Improving the biotransformation efficiency of soybean phytosterols in Mycolicibacterium neoaurum by the combined deletion of fbpC3 and embC in cell envelope synthesis.

Biotransformation of soybean phytosterols into 9α-hydroxy-4-androstene-3,17-dione (9-OHAD) by mycobacteria is the core step in the synthesis of adrenocortical hormone. However, the low permeability of the dense cell envelope largely inhibits the overall conversion efficiency of phytosterols. The antigen 85 (Ag85) complex encoded by fbpA, fbpB, and fbpC was proposed as the key factor in the combined catalysis of mycoloyl for producing mycolyl-arabinogalactan (m-AG) and trehalose dimycolate (TDM) in mycobacterial cell envelope. Herein, we confirmed that fbpC3 was essential for the biotransformation of trehalose monomycolate (TMM) to TDM in Mycolicibacterium neoaurum. The deficiency of this gene raised the cell permeability, thereby enhancing the steroid uptake and utilization. The 9-OHAD yield in the fbpC3-deficient 9-OHAD-producing strain was increased by 21.3%. Moreover, the combined deletion of fbpC3 and embC further increased the 9-OHAD yield compared to the single deletion of fbpC3. Finally, after 96 h of bioconversion in industrial resting cells, the 9-OHAD yield of 11.2 g/L was achieved from 20 g/L phytosterols and the productivity reached 0.116 g/L/h. In summary, this study suggested the critical role of the fbpC3 gene in the synthesis of TDM in M. neoaurum and verified the feasibility of improving the bioconversion efficiency of phytosterols through the cell envelope engineering strategy.

Introduction

Biotransformation of low value-added phytosterols mainly extracted from vegetable oil processing waste or pine tree biomass to important steroidal intermediates is the key step in the semi-synthetic route of current steroidal pharmaceutical industry [1]. Actinobacteria, especially mycobacteria, play an essential role in the above bioconversion process [2,3]. From steroidal intermediates, almost all kinds of steroidal drugs can be produced by chemical modifications [4]. The inactivation of 3-ketosteroid-Δ1-dehydrogenases (KstDs), which are key enzymes for initiating the degradation of steroidal nucleus, can block the sterol metabolic pathway of Mycolicibacterium neoaurum. Consequently, the obtained engineered strain could convert sterol substrates into an important steroidal intermediate, 9-OHAD [2]. However, due to its insufficient yield, the engineered strain was further improved to achieve a satisfying biotransformation efficiency of soybean phytosterols.

Sterols can be utilized as the sole carbon source and energy source for the basic physiological and metabolic activities of mycobacteria [5,6]. The uptake of sterols by cells can be divided in two phases: the mass transfer of sterol particles to cell surface and the transport of sterol molecules through cell envelope [7,8]. Hence, cell envelope plays an essential role in the uptake of sterols. The outermost covering of cell envelope is typically composed of a variety of non-covalent binding capsular lipids, such as trehalose monomycolate (TMM), trehalose dimycolate (TDM), phenolic glycolipids, sulfolipids, and phospholipids [9,10]. A covalent binding layer of dense mycolyl-arabinogalactan-peptidoglycan (m-AG-PG) can be found on the inner side of the non-covalent layer (Fig. 1) [10,11]. This unique cell envelope leads to a low permeability and forms an effective barrier, which can prevent toxic compounds from entering mycobacterial cells [12,13]. However, the feature negatively affects the uptake of sterol substrates at the bioconversion to steroidal intermediates [8,14]. The deletion of kasB encoding a β-ketoacyl-acyl carrier protein synthase likely shortened the length of keto-mycolic acids, the key components of mycolic acid methyl esters (MAMEs) in the core m-AG-PG structure [8]. Moreover, the inactivation of arabinofuranosyltransferase gene embC caused a synthesis defect of the lipoarabinomannan from lipomannan [14,15]. The above two changes remarkably improved the cell permeability and enhanced the uptake of steroids. As a result, the accumulation efficiency of the target steroidal intermediate was further increased, whereas the inactivation of other dispensable genes (hadA, hadC, mmaA1, mmaA3, mmaA4 or pks13) in the mycolic acid synthesis pathway did not show positive effects on steroid utilization [8].

Fig. 1

Schematic diagram of the uptake and bioconversion of phytosterols by mycobacteria. The outer layer of cells is mainly composed of the non-covalent binding components, including TMM, TDM, glycolipid, phospholipid, etc. The core of the cell envelope is the dense covalent mycolyl-arabinogalactan-peptidoglycan layer. The uptake of sterols mainly relies on the direct contact between phytosterol substrates and cell surface. The steroidal drug intermediate (9-OHAD) can be produced through the incomplete metabolism of phytosterols in the engineered M. neoaurum strain. PL, phospholipid; GL, glycolipid; TMM, trehalose monomycolate; TDM, trehalose dimycolate; MA, mycolic acid methyl esters; AG, arabinogalactan; PG, peptidoglycan.

TMM is the final product of mycolic acid synthesis pathway in the mycobacterial cytoplasm [16]. Synthesized TMM is then transported into periplasmic space for the mycoloyl residue transfer onto arabinogalactan to produce the m-AG, or used in the formation of the outer membrane cord factor TDM, the most abundant surface-exposed glycolipid of the mycobacterial cell envelope [10,[16], [17], [18], [19]]. The complex of Ag85 composed of three secreted transesterases (Ag85A, Ag85B and Ag85C respectively encoded by fbpA, fbpB, and fbpC) was identified as the catalyzing enzyme for the assembly of the two lipids [20,21]. Up to now, at least two trehalose binding sites (active site and secondary site) in Ag85 enzymes were identified [20]. TMM firstly binds to secondary site for inducing a conformational change of the side chain at Phe232 in Ag85A/Ag85B and Leu230 in Ag85C after being transported through the membrane to periplasm [21]. Then, the molecule can enter active site to release a free trehalose, whereas the residual α chain of ester-linked mycolic acid is buried in a hydrophobic hole [21]. Finally, mycolic acid is selectively transferred onto 6′-hydroxy of TMM to form TDM or onto the 5-hydroxy of the terminal arabinose to yield m-AG [16]. Actually, a catalytic triad composed of a nucleophilic serine (Ag85C: Ser124), histidine base (Ag85C: His260), and glutamic acid (Ag85C: Glu228) was employed to catalyze a ping–pong reaction mechanism for the mycolyl transfer [20,22,23]. A new potent inhibitor ebselen of M. tuberculosis Ag85 complex was identified to bind covalently to a Cys209 residue located near the Ag85C active site, thereby leading to a remarkably decreased enzymatic activity of Cys209 mutation [24,25]. Besides, nuclear magnetic resonance analysis revealed that another novel inhibitor I3-AG85 could bind Ag85C and thus block the biosynthesis of cord factor TDM in M. tuberculosis [17]. These studies suggested that the inactivation of fbpC suppressed the mycoloyl residue transfer in cell envelope synthesis of engineered M. neoaurum and might be a robust and efficient strategy to raise cell permeability. Since the fbp genes are likely synthetically lethal, the double deletion of them has not been reported [26]. It is necessary to evaluate the effects of the serial deletion of identified target genes involved in cell envelope synthesis of M. neoaurum [8,14]. As a consequence of these combined manipulations, the biotransformation efficiency of soybean phytosterols by the mutant strain might be boosted up to a new level.

In this study, the relationship between the located fbpC3 and the synthesis of mycolic acids-related components in M. neoaurum was firstly evaluated. The changes in cell permeability caused by the disordered assembly of cell envelope and the chain effect on steroid uptake and utilization were then determined. Then, the biotransformation efficiency of soybean phytosterols in the fbpC3-deficient 9-OHAD-producing M. neoaurum strain was measured. Finally, the combined manipulation of the previously screened gene targets that influenced the accumulation of steroidal intermediates [8,14] and the newly identified fbpC3 was performed to further increase 9-OHAD yield.

Materials and methods

Strains, plasmids, primers, media, and culture conditions

The strains used in this study are listed in Table 1. M. neoaurum ATCC 25795 (Mn) and E. coli DH5α were respectively purchased from American Type Culture Collection (ATCC) and TIANGEN Biotech. Co., Ltd. (Shanghai, China). The steroid drug intermediate 9-OHAD producer MnΔkstD1ΔkstD2ΔkstD3 (WI) was constructed by Kang Yao [2]. Plasmids and primers used for constructing the mutants are shown in Supplementary Table S1 and Table S2. Recombinant plasmids were transformed into E. coli DH5α by the heat shock method or into M. neoaurum by electroporation as described previously [27,28].

Table 1

Strains used in this study.

Names	Descriptions	Sources
E. coli DH5α	E. coli strain for cloning	TIANGEN Co., Ltd.
M. neoaurum ATCC 25795 (Mn)	Wild type strain, the starting strain	ATCC
MnΔfbpC3	fbpC3 deleted in M. neoaurum ATCC 25795	This study
MnΔfbpC3+fbpC3	fbpC3 complemented in MnΔfbpC3 strain	This study
WI	kstD1, kstD2 and kstD3 deleted in M. neoaurum ATCC 25795, 9-OHAD-producing strain	Yao et al., 2014 [2]
WIΔfbpC3	fbpC3 deleted in WI strain	This study
WIΔfbpC3ΔkasB	fbpC3 and kasB deleted in WI strain	This study
WIΔfbpC3ΔembC	fbpC3 and embC deleted in WI strain	This study

The media and culture conditions in previous descriptions were adopted in the study [28]. E. coli was cultured in Luria–Bertani (LB) medium at 37 °C and kanamycin (50 mg/L) or hygromycin (100 mg/L) was added in the corresponding media if necessary. M. neoaurum was firstly cultivated in LB medium and then inoculated in MYC/01 medium (20.0 g/L glycerol, 2.0 g/L citric acid, 2.0 g/L NH4NO3, 0.5 g/L K2HPO4, 0.5 g/L MgSO4·7H2O, and 0.05 g/L ammonium ferric citrate (pH 7.5)) to obtain seed cells. For the purpose of determining growth phenotypes, the cells were inoculated into minimal medium (MM) (2.0 g/L NH4NO3, 0.5 g/L K2HPO4, 0.5 g/L MgSO4·7H2O, and 0.05 g/L ammonium ferric citrate) containing 1 g/L cholesterol (purity >95.0%, Adamas Reagent, Co., Ltd., Shanghai, China). To assess the bioconversion efficiency, the cells were transferred into MYC/02 medium (10.0 g/L glucose, 2.0 g/L citric acid, 2.0 g/L NH4NO3, 0.5 g/L MgSO4·7H2O, and 0.05 g/L ferric ammonium citrate (pH 7.5)) containing 5 g/L phytosterols. In order to determine the biotransformation efficiency of resting cells, the cells were inoculated into MYC/02 medium for 48 h–72 h. Then, the cultivated cells were harvested, washed with 20 mM KH2PO4, and resuspended to obtain 200 g/L suspensions. The transformation was performed with 100 g/L cells, 20 g/L phytosterols, and 80 g/L hydroxypropyl-β-cyclodextrin. The transformation experiment was carried out in shake flasks (30 °C and 200 rpm).

Gene manipulation in M. neoaurum

Target gene-deleted mutants were acquired through the allelic homologous recombination as previously described [28]. The plasmids p2NIL and pGOAL19 were employed to construct the homologous recombination plasmids. Briefly, the upstream and downstream sequences of a target gene were cloned, digested with restriction enzymes and ligated into p2NIL. A selection marker cassette from pGOAL19 was dissected and inserted into the recombinant p2NIL to construct the knockout-plasmids (Supplementary Table S1), which could be electroporated into M. neoaurum to obtain the corresponding deficient strains.

The coding frame of a wild-type target gene was firstly cloned, digested and then inserted into pMV261 and the constructed p261-gene was used to overexpress the function of the corresponding deficient gene. In addition, the expression cassette could be digested from the p261-gene and integrated into pMV306 to obtain a complementary p306-gene. The constructed plasmid could be then integrated into the mycobacterial chromosome DNA in a single copy.

Analysis of mycolic acid methyl esters

Mycobacterial cell MAMEs were extracted and determined as previously described [8]. Cultivated cells (50 mg, wet weight) were firstly collected at 12,000×g for 10 min. Then, after adding 0.5 mL of methanol and chloroform (2:1, v/v), collected cells were incubated at 60 °C for 2 h. Polar lipids were dissolved in the supernatant after the extraction solution was centrifuged at 12,000×g for 10 min.

MAMEs could be acquired from defatted cells by adding 500 μL of 10% tetrabutylammonium hydroxide (Sigma-Aldrich LLC., MO, USA) at 100 °C overnight. After cooling, 500 μL of ddH2O, 250 μL of dichloromethane, and 62.5 μL of iodomethane (Sigma-Aldrich LLC., MO, USA) were added, then stirred for 30 min and centrifuged at 12,000×g for 10 min. The lower organic layer was washed respectively with 1.0 mL of hydrochloric acid (1 M) and 1.0 mL of ddH2O. Crude MAMEs would be obtained after the organic layer was dried under a nitrogen stream. In order to further purify MAMEs, the resulting residue was dissolved in a mixture of toluene (0.2 mL) and acetonitrile (0.1 mL), followed by the addition of 0.2 mL of acetonitrile and incubation for 1 h at 4 °C. The mixture solution was centrifuged at 12,000×g for 10 min and then the obtained MAMEs were re-suspended in 200 μL of dichlormethane for further analysis.

Mycolic acids were analyzed by silica gel TLC plates. The mean grayscale intensity was analyzed with Quantity One (Version 4.6.6, Bio-Rad Laboratories, CA, USA). Relative abundances of polar mycolic acids (TMM and TDM) and nonpolar MAMEs were respectively calculated.

Analysis of cell permeability

After the manipulation of target genes, cell permeability was assessed according to the previous procedure [29]. Cultivated cells with the same wet weight were suspended in 4.5 mL of phosphate buffer (106 cells/mL), mixed with 0.5 mL of fluorescein diacetate (FDA) acetone solution (2 mg/mL) and vibrated at 32 °C for 10 min. Then, the fluorescence value of the obtained suspension was detected at the maximum excitation wavelength of 485 nm and the emission wavelength of 538 nm.

The quantity of cholest-4-en-3-one entering cells per unit time was determined as described [30]. The cells were cultured in MYC/02 medium containing 1.0 g/L cholest-4-en-3-one (purity >95.0%, Shanghai TITAN Scientific Co., Ltd., China) for 12 h and then 5 mL of the culture solution was harvested, centrifuged at 12,000×g for 10 min and washed with ddH2O for two times, followed by the addition of 1.0 mL of the mixture of petroleum ether and ethyl acetate (6:4, v/v). Next, the cells were suspended in the mixture of acetonitrile and ddH2O (7:3, v/v) (50 mg wet weight/mL). After adding 0.8 g of glass beads, the cells were destroyed with FastPrep-24 instrument (MP Biomedicals, CA, USA) and centrifuged at 12,000×g for 10 min. Then, steroids entering cells could be released and determined with a reversed-phase C18-column (250 mm × 4.6 mm) at 254 nm in HPLC system. The mixture of methanol and water (8:2, v/v) was used as the mobile phase.

Analysis of sterol biotransformation

The change in the transformation efficiency of constructed strains was assessed in the vegetative cell and resting cell conversion systems [8,14]. Briefly, the cells in the suspension culture (0.5 mL) were extracted with the same volume of ethyl acetate (0.5 mL). The resting cell conversion suspension (0.1 mL) was extracted with 1.0 mL of ethyl acetate. After centrifugation at 12000g for 20 min, the upper organic phase containing steroids was analyzed with high-performance liquid chromatography (HPLC) or gas chromatography (GC).

A 1100 series HPLC system (Agilent Technologies, CA, USA) was used to analyze the target product 9-OHAD. The samples were determined with a reversed-phase C18-column (250 mm × 4.6 mm) (Agilent Technologies, CA, USA) at a wavelength of 254 nm. The mobile phase was the mixture of methanol and water (8:2, v/v). A GC system 7820A (Agilent Technologies, CA, USA) was used to determine sterol substrates. The samples were detected with a DB-5 column (30 m × 0.25 mm (i.d.) × 0.25 μm film thickness, Agilent Technologies, CA, USA). The oven temperature was controlled as follows: 200 °C for 2 min, 200 °C–280 °C within 4 min, 280 °C for 2 min, 280 °C–305 °C within 1.5 min, and 305 °C for 10 min. Inlet temperature and flame-ionization detector temperature were maintained at 320 °C. The flow rate of nitrogen carrier gas was 2 mL/min at 50 °C.

Results and discussion

FbpC3 is the key factor for the assembly of TDM in M. neoaurum

To further improve the bioconversion efficiency, the fbpC genes involved in the assembly of TMM after the transportation from cytoplasm to periplasm was explored. First of all, a gene fbpC3 (GenBank Accession No. NZ_JMDW01000024.1; Region: 26,764 … 27,720, 957-bp) annotated for encoding the Ag85C protein was located in the genome of M. neoaurum ATCC 25795 (GenBank Accession No. NZ_JMDW00000000.1). Then, the gene was tentatively deleted from the wild-type strain for the subsequent phenotype identification. Briefly, a 1065-bp upstream sequence and a 1053-bp downstream sequence of the fbpC3 coding sequence were cloned into the knockout plasmid for the allelic homologous recombination of the wild-type gene (Fig. 2A and B). Therefore, the gene fbpC3 was theoretically mutated and the MnΔfbpC3 strain was constructed accordingly. Interestingly, the deletion of fbpC3 also accelerated cell growth in the presence of cholesterol (Fig. 2C), similar to the effect of the deletion of kasB and embC in the strain [8,14]. This phenotypic change might be ascribed to the changed FbpC3-responsible cell envelope assembly.

Fig. 2

Inactivation of fbpC3 caused the assembly deficiency of TDM in M. neoaurum. (A) Schematic diagram of the deletion of fbpC3 gene from the genome of M. neoaurum ATCC 25795. A 1065-bp upstream sequence and a 1053-bp downstream sequence were designed to replace the wild-type gene. (B) Validation results of allelic replacement at the located fbpC3 in the strain. (C) Growth characteristics of the wide-type strain Mn, the fbpC3-deficient strain (MnΔfbpC3) and the fbpC3-complemented strain (MnΔfbpC3+fbpC3) in the MM containing 1.0 g/L cholesterol. (D) Content analysis of polar lipids (TMM and TDM) in the mycobacterial cell envelope. (E) Calculated percentage of the relative gray intensity of polar lipids. Deletion of fbpC3 caused the synthesis deficiency of TDM in M. neoaurum. The relative abundances of TMM and TDM were 96.0% and 4.0% in the fbpC3-deleted strain, whereas the percentages of the two components in the Mn strain was 86.2% and 13.8%, respectively.

Next, the mycolic acids-related components in the mutant strain were extracted and analyzed. The polar lipids displayed significant differences in the MnΔfbpC3 strain (Fig. 2D). The synthesis of TDM was thoroughly blocked, but the TMM component was not significantly changed. The relative abundances of polar TMM and TDM in the MnΔfbpC3 strain were 96.0% and 4.0%, whereas the percentages of the two components in the wild-type Mn strain were respectively 86.2% and 13.8% (Fig. 2E). In a word, the TDM content was decreased by 245% approximately after the deletion of fbpC3 and this phenotype was restored in the fpbC3-complemented strain, indicating that the located gene fbpC3 empowered the strain to synthesize TDM. Additionally, MAMEs showed no obvious difference between the MnΔfbpC3 strain and the parental strain (Supplementary Fig. S1), suggesting that fbpC3 was possibly irrelevant to the MAMEs synthesis.

Deleting fbpC3 raised cell permeability of M. neoaurum

After the deletion of fbpC3 in M. neoaurum, the MnΔfbpC3 strain showed a determined defect in TDM assembly compared to that of its parental strain. It was speculated that this change might improve cell permeability and enhance the utilization of sterol substrates. To confirm this speculation, cell permeability of fbpC3-deficient strain was evaluated through the fluorescein diacetate (FDA) assay [8]. The mutant strain displayed an obvious improvement in fluorescence intensity (Fig. 3A). When incubation time increased from 10 to 30 min, the improvement in fluorescence intensity of MnΔfbpC3 strain compared to that of the Mn strain increased from 15.0% to 23.3%, indicating that the penetrated FDA in the fbpC3-deficient cells was raised significantly. In other words, the permeability of the mutant strain might be improved by the deletion of fbpC3.

Fig. 3

Knockout of fbpC3 increased cell permeability of M. neoaurum. (A) Determination of cell permeability in the fbpC3-deficient strain. Mycobacterial cells were stained with FDA and analyzed with a fluorescence spectrophotometer. (B) Effects of the deletion of fbpC3 on the uptake of steroid cholest-4-en-3-one. Cells were cultivated in MM containing 1.0 g/L cholest-4-en-3-one. Data represent mean ± standard deviation of three measurements.

Based on the above results, the analog of cholesterol, cholest-4-en-3-one, was then used as a label to evaluate the cell permeability to steroids [30]. The determination of the steroids entering cells revealed that the enhanced cell permeability led to about an increase of 55% in the uptake of cholest-4-en-3-one by the MnΔfbpC3 strain compared to that of the Mn strain (Fig. 3B). Then, the utilization efficiency of the sterol substrate in the fbpC3-deleted strain was further analyzed. The utilization efficiency of cholesterol in the MnΔfbpC3 strain was raised indeed (Supplementary Fig. S2). Under the same cultivation conditions, the residual cholesterol content in the culture medium for the fbpC3-mutant strain was respectively 1.52 g/L at 48 h and 0.43 g/L at 96 h, whereas that of Mn strain was 1.72 g/L at 48 h and 0.66 g/L at 96 h. The phenotypic change might be interpreted as a chain effect caused by the improved cell permeability. In other words, it was confirmed that the improved cell permeability was associated with the deletion of fbpC3. Therefore, the uptake and utilization of steroids was accelerated.

Increasing the accumulation efficiency of 9-OHAD by deleting fbpC3

To assess the effect of the improved cell permeability caused by the decreased content of polar TDM on the bioconversion of target steroidal intermediates, a new WIΔfbpC3 strain was constructed based on the engineered 9-OHAD-producing strain WI [2]. In the biotransformation conditions, the vegetative cells of WIΔfbpC3 strain yielded 0.31 g/L 9-OHAD after 24 h of conversion, whereas its parental strain WI yielded 0.18 g/L 9-OHAD after 24 h of conversion. The accelerated accumulation of 9-OHAD was observed at all the subsequent three sampling time points. Ultimately, the deletion of fbpC3 increased the 9-OHAD yield from 1.31 g/L to 1.59 g/L after 96 h of conversion. The transformation efficiency of WIΔfbpC3 strain was at least 21.3% higher than that of the WI strain in the above biotransformation conditions (Fig. 4A).

Fig. 4

Assessment of the influences of the deletion of multiple genes involved in the cell wall synthesis on the bioconversion of phytosterols to 9-OHAD. (A) Preliminary determination of 9-OHAD yield in the vegetative cell transformation system containing 5 g/L phytosterols. (B) Assessment results of the constructed 9-OHAD-producing strain with a resting cell conversion system containing 20 g/L phytosterols. Data represent mean ± standard deviation of three measurements.

Subsequently, the resting cell transformation system was employed to measure the conversion efficiency of strains per unit wet weight. After 72 h of conversion, the highest increase was observed in WIΔfbpC3 strain and the yield reached 8.9 g/L (Fig. 4B), which was 35.1% higher than that of the parental strain. After 120 h of bioconversion, the WIΔfbpC3 strain produced 10.7 g/L 9-OHAD and the molar yield reached 68.2%, which was 20.3% higher than that of its parental strain WI. The above data showed that the deletion of fbpC3 in the engineered 9-OHAD-producing strain indeed enhanced the bioconversion of soybean phytosterols to the target steroidal intermediate 9-OHAD.

Effects of the combined deletion of fbpC3 and other modified sites on the 9-OHAD yield

In previous studies [8,14], the independent inactivation of kasB or embC raised the productivity of 9-OHAD-producing strain. The combined deletion of screened target genes might have an additive enhancement effect on the yield of target intermediate. Then, we deleted the gene kasB of strain WIΔfbpC3 to obtain the WIΔfbpC3ΔkasB strain. Unexpectedly, the yield of target product in the obtained strain WIΔfbpC3ΔkasB did not show the additive enhancement effect compared to the WIΔfbpC3 strain. The 9-OHAD yield in the WIΔfbpC3ΔkasB strain declined to about 0.79 g/L after 96 h of conversion under the same cultivation conditions (Fig. 4A). Therefore, the combined deletion of two genes (fbpC3 and kasB) involved in the synthesis and assembly of mycolic acids-related components might not be an efficient method for further enhancing the bioconversion efficiency.

The gene embC is responsible for the assembly of lipoarabinomannan, which is one of the non-covalent lipoglycan component in the cell envelope [14]. Hence, embC was tentatively deleted from the genome of WIΔfbpC3 strain to generate the strain WIΔfbpC3ΔembC. Interestingly, this mutant strain displayed some additive effect on the biotransformation of soybean phytosterols to target product 9-OHAD in the shake flask experiments (Fig. 4A). The highest improvement was determined in the newly constructed WIΔfbpC3ΔembC strain after 72 h of conversion and the 9-OHAD yield reached 1.25 g/L, which was about 105% higher than that of the WI strain. The serial deletion of fbpC3 and embC finally increased the 9-OHAD to 1.68 g/L after 96 h of conversion. In addition, during the subsequent assessment with the resting cell bioconversion system, after 96 h of conversion, the 9-OHAD yield in the WIΔfbpC3ΔembC strain increased to 11.2 g/L with the productivity of 0.116 g/L/h and the molar yield was 71.4%, which was higher than that the strain WIΔkasB (69.5%) [8] and the strain WIΔfbpC3 (68.2%) (Fig. 4B). In a word, after 96 h of biotransformation, the 9-OHAD yield in the WIΔfbpC3ΔembC strain was improved by 26% compared to that in the WI strain. When transformation time was increased to 144 h, the 9-OHAD yield of the WIΔfbpC3ΔembC strain reached about 11.9 g/L, which was the highest in all the reported data of existing 9-OHAD-producing M. neoaurum strains [8,14,28,31].

Actually, a total of four fbpC genes were annotated in the genome of M. neoaurum (Supplementary Table S3). However, the inactivation of fbpC1, fbpC2, or fbpC4 in the M. neoaurum did not show any observable phenotype change (data not shown). Only the fbpC3-deficient strain displayed an expected enhancement on the 9-OHAD production, indicating that fbpC3 possibly played an essential role in the synthesis of 9-OHAD. The serial inactivation of fbpC3 and kasB theoretically caused defects on the synthesis of the polar TDM as well as the nonpolar mero-mycolic acids in the major frame of the core m-AG-PG structure in the M. neoaurum strain [8]. Consequently, the growth of the WIΔfbpC3ΔkasB strain was extremely suppressed in the transformation medium in the presence of soybean phytosterols (data not shown). In contrast, the deletion of fbpC3 and embC resulted in the deficient mutant on the assembly of TDM and lipoarabinomannan [14]. The combination of the two defects in the cell envelope synthesis displayed different superposition effects on the transformation production of the engineered strain. Thus, the excessive modifications on the synthesis and assembly pathway of similar components in M. neoaurum might be not a reasonable strategy for improving the bioconversion of soybean phytosterols. The intracellular metabolism rate is another critical factor influencing the overall transformation efficiency in the engineered strain, suggesting that the combined enhancement of the mass transfer, the transport of sterols into cells as well as the intracellular catabolism of sterols might be a promising breakthrough for evolving the transforming strain.

Conclusions

In summary, this study confirmed that fbpC3 was a key gene responsible for the assembly of polar TDM in the cell envelope of M. neoaurum. The inactivation of this gene was proved to be an effective way to increase the production of the steroidal intermediate 9-OHAD. However, interestingly, the joint deletion of fbpC3 and kasB did not display the desired promotion effect on 9-OHAD accumulation, whereas the combined deletion of fbpC3 and embC achieved an expected additive enhancement effect on the production of 9-OHAD. Ultimately, in the WIΔfbpC3ΔembC strain under the resting cell transformation system containing 20 g/L of phytosterols, the 9-OHAD yield increased to 11.9 g/L and the molar yield reached 75.8%.

Funding

This work was supported by the (Nos. 21776075 and 32100067), the (No. 20ZR1415100), the (No. SQ2020YFC210061), the (No. 2020M671028), the Shanghai Municipal Health Commission (No. 20204Y0380), the Teacher's Professional Development Project of Shanghai Municipal Education Commission, and the Scientific Research Foundation of SUMHS.

CRediT authorship contribution statement

Liang-Bin Xiong: Methodology, Formal analysis, Investigation, Writing – original draft, Funding acquisition. Hao-Hao Liu: Methodology, Validation, Formal analysis, Investigation, Writing – original draft. Lu Song: Methodology, Validation, Formal analysis. Miao-Miao Dong: Investigation. Jie Ke: Investigation. Yong-Jun Liu: Investigation. Ke Liu: Investigation. Ming Zhao: Writing – review & editing, Funding acquisition. Feng-Qing Wang: Conceptualization, Methodology, Formal analysis, Investigation, Supervision, Writing – review & editing, Funding acquisition. Dong-Zhi Wei: Supervision, Funding acquisition.

Declaration of competing interest

The authors declared that they have no conflicts of interest to this work. The manuscript is approved by all authors for publication. I would like to declare on behalf of my co-authors that the work described was original research that has not been published previously, and not under consideration for publication elsewhere, in whole or in part. All the authors listed have approved the manuscript that is enclosed.