Skip to content


Open Access

Identification, function, and application of 3-ketosteroid Δ1-dehydrogenase isozymes in Mycobacterium neoaurum DSM 1381 for the production of steroidic synthons

Microbial Cell Factories201817:77

Received: 13 November 2017

Accepted: 3 May 2018

Published: 18 May 2018



3-Ketosteroid-Δ1-dehydrogenase (KstD) is a key enzyme in the metabolic pathway for chemical modifications of steroid hormones. Only a few KstDs have thus far been characterized biochemically and applied for the production of steroidal pharmaceutical intermediates. Three KstDs, KstD1, KstD2, and KstD3, were identified in Mycobacterium neoaurum DSM 1381, and they shared up to 99, 85 and 97% amino acid identity with previously reported KstDs, respectively. In this paper, KstDs from M. neoaurum DSM 1381 were investigated and exemplified their potential application for industrial steroid transformation.


The recombinant KstD2 from Bacillus subtilis exhibited higher enzymatic activity when 4-androstene-3,17-dione (AD) and 22-hydroxy-23, 24-bisnorchol-4-ene-3-one (4HP) were used as the substrates, and resulted in specific activities of 22.40 and 19.19 U mg−1, respectively. However, the specific activities of recombinant KstD2 from Escherichia coli, recombinant KstD1 from B. subtilis and E. coli, and recombinant KstD3, also fed with AD and 4HP, had significantly lower specific activities. We achieved up to 99% bioconversion rate of 1,4-androstadiene-3,17-dione (ADD) from 8 g L−1 AD after 15 h of fermentation using E. coli transformant BL21-kstD2. And in vivo transcriptional analysis revealed that the expression of kstD1 in M. neoaurum DSM 1381 increased by 60.5-fold with phytosterols as the substrate, while the mRNA levels of kstD2 and kstD3 were bearly affected by the phytosterols. Therefore, we attempted to create a 4HP producing strain without kstD1, which could covert 20 g L−1 phytosterols to 14.18 g L−1 4HP.


In vitro assay employing the recombinant enzymes revealed that KstD2 was the most promising candidate for biocatalysis in biotransformation of AD. However, in vivo analysis showed that the cellular regulation of kstD1 was much more active than those of the other kstDs in response to the presence of phytosterols. Based on the findings above, we successfully constructed E. coli transformant BL21-kstD2 for ADD production from AD and M. neoaurum DSM 1381 ΔkstD1 strain for 4HP production using phytosterols as the substrate.


Mycobacterium 3-Ketosteroid-∆1-dehydrogenase22-Hydroxy-23,24-bisnorchol-4-ene-3-one (4HP)1,4-Androstadiene-3,17-dione (ADD) steroids


Many actinobacteria, including Mycobacterium, Streptomyces, and Rhodococcus, can utilize natural sterols as carbon and energy sources [13], and interruption of those organisms’ unique catabolic pathways often led to the accumulation of steroid hormone derivatives [4, 5], some of which are important precursors, such as C19-steroids (4-androstene-3,17-dione [AD], 1,4-androstadiene-3,17-dione [ADD], and 9α-hydroxy-4-androsten-3,17-dione [9-OHAD]), for the production of steroidal drugs [68]. Phytosterols are found in plant seeds and can be utilized for the production of AD, ADD, and 9-OHAD using Mycobacterium sp. NRRL 3805B [9], Mycobacterium sp. NRRL 3683B [1], and Mycobacterium neoaurum NwIB-yV [10], respectively. However, besides the low conversion rate [11], another major drawback of microbial transformation of phytosterols is the concomitant accumulation of byproducts due to the excessive enzymatic bioprocessing of the industrial strains [12]. Although actinobacteria strains were reported to be able to perform chemical modifications on C22-steroids, such as 22-hydroxy-23, 24-bisnorchola-4-en-3-one (4HP), 22-hydroxy-23, 24-bisnorchola-1,4-dien-3-one (HPD) and 9,22-dihydroxy-23,24-bisnorchol-4-ene-3-one (9-OHHP) [1, 12, 13] which are all valuable precursors for the synthesis of progestational and adrenocortical hormones, more informations in detailed mechanisms are urgently needed for the full industrial application.

Earlier research has, to some extent, clarified the sterol metabolic pathways in the actinobacteria based on the identification of intermediates [2]. Generally, 3-oxidation and the partial or full removal of aliphatic chains at C17 of the sterols, a process similar to fatty acid β-oxidation, are initial steps for the degradation of sterols, leading to the synthesis of 3-keto compounds such as 4HP and AD (Fig. 1) [2, 14]. Enzymes, including cholesterol oxidase (CHO), 17β-hydroxysteroid dehydrogenase/β-hydroxyacyl-CoA dehydrogenase (Hsd4A), thiolase FadA5 (FadA5), and cytochrome P450 125 (CYP125), have been reported to be involved in that degradation of sterols [12, 15]. After the degradation, as shown in Fig. 1, 4HP and AD can then be converted to HPD and ADD, respectively, by 3-ketosteroid-Δ1-dehydrogenase (KstD) [10, 12]. And HPD or ADD enter downstream oxidative process in cells after the 9α-hydroxylation, which is catalysed by 3-ketosteroid-9α-hydroxylases (KSH) [16, 17].
Figure 1
Fig. 1

An overview of proposed pathway for phytosterols degradation in mycobacteria. Phytosterols can be transformed to vary valuable intermediates, including C19-steroids (4-androstene-3,17-dione [AD], 1,4-androstadiene-3,17-dione [ADD], and 9α-hydroxy-4-androsten-3,17-dione [9-OHAD]) and C22-steroids (22-hydroxy-23, 24-bisnorchola-4-en-3-one [4HP], 22-hydroxy-23, 24-bisnorchola-1,4-dien-3-one [HPD] and 9,22-dihydroxy-23,24-bisnorchol-4-ene-3-one [9-OHHP]). KstD, 3-ketosteroid-Δ1-dehydrogenase; KSH, 3-ketosteroid-9α-hydroxylases; CYP125, cytochrome P450 125; CHO, cholesterol oxidase; Hsd4A, 17β-hydroxysteroid dehydrogenase and β-hydroxyacyl-CoA dehydrogenase; FadA5, thiolase FadA5

KstD removes hydrogen atoms of the C-1 and C-2 in the A-ring of the polycyclic ring structure of 3-ketosteroids in substrates including AD, hydrocortisone acetate, and 9-OHAD (Fig. 1) [12, 18, 19]. Although recent studies also focused on the genetic removal of KstD from native cells [10], the genetic manipulation in some strains might be challenging since host cells could contain multiple kstDs with each gene playing a different role in engineering the bacteria and altering its metabolic pathway [10, 13]. For example, in M. neoaurum ATCC 25795, KstD3 and KstD1 contributed to the conversion of AD and 9-OHAD, respectively, whereas KstD2 was probably involved in Δ1- dehydrogenation of C22 intermediates [10, 12]. Besides, enormous attempts have been made to express these KstDs heterologously in Escherichia coli, Bacillus subtilis, Pichia pastoris, and Corynebacterium crenatum to accomplish biotransformation of sterols [2024]. For example, an E. coli transformant yielded 5.6 g L−1 ADD during fed-batch fermentation [23]. A B. subtilis, expressing codon-optimized kstD gene from M. neoaurum JC-12, produced 8.76 g L−1 ADD by whole-cell biocatalysis [24]. Prominently, a KstD overexpressing C. crenatum can transform 83.87% of the AD to ADD [22].

It has been reported that Mycobacterium neoaurum DSM 1381 (Mycobacterium parafortuitum complex MCI 0617) was able to transform phytosterols to 4HP and HPD, and the molar ratio of HPD/4HP reached 16.61:1, suggesting that M. neoaurum DSM 1381 owned KstDs with high catalytic activities [11]. In this study, transcriptional analysis and heterologous overexpression of KstDs were performed to determine the isoenzymes’ biochemical roles in the biotransformation of phytosterols to HPD. KstD2 can be used to construct recombinant strains that could efficiently transform AD to ADD, and ΔkstD1 mutant of M. neoaurum DSM 1381 was found to synthesize 4HP, all of which could bring a substantive impact on the current pharmaceutical industry.


Bacterial strains, plasmids, and reagents

Mycobacterium neoaurum DSM 1381 was purchased from Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ), and the MP01 medium used to maintain M. neoaurum DSM 1381 at 30 °C was (g L−1): corn steep powder 10.0, glucose 6.0, K2HPO4 · 3H2O 2.0, MgSO4 · 7H2O 1.0, NaNO3 2.0 and Tween 80 1.0 mL (v/v) (adjusted to pH 7.5). 5 g L−1 phytosterols were added to MP01 medium to measure the steroid bioconversion performances of the M. neoaurum DSM 1381 and related transformants. For high steroid concentration fermentation, phytosterols were prepared in (2-Hydroxypropyl)-β-cyclodextrin (HP-β-CD) (1:1.5). E. coli DH5α, E. coli BL21 (DE3) and B. subtilis 6051a were cultivated with Luria–Bertani medium (LB medium) at 37 °C and 200 rpm for molecular cloning and heterologous expression of kstD genes. All the other strains and plasmids were listed in Table 1. Oligonucleotides are listed in Additional file 1: Table S1. All the plasmids were constructed with ClonExpress® II One Step Cloning Kit (Vazyme Biotech Co.Ltd. Nanjing, China). Restriction enzymes and other molecular biology reagents were purchased from Thermo Fisher Scientific.
Table 1

Bacterial strains and plasmids used in this study





 M. neoaurum DSM 1381

Mutant of M. neoaurum ATCC 25790 which can accumulate HPD, 4HP and ADD



Mutant of M. neoaurum DSM 1381 deleted kstD1

This study


ΔkstD1 harboring pMV306hsp-kstD1/pMV306hsp-kstD2/pMV306hsp-kstD3

This study


ΔkstD1 harboring pMV306Pk1-kstD1/pMV306Pk2-kstD2/pMV306Pk3-kstD3

This study


E. coli BL21 (DE3) harboring empty plasmid pET-28a(+)

This study


E. coli BL21 (DE3) harboring plasmid pET28a-kstD1/pET28a-kstD2/pET28a-kstD3

This study


B. subtilis 6051a harboring empty plasmid pHT01

This study


B. subtilis 6051a harboring plasmid pHT01-kstD1/pHT01-kstD2/pHT01-kstD3

This study



pLAM12 carrying Che9c 60–61 under control of the acetamidase promoter



Source of hygromycin (Hyg) cassette


 pET-28a (+)

E. coli expression vector, KanR



B. subtilis expression vector pHT01; containing Pgrac promoter; AmpR for E. coli; CmR for B. subtilis



Mycobacterium integrative vector with single copy, without hsp60 promoter, KanR



pMV306 contains hsp60 promoter



pET-24a(+) contains upstream and downstream of kstD1 gene flanking Hyg cassette

This study


pHT01 possessing kstD1 or kstD2 or kstD3 gene

This study


pET-28a (+) possessing kstD1 or kstD2 or kstD3 gene

This study


pMV306hsp possessing kstD1 or kstD2 or kstD3 gene

This study

 pMV306Pk1- kstD1

pMV306 carrying the 441 bp upstream region together with kstD1 gene

This study


pMV306 carrying the 726 bp upstream region together with kstD2 gene

This study

 pMV306Pk3- kstD3

pMV306 carrying the 1347 bp upstream region together with kstD3 gene

This study

Phytosterols were purchased Jiangsu Yuehong Group Company (Jiangsu, China). AD and ADD were obtained from Sigma. 4HP was from Steraloids (Newport, RI, USA). Phenazine methosulphate (PMS) and 2,6-dichlorophenolindophenol (DCPIP) were obtained from Sigma-Aldrich (Shanghai, China). Isopropyl β-d-1-thiogalactopyranoside (IPTG), Ampicillin 100 µg mL−1, kanamycin 50 µg mL−1, Hygromycin 150 µg mL−1 or 10 µg mL−1 chloramphenicol was supplemented into medium when needed.

Bioinformatic analysis

The genome of M. neoaurum DSM 1381 was isolated and cut with Covaris M220 to fragments of 400–500 bp length, and libraries of 500 bp genomic DNA fragments were builded and sequenced using Illumina Miseq (Majorbio, Shanghai). Then GS De Novo Assembler v2.8 were employed to perform genome assembly. And the genes were predicted using Glimmer 3.02 ( and annotated with BLAST 2.2.25 + . The putative genes for KstD were identified by comparing with the known KstD protein sequences taken from the NCBI database. Then the amino acid (aa) sequences of the identified KstDs in Mycobacterium neoaurum ATCC 25795 [10], Mycobacterium neoaurum NwIB-01 [18], Mycobacterium sp. VKM Ac-1817D [25], Mycobacterium sp. VKM Ac-1816D [25], Mycobacterium smegmatis mc2155 [13], Rhodococcus rhodochrous DSM43269 [26], Rhodococcus ruber strain Chol-4 [27, 28], Rhodococcus erythropolis SQ1 [2931] were used to construct phylogenetic tree using the MEGA6 software with ClustalW and neighbor-joining algorithm. FgenesB was used to predict Operons and ORFs closed to the three kstDs ( The putative binding sites of transcription factors KstR [32] and KstR2 [33] were searched among the regions 500 bp upstream plus ORFs or operon of kstDs by software package UGENE 1.27.0. The positional weight matrices (PWM) were built from the known KstR operator motifs of mycobacteria and the sites with quality parameter score no less than 85% were used for further analysis.

Construction of recombinant BL21-kstD1/BL21-kstD2/BL21-kstD3 and 6051a-kstD1/6051a-kstD2/6051a-kstD3 strains

The kstD genes were amplified from the M. neoaurum DSM 1381 genome DNA with the corresponding primers by PCR. The PCR products were cloned into BamHI/HindIII-digested E. coli expression vector pET-28a (+) (Novagen) or BamHI/SmaI-digested B. subtilis expression vector pHT01 [34], all expression plasmids were purified from E. coli DH5α and verified by DNA sequencing. pET28a-kstD1/pET28a-kstD2/pET28a-kstD3 were transformed into E. coli BL21 (DE3) following standard protocols; pHT01-kstD1/pHT01-kstD2/pHT01-kstD3 were transformed into B. subtilis 6051a by the method described previously [35]. The positive transformants of the recombinant E. coli BL21 (DE3) strains BL21-kstD1/BL21-kstD2/BL21-kstD3 and B. subtilis 6051a strains 6051a-kstD1/6051a-kstD2/6051a-kstD3 were selected with the supplement of kanamycin and chloramphenicol in the LB agar plates, respectively and then verified by DNA sequencing. The expression of kstDs in E. coli BL21 (DE3) and B. subtilis 6051a were checked by SDS-PAGE, KstD enzymatic assay and whole-cell biotransformation of 4HP and AD.

Whole-cell steroid biotransformation with E. coli BL21 (DE3) and B. subtilis 6051a recombinants

The recombinant strains for heterologous expression of kstD genes were cultured in LB medium for 8 h at 37 °C. Then the cells inoculated in LB medium (30 mL per 250 mL flask) with 10% inoculum size. The LB mediums were previously supplied with 1 mM IPTG and 1 g L−1 AD or 4HP which was dissolved in HP-β-CD (final concentration, 0.7%) and Tween 80 (final concentration, 0.1%). The cultures were incubated at 37 °C and 200 rpm for 12 h and then sampled to detect the substrate conversion rates using HPLC. Transformation capacity of BL21-kstD2 on AD was tested in the fermentation processing mentioned above except the medium changed to Terrific Broth medium which is richer in nutrition. 8 g L−1 AD, HP-β-CD (1:7) and Tween 80 (final concentration, 0.1%) were added in Terrific Broth medium.

Expression analysis by RT–qPCR

Mid-log exponential phase cultures of M. neoaurum DSM 1381 and ΔkstD1 (fermentation time: 33 h-36 h) on MP01 medium added with 5 g L−1 phytosterols and MP01 medium were collected and used to extract total RNA. After wall-breaking with liquid nitrogen grinding, standard protocol for RNAiso Plus reagents was followed to isolate RNA. Recombinant DNase I (TAKARA) was employed to eliminate the contaminating genomic DNA. The quality and concentration of RNA were evaluated by agarose gel electrophoresis, PCR test and Nano Drop 2000 (Thermo Scientific). Besides, the trace residual genomic DNA in the total RNA was removed during the reverse transcription using PrimeScript™RT reagent Kit with gDNA Eraser. 0.8 μg total RNA with gDNA Eraser and its buffer was first incubated at 42 °C, and then PrimeScript RT Enzyme Mix I, RT Primer Mix, 5 × PrimeScript Buffer 2 (for Real Time) were added to 20 mL and reacted at 37 °C for 15 min; 85 °C for 5 s; 4 °C. The cDNA products were diluted and analyzed on ViiA™ 7 Real-Time PCR System (Applied Biosystems®). SYBR® Premix Ex Taq™ (Tli RNaseH Plus) and primers (16SrRNAf&r, RTkd1f&r, RTkd2f&r, RTkd3f&r) were included in the reaction mixture of RT-qPCR and all reaction mixtures were prepared in triplicate. The reaction program was as follow: 95 °C for 15 min; 40 cycles of 95 °C for 10 s; 60 °C for 40 s; Melting Curve Stage was from 60 to 95 °C. The relative fold change of each gene was calculated using the 2−∆∆Ct algorithm and 16S rRNA was used as a reference gene. All the RT-qPCR experiments repeated for three times. The MP01 medium was set as native control. The fold change in the Fig. 5 indicates the increase degree of the expression level of the kstDs after accounting for the level detected in the MP01 medium.

Deletion of kstD1 gene

Gene deletion was performed by the methods described by Alessandro Cascioferro et al. [36]. A 1.6-kb hygromycin (hyg) cassette was amplified from PGOAL19 [37] using the primers hygdif-f&r, which contain the reported dif sequence of M. smegmatis, and then the cassette was subcloned in the EcoRI site of the plasmid pET-24a (+), then the upstream and downstream of kstD1 were ligated to the sides of the excisable cassette via HindIII and XbaI restriction sites, respectively. The result pET24a-K1UHD plasmid was the template to amplify the recombineering DNA fragments using primers kd1Uf and kd1Dr. The PCR products were transformed into M. neoaurum DSM 1381 harboring pJV53 which was employed to increase recombination efficiency [38]. The successful recombination colonies were selected using both kanamycin and Hygromycin and then verified by DNA sequencing. The strain whose kstD1 was deleted is marked as ΔkstD1.

Overexpression of KstD1, KstD2 and KstD3 in ΔkstD1

pMV306 without promoter and with promoter (pMV306hsp) were used to construct plasmids for functional complementation of KstD1 and overexpression of KstD2 and KstD3 [39, 40]. Firstly, the KstD ORFs were cloned to the EcoRI and SalI sites of pMV306hsp. And the host promoter regions together with the ORFs of KstDs were amplified from M. neoaurum DSM 1381 genome and subcloned into pMV306 between BamHI and EcoRI sites. After being verified by DNA sequencing, the resulting pMV306hsp-kstD1/pMV306hsp-kstD2/pMV306hsp-kstD3 and pMV306Pk1-kstD1/pMV306Pk2-kstD2/pMV306Pk3-kstD3 were introduced in ΔkstD1 by electroporation. The colonies of ΔkstD1 harboring the right recombinant plasmids were picked and verified using the primers kan-f&r to check the presence of plasmids. The result recombinant strains were noted as HK1, HK2, HK3, PK1, PK2 and PK3, respectively. Then bioconversion of phytosterols using the recombinant ΔkstD1 strains was performed to study the characteristics of the three KstD isozymes and their promoters in ΔkstD1.

KstD enzymatic assay

Induced with 1 mM IPTG for 24 h, the cell pellets obtained at 6000 rpm for 10 min at 4 °C from 50 mL cultures of the recombination E. coli BL21 (DE3) and B. subtilis 6051a strains were resuspended in 4 mL 50 mM Tris–HCl buffer (pH 7.0) after twice washing and then sonicated for 10 min under the protection of ice-water bath. Then the supernatant of cell extracts (12,000 rpm, 4 °C, 5 min) was used for enzyme activity assay. And for the M. neoaurum DSM 1381 and ΔkstD1, after induced with 5 g L−1 phytosterols (33–36 h), the cells were collected in the same way. The KstD enzymatic activities of soluble part of both the cultures and cell extracts obtained at 12,000 rpm for 30 min at 4 °C were measured spectrophotometrically at 600 nm (ε600 = 18.7 × 103 cm−1 M−1) by Thermo Scientific Nano Drop 2000 at 30 °C [28]. The reaction mixture (1 mL) contained 50 mM Tris–HCl pH 7.0, 1.5 mM PMS, 0.12 mM DCPIP, cell extracts or supernatants of cultures, 500 μM AD/4HP and 5 g L−1 HP-β-CD. The substrates were previously dissolved in HP-β-CD. Three replicates were analyzed. The enzymatic activity of the specific substrate of each sample was calculated by subtracting the value of the activity of control (without any steroid). The total protein content in the supernatants of cell extracts and cultures were quantified by Bradford assay [41]. 1 U of enzyme activity is defined as the reduction of 1 µmol of DCPIP/min.

SDS-polyacrylamide gel electrophoresis of proteins (SDS-PAGE)

The samples used for SDS-PAGE were mixed with 5 × SDS loading buffer (Shanghai Songon) with ratio 4:1 (v/v). The mixture was then boiled in water for 10 min, and centrifuged for 10 min at 12,000 rpm. The samples were run on a SDS-PAGE as described by Laemmli [42].

Analytical methods

The fermentation experiments of M. neoaurum DSM 1381 were sampled every 12 or 24 h and three replications were used to measure steroids. The samples were extracted three times with an equal volume of ethyl acetate, and then the three extracts were mixed to analyze with thin layer chromatography (TLC), high-performance liquid chromatography (HPLC) and gas chromatography (GC). TLC was used as a qualitative approach to detect the steroids bioconversion products with ethyl acetate-hexane (6:4) as a developing solvent. Samples in ethyl acetate were re-dissolved in methanol after drying. HPLC with Agilent Extend-C18 column (4.6 × 250 mm; 40 °C) was used to determine the 3-ketosteroids with methanol/water (80:20, v/v) as the mobile phase at a flow rate of 1 mL min−1 and the wave length of ultraviolet detector was 254 nm. To quantify phytosterols having no ultraviolet absorption, the chromatographic method was performed on a Rtx-5 (30 m × 0.53 mm × 5.0 µm) using squalene (Sigma) as an internal reference standard. GC-2010Plus (Shimadzu, Japan) with a flame ionization detector was employed. The temperatures of inlet, column and flame-ionization detector were 320, 300 and 320 °C, respectively.

Accession numbers

The kstD1, kstD2, and kstD3 ORF sequences from M. neoaurum DSM 1381 have been deposited in the GenBank database with the Accession Numbers MG251735, MG251736, and MG251737, respectively.


In silico analysis of the three putative KstD isoenzymes

M. neoaurum DSM 1381’s genome were sequenced and annotated as described in “Methods”. Three putative kstD ORFs (orf04645, orf05164, and orf05167) were identified, and the ORF sequences have been deposited in GenBank database. The genetic organizations of the three kstD ORFs are shown in Additional file 2: Fig. S1. orf04645, hereinafter referred to as kstD1, was located within a gene cluster associated with steroid degradation and was surrounded by hsaF (4-Hydroxy-2-ketovalerate aldolase) and hsd4B (3β-Hydroxysteroid dehydrogenase/Δ54-isomerase). kstD2 (orf05167) was adjacent to kstD3 (orf05164) and kshA (3-Ketosteroid 9α-hydroxylase oxygenase subunit). According to FgenesB, kstD1 and kstD2 were regulated independently by their promoters, and kstD3 was located on an operon. Further, KstD expression is thought to be regulated through the binding of helix-turn-helix transcriptional repressors (KstRs) [32, 33]. UGENE 1.27.0 identified a putative KstR1 binding site before the kstD1 gene with a quality parameter score of 89.26%. Nevertheless, no KstR1 binding site was predicted before kstD2 and kstD3, and none of the three kstDs were under the control of KstR2.

The nucleotide sequences of the kstDs were translated into aa sequences. A dendrogram was subsequently composed to elucidate evolutionary relationships between the three putative KstD proteins and their homologs, obtained from recent literature. As shown in Fig. 2, the aa sequence of KstD1 of M. neoaurum DSM 1381 shared a high identity with KstD1 of M. neoaurum ATCC 25795 (97%) [10], KstD1 of Mycobacterium sp. VKM Ac-1816D (99%) [25], KstD3 of R. ruber Chol-4 (66%) [28], and KstD3 of R. erythropolis SQ1 (67%) [31]. KstD3 of M. neoaurum DSM 1381 showed 97, 68, and 46% identity with the homologous protein of M. neoaurum ATCC 25795, R. ruber Chol-4 and R. erythropolis SQ1, respectively [10, 28, 29]. Unexpectedly, KstD2 of M. neoaurum DSM 1381 shared only 85% aa identity with the KstD2 of M. neoaurum ATCC 25795, but 68% and 65% aa identity with the KstD2 of R. ruber Chol-4 and R. erythropolis SQ1, respectively [10, 28, 30]. Moreover, through NCBI BLAST, the highest identity found that corresponds to the KstD2 sequence of M. neoaurum DSM 1381 was 85%, suggesting a new unidentified KstD.
Figure 2
Fig. 2

Phylogenetic tree of three KstDs in M. neoaurum DSM 1381 and their orthologues. 1000 replicates were set to generate bootstrap values when conducting the analysis

Heterologous expression of KstD1, KstD2 and KstD3 of M. neoaurum DSM 1381 in E. coli BL21 (DE3) and B. subtilis 6051a

The expression of KstDs in E. coli BL21 (DE3) or B. subtilis 6051a were identified through SDS-PAGE (Additional file 3: Fig. S2). To detect the activities of the KstDs, the biochemical properties of recombinant KstD1, KstD2, and KstD3 from E. coli BL21 (DE3) or B. subtilis 6051a were investigated (Table 2). Host cells with the blank plasmid were used as controls. The intracellular and extracellular KstD activities of BL21-kstD1/BL21-kstD2/BL21-kstD3 and 6051a-kstD1/6051a-kstD2/6051a-kstD3 on AD and 4HP were separately measured. As shown in Table 2, the KstD activity of BL21-kstD2, with AD and 4HP as substrates, were 3.46 and 2.82 U mg−1, respectively. Unexpectedly, recombinant KstD1 and KstD3 from E. coli BL21 (DE3) showed very low level of activities (≤ 0.34 U mg−1) on both AD and 4HP. However, the activities of intracellular soluble parts from 6051a-kstD1 on AD and 4HP reached 2.12 and 1.81 U mg−1, respectively. The improvement on enzymatic activity of recombinant KstD1 was probably caused by poorly expressed enzyme E. coli due to the recombinant KstDs forming inclusion body in E. coli, as shown in Additional file 3: Fig. S2, which were also reported previously [21, 23]. And the same phenomenon was also noticed on the KstD2. The intracellular KstD activity of recombinant 6051a-kstD2 reached 22.40 and 19.19 U mg−1, respectively. Nevertheless, the recombinant KstD3’s activity was still negligible in B. subtilis, further verifying the poor activity of KstD3.
Table 2

The KstD enzyme activity (U mg−1 total soluble protoplast protein and extracellular protein) of M. neoaurum DSM 1381, ΔkstD1, recombinant E. coli BL21 (DE3), and B. subtilis 6051a strains





Intracellular enzyme

Extracellular enzyme

Intracellular enzyme

Extracellular enzyme

DSM 1381

2.45 ± 0.05


2.00 ± 0.19



0.14 ± 0.01


0.06 ± 0.01








2.12 ± 0.06


1.81 ± 0.05



22.40 ± 1.26


19.19 ± 0.96



0.18 ± 0.06


0.15 ± 0.06








0.34 ± 0.04


0.21 ± 0.05



3.46 ± 0.07


2.82 ± 0.12



0.29 ± 0.08


0.32 ± 0.14


All assays were performed with triplicate cultures. Standard deviations of the biological replicates are shown

DSM1381, M. neoaurum DSM 1381; ΔkstD1, mutant of M. neoaurum DSM 1381 deleted kstD1; BL21-kstD1/BL21-kstD2/BL21-kstD3, E. coli BL21 (DE3) containing pET28a-kstD1/pET28a-kstD2/pET28a-kstD3; BL21-pET28a, E. coli BL21 (DE3) containing empty plasmid pET28a(+); 6051a-kstD1/6051a-kstD2/6051a-kstD3, B. subtilis 6051a containing pHT01-kstD1/pHT01-kstD2/pHT01-kstD3; 6051a-pHT01, B. subtilis 6051a containing empty plasmid pHT01

u.d. undetectable enzyme activity

In conclusion, among the three KstDs, KstD2 showed the highest enzymatic activity when expressed heterogeneously, and KstD1 performed poorly, especially in E. coli. In addition, for KstD3, the KstD enzyme activities were hardly detected in either host.

Bioconversion of AD and 4HP by the recombinant E. coli BL21 (DE3) and B. subtilis 6051a

The gene expression and activity of the three KstDs of M. neoaurum DSM 1381 in E. coli and B. subtilis was tested in shake flask fermentation process (Fig. 3). Considering the low solubility of steroids in purified water, both HP-β-CD and Tween 80 were used to improve AD and 4HP solubility in the fermentation medium. The KstDs displayed different AD and 4HP conversion capacities in the different hosts. After 12 h of fermentation, 6051a-kstD1 transformed 20.48% of AD to ADD and 27.41% of 4HP to HPD. By comparison, BL21-kstD1 performed poorly. These observations were consistent with the KstD enzyme activity assay results. Further, KstD3 was only able to transform a maximum of 2.39% substrate in either of the host. Remarkably, BL21-kstD2 converted as much as 95% of AD and 63.41% of 4HP in 12 h. Even though 6051a-kstD2 was more active than BL21-kstD2, based on the enzyme activity assay, KstD2 in recombinant B. subtilis was only able to convert 88.52% of AD to ADD and 63.31% of 4HP to HPD. One of the reasons for the lower conversion rate with 6051a-kstD2 was that the KstD activity of BL21-kstD2 is higher than that of 6051a-kstD2 during the early stage after the IPTG induction, prior to the formation of inclusion bodies. The steroid conversion results further demonstrated that KstD2 displays a high dehydrogenation activity and that BL21-kstD2 is an excellent ADD producer. To maximize the AD dehydrogenation capacity of BL21-kstD2, AD fermentation was performed in Terrific Broth medium. Under the condition described in “Methods”, BL21-kstD2 was capable of transforming 8 g L−1 of AD to ADD at a 99% conversion rate in 15 h (Fig. 4).
Figure 3
Fig. 3

AD and 4HP bioconversion assays of recombinant strains. LB medium supplied with 1 g L−1 AD or 4HP was used to test the transformation ability of the recombinant E. coli BL21 (DE3) and B. subtilis 6051a strains. BL21-kstD1/BL21-kstD2/BL21-kstD3, E. coli BL21 (DE3) containing pET28a-kstD1/pET28a-kstD2/pET28a-kstD3; BL21-pET28a, E. coli BL21 (DE3) containing empty plasmid pET28a(+); 6051a-kstD1/6051a-kstD2/6051a-kstD3, B. subtilis 6051A containing pHT01-kstD1/pHT01-kstD2/pHT01-kstD3; 6051a-pHT01, B. subtilis 6051a containing empty plasmid pHT01

Figure 4
Fig. 4

Time course of ADD accumulation from AD by BL21-kstD2 and BL21- pET28a. BL21-kstD2, E. coli BL21 (DE3) containing pET28a-kstD2; BL21-pET28a, E. coli BL21 (DE3) containing empty pET28a(+); Green circle, conversion rate by BL21-pET28a; black inverse triangle, concentration of remained AD by BL21-kstD2; blue triangle, real-time yield of ADD by BL21-kstD2; red square, conversion rate by BL21-kstD2. Standard deviations of the biological replicates are represented by error bars

Transcriptional analysis of kstD genes in M. neoaurum DSM 1381

The transcription levels of the three kstD genes in M. neoaurum DSM 1381 were analyzed by RT-qPCR following an addition of phytosterols. Cultures grown in MP01 medium containing 5 g L−1 phytosterols and control cultures grown in same medium without added steroids were collected to extract total RNA. As shown in Fig. 5, kstD1 transcripts increased 60.5-fold in M. neoaurum DSM 1381 when induced with phytosterols. However, only a 1.6- and 2.5-fold increase in, respectively, kstD2 and kstD3 transcription was observed in response to the same treatment. These results were consistent with the gene sequencing analysis where a putative KstR binding site was only identified before kstD1. Taken together, only kstD1 in M. neoaurum DSM 1381 was believed to be involved in the transformation of phytosterols. Further, the genetic regulation of kstD homologs, belonging to the same phylogenetic tree branch, seemed to be conserved. For example, a 4.5-, 13-, and 240.5-fold upregulation of, respectively, kstD1 of M. neoaurum ATCC 25795 [10], MSMEG_5941 of M. smegmatis mc2155 [13], and kstD3 of R. ruber Chol-4 [28] were observed following an induction with cholesterol. Moreover, in M. neoaurum ATCC 25795, only kstD1 and kstD3 contributed to the accumulation of ADD from cholesterol. After deleting MSMEG_5941, a homolog of M. neoaurum DSM 1381 kstD1, the AD molar yield for phytosterols became 84%, and only a small amount of ADD and HPD were accumulated in ΔkshB1 M. smegmati mutant [13]. This may suggest that among the three kstDs, only kstD1 was involved in the transformation of phytosterols, which would agree with earlier reports for other Mycobacterium strains. Based on our results it was anticipated that the deletion of the kstD1 gene would inhibit the conversion of 4HP to HPD.
Figure 5
Fig. 5

Analysis of kstD genes transcription in M. neoaurum DSM 1381 and ΔkstD1 strains. The fold change indicates the ratio of phytosterols supplemented to control culture mRNA levels for the indicated strains. Data were calculated from three independent experiments and standard deviations are shown

Construction of the ΔkstD1 M. neoaurum DSM 1381 mutant

As previously reported, the ΔkstD1 M. neoaurum DSM 1381 mutant clones that lost their hyg resistance were verified by PCR and the DNA fragment of the desirable mutants was sequenced [36], from which it was found that only one dif site was left between the upstreasm and downstream of kstD1. Then, the phenotypes of the ΔkstD1 mutant were studied, and its transcription levels of kstD2 and kstD3 were quantified. Following the deletion of KstD1, the expressions of kstD2 and kstD3 rose 2.4- and 1.4-fold, respectively (Fig. 5). This strongly suggested that even if kstD1 was deleted, kstD2 and kstD3 could not be induced by phytosterols. Further, compared to the wild-type, the intracellular enzyme activity of ΔkstD1 on AD had decreased from 2.45 to 0.14 U mg−1 (Table 2). Also, ΔkstD1 showed only slight activity for 4HP (Additional file 4: Fig. S3). Finally, when fermented with 5 g L−1 phytosterols, the M. neoaurum DSM 1381 wild-type produced 0.2 g L−1 4HP, 3.0 g L−1 HPD, 17.8 mg L−1 AD, and 32.5 mg L−1 ADD over 132 h, while the ΔkstD1 accumulated mainly 4HP (2.8 g L−1). Yields of HPD, AD, and ADD were merely 0.1 g L−1, 48.4 mg L−1, and 9.4 mg L−1, respectively. Overall, kstD1 has been shown to play a major role in the production of HPD from 4HP. As expected, the ΔkstD1 mutant mainly produced 4HP.

kstD1, kstD2 and kstD3 expression affects the molar ratio of HPD/4HP and ADD/AD in ΔkstD1

The KstDs were overexpressed in the ΔkstD1 mutant, either using a strong constitutive promoter (hsp60), or under the control of their native promoter, to determine their specific function in the transformation of phytosterols to HPD. The molar yields of AD, ADD, 4HP and HPD after 144 h fermentation with 5 g L−1 phytosterols were measured (Table 3) and compared to the wild type M. neoaurum DSM 1381 (Fig. 6).
Table 3

Comparison of the molar yields (%) of products of ΔkstD1 and its recombinant strains using 5 g L−1 phytosterols


DSM 1381









0.99 ± 0.11

0.62 ± 0.17

1.01 ± 0.06

1.07 ± 0.27

0.49 ± 0.08

0.94 ± 0.16

1.38 ± 0.34

0.59 ± 0.22


0.28 ± 0.04

0.8 ± 0.15

0.61 ± 0.15

0.3 ± 0.06

0.57 ± 0.09

0.31 ± 0

0.8 ± 0.22

0.52 ± 0.12


68.79 ± 6.76

1.63 ± 0.13

59.37 ± 5.62

72.19 ± 2.03

2.22 ± 0.21

67.96 ± 3.9

8.84 ± 0.26

1.8 ± 0.31


3.8 ± 0.63

67.01 ± 0.37

12.52 ± 1.68

3.47 ± 0.54

68.07 ± 1.55

5.63 ± 0.4

66.42 ± 2.72

68.82 ± 2.94

DSM1381, M. neoaurum DSM 1381; ΔkstD1, mutant of M. neoaurum DSM 1381 deleted kstD1; HK1/HK2/HK3, ΔkstD1 harboring pMV306hsp-kstD1/pMV306hsp-kstD2/pMV306hsp-kstD3; PK1/PK2/PK3, ΔkstD1 harboring kstD1/kstD2/kstD3 expression plasmids under control of their own promoter

Figure 6
Fig. 6

Molar ratio of ADD/AD and HPD/4HP of M. neoaurum DSM 1381 and its recombinant strains. The assay was carried out with MP01 medium supplied with 5 g L−1 phytosterols. DSM1381, M. neoaurum DSM 1381; ΔkstD1, mutant of M. neoaurum DSM 1381 deleted kstD1; HK1/HK2/HK3, ΔkstD1 harboring pMV306hsp-kstD1/pMV306hsp-kstD2/pMV306hsp-kstD3; PK1/PK2/PK3, ΔkstD1 harboring kstD1/kstD2/kstD3 expression plasmids under control of their own promoters

Data for the constitutively expressed KstDs showed that KstD2 has a higher Δ1-dehydrogenation activity than KstD1 since the HPD/4HP molar ratio of HK2 (20.05:1) was much higher than HK1 (4.76:1) and even higher than M. neoaurum DSM 1381 (16.61:1). Further, the HPD/4HP molar ratios of HK3 (0.03:1) only slightly exceeded those of the ΔkstD1 mutant (0.02:1), implying that the Δ1-dehydrogenation activity of KstD3 is negligible. The HPD/4HP molar ratio of PK1, or in other words the native expression of KstD1 in ΔkstD1, could be recovered to 12.09:1. And it is worth noting that the HPD molar yield of PK2 was 8.84%, much higher than in ΔkstD1 (1.63%), and this might be due to the use of a weak native promoter and the concomitant high activity of KstD2, based on our RT-qPCR results and heterologous expression experiments. It is not thought that this lack of KstD3 activity in PK3 was related to the native promoter of KstD3 in the cassette used for KstD3 overexpression, as the KstD3 activity on 4HP and AD was also negligible in HK3. In conclusion, kstD1 had a strong promoter with high Δ1-dehydrogenation activity in M. neoaurum DSM 1381; kstD2 was found to play only a minor role in the phytosterols conversion; and the effect of kstD3 appeared negligible.

Phytosterols conversion capacity of the ΔkstD1 mutant

Figure 7 showed that the conversion rates of 5, 10, 15 and 20 g L−1 phytosterols over 168 h of fermentation was 100, 98.7, 98.6, and 96.3%, respectively. Specifically, in the space of 96 h, ΔkstD1 was shown to have transformed all 5 g L−1 phytosterols into 2.88 g L−1 4HP, 0.15 g L−1 HPD, and 82.10 mg L−1 AD. The yield of 4HP increased to 6.78, 9.80, and 14.18 g L−1 when fed with 10, 15, and 20 g L−1 phytosterols, respectively. The purities of 4HP were 92.3, 94, 95, and 96.12%, respectively, and this was thought to be the result of a continuous conversion from 4HP to HPD after most of the substrates were utilized. The former was deduced from our observations showing a similar 4HP/HPD molar ratio for 5 g L−1 phytosterols after 48 h (47.65:1) as for 20 g L−1 phytosterols after 168 h (48.99:1); and the latter being higher than the 4HP/HPD ratio for 5 g L−1 phytosterols after 168 h (18.5:1). Thus, elongation of the fermentation time would decrease the purity of 4HP but would benefit the conversion rate. This may have been due to the native expression of other KstDs and the pathway accumulating AD.
Figure 7
Fig. 7

Time course of products accumulation from phytosterols by ΔkstD1. 1.5 parts of HP-β-CD to 1 part phytosterols was used to enhance bioconversion. Black circle, 5 g L−1 phytosterols; green triangle, 10 g L−1 phytosterols; blue inverse triangle, 15 g L−1 phytosterols; red square, 20 g L−1 phytosterols. HP-β-CD with 1.5:1 ratio to phytosterols was employed to enhance the bioconversion. a The concentration of remained phytosterols. bd Real-time yield of 4HP, HPD and AD, respectively. All assays were performed in triplicate. Standard deviations of the biological replicates are represented by error bars


Many mutants have already been obtained using conventional breeding methods to produce a variety of drug intermediates and raw materials [1, 43]. However, the underlying mechanisms have only been studied for the last 20 years [3, 44, 45]. As whole-genome sequencing and transcriptome sequencing becomes more affordable, there is an increasing interest to identify the genes encoding key enzymes in actinobacteria [10, 25]. In fact, many mutants, such as M. neoaurum DSM 1381 [11] and Mycobacterium sp. VKM Ac-1815D [25], already have a valuable gene database that can be used as a tool for constructing novel strains to meet the demands of the pharmaceutical industry. M. neoaurum DSM 1381, a HPD/4HP producing M. neoaurum mutant [11], was considered a good candidate strain to explore the steroid degradation pathway responsible for the accumulation of HPD. To the best of our knowledge, this paper is the first time to present the performance of three KstDs isoenzymes during phytosterols conversion in M. neoaurum DSM 1381.

In this study, three kstD genes were identified in the M. neoaurum DSM 1381 genome. It is worth noting that, even though M. neoaurum DSM 1381 and M. neoaurum ATCC 25795 belong to the same species, the aa sequences of the KstDs in this study have highlighted certain differences. KstD1 shared a high aa sequence identity with its homologs of other M. neoaurum strains. For the KstD3, the shared aa identity of M. neoaurum ATCC 25795 and M. neoaurum DSM 1381 was as high as for KstD1 (97%). KstD2, on the other hand, only shared 85% aa identity with KstD2 from M. neoaurum ATCC 25795 [10] which was concomitantly also the highest shared aa identity following a blast search. KstD2 was therefore considered a new enzyme, different from previously reported KstDs. As shown in Fig. 8, the three isoenzymes are flavoproteins containing a consensus N-terminal flavin adenine dinucleotide (FAD)-dependent domain (GSGX5–6AX2AX8E) [10, 19]. Four residues, considered significant for flavoprotein functioning, were found to be highly conserved in the three KstDs: Tyr119, Tyr487, and Gly491, in the FAD-binding domain, and Tyr318 in the catalytic domain [45]. Nevertheless, mutations of other sites of KstD enzymes also have been shown to influence their activity. For example, p.S138L decreased the activity of KstD1 of M. neoaurum, whereas a p.V366S increased its activity [46, 47]. Furthermore, a p.Y125H substitution of the KstD1 from M. neoaurum ATCC 25795 had only a relatively small effect [10]. We would further study the function of the three isoenzymes to further our understanding and elucidate the reaction mechanisms of the KstD enzymes.
Figure 8
Fig. 8

The sequence alignment of known KstD enzymes. DSM1381-KstDs from M. neoaurum DSM 1381, ATCC25795-KstDs from M. neoaurum ATCC 25795, DSM43269-KstDs from R. rhodochrous DSM 43269, Cho1-4-KstDs from R. ruber strain Chol-4, SQ1-KstDs from R. erythropolis SQ1. Active site residues and residues involved in co-ordination of a FAD in SQ1-KstD1 are indicated by number sign. A conserved sequence for FAD-binding region is indicated by asterisk

According to our results, the performance of KstD1 from M. neoaurum DSM 1381 is much better in B. subtilis than that in E. coli. This is in agreement with the performance of its homolog from M. neoaurum JC-12 and implied that these two KstDs probably shared the same characteristics [21]. Compared with the KstD3 from M. neoaurum ATCC 25795, the KstD3 from M. neoaurum DSM 1381 hardly showed detectable activity on both AD and 4HP [10]. Following the analysis of its aa sequence, it is possible that these observations can be attributed to the eight substitutions near Tyr318 in the catalytic domain of KstD3. KstD2 of M. neoaurum DSM 1381 showed remarkably high activities on both AD and 4HP. The properties of KstD2 offer exciting prospects to increasing our knowledge of the catalytic mechanisms of KstDs and as an application for drug development. The latter is owing to the ease by which KstD2 can be expressed in both commonly used hosts, whereas most other well-studied KstDs can only be actively expressed at low temperature or by addition of osmolytes [19, 23, 28]. Taken together, the research conducted so far suggests that KstD2 of M. neoaurum DSM 1381 has several characteristics that would favor its use on an industrial scale.

Recently, several KstDs were expressed in common hosts to construct industrial stains [23]. In E. coli, 6 g L−1 hydrocortisone can currently be transformed into prednisolone when expressing MsKstD1 from M. smegmatis mc2155 [19]. P. pastoris expresssion KstDF from Aspergillus fumigatus CICC 40167 is able to transform 1 g L−1 AD to ADD after 4 h of fermentation, however, the cell culture needs a further 6 days to establish [20]. In addition, KstD1 from M. neoaurum JC-12 has been expressed in C. crenatum, E. coli and B. subtilis to produce ADD from AD [2224]. The highest yield (8.76 g L−1) was obtained in B. subtilis, with a fed-batch strategy in 50 h, after codon optimization and co-expression with catalase to remove H2O2 toxicity [24]. Most of the research was carried out using a whole-cell biocatalyst method. The optimum temperature of these KstDs is 30 or 25 °C, which does not promote bacterial growth [19, 22]. Nonetheless, these KstD activities are still quite low. The discovery of KstD2 of M. neoaurum DSM 1381 is therefore of great importance. Compared to previously studied KstDs, the BL21-kstD2 clone, expressing KstD2 of M. neoaurum DSM 1381, effectively transformed up to 8 g L−1 AD to ADD after 15 h of fermentation. In brief, BL21-kstD2 is believed to be a promising industrial strain for the effective transformation of 4-ene-3-oxosteroids.

In M. neoaurum DSM 1381, most of the KstD activities were contributed by kstD1 according to RT-qPCR and overexpression experiment results. In agreement with earlier reports of KstD1 homologues from Rhodococcus to Mycobacterium, kstD1 expression in M. neoaurum DSM 1381 was induced by phytosterols [13, 25, 28], a TetR-type transcriptional regulator involved in the steroid metabolism within actinobacteria, in the upstream region of kstD1. Compared to the kstD2 homologs in M. neoaurum ATCC 25795, KstD2 of M. neoaurum DSM 1381 showed a much higher activity on AD but only played a minor role in phytosterols conversion to 4HP, possibly due to its weak gene expression levels [10]. Its homolog in M. smegmatis also showed little effect on AD degradation. However, KstD2 homologs in R. rhodochrous DSM43269 [26], R. ruber strain Chol-4 [27, 28] and R. erythropolis SQ1 [2931] were the main contributors to AD degradation. The operon containing kstD3 and controlled by a promoter could not be induced by phytosterols. The effect of KstD3 on phytosterols conversion to HPD is considered negligible since both its transcription and activity were low. As reported for M. neoaurum ATCC 25795, KstD3 played a role in AD metabolism but had no obvious impact on the Δ1-dehydrogenation of 4HP [10, 12]. Its homolog in M. smegmatis showed high activities on multiple substrates but made little contribution to AD degradation. R. rhodochrous DSM43269 [26], R. ruber strain Chol-4 [27, 28] and R. erythropolis SQ1 [31] harbor an active KstD3 homolog (Fig. 8).

Compared to rhodococci, mycobacteria are better hosts for construct 4HP or HPD producers, as the corresponding C22 intermediate of Rhodococcus is 4-pregnen-3-one-20β-carboxylic acid (4-BNC) rather than 4HP, and the reason for this has remained elusive to date [17]. In this paper, the molar yield of 4HP increased from 3.8 to 67.01%, after the deletion of kstD1 in M. neoaurum DSM 1381, which is much higher than the XIIΔhsd4AΔkstD123 mutant of M. neoaurum ATCC 25795 (47–49%) [12]. In M. neoaurum ATCC 25795, 1–2% HPD accumulated even after kstD1, kstD2, kstD3 were removed and would suggest that there are other dehydrogenases that might contribute to the degradation of steroids [12]. Therefore, it was thought that ΔkstD1 could not be optimized further by a simple deletion of kstD2 to eliminate the appearance of HPD. Further research into the transformation mechanism of 4HP to HPD would be required to remove the small amount of AD (0.8% molar yield). Nevertheless, the ΔkstD1 mutant of M. neoaurum DSM 1381 was shown to be an excellent strain due to its high molar yield of 4HP (65–73%), a high capacity to utilize substrate (96.3% of 20 g L−1 phytosterols) and the low accumulation of byproducts.


In the transformation of phytosterols to HPD in M. neoaurum DSM 1381, KstD1 has been shown to play a dominant role, whereas KstD2 was only a minor contributor and KstD3 activity was negligible. KstD2, in particular, was shown to be a novel and strong candidate for industrial applications as it demonstrated high activity and could easily be expressed in E. coli and B. subtilis. And the recombinant BL21-kstD2 was proven to be a promising ADD producer. In addition, our research has led to the construction of an excellent 4HP producing strain, obtained by deleting kstD1 in M. neoaurum DSM 1381. This ΔkstD1 mutant could produce 14.18 g L−1 4HP from 20 g L−1 phytosterols after 168 h of fermentation.









22-Hydroxy-23, 24-bisnorchola-4-en-3-one


22-Hydroxy-23, 24-bisnorchola-1,4-dien-3-one




4-Pregnen-3-one-20β-carboxylic acid


open reading frame


phenazine methosulphate






thiolase FadA5


cytochrome P450 125


cholesterol oxidase


17β-Hydroxysteroid dehydrogenase and β-hydroxyacyl-CoA dehydrogenase




4-Hydroxy-2-ketovalerate aldolase


3β-Hydroxysteroid dehydrogenase/Δ54-isomerase


3-Ketosteroid 9α-hydroxylase oxygenase subunit




high performance liquid chromatography


gas chromatography


Authors’ contributions

RJZ, XCL, YSW, and YCH made the heterologous expression in E. coil and B. subtilis. RJZ, and XCL carried out the kstD1 deletion. RJZ performed the RT-qPCR experiment and overexpression study. RJZ and BGZ analyzed the data. RJZ, BGZ, JSS and JPS conceived the study and reviewed the manuscript. All authors read and approved the final manuscript.


We sincerely thank Wolfgang Schumann (Institute of Genetics, University of Bayreuth, D-95440, Bayreuth, Germany) for providing the plasmids pHT01 and and W.R. Jacobs Jr. (Howard Hughes Medical Institute) for providing the plasmids pMV306. pMV306hsp and pGOAL19 were gifts from Brian Robertson & Siouxsie Wiles (Addgene plasmid # 26155) and Tanya Parish (Addgene plasmid # 20190), respectively.

Competing interests

The authors declare that they have no competing interests.

Availability of data and materials

All data generated and analyzed during this study are included in this published article and its additional files.

Consent for publication

Not applicable.

Ethics approval and consent to participate

Not applicable.


The authors gratefully acknowledge the following funding support for this research: the National Natural Science Foundation of China (41106125).

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (, which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated.

Authors’ Affiliations

Lab of Biorefinery, Shanghai Advanced Research Institute, Chinese Academy of Sciences, Pudong, China
School of Life Science and Technology, ShanghaiTech University, Shanghai, China
Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences, Shanghai, China
University of Chinese Academy of Sciences, Beijing, China


  1. Donova MV, Egorova OV. Microbial steroid transformations: current state and prospects. Appl Microbiol Biotechnol. 2012;94:1423–47.View ArticleGoogle Scholar
  2. Garcia JL, Uhia I, Galan B. Catabolism and biotechnological applications of cholesterol degrading bacteria. Microb Biotechnol. 2012;5:679–99.View ArticleGoogle Scholar
  3. Van der Geize R, Yam K, Heuser T, Wilbrink MH, Hara H, Anderton MC, Sim E, Dijkhuizen L, Davies JE, Mohn WW, Eltis LD. A gene cluster encoding cholesterol catabolism in a soil actinomycete provides insight into Mycobacterium tuberculosis survival in macrophages. Proc Natl Acad Sci USA. 2007;104:1947–52.View ArticleGoogle Scholar
  4. Donova MV, Gulevskaya SA, Dovbnya DV, Puntus IF. Mycobacterium sp. mutant strain producing 9 alpha-hydroxyandrostenedione from sitosterol. Appl Microbiol Biotechnol. 2005;67:671–8.View ArticleGoogle Scholar
  5. Lo CK, Pan CP, Liu WH. Production of testosterone from phytosterol using a single-step microbial transformation by a mutant of Mycobacterium sp. J Ind Microbiol Biotechnol. 2002;28:280–3.View ArticleGoogle Scholar
  6. Angelova B, Mutafov S, Avramova T, Dimova I, Boyadjieva L. 9 alpha-hydroxylation of 4-androstene-3,17-dione by resting Rhodococcus sp. cells. Process Biochem. 1996;31:179–84.View ArticleGoogle Scholar
  7. Molchanova MA, Andryushina VA, Savinova TS, Stytsenko TS, Rodina NV, Voishvillo NE. Preparation of androsta-1,4-diene-3,17-dione from sterols using Mycobacterium neoaurum VKPM Ac-1656 strain. Russ J Bioorg Chem. 2007;33:354–8.View ArticleGoogle Scholar
  8. Mahato SB, Garai S. Advances in microbial steroid biotransformation. Steroids. 1997;62:332–45.View ArticleGoogle Scholar
  9. Liu WH, Kuo CW, Wu KL, Lee CY, Hsu WY. Transformation of cholesterol to testosterome by Mycobacterium sp. J Ind Microbiol. 1994;13:167–71.View ArticleGoogle Scholar
  10. Yao K, Xu LQ, Wang FQ, Wei DZ. Characterization and engineering of 3-ketosteroid-big up tri, open1-dehydrogenase and 3-ketosteroid-9alpha-hydroxylase in Mycobacterium neoaurum ATCC 25795 to produce 9alpha-hydroxy-4-androstene-3,17-dione through the catabolism of sterols. Metab Eng. 2014;24:181–91.View ArticleGoogle Scholar
  11. Imada YT. K: process for producing steroidal alcohols. US Patent 4,223,091. 1980.Google Scholar
  12. Xu L-Q, Liu Y-J, Yao K, Liu H-H, Tao X-Y, Wang F-Q, Wei D-Z. Unraveling and engineering the production of 23,24-bisnorcholenic steroids in sterol metabolism. Sci Rep. 2016;6:21928.View ArticleGoogle Scholar
  13. Galan B, Uhia I, Garcia-Fernandez E, Martinez I, Bahillo E, de la Fuente JL, Barredo JL, Fernandez-Cabezon L, Garcia JL. Mycobacterium smegmatis is a suitable cell factory for the production of steroidic synthons. Microb Biotechnol. 2017;10:138–50.View ArticleGoogle Scholar
  14. de las Fernandez Heras L, Mascaraque V, Garcia Fernandez E, Maria Navarro-Llorens J, Perera J, Drzyzga O. ChoG is the main inducible extracellular cholesterol oxidase of Rhodococcus sp. strain CECT3014. Microbiol Res. 2011;166:403–18.View ArticleGoogle Scholar
  15. Vrielink A. Cholesterol oxidase: structure and function. Sub-cell Biochem. 2010;51:137–58.View ArticleGoogle Scholar
  16. Dresen C, Lin LYC, D’Angelo I, Tocheva EI, Strynadka N, Eltis LD. A flavin-dependent monooxygenase from Mycobacterium tuberculosis involved in cholesterol catabolism. J Biol Chem. 2010;285:22264–75.View ArticleGoogle Scholar
  17. Yeh CH, Kuo YS, Chang CM, Liu WH, Sheu ML, Meng M. Deletion of the gene encoding the reductase component of 3-ketosteroid 9 alpha-hydroxylase in Rhodococcus equi USA-18 disrupts sterol catabolism, leading to the accumulation of 3-oxo-23,24-bisnorchola-1,4-dien-22-oic acid and 1,4-androstadiene-3,17-dione. Microb Cell Fact. 2014;13:10.View ArticleGoogle Scholar
  18. Wei W, Wang F-Q, Fan S-Y, Wei D-Z. Inactivation and augmentation of the primary 3-ketosteroid-delta(1)-dehydrogenase in Mycobacterium neoaurum NwIB-01: biotransformation of soybean phytosterols to 4-androstene-3,17-dione or 1,4-androstadiene-3,17-dione. Appl Environ Microbiol. 2010;76:4578–82.View ArticleGoogle Scholar
  19. Wang X, Feng J, Zhang D, Wu Q, Zhu D, Ma Y. Characterization of new recombinant 3-ketosteroid-delta1-dehydrogenases for the biotransformation of steroids. Appl Microbiol Biotechnol. 2017;101:6049–60.View ArticleGoogle Scholar
  20. Chen MM, Wang FQ, Lin LC, Yao K, Wei DZ. Characterization and application of fusidane antibiotic biosynthesis enzyme 3-ketosteroid-1-dehydrogenase in steroid transformation. Appl Microbiol Biotechnol. 2012;96:133–42.View ArticleGoogle Scholar
  21. Zhang W, Shao M, Rao Z, Xu M, Zhang X, Yang T, Li H, Xu Z. Bioconversion of 4-androstene-3,17-dione to androst-1,4-diene-3,17-dione by recombinant Bacillus subtilis expressing ksdd gene encoding 3-ketosteroid-Delta(1)-dehydrogenase from Mycobacterium neoaurum JC-12. J Steroid Biochem Mol Biol. 2013;135:36–42.View ArticleGoogle Scholar
  22. Zhang X, Wu D, Yang TW, Xu MJ, Rao ZM. Over-expression of Mycobacterium neoaurum 3-ketosteroid-delta 1-dehydrogenase in Corynebacterium crenatum for efficient bioconversion of 4-androstene-3,17-dione to androst-1,4-diene-3,17-dione. Electron J Biotechnol. 2016;24:84–90.View ArticleGoogle Scholar
  23. Shao M, Chen Y, Zhang X, Rao Z, Xu M, Yang T, Li H, Xu Z, Yang S. Enhanced intracellular soluble production of 3-ketosteroid-Δ1-dehydrogenase from Mycobacterium neoaurum in Escherichia coli and its application in the androst-1,4-diene-3,17-dione production. J Chem Technol Biotechnol. 2017;92:350–7.View ArticleGoogle Scholar
  24. Shao M, Sha Z, Zhang X, Rao Z, Xu M, Yang T, Xu Z, Yang S. Efficient androst-1,4-diene-3,17-dione production by co-expressing 3-ketosteroid-delta1 -dehydrogenase and catalase in Bacillus subtilis. J Appl Microbiol. 2017;122:119–28.View ArticleGoogle Scholar
  25. Bragin EY, Shtratnikova VY, Dovbnya DV, Schelkunov MI, Pekov YA, Malakho SG, Egorova OV, Ivashina TV, Sokolov SL, Ashapkin VV, Donova MV. Comparative analysis of genes encoding key steroid core oxidation enzymes in fast-growing Mycobacterium spp. strains. J Steroid Biochem Mol Biol. 2013;138:41–53.View ArticleGoogle Scholar
  26. Liu Y, Shen Y, Qiao Y, Su L, Li C, Wang M. The effect of 3-ketosteroid-delta(1)-dehydrogenase isoenzymes on the transformation of AD to 9alpha-OH-AD by Rhodococcus rhodochrous DSM43269. J Ind Microbiol Biotechnol. 2016;43:1303–11.View ArticleGoogle Scholar
  27. de las Fernandez Heras L, van der Geize R, Drzyzga O, Perera J, Navarro Llorens JM. Molecular characterization of three 3-ketosteroid-delta(1)-dehydrogenase isoenzynnes of Rhodococcus ruber strain Chol-4. J Steroid Biochem Mol Biol. 2012;132:271–81.View ArticleGoogle Scholar
  28. Guevara G, de Las Fernandez Heras L, Perera J, Navarro Llorens JM. Functional differentiation of 3-ketosteroid delta1-dehydrogenase isozymes in Rhodococcus ruber strain Chol-4. Microb Cell Fact. 2017;16:42.View ArticleGoogle Scholar
  29. Van der Geize R, Hessels G, Van Gerwen R, Vrijbloed J, van Der Meijden P, Dijkhuizen L. Targeted disruption of the kstD gene encoding a 3-Ketosteroid Δ1-dehydrogenase isoenzyme of Rhodococcus erythropolis strain SQ1. Appl Environ Microbiol. 2000;66:2029–36.View ArticleGoogle Scholar
  30. van der Geize R, Hessels GI, Dijkhuizen L. Molecular and functional characterization of the kstD2 gene of Rhodococcus erythropolis SQ1 encoding a second 3-ketosteroid delta(1)-dehydrogenase isoenzyme. Microbiology. 2002;148:3285–92.View ArticleGoogle Scholar
  31. Knol J, Bodewits K, Hessels GI, Dijkhuizen L, Van der Geize R. 3-Keto-5 alpha-steroid Delta’-dehydrogenase from Rhodococcus erythropolis SQ1 and its orthologue in Mycobacterium tuberculosis H37Rv are highly specific enzymes that function in cholesterol catabolism. Biochem J. 2008;410:339–46.View ArticleGoogle Scholar
  32. Kendall SL, Withers M, Soffair CN, Moreland NJ, Gurcha S, Sidders B, Frita R, ten Bokum A, Besra GS, Lott JS, Stoker NG. A highly conserved transcriptional repressor controls a large regulon involved in lipid degradation in Mycobacterium smegmatis and Mycobacterium tuberculosis. Mol Microbiol. 2007;65:684–99.View ArticleGoogle Scholar
  33. Kendall SL, Burgess P, Balhana R, Withers M, ten Bokum A, Lott JS, Gao C, Uhia-Castro I, Stoker NG. Cholesterol utilization in mycobacteria is controlled by two TetR-type transcriptional regulators: kstR and kstR2. Microbiology. 2010;156:1362–71.View ArticleGoogle Scholar
  34. Nguyen HD, Phan TT, Schumann W. Expression vectors for the rapid purification of recombinant proteins in Bacillus subtilis. Curr Microbiol. 2007;55:89–93.View ArticleGoogle Scholar
  35. Xue G-P, Johnson JS, Dalrymple BP. High osmolarity improves the electro-transformation efficiency of the gram-positive bacteria Bacillus subtilis and Bacillus licheniformis. J Microbiol Methods. 1999;34:183–91.View ArticleGoogle Scholar
  36. Cascioferro A, Boldrin F, Serafini A, Provvedi R, Palu G, Manganelli R. Xer site-specific recombination, an efficient tool to introduce unmarked deletions into mycobacteria. Appl Environ Microbiol. 2010;76:5312–6.View ArticleGoogle Scholar
  37. Parish T, Stoker NG. Use of a flexible cassette method to generate a double unmarked Mycobacterium tuberculosis tlyA plcABC mutant by gene replacement. Microbiology. 2000;146:1969–75.View ArticleGoogle Scholar
  38. van Kessel JC, Hatfull GF. Recombineering in Mycobacterium tuberculosis. Nat Methods. 2007;4:147–52.View ArticleGoogle Scholar
  39. Stover C, De La Cruz V, Fuerst T, Burlein J, Benson L, Bennett L, Bansal G, Young J, Lee M, Hatfull G, et al. New use of BCG for recombinant vaccines. Nature. 1991;351:456–60.View ArticleGoogle Scholar
  40. Andreu N, Zelmer A, Fletcher T, Elkington PT, Ward TH, Ripoll J, Parish T, Bancroft GJ, Schaible U, Robertson BD, Wiles S. Optimisation of bioluminescent reporters for use with mycobacteria. PLoS ONE. 2010;5:e10777.View ArticleGoogle Scholar
  41. Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem. 1976;72:248–54.View ArticleGoogle Scholar
  42. Laemmli UK. Cleavage of structural proteins during assembly of head of bacteriophage-T4. Nature. 1970;224:680–5.View ArticleGoogle Scholar
  43. Szentirmai A. Microbial physiology of side-chain degradation fo sterols. J Ind Microbiol. 1990;6:101–15.View ArticleGoogle Scholar
  44. Donova MV. Transformation of steroids by actinobacteria: a review. Appl Biochem Microbiol. 2007;43:1–14.View ArticleGoogle Scholar
  45. Rohman A, van Oosterwijk N, Thunnissen A-MWH, Dijkstra BW. Crystal structure and site-directed mutagenesis of 3-ketosteroid delta(1)-dehydrogenase from Rhodococcus erythropolis SQ1 explain its catalytic mechanism. J Biol Chem. 2013;288:35559–68.View ArticleGoogle Scholar
  46. Xie R, Shen Y, Qin N, Wang Y, Su L, Wang M. Genetic differences in ksdD influence on the ADD/AD ratio of Mycobacterium neoaurum. J Ind Microbiol Biotechnol. 2015;42:507–13.View ArticleGoogle Scholar
  47. Shao M, Zhang X, Rao Z, Xu M, Yang T, Li H, Xu Z, Yang S. A mutant form of 3-ketosteroid-delta(1)-dehydrogenase gives altered androst-1,4-diene-3, 17-dione/androst-4-ene-3,17-dione molar ratios in steroid biotransformations by Mycobacterium neoaurum ST-095. J Ind Microbiol Biotechnol. 2016;43:691–701.View ArticleGoogle Scholar


© The Author(s) 2018