English

• Steroids are widely distributed in natural organisms and have great significance on clinical application due to their important and broad pharmacological activities, such as anti-microbial [1], anti-tumor [2], anti-inflammatory [3], and anti-human immunodeficiency virus (HIV) effects [4]. Currently, over 400 steroid-based drugs are extensively used for treatment on inflammation, cancer, rheumatic, arthritis, and contraception [5–8]. The bioactivity of steroid drugs largely depends on the type and number of functional groups in specific positions and stereo-configurations on the skeleton [9]. In that way, the hydroxyl groups in specific positions become the vital building blocks in the synthesis of steroid drugs. However, such hydroxylations are challenging through chemical catalysis due to the predominance of inert C−H bonds in steroids.

  Microbial steroid transformation has been utilized for years as a powerful tool to generate novel steroidal drugs and routinely incorporated into the synthetic routes as a key to affect specific oxygenation of the steroids for the reason of the high regio- and stereo-selectivity, better process efficiency, and eco-friendliness [7,10–15]. More importantly, fungal biotransformation occupies an important position in steroid biotransformation for its affluent enzyme system, and some fungal cytochrome P450 monooxygenases (CYPs) with varying steroid hydroxylation activities have been identified (Fig. S2 in Supporting information) [16]. However, the number of fungal steroidal CYPs and their range of hydroxylation positions remain limited, so digging more fungal CYPs and applying them to the steroid production will undoubtedly drive the industrial progress of steroidal drugs forward.

  Beauveria bassiana is a crucial microbial catalyst used in a variety of chemical biotransformation, which is capable of accepting multiple types of substrates and catalyzing diverse types of reactions [17–21], especially the hydroxylation of steroids [22]. Despite the completion of the genome sequence of B. bassiana, which revealed 83 putative CYPs [23], the related steroid CYPs have not yet been identified. Therefore, there is a considerable interest in mining the putative CYPs of B. bassiana that have high catalytic hydroxylation ability.

  Androst-4-ene-3,17-dione (4-AD, 1) is a key intermediate in steroid synthesis, whose hydroxylation products usually have pharmacological activities or play important roles in industrial production of steroid drugs [24,25]. Firstly, we carried out the biotransformation of 4-AD by B. bassiana in order to give more detailed time-course for steroid transformation and lay a foundation for the discovery of steroidal CYPs from B. bassiana. When 4-AD was fed to B. bassiana, eight metabolites (1a–1h) were observed after 3 days incubation (Fig. 1A), and thus were isolated from the extracts of the culture medium. We identified their structures by mass spectrometry (MS) and nuclear magnetic resonance (NMR) spectra, and divided them into two groups as the hydroxylated derivatives of 4-AD (1f, 1a, 1b, 1d, 1e) and those of testosterone (TES, 2) which was the C17 keto reduction product of 4-AD (1h, 1c, 1g) (Fig. 1C). It is worth mentioning that 1β, 11α-dihydroxyandrost-4-ene-3,17-dione (1e) and 1β, 11α-dihydroxytestosterone (1g) were two new compounds (the detailed structure elucidation of new compounds and the spectral data of known compounds were described in Figs. S3 and S4 in Supporting information). The position and relative configuration of two hydroxyl groups in 1e and 1g were determined by analysis of ¹H NMR, ¹³C NMR, heteronuclear single quantum coherence (HSQC), heteronuclear multiple-bond correlation (HMBC) and ¹H–¹H correlation spectroscopy (¹H–¹H COSY) (Figs. S20–S31 in Supporting information).

  Figure 1

  

  Figure 1.  Biotransformation of 4-AD (1) by B. bassiana. (A, B) Time course profile for the biotransformation of 1 by B. bassiana. Culture control consisted of fermentation blanks in which microorganisms were grown under identical conditions but without the substrate. Substrate control consisted of sterile medium containing the same amount of substrate and were incubated under the same conditions. (C) A proposed biotransformation pathway of 1 by B. bassiana. The color depth of arrowhead represents the conversion rate, the darker the color, the higher the conversion rate. ^★ New compounds.

  DownLoad: Full-Size Img PowerPoint

  The transformation route of 4-AD by B. bassiana was summarized by analyzing the time-course changes of the products. Initially, the substrate was almost completely transformed to 11α-hydroxylated products 1f and 1h after 24 h incubation. The second oxidation of 1f and 1h including 6β-hydroxylation, 1β-hydroxylation or Baeyer-Villiger oxidation occurred gradually in the next 96 h, leading to the corresponding production of 1b, 1d, 1e and 1c, 1g, respectively. Besides, 1b underwent an elimination to produce 1a (Fig. 1C). The above results clearly indicated the presence of at least one efficient steroid oxygenase in B. bassiana.

  Since steroidal hydroxylases of B. bassiana may be induced by substrate at the transcriptional level, the substrate 4-AD was added to the culture medium to up-regulate the expression of related hydroxylases. Strategy of RNA sequencing was chosen to screen for the genes with differential expression levels in different states, especially up-regulated genes after the substrate induction.

  Moreover, we focused on CYPs, because almost all steroidal hydroxylases reported are CYPs [16]. Among the 70 up-regulated genes (Fig. S5 in Supporting information), two CYP genes were selected for further analysis which exhibited a 2.48-fold increase (BBA_03078) and 8.75-fold (BBA_08698) increase in their transcript abundance after induction with 4-AD (1) (Fig. 2, Table S1 in Supporting information).

  Figure 2

  

  Figure 2.  Volcano map of differential genes. Points filled with light-pink or light-blue represent gene expressions with increased or decreased abundance (|log₂ Fold change| ≥ 1) in response to 4-AD induction, while gray points represent genes with no differences in the expression. Two CYP genes are colored as red, respectively.

  DownLoad: Full-Size Img PowerPoint

  The two genes were successfully amplificated from the total cDNA of B. bassiana, which were named CYP68BE1 (BBA_03078) and CYP58A5 (BBA_08698) by International CYP Nomenclature Committee. To explore the steroid hydroxylation activities of two candidate CYP genes, cyp68be1 and cyp58a5 (Fig. S6 and Table S3 in Supporting information) were respectively introduced into S. cerevisiae INVSc1 cells, yielding the recombinant strains SC-1 (harboring cyp68be1) and SC-3 (harboring cyp58a5). Catalytic activities of the vast majority of CYPs require redox partner proteins to sequentially deliver electrons from nicotinamide adenine dinucleotide phosphate hydrogen (NADPH) [26]. To successfully perform the function of the two CYPs, a recombinant vector containing AoCYP reductase (AoCPR) and cyt b5 genes was constructed and transformed into SC-1 and SC-3, respectively, resulted in the generation of the recombinant strains SC-2 and SC-4 (Tables S4 and S5, Figs. S7 and S8 in Supporting information). Additionally, the mock transformants harboring the empty vectors were also generated as the controls.

  Biotransformations of 4-AD were performed with the aforementioned recombinants for 48 h (Fig. S8), and the results were analyzed by high-performance liquid chromatography (HPLC)-MS. Compared with the controls SC-5 and SC-6, the SC-3 and SC-4 strains did not produce any products, indicating that CYP58A5 was not the target CYP. To our delight, the SC-1 and SC-2 recombinant yeasts produced two major products (1f and 1b) and four minor ones (1h, 1a, 1e and 1g) based on comparison with the standards, demonstrating that CYP68BE1 is a multifunctional CYP performing consecutive catalytic processes. This enzyme could hydroxylate 4-AD at the C11α, C6β or C1β-positions (Fig. 3). Meanwhile, SC-1 and SC-2 had the same capacities to hydroxylate substrate 1, which indicated that CYP68BE1 could interact with the endogenous CPR of S. cerevisiae.

  Figure 3

  

  Figure 3.  HPLC results of in vivo functional verification of two CYPs in recombinant S. cerevisiae.

  DownLoad: Full-Size Img PowerPoint

  To explore the catalytic promiscuity of CYP68BE1 on different steroids, we selected TES (2), 1,4-androstenedione (1,4-AD, 3), progesterone (P4, 4), estrone (E1, 5), and estra-4,9-diene-3,17-dione (E-4,9, 6) (Fig. S1 in Supporting information) for in vivo biotransformation assays by the recombinant strain SC-1. All the above five substrates are frequently used as the sexual hormones or vital industrial intermediates. The HPLC-MS results (Figs. S9–S16 in Supporting information) showed that all of them could be mono- or di-hydroxylated with high capabilities by SC-1. To elucidate the structures of the products, the large-scale fermentations were carried out. A total of 12 mono- or di-hydroxylated products were isolated, in which three products (5b, 6b, and 6d) were novel compounds. These structures were identified by MS, 1D and 2D NMR spectroscopic data analysis (structure elucidation section was showed in Figs. S13–S16 and S32–S49 in Supporting information). Besides, the NMR data of other three metabolites (5a, 6a and 6c) were reported for the first time. The above results indicated that C11α was the most common position for steroid oxidation by CYP68BE1, but different binding orientations (C6β, C1α or C1β) of CYP68BE1 could be observed with different substrates. Since some accessory products (1a, 4b, 5b) yielded in vivo affected the functional research of CYP68BE1, we conducted an in vitro functional assay on CYP68BE1 by extracting microsomes to avoid the influence of internal factors of S. cerevisiae on the experimental results.

  For substrates 1–4, with highly structural similarity, 11α-monohydroxylation process occurred in catalysis of the all four substrates by CYP68BE1 (Figs. 4A–D, Figs. S7–S10). Similarly, the monohydroxylated products of the four substrates continued to be hydroxylated under the catalysis of CYP68BE1, generating the corresponding 6β, 11α-dihydroxylated products. But the regio-selectivity of the enzyme is slightly different in the process of catalyzing the second-hydroxylation step of 1 and 2. TES (2) was also hydroxylated by CYP68BE1 to yield extra 1g both in vivo and in vitro (Fig. 4B, Fig. S10). The dihydroxylation product of 1 also included small amount of 1β, 11α-type (1e). Possibly due to the decrease in the activity of CYP68BE1, no 1e was detected in the HPLC analysis of in vitro biotransformation. (Fig. 4A, Fig. S9). Based on the yields of different products, it was suggested that the regio-selectivity of CYP68BE1 preferred the C6β position of these four substrates during the dihydroxylation process.

  Figure 4

  

  Figure 4.  Results of catalytic promiscuity tests of CYP68BE1. (A–F) In vitro functional verification and conversion route of CYP68BE1 to substrates 1–6. (G) Conversion products distribution of 1–6 from CYP68BE1 in vitro assay. MP-Control refers to the microsome protein of SC-5, and MP-CYP68BE1 refers to the microsome protein of SC-1. ^★ New compounds. ^* Hydroxylated steroids products in which C17 keto has been reduced. Products 3c, 3d and 5d were not determined due to separation problems, but their structures were deduced to be hydroxylated products, considering their molecular weights.

  DownLoad: Full-Size Img PowerPoint

  Certainly, the structure of substrate 5 is quite different from other substrates, characterized by a benzene ring in the A ring and the absence of 19-CH₃. The catalytic activity of CYP68BE1 on this substrate is noteworthy, as it produced not only the expected 11α-monohydroxylated metabolite (5a) and 6β, 11α-dihydroxylated metabolite (5d), but also a 6β-monohydroxylated metabolite (5c) (Fig. 4E, Fig. S13). The catalytic hydroxylation of substrate 6 only produced the monohydroxylated products: 11α-hydroxyestra-4,9-diene-3,17-dione (6a), 1α-hydroxyestra-4,9-diene-3,17-dione (6b), 11β, 17β-dihydroxyestra-4,9-diene-3-one (6c) and 1α, 17β-dihydroxyestra-4,9-diene-3-one (6d), which were significantly different from those of the other substrates (Fig. 4F, Fig. S14).

  To systematically explain the mechanisms of the production of different hydroxylated products above, we used Alphafold2 to perform three-dimensional structure prediction, and molecular dynamics (MD) simulation to explore the molecular basis for regio-selectivity of CYP68BE1 to different steroid substrates (Figs. S17–S19 in Supporting information). The 11α-hydroxylation product (1h) of 2 was docked into the pocket of CYP68BE1 with C11 pointing to the ferrous of heme. Then 200 ns MD simulations were performed and the evolution of distances of Fe-C11 and Fe-C6 was recorded. The distance of Fe-C11 approximately maintained within 5 Å (4.58 ± 0.33 Å) from 0 to 75 ns, followed by the distance of Fe-C11 increased immediately with Fe-C6 distance decreased to about 5 Å (4.58 ± 0.33 Å) (Fig. 5A). The above-mentioned change of distance held the possibility that 1h, as the intermediate of modification reaction to 2, could flip to the C6 near-attack pose in favor of the next modification of C6-hydroxylation. However, we could not exclude another possibility that the intermediate 1h firstly dissociated from the pocket and then bound into the pocket again with C6 near-attack pose.

  Figure 5

  

  Figure 5.  MD simulations of CYP68BE1 in the reaction with different substrates. (A) The evolution of distance over simulation times of compound 1h. (B) Representative C11 near-attack pose generated by cluster analysis from MD simulations of 5. (C) Representative C6 near-attack pose generated by cluster analysis from MD simulations of 5. (D) Representative C11/C1 near-attack pose from MD simulations of 6.

  DownLoad: Full-Size Img PowerPoint

  Due to the identification of 5a and 5c, both of which were monohydroxylated products, it is suggested that 5 could take two different near-attack poses in the pocket of CYP68BE1. MD simulations based on two initial near-attack poses of C11 and C6 were conducted respectively (Figs. 5B and C, Fig. S18). Within the 500 ns of trajectories of two systems, C11 and C6 near-attack poses were basically retained with both distances keeping around 5 Å (5.02 ± 0.30 Å and 4.05 ± 0.30 Å). Compared with simulations of 1h, the flip of conformation was not observed in simulation of 5. The relative stable conformations of 5 suggested that the substrate could adopt a constant near-attack pose regardless of C6 or C11 once binding into the protein pocket. However, no observations of transformation between different near-attack poses were likely caused by insufficient time scale of MD simulations.

  E-4,9 (6) preserved permanent conformation with Fe-C11 (4.75 ± 0.30 Å) and Fe-C1 (4.74 ± 0.37 Å) distance keeping around 5 Å, implying that CYP68BE1 could modify C11 or C1 sites of 6 (Fig. 5D, Fig. S19). Lacking obvious distinction between Fe-C11 and Fe-C1 distances, CYP68BE1 did not have regioselectivity to the two sites of 6. The substrate probably was hydroxylated at either C11 or C1 site, followed by being released from the pocket, due to obtaining of only monohydroylated products from experiments. There was the possibility that steric hindrance owing to formation of monohydroxylated products hampered the further attack to either C11 or C1 from heme.

  In summary, B. bassiana showed a strong capacity for 4-AD (1) transformation, forming two new 1β, 11α-hydroxylated metabolites (1e and 1g), except for its previously reported function to catalyze C11α and C6β, 11α hydroxylation [17]. Employing strategies of transcriptomics, heterologous expression, in vivo and in vitro functional characterization, we excavated and identified a new steroid hydroxylase CYP68BE1, which was responsible for the hydroxylation on multiple positions (11α; 1α; 6β; 1β, 11α; 6β, 11α) of six steroid substrates with more than 70% conversion rate (Fig. 4G), indicating that CYP68BE1 was promiscuous and catalytically versatile. We obtained 12 mono- or di-hydroxylated products, and three of them are new compounds. Notably, of the products involved in this work, 11α-hydroxyandrostenedione (1f) is an important intermediate used to produce diuretics eplerenone [27]; 1β, 11α-dihydroxytestosterone (1g) is expected to be a significant intermediate in the synthesis of ouabain; 11β, 17β-dihydroxyestra-4,9-diene-3-one (6c) is regarded as a vital intermediate for the production of trenbolone [28]; 11α-hydroxyprogesteron (4c) is a potent inhibitor of 11β-hydroxysteroid dehydrogenase type 2 (11β HSD-2) [29]. Clearly, steroidal CYP with sufficient substrate promiscuity and high regioselectivity is usually regarded as eco-friendly and efficient biocatalyst for the synthesis of high-value steroid drugs and the enrichment of nature product library by enzymatic way or metabolic engineering technology. Therefore, we performed MD simulation with the aim of understanding the structural basis of the regioselectivity and substrate specificity of CYP68BE1. On the one hand, further designing mutation studies are needed to bias the catalytic reactions toward desired products, facilitating its industrial applications. On the other hand, combining with other modification enzymes, such as methylases and glycosyltransferases, more structurally diverse steroids could be created for the pharmacological screening.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  Acknowledgments

  The study was supported by the National Key Research and Development Program of China (Nos. 2020YFA0908003 and 2018YFA0901900), CAMS Innovation Fund for Medical Sciences (No. CIFMS2021-I2M-1–029). We are grateful to the Department of Instrumental Analysis of our institute for measurement of the NMR and HRESIMS data. We acknowledge the Network and Information Center of our institute for providing computing resources. We would like to thank Yufan Chou for support with MD services.

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.108818.

  
  
• 1. [1]

     F. Carvalhall, M. Correia-da-Silva, E. Sousa, M. Pinto, A. Kijjoa, J. Mol. Endocrinol. 61 (2018) T211–T231. doi: 10.1530/JME-17-0252

  2. [2]

     V.M. Dembitsky, Prog. Lipid Res. 79 (2020) 101048. doi: 10.1016/j.plipres.2020.101048

  3. [3]

     F.R.S. Passos, H.G. Araújo-Filho, B.S. Monteiro, et al., Phytomedicine 96 (2022) 153842. doi: 10.1016/j.phymed.2021.153842

  4. [4]

     X. Pang, L.P. Kang, X.M. Fang, et al., J. Nat. Med. 72 (2018) 166–180. doi: 10.1007/s11418-017-1126-1

  5. [5]

     W.Y. Tong, X. Dong, Recent Pat. Biotechnol. 3 (2009) 141–153. doi: 10.2174/187220809788700157

  6. [6]

     K. Yildirim, A. Kuru, E. Küçükbaşol, Biocatal. Biotransform. 38 (2020) 7–14. doi: 10.1080/10242422.2019.1604690

  7. [7]

     M.V. Donova, D.V. Egorova, Appl. Microbiol. Biotechnol. 94 (2012) 1423– 1447. doi: 10.1007/s00253-012-4078-0

  8. [8]

     Y. Ma, J.M. Liu, R.D. Chen, X.Q. An, J.G. Dai, Chin. Chem. Lett. 28 (2017) 1200–1204. doi: 10.1016/j.cclet.2017.02.019

  9. [9]

     K. Tantuco, E. Deretey, I.G. Csizmadia, J. Mol. Struct. 503 (2000) 97–111. doi: 10.1016/S0166-1280(99)00366-8

  10. [10]

      N. Liu, J.H. Fen, R. Zhang, et al., Green Chem. 21 (2019) 4076–4083. doi: 10.1039/c9gc01267b

  11. [11]

      A. Bartmanska, J. Dmochowska-Gladysz, E. Huszcza, Steroids 70 (2005) 193–198. doi: 10.1016/j.steroids.2004.11.011

  12. [12]

      N. Nassiri-Koopaei, M.A. Faramarzi, Biotechnology 33 (2015) 1–28. doi: 10.3109/10242422.2015.1022533

  13. [13]

      M.S. Smitha, S. Singh, R. Singh, J. Bacteriol. Mycol. 4 (2017) 47–51.

  14. [14]

      C.X. Wu, J.Q. Xu, J.P. Xie, Z.Y. Wang, Chin. Chem. Lett. 29 (2018) 1251–1253. doi: 10.1016/j.cclet.2018.06.025

  15. [15]

      H.N. Bhatti, R.A. Khera, Steroids 77 (2012) 1267–1290. doi: 10.1016/j.steroids.2012.07.018

  16. [16]

      L.B. Xiong, L. Song, Y.Q. Zhao, et al., Synth. Biol. J. 2 (2021) 942–963.

  17. [17]

      Z. Xiong, Q. Wei, H. Chen, et al., Steroids 71 (2006) 979–983. doi: 10.1016/j.steroids.2006.07.007

  18. [18]

      A. Swizdor, T. Kolek, A. Panek, A. Białońska, Biochim. Biophys. Acta 1811 (2011) 253–262. doi: 10.1016/j.bbalip.2011.01.005

  19. [19]

      F. Nicolau, T. Peeples, Biotechnol. Adv. 5 (2016) 1–14.

  20. [20]

      A. Cano-Flores, G. Delgado, Chem. Biodivers. 14 (2017) e1700211. doi: 10.1002/cbdv.201700211

  21. [21]

      L.N. Xie, L.W. Zhang, C. Wang, et al., Proc. Nat. Acad. Sci. U. S. A. 115 (2018) E4980–E4989.

  22. [22]

      J.X. Zhan, E.M.K. Wijeratne, A.A.L. Gunatilak, Nat. Prod. Commun. 5 (2010) 801–804.

  23. [23]

      G.H. Xiao, S.H. Ying, P. Zheng, et al., Sci. Rep. 2 (2012) 483. doi: 10.1038/srep00483

  24. [24]

      Y. Zhao, B. Zhang, Z.Q. Sun, et al., ACS Catal. 12 (2022) 9839–9845. doi: 10.1021/acscatal.2c02185

  25. [25]

      A. Li, C.G. Acevedo-Rocha, L. D'Amore, et al., Angew. Chem. Int. Ed. 59 (2020) 12499–12505. doi: 10.1002/anie.202003139

  26. [26]

      W. Zhang, Y. Liu, J.Y. Yan, et al., J. Am. Chem. Soc. 136 (2014) 3640–3646. doi: 10.1021/ja4130302

  27. [27]

      K.L. Davis, J.M. Nappi, Clin. Ther. 25 (2003) 2647–2668. doi: 10.1016/S0149-2918(03)80326-0

  28. [28]

      Y. Peng, C. Gao, Z. Zhang, et al., ACS Catal. 12 (2022) 2907–2914. doi: 10.1021/acscatal.1c05776

  29. [29]

      H.M.M. Zhou, M.F.F.E.P. Gomez-Sanchez, E.N. Cozza, et al., Clin. Ther. 137 (2003) 2308–2314.

• 

  Figure 1  Biotransformation of 4-AD (1) by B. bassiana. (A, B) Time course profile for the biotransformation of 1 by B. bassiana. Culture control consisted of fermentation blanks in which microorganisms were grown under identical conditions but without the substrate. Substrate control consisted of sterile medium containing the same amount of substrate and were incubated under the same conditions. (C) A proposed biotransformation pathway of 1 by B. bassiana. The color depth of arrowhead represents the conversion rate, the darker the color, the higher the conversion rate. ^★ New compounds.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Volcano map of differential genes. Points filled with light-pink or light-blue represent gene expressions with increased or decreased abundance (|log₂ Fold change| ≥ 1) in response to 4-AD induction, while gray points represent genes with no differences in the expression. Two CYP genes are colored as red, respectively.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  HPLC results of in vivo functional verification of two CYPs in recombinant S. cerevisiae.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Results of catalytic promiscuity tests of CYP68BE1. (A–F) In vitro functional verification and conversion route of CYP68BE1 to substrates 1–6. (G) Conversion products distribution of 1–6 from CYP68BE1 in vitro assay. MP-Control refers to the microsome protein of SC-5, and MP-CYP68BE1 refers to the microsome protein of SC-1. ^★ New compounds. ^* Hydroxylated steroids products in which C17 keto has been reduced. Products 3c, 3d and 5d were not determined due to separation problems, but their structures were deduced to be hydroxylated products, considering their molecular weights.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  MD simulations of CYP68BE1 in the reaction with different substrates. (A) The evolution of distance over simulation times of compound 1h. (B) Representative C11 near-attack pose generated by cluster analysis from MD simulations of 5. (C) Representative C6 near-attack pose generated by cluster analysis from MD simulations of 5. (D) Representative C11/C1 near-attack pose from MD simulations of 6.

  下载: 全尺寸图片 幻灯片
•