Bile salts are amphiphilic steroids with digestive functions in vertebrates. Upon excretion, bile salts are degraded by environmental bacteria. Degradation of the bile salt steroid skeleton resembles the well-studied pathway for other steroids, like testosterone, while specific differences occur during side chain degradation and the initiating transformations of the steroid skeleton. Of the latter, two variants via either Δ1,4- or Δ4,6-3-ketostructures of the steroid skeleton exist for 7-hydroxy bile salts. While the Δ1,4 variant is well known from many model organisms, the Δ4,6 variant involving a 7-hydroxysteroid dehydratase as a key enzyme has not been systematically studied. Here, combined proteomic, bioinformatic, and functional analyses of the Δ4,6 variant in Sphingobium sp. strain Chol11 were performed. They revealed a degradation of the steroid rings similar to that of the Δ1,4 variant except for the elimination of the 7-OH as a key difference. In contrast, differential production of the respective proteins revealed a putative gene cluster for the degradation of the C5 carboxylic side chain encoding a CoA ligase, an acyl-CoA dehydrogenase, a Rieske monooxygenase, and an amidase but lacking most canonical genes known from other steroid-degrading bacteria. Bioinformatic analyses predicted the Δ4,6 variant to be widespread among the Sphingomonadaceae, which was verified for three type strains which also have the predicted side chain degradation cluster. A second amidase in the side chain degradation gene cluster of strain Chol11 was shown to cleave conjugated bile salts while having low similarity to known bile salt hydrolases. This study identifies members of the Sphingomonadaceae that are remarkably well adapted to the utilization of bile salts via a partially distinct metabolic pathway.
IMPORTANCE This study highlights the biochemical diversity of bacterial degradation of steroid compounds, in particular bile salts. Furthermore, it substantiates and advances knowledge of a variant pathway for degradation of steroids by sphingomonads, a group of environmental bacteria that are well known for their broad metabolic capabilities. Biodegradation of bile salts is a critical process due to the high input of these compounds from manure into agricultural soils and wastewater treatment plants. In addition, these results may also be relevant for the biotechnological production of bile salts or other steroid compounds with pharmaceutical functions.
Bile salts are multifunctional steroidal compounds that act as detergents in the digestion of lipophilic nutrients and exhibit signaling function in vertebrates (1, 2). The amphiphilic character of mammalian bile salts is determined by a carboxylic C5 side chain at C17 and one to three hydroxy groups on the steroid nucleus. Bile salts are produced from cholesterol in the liver, conjugated to taurine or glycine via amide bonds, and excreted into the gastrointestinal tract. In the intestine, free bile salts are released by deconjugation catalyzed by bile salt hydrolases produced by intestinal bacteria (3). Although most bile salts are reabsorbed (4, 5), about 0.4 to 0.6 g of bile salts are excreted per day by each human (6), adding up to about 18 metric tons of bile salts excreted per year by the population of a city with 100,000 inhabitants.
Upon excretion, bile salts become an energy and carbon source for environmental bacteria (7, 8), and several bile salt-degrading bacteria have been isolated from soils and aquatic habitats (9–12). These include Rhodococcus jostii RHA1 (13), Comamonas testosteroni CNB-2 and TA441 (14), Pseudomonas stutzeri Chol1 (9), Pseudomonas sp. strain DOC21 (11), Azoarcus sp. strain Aa7 (12), and Sphingobium sp. strain Chol11, formerly Novosphingobium sp. strain Chol11 (10). Aerobic bile salt degradation proceeds similarly to the degradation of other steroids such as cholesterol and can be divided into different phases (Fig. 1) (7, 8, 15): (1) oxidation of the A-ring, (2) side chain degradation, (3) oxygenolytic cleavage of ring B, and (4) oxygenolytic and hydrolytic degradation of the remaining seco-steroid. The first three steps may occur simultaneously (16–18).
In R. jostii RHA1, C. testosteroni TA441, and P. stutzeri Chol1, bile salt degradation proceeds through the well-elucidated 9,10-seco pathway. In phase 1, oxidative reactions at the A-ring generate intermediates with a Δ1,4-3-keto structure of the steroid skeleton (7, 13, 14). During degradation of the trihydroxy bile salt model-substrate cholate (I in Fig. 1), this leads to formation of Δ1,4-3-ketocholate (III in Fig. 1) (16). In phase 2, the bile salt side chain is degraded by the successive release of acetyl coenzyme A (acetyl-CoA) and propionyl-CoA (16, 19, 20). In actinobacteria such as R. jostii RHA1, acetyl-CoA is predicted to be released by β-oxidation (13). In proteobacteria such as P. stutzeri Chol1, acetyl-CoA is released by an aldolase-mediated cleavage reaction and subsequent oxidation of the resulting aldehyde group (19, 21). Both mechanisms of side chain degradation result in intermediates with C3 carboxylic side chains (16, 17, 21, 22). In actinobacteria as well as proteobacteria, this C3 side chain is released as propionyl-CoA by a second cycle of aldolytic cleavage reactions (22–24) resulting in C19 steroids, so-called androsta-1,4-diene-3,17-diones (ADDs) (9, 13, 20). In the case of cholate, 7,12β-dihydroxy-ADD (12β-DHADD, IV in Fig. 1) is formed by side chain degradation and subsequent stereoinversion of the 12-OH that is a requirement for further degradation (25).
In phase 3, degradation of the steroidal ring system is initiated by the introduction of a hydroxy group at C9 by the monooxygenase KshAB (17), which leads to spontaneous opening of the B-ring driven by the aromatization of ring A. This produces 9,10-seco intermediates such as 3,7,12-trihydroxy-9,10-seco-androsta-1,3,5-triene-9,17-dione (THSATD, V in Fig. 1). Phase 4 starts with the meta-cleavage of the aromatic A-ring and hydrolytic cleavage of the former A-ring, which results in differently hydroxylated H-methyl-hexahydro-indanone-propanoate (HIP) intermediates such as 3′,7-dihydroxy-HIP (DH-HIP, VI in Fig. 1) for cholate (7, 8). At this stage of degradation, intermediates from differently hydroxylated bile salts are channeled into a common pathway in P. stutzeri Chol1 (23). In this process, the former 12-OH is removed and a hydroxy group at former C7 is introduced into 7-deoxy bile salt derivatives during β-oxidation of the former B-ring. Further degradation of HIPs proceeds via β-oxidation of the former B-ring and hydrolytic cleavages of rings C and D (26, 27).
In contrast to this well-elucidated pathway, degradation of 7-hydroxy bile salts such as cholate (I in Fig. 1) proceeds differently in Sphingobium sp. strain Chol11 (10) but can also be divided into the four phases. After the initial formation of Δ4-3-keto-intermediates such as Δ4-3-ketocholate (II in Fig. 1) in phase 1, the hydroxy group at C7 is eliminated by the 7α-hydroxy steroid dehydratase Hsh2 (28). This leads to the formation of a double bond in the B-ring and to Δ4,6 intermediates such as 12-hydroxy-3-oxo-chol-4,6-dienoate (HOCDA, VII in Fig. 1) (10). This variant of the pathway will be referred to as Δ4,6 variant, in contrast to the Δ1,4 variant described above (29). As Δ4,6 derivatives of ADDs such as 12-hydroxy-androsta-1,4,6-triene-3,17-dione (HATD, VIII in Fig. 1) can be found in culture supernatants of strain Chol11 growing with cholate (10), side chain degradation seems to be the next phase (phase 2) of degradation. This is initiated by CoA activation catalyzed by CoA ligase SclA (30). In contrast to the model organisms using the Δ1,4 variant, Sphingobium sp. strain Chol11 growing with cholate produces no intermediates with a shortened side chain that can be found in culture supernatants (10). Interestingly, many genes for side chain degradation known from other model organisms are missing in strain Chol11; in particular, no homologs of the known steroid side chain aldolases or thiolases could be found by reciprocal BLASTp analyses (30, 31). The fact that several KshA homologs and many HIP degradation proteins are encoded in the genome of strain Chol11 suggests that phases 3 and 4, cleavage of the B-ring and further degradation of the seco-steroid, proceed similarly to the 9,10-seco pathway (30). Most homologs of steroid-degradation proteins are encoded in several clusters on the smaller chromosome 2 of strain Chol11, whereas the larger chromosome 1 and the two plasmids seems to be less involved in steroid degradation in this strain (30, 31).
To further elucidate bile salt degradation in strain Chol11 via the Δ4,6 pathway variant, differential proteome analyses of substrate-adapted cells were performed. For this, cholate (I in Fig. 1) was used as a model substrate for the Δ4,6 variant. This was compared to the 7-deoxy bile salt deoxycholate, since 7-deoxy bile salts cannot be degraded via Δ4,6 intermediates (28) and therefore require either completely different pathways or variations of one common pathway. As a further reference substrate, 12β-DHADD (IV in Fig. 1) was used because it does not possess a side chain and therefore might reveal proteins that are specific for side chain degradation.
RESULTS AND DISCUSSION
Proteome analyses reveal three gene clusters encoding bile salt degradation.
To assess inducibility of cholate degradation in Sphingobium sp. strain Chol11, cholate- and glucose-grown cells were compared. Suspensions of cholate-adapted cells depleted cholate more quickly than suspensions of glucose-adapted cells (Fig. S1A). When protein synthesis was inhibited by chloramphenicol, cholate was still completely degraded by cholate-adapted cells, but only a low percentage was depleted by glucose-adapted cells paralleled by continuous production of HOCDA (VII in Fig. 1) (Fig. S1A and B). These findings indicate inducibility of cholate degradation in strain Chol11, with enzymes for A-ring oxidation and 7α-dehydroxylation constitutively produced but those for degradation of the side chain and the steroid nucleus requiring de novo synthesis in glucose-grown cells.
Thus, the proteomic profiles of cholate-, deoxycholate-, and 12βDHADD-adapted cells were compared to that of glucose-adapted cells using two-dimensional difference gel electrophoresis (2D DIGE), whole-cell shotgun proteomics, and analyses of the membrane protein-enriched fractions (Table S1). In total, 44.6% of the 3,550 predicted proteins of strain Chol11 were detected (Fig. S2A). Proteins from the database of clusters of orthologous genes (COG) categories “inorganic ion transport and metabolism” and “lipid transport and metabolism” showed the highest increase in abundance in cholate-grown cells compared to that in glucose-grown cells (Fig. S2B).
The majority of proteins detected with significantly higher abundances in steroid- versus glucose-grown cells are encoded on chromosome 2, where most predicted steroid-degradation genes in strain Chol11 are located (31). Seventy-five proteins had significantly higher abundances in cells grown on bile salts versus cells grown on glucose, according to 2D-DIGE analysis. Most of these proteins are encoded in three gene clusters (Fig. 2), of which one (cluster 3) was previously predicted to encode steroid degradation (30).
The finding that the same set of proteins is produced in higher quantities during growth with both bile salts and most proteins were also produced for 12β-DHADD degradation indicates that degradation of these steroids generally involves the same proteins. This implies that both 7-hydroxy and 7-deoxy bile salts are degraded via the same pathway. Notably, however, a subset of proteins encoded in gene cluster 2 is differentially abundant, depending on the presence of a side chain. This implies that the proteins encoded in gene cluster 2 might be specific for side chain degradation. Based on proteome and bioinformatic analyses, we compiled a model of bile salt degradation in strain Chol11 (Fig. 3).
Several TonB-dependent receptor proteins (COG category “inorganic ion transport and metabolism”) were among the membrane proteins with the highest increases in abundance, including Nov2c232 (gene cluster 2), Nov2c378 (near gene cluster 3), and Nov2c659 (near gene cluster 4). TonB-dependent outer membrane transporter proteins (called TonB-dependent receptor protein) require the accessory proteins TonB, ExbB, and ExbD in the inner membrane to form a functional TonB system (32). These accessory proteins were identified in the membrane protein-enriched fractions of cells grown on all substrates with similar Mascot scores, indicating constitutive formation (Nov1c1853 to 1856). Transporters for the uptake of bile salts are not known in Proteobacteria, but there are indications that TonB-dependent receptors could be involved. First, they are generally known to participate in the import of complex growth substrates, such as lignin degradation compounds (33). Second, in Novosphingobium tardaugens NBRC16725, which was recently renamed Caenibius tardaugens NBRC16725 (34), a TonB-dependent receptor was upregulated in estradiol-grown cells, and a corresponding deletion mutant showed reduced growth with estradiol (35).
The TonB-dependent receptor Nov2c232 (28% identity to the TonB-dependent receptor that was deleted in C. tardaugens NBRC16725) is encoded in the putative side chain degradation cluster in close vicinity to a gene encoding a transporter of the major facilitator superfamily (MFS) (Nov2c225), and both were more abundant in cholate- and deoxycholate-grown cells (Fig. 4 and Table S1). These could be involved in transport of bile salts as well as early degradation intermediates such as Δ4-3-ketocholate (II in Fig. 1) or HOCDA (VII in Fig. 1), which are found in strain Chol11 culture supernatants (10, 29). Such transient extracellular accumulation of intermediates (10) is common for bile salt-degrading Proteobacteria and has been observed during bile salt degradation in soil (36). Thus, the above transporters could alternatively be involved in intermediate efflux.
Phase 1: A-ring oxidation and B-ring dehydration (blue section inFig. 3). (i) 3α-Hydroxysteroid dehydrogenase. The first step of bile salt degradation is the oxidation of 3-OH to a keto group by a 3α-hydroxy steroid dehydrogenase (3α-HSD) (28, 37). A putative 3α-HSD (Nov2c6) is encoded in gene cluster 1 (Fig. 4 and 5). This protein was present in lower abundances in glucose- versus steroid-grown cells (Fig. 4). Additional putative 3α-HSDs, Nov2c397 and Nov2c683, are encoded close to gene clusters 3 and 4, respectively, and were not differentially expressed. This, as well as Nov2c6 being present in low abundances in glucose-adapted cells, is consistent with biotransformation experiments showing that oxidation of the 3-OH is constitutive in strain Chol11.
(ii) 5β-Δ4-3-Ketosteroid dehydrogenase. The next step is the introduction of a double bond in the A-ring by 5β-Δ4-3-ketosteroid dehydrogenase (5β-Δ4-KSTD, named 5β-Δ4-KSTD1) (29). 5β-Δ4-KSTD1 is encoded in cluster 1 (Nov2c19) and is much more abundant in steroid- versus glucose-grown cells (>5.5-fold, Fig. 4).
(iii) 7α-Hydroxysteroid dehydratase. The key enzyme of the Δ4,6 pathway in strain Chol11, the 7α-hydroxysteroid dehydratase Hsh2 (Nov2c400), introduces a double bond in the B-ring by elimination of water (28). The hsh2 gene is located in close proximity to cluster 3. Hsh2 was shown previously to be active in glucose-grown cells (29), in agreement with its similar abundance in all cells (Fig. 4).
(iv) Δ1-3-Ketosteroid dehydrogenase. In the Δ1,4 variant bile salt degradation pathway, the next step is the introduction of a second double bond in the A-ring at C1 by a Δ1-3-ketosteroid dehydrogenase (Δ1-KSTD), which is a structural prerequisite for subsequent cleavage of the B-ring. The formation of Δ1,4,6 intermediates such as HATD (VIII in Fig. 1) during growth with bile salts shows that this reaction also occurs in strain Chol11 (10). Putative Δ1-KSTDs (38, 39) are Nov2c82 (encoded between gene clusters 1 and 2, not detected in any condition) and Nov2c695 (encoded close to gene cluster 4, about 2- to 3-fold increased abundance in steroid-grown cells) (Fig. 4). This expression pattern is in line with the whole-cell biotransformation assays for testing inducibility of bile salt degradation in which no Δ1-KSTD activity could be observed in glucose-adapted cells (Fig. S1). As no Δ1,4,6 intermediates with a side chain were reported for strain Chol11 during growth on bile salts (10, 28), this step might occur after side chain degradation in strain Chol11.
(v) Oxidation of the 12-OH-group. The previously observed formation of the 3,12-dioxo-chol-4,6-dienoate (DOCDA, XIII in Fig. 3) (10, 28) suggests the involvement of a 12α-dehydrogenase in the degradation of 12-hydroxy bile acids in strain Chol11. A similar reaction is catalyzed in C. testosteroni by SteA, which is active on 12α-hydroxy steroids without a side chain (25). In C. testosteroni, the resulting 12-oxo-steroids are then reduced to the corresponding 12β-steroids by SteB (Fig. 5B) before cleavage of the A-ring. The presence of a 12β-OH in the degradation intermediate HATD (VIII in Fig. 1) indicates that the reduction to a 12β-OH also takes place in strain Chol11. Homologs to SteA and SteB are encoded in cluster 1 (Nov2c15 and Nov2c16, respectively) and were detected with higher Mascot scores in steroid-grown cells (Fig. 4).
Phase 2: Side chain degradation (green section inFig. 3). (i) CoA activation of the side chain. The steroid-C24-CoA ligase SclA catalyzing the initial step of side chain degradation in strain Chol11 was previously described (30) and is encoded in the putative side chain degradation gene cluster 2 (Fig. 6A). SclA abundance, relative to that in glucose-grown cells, was 3.3-fold higher in cholate- and deoxycholate-grown cells and 1.3-fold higher in 12β-DHADD-grown cells (Fig. 4).
(ii) Desaturation of the side chain. Previous enzyme assays with cell extract of strain Chol11 indicated that a double bond is introduced into the CoA activated side chain (30). During the degradation of steroids, this dehydrogenation is catalyzed by α2β2-heterotetrameric acyl-CoA dehydrogenases (ACADs) (40, 41) or by homodimeric ACADs in which the two subunits are fused (42). Two predicted ACAD proteins in gene cluster 2 (Nov2c221 and Nov2c222) had 4- to 5-fold increased abundances in cholate- and deoxycholate-grown cells but not in 12β-DHADD-grown cells (Fig. 4). This suggests that they comprise an ACAD involved in side chain degradation of bile salts, but the location of any double bond formed by this ACAD is unknown.
(iii) Further side chain degradation. The next step in bile salt side chain degradation is typically the addition of water to the double bond (Fig. 6B). The enzymes catalyzing this hydration belong to the MaoC family, containing a hotdog fold domain, or the crotonase family (19, 43–45). These hydratases consist either of a single protein, such as the C5 side chain hydratases of P. stutzeri Chol1 and Mycobacterium tuberculosis (19, 45), or of two subunits, such as the C3 side chain hydratase of M. tuberculosis (43). Two proteins from the thioesterase superfamily with hotdog fold domains are encoded in cluster 2 (Nov2c219 and Nov2c220) adjacent to the ACAD genes. Nov2c219 was detected in only cholate- and deoxycholate-grown cells, and Nov2c220 was detected only in cholate-grown cells (Fig. 4). However, these proteins show only very low similarity (less than 20% identity) to known side chain hydratases, which suggests a different function for these proteins.
Homologs to other known side chain degradation enzymes, such as thiolases and aldolases, are not encoded in cluster 2 (30). In P. stutzeri Chol1 and R. jostii RHA1, the first cycle of side chain degradation leads to the release of acetyl-CoA and a shortened C3 side chain (17, 19). For the degradation of the C3 side chain, a second cycle of aldolytic cleavage with similar steps, including introduction of a double bond, addition of water, and aldolytic cleavage, is necessary and encoded in the genomes of both organisms (13, 16, 19, 23, 46, 47).
Our proteome and bioinformatic analyses of strain Chol11 revealed no enzymes that could potentially be involved in side chain cleavage or degradation of the C3 side chain. This suggests that side chain degradation in strain Chol11 is a mechanism other than aldolytic or thiolytic cleavage. A so-far unknown alternative mechanism might involve other proteins encoded in cluster 2, including putative hydroxysteroid dehydrogenases, amidases, and a Rieske monooxygenase (Fig. 6).
Phase 3: B-ring cleavage by the monooxygenase KshAB (yellow section inFig. 3). The first step in the degradation of the steroid nucleus is the cleavage of the B-ring by the KshAB monooxygenase system (9, 48). Five homologs of the oxygenase component KshA are encoded on chromosome 2 (Nov2c66, Nov2c228, Nov2c407, Nov2c430, and Nov2c440 with 27% to 32% identity to KshA of P. stutzeri Chol1) (Fig. 4). The numerous B-ring-cleaving KshA homologs in the genome of strain Chol11 strongly suggest that steroid nucleus degradation starts with 9,10-seco cleavage, although the resulting 9,10-seco-steroids have so far never been detected in cell-free supernatants of strain Chol11 cultures. Similar multiplicity of KshA homologs is also known from steroid-degrading Rhodococci (49, 50). Nov2c228, Nov2c407, Nov2c430, and Nov2c440 had increased abundances in steroid-grown cells and therefore are candidates for this reaction. However, Nov2c228 has the lowest similarity to KshAChol1, its encoding gene is localized in the side chain degradation gene cluster, and the abundance of Nov2c228 was increased only in bile salt-grown cells but not in 12β-DHADD-grown cells. Thus, a different role for this enzyme appears feasible. In this context, the similarities of the KshA oxygenases and Neverland oxygenases (22 to 24% identity of Nov2c228 and Nov2c407 to the Neverland oxygenases Nvd from Drosophila melanogaster, Caenorhabditis elegans, Bombyx mori, Xenopus laevis, and Danio rerio), which are involved in the production of ecdysteroids in arthropods (51, 52), could indicate a wider function for Rieske monooxygenases in steroid metabolism.
Interestingly, there are no distinct homologs of the reductase component KshB encoded in the genome of strain Chol11. This was also reported for C. tardaugens NBRC16725 (53). A flavodoxin reductase Novbp123 with 29% identity to KshB of P. stutzeri Chol1 is encoded on plasmid pSb of strain Chol11 (31). This enzyme was detected in the membrane protein-enriched fractions of all tested cells with similar Mascot scores, indicating that its synthesis is not regulated in response to steroid degradation. Novbp123 is encoded in a gene cluster together with a ferredoxin, a cytochrome c, and several exported and membrane proteins, suggesting that it is involved in membrane electron transport rather than bile salt degradation specifically (Fig. S3). This points at a different, KshB-independent electron shuttling mechanism for the KshA homologs in strain Chol11.
Phase 4: Complete degradation of the 9,10-seco intermediates (orange section inFig. 3). Most proteins required for the degradation of the 9,10-seco degradation intermediates, derived from both cholate and deoxycholate, and the respective HIP intermediates (27) are encoded in gene cluster 3 (Fig. 7A). All of these proteins are increased in abundance at least 1.5-fold in steroid-grown cells (Fig. 4). This confirms that degradation of the steroid nucleus proceeds via the 9,10-seco pathway. Cluster 3 is very similar to the cluster encoding testosterone degradation in C. tardaugens NBRC16725 (53). In both organisms, genes encoding homologs for the reductase component TesA1 of the 9,10-seco-steroid monooxygenase, TesA1A2, are missing, which could be a further hint at a different electron shuttling mechanism. However, in cluster 3 of strain Chol11, a flavin reductase (Nov2c347) is encoded near the gene for the oxygenase component, TesA2 (Nov2c349), indicating that Nov2c347 could serve as a TesA1 substitute. Regarding HIP degradation, a homolog for the gene encoding HIP-CoA ligase ScdA (54) is missing in strain Chol11, but the CoA transferase Nov2c359 could have this function.
(i) Fate of the 12-OH. In P. stutzeri Chol1, the 12β-OH is removed during C- and D-ring degradation by the elimination of water catalyzed by Hsh1 and subsequent reduction of the resulting double bond by Sor1 (23) (Fig. 5C). In strain Chol11, homologs to Hsh1 and Sor1 are encoded in cluster 1 near steA and steB (nov2c12 and nov2c13, respectively; Fig. 5A) and were found in increased abundances in steroid-grown cells (Fig. 4). A gene cluster with the same order of genes for the 12-OH transforming enzymes SteA, SteB, Hsh1, and Sor1 is present in C. tardaugens NBRC16725, and a similar cluster is present in C. testosteroni CNB-2, P. stutzeri Chol1, and Azoarcus sp. strain Aa7 (Fig. 5A), implicating a general role of these enzymes in bile salt degradation.
(ii) Channeling of 7-hydroxy and 7-deoxy bile salts into C- and D-ring degradation. During degradation of HIP intermediates (such as VI, XV, and XXII in Fig. 7), the remainder of the B-ring is degraded via β-oxidation, which requires a hydroxy group at the former C7 (23) (Fig. 7B). This hydroxy group is present in 7-hydroxy bile salts such as cholate, but during the degradation of 7-deoxy bile salts, such as deoxycholate, it has to be introduced into the propanoate side chain of the respective HIP intermediates. This is initiated by the introduction of a double bond by the heteromeric ACAD ScdC1C2 followed by the addition of water by the hydratase ScdD (23, 55, 56).
In Δ4,6 intermediates, the hydroxy group at C7 is eliminated. Thus, bile salt degradation by the Δ4,6 pathway variant results in HIP-like intermediates with a double bond in the propanoate side chain attached to ring C (XV in Fig. 1). This is the same intermediate as that found during degradation of 7-deoxy bile salts after introduction of a double bond (23), and the needed hydroxy group could be added by the hydratase, Nov2c364, which was detected in all steroid-grown cells in similar abundance (Fig. 4). Although the ACAD reaction catalyzed by ScdC1C2 is not needed for the degradation of cholate via the Δ4,6 variant, both subunits, Nov2c367 and Nov2c361, were found in all steroid-grown cells with similar Mascot scores (Fig. 4). Thus, it is possible that the Δ6 double bond had meanwhile been reduced, necessitating the ACAD.
As the hydroxy group eliminated by Hsh2 must again be added at the stage of HIP intermediates, the benefit of the elimination remains unclear. It might be related to the fact that many intermediates of bile salt degradation are excreted in significant amounts during growth of bile salt-degrading bacteria not only in laboratory cultures but also in soil samples (10, 36). Other bile salt-degrading strains, such as P. stutzeri Chol1, that exclusively use the Δ1,4 degradation pathway are unable to utilize Δ4,6 compounds as growth substrates (10, 28). To this end, the dehydration might provide a way for strain Chol11 to secure the individual availability to these important carbon sources in natural habitats.
Widespread distribution of the Δ4,6 variant within the family Sphingomonadaceae.
(i) Prediction of the Δ4,6 variant pathway in steroid-degrading bacteria. Apart from strain Chol11, steroid degradation has been reported in other sphingomonads such as C. tardaugens NBRC16725 (53, 57) and Sphingomonas sp. strain KC8 (58). In addition, Sphingomonas sp. strain Chol10 was also isolated as an HOCDA-degrading bacterium (10).
To investigate the prevalence of the Δ4,6 variant pathway within the family Sphingomonadaceae, we searched all 398 complete and draft genomes from the genera Sphingobium, Novosphingobium, and Sphingomonas available from the NCBI RefSeq database for the simultaneous presence of key steroid-degradation proteins and Hsh2, the key protein of the Δ4,6 pathway. First, steroid-degrading bacteria were predicted using 23 hidden Markov models (HMMs) representing 10 key proteins of canonical steroid nucleus degradation (59). Based on the presence of 7 out of 10 key proteins, including KshA and TesB, 53 genomes were predicted to encode steroid degradation. Second, Hsh2 orthologs were determined in these genomes by BLASTp analyses using Hsh2 of Sphingobium sp. strain Chol11 (28) as query. Thirty-nine genomes containing both the steroid-degradation genes and hsh2 were found. To further confirm the prediction of these proteins being involved in steroid degradation in these organisms, a reciprocal BLASTp analysis was performed using steroid-degradation proteins from P. stutzeri Chol1 and Hsh2Chol11 as queries (Fig. 8). Thirty-eight genomes were confirmed to encode Hsh2 orthologs, and all of them contained orthologs for the majority of key proteins for steroid nucleus degradation. However, only a small subset of side chain degradation proteins is encoded in most of the Sphingobium and Novosphingobium genomes. This was in contrast to the Sphingomonas genomes where orthologs of the genes encoding CoA ligase, heteromeric ACAD, heteromeric hydratase, and aldolase for the degradation of the C3 side chain from P. stutzeri Chol1 are present.
This suggests that bile salt degradation via Δ4,6 intermediates is widely distributed among members of Sphingomonas, Sphingobium, and Novosphingobium, while the distinct side chain degradation mechanism proposed for strain Chol11 may occur in most members of Sphingobium and Novosphingobium.
(ii) Bile salt degradation in strains predicted to use the Δ4,6 variant. To investigate whether strains predicted to degrade bile salts via the Δ4,6 pathway variant did so, a selection of type strains with complete genome sequences was analyzed for bile salt degradation. Novosphingobium aromaticivorans F199, Sphingobium herbicidovorans MH, and C. tardaugens NBRC16725 were tested for the degradation of the 7α-hydroxy bile salts, cholate and chenodeoxycholate, and the 7-deoxy bile salt, deoxycholate (structures in Fig. 9). Accumulation of Δ4,6 intermediates was also monitored.
Strains MH, NBRC16725, and F199 grew on all tested bile salts (Fig. 9B, D, and F). While all bile salts were completely degraded, all three strains transiently accumulated the characteristic Δ4,6 intermediates HOCDA (VII in Fig. 1) and DOCDA (XIII in Fig. 3) (Fig. 9C, E, G, and I). In addition, during degradation of the 7α-hydroxy bile salts, strain F199 formed the intermediates XXIV and XXV that also have a Δ4,6 structure and lack a side chain according to their UV and mass spectra, respectively (Fig. 9I). This supports our prediction that the Δ4,6 variant is the predominant pathway for bile acid degradation in Sphingobium and Novosphingobium strains.
Like strain Chol11, strains NBRC16725, MH, and F119 were able to fully metabolize side chain-bearing bile salts, despite not encoding all side chain catabolic enzymes that are known from other bile salt-degrading bacteria. A BLASTp analysis revealed the presence of gene clusters with high similarity to gene cluster 2 of strain Chol11 in all three strains (Fig. 6A), although the clusters from the Novosphingobium strains, NBRC16725 and F119, were notably missing several genes. In particular, homologs of the thioesterase superfamily proteins Nov2c219 and Nov2c220, the only candidates for side chain hydratases in strain Chol11, were absent in the genome of NBRC16725. Homologs of SclA, the putative ACAD Nov2c221/Nov2c222, the Rieske monooxygenase Nov2c228, and the putative amidases Nov2c227 and Nov2c229 were found in the predicted side chain degradation clusters of all three strains.
Degradation of conjugated bile salts.
Strain Chol11 grew on the conjugated bile salts, taurocholate and glycocholate (X and XI, respectively, in Fig. 3), but not on the free amino acids glycine or taurine (Fig. 10B and E). The inability to grow on these amino acids agrees with the similar biomass yields on taurocholate and glycocholate versus on unconjugated cholate (28). These results suggest that glycine or taurine is removed by amidases prior to metabolism of the free bile salts. Two candidate amidases, Nov2c227 and Nov2c229, encoded in the predicted side chain degradation gene cluster, were 2.3- to 8.9-fold more abundant in cholate- and deoxycholate-grown cells.
To test their potential role in deconjugation of bile salts, Nov2c227 and Nov2c229 were each heterologously expressed in Escherichia coli MG1655. Cell suspensions and cell extracts of E. coli expressing Nov2c227 transformed both taurocholate and glycocholate to cholate (Fig. 10). Cell suspensions and cell extracts of E. coli expressing the other amidase candidate, Nov2c229, as well as the empty vector control did not substantially catalyze this transformation. Additionally, compounds 2 Da lighter than the conjugated bile salts and cholate were present in almost all assays. E. coli possesses a 7α-hydroxy steroid dehydrogenase, which catalyzes the oxidation of the 7α-OH of conjugated and free bile salts (60). Therefore, it is likely that the additional compounds are the 7-keto derivatives of the conjugated bile salts and cholate (XXVI, XXVII, and XXVIII in Fig. 10). Since Nov2c227 cleaves the conjugated bile salts, glycocholate and taurocholate, it was named bile salt amidase (Bsa).
Interestingly, Bsa contains a TAT signal peptide as predicted by SignalP, indicating that this protein is secreted, and deconjugation of conjugated bile salts takes place in the periplasmatic or extracellular space prior to bile salt uptake. In contrast to the well-elucidated N-terminal nucleophile family bile salt hydrolases from probiotic lactic acid bacteria such as Bifidobacterium longum (61), Nov2c227 belongs to the large family of amidases. This indicates a different evolutionary origin for bile salt hydrolases in steroid degradation as opposed to those in intestinal bacteria, despite their common function. Homologs of Bsa are encoded in genomes of many strains of the family Sphingomonadaceae, e.g., EGO055_026080 from C. tardaugens NBRC16725. Furthermore, C. testosteroni KF-1 was reported to degrade taurocholate (62), and its genome encodes two Bsa homologs (identity for both, 44%). The respective homologs ORF25 and ORF26 from model organism C. testosteroni TA441 are encoded in its steroid-degradation megacluster (14, 27). The function of Nov2c229 remains unknown so far.
Conclusion and general discussion.
Differential proteome analyses of Sphingobium sp. strain Chol11 together with bioinformatic analyses revealed a comprehensive set of candidate proteins for the complete degradation of bile salts. From these data, the complete degradation pathway for the steroid nucleus could be deduced, which is a mosaic of unknown reactions (especially side chain degradation) and reactions known from other steroid-degrading organisms (Fig. 3). Apparently, despite the variation in the first steps of 7-hydroxy bile salt degradation in strain Chol11, which leads to the introduction of a double bond in the B-ring by water elimination catalyzed by Hsh2, further degradation of the Δ4,6 intermediates is very similar to the 9,10-seco pathway known from other organisms and steroids. However, the Δ4,6 variation produces steroid intermediates that cannot be utilized by other organisms such as P. stutzeri Chol1 (10), suggesting the involvement of specialized enzymes in organisms that are able to utilize Δ4,6 intermediates. The prediction and identification of further organisms degrading bile salts via this Δ4,6 variant of the 9,10-seco pathway demonstrate a wide distribution of this variant in the family Sphingomonadaceae.
In addition, the interpretation of the proteome analyses in conjunction with bioinformatic analyses resulted in the identification of a side chain degradation gene cluster encoding key proteins that are presumably responsible for side chain removal by a yet-unknown mechanism. The bile salt hydrolase activity of Bsa, as well as the CoA ester formation by the CoA ligase SclA (30), encoded in this gene cluster further corroborates the involvement of this cluster in bile salt side chain degradation. The subsequent introduction of a double bond into a yet-unknown position is presumably catalyzed by the heteromeric ACAD Nov2c221/222. The absence of several other side chain degradation genes from the genomes of strain Chol11, C. tardaugens NBRC16725, and N. aromaticivorans F199 suggests a mechanism different from thiolytic and aldolytic cleavage. Interestingly, homologs of SclA, the putative ACAD Nov2c221/Nov2c222, the Rieske monooxygenase Nov2c228, and the putative amidase Nov2c229 were found in all Sphingomonads confirmed to use the Δ4,6 variant for bile salt degradation. However, the functions of this Rieske monooxygenase and the putative amidase during bile salt side chain degradation remain unclear. In addition to being highly conserved and encoded near confirmed side chain degradation genes in all four tested strains, they seem to be formed exclusively in response to side chain-containing bile salts in strain Chol11. Further investigations regarding this gene cluster using molecular methods are under way. Unfortunately, these analyses are impaired by the complicated genetic modification of strain Chol11.
Interestingly, several members of the family Sphingomonadaceae are adapted to growth with steroids and bile salts. Moreover, both Sphingobium sp. strain Chol11 and C. tardaugens NBRC16725 grow only slowly with nonsteroidal substrates (57), and some genes for early steps of bile salt degradation are apparently constitutively induced. Together with the prevalence of bile salt degradation in strains that had originally been isolated with xenobiotic compounds (63, 64), this indicates that bile salt degradation may be a conserved property of these organisms and calls attention to its evolutionary origin.
MATERIALS AND METHODS
Cultivation of bacteria.
If not indicated otherwise, Sphingobium sp. strain Chol11 (DSM 110934) (10), S. herbicidovorans MH (DSM 11019) (64), N. aromaticivorans F199 (DSM12444) (65), C. tardaugens NBRC16725 (DSM 16702) (57), and E. coli MG1655 (DSM 18039) (66) were cultivated in HEPES-buffered medium B (MB) (67). If not indicated otherwise, wild-type strains other than E. coli were grown with 1 mM cholate as the sole carbon source, whereas E. coli MG1655 was grown with 15 mM glucose. For maintenance, strain Chol11 and C. tardaugens NBRC16725 were grown on MB agar with 1 mM cholate, S. herbicidovorans MH was grown on CASO agar (Merck Millipore, Burlington, MA, USA), N. aromaticivorans F199 was grown on MB agar with 15 mM glucose, and E. coli MG1655 was grown on LB agar (68). For cultivation of strains containing pBBR1MCS-5 (69), 20 μg ml−1 gentamicin was added. When bile salts were added, gentamicin was omitted. Liquid cultures with volumes up to 5 ml were cultivated in 10-ml test tubes at 200 rpm; larger cultures were cultivated in 500-ml Erlenmeyer flasks without baffles at 120 rpm. All strains were incubated at 30°C, except for strain maintenance of E. coli strains (37°C). For agar plates, 1.5% (wt/vol) Bacto agar (BD, Sparks, USA) was added.
Cholate (≥99%), deoxycholate (≥97%), and glycocholate (≥97%) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Chenodeoxycholate (≥98%) was purchased from Carl Roth (Karlsruhe, Germany). Taurocholate (>98%) was purchased from Fluka (Buchs, Switzerland).
For growth experiments, 3- to 5-ml main cultures were inoculated to a predefined optical density at 600 nm (OD600; about 0.02) directly from liquid starter cultures and growth was determined by measurement of OD600 (Camspec M107, Spectronic Camspec, Leeds, UK). Growth with bile salts was tested using 1 mM cholate, 1 mM chenodeoxycholate, or 1 mM deoxycholate. Growth with conjugated bile salts was tested using 1 mM taurocholate, 1 mM glycocholate, 15 mM glycine, or 15 mM taurine. At suitable time points, samples for high-performance liquid chromatography mass spectrometry (HPLC-MS) measurements were withdrawn.
Starter cultures were grown with 1 mM cholate for strain Chol11, N. aromaticivorans F199, and C. tardaugens NBRC16725 or 12 mM succinate for S. herbicidovorans MH. Starter cultures were incubated overnight for about 15 h.
Induction of cholate degradation in strain Chol11 was tested using suspensions of cholate- and glucose-grown cells with 1 mM cholate and 10 μg ml−1 chloramphenicol to inhibit de novo protein synthesis as described in reference 29.
For determining whole-cell biotransformation of various steroid compounds by E. coli MG1655 expressing amidase genes, 5-ml starter cultures of E. coli MG1655 pBBR1MCS-5 as the empty vector control, E. coli MG1655 pBBR1MCS-5::nov2c227, or E. coli MG1655 pBBR1MCS-5::nov2c229 in LB with 20 μg ml−1 gentamicin were incubated for about 15 h. Cells were harvested by centrifugation in 2-ml reaction tubes (8,000 × g, ambient temperature, 3 min), washed with MB, and resuspended in MB to an OD600 of about 1. Cell suspensions were incubated for several days at 30°C at 200 rpm after addition of either 1 mM taurocholate or 1 mM glycocholate. Thirty millimolar glucose was added to all preparations as a carbon source.
For monitoring biotransformations of conjugated bile salts by cell extracts of E. coli MG1655 expressing amidase genes, 50 ml LB with 20 μg ml−1 gentamicin was inoculated to an initial OD600 of 0.015 with the aforementioned strains of E. coli MG1655 and incubated for about 18 h with addition of 0.2 mM isopropyl-β-d-thiogalactopyranoside after about 3 h. Cells were harvested (8,000 × g, 4°C, 8 min), washed with 10 mM MOPS buffer (morpholinepropanesulfonic acid; pH 7.8), and resuspended in about 2 ml 50 mM MOPS buffer (pH 7.8). Cells were disrupted in 15-ml conical centrifugation tubes by ultrasonication on ice (amplitude 60%, cycle 0.5, UP200S, Hielscher Ultrasonics, Teltow, Germany) for 8 min with a 1-min break after 4 min. Cell debris was removed by centrifugation (25,000 × g, 4°C, 30 min) in 2-ml reaction tubes. Cell extracts were used immediately in enzyme assays or stored at −20°C for later use. Enzyme assays (1 ml) contained 50 mM MOPS (pH 7.8), 1 mM substrate, and 100 μl cell extract. Assays were incubated for 30 min to 15 h at 30°C. Samples of all enzyme assays were subjected to HPLC-MS measurements.
Generation of substrate-adapted cells. Strain Chol11 was freshly thawed from a cryopreservation culture for each cultivation on MB agar and streaked twice onto plates with 1 mM cholate. After 1 to 4 days, the strain was transferred to new agar plates containing 1 mM cholate, 1 mM deoxycholate, 2 mM 12β-DHADD, or 15 mM glucose. Cells were further transferred twice to the same medium after incubation for 2 days with steroidal substrates or 3 to 4 days with glucose. From these plates, 12 5-ml MB starter cultures containing the same carbon source as the respective solid medium were inoculated and incubated for about 15 h for steroidal compounds or 30 h for glucose. Subsequently, 1 mM bile salt, 2 mM 12β-DHADD, or 15 mM glucose was added, and the starter cultures were incubated for further 1 to 1.5 h. Starter cultures with the same substrate were pooled and used for inoculation of 12 100-ml main cultures in 500-ml Erlenmeyer flasks without baffles containing the same carbon sources at an OD600 of 0.015. Main cultures were incubated at 30°C and 200 rpm, and growth was monitored until the half maximal OD600 was reached. Cultures were harvested in two 50-ml conical centrifugation tubes by centrifugation (5,525 × g, 4°C, 30 min) and kept on ice. Cells were washed with 25 ml 100 mM Tris buffer (pH 7.5 with HCl) containing 5 mM MgCl2, and cells of the same culture were pooled. After centrifugation, cells were resuspended in 625 μl of the same buffer, transferred to 2-ml reaction tubes, and harvested by centrifugation (14,000 × g, 4°C, 15 min). After weighing, pellets were snap-frozen using liquid nitrogen and stored at −70°C.
Profiling of soluble proteins by 2D DIGE and protein identification by MALDI-TOF/TOF. Soluble proteins were extracted from cells of strain Chol11, and 2D DIGE was performed essentially as described previously (70). Per growth condition, four biological replicate samples were prepared and 50 μg total protein was used for minimal labeling with 200 pmol of Lightning SciDye DIGE fluorescence dyes (SERVA Electrophoresis GmbH, Heidelberg, Germany). Glucose-adapted cells served as the reference state and were labeled with Sci5. Protein extracts of the other three (test) states were each labeled with Sci3. The internal standard contained equal amounts of all test and the reference state(s) and was labeled with Sci2. First-dimension separation by isoelectric focusing (IEF) was conducted with 24-cm-long IPG strips (pH 3 to 11 nonlinear; GE Healthcare) run in a Protean i12 system (Bio-Rad, Munich, Germany). The IEF program used was as follows: 50 V for 13 h, 200 V for 1 h, 1,000 V for 1 h, gradual gradient to 10,000 V within 2 h, and 10,000 V until 80,000 V · h was reached. Second-dimension separation according to molecular size was done by SDS-PAGE (12.5% gels, vol/vol) using an Ettan DALT twelve system (GE Healthcare).
Directly after electrophoresis, 2D DIGE gels were digitalized using a CCD camera system (Intas Advanced 2D Imager; Intas Science Imaging Instruments GmbH, Göttingen, Germany) (71). Cropped gel images were analyzed with the SameSpots software (version 18.104.22.168, TotalLab, Newcastle upon Tyne, UK), and spots with changes in abundance of ≥|1.5|-fold and an analysis of variance (ANOVA) P value of ≤1 × 10−4 were accepted as significant (72). Spots of interest were excised from at least two separate, preparative colloidal Coomassie brilliant blue-stained gels (300 μg protein load) using the EXQuest spot cutter (Bio-Rad) and subsequently washed and tryptically digested as described recently (73).
Sample digests were spotted onto Anchorchip steel targets (Bruker Daltonik GmbH, Bremen, Germany) and analyzed with an UltrafleXtreme matrix-assisted laser desorption ionization–tandem time of flight (MALDI-TOF/TOF) mass spectrometer (Bruker Daltonik GmbH) as described recently (73). Peptide mass fingerprint (PMF) searches were performed with a Mascot server (version 2.3; Matrix Science, London, UK) against the translated genome of strain Chol11 with a mass tolerance of 25 ppm. Five lift spectra were collected to confirm PMF identification, and three additional spectra were acquired of unassigned peaks applying feedback by the ProteinScape platform (version 3.1; Bruker Daltonik GmbH). In case of failed PMF identification, eight lift spectra of suitable precursors were acquired. Tandem mass spectrometry (MS/MS) searches were performed with a mass tolerance of 100 ppm. For both MS and MS/MS searches, Mascot scores not meeting the 95% certainty criterion were not considered significant. A single miscleavage was allowed (enzyme trypsin), and carbamidomethyl (C) and oxidation (M) were set as fixed and variable modifications, respectively.
Shotgun proteomic analysis. For shotgun analysis, cell pellets of three biological replicate samples per growth condition were suspended in lysis buffer, cells were disrupted, and the debris-free fraction was reduced, alkylated, and subjected to tryptic digest as described previously (73). Obtained peptides were separated by nanoLC (UltiMate 3000; ThermoFisher Scientific, Germering, Germany) using a trap-column (C18, 5 μm bead size, 2 cm length, 75 μm inner diameter; ThermoFisher Scientific) and a 25-cm analytical column (C18, 2 μm bead size, 75 μm inner diameter; ThermoFisher Scientific) applying a 360-min linear gradient (74). The nanoLC eluent was continuously analyzed by an online-coupled ion-trap mass spectrometer (amaZon speed ETD; Bruker Daltonik GmbH) using the captive spray electrospray ion source (Bruker Daltonik GmbH). The instrument was operated in positive mode with a capillary current of 1.3 kV and drygas flow of 3 liters min −1 at 150°C. Active precursor exclusion was set for 0.2 min. Per full scan MS, 20 MS/MS spectra of the most intense masses were acquired. Protein identification was performed with ProteinScape as described above, including a mass tolerance of 0.3 Da for MS and 0.4 Da for MS/MS searches and applying a target decoy strategy (false discovery rate < 1%).
Analysis of the membrane protein-enriched fraction. Total membrane fractions were prepared from two biological replicates per substrate condition as described in reference 73. The obtained protein content was determined with the RC-DC assay (Bio-Rad) and 8 μg total protein separated by SDS-PAGE gels (∼7 cm separation gel). Following staining with Coomassie brilliant blue (75), each sample lane was cut into 8 slices and each slice into small pieces of ∼1 to 2 mm³ prior to washing, reduction, alkylation, and tryptic digest (73). Separation and mass determination were performed as described above, using a 100-min linear gradient. Identified proteins (performed as described above) of each slice per sample were compiled using the protein extractor of the ProteinScape platform.
Cloning was performed according to standard procedures and as described elsewhere (28).
For expression of amidase genes in E. coli MG1655, genes were amplified using the respective primer pairs expfor/exprev (Table 1) and ligated into vector pBBR1MCS-5. The respective ligation products were transferred to E. coli MG1655 by heat shock transformation. Presence and correct ligation of plasmids were confirmed by colony PCR and sequencing using M13 primers.
TABLE 1 Primers used for construction of plasmids for heterologous expression
Steroid compounds were analyzed by HPLC-MS. For this, samples were centrifuged (>16,000 × g, ambient temperature, 5 min) to remove cells and particles. Supernatants were stored at −20°C and centrifuged again prior to measurement. Samples from cell suspension experiments with E. coli MG1655 were directly frozen at −20°C and centrifuged only after thawing to break the cells. HPLC-MS measurements were performed with a Dionex Ultimate 3000 HPLC (ThermoFisher Scientific, Waltham, MA, USA) with a UV/visible light diode array detector and coupled to an ion-trap amaZon speed mass spectrometer (Bruker Daltonik, Bremen, Germany) with an electrospray ion source. Compounds were separated over a reversed phase C18 column (150 × 3 mm, Eurosphere II, 100-5 C18; Knauer Wissenschaftliche Geräte, Berlin, Germany) at 25°C. Samples of cell suspensions for testing the induction of steroid degradation in Sphingobium sp. strain Chol11 were measured as described in reference 28, whereas all other samples were measured as described in reference 29.
Bile salt concentrations were determined as peak area from base peak chromatograms measured in negative ion mode. Intermediates were identified according to retention time, UV and MS spectra, and comparison with known compounds. Structures of unknown metabolites were proposed on the basis of retention time as well as UV and MS spectra.
Searches for homologous proteins and determinations of protein similarities were performed using the BLASTp algorithm (76, 77). Protein similarities were calculated from global alignments in the BLAST suite using the Needleman-Wunsch algorithm (76, 78). Protein domains and families were predicted using Interpro (79) and the Conserved Domains Database (80, 81). For functional annotation of strain Chol11, the eggNOG database (82) was used.
For the bioinformatic identification of other sphingomonads using the Δ4,6 variant of the 9,10-seco pathway, first a database of putative steroid degraders was set up similar to that set up in reference 83. On 18 October 2018, all complete and draft genomes available for the genera Sphingobium, Novosphingobium, and Sphingomonas were downloaded from the RefSeq database (version 89). Using 23 hidden Markov models (HMMs) (59), these genomes were searched for 10 homologs of steroid-degradation proteins using HMMER version 3 (84) using a maximum E value of 10−25. HMMs for Δ1-KSTD (KstD), KshA, TesA2 (HsaA in R. jostii RHA1), TesB (HsaC), TesE (HsaE), TesF (HsaG), TesG (HsaF), ScdK (IpdC), ScdL1 (IpdA), and ScdL2 (IdpB) were used. Bacteria were predicted to be able to degrade steroids when their genomes encoded homologs of 7 out of the 11 steroid-degradation key enzymes, including KshA and TesA2. With these genomes, a reciprocal BLASTp search (83) was conducted using the key enzyme of the Δ4,6 pathway variant Hsh2 from strain Chol11 (28) as the query using E value and identity cutoffs of 10−25 and 35%, respectively. These values were optimized empirically comparing analyses using Hsh2Chol11 as well as BaiE from C. scindens (UniProt ID P19412), which has a similar function in a different pathway (85). The results of both analyses were compared, and E value and identity cutoffs were chosen to ensure that proteins were identified as homologs of only one of these dehydratases. All genomes from putative steroid degraders containing Hsh2 homologs were subjected to a reciprocal BLASTp analysis using known steroid-degradation proteins from P. stutzeri Chol1 as queries. For data analysis and preparation of figures, R (v3.5.1) was used together with the packages circlize (v0.4.8), genoPlotR (v0.8.9), ggplot2 (v3.2.1), ComplexHeatmap (v1.18.1), gplots (v3.0.3), RColorBrewer (v1.1-2), VennDiagram (v1.6.20), ape (v5.3), reshape2 (v1.4.3), tidyverse (v1.3.0), and readxl (v1.3.1).
We thank Karin Niermann and Kirsten Heuer (both Münster) as well as Christina Hinrichs (Oldenburg) for excellent experimental support and Florentin Schmidt for help with planning of cloning.
This work was funded by two grants of the Deutsche Forschungsgemeinschaft (DFG projects PH71/3-2 and INST 211/646-1 FUGG) to B.P. and a scholarship of the DAAD Stiftung in cooperation with the Prof. Dr. Bingel-Stiftung to F.F.
Jones BV, Begley M, Hill C, Gahan CGM, Marchesi JR. 2008. Functional and comparative metagenomic analysis of bile salt hydrolase activity in the human gut microbiome. Proc Natl Acad Sci U S A 105:13580–13585.
Holert J, Yücel O, Suvekbala V, Kulić Ž, Möller H, Philipp B. 2014. Evidence of distinct pathways for bacterial degradation of the steroid compound cholate suggests the potential for metabolic interactions by interspecies cross-feeding. Environ Microbiol 16:1424–1440.
Merino E, Barrientos A, Rodríguez J, Naharro G, Luengo JM, Olivera ER. 2013. Isolation of cholesterol- and deoxycholate-degrading bacteria from soil samples: evidence of a common pathway. Appl Microbiol Biotechnol 97:891–904.
Yücel O, Borgert SR, Poehlein A, Niermann K, Philipp B. 2019. The 7α-hydroxysteroid dehydratase Hsh2 is essential for anaerobic degradation of the steroid skeleton of 7α-hydroxyl bile salts in the novel denitrifying bacterium Azoarcus sp. strain Aa7. Environ Microbiol 21:800–813.
Birkenmaier A, Holert J, Erdbrink H, Moeller HM, Friemel A, Schoenenberger R, Suter MJ-F, Klebensberger J, Philipp B. 2007. Biochemical and genetic investigation of initial reactions in aerobic degradation of the bile acid cholate in Pseudomonas sp. strain Chol1. J Bacteriol 189:7165–7173.
Holert J, Jagmann N, Philipp B. 2013. The essential function of genes for a hydratase and an aldehyde dehydrogenase for growth of Pseudomonas sp. strain Chol1 with the steroid compound cholate indicates an aldolytic reaction step for deacetylation of the side chain. J Bacteriol 195:3371–3380.
Holert J, Kulić Ž, Yücel O, Suvekbala V, Suter MJF, Möller HM, Philipp B. 2013. Degradation of the acyl side chain of the steroid compound cholate in Pseudomonas sp. strain Chol1 proceeds via an aldehyde intermediate. J Bacteriol 195:585–595.
Birkenmaier A, Möller HM, Philipp B. 2011. Identification of a thiolase gene essential for β-oxidation of the acyl side chain of the steroid compound cholate in Pseudomonas sp. strain Chol1. FEMS Microbiol Lett 318:123–130.
Holert J, Yücel O, Jagmann N, Prestel A, Möller HM, Philipp B. 2016. Identification of bypass reactions leading to the formation of one central steroid degradation intermediate in metabolism of different bile salts in Pseudomonas sp. strain Chol1. Environ Microbiol 18:3373–3389.
Barrientos Á, Merino E, Casabon I, Rodríguez J, Crowe AM, Holert J, Philipp B, Eltis LD, Olivera ER, Luengo JM. 2015. Functional analyses of three acyl-CoA synthetases involved in bile acid degradation in Pseudomonas putida DOC21. Environ Microbiol 17:47–63.
Horinouchi M, Hayashi T, Koshino H, Malon M, Yamamoto T, Kudo T. 2008. Identification of genes involved in inversion of stereochemistry of a C-12 hydroxyl group in the catabolism of cholic acid by Comamonas testosteroni TA441. J Bacteriol 190:5545–5554.
Crowe AM, Casabon II, Brown KL, Liu J, Lian J, Rogalski JC, Hurst TE, Snieckus V, Foster LJ, Eltis LD. 2017. Catabolism of the last two steroid rings in Mycobacterium tuberculosis and other bacteria. mBio 8:1–16.
Horinouchi M, Koshino H, Malon M, Hirota H, Hayashi T. 2019. Steroid degradation in Comamonas testosteroni TA441: identification of the entire β-oxidation cycle of the cleaved B ring. Appl Environ Microbiol 85:1–17.
Yücel O, Drees S, Jagmann N, Patschkowski T, Philipp B. 2016. An unexplored pathway for degradation of cholate requires a 7α-hydroxysteroid dehydratase and contributes to a broad metabolic repertoire for the utilization of bile salts in Novosphingobium sp. strain Chol11. Environ Microbiol 18:5187–5203.
Feller FM, Marke G, Drees SL, Wöhlbrand L, Rabus R, Philipp B. 2021. Substrate inhibition of 5β-Δ4–3-ketosteroid dehydrogenase in Sphingobium sp. strain Chol11 acts as circuit breaker during growth with toxic bile salts. Front Microbiol 12:655312.
Yücel O, Holert J, Ludwig KC, Thierbach S, Philipp B. 2018. A novel steroidcoenzyme A ligase from Novosphingobium sp. strain Chol11 is essential for an alternative degradation pathway for bile salts. Appl Environ Microbiol 84:16.
Yücel O, Wibberg D, Philipp B, Kalinowski J. 2018. Genome sequence of the bile salt-degrading bacterium Novosphingobium sp. strain Chol11, a model organism for bacterial steroid catabolism. Genome Announc 6:e01372-17.
Mendelski MN, Dölling R, Feller FM, Hoffmann D, Ramos Fangmeier L, Ludwig KC, Yücel O, Mährlein A, Paul RJ, Philipp B. 2019. Steroids originating from bacterial bile acid degradation affect Caenorhabditis elegans and indicate potential risks for the fauna of manured soils. Sci Rep 9:11120.
Knol J, Bodewits K, Hessels GI, Dijkhuizen L, van der Geize R. 2008. 3-Keto-5α-steroid Δ1-dehydrogenase from Rhodococcus erythropolis SQ1 and its orthologue in Mycobacterium tuberculosis H37Rv are highly specific enzymes that function in cholesterol catabolism. Biochem J 410:339–346.
Yang M, Lu R, Guja KE, Wipperman MF, St Clair JR, Bonds AC, Garcia-Diaz M, Sampson NS. 2015. Unraveling cholesterol catabolism in Mycobacterium tuberculosis: ChsE4-ChsE5 α2β2 acyl-CoA dehydrogenase initiates β-oxidation of 3-oxo-cholest-4-en-26-oyl CoA. ACS Infect Dis 1:110–125.
Yang M, Guja KE, Thomas ST, Garcia-Diaz M, Sampson NS. 2014. A distinct MaoC-like enoyl-CoA hydratase architecture mediates cholesterol catabolism in Mycobacterium tuberculosis. ACS Chem Biol 9:2632–2645.
Yuan T, Werman JM, Yin X, Yang M, Garcia-Diaz M, Sampson NS. 2021. Enzymatic β-oxidation of the cholesterol side chain in Mycobacterium tuberculosis bifurcates stereospecifically at hydration of 3-oxo-cholest-4,22-dien-24-oyl-CoA. ACS Infect Dis 7:1739–1751.
Petrusma M, Hessels G, Dijkhuizen L, van der Geize R. 2011. Multiplicity of 3-ketosteroid-9α-hydroxylase enzymes in Rhodococcus rhodochrous DSM43269 for specific degradation of different classes of steroids. J Bacteriol 193:3931–3940.
Yoshiyama-Yanagawa T, Enya S, Shimada-Niwa Y, Yaguchi S, Haramoto Y, Matsuya T, Shiomi K, Sasakura Y, Takahashi S, Asashima M, Kataoka H, Niwa R. 2011. The conserved Rieske oxygenase DAF-36/Neverland is a novel cholesterol-metabolizing enzyme. J Biol Chem 286:25756–25762.
Zhu Z, Li C, Cheng X, Chen Y, Zhu M, Liu X, Mao S, Qin HM, Lu F. 2019. Soluble expression, purification and biochemical characterization of a C-7 cholesterol dehydrogenase from Drosophila melanogaster. Steroids 152:108495.
Horinouchi M, Hayashi T, Koshino H, Kudo T. 2006. ORF18-disrupted mutant of Comamonas testosteroni TA441 accumulates significant amounts of 9,17-dioxo-1,2,3,4,10,19-hexanorandrostan-5-oic acid and its derivatives after incubation with steroids. J Steroid Biochem Mol Biol 101:78–84.
Horinouchi M, Hayashi T, Koshino H, Malon M, Hirota H, Kudo T. 2014. Identification of 9α-hydroxy-17-oxo-1,2,3,4,10,19-hexanorandrostan-5-oic acid in steroid degradation by Comamonas testosteroni TA441 and its conversion to the corresponding 6-en-5-oyl coenzyme A (CoA) involving open reading frame 28 (ORF28)- and ORF30-e. J Bacteriol 196:3598–3608.
Horinouchi M, Hayashi T, Koshino H, Malon M, Hirota H, Kudo T. 2014. Identification of 9α-hydroxy-17-oxo-1,2,3,4,10,19-hexanorandrost-6-en-5-oic acid and β-oxidation products of the C-17 side chain in cholic acid degradation by Comamonas testosteroni TA441. J Steroid Biochem Mol Biol 143:306–322.
Fujii K, Satomi M, Morita N, Motomura T, Tanaka T, Kikuchi S. 2003. Novosphingobium tardaugens sp. nov., an oestradiol-degrading bacterium isolated from activated sludge of a sewage treatment plant in Tokyo. Int J Syst Evol Microbiol 53:47–52.
Holert J, Cardenas E, Bergstrand LH, Zaikova E, Hahn AS, Hallam SJ, Mohn WW. 2018. Metagenomes reveal global distribution of bacterial steroid catabolism in natural, engineered, and host environments. mBio 9:e02345-17.
Yoshimoto T, Higashi H, Kanatani A, Lin XS, Nagai H, Oyama H, Kurazono K, Tsuru D. 1991. Cloning and sequencing of the 7α-hydroxysteroid dehydrogenase gene from Escherichia coli HB101 and characterization of the expressed enzyme. J Bacteriol 173:2173–2179.
Zipper C, Nickel K, Angst W, Kohler H-PE. 1996. Complete microbial degradation of both enantiomers of the chiral herbicide Mecoprop [(RS)-2-(4-chloro-2-methylphenoxy)propionic acid] in an enantioselective manner by Sphingomonas herbicidovorans sp. nov. Appl Environ Microbiol 62:4318–4322.
Fredrickson JK, Brockman FJ, Workman DJ, Li SW, Stevens T. 1991. Isolation and characterization of a subsurface bacterium capable of growth on toluene, naphthalene, and other aromatic compounds. Appl Environ Microbiol 57:796–803.
Blattner FR, Plunkett G, Bloch CA, Perna NT, Burland V, Riley M, Collado-Vides J, Glasner JD, Rode CK, Mayhew GF, Gregor J, Davis NW, Kirkpatrick HA, Goeden MA, Rose DJ, Mau B, Shao Y. 1997. The complete genome sequence of Escherichia coli K-12. Science 277:1453–1462.
Kovach ME, Elzer PH, Steven Hill D, Robertson GT, Farris MA, Roop RM, Peterson KM, Elzer AP, Steven Hill D, Robertson GT, Farris MA, Martin R, Ii R, Peterson KM, Peterson KM. 1995. Four new derivatives of the broad-host-range cloning vector pBBR1MCS, carrying different antibiotic-resistance cassettes. Gene 166:175–176.
Wöhlbrand L, Rabus R, Blasius B, Feenders C. 2017. Influence of nanoLC column and gradient length as well as MS/MS frequency and sample complexity on shotgun protein identification of marine bacteria. J Mol Microbiol Biotechnol 27:199–212.
Neuhoff V, Arold N, Taube D, Ehrhardt W. 1988. Improved staining of proteins in polyacrylamide gels including isoelectric focusing gels with clear background at nanogram sensitivity using Coomassie brilliant blue G‐250 and R‐250. Electrophoresis 9:255–262.
Mitchell AL, Attwood TK, Babbitt PC, Blum M, Bork P, Bridge A, Brown SD, Chang HY, El-Gebali S, Fraser MI, Gough J, Haft DR, Huang H, Letunic I, Lopez R, Luciani A, Madeira F, Marchler-Bauer A, Mi H, Natale DA, Necci M, Nuka G, Orengo C, Pandurangan AP, Paysan-Lafosse T, Pesseat S, Potter SC, Qureshi MA, Rawlings ND, Redaschi N, Richardson LJ, Rivoire C, Salazar GA, Sangrador-Vegas A, Sigrist CJA, Sillitoe I, Sutton GG, Thanki N, Thomas PD, Tosatto SCE, Yong SY, Finn RD. 2019. InterPro in 2019: improving coverage, classification and access to protein sequence annotations. Nucleic Acids Res 47:D351–D360.
Lu S, Wang J, Chitsaz F, Derbyshire MK, Geer RC, Gonzales NR, Gwadz M, Hurwitz DI, Marchler GH, Song JS, Thanki N, Yamashita RA, Yang M, Zhang D, Zheng C, Lanczycki CJ, Marchler-Bauer A. 2020. CDD/SPARCLE: the conserved domain database in 2020. Nucleic Acids Res 48:D265–D268.
Huerta-Cepas J, Szklarczyk D, Heller D, Hernández-Plaza A, Forslund SK, Cook H, Mende DR, Letunic I, Rattei T, Jensen LJ, Von Mering C, Bork P. 2019. EggNOG 5.0: a hierarchical, functionally and phylogenetically annotated orthology resource based on 5090 organisms and 2502 viruses. Nucleic Acids Res 47:D309–D314.
Wells JE, Hylemon PB. 2000. Identification and characterization of a bile acid 7α-dehydroxylation operon in Clostridium sp. strain TO-931, a highly active 7α-dehydroxylating strain isolated from human feces. Appl Environ Microbiol 66:1107–1113.
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.