Engineered 3-Ketosteroid 9α-Hydroxylases in Mycobacterium neoaurum: an Efficient Platform for Production of Steroid Drugs

Steroidal drugs are widely used for anti-inflammation, anti-tumor action, endocrine regulation, and fertility management, among other uses. The two main starting materials for the industrial synthesis of steroid drugs are phytosterol and diosgenin. The phytosterol processing is carried out by microbial transformation, which is thought to be superior to the diosgenin processing by chemical conversions, given its simple and environmentally friendly process. However, diosgenin has long been used as the primary starting material instead of phytosterol. This is in response to challenges in developing efficient microbial strains for industrial phytosterol transformation, which stem from complex metabolic processes that feature many currently unclear details. In this study, we identified two oxygenase homologues of 3-ketosteroid-9α-hydroxylase, KshA1_N and KshA2_N, in M. neoaurum and demonstrated their crucial role in determining the yield and variety of products from phytosterol transformation. This work has practical value in developing industrial strains for phytosterol biotransformation.

ABSTRACT

3-Ketosteroid 9α-hydroxylase (Ksh) consists of a terminal oxygenase (KshA) and a ferredoxin reductase and is indispensable in the cleavage of steroid nucleus in microorganisms. The activities of Kshs are crucial factors in determining the yield and distribution of products in the biotechnological transformation of sterols in industrial applications. In this study, two KshA homologues, KshA1_N and KshA2_N, were characterized and further engineered in a sterol-digesting strain, Mycobacterium neoaurum ATCC 25795, to construct androstenone-producing strains. kshA1_N is a member of the gene cluster encoding sterol catabolism enzymes, and its transcription exhibited a 4.7-fold increase under cholesterol induction. Furthermore, null mutation of kshA1_N led to the stable accumulation of androst-4-ene-3,17-dione (AD) and androst-1,4-diene-3,17-dione (ADD). We determined kshA2_N to be a redundant form of kshA1_N. Through a combined modification of kshA1_N, kshA2_N, and other key genes involved in the metabolism of sterols, we constructed a high-yield ADD-producing strain that could produce 9.36 g liter⁻¹ ADD from the transformation of 20 g liter⁻¹ phytosterols in 168 h. Moreover, we improved a previously established 9α-hydroxy-AD-producing strain via the overexpression of a mutant KshA1_N that had enhanced Ksh activity. Genetic engineering allowed the new strain to produce 11.7 g liter⁻¹ 9α-hydroxy-4-androstene-3,17-dione (9-OHAD) from the transformation of 20.0 g liter⁻¹ phytosterol in 120 h.

IMPORTANCE Steroidal drugs are widely used for anti-inflammation, anti-tumor action, endocrine regulation, and fertility management, among other uses. The two main starting materials for the industrial synthesis of steroid drugs are phytosterol and diosgenin. The phytosterol processing is carried out by microbial transformation, which is thought to be superior to the diosgenin processing by chemical conversions, given its simple and environmentally friendly process. However, diosgenin has long been used as the primary starting material instead of phytosterol. This is in response to challenges in developing efficient microbial strains for industrial phytosterol transformation, which stem from complex metabolic processes that feature many currently unclear details. In this study, we identified two oxygenase homologues of 3-ketosteroid-9α-hydroxylase, KshA1_N and KshA2_N, in M. neoaurum and demonstrated their crucial role in determining the yield and variety of products from phytosterol transformation. This work has practical value in developing industrial strains for phytosterol biotransformation.

INTRODUCTION

Some microorganisms can naturally use sterols as carbon and energy sources. During the microbial metabolism of sterols, several metabolic intermediates can be readily used to synthesize useful steroid drugs, including androst-4-ene-3,17-dione (AD), androst-1,4-diene-3,17-dione (ADD), 9α-hydroxy AD (9α-OHAD), and 22-hydroxy-23,24-bisnorchol-4-en-3-one (4-HBC), among others. Thus, an important means for the commercial production of steroid drugs has been developed based on the biotechnological modification of these microbial metabolic processes, primarily in the mycobacteria (1, 2). To develop a strain that can produce one of these valuable metabolites in high yield, several enzymatic steps must be considered. The 9α-hydroxylation and Δ1-dehydrogenation of 3-ketosteroids, which are carried out by 3-ketosteroid 9α-hydroxylases (Kshs) and 3-ketosteroid Δ1-dehydrogenases (KstDs), respectively, are two key steps that must be precisely controlled because of their essential roles in the selective formation and further degradation of target metabolites (3, 4). According to the proposed pathway of sterol catabolism in these microorganisms (Fig. 1a), genetic inactivation of Kshs or KstDs will block the degradation of steroid nucleus and thereby lead to the accumulation of steroidal intermediates, with three possible nucleus structures: AD, ADD, or 9α-OHAD (Fig. 1b). Theoretically, full inactivation of Kshs can achieve the production of ADD while maintaining sufficient activation of other enzymes involved in sterol metabolism, especially KstDs. Likewise, complete inactivation of KstDs can result in the production of 9α-OHAD while maintaining sufficient activation of other enzymes, especially Kshs. The inactivation of both Kshs and KstDs can result in AD as the main product. Similarly, various C₂₂ metabolites also can be achieved by combined modification of Kshs, KstDs, and other key enzymes involved in side chain degradation (Fig. 1b). Therefore, Kshs and KstDs are important targets for modification by random mutagenesis or genetic engineering to develop microorganisms that can transform sterols to valuable steroidal intermediates.

FIG 1.

Role of Ksh in the sterol catabolism pathway under aerobic conditions. (a) Metabolism of the steroid nucleus. (b) Proposed catabolic pathway of sterols (e.g., β-sitosterol). C₁₉ metabolites are shown in the blue box, and C₂₂ metabolites are in the green box. The conversion from 22HOBNC-CoA to AD was designated the AD pathway (blue arrows), and the conversion from 22HOBNC-CoA to 4-HBC was designated the HBC pathway (green arrows). Abbreviations: 22OBNC-CoA, 3,22-dioxo-25,26-bisnorchol-4-ene-24-oyl CoA; 22HOBNC-CoA, 22-hydroxy-3-oxo-25,26-bisnorchol-4-en-24-oyl CoA; 4-BNC, 3-oxo-23,24-bisnorchol-4-en-22-oic acid; 4-BNC-CoA, 3-oxo-23,24-bisnorchol-4-en-22-oyl-coenzyme A thioester; AD, 4-androstene-3,17-dione; ADD, 1,4-androstadiene-3,17-dione; T, testosterone; DHT, boldenone; 9-OHAD, 9α-hydroxy-4-androstene-3,17-dione; 9-OHADD, 9α-hydroxy-1,4-androstadiene-3,17-dione; 4-HBC, 22-hydroxy-23,24-bisnorchol-ene-3-one; 1,4-HBC, 22-hydroxy-23,24-bisnorchol-1,4-dien-3-one; 9-OHHBC, 9,22-dihydroxy-23,24-bisnorchol-4-ene-3-one.

Ksh is a two-component oxygenase, including a Rieske oxygenase (KshA) and a ferredoxin reductase (KshB) that consumes O₂ and 3-ketosteroids as substrates and NAD as the electron donor (3, 5). The first Ksh system was identified and molecularly characterized in Rhodococcus erythropolis SQ1, which confirmed the previously speculated catabolic processing of steroid nucleus and revealed the physiological importance of the Ksh system (6). The second Ksh system was subsequently isolated and identified in this strain (7). The cholesterol metabolism in Mycobacterium tuberculosis has been extensively studied and is of substantial interest because cholesterol is a required nutrient for the survival and persistence of this pathogen in the host (8, 9). In light of the importance of the Ksh system in biotechnology, pathogenesis, and medicine development, the biochemical properties of Ksh subsequently have been characterized in detail using a reconstituted Ksh system from R. rhodochrous DSM 43269 and crystal structure analysis of KshA from M. tuberculosis H37Rv (5). The structure of KshA clearly showed an extended substrate-binding pocket and an active-site-accessing funnel that was consistent with the binding requirement of large steroid substrates (3). The physiological roles of Ksh in these two strains were subsequently investigated in detail (10, 11). In M. tuberculosis H37Rv, only one Ksh is encoded by Rv3526 (kshA) and Rv3571 (kshB). The deletion of either gene resulted in rapid death of the pathogen in the host, indicating that the Ksh system is an essential factor for the infectivity of M. tuberculosis H37Rv (10). In R. rhodochrous DSM 43269, five KshA homologues (KshA1 to KshA5) have been identified (12), and their enzymatic properties have been characterized. Each Ksh system displays unique steroid induction patterns and substrate ranges, indicating an interesting multiplicity of the Ksh system in certain microorganisms that employ dynamic and finely tuned steroid catabolism (11).

Although detailed information has been obtained about Ksh in some microorganisms, such as R. rhodochrous DSM 43269, attempts to modify the Ksh system toward engineered microorganisms that can produce desirable sterol metabolites in high purity and productivity have rarely been successful. Nevertheless, genetic modification of the Ksh system in some sterol-digesting microorganisms has shown remarkable biotechnological significance in the development of valuable steroid producers (13). For example, the five-KshA-null mutant of R. rhodochrous DSM 43269 could transform some sterols into ADD and 3-oxo-23,24-bisnorchol-1,4-dien-22-oic acid (1,4-BNC) at molar ratios of 1% to 7% and 50% to 80%, respectively, depending on the substrate that was used (12). Furthermore, a blocked kshA mutant of M. smegmatis mc²155 that was isolated by transposon mutagenesis was able to produce both AD and ADD from sitosterol (13). These examples demonstrate that the modification of the Ksh system can result in the production of valuable C₁₉ and C₂₂ metabolites from sterols, which is consistent with theoretical speculation. These studies, however, also highlighted the multiplicity and variability of Ksh systems in different microorganisms and illustrated the complexity of Ksh systems with respect to the overall catabolism of sterols. Consequently, a systematic metabolic engineering strategy is required to further develop desirable steroid-producing microorganisms by genetically modifying key reactions in the sterol catabolism, including the Ksh system.

Previously, we characterized and engineered genes encoding key enzymes that participate in sterol catabolism within a fast-growing mycobacterial strain, Mycobacterium neoaurum ATCC 25795 (designated Mn25795 in this study). The engineered genes included the product-determining KstD genes (kstD genes) and two productivity-related cholesterol oxidase genes (choM genes) (1, 4). These experiments successfully developed AD- and ADD-producing strains that exhibit high product purity and productivity (1, 14). In this study, we identified and characterized the Ksh system in Mn25795 to investigate its physiological role in determining the conversion of intermediates during the catabolism of sterols. We then modified the Ksh system using a combined metabolic engineering strategy involving kstD genes and hsd4A (encoding Hsd4A, a dual-function enzyme exhibiting NAD⁺-dependent 17β-hydroxysteroid dehydrogenase and β-hydroxy-acyl-coenzyme A [CoA] dehydrogenase activity) to develop 9α-OHAD- and ADD-producing strains.

RESULTS

The Ksh system of Mn25795.

KshA is the functional component of Ksh that performs the specific 9α-hydroxylation of 3-ketosteroids. Among characterized KshAs, the two highly conserved domains are the Rieske [2Fe-2S] domain and the nonheme Fe(II)-binding domain (3). Genome-wide sequencing and investigation of the Mn25795 genome revealed two putative KshA-encoding homologues that were designated kshA1_N and kshA2_N. kshA1_N encodes a protein (KshA1_N) with 395 amino acids and an estimated molecular mass of 45.2 kDa that shares high amino acid sequence identity with characterized KshAs from M. tuberculosis H37Rv (KshA_H37Rv, 84%) and M. smegmatis mc²155 (KshA_MS, 89%). kshA2_N encodes a protein (KshA2_N) with 380 amino acids and an estimated molecular mass of 43.0 kDa that displays high amino acid sequence identity with a putative KshA from M. abscessus (81%) but otherwise exhibits a low sequence identity (<65%) with characterized KshAs. KshA2_N and KshA1_N share only 62% sequence identity, but the Rieske type [2Fe-2S]_R domain (C-X-H-X_15–17-C-X₂-H) (15) and the mononuclear, nonheme Fe(II) binding motif (D/E-X3-D-X2-H-X4-H) (16), which are typical of KshA, are conserved in both homologues.

kshA1_N is located in the sterol catabolism gene cluster, and its promoter region has a palindrome sequence (TnnAACnnGTTnnA, where “n” represents any base), a typical recognition sequence of the highly conserved transcriptional repressor (KstR) of sterol catabolism in mycobacteria (17). The genetic organization of kshA1_N and its adjacent genes shares high conservation with its orthologs in M. vanbaalenii PYR-1, M. smegmatis mc²155, and M. tuberculosis H37Rv (Fig. 2a). Consequently, these properties indicated that kshA1_N plays a key role in sterol catabolism. Unlike kshA1_N, kshA2_N was not arranged in the recognized sterol catabolism gene cluster, and no identifiable binding site for KstR was identified in its promoter region, which indicates that kshA2_N does not function in a KstR-dependent regulatory mode (18). Nevertheless, the genetic organization of kshA2_N and its adjacent genes also shows high conservation in some mycobacteria. Intriguingly, the conserved adjacent genes of kshA2_N include two characterized genes, kstD2 and kstD3, which are redundant genes of kstD1 and have been confirmed to play important roles in sterol catabolism (4). Although the function of this gene cluster is unknown, it appears to be related to the degradation of steroids. These observations indicate that kshA2_N also is important in the catabolism of sterols.

FIG 2.

Molecular characterization of KshA1_N and KshA2_N. (a) Schematic description of the locations of kshA1_N and kshA2_N in the Mn25795 chromosome. Arrow directions indicate the transcriptional direction of genes. Orthologous genes are indicated by gray-shaded regions, kshA1_N and kshA2_N are shown in green, and kstD genes are shown in blue. Intergenic areas containing the conserved palindromic motif for KstR1 binding in Mn25795 are labeled with asterisks. The genes from M. tuberculosis H37Rv that we predicted to be essential for survival in the host are indicated by triangles. (b) ClustalW alignment of the partial sequence from the putative β-sheet region located at the entrance of the KshA active site from M. tuberculosis H37Rv and Mn25795 and R. rhodochrous DSM 43269. Highly conserved residues and the four mutation points in the chimeric KshA1_N are indicated by filled and open rhombuses, respectively. (c) Three-dimensional model structure of KshA_H37Rv and the loop regions of KshA_H37Rv, KshA1_N, KshA2_N, KshA1_DSM43269, and KshA5_DSM43269. Green indicates the putative β-sheet domain; red and blue indicate the first and second α-helix domains, respectively; the orange sphere shows the nonheme ferrous iron; and the orange and yellow cluster spheres indicate the Rieske [2Fe-2S] cluster. The loop region of KshA_H37Rv is indicated with a box. The loop regions from KshA1_N, KshA2_N, KshA1_DSM43269, and KshA5_DSM43269 were homologously modeled with KshA1_H37Rv as a template and are shown as enlarged stick representations.

On the basis of the structure of KshA_H37Rv and the structural comparison of substrate preference between KshA1 and KshA5 from R. rhodochrous DSM 43269 (Fig. 2b), a highly variable 19-amino-acid-residue loop region was confirmed to be the essential determinant of substrate preference, and it is located at the entrance of the active site of KshA (6, 19). Alignment indicated that the variable loop region of KshA1_N was highly consistent with that of KshA_H37Rv, except for a few amino acids, including Asp₂₁₅ and Ser₂₂₀ in KshA_H37Rv, which corresponded to Asn₂₁₃ and Thr₂₁₈ in KshA1_N, respectively. Nevertheless, the corresponding loop of KshA2_N displays substantial differences from KshA_H37Rv, which implies enzymatic differences between KshA1_N and KshA2_N. To further characterize their properties, the specific loop regions of KshA1_N (Glu₁₉₆ to Val₂₆₀) and KshA2_N (Glu₁₉₂ to Val₂₅₈) were homologically modeled using KshA_H37Rv as a template. We predicted the loop structure of KshA1_N to be significantly different from that of KshA2_N and others (Fig. 2c), which further suggests enzymatic differences between KshA1_N and KshA2_N.

KshB is the auxiliary component of Ksh that provides electrons to KshA. Only one copy of kshB (kshB_N) is in the genome, which contains an open reading frame (ORF) of 1,056 bp encoding 351 amino acids with an estimated molecular mass of 37.8 kDa. The protein is a typical class IA monooxygenase reductase harboring three highly conserved domains: a flavin-binding domain (RCYSL, from residues 65 to 69), an NAD binding domain (GSGITP, from residues 129 to 134), and a 2Fe-2S iron-sulfur cluster binding domain (CX₄CX₂CX₂₉C, from residues 298 to 336). BLAST analysis demonstrated that KshB_N shares high amino acid sequence identity with homologues from M. smegmatis mc²155 (KshB_MS, MSMEG_5925, 79%) and M. tuberculosis H37Rv (KshB_H37Rv, Rv3571, 72%). Like kshB_MS and kshB_H37Rv, kshB_N is located in the gene cluster encoding sterol catabolism-related genes, but it is located several kilobases away from kshA1_N. Interestingly, KshB_N, along with KshA1_N and KshA2_N, exhibited more than 99% sequence identity with the putative KshB and two putative KshAs from Mycobacterium sp. strain VKM Ac-1815D (18). Phylogenetic analysis of 16S rRNA genes indicated that Mycobacterium sp. strain VKM Ac-1815D is identical to Mn25795, with 100% sequence similarity.

Determination of the roles of kshA1_N and kshA2_N in sterol catabolism.

To distinguish the roles of kshA1_N and kshA2_N in sterol catabolism, we constructed in-frame deletion mutants of kshA1_N (Mut_kshA1N), kshA2_N (Mut_kshA2N), and both together (Mut_KshA1&2N), in addition to their functional complements (Com_kshA1N, Com_kshA2N, and Com_kshA1&2N, respectively), in Mn25795. The deletion of kshA1_N significantly inhibited the growth of Mn25795 in minimum medium (MM) with cholesterol as the sole carbon source (Fig. 3a). Moreover, the inhibitory effect was clearly observed in Mut_kshA1N grown in MM with two carbon sources (i.e., cholesterol and glycerol), demonstrating the critical role of kshA1_N in the catabolism of cholesterol. In contrast, kshA2_N appeared to have no obvious effect on cholesterol catabolism, as the single deletion of kshA2_N did not disturb the growth of Mn25795 using cholesterol or cholesterol-glycerol as carbon sources. To exclude the shielding effect from the high activity of KshA1_N, the role of kshA2_N was again analyzed after the deletion of kshA2_N and kshA1_N, which showed that the double deletion of kshA1_N and kshA2_N resulted in even worse growth on cholesterol. Therefore, kshA2_N may provide functional redundancy for kshA1_N.

FIG 3.

Determination of the role of kshA1_N and kshA2_N in Mn25795. (a) Growth curves of Mn25795 and its derivative strains using cholesterol (a) or cholesterol-glycerol (b) as carbon sources, respectively. Symbols: Mn25795, ■; Mut_kshA1N, △; Mut_kshA2N, ▽; Mut_kshA1&2N, ♢; Com_kshA1N, ▲; Com_kshA1&2N, ◆; Com-blank, •. The concentration of cholesterol was 2 g liter⁻¹. The concentration of glycerol was the same as that for MYC/01. (c) High-performance liquid chromatography (HPLC) analyses of the metabolites resultant from the metabolism of 2 g liter⁻¹ cholesterol by Mut_kshA1N in cholesterol-glycerol medium at 120 h. (d) Transcriptional changes in kshA1_N and kshA2_N in glycerol medium with the addition of 2.0 mM cholesterol and 1.5 mM AD, respectively. Data are shown as averages from triplicate experiments ± standard errors.

Two hypothetical mechanisms may lead to growth impairment caused by KshA1_N inactivation. One mechanism occurs by blocking the degradation of the cholesterol steroid nucleus, such that no efficient carbon source would be available to support growth. The other mechanism involves the accumulation of some cholesterol metabolites, such as AD and ADD, that are toxic to respiration, viability, and duplication in microorganisms (20). To assess whether these putative metabolic intermediates exhibited strong toxicity and inhibited the growth of Mut_kshA1N, we added glycerol as a supplemental carbon source to the MM containing cholesterol to evaluate growth differences. Growth of Mut_kshA1N in the cholesterol-glycerol medium resulted in considerable accumulation of metabolites, including AD and ADD, whereas growth of Mut_kshA1N only slightly changed when grown in glycerol medium and the cholesterol-glycerol medium (Fig. 3b and c). These results indicate that cholesterol metabolism intermediates had no significant inhibitory effects on the growth of Mut_kshA1N. Therefore, the significant inhibition of growth caused by the deletion of kshA1_N could mainly be attributed to the lack of an appropriate carbon source to support growth. Compared with Mut_kshA1N, the growth of Mut_kshA2N in cholesterol-glycerol medium showed no obvious accumulation of cholesterol metabolites. These results clearly demonstrated that kshA1_N is the main kshA gene responsible for cholesterol metabolism.

The potential role of kshA2_N in sterol catabolism was thus unclear, as kshA2_N is not a member of the recognized gene cluster encoding sterol catabolism enzymes, and these results did not reveal its definite function in sterol metabolism. Therefore, to resolve this uncertainty and to distinguish functional differences between kshA1_N and kshA2_N, we measured transcription levels of kshA1_N and kshA2_N by reverse transcription-quantitative PCR (RT-qPCR) with the addition of cholesterol and AD to culture medium and with glycerol as the control. Both cholesterol and AD significantly induced the transcription of kshA1_N and kshA2_N, suggesting that kshA2_N is also involved in sterol metabolism, similar to kshA1_N (Fig. 3d). Cholesterol addition induced an approximately 2.0-fold increase in ksh2_N transcription and an approximately 4.7-fold increase in kshA1_N transcription, confirming that kshA1_N was more important than kshA2_N in sterol catabolism. Moreover, both kshA1_N (1.7-fold) and kshA2_N (1.6-fold) were induced by AD, which is a common C₁₉ steroid intermediate in sterol catabolism. Furthermore, the induction effect of cholesterol was greater than that of AD for kshA1_N and also higher than that of AD for kshA2_N but to a lesser extent, indicating that AD is a poor inducer of kshA1_N and kshA2_N.

Enzymatic characterization of KshA1_N and KshA2_N in vitro.

To determine whether KshB_N could form the active Ksh enzyme complex with KshA1_N and KshA2_N, the Ksh enzymes were reconstituted from separately expressed and purified KshB_N and KshA_N (KshA1_N or KshA2_N). However, no Ksh activity was detected, though the purified KshB_N maintained reductase activity of 6.3 μmol min⁻¹ mg⁻¹ in a 2,6-dichlorophenolindophenol (DCPIP) assay. As suggested by Petrusma et al., the protein-protein interactions between KshA and KshB subunits are critically important for maintaining Ksh activity (5). Therefore, the copurification of KshA_N and KshB_N was conducted, and the active Ksh system was successfully reconstituted (Fig. 4a and b). To ensure the full activity of KshA_N in the reconstructed Ksh system, an additional amount of purified KshB_N was used to supplement the copurified KshA_N and KshB_N to achieve a KshA_N-to-KshB_N molar ratio of approximately 1:2. Under these conditions, we analyzed the enzymatic properties of the Ksh enzymes consisting of KshA1_N or KshA2_N copurified with KshB (Tables 1 and 2).

FIG 4.

SDS-PAGE analysis of E. coli expression and copurification of KshA_N and KshB_N. (a) KshA1_N and KshB_N. (b) KshA2_N and KshB_N. Lane 1, standard marker protein; 2, cell extracts from BL21-kshA_N; 3, cell extracts from BL21-kshB_N; 4, mixture of cell extracts of BL21-kshA_N and BL21-kshB_N in a KshA_N:kshB_N molar ratio of approximately 1:2; 5, the copurified fraction of KshA_N and KshB_N with 150 mM imidazole.

TABLE 1.

Steroid substrate profiles of KshA1_N and KshA2_N

Steroid substratea	KshA1N		KshA2N
	Enzyme activitya (nmol min−1 mg−1)	Relative activityb (%)	Enzyme activity (nmol min−1 mg−1)	Relative activity (%)
4-Androstene-3,17-dione (AD)	2.3 × 102 ± 21 × 102	100	1.2 × 102 ± 15 × 102	100
4-Androstene-17β-ol-3-one (testosterone)	3.1 × 102 ± 17 × 102	135	90 ± 10	75
1,4-Androstadiene-3,17-dione (ADD)	6.5 × 102 ± 31 × 102	282	1.5 × 102 ± 29 × 102	125
1,4-Androstadiene-17β-ol-3-one (boldenone)	5.8 × 102 ± 47 × 102	252	1.3 × 102 ± 27 × 102	108
1-(5α)-Androstene-17β-ol-3-one	1.8 × 102 ± 39 × 102	78	23 ± 7	19
1-(5α)-Androstene-3,17-dione	2.1 × 102 ± 33 × 102	91	27 ± 9	23
5α-Androstane-3,17-dione	52 ± 11	23	NDc	ND
5β-Androstane-3,17-dione	75 ± 12	33	ND	ND
5α-Androstane-17β-ol-3-one (androstanolone)	42 ± 15	18	ND	ND
5α-Androstane-3α-ol-17-one	ND	ND	ND	ND
5α-Androstane-3β-ol-17-one (epiandrosterone)	ND	ND	ND	ND
5-Androstene-3β-ol-17-one (dehydroepiandrosterone)	ND	ND	ND	ND
9α-Hydroxy-4-androstene-3,17-dione (9-OHAD)	ND	ND	ND	ND
4-Androstene-3,17-dione-19-ol (19-OHAD)	23 ± 16	10	ND	ND
1,3,5(10)-Estratrien-3-ol-17-one	ND	ND	ND	ND
4-Pregnene-3,20-dione (progesterone)	7.7 × 102 ± 49 × 102	335	1.8 × 102 ± 21 × 102	150
11β-Hydrocortisone	1.2 × 102 ± 40 × 102	52	38 ± 11	32
23,24-Bisnorchol-1,4-diene-22-oic acid (1,4-BNC)	9.8 × 102 ± 55 × 102	426	2.1 × 102 ± 38 × 102	175
4-Cholestene-3-oned (cholesterone)	45 ± 7	21	ND	ND
5-Cholestene-3β-old (cholesterol)	ND	ND	ND	ND

^aSpecific Ksh activity levels toward substrates were analyzed at 200 μM and represented as means ± standard errors from three independent experiments.

^bRelative activity signifies the percentage of the activity toward substrates relative to that of AD, which was set at 100%.

^cND, no detectable activity toward specific steroid.

^dSteroid concentration of 25 μM due to the low steroid solubility.

TABLE 2.

Relative initial activities of wild-type and mutated KshA1_N with a range of steroid substrates

Steroid substrate	Relative initial activity of Ksh (%)a
	A1N	A1V207T	A1P210E	A1D225R	A1N238D
4-Androstene-3,17-dione (AD)	100	100	100	100	100
4-Androstene-17β-ol-3-one (testosterone)	135 ± 10	161 ± 30	154 ± 20	133 ± 37	139 ± 24
1,4-Androstadiene-3,17-dione (ADD)	282 ± 40	289 ± 38	275 ± 47	178 ± 29	252 ± 50
1,4-Androstadiene-17β-ol-3-one (boldenone)	252 ± 31	258 ± 33	255 ± 51	174 ± 18	267 ± 29
1-(5α)-Androstene-3,17-dione	91 ± 19	112 ± 20	100 ± 22	63 ± 18	95 ± 27
5α-Androstane-3,17-dione	23 ± 4	32 ± 16	32 ± 6	NDc	26 ± 9
5β-Androstane-3,17-dione	33 ± 6	42 ± 18	41 ± 12	15 ± 5	52 ± 19
5α-Androstane-17β-ol-3-one (androstanolone)	18 ± 3	35 ± 11	13 ± 7	ND	25 ± 9
17α-Methyl-1-(5α)-androstane-17β-ol-3-one	78 ± 8	92 ± 21	82 ± 10	32 ± 13	73 ± 32
4-Androstene-3,17-dione-19-ol	10 ± 2	12 ± 4	16 ± 2	ND	ND
4-Pregnene-3,20-dione (progesterone)	335 ± 38	430 ± 55	323 ± 40	215 ± 61	320 ± 40
11β-Hydrocortisone	52 ± 9	82 ± 26	42 ± 11	22 ± 8	57 ± 19
23,24-Bisnorchol-4-ene-22-oic acid (4-BNC)	426 ± 52	499 ± 58	447 ± 65	260 ± 43	440 ± 52
4-Cholestene-3-oneb (cholestenone)	21 ± 2	56 ± 17	30 ± 9	ND	25 ± 12

^aRelative activity signifies the percentage of activity toward substrates at 200 μM relative to that of AD, which was set at 100%. Errors were calculated as standard errors of the means (from three experiments).

^bSteroid concentration of 25 μM due to the low solubility of cholesterone.

^cND, no detectable activity toward specific steroid.

Both KshA_N homologues showed a broad preference for steroid substrates with no substantial differences, although the specific activity of KshA1_N was higher than that of KshA2_N for suitable substrates (Table 1). Clearly, steroids with 4-ene-3-one or 1-ene-3-one structures were the proper substrates for both KshA_N homologues, whereas steroids with 3-alcohol structures or aromatic A rings were not. Moreover, 1,4-diene-3-one structures, such as those of ADD and boldenone, were the more suitable substrates for both KshA_N homologues than 4-ene-3-one substrates, such as AD and testosterone. These substrate preferences indicate that the 3-keto-4-ene-conjugated moiety was the required structure for proper substrates of both KshA_Ns. The thing of note, however, was that KshA1_N also showed a small amount of activity with some steroids with 3-keto structures and without double bonds on the A ring, including androstanolone and androstane-3,17-dione.

Interestingly, both KshA_N homologues displayed the highest activities toward 23,24-bisnorchol-4-ene-22-oic acid (4-BNC) among the tested substrates, followed by progesterone, another steroid with a short C-17 side chain, which were much higher than their activities toward the C₁₉ steroids. These results indicated that the optimal substrates of KshA_N should also contain C-17 side chains. Cholesterone, a sterol intermediate with a full C-17 side chain, was not a good substrate for both KshA_N homologues. Thus, only certain intermediates in the sterol catabolism pathway with a truncated (partially degraded) C-17 side chain, such as 1,4-HBC and its derivatives or precursors, were optimal substrates for KshA1_N and KshA2_N rather than C₁₉ steroids. This clearly demonstrates that the cleavage of the steroid nucleus should occur concurrently with the cleavage of the C-17 side chain rather than afterward, which is consistent with what has been proposed for KshAB in M. tuberculosis (21).

To determine substrate affinities, we performed kinetic studies to determine the affinity constants of KshA1_N and KshA2_N. The enzyme activities of KshA1_N and KshA2_N toward different concentrations of substrates are given in Tables S3 and S4 in the supplemental material. Enzyme activity of KshA1_N toward most of the tested substrates showed a K_m value below 10 μM, except for androstane structures, cholesterone, and 4-androstene-3,17-dione-19-ol, indicating a desirable enzyme-substrate binding relation. The substrate affinities for KshA2_N had a K_m value of 33 ± 8 μM for 1-(5α)-androstene-17β-ol-3-one, 25 ± 6 μM for 1-(5α)-androstene-3,17-dione, and 13 ± 3 μM for 11β-hydrocortisone; the remaining substrates that were tested had K_m values of <10 μM (Table 1).

Modification of KshA1_N.

Given the pivotal role of Ksh in the metabolism of sterols, its activity must be modified to develop engineered strains that produce steroidal metabolites with sufficient yield. In some cases, such as in the development of 9-OHAD-producing strains, sufficient activity of Ksh is necessary to reduce the generation of side products (e.g., AD). Therefore, this study further investigated the modification of KshA1_N.

Previous studies have demonstrated that KshA homologues have similar tertiary structures with a variable β-sheet, including 58 amino acids in the substrate-binding pocket. In addition, a highly variable loop region of the β-sheet at the entrance of the active site contributes to substrate preference (Fig. 2b and c) (19, 22). Petrusma et al. demonstrated that the β-sheet could be exchanged between two KshA homologues, and the resulting chimeric enzymes exhibited substrate preferences similar to those of the donor enzymes (19). Hence, the β-sheet and loop regions of KshA1 and KshA2 were exchanged in this study to evaluate whether the enzymatic properties could be modified, which resulted in the chimeric proteins KshA1_A2β, KshA2_A1β, KshA1_A2loop, and KshA2_A1loop. Unfortunately, the four chimeras exhibited no enzymatic activity, indicating that the β-sheet and loop structures of KshA1 and KshA2 are not exchangeable.

Consequently, we attempted a strategy of point mutation of the β-sheet to modify the activity of KshA. On the basis of previous reports and alignment of the KshA_N β-sheets with well-characterized KshA counterparts from M. tuberculosis H37Rv and R. rhodochrous DSM 43269 (Fig. 2b), we selected four putatively crucial but nonconserved amino acids in the β-sheet of KshA1_N (two in the loop region) as the mutation sites (Fig. 2b): V207, P210, D225, and N238. According to the mutation of Q209T in the KshA1_A5 chimera that was reported previously (19), V207 was replaced by T207, resulting in a V207T mutant. We conducted the other three point mutations (P210E, N238D, and D225R) based on the differences in amino acids between Mycobacterium and Rhodococcus in the alignment analysis (Fig. 2b). The point mutations P210E (activity toward AD, 2.1 × 10² ± 18 × 10² nmol min⁻¹ mg⁻¹), N238D (activity toward AD, 1.9 × 10² ± 25 × 10² nmol min⁻¹ mg⁻¹), and D225R (activity toward AD, 1.6 × 10² ± 15 × 10² nmol min⁻¹ mg⁻¹) displayed no obvious improvement in activity relative to wild-type KshA1_N (Table 2). The KshA1_V207T construct showed elevated Ksh activity (AD, 2.4 × 10² ± 23 × 10² nmol min⁻¹ mg⁻¹) relative to the wild-type enzyme, however, and this was particularly evident with increases of 17.1% and 28.4% toward 4-BNC and progesterone, respectively (Table 2). As described earlier, sterol intermediates with a truncated side chain are the proper physiological substrates of Ksh. Thus, the significantly enhanced activity of KshA1_V207T toward 4-BNC indicated an apparently improved activity of KshA1_V207T relative to KshA1_N. Therefore, we employed KshA1_V207T in further modification of Mn25795.

Development of an ADD-producing strain.

ADD is a valuable C₁₉ metabolite that can be used as a precursor for the industrial synthesis of 19-nor-steroids, estrogens, and some adrenocortical hormones. As the abrogation of Ksh activities can theoretically block the continuous degradation of ADD and lead to ADD accumulation during sterol catabolism, we attempted to develop an ADD-producing strain of Mn25795 via inactivation of KshA_Ns. As indicated, the kshA1_N-kshA2_N double-deletion strain Mut_kshA1NA2N (termed strain I) could transform phytosterols with ADD constituting the major metabolic product. We added 2 g liter⁻¹ phytosterols to fermentation media with strain I, followed by incubation in a shake flask for 168 h. The resultant conversion rate of phytosterols was nearly 95%, and the ADD yield reached 0.46 g liter⁻¹, which represents a 33.20% molar yield. The strain also produced other by-products in addition to ADD, including a 20.6% yield (0.29 g liter⁻¹) of AD, a 15.1% (0.21 g liter⁻¹) yield of DHT, and a 4.4% yield (0.06 g liter⁻¹) of T. Thus, the construct Mut_kshA1NA2N was not a great ADD producer and could be further improved by reducing the production of these by-products. Considering the role of KstD in transforming AD to ADD (Fig. 1b), the undesired production of AD was attributed to insufficient activity of KstD. Previously, three KstD homologues, and their physiological roles, were characterized in Mn25795, with KstD₁ representing the primarily active homologue (4).

Following these observations, kstD₁ was overexpressed in Mut_kshA1NA2N, resulting in the strain Mut_kshA1NA2N-p261kstD₁, which was termed strain II. After incubation for 168 h with 2 g liter⁻¹ phytosterol additions, strain II gave a markedly enhanced yield of ADD (56.3% molar yield, 0.78 g liter⁻¹) with a significantly decreased yield of AD (5.4% molar yield, 0.08 g liter⁻¹) (Table 3) compared with those of strain I. Moreover, no detectable DHT and T were produced. The ADD yield ratio among the metabolic products of strain II reached 91.2%, representing more than a 2.0-fold increase (45.3%) over strain I (Fig. 5b). These results indicate the necessity of combined modification of kshA and kstD genes in the development of ADD production strains. To evaluate the potential for industrial application of strain II, we employed a resting cell biotransformation system in a shake flask with 20 g liter⁻¹ phytosterols. Surprisingly, we obtained only a 38.2% molar yield (5.3 g liter⁻¹) of ADD, along with abundant 1,4-HBC (31.0% yield, 4.3 g liter⁻¹) (Fig. 5c). Intriguingly, these results differed from previous results, yielding the new by-product 1,4-HBC and indicating that the sterol transformation conditions have an important effect on the resultant by-products.

TABLE 3.

Profiles of phytosterol transformation by ADD-producing strains for 168 h

Strain and concn of substratea (g liter−1)	Major product	Concn (g liter−1)	Molar yieldb (%)	Yield ratio of ADDc (%)
I
2	ADD	0.46 ± 0.03	33.2 ± 1.8	45.3 ± 0.5
	AD	0.29 ± 0.04	20.6 ± 2.7
	DHT	0.21 ± 0.01	15.1 ± 1.0
	T	0.06 ± 0.01	4.4 ± 1.1
II
2	ADD	0.78 ± 0.05	56.3 ± 3.2	91.2 ± 0.3
	AD	0.08 ± 0.02	5.4 ± 0.5
20 (resting cells)	ADD	5.31 ± 0.25	38.2 ± 1.8	55.2 ± 0.6
	1,4-HBC	4.29 ± 0.20	31.0 ± 1.5
20 (growing cells)	ADD	7.78 ± 0.15	55.4 ± 1.2	89.1 ± 0.2
	AD	0.60 ± 0.03	4.3 ± 0.2
	1,4-HBC	0.41 ± 0.01	2.5 ± 0.1
III
20 (resting cells)	ADD	7.51 ± 0.31	53.8 ± 2.1	72.6 ± 0.4
	1,4-HBC	2.82 ± 0.29	20.3 ± 1.7
20 (growing cells)	ADD	9.36 ± 0.25	66.7 ± 1.6	92.6 ± 0.3
	AD	0.74 ± 0.05	5.3 ± 0.3

^aThe conversion of 2 g liter⁻¹ phytosterols was performed in shake flasks using growing cells, and the conversion of 20 g liter⁻¹ phytosterols was performed in a 7-liter bioreactor using resting cells or growing cells. Data are shown as the averages from triplicate experiments.

^bAmounts of major metabolites and their molar yields were determined from three independent transformations and expressed as the actual range from the least to the topmost.

^cThe yield ratio of ADD was obtained by calculating the molar yield of ADD among all the major products.

FIG 5.

ADD synthesis metabolic pathway and metabolite analysis of constructed ADD-producing strains. (a) The ADD synthesis metabolic pathway from β-sitosterol in Mn25795. Red arrows indicate strengthened steps, and the crossed arrows represent blocked steps. The structure of the desired product ADD is also indicated in red. (b) HPLC comparison of the products derived from transformation of 2 g liter⁻¹ phytosterol by strains I and II in the growing cell system at the shake flask level. (c) HPLC comparison of the products derived from transformation of 20 g liter⁻¹ phytosterols by strains II and III in the resting cell system at the shake flask level.

On the basis of our previous research (23), the accumulation of 1,4-HBC during the sterol catabolism can be attributed to the insufficient activity of a key enzyme, Hsd4A, that is involved in the degradation of the sterol side chain (Fig. 5a). Hsd4A is a dual-function enzyme in sterol catabolism, exhibiting NAD⁺-dependent 17β-hydroxysteroid dehydrogenase and β-hydroxy-acyl-CoA dehydrogenase activity. The latter activity was considered to be responsible for the conversion of 22-hydroxy-3-oxo-25,26-bisnorchol-4-en-24-oyl CoA (22HOBNC-CoA) to 3,22-dioxo-25,26-bisnorchol-4-ene-24-oyl CoA (22OBNC-CoA), and the deletion of Hsd4A thus could result in the production of 1,4-HBC or its derivatives (23). Therefore, we further modified strain II by increasing the expression of Hsd4A to reduce the yield of 4-HBC, thereby resulting in Mut_kshA1NA2N-p261kstD₁-p60hsd4A, termed strain III. After incubation with 20 g liter⁻¹ phytosterols in the resting cell system for 168 h, strain III gave a 53.8% molar yield (7.5 g liter⁻¹) of ADD and maintained a 20.3% molar yield (2.8 g liter⁻¹) of 1,4-HBC (Table 3). Compared with that of strain II, the ADD yield ratio of strain III increased from approximately 55.2% to 72.6%, indicating the effectiveness of hsd4A overexpression toward the production of ADD. Consequently, we further evaluated the productive properties of strains II and III in the growing cell system. The results indicated that strain II can ferment 20 g liter⁻¹ phytosterols in 168 h to yield 55.4% ADD (7.78 g liter⁻¹), 4.3% AD (0.60 g liter⁻¹), and 2.5% 1,4-HBC (0.41 g liter⁻¹). In contrast to the results from the resting cell system, the 1,4-HBC yield was significantly reduced from a 31.0% molar yield to 2.5%, although a 4.3% molar yield of AD still occurred, and the yield ratio of ADD was greatly enhanced, from 55.2% to 89.1%.

These results highlight the importance of transformation growth conditions in determining the composition of metabolic products from the conversion of phytosterols by ADD-producing strains. We also converted 20 g liter⁻¹ phytosterols using strain III in the growing cell system and obtained a 66.7% molar yield of ADD (9.36 g liter⁻¹), the yield ratio of which reached 92.5%, along with a 5.3% molar yield of AD (0.74 g liter⁻¹) and no measurable 1,4-HBC yield (Table 3). These results demonstrate that strain III is a good ADD-producing strain and that the growing cell system is more suitable than the resting cell transformation system for the production of ADD from phytosterol conversion.

Development of a 9-OHAD-producing strain.

9-OHAD is a valuable C₁₉ metabolite and is a good precursor for the industrial production of adrenocortical hormones. The construction of a 9-OHAD-producing strain also involves the modification of Ksh activity. In our previous work (4), M. neoaurum NwIB-yV was constructed via the overexpression of KshA1_N in strain M. neoaurum Mut_MN-kstD1&2&3, which is designated strain IV in this study. After 120 h of incubation with 2 g liter⁻¹ of phytosterols in a shake flask, strain IV exhibited a conversion rate of >95% over 120 h. The conversion rate of this 9-OHAD-producing strain was higher than that of ADD-producing strains, for which we obtained a 74.8% molar yield (1.10 g liter⁻¹) of 9-OHAD, along with a 3.4% yield (0.05 g liter⁻¹) of AD (Table 4). The 9-OHAD yield ratio of strain IV was more than 95%, indicating that strain IV was a good 9-OHAD producer. Subsequently, we investigated its industrial application potential in the resting cell system with 20 g liter⁻¹ phytosterols. However, the 9-OHAD yield ratio of strain IV decreased significantly from 95.7% to 89.0%, with a 68.3% molar yield (10.04 g liter⁻¹) of 9-OHAD and an 8.4% molar yield (1.24 g liter⁻¹) of AD. These results indicate that in the presence of high concentrations of phytosterols, the activity of Ksh is insufficient and should be further enhanced to improve the yield ratio of 9-OHAD. The transcript levels of kshA1_N in strain IV were 1.6-fold and 2.0-fold higher than that of strain Mut_MN-kstD1&2&3 after 72 h and 120 h of phytosterol transformation, respectively (Fig. 6b).

TABLE 4.

Profiles of phytosterol transformation by 9-OHAD-producing strains for 120 h

Strain and concn of substratea (g liter−1)	Major product	Concnb (g liter−1)	Molar yieldc (%)	Yield ratio of 9-OHAD (%)
IV
2	9-OHAD	1.10 ± 0.05	74.8 ± 3.1	95.7 ± 0.2
	AD	0.05 ± 0.02	3.4 ± 1.0
20	9-OHAD	10.04 ± 0.22	68.3 ± 1.5	89.0 ± 0.5
	AD	1.24 ± 0.11	8.4 ± 0.5
V
2	9-OHAD	1.32 ± 0.03	90.2 ± 1.9	97.6 ± 0.3
	AD	0.03 ± 0.01	2.2 ± 0.3
20	9-OHAD	11.71 ± 0.15	79.8 ± 1.0	94.2 ± 0.4
	AD	0.71 ± 0.10	4.9 ± 0.6

^aThe conversion of 2 g liter⁻¹ phytosterols was performed in shake flasks using growing cells, and the conversion of 20 g liter⁻¹ phytosterols was performed in a 7-liter bioreactor using resting cells. Data are shown as the averages from triplicate experiments.

^bAmounts of major metabolites and their molar yields were determined from three independent transformations and are expressed as the actual range from the least to the topmost.

^cThe yield ratio of 9-OHAD was obtained by calculating the molar yield of 9-OHAD among all the products.

FIG 6.

Comparative analyses of 9-OHAD-producing strains. (a) The role of Ksh in the production of 9-OHAD. (b) Transcription profiles of kshA1_N in strains IV and V at 72 h (gray bars) and 120 h (dark bars) during the metabolism of 2.0 mM cholesterol. The fold change values indicate mRNA levels in strains IV and V relative to that of Mut_MN-kstD1&2&3. The data represent averages from triplicate experiments, and the error bars indicate ± standard deviations. (c) HPLC comparison of metabolites from the transformation of 2 g liter⁻¹ phytosterols by strains IV and V in the growing cell system at the shake flask level. (d) HPLC comparison of the metabolites from the transformation of 20 g liter⁻¹ phytosterols by strains IV and V in the resting cell system at the shake flask level.

These observations suggest that KshA1_N already is well overexpressed in strain IV, and increasing the expression level of KshA1 further likely will not improve the conversion of AD to 9-OHAD. Therefore, to enhance activity compared with that of KshA1, we overexpressed the mutant KshA1_V207T in Mut_MN-kstD1&2&3 to construct an improved 9-OHAD-producing strain relative to strain IV, thereby resulting in Mut_MN-kstD1&2&3-p261-kshA1_V207T, termed strain V. As expected, strain V showed improved performance in the resting cell system, converting 20 g liter⁻¹ of phytosterols to 11.7 g liter⁻¹ 9-OHAD within 120 h while also producing only 0.71 g liter⁻¹ AD. Compared with strain IV, the 9-OHAD yield ratio of strain V was enhanced from 89.0% to 94.2%. These results clearly show that strain V is a more promising 9-OHAD producer for industrial application than strain IV based on its higher 9-OHAD yield ratio and decreased by-product formation.

DISCUSSION

In the industrial process of phytosterol transformation to produce valuable steroid intermediates, unavoidable generation of by-products often leads to high purification costs and a low recovery rate of the target product, and therefore it is one of the main barriers limiting the successful development of engineered strains for industrial application. Generally, side products arise from specific rate-limiting reactions that often are concurrent and promiscuous. The 3-ketosteroid-Δ1-dehydrogenation and 3-ketosteroid-9α-hydroxylation that are performed by KstDs and Kshs, respectively, are two key reactions that directly determine the yield and yield ratio of products during sterol catabolism. The two reactions proceed concurrently and can cross-react, which involves many intermediates in the catabolic process of sterols.

The complex multiplicity of kshA is a common feature in rhodococci and mycobacteria, and it supports the utilization of a variety of environmental steroids as nutrition sources. For example, there are five kshA genes in R. rhodochrous DSM 43269 and Mycobacterium sp. strain VKM Ac-1817D and seven kshA genes in R. equi (24). In contrast to the multiplicity in these strains, the Ksh system in Mn25795 exhibited moderate redundancy with two KshA homologues: KshA1_N and KshA2_N. This suggests that the strain is more easily modified for industrial use by mutation screening or genetic engineering. In line with previously reported KstDs and KshAs (4, 11), KshA1_N and KshA2_N clearly exhibited wide and differentiated substrate preferences in this study, which suggests promiscuous reaction orders are carried out by Kshs with metabolic intermediates of phytosterols. Among the tested intermediates, both KstDs and KshAs showed greater activity toward the metabolic intermediates with truncated side chains, most notably BNC-CoA, 1,4-BNC-CoA, or 9α-OH-BNC-CoA (4), instead of the other metabolites, including C₁₉ metabolites AD, ADD, and 9-OHAD. For strains engineered to excessively accumulate certain C₁₉ metabolites, some precursors of the expected metabolite, which is not the optimal substrate for KstDs and KshAs, will be inevitably and excessively produced because of the concurrent and promiscuous reactions in the catabolic process of sterols. As most of the precursors are not the optimal substrates of KstDs and KshAs, these excess poor substrates could easily result in the generation of by-products resulting from the competitiveness effect, when KstD and KshA expression does not increase accordingly. This process is not a rare event, as some side products often are generated in industrial sterol biotransformation. For example, AD often is generated as a main by-product in the industrial production of 9-OHAD or ADD (4, 14). The activities of KstDs and Kshs are therefore of substantial interest, and their expression must be accordingly modified in the development of engineered strains for industrial application.

In view of the limiting conditions for cell growth in the shake flasks, such as lowered dissolved oxygen and pH fluctuation, the transformation capacity of growing mycobacterial cells at the shake flask level is generally limited far below their practical transformation capacity in bioreactors. In this study, we employed a resting cell transformation system, which has been developed to rapidly assess the application potential of newly constructed strains in previous studies (25). Interestingly, the resting cell transformation system showed another use as an assessment tool to identify possible rate-limiting steps among the metabolic processes of sterols. The constructed ADD-producing strain II showed a desirable transformation capacity for 20 g liter⁻¹ phytosterols over 168 h when grown in the resting cell transformation system but produced an unexpected by-product, 1,4-HBC. This result is significantly different from the metabolite profile obtained in the growing cell transformation system with 2 g liter⁻¹ phytosterols at the same shake flask level (Table 3). The yield of 1,4-HBC indicated that a rate-limiting step must exist in the degradation of sterol side chains. On the basis of our previous research (23), the rate-limiting step could be attributed to the conversion of 22HOBNC-CoA to 22OBNC-CoA in the degradation of the sterol side chain that is catalyzed by Hsd4A (Fig. 5a). Augmented expression of Hsd4A in strain II greatly decreased the molar yield of 1,4-HBC in the resting cell transformation system and completely inhibited its yield in the growing cell transformation system. These results confirmed that the conversion of 22HOBNC-CoA to 22OBNC-CoA is the rate-limiting step for the production of ADD in strain II. Consequently, hsd4A should be considered for modification in combination with kstD and kshA, which will together be useful in the development of ADD-producing strains via genetic engineering. Obviously, the usefulness of the resting cell transformation system to evaluate possible rate-limiting steps may be attributed to the magnification of effects resulting from specific rate-limiting steps in the resting cell system compared with the growing cell system.

In the phytosterol transformation of strain IV in the resting cell transformation system, the yield ratio of 9-OHAD was significantly decreased because of the increase of AD as a by-product compared with that in the growing cell transformation system, indicating that the activity of Ksh was still insufficient for the transformation of sterols when provided in high concentrations. This was particularly intriguing because the transcript levels of kshA1_N in strain IV were 1.6- to 2.0-fold greater than its original levels for the strain without KshA_N overexpression. To address this problem, we constructed a mutant, kshA1_V207T, with significantly elevated Ksh activity. The mutant KshA1_V207T exhibited improved activity in the conversion of phytosterols to 9-OHAD compared with KshA1. Consequently, this development provides an alternative means to improve the industrial potential of 9-OHAD-producing strains.

A summary of the strains that were constructed in this study and the effects of various genotypes on sterol catabolism are shown in the Fig. 7. Overall, this study profiled the properties of 3-ketosteroid 9α-hydroxylases in Mn25795 in detail and investigated specific means to develop ADD-producing and 9-OHAD-producing strains by investigating the production of side products and accordingly modifying kshA_N and other key genes that are associated with the yield of by-products. In combination with our previous studies on the characterization and engineering of kstD genes in Mn25795 (14, 23), our present analysis of kshA1_N and kshA2_N genes provides a relatively complete engineering solution that can be used to develop optimal ADD-producing and 9-OHAD-producing strains for desired industrial applications.

FIG 7.

Metabolic pathways of ADD and 9-OHAD synthesis. (a) Metabolic pathway to produce ADD; (b) metabolic pathway to produce 9-OHAD; (c) description of the strains that were established in this study. ADD and 9-OHAD are denoted in red. The key steps catalyzed by Kshs, KstDs, and Hsd4A were labeled by red, purple, and blue arrows, respectively. The symbol X represents the block of the metabolic pathway, and the thick red arrows, purple arrows, and the blue arrow represent the overexpression of Ksh, KstDs, and Hsd4A in the metabolic pathway. The successive thick arrows represent certain unspecified enzymes.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions.

All strains, plasmids, and primers used in this study are listed in Table 5 and Table S1 in the supplemental material. Escherichia coli DH5α was employed for plasmid replication, and Escherichia coli BL21(DE3) was used for heterologous expression. Luria-Bertani (LB) medium was used for cultivating E. coli strains, and glycerol medium MYC/01 and glucose medium MYC/02 were used for cultivating M. neoaurum strains as previously reported (1). Before being added to the media, phytosterols were prepared as a 240 mM stock solution, emulsified by Tween 80 via ultrasonication (20 min, 300 W).

TABLE 5.

Primers, plasmids, and strains used in this study

Namea	Description	Source or reference
Primers
pETkshA1N-f&r	TATCCATGGTGACTACCGAGACAGCCGGC/TATCTCGAGGCTCGGCTGAGCCGGTTCTTT	This study
pETkshA2N-f&r	TATCCATGGTGACCGATATCCGCGAGATCG/TATCTCGAGCGCAGTCGAACCGCGACCG	This study
pETkshBN-f&r	TGGCCATGGTGACGGAGGAACCGCTCGGCA/TCACTCGAGTTCGTCGTAGGTGACTTCCA	This study
p261kshA1N-f&r	CAAGGATCCAGTGACTACCGAGACAGCCGGC/TAGAAGCTTCTACAGGTTTTCCTCGACCTCGA	This study
p261kshA2N-f&r	CAAGGATCCAATGACCGATATCCGCGAGATCG/TAGAAGCTTTTACGCAGTCGAACCGCGACCG	This study
p261kstD1-f&r	TACGGATCCAGTGTTCTACATGACTGCCCAGGA/GCCAAGCTTTCAGGCCTTTCCAGCGAGATGC	This study
o-p261-f&r	TAGGCGAGTGCTAAGAATAACGTTG/TCGTTTTATTTGATGCCTGGCAGT	This study
16SrRNA-f&r	CCTATGTTGCCAGCGGGTTATGC/GCGATTACTAGCGACTCCGACTTCA	This study
RTqA1-f&r	GCGTCCTACTTCGGTCCGTCGTTCA/GTCGATCCGGGTCTTGTGCTTCCAG	This study
RTqA2-f&r	GCCTATCCGACGTACTTCAAGAAC/CACTGCAGGACGAACGAATCATGG	This study
kshA1V207T-f&r	TACCTGCACAACACCGGCCGTCCGGAT/ATCCGGACGGCCGGTGTTGTGCAGGTA	This study
kshA1P210E-f&r	AACGTCGGCCGTGAGGATGTCAACGAC/GTCGTTGACATCCTCACGGCCGACGTT	This study
kshA1D225R-f&r	GAGGCGCACCTGCGTTCCGAGGCGTCC/GGACGCCTCGGAACGCAGGTGCGCCTC	This study
kshA1N238D-f&r	TCGTTCATGATCGACTGGCTGCACAAC/GTTGTGCAGCCAGTCGATCATGAACGA	This study
Plasmids
pET-28a(+)	E. coli expression vector, Kanr	Novagen
pET28a-kshA1N	Expression plasmid pET28a possessing ORF of kshA1N	This study
pET28a-kshA1V207T	Expression plasmid pET28a possessing ORF of kshA1V207T	This study
pET28a-kshA1P210E	Expression plasmid pET28a possessing orf of kshA1P210E	This study
pET28a-kshA1D225R	Expression plasmid pET28a possessing ORF of kshA1D225R	This study
pET28a-kshA1N238D	Expression plasmid pET28a possessing ORF of kshA1N238D	This study
pET28a-kshA2N	Expression plasmid pET28a possessing ORF of kshA2N	This study
pET28a-kshBN	Expression plasmid pET28a possessing ORF of kshBN	This study
pMV261	Shuttle vector of Mycobacterium and E. coli carrying the heat shock (hsp60) promoter, Kanr	W. R. Jacobs, Jr.
p261-16srRNA	pMV261 carrying the complete ORF of the 16S rRNA gene, for calibration in absolute quantification	This study
p261-kshA1N	pMV261 derived, possessing kshA1N under the control of the hsp60 promoter for augmentation	This study
p261-kshA1V207T	pMV261 derived, possessing kshA1V207T under the control of the hsp60 promoter for augmentation	This study
p261-kshA2N	pMV261 derived, harboring Phsp60-kshA2N operon	This study
p261-kstD1	pMV261 derived, harboring Phsp60-kstD1 operon	This study
Strains
M. neoaurum
ATCC 25795	Type strain, designated Mn25795 in this study	ATCC
MutkshA1N	Mutant of Mn25795 with inactivation of KshA1N	This study
MutkshA2N	Mutant of Mn25795 with inactivation of KshA2N	This study
Strain I	MutkshA1NA2N, mutant of Mn25795 with inactivation of KshA1N and KshA2N	This study
MutkstDN	Mutant of Mn25795 with inactivation of KstD1, KstD2, and KstD3	This study
E. coli
BL21-kshA1N&kshA2N	Recombinant BL21(DE3) possessing pET28a-kshA1N or pET28a-kshA2N	This study
BL21-kshBN	Recombinant BL21(DE3) possessing pET28a-kshBN	This study
BL21-kshA1V207T	Recombinant BL21(DE3) possessing pET28a-kshA1V207T	This study
BL21-kshA1P210E	Recombinant BL21(DE3) possessing pET28a-kshA1P210E	This study
BL21-kshA1D225R	Recombinant BL21(DE3) possessing pET28a-kshA1D225R	This study
BL21-kshA1N238D	Recombinant BL21(DE3) possessing pET28a-kshA1N238D	This study
Strain II	Augmented strain containing p261kstD1, derived from strain I	This study
Strain III	Augmented strain containing p261kstD1-hsd4A, derived from strain I	This study
Strain IV	Augmented strain containing p261kshA1N, derived from MutkstDN	Yao et al. (4)
Strain V	Augmented strain containing p261kshA1V207T, derived from MutkstDN	This study

^aMn25795 is a soil isolate deposited in the American Type Culture Collection.

For the conversion of sterol, two basic processes were induced, activation and proliferation of strains and implementation of sterol conversion. For the activation and proliferation of strains, colonies that formed on LB agar medium were usually inoculated into 5 ml of MYC/01 medium and then cultured up to an optical density at 600 nm (OD₆₀₀) of 0.8 to 1.0; these cultures were directly used as seed cultures or further expanded to prepare 500 ml of seed culture for scale-up. For the implementation of sterol conversion, two methods were employed: growing cell transformation and resting cell transformation (26). Growing cell transformation was carried out in MYC/02 medium with the addition of cholesterol or phytosterols as described before (1, 25); the medium was inoculated with 10% (vol/vol) seed culture and grown at 30°C with shaking at 220 rpm or in a 7-liter bioreactor (7BG; Baoxing, Shanghai) at 250 rpm and airflow at 1 volume of air per unit of medium per minute (vvm). Resting cell transformation was carried out in 0.02 mM KH₂PO₄ buffer containing 20 g/liter phytosterols, 80 g/liter hydroxypropyl-β-cyclodextrin (HP-β-CD; RSC Chemical Industries Co. Ltd., Jiangsu, China), as previously described (26), and 100 g/liter wet mycobacterial cells harvested from the activation and proliferation culture of strains. Resting cell transformation was also performed under conditions similar to those used for growing cell transformation.

To determine the transcription and expression levels of kshA_N, as well as the growth phenotypes of kshA_N mutants, minimum medium (MM) [1.52 g liter⁻¹ KH₂PO₄, 2.44 g liter⁻¹ Na₂HPO₄, 0.5 g liter⁻¹ (NH₄)₂SO₄, 0.2 g liter⁻¹ MgSO₄ · 7H₂O, 0.05 g liter⁻¹ CaCl₂ · 2H₂O, pH 6.9] plus 18 mM glycerol (MM-glycerol) or 1.8 mM cholesterol (MM-sterol) was used. Culture in MM was also carried out at 30°C with shaking at 220 rpm, and growth was denoted by OD₆₀₀ values that were determined against blank medium. To analyze the transcription of kshA_N paralogous genes in Mn25795, cholesterol, rather than phytosterols, was used because it can be stably emulsified by Tween 80 in MM, which was diluted to 1.8 mM using stock solution (240 mM) emulsified by Tween 80 as described previously (1). To eliminate differences of Tween 80 between MM-glycerol medium and MM-sterol medium, 0.03% Tween 80 was also added into MM-glycerol medium. In addition, the induction of kshA_N by AD was determined using 1.8 mM AD diluted from an 80 mM stock solution dissolved in dimethyl sulfoxide (DMSO). To eliminate the difference, an equal amount of DMSO was also added to MM-glycerol medium as a control. To assess the growth of mutants on sterol, 35 mg of cholesterol powder was directly added into MM with or without 18 mM glycerol as a control to exclude the putative toxicity of metabolites.

DNA and RNA extraction and transcriptional analysis of kshA homologues.

The isolation of genomic DNA from Mn25795 was conducted strictly according to Uhía et al. (27). The transcription levels of kshA homologues in wild-type, disrupted, and augmented strains were determined by reverse-transcription quantitative PCR (RT-qPCR). RNA extraction and real-time quantitative PCR analyses of cDNA samples were performed as previously reported (4). The quantity and quality of DNA and RNA were determined using a NanoDrop spectrophotometer (NanoDrop Technologies) and 2% (wt/vol) agarose gel electrophoresis. The specific RNA extraction process is given in the supplemental material. When the transcription of kshA_N was analyzed in the engineered strains, the amount of kshA_N mRNA present during cholesterol transformation was quantified in comparison with that of the parent strain ATCC 25795.

The following reverse transcription process was used: 2 μl of pd(N)6 random hexamers (10 μM), 1 μg of total RNA, and 2 μl of deoxynucleoside triphosphates (dNTPs) (10 mM) were first adjusted to 15 μl using double-distilled water (ddH₂O), incubated at 70°C for 5 min, and immediately cooled on ice for 2 min. Twenty units of RNA inhibitor, 4 μl of 5× buffer with dithiothreitol (DTT), and 200 U of Super M-MLV reverse transcriptase then were added to 20 μl and incubated under the following conditions: 25°C, 10 min; 42°C, 60 min; and 95°C, 5 min. The mixture of cDNA was diluted to a final volume of 50 μl before quantification on a NanoDrop. Real-time qPCR analyses of cDNA samples were performed on a StepOne real-time PCR system (Applied Biosystems, CA). A 50-ng cDNA sample was completely mixed with 12.5 μl of SuperReal PreMix Plus with SYBR green I, 0.3 μM specific oligonucleotides (16SrRNA-f&r, RTqA1-f&r, and RTqA2-f&r) (Table 5), and 2 μl of ROX reference dye (for calibration of fluorescent signals from background and different wells) into 25 μl of reaction mixture. Amplification was performed with 15 min of predenaturing and then 40 cycles of 95°C for 10 s, 60°C for 30 s, and 72°C for 30 s. Melting curve analyses were performed at the end of each cycle. Each gene was measured in triplicate from three independent tests. The cDNA amplification efficiencies of samples, internal standards (16S rRNA), and calibrators (samples without cholesterol induction) were equivalently modulated, such that the relative amount of mRNA could be processed using the 2^−△△CT algorithm (28). Each gene was determined in triplicate in three independent tests.

Bioinformatic analyses.

Comprehensive protein information on KshA is available at the UniProtKB database (http://www.uniprot.org/uniprot/P71875), and the mapping of its secondary structure was based on data provided from PDBsum services at the EMBL-EBI website (http://www.ebi.ac.uk/pdbsum/2ZYL).

The tertiary structures of two KshAs from Mn25795 (designated KshA1_N and KshA2_N, respectively) were modeled based on alignments of their coding sequences with that of KshA_H37Rv, whose three-dimensional (3D) structure has been characterized and reported with the accession number 2ZYL in the PDB (http://www.ebi.ac.uk/pdbe-srv/view/entry/2zyl/summary.html) (3, 19). The SCWRL server assisted the analysis of differences between KshA_N homologues and KshA_H37Rv. These differences were visualized and analyzed using the PyMOL Molecular Graphics System.

Heterologous expression of Ksh enzymes in E. coli, protein purification, and activity assay.

In this study, the KshA and KshB components from Mn25795 (designated KshA_N and KshB_N, respectively) were expressed independently with a His tag at their C termini. E. coli cells harboring recombinant plasmids were developed as previously reported (4), and the primers used are shown in Table 5. Reconstituted cells were grown at 37°C at 220 rpm to an OD₆₀₀ of 0.4 to 0.6 (for KshA1_N and KshA2_N) or OD₆₀₀ of 0.2 (for an optimal soluble fraction of KshB_N), followed by a 12-h induction using 0.5 mM isopropyl-β-d-thiogalactopyranoside (IPTG) at 30°C (for KshA1_N) or 25°C (for KshA2_N). The cell pellets were first independently harvested by centrifugation at 5,000 × g for 15 min at 4°C, resuspended in 10 ml of buffer A (50 mM Tris-HCl buffer, pH 7.5, 0.5 M NaCl, 20 mM imidazole) with 1 mM phenylmethylsulfonyl fluoride (PMSF), and disrupted using a French press under 20,000 lb/in² pressure at 4°C (APV-2000-1; Germany). Final supernatants free of cell debris were obtained by centrifugation at 12,000 × g for 30 min at 4°C. As concluded by Petrusma et al. (19), the activity of KshA could not be maintained in vitro without KshB. Thus, the extracts of KshA_N and KshB_N and their premixed extracts were prepared for purification. Subsequent nickel-nitrilotriacetic acid (Ni-NTA) resin purification and dialysis steps were conducted at 4°C (29).

The Bradford method was used to determine the protein concentrations of purified KshA_N and KshB_N, with bovine serum albumin as the standard. Quantity One 1-D analysis software (version 4.62; Bio-Rad) was applied to determine the ratio of KshA_N to KshB_N in an SDS-PAGE image. Trace tracking in association with Gauss model bands, a standard processing method, assisted in eliminating background disturbances from identical lanes. Therefore, the absolute concentrations of KshA_N and KshB_N were quantified according to the signal ratio calculated by Quantity One.

Standard Ksh enzyme activity assay.

The KshB_N assay system contained 0.25 mM NADH, 0.1 mM 2,6-dichlorophenolindophenol (DCPIP; J&K), and 1 to 2 μg of KshB_N in 50 mM Tris-HCl (1 ml, pH 7.0). The system remained at room temperature for 15 min before thorough mixing. Assays were conducted by measuring the OD₆₀₀ of DCPIP (ε = 21 mM⁻¹ cm⁻¹) at 22°C (5).

According to previous reports, the optimal Ksh activity assay mixture comprised 105 μM NADH, 200 μM steroid substrate dissolved in 100% isopropanol, and 50 mM Tris-HCl (200 μl, pH 7.0) (5, 11). Up to 10 μg of KshA_N with an additional amount of KshB_N to reach saturation (ultimately, the ratio of KshA_N to KshB_N was equal to 1:2) was included in the assay mixture. Thus, the specific activity of Ksh was characterized as the amount of steroids (equal to the oxidation of NADH) hydroxylated per minute using 1 mg of total Ksh protein at pH 7.0 and 35°C (μmol min⁻¹ mg⁻¹). The relationship between the substrate and reaction velocity was processed using SigmaPlot software (version 12.1) based on the Michaelis-Menten formula. The substrate concentration varied from 10 μM to 200 μM, and V_max and K_m were calculated according to the Lineweaver-Burk plot. Recording began upon the addition of substrate, which had been subjected to a 5-min preincubation at 35°C. A SpectraMax 190 (Molecular Devices) with the Soft-max PRO program was used to detect and analyze the oxidation of NADH at 340 nm (ε = 6.22 mM⁻¹ cm⁻¹) at 35°C (5, 12). The steroids dissolved poorly in water and were uniformly adjusted to 200 μM (25 μM for 4-cholestene-3-one) in the final mixture. The control was the same mixture without addition of substrate. Products with the 9α-hydroxyl group were visualized via high-performance liquid chromatography-ultraviolet (HPLC-UV) analysis.

Construction of chimeras of KshA homologues and mutants with amino acid substitutions.

The process of exchanging β-sheets and loop regions between KshA1_N and KshA2_N is shown in the supplemental material.

Site-directed mutagenesis of KshA1_N.

The mutagenesis of KshA1_N was developed with reference to the protein alignment of KshA1 and KshA5 from Rhodococcus rhodochrous DSM 43269 (19). A QuikChange site-directed mutagenesis kit (Stratagene) was employed to generate the kshA1_V207T (GTC-ACC), kshA1_P210E (CCG-GAG), kshA1_D225R (GAT-CGT), and kshA1_N238D (AAC-GAC) mutant strains.

Deletion of kshA1_N and kshA2_N in Mn25795 and complementation thereof.

Gene deletions in Mn25795 were performed as described in previous reports (1, 14). The process of developing suicide plasmids for genetic deletion and plasmids for functional complementation is shown in the supplemental material. Pairs of primers, kshA1_N-del-U-f/kshA1_N-del-D-r and kshA2_N-del-U-f/kshA2_N-del-D-r (Table S1), were employed for PCR amplification and sequenced for an expected knockout. Since no overlapping regions were found in the homologous arms of kshA1_N and kshA2_N, Mut_kshA1NA2N was achieved by the disruption of kshA2_N on Mut_kshA1N.

In this study, complementation in mycobacteria was achieved by using pMV306-derived plasmids (p306kshA1_N [Com_kshA1N] and p306kshA2_N), and a blank pMV306 plasmid was used as the control (Com-blank). Additionally, Mut_kshA1NA2N was complemented using p306kshA1_N and p306kshA2_N, creating Com_kshA1&2N. All complemented strains with plasmids inserted were verified using PCR with the universal primers o-p306-f and o-p306-r. Complementation was confirmed by RT-qPCR to ensure the expression of complemented genes.

All plasmids constructed for gene knockout, overexpression, and complementation were electrointroduced into mycobacterial cells using electroporation protocols described previously (30).

Construction of augmented strains capable of producing 9-OHAD.

PCR products amplified from the plasmid pET28-kshA1_V207T (Table 5) using the primers p261kshA1_N-f and p261kshA1_N-r were digested with DpnI (Fermentas) to remove the template, purified, further digested with BamHI-HindIII, and cloned into pMV261 to generate pMV261-kshA1_V207T (Table 5). The primers and the plasmids used to construct other mutants are shown in Table 5. Mut_kstDN (inactivation of kstD1, kstD2, and kstD3 in Mn25795) and M. neoaurum NwIB-yV (pMV261-kshA1_N transformed into Mut_kstDN, named strain IV in this study) were introduced according to Yao et al. (4). pMV261-kshA1_V207T then was transformed into Mut_kstDN to create strain V. The primers o-p261-f and o-p261-r were used to confirm the introduction of those recombinant plasmids, and their expression levels were assessed by RT-qPCR.

Analyses of steroidal metabolites. (i) Sample preparation.

Phytosterol conversion in flasks either by growing cells or resting cells was sampled every 24 h. The growing cell transformation of phytosterols in the bioreactor was sampled every 12 h. The 9α-hydroxylation of various steroidal substrates by the reconstituted Ksh in vitro was sampled once the reaction was over. One milliliter of sample was extracted three times on a vortex mixer with 0.5 ml of ethyl acetate and 10% H₂SO₄ for 2 min before centrifugation at 12,000 × g for 20 min. At this point, thin-layer chromatography (TLC) was employed to initially evaluate the products, and gas chromatography (GC) was used to determine the levels of cholesterol or components of phytosterol mixture, as described previously (1). For liquid chromatography-mass spectrometry (LC-MS) analysis, the organic phase was transferred into a clean tube, dried under vacuum, and redissolved in 200 μl of methanol.

(ii) LC-MS.

9-OHAD dissolved in methanol was filtered through a 0.2-μm membrane and then analyzed by HPLC under the following chromatographic conditions: an Agilent XDB-C₁₈ column (4.6 by 250 mm) with a constant temperature of 40°C, a mobile phase of methanol-water (70:30), and a flow rate of 1 ml min⁻¹. The absorbance of UV₂₅₄ represented the absolute concentration of products against the standard curve of 9-OHAD, ranging from 20 μM to 5 mM. The products of Mut_kshA1N transformation culture and the in vitro assay mixtures were analyzed by electrospray ionization-LC-mass spectrometry (LC-MS; Agilent 1100LC/MSD; Agilent Technologies) with DHT, ADD, AD, T, 6-OHAD, and 9-OHAD as authentic references. The MS parameters were the following: ion spray voltage, 30 V; ion source temperature, 100°C; desolvation temperature, 300°C; and a full scan from m/z 100 to m/z 1,000 in positive ionization mode.

GC-MS plus semipreparative HPLC.

The following analysis conditions were used for gas chromatography (GC)-MS: an Agilent ZORBAX Eclipse XDB C₁₈ preparation column (4.6 by 150 mm; particle size, 5 μm) at a 1-ml min⁻¹ flow rate with the methanol-water (7/3, vol/vol) mobile phase. Peaks were collected individually and evaporated under nitrogen flow. The pellets were redissolved in acetonitrile, derivatized at 60°C with an equivalent amount of BSTFA-1% TMCS for 30 min, and then analyzed on an Agilent 6890 series GC according to Yao et al. (4). An HP 5973 mass-selective detector was operated in standby in selected ion monitoring (SIM) mode with full scans from 100 m/z to 1,000 m/z. Molecular mass was determined against the trimethylsilyl derivatization reagents.

Accession number(s).

The genome sequencing information of Mn25795 has been deposited in the GenBank database with the accession number NZ_JMDW00000000. The accession numbers of kshA1, kshA2, kstD1, kstD2, and kstD3 gene sequences from Mn25795 are KF573736, KF573737, GQ411074.1, KF772209.1 and KF772210.1, respectively.