Functional Characterization of Three Specific Acyl-Coenzyme A Synthetases Involved in Anaerobic Cholesterol Degradation in Sterolibacterium denitrificans Chol1S

ABSTRACT

The denitrifying betaproteobacterium Sterolibacterium denitrificans Chol1S catabolizes steroids such as cholesterol via an oxygen-independent pathway. It involves enzyme reaction sequences described for aerobic cholesterol and bile acid degradation as well as enzymes uniquely found in anaerobic steroid-degrading bacteria. Recent studies provided evidence that in S. denitrificans, the cholest-4-en-3-one intermediate is oxygen-independently oxidized to Δ⁴-dafachronic acid (C₂₆-oic acid), which is subsequently activated by a substrate-specific acyl-coenzyme A (acyl-CoA) synthetase (ACS). Further degradation was suggested to proceed via unconventional β-oxidation, where aldolases, aldehyde dehydrogenases, and additional ACSs substitute for classical β-hydroxyacyl-CoA dehydrogenases and thiolases. Here, we heterologously expressed three cholesterol-induced genes that putatively code for AMP-forming ACSs and characterized two of the products as specific 3β-hydroxy-Δ⁵-cholenoyl-CoA (C₂₄-oic acid)- and pregn-4-en-3-one-22-oyl-CoA (C₂₂-oic acid)-forming ACSs, respectively. A third heterologously produced ATP-dependent ACS was inactive with C₂₆-, C₂₄-, or C₂₂-oic-acids but activated 3aα-H-4α-(3′propanoate)-7aβ-methylhexahydro-1,5-indanedione (HIP) to HIP-CoA, a rather late intermediate of aerobic cholesterol degradation that still contains the CD rings of the sterane skeleton. This work provides experimental evidence that anaerobic steroid degradation proceeds via numerous alternate CoA-ester-dependent or -independent enzymatic reaction sequences as a result of aldolytic side chain and hydrolytic sterane ring C—C bond cleavages. The aldolytic side chain degradation pathway comprising highly exergonic ACSs and aldehyde dehydrogenases is considered to be essential for driving the unfavorable oxygen-independent C₂₆ hydroxylation forward.

IMPORTANCE The biological degradation of ubiquitously abundant steroids is hampered by their low solubility and the presence of two quaternary carbon atoms. The degradation of cholesterol by aerobic Actinobacteria has been studied in detail for more than 30 years and involves a number of oxygenase-dependent reactions. In contrast, much less is known about the oxygen-independent degradation of steroids in denitrifying bacteria. In the cholesterol-degrading anaerobic model organism Sterolibacterium denitrificans Chol1S, initial evidence has been obtained that steroid degradation proceeds via numerous alternate coenzyme A (CoA)-ester-dependent/independent reaction sequences. Here, we describe the heterologous expression of three highly specific and characteristic acyl-CoA synthetases, two of which play key roles in the degradation of the side chain, whereas a third one is specifically involved in the B ring degradation. The results obtained shed light into oxygen-independent steroid degradation comprising more than 40 enzymatic reactions.

INTRODUCTION

Steroids constitute a highly abundant class of natural compounds with a common sterane ring system. They play essential roles as components of biological membranes and signaling molecules (1). Steroids serve as a growth substrate for certain Actinobacteria and Proteobacteria, which use them as a carbon source and, together with an electron acceptor, an energy source. Bacterial degradation represents the major means of eliminating steroids from the environment (2, 3, 4). Research on microbial steroid degradation also has a biotechnological impact, as enzymes that transform steroids may be used for the synthesis of value-added products, e.g., the specific conversion of cholecalciferol to 25-hydroxyvitamin D₃ by monooxygenases (5) or oxygen- and electron donor-independent hydroxylases (6). Finally, steroid degradation is medically relevant, because Mycobacterium tuberculosis is known to use cholesterol from macrophages as a growth substrate during intracellular survival (7, 8).

Aerobic steroid degradation has mainly been studied in several cholesterol-degrading Actinobacteria such as the model organisms Rhodococcus jostii RHA1 and Mycobacterium tuberculosis H37Rv (9), while the aerobic degradation of the steroidal core ring has mainly been studied in Comamonas testosteroni (10). Bile salt degradation was originally studied in Pseudomonas species but recently also in R. jostii (11, 12, 13, 14).

During both aerobic and anaerobic cholesterol degradation, catabolism is initiated by the oxidation of ring A by 3β-hydroxysteroid dehydrogenases or cholesterol oxidases and 3-ketosteroid-Δ¹-dehydrogenases, yielding cholest-4-en-3-one and cholesta-1,4-diene-3-one, respectively (15, 16, 17). Further degradation involves the oxidation of the C-26 primary carbon atom to yield Δ⁴-dafachronic acid (C₂₆-oic acid) (Fig. 1). The enzymes involved in this reaction sequence fundamentally differ in aerobic and anaerobic bacteria. In aerobic Actinobacteria, the cytochrome P450 oxygenase Cyp125 or Cyp142 catalyzes C-26 hydroxylation and further oxidation to C₂₆-oic acid (18, 19). In contrast, the denitrifying Sterolibacterium denitrificans first hydroxylates the tertiary C-25 with water by a Mo-dependent steroid C-25 dehydrogenase (C25DH) (20) followed by an apparent hydroxyl shift from the tertiary C-25 to the primary C-26 by an unknown enzyme (21). The subsequent oxidation of the primary alcohol to C₂₆-oic acid is then achieved by the action of putative cholesterol-induced alcohol and aldehyde dehydrogenases (22). In both aerobic and anaerobic cholesterol-degrading bacteria, ATP-dependent acyl-coenzyme A (acyl-CoA) synthetases (ACSs) specific for C₂₆-oic acid were identified and characterized (22, 23) (Fig. 1).

FIG 1.

Aerobic (A) and anaerobic (B) cholesterol degradation pathways. The involvement of ATP-dependent acyl-CoA synthetases (ACSs) is indicated; in the anaerobic pathway, only the first ACS acting on C₂₆-oic acid had been isolated and characterized at the beginning of this work. ACDH, acyl-CoA dehydrogenase; ECH, enoyl-CoA hydratase; HADH, 3-hydroxyacyl-CoA dehydrogenase; THL, thiolase; ALD, aldolase; ALDH, aldehyde dehydrogenase.

In all cholesterol-degrading bacteria, β-oxidation of the C₂₆-oyl-CoA intermediate yields androsta-1,4-diene-3,17-dione (ADD), two propionyl-CoAs, and one acetyl-CoA (Fig. 1). In Actinobacteria, classical acyl-CoA dehydrogenases, enoyl-CoA hydratases, 3-hydroxyacyl-CoA dehydrogenases, and thiolases are involved (8). The only exception is the last C—C bond cleavage at a tertiary carbon atom, which is catalyzed by an aldolase at the level of a β-hydroxyacyl-CoA intermediate (24). In contrast, a recent integrated multiomics study suggested that cholesterol side chain degradation in S. denitrificans proceeds via aldolytic cleavage of all 3-hydroxacyl-CoA intermediates to propionyl-CoA/acetyl-CoA and the corresponding C₂₄/C₂₂ aldehydes. The latter were proposed to be oxidized and activated to the corresponding C₂₄-/C₂₂-oyl-CoAs (22). Consequently, the conventional 3-hydroxyacyl-CoA dehydrogenase and thiolases would be replaced by aldolases, aldehyde dehydrogenases, and ACSs (Fig. 1). However, besides the C₂₆-oic acid-activating ACSs, none of the proposed enzymes involved in anoxic cholesterol side chain degradation have been isolated and characterized yet. Notably, the aldolytic pathway proposed for cholesterol degradation in S. denitrificans for steroid side chain degradation resembles that reported for aerobic cholate degradation (4, 14), demonstrating the composite character of the anaerobic cholesterol degradation pathway.

After side chain removal, the degradation of the ADD formed again greatly differs in aerobic and anaerobic steroid-degrading bacteria. In the aerobic pathway, a series of oxygenase-dependent reactions results in ring A and B cleavage via a name-giving 9,10-seco intermediate, finally yielding the CD rings containing 3aα-H-4α(3′propanoate)-7aβ-methylhexahydro-1,5-indanedione (HIP) and its CoA ester HIP-CoA (8). In contrast, denitrifying bacteria degrade ring A by hydrolysis to 17-hydroxy-1-oxo-2,3-seco-androstan-3-oic acid via the 2,3-seco pathway (25). Nothing is known about ring B cleavage in anaerobic cholesterol-degrading denitrifying bacteria. But, the identification of high-resolution masses matching that for HIP-CoA suggested that this compound results from ring B cleavage in S. denitrificans and represents an intermediate in both aerobic and anaerobic cholesterol degradation (22) (Fig. 1). Very recently, the degradation of HIP-CoA to central intermediates was revealed in Actinobacteria, which involves two distinct ring-cleaving hydrolases (26). Genomic analyses suggested that HIP-CoA is similarly degraded in S. denitrificans and possibly other denitrifying bacteria (22).

In this work, we heterologously produced and characterized three ACSs from S. denitrificans and demonstrate that two of them are involved in side chain degradation (specific for C₂₄- and C₂₂-oic acids), whereas the third ACS is involved in B ring degradation (specific for HIP). The work provides enzymatic evidence for the omics-based proposal for side chain degradation and CD ring cleavage in an anaerobic cholesterol-degrading model organism; it sheds light into the largely still obscure anaerobic steroid degradation pathway.

RESULTS

ACS genes involved in anaerobic cholesterol degradation.

Recent mass spectrometry (MS)-based analyses of metabolites in S. denitrificans cells grown with cholesterol suggested a β-oxidation-like side chain degradation pathway (21, 22, 27). With the cholesterol-induced gene product SDENChol_v1_11189, which specifically catalyzed the ATP- and CoA-dependent formation of C₂₆-oyl-CoA, the first enzyme of the proposed β-oxidation sequence was isolated and characterized (22). Analyses of further metabolites suggested aldolytic cleavage of C—C bonds during side chain degradation, which would involve two additional C₂₄- and C₂₂-oic acid-converting ACSs. Finally, a CoA ester matching the authentic HIP-CoA standard was identified in whole cells during ultraperformance liquid chromatography (UPLC)-coupled high-resolution mass spectrometry (HRMS); its formation may be catalyzed by a further ACS (22).

Previous proteome analyses identified three cholesterol-induced ACS-like gene products next to the C₂₆-oyl-CoA-forming SDEN_v1_11189 (see Table S1 in the supplemental material). Two of them, SDENChol_v1_10299 and SDENChol_v1_10305, were induced by 1.45 and 2.35 (log2 ratios), respectively, in cells grown with cholesterol compared with that grown with testosterone, which contains the sterane skeleton but which lacks the side chain (22). This finding suggests that the two ACSs are involved in side chain degradation. In contrast, a third ACS candidate (SDENChol_v1_10766) was more abundant in cells grown with cholesterol than in those grown with propionate but was not induced in cells grown with cholesterol versus those grown with testosterone. Consequently, the latter is assumed to be involved in sterane ring system degradation.

Heterologous production, purification, and characterization of three ACSs involved in anaerobic cholesterol degradation.

The genes encoding three putative ACSs involved in anaerobic cholesterol degradation, SDENChol_v1_10299, SDENChol_v1_10305, and SDENChol_v1_10766, were heterologously expressed in soluble forms in Escherichia coli BL21 with N-terminal Strep tags. The products were purified via Strep-Tactin affinity chromatography (see Materials and Methods). An SDS-PAGE analysis of the three highly enriched enzymes revealed three soluble proteins at molecular masses of 65, 70, and 55 kDa matching the theoretical masses of 65.1, 71.3, and 58.6 kDa with enrichment factors of 43.4, 18.4, and 33, respectively (see Fig. S2).

Kinetic properties of SDENChol_v1_10299, SDENChol_v1_10305, and SDENChol_v1_10766.

The heterologously produced ACSs were tested for their substrate preference using numerous carboxylic acids, including Δ⁴-dafachronic acid (C₂₆), 3β-hydroxy-Δ⁵-cholenic acid (C₂₄), cholic acid and its analogues deoxy- and lithocholic acids (C₂₄ each), pregn-4-en-3-one-22-oic acid (C₂₂), HIP (C₁₃), palmitic acid (C₁₈), decanoic acid (C₁₀), hexanoic acid (C₆), and propionate (C₃), as potential substrates. The concentrations of all carboxylic acids tested were 0.3 mM in the presence of 7.8% (wt/vol) 2-hydroxypropyl-β-cyclodextrin as the solubilizing agent. The addition of the latter largely increased the solubility of the carboxylic acid substrates and in turn the activity of the individual ACSs. The time-dependent formation of CoA ester products was followed by UPLC analysis in the presence of Mg-ATP, CoA, and the individual heterologously produced ACSs. The identities of the reaction products were confirmed by coelution during UPLC analyses with standards, characteristic UV-visible (UV-vis) spectra, and liquid chromatography-electrospray ionization MS (LC-ESI/MS) analyses (see Fig. S1 and S3 and Table S2). No conversion was observed in any case when Mg-ATP was omitted from the assay.

SDENChol_v1_10299 showed the highest activity with the 3β-hydroxy-Δ⁵-cholenic acid (C₂₄-oic acid), whereas virtually no conversion of the C₂₆- and C₂₂-oic acids and shorter carboxylic acids was observed (Table 1). These results strongly indicate that cholesterol-induced SDENChol_v1_10299 is specifically involved in the activation of the side chain degradation product 3β-hydroxy-Δ⁵-cholenic acid to its CoA ester. It also activated the C₂₄ bile acids lithocholic acid, cholic acid, and deoxycholic acid to their respective CoA thioesters at lower rates, suggesting that the chain length rather than modifications at the sterane ring system governs the substrate preference. In contrast, SDENChol_v1_10305 only converted pregn-4-en-3-one-22-oic acid to its CoA ester, whereas it was apparently inactive (<0.1%) with all other carboxylic acids tested (Table 1). In conclusion, the gene product is now assigned as a specific pregn-4-en-3-one-22-oyl-CoA synthetase. Notably, Δ⁴-dafachronic acid (C₂₆-oic acid), the preferred substrate of the recently characterized SDENChol_v1_11189, was not converted by SDENChol_v1_10299 or SDENChol_v1_10305. These results suggest that all three ACSs are involved in side chain degradation with a strong preference for the chain lengths of the individual substrates. SDENChol_v1_10766 converted neither the C₂₆-/C₂₄-/C₂₂-oic acid substrates nor any of the aliphatic fatty acids. Instead, it showed a high activity with HIP, indicating that it plays a role in the activation of the CD rings containing carboxylic acid (Table 1).

TABLE 1.

Substrate preference of the heterologously produced SDENChol_v1_10299, SDENChol_v1_10305, and SDENChol_v1_10766

Using the coupled spectrophotometric assay, the initial rates of all three ACSs strongly depended on the substrate concentration. A fit to Michaelis-Menten curves revealed K_m values of SDENChol_v1_10299, SDENChol_v1_10305, and SDENChol_v1_10766 for their individual preferred carboxylic acid substrates (74 to 156 μM) as well as k_cat values (1.8 to 10.8 s⁻¹); the results are summarized in Table 2.

TABLE 2.

Apparent steady-state kinetic parameters for ACSs involved in cholesterol degradation of S. denitrificans^a

Enzyme	Substrate	Km (μM)	kcat (s−1)	kcat/Km (105 M−1 · s−1)	Reference
SDENChol_v1_11189	C26-Δ4-Dafachronic acid	310 ± 50	1.1 ± 0.19	0.035	(5)
SDENChol_v1_10299	C24-3β-Hydroxy-Δ5-cholenic acid	73 ± 14	10.78 ± 0.53	1.5	This work
SDENChol_v1_10305	C22-Pregn-4-en-3-one-22-oic acid	172 ± 20	1.83 ± 0.08	0.11	This work
SDENChol_v1_10766	C13-3aα-H-4α(3′-propanoate)-7aβ-methylhexahydro-1,5-indanedione (HIP)	156 ± 25	3.76 ± 0.24	0.24	This work

^aMean values from three biological replicates are shown with the respective standard errors.

ACS activities in extracts from cells grown with different substrates.

The substrate preferences of the three ACSs for side chain-containing C₂₄- and C₂₂-oic acids or for HIP suggested a specific role in steroid degradation, whereas they should not be required during growth with propionate. This finding is in full agreement with the differential induction of the individual genes during growth on different carbon sources (22). To further analyze the differential induction of ACS activities, we tested extracts grown with cholesterol and propionate using different carboxylic acid substrates. The activities with the individual carboxylic substrates of all three ACS were induced by a factor of 6 to 9 in cells grown with cholesterol versus with propionate, strongly confirming their specific role in anaerobic cholesterol degradation (Table 3).

TABLE 3.

Kinetic parameters of ACS activities in extracts from cells grown with cholesterol and propionate^a

ACS substrates	Vmax (nmol · min−1 · mg−1)
	Cholesterol	Propionate
C24-3β-Hydroxy-Δ5-cholenic acid	244 ± 11.2	31 ± 3.8
C22-Pregn-4-en-3-one-22-oic acid	98 ± 6.5	11 ± 2.1
C13-3aα-H-4α(3′-propanoate)-7aβ-methylhexahydro-1,5-indanedione (HIP)	112 ± 6.9	12 ± 1.4

^aThe mean values from three biological replicates with respective standard deviations are shown.

Phylogenetic analyses of ACSs involved in steroid degradation.

Using BLAST, a total of 14 putative genes encoding ACSs were identified in the genome of S. denitrificans (expect value threshold = e⁻¹⁴). A multiple sequence alignment with the clustalW program revealed that the experimentally verified ACSs from S. denitrificans specific for C₂₆-oic acid and HIP cluster with the corresponding FadD19 and FadD3 enzymes involved in cholesterol degradation in M. tuberculosis H37v and R. jostii RHA1, respectively (Fig. 2). In contrast, the closest related enzymes to the S. denitrificans ACS specific for C₂₄- and C₂₂-oic acids are those involved in bile acid degradation in R. jostii RHA1 and Pseudomonas sp. strain DOC21. According to the relationship with corresponding enzymes from E. coli, two further putative ACSs of S. denitrificans most likely function as acetate- and propionate-activating enzymes, whereas the function of the remaining five putative ACSs remains elusive.

FIG 2.

Phylogenetic analysis of ACSs from S. denitrificans (bold) and other steroid-degrading organisms. The clustering of ACSs specific for individual substrates is shown, with the four experimentally verified ACSs in red: C26ACS, Δ⁴-dafachronic acid; C24ACS, 3β-hydroxy-Δ⁵-cholenic acid; C22ACS, pregn-4-en-3-one-22-oic acid; HIPACS, 3aα-H-4α-(3′propanoate)-7aβ-methylhexahydro-1,5-indanedione (HIP). The accession numbers are given in parentheses. Scale bar, number of substitutions per site.

DISCUSSION

The three heterologously produced and characterized ACSs from S. denitrificans Chol1S in this work, together with a recently identified C₂₆-oyl-CoA-forming ACS, all show remarkable specificities for their individual C₂₆-, C₂₄-, and C₂₂-oic acid and HIP substrates. Two show similarities to enzymes involved in aerobic cholesterol degradation (C₂₆-oyl-CoA and HIP-CoA forming), whereas the other two are rather related to enzymes involved in bile acid degradation and are absent in aerobic cholesterol degradation (C₂₄- and C₂₂-oyl-CoA forming). This finding corroborates the composite anaerobic steroid degradation pathway involving numerous CoA-ester-dependent/independent reaction sequences known from aerobic cholesterol or bile acid degradation.

The three ACSs involved in side chain degradation strictly discriminate between the structurally related C₂₆-, C₂₄-, and C₂₂-oic acids, which strongly supports their assignment to individual carboxylic acid substrates of cholesterol side chain degradation. The assignment of SDENChol_v1_10766 to a highly specific HIP-CoA synthetase appears to be less clear, as no structurally related substrate analogues that could serve as potential intermediates in anaerobic steroid degradation were available. However, the k_cat of the enzyme from S. denitrificans is in a similar range (3.7 s⁻¹) as the one reported for HIP-CoA synthetase from R. jostii RHA1 (k_cat, 7.1 s⁻¹) (28). Thus, the steady-state kinetic properties of SDENChol_v1_10766 support that HIP or a structurally closely related analogue acts as an intermediate of anaerobic cholesterol degradation. Moreover, previous studies identified clustered cholesterol-induced genes in S. denitrificans (22) that are highly similar to those involved in β-oxidation of HIP-CoA to 4-methyl-5-oxooctanedioyl-CoA in other organisms (26). In conclusion, both findings together suggest that CD ring degradation proceeds via identical or highly similar pathways in aerobic and anaerobic cholesterol-degrading bacteria.

The identification of ACSs specific for C₂₄-oic and C₂₂-oic acids in S. denitrificans is in full agreement with the recently proposed aldolytic C—C bond cleavage during isoprenoid side chain β-oxidation (22). The classical thiolytic C—C cleavage of β-oxidation yields two CoA esters from a β-ketoacyl-CoA intermediate, whereas the aldolytic C—C-cleaving mechanism yields an aldehyde plus acetyl-CoA or propionyl-CoA. As a result, the aldehyde has to be oxidized to a carboxylic acid, followed by ACS (AMP-forming)-dependent thioesterification to initiate a new round of β-oxidation. On the first view, this strategy appears to be energetically less efficient than thiolytic cleavage, as two additional phosphoanhydride bonds are hydrolyzed for each β-oxidation reaction sequence, assuming that the pyrophosphate formed by AMP-forming ACS is readily hydrolyzed by a pyrophosphatase. Consequently, anaerobic cholesterol degradation via two aldolytic side chain cleavages yields four ATP less than that via the thiolytic pathway. However, the insights into the anaerobic cholesterol degradation pathway obtained so far suggest that in total, around 10 acetyl-CoA/propionyl-CoA units plus one succinyl-CoA are formed from cholesterol, which are fully oxidized in the tricarboxylic acid cycle. Taking into account that NADH oxidation coupled to denitrification yields more than one ATP per NADH oxidized (29), the four additional ATP consumed during aldolytic side chain degradation will only slightly affect the overall ATP yield. Thus, energetic considerations appear to be rather marginal when comparing the employment of the aldolytic versus the thiolytic side chain degradation pathway. A possible much more important point is that the aldolytic pathway comprises, with the NAD⁺-dependent aldehyde oxidation to a carboxylic acid (ΔG°' ≈ −45 kJ mol⁻¹) (30) and the CoA ester formation by AMP-forming ACS/pyrophosphatase (ΔG°′ ≈ −20 kJ · mol⁻¹), two highly exergonic steps that are missing in the thiolytic pathway. In contrast, the latter involves an unfavorable reaction that is missing in the aldolytic pathway: the oxidation of β-hydroxyacyl-CoA to β-ketoacyl-CoA with NAD⁺ as the electron acceptor (ΔG°′ ≈ 25 kJ mol⁻¹), which almost compensates for the subsequent exergonic β-ketothiolase reaction (ΔG = −25.1 kJ · mol⁻¹) (31). In conclusion, the equilibrium of the aldolytic pathway is clearly shifted more towards the oxidative direction than the thiolytic pathway is, which will strongly promote the unidirectionality of the overall catabolic pathway.

The question rises why denitrifying steroid-degrading bacteria employ an aldolytic pathway? The most probably rate-limiting step in anaerobic cholesterol degradation is the formation of the C₂₆-hydroxy species via a tertiary C₂₅-hydroxy intermediate. This conversion of a tertiary to a primary alcohol is unprecedented in biology and probably highly unfavorable; it might be only possible if highly exergonic downstream β-oxidation reactions pull the reaction forward. This assumption is in accordance with 25-OH-cholest-4-en-3-one (and its ring A 1,4-diene analogue) representing the most abundant metabolite in MS-based analyses (22), indicating that its conversion is rate limiting in the overall degradation pathway. Thus, the rationale of using the energetically less efficient aldolytic pathway instead of the thiolytic pathway may be to provide a virtually irreversible reaction sequence to drive an unfavorable rate-limiting reaction forward.

It is obvious that in bacteria specialized for using fatty acids and steroids as growth substrates, the number of ACSs is usually high (≫10). Indeed, phylogenetic analyses with S. denitrificans identified 10 further genes encoding putative ACSs next to the four experimentally verified ACSs. Among them, the SDENCholv1_20262 and SDENCholv1_20218 most possibly encode propionyl-CoA and acetyl-CoA synthetases, respectively, according to their similarities to experimentally verified enzymes (Fig. 2). The function of the remaining seven putative ACSs cannot be easily predicted. Likely functions are the activation of aliphatic fatty acids with chain lengths longer than C₃, e.g., palmitic acid, a known growth substrate of S. denitrificans (32), or the activation of C₂₆-oic acid analogues during growth with β-sitosterol, stigmasterol, or ergosterol (6). Finally, hydrolytic cleavage of ring A yields 1,17-dioxo-2,3-seco-androstan-3-oic acid (DSAO) in denitrifying steroid degraders (25), which needs to be activated to a CoA thioester to initiate a new round of β-oxidation (22). Altogether, S. denitrificans contains a highly versatile inventory of ACSs as a result of its specialization for steroid and fatty acid growth substrates.

MATERIALS AND METHODS

Materials.

Cholest-4-en-3-one-26-oic acid (analytical standard), pregn-4-en-3-one-22-oic acid, 3β-hydroxycholenic acid, lithocholic acid, cholic acid, deoxycholic acid, palmitic acid, decanoic acid, hexanoic acid, propionic acid, and CoA were purchased from Sigma-Aldrich (Darmstadt, Germany), Cayman Chemicals (Ann Arbor, MI, USA), or Santa Cruz Biotechnology (Heidelberg, Germany). In addition to the HIP synthesized by the authors (see below), it was kindly provided by L. Eltis (Department Microbiology & Immunology, University of British Columbia, Vancouver, Canada). Other chemicals and reagents were of analytical or high-performance liquid chromatography (HPLC) grade. Sterolibacterium denitrificans Chol1S (DSM 13999) was obtained from the Deutsche Sammlung für Mikroorganismen und Zellkulturen ([DSMZ] Braunschweig, Germany).

Cloning of acs genes.

Cholesterol-grown cells were used as the template for the extraction of genomic DNA using the Illustra bacteria genomicPrep kit (GE Healthcare). The genes encoding putative ACSs were amplified with specific primers (Table 4); an N-terminal Strep tag was fused to each subunit. After restriction and ligation into the vector pASK-IBA15plus (IBA Lifesciences), the construct was transformed into Escherichia coli BL21(DE3) (New England BioLabs).

TABLE 4.

Primers used for heterologous expression of ACS genes

Primer name (target)	Sequence (5′→3′)	Ta (°C)a	Restriction site
SDENv1_10299 forw (C24-ACS)	GGAGGCGGTACCACGGAAGCGCTGAAAC	56	KpnI
SDENv1_10299 rev (C24-ACS)	GGTGGAAAGCTTCTCAGTCGCGTTCAAAATC	55	HindIII
SDENv1_10305 forw (C22-ACS)	GGAGGCGGTACCGCAATTCCCCGCGAAG	56	KpnI
SDENv1_10305 rev (C22-ACS)	GGTGGAAAGCTTATCGGTAGTTTAGACGCC	54	HindIII
SDENv1_10766 forw (HIP-ACS)	GGAGGCGAATTCTCGCCGTTGCCGCAAAC	60	EcoRI
SDENv1_10766 rev (HIP-ACS)	GCCTCCAAGCTTGGAGTACGAATCAGCGCAACTGATAC	61	HindIII

^aT_a, annealing temperature.

Heterologous expression of genes and protein purification.

E. coli cells expressing the individual genes encoding ACSs were grown in LB medium at 30°C. At an optical density at 600 nm (OD₆₀₀) of 0.6, gene expression was induced with 250 μM isopropyl β-d-1-thiogalactopyranoside. The growth temperature was subsequently set to 16°C, and cells were harvested after 18 h. Five grams of cells was suspended in buffer A (50 mM HEPES-HCl [pH 7.5], 150 mM potassium chloride) with 0.1 mg · ml⁻¹ DNase I and lysed using a French pressure cell press. The cell lysate was ultracentrifuged and applied to a Strep-Tactin affinity column (GE Healthcare) using a fast protein liquid chromatography (FPLC) system (ÄKTA purifier; GE Healthcare). The removal of nonspecifically bound proteins and the elution of the heterologously expressed proteins were conducted using buffer A and buffer A containing 5 mM desthiobiotin, respectively. The eluted proteins were concentrated using a centrifugal concentrator (Sartorius). Protein concentrations were determined with the Bradford method (33), and protein purity was determined by SDS-PAGE.

Spectrophotometric enzyme assays.

For the determination of specific ACS activities in cell extracts, the time-dependent formation of CoA esters was followed using a discontinuous ultraperformance liquid chromatography (UPLC)-based assay (see below). Specific activities of heterologously produced ACSs were determined using a spectrophotometric assay that coupled AMP formation to NADH oxidation, as described previously (28) with slight modifications. Reactions were performed in a total volume of 0.1 ml with 0.1 M HEPES buffer (pH 8) containing 1 mM CoA, 1.5 mM ATP, 1 mM phosphoenolpyruvate (PEP), 4 U pyruvate kinase, 4 U adenylate kinase, 4 U lactate dehydrogenase, 500 μM NADH, 2.5 mM MgCl₂, 2 mM dithioerythritol (DTE), 7.8% (wt/vol) 2-hydroxypropyl-β-cyclodextrin, and 787 nM SDENchol_10299, 328 nM SDENchol_10305, or 820 nM SDENchol_10766. The mixture was incubated for 5 min at 30°C, and the reaction was initiated by adding 300 μM of the respective carboxylic acid substrates (from 5 to 10 mM stocks, dissolved in 70% isopropanol). All measurements were recorded on a Varian Cary 100 BIO UV-vis spectrophotometer with the standard software CaryWinUV kinetics (version 3.00).

Synthesis of HIP.

A mixture of 5α-OH-HIP and 7β-OH-HIP was purified from cultures of Pseudomonas stutzeri Chol1 Δscd3A with 12 mM succinate and 1 mM deoxycholate, as described previously (34). After the complete transformation of deoxycholate to the HIPs, the supernatant was acidified to pH 2 to 3 with HCl and extracted with ethyl acetate. The HIPs were resolved in Milli-Q H₂O. The complete transformation of deoxycholate and the purity of HIPs were determined by liquid chromatography-mass spectrometry (LC-MS) analysis as described for other steroid compounds (35).

UPLC analysis of CoA esters.

The products formed by purified ACSs or cell extracts were analyzed by an UPLC-based assay using a Waters Acquity H-class UPLC system equipped with a diode array detector (Waters, Eschborn, Germany) and a Knauer Bluespher 100-2 C₁₈ column (2 mm by 100 mm, 2-μm particle size; Knauer, Berlin, Germany). Samples (50 μl) were precipitated by the addition of 25 μl 50% (vol/vol) 1 M HCl or 100 μl methanol (MeOH), and the supernatants were applied to the column. If necessary (low concentrations), CoA esters were enriched by solid-phase extraction as described elsewhere (36). For UPLC analyses, a gradient of 15 to 90% acetonitrile in 10 mM ammonium acetate buffer and a flow rate of 0.3 ml · min⁻¹ were used for separation. The identification of the individual CoA esters was accomplished by their retention times, their UV-vis spectra, and by HRMS (see below and Table S2).

Liquid chromatography-mass spectrometry.

CoA esters were analyzed using a Waters Acquity I-class UPLC system with a Knauer Bluespher 100-2 C₁₈ column (2 mm by 100 mm, 2-μm particle size; Knauer, Berlin, Germany) coupled to a Waters Synapt G2-Si HDMS electrospray ionization (ESI)-quadrupole time of flight (Q-TOF) system (Waters, Eschborn, Germany). A gradient of 15 to 90% acetonitrile in 10 mM ammonium acetate buffer and a flow rate of 0.3 ml/min were used for separation. Any compounds were measured in MS positive mode with a capillary voltage of 3 kV, 150°C source temperature, 450°C desolvation temperature, 1,000 liter · min⁻¹ N₂ desolvation gas flow, and 100 liter · min⁻¹ N₂ cone gas flow. Collision-induced dissociation of precursor ions was performed using a collision energy of 20 V. LC-UV-vis analyses were conducted on a Waters Acquity H-class UPLC with a Knauer Eurospher C₁₈ column and the same gradient as described above. The evaluation of LC-MS metabolite data was performed using MassLynx (Waters); for the evaluation of LC-UV-vis data, MassLynx or Empower (Waters) was used. CoA esters were verified by the detection of their characteristic fragment ions at m/z 428.0367 (35).

Phylogenetic analyses.

The protein sequences of ACSs detected in the genome of S. denitrificans were aligned with a set of published sequences. The phylogenetic analyses were conducted on the server at www.phylogeny.fr (37) by using the MUSCLE algorithm (default parameters); Gblocks was used for data curation (default parameters) and PhyML for phylogenetic tree construction by the maximum likelihood method, built with 3,000 bootstrap replicates. The resulting tree was exported as a newick format and visualized in MEGA 7 (38).