chromatograms, the vertical axis indicates intensity (counts/second [cps]) and the horizontal axis indicates RT (min). In mass spectra, the vertical axis indicates intensity (counts/second) and the horizontal axis indicates mass (m/z).
UHPLC/MS analysis of the culture of complementation experiment with the ORF31-null mutant, ORF4-null mutant, and ORF5-null mutant. Shown are HPLC chromatograms and MS chromatograms of cultures of the ORF31-null mutant harboring broad-host-range plasmid pMFY42 (ORF31− with pMFY42) (A) or harboring pMFY42 carrying ORF31 (ORF31− with pMFYORF31) (B), the ORF4-null mutant harboring pMFY42 (ORF4− with pMFY42) (C) or harboring pMFY42 carrying ORF4 (ORF4− with pMFYORF4) (D), and the ORF5-null mutant harboring pMFY42 (ORF5− with pMFY42) (E) or harboring pMFY42 encoding four ORF5 copies (ORF5− with pMFYORF5m) (F) incubated with 0.1% ADD for 7 days. Filled arrowheads indicate the dominant metabolite accumulating in the culture of each gene disruption mutant, and open arrowheads indicate a putative product of complementation. Graphs show m/z 211 (compound IX), m/z 209 (compound XIII), m/z 207 (compound XI), m/z 163 (a fragment of compound XI), and m/z 201 (compound V) (cf. Fig. 1). *, graphs for m/z 207 and m/z 163 are scaled up 10-fold on the vertical axis. In UHPLC chromatograms, the vertical axis indicates wavelength, the horizontal axis indicates RT, and the absorbance is plotted in contour. In mass chromatograms, the vertical axis indicates intensity (counts/second) and the horizontal axis indicates RT.
Speculation of compound XIII; conversion of 9,17-dioxo-1,2,3,4,10,19-hexanorandrostan-5-oic acid (compound I) by the ORF31-encoded enzyme.Complementation experiments indicated that the compound IX CoA ester is converted by the ORF31-encoded enzyme to the compound XIII CoA ester and then to the compound XI CoA ester by the ORF4-encoded enzyme. To identify compound XIII, metabolites in the culture of the ORF4-null mutant were isolated. From 1 liter of the culture, 104 mg of a compound detected at an RT of 5.65 min in HPLC analysis was isolated and identified as compound IX. A smaller amount of another putative metabolite detected at almost the same RT was presumed to be compound XIII, but the isolation was not successful. Compound IX and putative compound XIII were separated by HPLC analysis after methylation (RTs shifted to 7.7 min and 8.3 min, respectively). However, the putative compound XIII decomposed during the methylation and purification procedure, and only 0.68 mg of the methyl ester of the putative compound XIII was obtained, which was not enough for identification.
A homology search indicated that the ORF31-encoded enzyme is a short-chain dehydrogenase/3-ketoacyl-(acyl-carrier protein) (ACP) reductase (FabG)/enoyl-ACP reductase and that the ORF4-encoded enzyme is an enoyl-ACP reductase belonging to the nitrate monooxygenase superfamily. The putative orthologs of the ORF31- and ORF4-encoded enzymes, IpdF and IpdC in M. tuberculosis H37Rv, showed the conversion of the compound IX CoA ester to the compound XI CoA ester when IpdF and IpdC act together, but the putative product was not detected when the compound IX CoA ester was treated with each one of the enzymes (40). Meanwhile, purified IpdF solely converts the compound I CoA ester to the compound II CoA ester. Based on this information, compound I and the related metabolites in the culture of ORF4 mutants and ORF31 mutants in the complementation experiments were examined in detail (Fig. S5). The amount of compound II detected in the culture of the complemented mutant of ORF4 was smaller than that in the ORF4-null mutant, while the amount of compound II in the culture of the complemented mutant of ORF31 was almost the same as that in the ORF31-null mutant. This result likely suggests the conversion of the compound II CoA ester by ORF4-encoded enzyme, but it is not clear whether the data suggest the involvement of ORF4 in the conversion of the CoA ester of compound II to the CoA ester of compound III or the involvement of ORF31 in the conversion of the CoA ester of compound I to the CoA ester of compound II, because some metabolites upstream of the degradation pathway decreased by downstream complementation (cf. Fig. 5E and F). Therefore, we constructed an (stdC2 ORF31 stdD)-null mutant in which stdC2, ORF31, and stdD were disrupted. The bottom of Fig. 6 shows how the genes were disrupted. Then (stdC2 ORF31 stdD)-null mutants carrying each one of pMFYORF31, pMFYORF4, a pMFY-based plasmid carrying stdC2 (pMFYstdC2; positive control for pMFYORF4), and pMFY42 (negative control) were constructed and incubated with ADD (Fig. 6). UHPLC/MS analysis of the culture showed a distinct increase of the amount of compound II in the culture of the (stdC2 ORF31 stdD)-null mutant carrying pMFYORF31 (Fig. 6B, m/z 239), indicating the involvement of the ORF31-encoded enzyme in conversion of a ketone moiety to a hydroxyl moiety at C-9. The metabolites detected in the culture of the (stdC2 ORF31 stdD)-null mutant carrying pMFYORF4 (Fig. 6C) were almost identical to those detected in the negative control (Fig. 6A).
MS chromatograms (total mass [MS], m/z 239, m/z 237, and m/z 275) of cultures of the (stdC2 ORF31 stdD)-null mutant (gene disruption is briefly shown below the panels) with pMFY42 (A, negative control), with pMFYORF31 (B), with pMFYORF4 (C), and with pMFY42 carrying stdC2 (pMFYScdC2 [D]; positive control for the mutant in panel C) incubated with ADD for 5 days. Culture of the ORF18-null mutant (35) is presented as authentic for compound I (m/z 237; RT, 0.83 to 0.88 min; multiple peaks of isomers) with the small amount of a derivative (compound XIV [m/z 239; RT, 0.63 min]) (E). Culture of the ScdC1C2-null mutant (36) is presented as authentic for II (m/z 239; RT, 0.73 to 0.76 min; multiple peaks of isomers) (F). Culture of the ScdD-null mutant (37) is presented as authentic for compound III (m/z 237; RT, 0.83 to 0.88 min, multiple peaks of isomers) (G). In panel G m/z 237, compound I and compound III are detected at almost the same RT (ca. 0.83 to 0.88 min). m/z 175 graphs indicate a fragment of compound III. This fragment is not detected from compound I. In mass chromatograms, the vertical axis indicates intensity (counts/second) and the horizontal axis indicates RT. Diagrams below panel G show the conversion of compound I to the CoA ester of compound III in TA441 with ORF31 and ORF4, indicating reactions in which they might be involved.
DISCUSSION
Most of the genes in the gene cluster from scdL1 to ORF33 in TA441 (Fig. 1) are involved in degrading steroidal C- and D-rings by β-oxidation, but their function is still under investigation. In this study, an ORF5-null mutant was shown to accumulate 9,17-dioxo-1,2,3,4,5,6,10,19-octanorandrost-8(14)-en-7-oic acid (compound XI) as the dominant metabolite, and the levels of compound XI decreased by complementation of the mutant with an ORF5-encoded enzyme. Homology searching indicated that the ORF5-encoded enzyme is a member of the crotonase/enoyl-CoA hydratase superfamily. Based on this information, the ORF5-encoded enzyme was elucidated to be the enoyl-CoA hydratase for the CoA ester of compound XI. The D-ring is supposed to be cleaved after hydration by the ORF5-encoded enzyme. However, the possible product of the ORF5-encoded hydratase (a compound with a molecular weight of 226) did not accumulate in any one of the examined gene disruption mutant cultures (data not shown), which might imply that the D-ring is nonenzymatically cleaved after hydration.
9α-Hydroxy-17-oxo-1,2,3,4,5,6,10,19-octanorandrostan-7-oic acid (compound IX) was also isolated from the ORF5-null culture. This compound was detected in the cultures of several mutants in which one of the genes from scdL1 to ORF33 was disrupted, suggesting that compound IX is one of the major metabolites that accumulates during the degradation of steroid compounds in TA441. The ORF4-null mutant and ORF31-null mutant accumulated more compound IX than other mutants did, and a putative metabolite showing a molecular weight of 210 (compound XIII) was detected together with compound IX in the culture of the ORF4-null mutant. Complementation experiments indicated that the ORF31-encoded enzyme is involved in conversion of the CoA ester of compound IX to the CoA ester of compound XIII and that the ORF4-encoded enzyme is involved in the subsequent conversion to the CoA ester of compound XI. However, isolation and identification of XIII from the culture of the ORF4-null mutant were not successful because the amount isolated was too small, probably due to the unstable β-diketo structure. Compound XI also has a β-diketo structure and NMR data were barely obtained, but the EI-MS data were taken with the methyl ester. The isolated amount of compound XI was 33 mg, and that of the methyl ester of compound XIII was 0.68 mg. Nonmethylated compound XIII was not isolated because it was detected at the same RT as compound IX during HPLC separation.
By a homology search, the ORF31-encoded enzyme was found to be a protein that belongs to the provisional short-chain dehydrogenase (SDR) family (PRK07831). This family includes FabG, a 3-oxoacyl-acyl carrier protein (ACP) reductase in bacterial type II fatty acid synthesis, and other 3-oxoacyl-ACP reductases. In the phylogenetic tree without highly similar enzymes, the ORF31-encoded enzyme formed a cluster with IpdF of M. tuberculosis H37Rv (40) (Fig. 7, top; the tree was constructed by SmartBLAST [https://blast.ncbi.nlm.nih.gov/Blast.cgi]). Purified IpdF, a putative ortholog of the ORF31-encoded enzyme, was shown to convert the CoA ester of compound I to the CoA ester of compound II (40). Our experiments using gene disruption mutants indicated that the ORF31-encoded enzyme has the same activity: conversion of a ketone moiety to a hydroxyl moiety at C-9. Therefore, the ORF31-encoded enzyme is assumed to be the hydroxylacyl dehydrogenase that converts the CoA ester of compound IX to the CoA ester of compound XIII, and this conversion is presumed to be a reversible reaction. Experiments using the gene disruption mutants indicated that the ORF31-encoded enzyme is also involved in conversion of the CoA ester of compound I to the CoA ester of compound II. However, as the dominant metabolite accumulated by the ORF31 disruption mutant is compound IX, the primary function of the ORF31-encoded enzyme is converting the CoA ester of compound IX to the CoA ester of compound XIII, and there should be at least one more enzyme that catalyzes the conversion of the CoA ester of compound I to the CoA ester of compound II in TA441.
(Top) Phylogenetic tree of ORF31-encoded enzyme and related enzymes (highly similar enzymes were excluded). Enzymes are 3-oxoacyl-ACP reductase of Clostridium difficile 630 (50), predicted 3-oxoacyl-ACP reductase of soybean (GenBank accession no. XM_003538516.3), 3-oxoacyl-ACP reductase FabG of Shewanella oneidensis MR-1 (51), 3-oxoacyl-ACP reductase of Thermotoga maritima MSB8 (52), and IpdF of M. tuberculosis H37Rv (40). (Bottom) Phylogenetic tree of ORF4-encoded enzyme and related enzymes belonging to the nitrate monooxygenase (NMO) superfamily (highly similar enzymes were excluded). Enzymes are NMO of Pseudomonas aeruginosa PAO1 (53), FabK of C. difficile 630 (50), FabK of Streptococcus pneumoniae R6 (54), putative protein of T. maritima MSB8 (52), and IpdC of M. tuberculosis H37Rv (40). The trees were constructed with SmartBLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi).
A homology search indicated that the ORF4-encoded enzyme belongs to the nitronate monooxygenase (NMO) superfamily. In the phylogenetic tree without highly similar enzymes, the ORF4-encoded enzyme formed a cluster with IpdC of M. tuberculosis H37Rv (40) near the cluster consists of FabK, an enoyl-ACP reductase in bacterial fatty acid biosynthesis (48) (Fig. 7, bottom; the tree was constructed by SmartBLAST). Among the enoyl-ACP reductases in type II fatty acid synthesis, FabI is widely distributed in bacteria and plants, whereas FabK is a distinctly different member of the FabI group. fabK is generally found in an operon for fatty acid synthesis which lacks fabI. Based on this information and all the data obtained in this study, the ORF4-encoded enzyme was elucidated to be the acyl-CoA dehydrogenase for conversion of the CoA ester of compound XIII to the CoA ester of compound XI.
Accordingly, the steroidal C- and D-ring degradation in TA441 determined in this study is as follows. The CoA ester of compound IX is converted to the CoA ester of compound XIII by RF31-encoded hydroxylacyl dehydrogenase (named ScdG), followed by conversion to the CoA ester of compound XI by ORF4-encoded acyl-CoA dehydrogenase (named ScdK). Then a water molecule is added to the resultant CoA ester of compound XI by the ORF5-encoded enoyl-CoA hydratase (named ScdY) (Fig. 8). Compound XIII was assumed to be 9,17-dioxo-1,2,3,4,5,6,10,19-octanorandrostan-7-oic acid, and conversion of the CoA ester of compound IX to the CoA ester of compound XIII is likely to be a reversible reaction. ScdG might be a 3-hydroxylacyl-CoA dehydrogenase which also acts on a hydroxyl group of compounds other than 3-hydroxylacyl-CoA. The elucidated pathway in C. testosteroni TA441 is different from the corresponding pathways reported for M. tuberculosis H37Rv (Fig. 8) (40, 45). ScdG is involved in conversion of the CoA ester of compound I to the CoA ester of compound II, but this is not the primary function of ScdG, and there is at least one other enzyme for catalyzing this reaction in TA441.
Metabolites and genes in degradation of the steroidal C- and D-rings in C. testosteroni TA441 elucidated in this study. The corresponding degradation pathway in M. tuberculosis H37Rv is presented at the bottom (40, 45). Compounds are 9α-hydroxy-17-oxo-1,2,3,4,5,6,10,19-octanorandrostan-7-oic acid (compound IX), 9,17-dioxo-1,2,3,4,5,6,10,19-octanorandrostan-7-oic acid (compound XIII), and 9,17-dioxo-1,2,3,4,5,6,10,19-octanorandrost-8(14)-en-7-oic acid (compound XI). ORF31, ORF4, and ORF5 are named scdG, scdK, and scdY, respectively.
MATERIALS AND METHODS
General experimental procedures.Fast-atom bombardment-mass spectrometry (FAB-MS) (negative-ion mode) was recorded on a JEOL JMS-700 mass spectrometer (JEOL Ltd., Tokyo, Japan) using a glycerin matrix. Electron ionization MS was recorded on a JEOL JMS-SX102 mass spectrometer (JEOL Ltd.). One- and two-dimensional nuclear magnetic resonance (NMR) spectra were recorded on a JNM-ECP500 or JNM-ECA600 spectrometer (JEOL Ltd.). Tetramethylsilane at 0 ppm in deuterated chloroform solution and the residual proton signal at 2.49 ppm in deuterated dimethyl sulfoxide (DMSO-d6) solution or the residual proton signal at 3.30 ppm in deuterated methanol solution were used as internal references for 1H chemical shifts. 13C chemical shifts were obtained with reference to DMSO-d6 (39.5 ppm) or deuterated chloroform (77.0 ppm) at 25°C.
Culture conditions.Mutant strains of C. testosteroni TA441 were grown at 30°C in a mixture of equal volumes of Luria-Bertani (LB) medium and C medium (a mineral medium for TA441) (29) with suitable carbon sources. Cholic acid, ADD, and other steroids were added as filter-sterilized DMSO solutions with a final concentration of 0.1% (wt/vol).
Construction of gene disruption mutants, plasmids, and mutants for complementation experiments.ORF5-, ORF4-, ORF31-, and stdC2 stdD-null mutants were constructed by insertion of a kanamycin resistance gene without a terminator into the SmaI site, SacII site, SmaI site, and ApaI-EcoRV site in each ORF or gene. Insertion of the kanamycin resistance gene was confirmed by Southern hybridization. DNA fragments containing each one of ORF5, ORF4, and ORF31 were obtained by PCR amplification and introduced into broad-host-range plasmid pMFY42 (49), which can be maintained in Pseudomonas and its relatives and gives them tetracycline resistance, to construct pMFYORF5, pMFYORF4, and pMFYORF31, respectively. Plasmid pMFYORF5m, encoding four copies of ORF5 (at least two were in the same direction as the promoter of pMFY42), was constructed in the same manner. Retention of the plasmids by the gene disruption mutants and transformants was confirmed by Southern hybridization and/or PCR amplification.
HPLC analysis.After the addition of a double volume of methanol to the culture, the mixture was centrifuged, and the supernatant was directly injected into a high-performance liquid chromatograph (Alliance 2695, with UV and 996-photodiode array detectors; Nihon Waters, Tokyo, Japan) equipped with an Inertsil ODS-3 column (4.6 by 250 mm; GL Sciences Inc., Tokyo, Japan). Elution was carried out using a linear gradient from 20% solution A (CH3CN-CH3OH-trifluoroacetic acid [TFA] at 95:5:0.05) and 80% solution B (H2O-CH3OH-TFA at 95:5:0.05) to 65% solution A and 35% solution B over 10 min; these conditions were maintained for 3 min and then changed to 20% solution A. The flow rate was 1.0 ml/min at 40°C. For the isolation of intermediate compounds, a Waters 600 HPLC with an Inertsil ODS-3 column (20 by 250 mm) was used, and elution was carried out using a linear gradient from 20% solution A and 80% solution B to 65% solution A and 35% solution B over 25 min, which was maintained for 10 min, followed by 20% solution A for 5 min. The accumulated compound was detected at 206 nm. The flow rate was 8 ml/min at 40°C. When other conditions were used, they are described.
Isolation of compounds in the culture of the ORF5-null mutant.The ORF5 disruption mutant was incubated in 100 ml of culture with 0.1% (wt/vol) cholic acid for 4 days, and the total volume (500 ml) of the culture was extracted with 550 ml of ethyl acetate under neutral conditions. The ethyl acetate layer was treated with Na2SO4 and concentrated in vacuo. Then it was dissolved in a small amount of methanol and subjected to HPLC (Waters 600), which revealed the presence of compound X (Fig. 2; RT, 5.7 min) and compound XII (RT, 6.9 min) (9,17-dihydroxy-1,2,3,4,10,19-hexanorandrost-6-en-5-oic acid was detected at an RT of 5.7 min, but it was not contained in the layer extracted under neutral conditions). The fractions containing each of these compounds were purified by HPLC (Waters 600). Then the water layer was extracted twice with 550 ml of ethyl acetate under acidic conditions (pH 2 with HCl), and the compounds in the ethyl acetate layer extracted under acidic conditions were isolated in the same manner as the ethyl acetate layer extracted under neutral conditions. Compounds IX (RT 5.6 min) and XI (RT 5.7 min) were isolated. Fractions containing each of these compounds were purified with HPLC (Waters 600), dried, and resolved in a suitable solvent for NMR analysis (CD3OD). Compound XI was unstable and decomposed to compound XII during storage in the freezer (−20°C). The peaks of compound XI detected by high-resolution EI-MS were not clear for identification. FAB-MS data indicated the molecular weight of compound XI to be 208. Compound XI was methylated and purified, and high-resolution EI-MS data for the methyl ester of compound XI were successfully obtained. But the HPLC analysis indicated that the methyl ester of compound XI was also rather unstable and spontaneously converted to a compound detected at an RT of 13.6 min.
UHPLC/MS analysis.The 1-ml culture was extracted twice with a double volume of ethyl acetate under acidic conditions (pH 2 with HCl). The ethyl acetate layer was dried and dissolved in 550 μl of methanol, and 1 μl of the methanol solution was injected into a UHPLC/MS. UHPLC/MS was carried out using an Applied Biosystems Q Trap liquid chromatography-tandem MS system with a reverse-phase column (XTerra MSC18, 2.1 by 150 mm; Waters) at a flow rate of 0.4 ml/min at 40°C. Elution was carried out using a linear gradient from 20% solution C (CH3CN) and 80% solution D (H2O-HCOOH at 100:0.05) to 100% solution C over 4.5 min, which was maintained for 2 min. Electrospray ionization (negative-ion mode) was used for detection. The conditions for MS were an ion spray voltage of 5.0 kV, a curtain gas pressure of 15 lb/in2, a nebulizer gas pressure of 40 lb/in2, an auxiliary gas pressure of 60 lb/in2, and an ion source temperature of 400°C.
Methylation.Compounds were resolved in 5 ml of toluene-MeOH (3:1 solution). A total of 500 μl of TMS (trimethylsilyl)-diazomethane solution (0.6 mol/liter of hexane) was added to the solution and placed at room temperature for ca. 5 min. Acetic acid was added to the sample until the yellow color of the solution disappeared.
ACKNOWLEDGMENTS
M.H. appreciates R. Kato (head of the Condensed Molecular Materials Laboratory at RIKEN) for thoughtful support and advice.
FOOTNOTES
- Received 7 June 2018.
- Accepted 23 July 2018.
- Accepted manuscript posted online 7 September 2018.
Supplemental material for this article may be found at https://doi.org/10.1128/AEM.01324-18.
- Copyright © 2018 American Society for Microbiology.
