Distinct Regioselectivity of Fungal P450 Enzymes for Steroidal Hydroxylation

The steroidal hydroxylases CYP5150AP3 and CYP5150AN1 together with the previously characterized CYP5150AP2 belong to the CYP5150A family of P450 enzymes with high amino acid sequence identity, but they showed completely different regioselectivities toward 11-deoxycortisol, suggesting the regioselectivity diversity of steroidal hydroxylases of CYP5150 family. They are also distinct from the known bacterial and fungal steroidal hydroxylases in substrate specificity and regioselectivity. Biocatalytic hydroxylation is one of the important transformations for the functionalization of steroid nucleus rings but remains a very challenging task in organic synthesis. These hydroxylases are useful additions to the toolbox of hydroxylase enzymes for the functionalization of steroids at various positions.

ABSTRACT

In this study, we identified two P450 enzymes (CYP5150AP3 and CYP5150AN1) from Thanatephorus cucumeris NBRC 6298 by combination of transcriptome sequencing and heterologous expression in Pichia pastoris. The biotransformation of 11-deoxycortisol and testosterone by Pichia pastoris whole cells coexpressing the cyp5150ap3 and por genes demonstrated that the CYP5150AP3 enzyme possessed steroidal 7β-hydroxylase activities toward these substrates, and the regioselectivity was dependent on the structures of steroidal compounds. CYP5150AN1 catalyzed the 2β-hydroxylation of 11-deoxycortisol. It is interesting that they display different regioselectivity of hydroxylation from that of their isoenzyme, CYP5150AP2, which possesses 19- and 11β-hydroxylase activities.

IMPORTANCE The steroidal hydroxylases CYP5150AP3 and CYP5150AN1 together with the previously characterized CYP5150AP2 belong to the CYP5150A family of P450 enzymes with high amino acid sequence identity, but they showed completely different regioselectivities toward 11-deoxycortisol, suggesting the regioselectivity diversity of steroidal hydroxylases of CYP5150 family. They are also distinct from the known bacterial and fungal steroidal hydroxylases in substrate specificity and regioselectivity. Biocatalytic hydroxylation is one of the important transformations for the functionalization of steroid nucleus rings but remains a very challenging task in organic synthesis. These hydroxylases are useful additions to the toolbox of hydroxylase enzymes for the functionalization of steroids at various positions.

INTRODUCTION

Steroids are a family of biologically and pharmaceutically important compounds (1). Functionalization of the steroidal nucleus rings often generates valuable compounds either as key intermediates for the synthesis of steroidal drugs or with special biological activities themselves (1–4). Steroid hydroxylation, which directly introduces an oxygen atom into the inactivated C-H bond on the steroidal core, is one of the most important reactions for steroidal functionalization. The resulting hydroxylated steroids can be used as high-value drugs or intermediates for further chemical synthesis (2, 5).

In contrast to the classical chemical methods that usually use strong oxidizing agents for the oxidation of the inactivated C-H bond of steroids, microbial catalysis can promote hydroxylation of steroidal compounds under mild conditions. Microorganisms have been reported to hydroxylate steroids at various positions of the steroidal nucleus (2, 5, 6). Some successful examples with high regio- and stereoselectivity have appeared, and the resulting hydroxylated products have been used in the production of marketed steroidal drugs. For example, dehydroepiandrosterone is dihydroxylated by using the Colletotrichum lini strain to furnish 7α,15α-dihydroxyandrostenolone at an 80% yield. This process is currently carried out at industrial scale as a key step in the production of drospirenone, a widely used “fourth-generation” progestin (7). However, in most cases, microbial hydroxylation of steroids shows low regio- and/or stereoselectivity, leading to a mixture of hydroxylated products at different positions of the steroidal core and/or with different stereochemistry.

The multiple P450 enzymes with undesired regiospecific and/or stereospecific hydroxylation activities in microbes are usually responsible for these drawbacks in microbial transformations. A straightforward approach to solving these problems is to use the recombinant regio- and stereospecific steroidal hydroxylase as the biocatalyst for the desired steroidal hydroxylation. Therefore, it is highly desirable to identify and functionally characterize regiospecific and stereospecific P450 enzymes and apply them in the selective hydroxylation of steroids. In this context, various steroidal hydroxylases, especially from bacterial sources, have recently been reported. For example, CYP106A family P450s (CYP106A1 and CYP106A2) from Bacillus hydroxylate a multitude of steroidal compounds, mainly at the 15β and 7α/β positions (8–11). The CYP109 family member CYP109E1 has been characterized as a 16β-hydroxylase of 3-ketosteroids (12). CYP154C family P450s show 16α-hydroxylation activities toward various steroidal substrates (13–15). CYP260A1 of the CYP260A family of P450s is able to convert various C₁₉ steroids at the 1α position, while CYP260B1 has been identified as a 6β- and 9α-hydroxylase of C₂₁ steroids (16, 17). On the other hand, although many fungal P450s are referred to as being involved in steroid hydroxylations, not many of them have been biochemically characterized.

It is known that the fungus Thanatephorus cucumeris catalyzes the hydroxylation of 17,21-dihydroxypregn-4-ene-3,20-dione (11-deoxycortisol, RSS, cortexolone, or cortodoxone) to produce 19-hydroxy- and 11β-hydroxy-11-deoxycortisol (18). Recently, we have found that T. cucumeris could also catalyze 7β-hydroxylation of 11-deoxycortisol, producing 7β-hydroxy-11-deoxycortisol when the biotransformation was performed under neutral or acidic conditions (19). Therefore, we envisioned that T. cucumeris might contain different P450 enzymes responsible for these hydroxylation reactions of 11-deoxycortisol. Because these hydroxylation reactions are useful in the functionalization of steroids, we were interested in identifying the individual P450 enzyme specific for each hydroxylation. In our previous study, we successfully cloned and identified a novel P450 gene responsible for C-19 and 11β-hydroxylation (20) in this strain. In this study, de novo RNA sequencing after exposure of the T. cucumeris to 11-deoxycortisol at different initial pHs was performed. Based on the bioinformatics analysis, a few inducible cytochrome P450 (CYP) genes were identified in the sequence of T. cucumeris. Among them, a highly upregulated CYP gene, cyp5150ap3, was cloned and expressed in Pichia pastoris. The recombinant enzyme was functionally characterized with 7β-hydroxylase activity toward 3-keto-4-en-steroids. In addition, the enzyme encoded by a downregulated gene, cyp5150an1, a member of the same P450 family, was also characterized with 2β-hydroxylase activity.

RESULTS

Transcriptome assembly of T. cucumeris.

Based on the results showing that 7β-hydroxyl-11-deoxycortisol was produced when biotransformation was carried out at the initial pH below 7.0 (Table 1), we inferred that the 7β-hydroxylase of T. cucumeris was inducible by 11-deoxycortisol under acidic biotransformation conditions and the target gene responsible for 7β-hydroxylation of 11-deoxycortisol would be upregulated at the transcriptional level. The total RNAs from the cultures with 11-deoxycortisol at initial pHs of 6.0 (sample D) and 9.0 (sample B) were isolated and sequenced (Table 2). Transcriptome data assembly and candidate gene annotation and classification were performed as described before (20). A total of 49 candidate P450 genes with completed open reading frames (ORFs) were identified. All the sequences were delivered to Cytochrome P450 Nomenclature Committee (http://drnelson.uthsc.edu/CytochromeP450.html) for naming. Differential transcript analysis revealed that 5 candidate cytochrome P450 genes were transcribed at levels 2- to 5-fold higher in sample D than in sample B. Four genes (cyp5144aa1, cyp5150ap2, cyp63aq1, and cyp5150an1) appeared to be downregulated in biotransformation at pH 6.0 (Table 3).

TABLE 1.

Product concentrations in the biotransformation by T. cucumeris at different initial pH values^a

Initial pH	Product concn (mg/liter)
	19-OH-RSS	11β-OH-RSS	7β-OH-RSS
9.0	118 (±10)	80 (±10)	17 (±2)
6.0	77 (±10)	62 (±5)	60 (±5)

^aThe initial RSS concentration was 0.5 g/liter. The biotransformation experiments were carried out for 4 days and repeated three times (19).

TABLE 2.

An overview of the RNA sequencing outcome, including numbers for total high-quality raw reads, nucleotides, and the statistical Q20, Q30, and GC percentages^a

Transcriptome	No. of:			% of nucleotide bases with:		GC (%)
	Raw reads	Clean reads	Clean bases	Q20	Q30
Sample B	20,106,548	19,791,168	2,470,000,000	96.18	92.03	51.82
Sample D	18,486,306	18,139,002	2,270,000,000	96.02	91.92	50.86

^aQ20 and Q30 represent Phred quality scores (Q) of 20 and 30, respectively.

TABLE 3.

The differently regulated P450 genes in the biotransformations at different initial pHs

Candidate gene	Sample Ba FPKMd	Sample Db FPKM	log2 FCc
cyp5150ap3	17.39	335.2	4.70
cyp52l2	0.6400	6.680	3.99
cyp63ar2	11.91	72.39	3.03
cyp5139ag1	1.03	3.870	2.34
cyp5637c3	1.500	5.470	2.29
cyp5144aa1	13.18	3.580	−1.43
cyp5150ap2	2,646	625.5	−1.64
cyp63aq1	4.370	0.8100	−1.99
cyp5150an1	1037	19.11	−5.32

^aSample B: T. cucumeris cells cultured with 0.5 g/liter of 11-deoxycortisol under initial pH 9.0.

^bSample D: T. cucumeris cells cultured with 0.5 g/liter of 11-deoxycortisol under initial pH 6.0.

^cFC, fold change.

^dFPKM, fragments per kilobase of transcript per million fragments mapped.

Analysis and expression of two hydroxylase candidate genes in P. pastoris.

Since the 7β-hydroxylase activity was enhanced during the biotransformation at the initial pH of 6.0, the upregulated genes might encode 7β-hydroxylase. Therefore, the most upregulated gene, cyp5150ap3, was further analyzed. The cyp5150ap3 gene showed significant homology with two downregulated genes (cyp5150ap2 and cyp5150an1) in the biotransformation at pH 6.0 compared to that at pH 9.0 (Table 3). Primary structure sequence alignment showed that CYP5150AP3 exhibits high identity (68%) with CYP5150AP2, a steroid hydroxylase with steroidal 19-hydroxylase and 11β-hydroxylase activities identified in our previous work (20). It also has 46% identity with CYP5150AN1, a member of CYP5150 P450 family of unknown function, which was also characterized in this study.

The full-length cyp5150ap3 gene was amplified and ligated into the P. pastoris methanol-inducible expression plasmid pPICZA together with the P450 oxidoreductase gene por of the same origin (20), and the plasmid was designated picZ-CYP5150AP3-POR. The resulting vector was used to transform P. pastoris X33 cells with the cyp5150ap3 and por genes. The corresponding recombinant strain was named X33-CYP5150AP3-POR. Recombinant plasmid picZ-CYP5150AP3 harboring the cyp5150ap3 gene alone was also constructed and used to generate P. pastoris strain X33-CYP5150AP3 as a control.

Similarly, the full-length cyp5150an1 gene, which is most downregulated at pH 6.0, was amplified and heterologously coexpressed with the por gene in P. pastoris to generate the recombinant strain X33-CYP5150AN1-POR. The strain X33-CYP5150AN1 expressing cyp5150an1 alone was used as a control.

Steroid biotransformation and product identification.

When the biotransformation of 11-deoxycortisol (RSS, cortexolone, or cortodoxone) was performed at pH 7.5 with the recombinant strain X33-CYP5150AP3-POR for 72 h, a major product, product 1, with the same retention time as 7β-hydroxy-11-deoxycortisol (7.4 min) was detected by high-performance liquid chromatography (HPLC) (Fig. 1, peak E). The reaction was also carried out at pHs 6.0 and 9.0, and the conversion was slightly lower. The conversion increased first but did not change after 72 h. This product was not detected in the control reactions with yeast strains containing the empty vector expressing CYP5150AP3 or POR alone (Fig. 1, peaks B to D). A major product, product 1, was purified and characterized as 7β-hydroxy-11-deoxycortisol by high-resolution mass spectrometry (HR-MS) and nuclear magnetic resonance (NMR) analysis. Besides product 1, two minor products with retention times of 16.1 and 25.3 min were detected by HPLC. The minor product at 16.1 min was not detected in the control experiments (Fig. 1, peaks B to D), suggesting that it might be the transformation product by this enzyme. However, we did not isolate enough sample for full structure characterization. It might be the 6β-hydroxylated product, because 6β-hydroxytestosterone was produced in the transformation of testosterone by the recombinant strain X33-CYP5150AN1-POR (see below). The minor product at 25.3 min was also detected in the control experiments (Fig. 1, peaks B to D), suggesting that it might be the metabolite of the host Pichia pastoris cells. These results demonstrate that the cyp5150ap3 gene encodes a P450 enzyme that is responsible for the formation of 7β-hydroxy-deoxycortisol in the biotransformation of 11-deoxycortisol by T. cucumeris under neutral or acidic conditions.

FIG 1.

HPLC analysis for the biotransformation products of 11-deoxycortisol (1 mM) by the recombinant strain X33-CYP5150AP3-POR at pH 7.5 and 72 h of incubation. A, standard references of 7β-hydroxy-11-deoxycortisol (1) and 11-deoxycortisol (RSS); B, transformation with strain X33 harboring empty vector; C, transformation with strain X33 expressing cyp5150ap3 alone; D, transformation with strain X33 expressing por alone; E, transformation with strain X33 coexpressing cyp5150ap3 and por.

The activities of X33-CYP5150AP3-POR toward other steroids, such as pregn-4-ene-3,20-dione (progesterone), 17α-hydroxypregn-4-ene-3,20-dione (17α-OH-progesterone), 16,17-epoxyprogesterone, and 17β-hydroxyandrost-4-en-3-one (testosterone), were also evaluated. Product formation was only obviously observed with testosterone as the substrate. As shown in Fig. 2, products 2 and 3 were detected by HPLC analysis, while none of them was detected in the control reactions (Fig. 2). Products 2 and 3 were purified and identified by HR-MS and NMR analysis as 7β-hydroxytestosterone and 6β-hydroxytestosterone, respectively. The results showed that CYP5150AP3 catalyzed the 7β- and 6β-hydroxylation of testosterone to give the corresponding hydroxylated products in about a 1:1 ratio.

FIG 2.

HPLC analysis for the biotransformation products of testosterone (1 mM) by the recombinant strain X33-CYP5150AP3-POR at pH 7.5 and 72 h of incubation. A, standard reference testosterone (TS); B, transformation with strain X33 harboring empty vector; C, transformation with strain X33 expressing cyp5150ap3 alone; D, transformation with strain X33 expressing por alone; E, transformation with strain X33 coexpressing cyp5150ap3 and por. Numbers 2 and 3 refer to new products 2 and 3.

The biotransformation of 11-deoxycortisol by the recombinant strain X33-CYP5150AN1-POR was also investigated. A major product, product 4, with a retention time at 18.6 min was detected by HPLC (Fig. 3, peak E), and no product was accumulated by control strains. Product 4 was purified and characterized as 2β-hydroxy-11-deoxycortisol by HR-MS and NMR analysis. A minor peak at a retention time of 12.8 min was detected by HPLC but not observed in the control experiments (Fig. 3, peaks B to D). This minor product was not characterized due to the difficulty of obtaining pure product.

FIG 3.

HPLC analysis for the biotransformation products of 11-deoxycortisol (1 mM) by the recombinant strain X33-CYP5150AN1-POR at pH 7.5 and 72 h of incubation. A, standard references of 11-deoxycortisol; B, transformation with strain X33 harboring empty vector; C, transformation with strain X33 expressing cyp5150an1 alone; D, transformation with strain X33 expressing por alone; E, transformation with strain X33 coexpressing cyp5150an1 and por. The number 4 refers to new product 4.

Other steroids such as progesterone, 17α-OH-progesterone, 16,17-epoxyprogesterone, and testosterone were also used as substrates for the biotransformation with X33-CYP5150AN1-POR resting cells, and no obvious product was detected. Therefore, CYP5150AN1 was identified with the ability of catalyzing the 2β-hydroxylation of 11-deoxycortisol. The 2β-hydroxylated product might exist as a minor product in the microbial transformation of RSS using T. cucumeris due to the low expression level of the corresponding gene but has not been characterized in the previous studies.

DISCUSSION

Fungal organisms harbor a large number of P450 genes with an enormous diversity of functions, performing numerous important physiological and ecological roles in cell growth, defense, and detoxification (21, 22). In order to respond to the environment, the expression of P450s is also differently regulated by fungi under diversified conditions such as dissolved oxygen (23), pH (24), and temperature and medium composition (25, 26). In our studies, we found that the expression levels of cyp5150ap2 and cyp5150an1 were higher under an alkaline condition (pH 9.0) than in acidic medium (pH 6.0). In contrast, cyp5150ap3 was found to be highly expressed in acidic medium. The pH level of medium is one of the main stresses affecting microbial behavior. For example, the gene expression levels of enzymes involved in the oxidative phosphorylation and electron transfer chain are differently regulated when the pH changes, and this is proposed to be responsible for the high level of acid stress tolerance of P. anomala (27, 28). It is interesting that differential expression patterns in response to the pH level of biotransformation medium are observed among steroidal hydroxylases of the same family, and the underlying causes and/or physiological functions need further study.

Since a suitable NAD(P)H-cytochrome P450 reductase (POR) is usually required for the activity of P450 enzymes and coexpression of P450s with a POR of the same origin is a common strategy for P450 functional identification (29–31), the previously characterized POR from T. cucumeris was used in this study (20). Recombinant P. pastoris X33 whole cells coexpressing cyp5150ap3 with por catalyzed the hydroxylation of 11-deoxycortisol to give 7β-hydroxy-11-deoxycortisol as the major product (about 90%) and a minor product (about 10%) at 16.1 min, which was proposed to be the 6β-hydroxylated product (Fig. 1). Therefore, CYP5150AP3 should be the hydroxylase responsible for 7β-hydroxylation of 11-deoxycortisol by T. cucumeris. For the 3-keto-4-en-C₁₉ steroid testosterone, this enzyme catalyzed 7β- and 6β-hydroxylation with nearly 1:1 regioselectivity (Fig. 2). These results indicate that the side chain at the 17 position of the steroid core affects the regioselectivity of CYP5150AP3 in the hydroxylation of 3-keto-4-en-steroids (Fig. 4). The structure of steroid substrate also exerted great effect on the activity of CYP5150AP3 enzyme, and it exhibited little or no activity toward progesterone, 17α-hydroxy-progesterone, and 16,17-epoxyprogesterone.

FIG 4.

Hydroxylation reactions of 3-keto-4-en steroids catalyzed by the recombinant strain X33-CYP5150AP3-POR. The product 6β-hydroxy-11-deoxycortisol was not characterized but proposed based on the reaction of testosterone with strain X33-CYP5150AP3-POR.

A bacterial P450, CYP106A1, with low identity to CYP5150AP3 has been reported to have 6β- and 7β-hydroxylase activities toward the C₁₉ steroid androstenedione but no activity toward 3-hydroxysteroids. When C₂₁ steroids, such as desoxycortone, 11-deoxycortisol, and cortisone, were used as substrates of CYP106A1, 15β-hydroxylation was observed (10). In contrast, its isozyme CYP106A2, known as a 15β-hydroxylase toward 3-keto-4-en-steroid, mediates the 7β- and 7α-hydroxylation of dehydroepiandrosterone (DHEA), a 3-hydroxy-5-ene-C₁₉ steroid. The structure of the A ring of steroid molecules significantly influences the position of the introduced hydroxyl group by this enzyme (9, 11). Although CYP5150AP3 and the two bacterial P450 enzymes all catalyze 6β- and 7β-hydroxylations, they are distinct from each other in substrate specificity and regioselectivity.

Hydroxylation at C-7 produces very useful hydroxysteroids of pharmacological value (32, 33). Although many fungi have been reported to mediate 7-hydroxylation of steroids (34–42), low stereoselectivity (7α/7β) (38, 41) and side hydroxylation reactions (37, 42) have often been observed, thus presenting great challenges in developing highly regio- and stereoselective 7-hydroxylation processes of steroidal compounds. CYP5150AP3 favors the 7β-hydroxylation of 11-deoxycortisol, a 3-keto-4-en-C₂₁ steroid, and may serve as a starting point for addressing these challenges.

Hydroxylation at the C-2 position of the steroid core has been less studied. In animals, testosterone 2β-hydroxylation catalyzed by human and rat liver microsomes was identified. As for other biocatalysts, a bacterial P450 CYP102A1 (P450-BM3) mutant F87V was reported to have 2β- and 15β-hydroxylase activities toward progesterone and testosterone (43). Recently, two other bacterial hydroxylases, CYP154C4-1 and CYP154C4-2, have been reported to convert testosterone to 2α- and 2β-hydroxylated products (15). Interestingly, they showed hydroxylation activities mainly at the 16α position when progesterone was used as the substrate. In this study, CYP5150AN1 was distinct from P450s mentioned above, consistent with the fact that the amino acid sequence identity between them is lower than 15%.

Three fungal steroid hydroxylases have been reported to have different hydroxylation regioselectivities from CYP5150AP3 and CYP5150AN1. CYP509C12 from Rhizopus oryzae showed 11α- and minor 6β-hydroxylation activities toward 11-deoxycortisol and testosterone (44). CYP5311B1 from Absidia coerulea was capable of hydroxylating 16,17α-epoxyprogesterone at the 11α position (45). P450pra from Penicillium raistrickii was identified as a steroid hydroxylase with 15α-hydroxylation activity toward d-ethylgonendione (13-ethyl-gon-4-ene-3,17-dione) (46). Considering the rather low homology with each other, it is reasonable that these enzymes show diverse regioselectivity of steroid hydroxylation. In the current study, the steroidal hydroxylase CYP5150AP3 showed 68% amino acid sequence identity and 76% similarity with its sibling, CYP5150AP2, and 46% identity and 52% similarity with CYP5150AN1. Surprisingly, these enzymes showed a completely different regioselectivity toward 11-deoxycortisol. CYP5150AP2 shows steroidal 19-hydroxylase and 11β-hydroxylase activities, and no 7β-hydroxylase activity has been observed (20). CYP5150AP3 hydroxylates the 11-deoxycortisol mainly at the 7β position, without any 19-hydroxylase activity, while CYP5150AN1 favors the 2β-hydroxylation of 11-deoxycortisol (Fig. 5).

FIG 5.

Different regioselectivity of the CYP5150 family P450s toward hydroxylation of 11-deoxycortisol. CYP5150AP2 was identified in previous work (20); CYP5150AP2 and CYP5150AN1 were studied in the current investigation.

Several steroid hydroxylases have been classified into specific families (47). Combined with our previous work (20), we successfully identified three steroidal hydroxylases (CYP5150AP2, CYP5150AP3, and CYP5150AN1) of the CYP5150A family from T. cucumeris with different hydroxylation regioselectivities. In order to get insight into the differences in steroidal transformation between the CYP5150A enzymes, an alignment was performed; all conserved features of P450s were recognized in all enzymes. Focusing on the substrate recognition sites (SRSs), previously identified by Gotoh (48), 6 SRSs were predicted in each of these P450s (Fig. 6). We believe that certain domain or key sites should determine the regioselectivity of these enzymes, and further studies are needed to identify these structural determinants.

FIG 6.

Different substrate recognition site domains between CYP5150AP2, CYP5150AP3, and CYP5150AN1.

In summary, transcriptome sequencing was used to identify a few cytochrome P450 genes from the fungus T. cucumeris, which were differently regulated by the initial pH for biotransformation. Among them, a gene highly upregulated under acidic conditions, cyp5150ap3, was coexpressed with a redox partner por gene in P. pastoris strain X33. The biotransformation with the recombinant whole cells demonstrated that CYP5150AP3 catalyzed the 7β-hydroxylation of 11-deoxycortisol and 6β- and 7β-hydroxylations of testosterone. In addition, a gene downregulated under acidic conditions, cyp5150an1, was identified to encode an enzyme catalyzing the 2β-hydroxylation of 11-deoxycortisol. These enzymes show different substrate specificity and regioselectivity from those of the known bacterial and fungal steroidal hydroxylases. These newly characterized P450s would enrich the library of hydroxylase enzymes for the functionalization of steroidal compounds. CYP5150AP3 shows high identity with CYP5150AP2 (68%) and CYP5150AN1 (46%) but completely different regioselectivity. Further studies are needed to understand the structural basis underlying the diverse regioselectivity and substrate specificity of these P450 enzymes for the hydroxylation of steroidal compounds.

MATERIALS AND METHODS

Materials.

11-Deoxycortisol (purity >98%) and testosterone (purity, >98%) were obtained from Toronto Research Chemicals. 7β-Hydroxy-11-deoxycortisol (purity, >95%) was produced with T. cucumeris NBRC 6298 as described previously (18). Peptone and yeast extract were purchased from Oxoid Ltd. Yeast nitrogen base without amino acids (YNB) was purchased from Thermo Fisher Scientific Ltd. All other chemicals were of analytical grade and bought from Merck.

Microorganisms, plasmids, and cultural conditions.

Fungal strain T. cucumeris NBRC 6298 was obtained from the NITE Biological Resource Center (http://www.nbrc.nite.go.jp/NBRC2/NBRCCatalogueDetailServlet?ID=NBRC&CAT=00006298) and routinely maintained on potato dextrose agar (PDA) slants. Escherichia coli DH5α was grown at 37°C in Luria-Bertani (LB) medium (10 g/liter of tryptone, 5 g/liter of yeast extract, 10 g/liter of NaCl, and 15 g/liter of agar [pH 7.0]) and used as a host for molecular cloning of candidate genes. Pichia pastoris X33 (Invitrogen) was used as a host organism for the expression of candidate CYP genes and grown at 30°C in YPD medium (20 g/liter of peptone, 10 g/liter of yeast extract, 20 g/liter of dextrose, and 15 g/liter of agar), buffered minimal glycerol medium (BMG; 100 mM potassium phosphate [pH 7.5], 13.4 g/liter of YNB, 0.4 mg/liter of biotin, and 1% glycerol), or buffered minimal methanol medium (BMM; 100 mM potassium phosphate [pH 7.5], 13.4 g/liter of YNB, 0.4 mg/liter of biotin, and 1% methanol) with Zeocin (100 mg/liter).

Fungal sample preparation, RNA sequencing, and data preprocessing.

For RNA sequencing, T. cucumeris was cultured with 0.5 g/liter of 11-deoxycortisol at pH 9.0 or 6.0 for 24 h. The cells were collected and the total RNA was isolated using TRIzol reagent (Promega, USA) followed by purification using RNeasy separation columns (RNeasy kit; Qiagen). The purified total RNA was sequenced with Hiseq 2000 (Illumina, USA) at Novogene Co., Ltd. (Tianjin, China). Transcriptome assembly, annotation, and gene expression analysis were performed as previously described (20). Briefly, the assembly of clean data was performed using Trinity (49). Gene expression levels were estimated by fragments per kilobase per million (FPKM) (50). Differential expression analysis of two samples was performed using the bioconductor package edgeR (51). The read counts were adjusted by edgeR through one scaling-normalized factor. P value was adjusted using q value (52). A q value of <0.005 and a fold change of >1.5 were set as the threshold for significantly differential expression (53).

Construction and transformation of recombinant plasmids.

The putative CYP gene cyp5150ap3 and cyp5150an1 were amplified by PCR using cDNA as the templates and the primers in Table 4 under the following conditions: initial denaturation at 95°C for 2 min, followed by 30 cycles of denaturation at 95°C for 45 s, annealing at 59°C for 45 s, and extension at 72°C for 1 min 30 s, with a final extension step at 72°C for 10 min. The amplified gene was ligated into the KpnI/NotI-digested sites of the E. coli/P. pastoris shuttle vector pPICZ A to form pPICZ-CYP5150AP3 and pPICZ-CYP5150AN1. The identity of the gene in the plasmid was verified by sequencing (BGI, Shenzhen, Guangdong, China). The cytochrome P450 oxidoreductase (POR) from T. cucumeris NBRC 6298 was amplified with primers in Table 4. The coexpressing plasmid pPICZ-CYP5150AP3-POR was constructed in accordance with our previous study (20). Briefly, the por cDNA was amplified and then ligated to pUC57 vector at KpnI and NotI restriction enzyme sites to form pUC57-POR. One silent mutation at the EcoRI restriction enzyme site was created for elimination of the restriction site in por using PCRs. The resulting sequence was ligated into pPICZA at EcoRI and ApaI restriction sites to form the pPICZ-MPOR recombination vector. To construct the CYP gene expression cassette, the original PmeI site in the pPICZA vector was mutated; the resulting construct was named dePmeI-pPICZA. Binary vector binary-pPICZ-MPOR was constructed by ligating the pPICZA-MPOR BamHI-digested fragment with the BglII-BamHI-digested dePmeI-pPICZA fragment. A clone positive for tail-to-head orientation (Pichia expression manual; Invitrogen) was selected. The target gene cyp5150ap3 and cyp5150an1 were cloned into the por-containing binary vector at KpnI-NotI restriction sites to construct the coexpressing plasmids pPICZ-CYP5150AP3-POR and pPICZ-CYP5150AN1-POR. The coexpressing plasmids were linearized with PmeI and then electrotransformed into P. pastoris X33 cells according to the manufacturer’s instructions for the Pichia expression kit. The recombinant P. pastoris X33 cells harboring coexpressing plasmids were selected from YPD medium supplemented with 1 mg ml⁻¹ of Zeocin, and the integration of the cyp5150ap3-por or cyp5150an1-por expression cassette was verified by PCR techniques. The correct strains were named X33-CYP5150AP3-POR and X33-CYP5150AN1-POR. The empty vector pPICZA, pPICZ-CYP5150AP3, pPICZ-CYP5150AN1, and pPICZ-MPOR were also transformed into P. pastoris X33 to obtain control strains, which were named X33-CK, X33-CYP5150AP3, X33-CYP5150AN1, and X33-POR, respectively. The names of all the relative strains and vectors used and created in this work are listed in Table 5.

TABLE 4.

Primers for PCRs in this study

Name	Sequence (5′–3′)a
F-KpnI-cyp5150ap3	CATCCGGTACCATGGACCCCCTCTTGAAATACTTTC
R-NotI-cyp5150ap3	CGCGGCGGCCGCTCAATAATCGACCAGTGTGATCTTC
F-KpnI-cyp5150an1	CATCCGGTACCATGTCCGAAACCCGCTTTG
R-NotI-cyp5150an1	CGCGGCGGCCGCTTACAATACAGATACTTTAAGAGGCATAGTG
F-KpnI-por	CATCCGGTACCATGGCTCCTGCTCTCTCGAC
R-NotI-por	CGCGGCGGCCGCCTATGACCAGACATCCAACAACAAC
F-MEcoRI-por	ACCGATAATGCAGTCGAGTTCATGAATAACATCAAC
R-MEcoRI-por	GTTGATGTTATTCATGAACTCGACTGCATTATCGGT
F-EcoRI-por	CATCCGAATTCATGGCTCCTGCTCTCTCGAC
R-ApaI-por	GTTCGGGCCCCTATGACCAGACATCCAACAACAAC
F-dePmeI-pICZA	GGCCCAAAACTGACAGTTGAAACGCTGTCTTGGAACCT
R-dePmeI-pICZA	AGGTTCCAAGACAGCGTTTCAACTGTCAGTTTTGGGCC

^aRestriction sites in primers are underlined, and mutations are shown in bold.

TABLE 5.

Strains and plasmids used in this study

Strain or plasmid	Genotype or phenotype	Source or reference
Strains
Thanatephorus cucumeris NBRC 6298	Wild type	NBRC
Pichia pastoris X33	Wild type	Invitrogen
E. coli TOP 10	F– mcrA Δ(mrr-hsdRMS-mcrBC) Φ80lacZΔM15 ΔlacX74 recA1 araD139 Δ(ara leu) 7697 galU galK rpsL (StrR) endA1 nupG	Invitrogen
X33-CK	Pichia pastoris X33 harboring pICZA empty vector	21
X33-POR	Pichia pastoris X33 expressing por alone	21
X33-CYP5150AP3	Pichia pastoris X33 expressing cyp5150ap3 alone	This work
X33-CYP5150AN1	Pichia pastoris X33 expressing cyp5150an1 alone	This work
X33-CYP5150AP3-POR	Pichia pastoris X33 coexpressing cyp5150ap3 and por	This work
X33-CYP5150AN1-POR	Pichia pastoris X33 coexpressing cyp5150an1 and por	This work
Plasmids
pPICZA	Yeast expression vector (Pichia pastoris); Zeor	Invitrogen
pUC57	Cloning vector; Ampr	Invitrogen
pPICZ-CYP5150AP3	T. cucumeris cyp5150ap3 cDNA cloned in pPICZA vector at KpnI-NotI site; Zeor	This work
pPICZ-CYP5150AN1	T. cucumeris cyp5150an1 cDNA cloned in pPICZA vector at KpnI-NotI site; Zeor	This work
pUC57-POR	T. cucumeris P450 reductase gene por cloned in pUC57 vector at KpnI-NotI site; Ampr	21
pPICZ-MPOR	pPICZA-por vector containing por gene mutated (silent mutation) at EcoRI; Zeor	21
dePmeI-pPICZA	pPICZA vector mutated at PmeI site; Zeor	21
binary-pPICZ-MPOR	BglII-BamHI-digested dePmeI-pPICZA fragment cloned in pPICZ-MPOR at BamHI site; Zeor	21
pPICZ-CYP5150AP3-POR	cyp5150ap3 gene cloned in binary-pPICZ-MPOR at KpnI-NotI site; Zeor	This work
pPICZ-CYP5150AN1-POR	cyp5150an1 gene cloned in binary-pPICZ-MPOR at KpnI-NotI site; Zeor	This work

Heterologous expression of two hydroxylase candidate genes and por in P. pastoris and steroid biotransformation.

X33-CYP5150AP3-POR colonies with robust growth on YPD plates containing 1 mg ml⁻¹ of Zeocin were inoculated into 25 ml of BMG medium (pH 7.5) and cultured (48 h) until the optical density at 600 nm (OD₆₀₀) reached 10. The cells were collected by centrifugation (4,000 × g for 5 min) and diluted to an OD₆₀₀ of 1.0 in BMM medium (pH 7.5) containing aminolevulinic acid (2 mM), and 1% (vol/vol) methanol was added to induce expression of candidate genes every 24 h for 5 days at 20°C on a rotary shaker (250 rpm). The colonies of X33-CK, X33-CYP5150AP3, and X33-POR were used as controls.

Similarly, X33-CYP5150AN1-POR colonies were inoculated and induced as described above. The colonies of X33-CK, X33-CYP5150AN1, and X33-POR were used as controls.

Whole-cell steroid transformation was performed at 30°C on a rotary shaker (200 rpm) as described previously (20). Briefly, the methanol-induced recombinant strains were first collected by centrifugation (4,000 × g for 5 min) and resuspended in 30 ml of potassium phosphate buffer (50 mM; pH 6.0, 7.5, or 9.0) containing aminolevulinic acid (2 mM) in 250-ml shake flasks. Then the substrates were added to the reaction mixture at a final concentration of 1 mM. Methanol (1%, vol/vol) was added at each 24-h interval. The samples were taken from the reaction mixtures at intervals of 24, 48, 72, 96, and 120 h and extracted with ethyl acetate for high-performance liquid chromatography (HPLC) analysis.

Preparative reactions and product identification.

P. pastoris X33-CYP5150AP3-POR and X33-CYP5150AN1-POR colonies were inoculated into 500 ml of BMG medium (pH 7.5) in 2-liter shake flasks and incubated for 48 h at 30°C and 250 rpm until the OD₆₀₀ reached 10.0. Five milliliters of 11-deoxycortisol or testosterone solution in methanol was added into the culture at a final substrate concentration of 1 mM. Methanol (1%, vol/vol) was added to induce expression of candidate genes every 24 h for 4 days.

The transformation products were analyzed by HPLC with an Eclipse XDB C₁₈ column (250 mm by 4.6 mm by 5 μm). The mobile phase was methanol-water at a ratio of 45/55 or 70/30 (vol/vol) for the transformation of 11-deoxycortisol or testosterone, respectively. The flow rate was maintained at 0.6 ml/min with a column temperature of 30°C. The substrate and the hydroxylation products were detected by absorbance at 254 nm.

To isolate the products, the reaction mixtures were extracted with equivalent volumes of ethyl acetate twice. The extracts were concentrated and evaporated under reduced pressure. The residue was resuspended in appropriate volumes of dichloromethane. The products were isolated by thin-layer chromatography (TLC) (TLC plate, 100 mm by 200 mm by 0.5 mm; Anhui Liangchen Silicon Material Co., Ltd., Anhui, China). For the hydroxylation of 11-deoxycortisol, the plate was developed with dichloromethane-methanol (15:1, vol/vol). The product band was collected and extracted with dichloromethane, and further purification was performed by a preparative reverse-phase recycling HPLC with an XDB C₁₈ column using methanol-water (50/50, vol/vol) as the eluent. The flow rate was maintained at 11 ml/min at a column temperature of 30°C. For the hydroxylation of testosterone, the plate was developed with petroleum ether-ethyl acetate (45:55, vol/vol). Each product band was collected and extracted with dichloromethane, and further purification was performed by a preparative reverse-phase recycling HPLC with an XDB C₁₈ column using methanol-water (70/30, vol/vol) as an eluent. The flow rate was maintained at 11 ml/min at a column temperature of 30°C. The purified products were characterized by HR-MS and by ¹H and ¹³C nuclear magnetic resonance (NMR) analyses.

Product 1, 7β-hydroxy-11-deoxycortisol: HR-MS [M + Na]⁺ m/z 385.1996 (calculated, 385.1991). ¹H NMR (CD₃OD, 400 MHz) δ (ppm) 5.78 (s, 1H), 4.63 (m, 1H), 4.31 (m, 1H), 3.37 to 3.32 (m, 1H), 2.61 to 2.68 (m, 1H), 2.42 to 2.58 (m, 3H), 2.25 to 2.33 (m, 1H), 2.04 to 2.14 (m, 2H), 1.82 to 1.98 (m, 2H), 1.74 to 1.61 (m, 4H), 1.38 to 1.56 (m, 3H), 1.24 (s, 3H), 0.92 to 1.03 (m, 2H), 0.69 (s, 3H). ¹³C NMR (CD₃OD, 100 MHz) δ (ppm) 214.0, 202.3, 171.7, 124.9, 89.7, 75.8, 68.1, 52.0, 51.3, 49.8, 44.3, 43.9, 39.5, 36.8, 35.7, 34.8, 31.6, 27.4, 21.8, 17.7, 15.6.

Product 2, 7β-hydroxytestosterone: HR-MS [M + Na]⁺ m/z 327.1940 (calculated, 327.1940). ¹H NMR (600 MHz, CD₃OD) δ (ppm) 5.71 (s, 1H), 4.16-4.19 (m, 1H), 3.58 (t, J = 8.4 Hz, 1H), 2.58 to 2.62 (m, 1H), 2.29 to 2.33 (m, 1H), 2.23 to 2.28 (m, 1H), 1.94 to 2.02 (m, 2H), 1.86 to 1.91 (m, 1H), 1.70 to 1.78 (m, 2H), 1.54 to 1.65 (m, 1H), 1.44 to 1.53 (m, 1H), 1.36 to 1.41 (m, 1H), 1.27 to 1.35 (m, 1H), 1.22 (s, 3H), 1.11 to 1.18 (m, 1H), 0.97 to 1.05 (m, 2H), 0.79 (s, 3H). ¹³C NMR (125 MHz, CD₃OD) δ (ppm) 201.6, 171.6, 126.8, 82.4, 73.7, 55.4, 51.4, 44.1, 39.5, 39.4, 38.5, 37.8, 35.1, 31.2, 30.7, 24.3, 21.8, 19.8, 11.8.

Product 3, 6β-hydroxytestosterone: HR-MS [M + Na]⁺ m/z 327.1940 (calculated, 327.1940). ¹H NMR (600 MHz, CD₃OD) δ (ppm) 5.78 (s, 1H), 4.25 to 4.27 (m, 1H), 3.59 (t, J = 8.4 Hz, 1H), 2.52 to 2.60 (m, 1H), 2.28 to 2.33 (m, 1H), 2.05 to 2.10 (m, 1H), 1.94 to 2.04 (m, 3H), 1.86 to 1.91 (m, 1H), 1.59-1.74 (m, 3H), 1.46 to 1.56 (m, 2H), 1.39 (s, 3H), 1.30 to 1.38 (m, 1H), 1.18 to 1.24 (m, 1H), 1.05 to 1.12 (m, 1H), 0.96 to 1.03 (m, 1H), 0.91 to 0.96 (m, 1H), 0.81 (s, 3H). ¹³C NMR (125 MHz, CD₃OD) δ (ppm) 201.6, 176.3, 120.6, 82.2, 59.9, 52.2, 51.8, 44.5, 42.5, 41.6, 37.8, 37.1, 35.7, 33.9, 30.8, 24.3, 23.5, 23.2, 11.8.

Product 4, 2β-hydroxy-11-deoxycortisol: HR-MS [M + Na]⁺ m/z 385.2017 (calculated, 385.1991). ¹H NMR (600 MHz, CD₃OD) δ (ppm) 5.71 (s, 1H), 4.68 (m, 1H), 4.36 (m, 1H), 4.24 (t, J = 7.8 Hz, 1H), 2.70 to 2.73 (m, 1H), 2.60 to 2.64 (m, 1H), 2.47 to 2.52 (m, 1H), 2.29 to 2.37 (m, 2H), 1.94 to 2.07 (m, 2H), 1.82 to 1.86 (m, 3H), 1.52 to 1.79 (m, 4H), 1.33 to 1.45 (m, 2H), 1.23 (s, 3H), 1.09 to 1.12 (m, 1H), 0.6 (s, 3H). ¹³C NMR (125 MHz, CD₃OD) δ (ppm) 212.0, 200.2, 174.7, 119.1, 88.9, 68.3, 66.4, 50.3, 50.2, 48.1, 40.9, 40.2, 35.6, 34.7, 33.6, 32.4, 30.3, 23.1, 22.0, 21.6, 14.1.

Accession number(s).

The RNA-seq data reported here were deposited in the NCBI Sequence Read Archive (SRA) under accession numbers SRX3599177 and SRX4526378. The nucleotide sequences of cyp5150ap3 and cyp5150an1 were deposited in the GenBank database under accession numbers MG721490.1 and MF987471.1.