Construction of the one-step conversion pathway from PS to 7β-OH-AD in M. neoaurum
The conversion of PS to AD by Mycolicibacteria is a basic way to supply the universal synthon for production of diverse steroids by further structure modification. To simplify the conventional way to produce 7β-OH-AD, here, an integrated microbial cell factory was developed to directly convert PS to 7β-OH-AD in one-step. Therefore, M3 derived from M. neoaurum ATCC 25795 by deleting all the characterized 3-ketosteroid-Δ1-dehydrogenase (kstD1, kstD2, and kstD3) and the 3-ketosteroid-9α-hydroxylase isoenzymes (kshA1 and kshA2), which can effectively transform PS to AD, was selected as a chassis to construct a 7β-OH-AD producing strain [30]. According to the reported mutant of P450-BM3 with 7β-hydroxylation activity to AD (mP450-BM3), the codon-optimized genes of P450-BM3 and mP450-BM3 were firstly introduced into strain M3 by an episomal plasmid pMV261 and verified by colony PCR and enzyme digestion techniques (Fig. S1), thus resulting in the recombinant strains M3-BM3 and M3-mBM3 (Table 1). Compared to M3-261 and M3-BM3, the strain M3-mBM3 showed a firm activity in the conversion of PS and AD and generated an extra product, which appeared in the position of standard 7β-OH-AD on the thin layer chromatography (TLC) plate (Fig. S2). High resolution mass spectrometry (HRMS) analysis showed that the molecular weight of the purified product was 302.1954m/z (Fig. S3A), which was the same as that of standard 7β-OH-AD [25, 26]. High performance liquid chromatography (HPLC) showed that the extra product was consistent with the standard 7β-OH-AD in the peak time (Fig. S3B). These results indicated that a 7β-OH-AD producing strain was successfully constructed. The 7β-hydroxylase activity in the M3-mBM3 was determined to be 1.42 ± 0.13 U g− 1, whereas no 7β-hydroxylase activity was observed in M3-261 and M3-BM3 (Table 2).
Table 1
Strains and plasmids used in this study.
Name
|
Description
|
Sources
|
Strains
|
|
|
Escherichia coli DH 5α
|
E. coli cloning host
|
Transgen Biotech
|
Mycolicibacterium neoaurum
|
Source of G6PDH and NADK genes
|
This lab
|
M3
|
Ksdd and KshA deletion mutant of MNR ATCC 25795
|
This lab
|
M3-261
|
M3 containing pMV261 as control
|
This study
|
M3-BM3
|
M3 expressing original P450-BM3 gene
|
This study
|
M3-mBM3
|
M3 expressing mP450-BM3 gene (F87A/T260G)
|
This study
|
M3-mBM3-0
|
M3 expressing mP450-BM3-0 gene (S72W/V78L/A82L
|
This study
|
|
/T88S/A328G/A330W)
|
|
M3-mBM3-0-NADK2
|
M3 expressing mP450-BM3-0 and NADK2 genes
|
This study
|
M3-mBM3-0-G6PDH2
|
M3 expressing mP450-BM3-0 and G6PDH2 genes
|
This study
|
M3-mBM3-0-NADK2-G6PDH2
|
M3 expressing mP450-BM3-0, NADK2 and G6PDH2 genes
|
This study
|
Plasmids
|
|
|
pUC57-BM3
|
The codon-optimized original P450-BM3 gene delivered
|
Shanghai Generay
|
|
by pUC57, AmpR
|
Biotech Co. Ltd
|
pUC57-mBM3
|
The codon-optimized mP450-BM3 gene delivered by
|
Shanghai Generay
|
|
pUC57, AmpR
|
Biotech Co. Ltd
|
pMV261
|
Shuttle vector of Mycobacterium and E. coli, Phsp60, KanR
|
Dr. W. R. Jacobs Jr. for
|
|
|
providing pMV261
|
pMV261-BM3
|
pMV261 containing original P450-BM3 gene, KanR
|
This study
|
pMV261-mBM3
|
pMV261 containing mP450-BM3 gene, KanR
|
This study
|
pMV261-mBM3-0
|
pMV261 containing mP450-BM3-0 gene, KanR
|
This study
|
pMV261-NADK
|
pMV261 containing NADK gene from M. neoaurum, KanR
|
This study
|
pMV261-G6PDH
|
pMV261 containing G6PDH gene from M. neoaurum, KanR
|
This study
|
pMV261-mBM3-0-NADK2
|
pMV261 containing mP450-BM3-0 gene and NADK2
|
This study
|
|
gene from M. neoaurum, KanR
|
|
pMV261-mBM3-0-G6PDH2
|
pMV261 containing mP450-BM3-0 gene and G6PDH2
|
This study
|
|
gene from M. neoaurum, KanR
|
|
pMV261-mBM3-0-NADK2-
|
pMV261 containing mP450-BM3-0 gene, NADK2 and
|
This study
|
G6PDH2
|
G6PDH2 genes from M. neoaurum, KanR
|
|
Notes: AmpR ampicillin-resistant, KanR kanamycin-resistant.
Table 2
The specific activities of 7β-hydroxylase, NADK and G6PDH in recombinant M. neoaurum strains.
|
Enzyme activity
|
Strains
|
7β-hydroxylase
|
NADK
|
G6PDH
|
(U g− 1)
|
(U g− 1)
|
(U g− 1)
|
M3-261
|
0
|
1.24 ± 0.06
|
1.37 ± 0.07
|
M3-BM3
|
0
|
1.17 ± 0.04
|
1.31 ± 0.12
|
M3-mBM3
|
1.42 ± 0.13
|
1.12 ± 0.07
|
1.25 ± 0.04
|
M3-mBM3-0
|
2.34 ± 0.11
|
1.08 ± 0.09
|
1.18 ± 0.05
|
M3-mBM3-0-NADK2
|
2.57 ± 0.08
|
2.56 ± 0.12
|
1.21 ± 0.15
|
M3-mBM3-0-G6PDH2
|
2.72 ± 0.12
|
1.24 ± 0.05
|
3.42 ± 0.13
|
M3-mBM3-0-NADK2-G6PDH2
|
3.03 ± 0.06
|
2.35 ± 0.15
|
3.34 ± 0.09
|
With PS and AD as substrates, M3-mBM3 generated 34.24 and 47.39 mg L− 1 7β-OH-AD, respectively. The conversion capacity of PS to AD was higher, but a large quantity of AD was not converted into 7β-OH-AD (Fig. S4). The results indicated that the expression activity of mP450-BM3 in M3-mBM3 was much lower than that in the E. coli [23]. Given that it was expressed by a high-copy plasmid pMV261 with a powerful promoter Phsp60 in M3-mBM3, the expression level of mP450-BM3 might not be a key factor limiting its activity. Apart from that, two factors might be closely related to the expression activity of mP450-BM3 in M3-mBM3. Firstly, the specific activity of mP450-BM3 in M3-mBM3 was not high enough and further evolution might boost its activity. Secondly, NADPH in M3-mBM3 was not enough for mP450-BM3 because the catabolic process of PS to AD generated NADH other than NADPH [28, 31].
Evolving the activity of 7β-hydroxylase by site-specific mutagenesis
It was considered a feasible method for promoting protein catalytic performance by semi-rational design and molecular remodeling. In order to elucidate the interaction of mP450-BM3 with substrate AD, a homology model was constructed by using NPG-P450-BM3 (PDB ID:4kpa) as the initial search model. The combined mutation of F87A and T260G in P450-BM3 (mP450-BM3) generated a new product, 7β-hydroxy-AD [22]. Therefore, this mutant was used as the starting template for further mutagenesis (Fig. 2A).
The following active sites residues (S72, V78, A82, T88, A328, A330) were subjected to saturation mutagenesis, and the results showed that S72W, V78L, A82M, A82L, T88S, A328G, A330W and A330P mutants resulted in higher activities than the wild type (mP450-BM3) (Fig. S5). Afterward, the variants were divided into two groups based on the distance from the heme. The amino acid (S72W, V78L, A82M and A82L) that were away from the heme were used to construct double and triple mutants. In these mutants, the resulting S72W/V78L/A82L achieved the higher reactivity (Fig. 2B). Similarly, the other double and triple mutants were constructed based the residues (T88S, A328G, A330W and A330P) that were close to the heme, and among which better 7β-hydroxylation activity to substrate AD was observed in the mutant T88S/A328G/A330W (Fig. 2C). Mutation of these residues altered the size and shape of the substrate binding pocket. For example, polar residue T88 located in the substrate-binding pocket was mutated to serine, which removed the steric hindrance of the methyl groups between F87A and T88, ensuring that AD could be oriented along the B’C loop. The mutation of A82L increased the space of active site and made a conformational change in the mP450-BM3, thus enabling the oxidation of omeprazole at the 7-methyl position. Finally, a combinatorial mutation between S72W/V78L/A82L and T88S/A328G/A330W was carried out. Fortunately, but not surprisingly, the highest activity was achieved in the mutant S72W/V78L/A82L/T88S/A328G/A330W (mP450-BM3-0, Fig. 2D). The 7β-hydroxylase activity in the mutant (mP450-BM3-0) reached to 2.34 ± 0.11 U g− 1 (Table 2), and the yield of 7β-OH-AD increased to 66.25 ± 2.42 mg L− 1 (Table 3). What’s more, the stereoselectivity was also improved to a certain extent (Fig. S3B).
Table 3
The durations and titer of PS conversion by different recombinant M. neoaurum strains.
Strains
|
Durations (d)
|
Titer (mg L− 1)
|
M3-261
|
—
|
0
|
M3-BM3
|
—
|
0
|
M3-mBM3
|
7
|
34.24 ± 1.34
|
M3-mBM3-0
|
7
|
66.25 ± 2.42
|
M3-mBM3-0-NADK2
|
6
|
94.63 ± 2.27
|
M3-mBM3-0-G6PDH2
|
6
|
117.46 ± 1.49
|
M3-mBM3-0-NADK2-G6PDH2
|
6
|
139.87 ± 3.73
|
Regeneration and balance of cofactors in the conversion of PS to 7β-OH-AD
As an NADPH-dependent enzyme, mP450-BM3-0 requires the participation of cofactor NADPH to convert AD into 7β-OH-AD. It was found that the content of NADPH in M3-mBM3-0 was much less than that of NADH on the fifth day (Fig. 3A), the ratio of NADH/NADPH and NADP+/NADPH sharply increased along with the conversion of PS to 7β-OH-AD (Fig. 3B and Fig. 3C). These data demonstrated that the lack of NADPH supply caused by the low content of NADPH and the frustrated regeneration of NADPH was responsible for the low activity of the mutants of P450-BM3 in the engineered 7β-OH-AD strains. To enhance the supply of NADPH, therefore, a combined strategy was adopted to increase the content of NADPH.
21 mole NADH would be generated during the conversion of one mole PS into AD, and reducing the ratio NADH/NAD+ has been demonstrated to boost the conversion of PS to AD [28]. Therefore, the transformation of NADH to NADPH may promote the conversion of PS to 7β-OH-AD via AD. The strain M3-mBM3-0-NADK2 was obtained by augmenting the gene NADK in strain M3-mBM3-0, which could convert partial NAD(H) into NADP(H). The NADK enzyme activity in M3-mBM3-0-NADK2 was more than two times that in M3-mBM3-0 (Table 2). The content of NADPH increased from 100.79 µM to 298.53 µM on the fifth day, whereas the content of NADH was significantly reduced in the conversion process of PS (Fig. 3A). In addition, the augmentation of NADK in M3-mBM3-0-NADK2 showed no ang effects on the ratio of NAD+/NADH in the conversion process of PS to 7β-OH-AD (Fig. 3C), but it profoundly increased the ratio of NADPH/NADP+ in the conversion process of PS, especially in the later stages (Fig. 3D). These data fully demonstrated that converting a small part of NAD(H) to NADP(H) was an effective approach to enhance the supply of NADPH in the conversion process of PS to 7β-OH-AD. The approach could increase both the content of NADP(H) and the ratio of NADPH/NADP+ without a negative effect on cell growth and the conversion of PS (Fig. 4A). As expected, therefore, the titer of 7β-OH-AD was significantly enhanced by 30.01% (Fig. S4B).
It was found that abundant NADP+ was generated during the conversion from PS to 7β-OH-AD, accompanied by rapid depletion of NADPH (Fig. S6), resulting in the ratio of NADP+/NADPH increased continuously, which indicated that regeneration rate of NADPH from NADP+ was slow in M3-mBM3-0 (Fig. 3D). Lee et al. had enhanced the production of ɛ-caprolactone by overexpressing an NADPH-regenerating glucose-6-phosphate dehydrogenase in recombinant Escherichia coli [32]. In order to promote the cyclic regeneration of NADPH from NADP+, a native NADP+ dependent G6PDH from M. neoaurum was expressed in strain M3-mBM3-0 to generate strain M3-mBM3-0-G6PDH2, which could increase G6PDH activity by 2.9-folds (Table 2). As expected, the ratio of NADP+/NADPH was significantly reduced in the conversion process of PS, displaying the augmentation effect on NADK (Fig. 3D), though the ratio of NAD+/NADH was unexpectedly significantly reduced due to the side effect of G6PDH on the conversion of NAD+ to NADH (Fig. 3C). In contrast to that of strain M3-mBM3-0, the 7β-OH-AD production titer of M3-mBM3-0-G6PDH2 was enhanced by 43.68% and only slight negative effects on cell growth and PS conversion were observed (Fig. 4A). Therefore, increasing the cyclic regeneration of NADPH by the expression of G6PDH could also boost the activity of 7β-hydroxylase for the conversion of AD to 7β-OH-AD.
The co-expression of NADK and G6PDH was subsequently preformed in strain M3-mBM3-0 to generate strain M3-mBM3-0-NADK2-G6PDH2, which showed a similar growth phenotype to strain M3-mBM3-0 (Fig. 4A). The combined expression of both NADK2 and G6PDH2 resulted in an increased content of NADPH (Fig. 3A), a reduced ratio of NAD+/NADH (Fig. 3C), and an enhanced ratio of NADPH/NADP+ (Fig. 3D).
Finally, due to the significant increase in NADPH supply, the strain accordingly achieved a further increase in the production of 7β-OH-AD, 139.87 mg L− 1, which was 52.7% higher than that in M3-mBM3-0.The above results confirmed that the lack of NADPH supply was a key limiting factor in the conversion of PS to 7β-OH-AD. Transforming part of abundant NADH generated in the PS conversion process to NADPH and promoting the cyclic regeneration of NADPH could boost the production of 7β-OH-AD without affecting the growth of engineered strains and the conversion performance of PS.
Optimization of conversion conditions of PS to 7β-OH-AD
The biotransformation of PS to steroidal synthons by Mycolicibacterium sp. is a co-metabolism process with glucose and PS as well as its main metabolites are hydrophobic. The uncommon traits make the conversion of PS to 7β-OH-AD be more complex in fermentation conditions. Therefore, some key factors affecting the conversion of PS to 7β-OH-AD were further investigated in order to enhance its application potential.
Mycolicibacterium species experience complex physiological and morphologic changes along with the consumption of nutrition ingredients and their generation period is generally longer than that of common microorganisms [33]. Thus, the growth status was closely related to the conversion capacity of PS. Nevertheless, it was difficult to precisely determine the growth curve of Mycolicibacteria as PS and its metabolites were dissolved in the fermentation process. Therefore, the initial inoculation dosage of the fermentation process was investigated. The result indicated that 9% inoculation dosage with pre-cultured cells (OD600 value of 3.0) in the logarithmic phase as seeds was the optimal dosage for production of 7β-OH-AD and the further increase in the inoculation dosage was no more conducive to the production of 7β-OH-AD (Fig. 5A).
The conversion process of steroids as well as PS in an aqueous fermentation system is usually recognized as a pseudo-crystalline fermentation process and the solubility or dispersion of PS is the key factor limiting its conversion efficiency. PS can be partially dissolved in some co-solvents, such as tween, ethanol, and cyclodextrin. However, excessive substrates are difficult to be dissolved and dispersed and have an inhibitory effect on strains, thus resulting in low cell viability and conversion. Therefore, it is crucial to select a suitable substrate concentration for the conversion of PS into 7β-OH-AD. The conversion efficiency decreased gradually with the increase of substrate concentration (Fig. 5B). The titer of 7β-OH-AD was only 73.28 mg L− 1 when the substrate concentration reached 7 g L− 1. The low substrate concentration may be due to the restriction of activity of mP450-BM3, which limited its use of the substrates PS. The viscosity of the solution increased with the increase of substrate concentration, thus affecting the transfer rate of oxygen [34]. Thus, some efforts such as DNA shuffling and oxygen improvement should be made to increase the activity of mP450-BM3 for further improving conversion efficiency.
Glucose can not only be metabolized with sterols to support the growth of strains, but also can used as a co-substrate for the regeneration of NADPH. Therefore, it is necessary to explore the concentration and feeding time of glucose. The consumption rate of glucose increased significantly on the third day, whereas the recombinant strains grew rapidly (Fig. 4). Therefore, it was proper to add glucose after the third day to ensure the optimal conditions for the growth and metabolism of recombinant strains. The highest conversion efficiency was reached when glucose was added on the fourth day (Fig. 5C). Most of glucose was consumed after the fourth day, thus leading to the declined efficiency in the later fermentation stage.
A suitable concentration of glucose is also a crucial factor for the conversion. The conversion was improved by the increased glucose concentration, and the highest conversion efficiency was reached when the supplement concentration of glucose was 30 g L− 1 (Fig. 5D). When the concentration of glucose was further increased, the conversion efficiency was no longer increased. The state of fermentation broth would be changed by excessive glucose, thus adversely affecting the accumulation of fermentation products.