Highly Regioselective and Stereoselective Biohydroxylations of Oxandrolone

: Microbially catalyzed reactions are a powerful and valuable tool for organic synthesis of many compounds with potential biological activity. Herein, we report efﬁcient hydroxylations of the steroidal anabolic-androgenic lactone, oxandrolone, in the cultures of three strains of fungi, Fusarium culmorum , Mortierella isabellina , and Laetiporus sulphureus . These reactions resulted in the production of four metabolites identiﬁed as 12 β -hydroxyoxandrolone ( 2 ), 9 α -hydroxyoxandrolone ( 3 ), 6 α -hydroxyoxandrolone ( 4 ), and 15 α -hydroxyoxandrolone ( 5 ), the latter being a new compound. The high substrate conversion rates and the product yields achieved indicate that these strains offer a new way to generate steroidal hydroxylactones with potential pharmaceutical interest. The structures of the isolated derivatives were characterized on the basis of spectroscopic data. The effect of modiﬁcation of the A-ring structure of the steroid by the lactone group on the selectivity of hydroxylation in cultures of the tested fungi is also discussed.


Introduction
The ability of microorganisms to transform organic compounds makes them a suitable instrument for obtaining valuable molecules that cannot be formed easily by traditional chemical methods. As they occur at neutral pH and ambient temperature, and do not require expensive and harmful chemicals, biocatalytic reactions are easy to carry out, cost-effective and safe for the environment. Most fungal strains contain cytochrome P450 monooxygenases, similar to those of mammals, so they can carry out regioselective and stereoselective hydroxylation of molecules, such as steroids, which may increase or modify their biological activity [1]. The presence of a hydroxyl group in organic compounds affects their polarity, toxicity, extraction from the cell, and translocation via the cell membranes, while creating possibilities for various subsequent structural modifications in the search for new substances with healing properties [1,2].
Microbial hydroxylations are used commercially for the production of steroidal drugs and to produce steroidal key intermediates for their synthesis. The 11α-, 11β-, 9α-, and 16α-hydroxylations are utilized in the global manufacture of corticosteroids [2,3]. The introduction of a hydroxyl group at the 11α-position of progesterone by Rhizopus nigricans decreased the number of chemical steps from 36 to 11 in cortisone synthesis from deoxycholic acid and caused a remarkable reduction in the price of the drug [4]. As another example, 11β-hydroxylation of cortexolone by Culvularia lunata is used to obtain hydrocortisone [2,4,5], and 9α-hydroxy steroids are intermediates for the production of more active 9α-halogenated corticoids. The presence of a 16α-hydroxyl group in halogen corticosteroids enables a reduction of the undesirable mineralocorticoid activity of these derivatives [2,3,6]. Steroid hydroxylations occurring in other sites, including the 7α-, 7β-, 14α-and 15α-positions, may have the potential for industrial operation. The 7α-hydroxylation of cholesterol is essential for the production of bile acids, which have a

Results and Discussion
Our results indicated that F. culmorum AM282, M. isabellina AM212 and L. sulphureu AM498 were able to convert oxandrolone (1) into its polar metabolites. The time-depend ent progress of the transformations is compiled in Table 1. The structures of the obtained derivatives were established by spectroscopic methods by comparison of the characteris tic shift values of selected diagnostic signals of products and the starting compound. Th stereochemistry of the hydroxyl group was deduced on the basis of NOESY experiments The bioconversion of oxandrolone (1) with F. culmorum yielded only one main me tabolite (Figure 1), which was obtained with a yield of 62% mol. (69% determined by GC analysis, see Table 1).  Mass spectrometry (MS) data of metabolite 2 revealed an [M] + at m/z 323.2 (C19H30O4) reflecting the additional OH group in the molecule. The 1 H-NMR spectrum showed a new downfield methine proton signal at δH 3.73-3.76 ppm, geminal to the hydroxyl group. In comparison to the spectrum of the substrate, the C-18 methyl signal of 2 demonstrated downfield shift (Δ 0.07 ppm), suggesting that hydroxylation had occurred at the 12β po sition. The appearance of a new methine carbon signal at δC 73.1 ppm in the 13 C NMR spectrum, in combination with the downfield shifts of C-11 (Δ 10.9 ppm) and C-13 (Δ 4. ppm), and upfield shift of C-18 (Δ 5.5 ppm) when compared to the spectrum of 1, was an important confirmation of 12β-hydroxylation. Additionally, the NOESY spectrum showed correlations between 12α-H and 9α-(δH 0.90 ppm), 14α-(δH 1.18 ppm) proton and C-17α (δH 1.35 ppm) methyl group signals. Thus, product 2 was identified as 12β hydroxy-oxandrolone. Its NMR data were comparable with those reported in the litera ture [28]. The effective introduction of 12β-OH group into the structure of 17α-alkylated androstanes by F. culmorum was not surprising. Although this strain converts C19 3-oxo 4-ene and 3β-hydroxy-5α-saturated steroids to a mixture of their 6β-and 15α-hydroxy  In comparison to the spectrum of the substrate, the C-18 methyl signal of 2 demonstrated a downfield shift (∆ 0.07 ppm), suggesting that hydroxylation had occurred at the 12β position. The appearance of a new methine carbon signal at δ C 73.1 ppm in the 13 C NMR spectrum, in combination with the downfield shifts of C-11 (∆ 10.9 ppm) and C-13 (∆ 4.1 ppm), and upfield shift of C-18 (∆ 5.5 ppm) when compared to the spectrum of 1, was an important confirmation of 12β-hydroxylation. Additionally, the NOESY spectrum showed correlations between 12α-H and 9α-(δ H 0.90 ppm), 14α-(δ H 1.18 ppm) proton, and C-17α (δ H 1.35 ppm) methyl group signals. Thus, product 2 was identified as 12β-hydroxy-oxandrolone. Its NMR data were comparable with those reported in the literature [28]. The effective introduction of 12β-OH group into the structure of 17αalkylated androstanes by F. culmorum was not surprising. Although this strain converts C19 3-oxo-4-ene and 3β-hydroxy-5α-saturated steroids to a mixture of their 6β-and 15αhydroxy derivatives [13][14][15]29], in the presence of an alkyl substituent at the 17α-position, which is a drawback for 1-3 cis hydroxylation at the 15α-position, the reaction occurs at position 12β, which is its equivalent in enzyme-substrate complexes [2]. Thus, after transformation of 17α-methyltestosterone in F. culmorum AM282 culture, the mixture of 6β-, 15α-, and 12β-hydroxy derivatives at a 5:2:2 ratio was obtained [14]. We also found that the modification of the ring A of steroids and replacement of 3-oxo-4-ene group with an unusual δ lactone moiety makes such orientation of the substrate in the complex with the enzyme, that 12β-hydroxylation occurs. [2]. 12β-hydroxy-oxandrolone (2) was previously synthesized after 18 days of the transformation of 1 using Cunninghamella blakesleeana strain but it was isolated with only a 0.8% yield [28].
After seven days of incubation of oxandrolone (1) with M. isabellina AM212, three potential products were detected in the post-transformation extracts, but only one metabolite, 3 (80% based on GC analysis, see Table 1), was present in sufficient quantity for characterization ( Figure 2).
Catalysts 2021, 11, x FOR PEER REVIEW 4 of derivatives [13][14][15]29], in the presence of an alkyl substituent at the 17α-position, which i a drawback for 1-3 cis hydroxylation at the 15α-position, the reaction occurs at position 12β, which is its equivalent in enzyme-substrate complexes [2]. Thus, after transformation of 17α-methyltestosterone in F. culmorum AM282 culture, the mixture of 6β-, 15α-, and 12β-hydroxy derivatives at a 5:2:2 ratio was obtained [14]. We also found that the modifi cation of the ring A of steroids and replacement of 3-oxo-4-ene group with an unusual δ lactone moiety makes such orientation of the substrate in the complex with the enzyme that 12β-hydroxylation occurs. [2]. 12β-hydroxy-oxandrolone (2) was previously synthe sized after 18 days of the transformation of 1 using Cunninghamella blakesleeana strain bu it was isolated with only a 0.8% yield [28]. After seven days of incubation of oxandrolone (1) with M. isabellina AM212, thre potential products were detected in the post-transformation extracts, but only one metab olite, 3 (80% based on GC analysis, see Table 1), was present in sufficient quantity fo characterization ( Figure 2). The MS data of metabolite 3 exhibited an [M] + at m/z 323.2, which implied the incor poration of an oxygen atom to the substrate. In the 1 H NMR spectrum of 3, the CHOH signal was absent. The signal of OH-bearing quaternary carbon at δC 74.2 ppm, the down field shift of signals C-10 (Δ 4.4 ppm), C-11 (Δ 2.1 ppm), and C-8 (Δ 2.7 ppm), and the γ gauche upfield shifts of C-1 (Δ 4.4 ppm), C-5 (Δ 8.1 ppm), C-7 (Δ 3.5 ppm), C-12 (Δ 4.1 ppm), and C-14 (Δ 6.8 ppm) signals in comparison with the spectrum of metabolite 1 in dicated that 9α-hydroxylation had occurred. The proposed structure, 9α-hydroxy-oxan drolone (3), was supported by the downfield shift of C-19 methyl signal by 0.10 ppm with respect to the substrate. The spectroscopic data of this metabolite were consistent with those described in the literature [21] and 9α-hydroxyoxandrolone (3) was previously re ported as one of three products of a transformation using Rhizopus stolonifer with an iso lated yield of 8% after 10 days of cultivation [21]. In the current study with M. isabellin AM212, the yield was improved to 71%. In comparison to the transformation of epiandros terone (3β-hydroxy-5α-androstan-17-one) by M. isabellina, which resulted in a mixture o 7α-, 9α-, and 11α-hydroxy derivatives [9], there is another example that the introduction of a lactone moiety into the A-ring of 5α-saturated substrate enhances the regioselectivity of the hydroxylation.
Two others metabolites (4 and 5 at a 1:4.2 ratio) were detected after ten days of incu bation of 1 with Laetiporus sulphureus AM498 (Table 1, Figure 3).  The MS data of metabolite 3 exhibited an [M] + at m/z 323.2, which implied the incorporation of an oxygen atom to the substrate. In the 1 H NMR spectrum of 3, the CHOH signal was absent. The signal of OH-bearing quaternary carbon at δ C 74.2 ppm, the downfield shift of signals C-10 (∆ 4.4 ppm), C-11 (∆ 2.1 ppm), and C-8 (∆ 2.7 ppm), and the γ-gauche upfield shifts of C-1 (∆ 4.4 ppm), C-5 (∆ 8.1 ppm), C-7 (∆ 3.5 ppm), C-12 (∆ 4.1 ppm), and C-14 (∆ 6.8 ppm) signals in comparison with the spectrum of metabolite 1 indicated that 9α-hydroxylation had occurred. The proposed structure, 9αhydroxy-oxandrolone (3), was supported by the downfield shift of C-19 methyl signal by 0.10 ppm with respect to the substrate. The spectroscopic data of this metabolite were consistent with those described in the literature [21] and 9α-hydroxyoxandrolone (3) was previously reported as one of three products of a transformation using Rhizopus stolonifer with an isolated yield of 8% after 10 days of cultivation [21]. In the current study with M. isabellina AM212, the yield was improved to 71%. In comparison to the transformation of epiandrosterone (3β-hydroxy-5α-androstan-17-one) by M. isabellina, which resulted in a mixture of 7α-, 9α-, and 11α-hydroxy derivatives [9], there is another example that the introduction of a lactone moiety into the A-ring of 5α-saturated substrate enhances the regioselectivity of the hydroxylation.
derivatives [13][14][15]29], in the presence of an alkyl substituent at the 17α-position, which is a drawback for 1-3 cis hydroxylation at the 15α-position, the reaction occurs at position 12β, which is its equivalent in enzyme-substrate complexes [2]. Thus, after transformation of 17α-methyltestosterone in F. culmorum AM282 culture, the mixture of 6β-, 15α-, and 12β-hydroxy derivatives at a 5:2:2 ratio was obtained [14]. We also found that the modification of the ring A of steroids and replacement of 3-oxo-4-ene group with an unusual δ lactone moiety makes such orientation of the substrate in the complex with the enzyme, that 12β-hydroxylation occurs. [2]. 12β-hydroxy-oxandrolone (2) was previously synthesized after 18 days of the transformation of 1 using Cunninghamella blakesleeana strain but it was isolated with only a 0.8% yield [28].
After seven days of incubation of oxandrolone (1) with M. isabellina AM212, three potential products were detected in the post-transformation extracts, but only one metabolite, 3 (80% based on GC analysis, see Table 1), was present in sufficient quantity for characterization ( Figure 2). The MS data of metabolite 3 exhibited an [M] + at m/z 323.2, which implied the incorporation of an oxygen atom to the substrate. In the 1 H NMR spectrum of 3, the CHOH signal was absent. The signal of OH-bearing quaternary carbon at δC 74.2 ppm, the downfield shift of signals C-10 (Δ 4.4 ppm), C-11 (Δ 2.1 ppm), and C-8 (Δ 2.7 ppm), and the γgauche upfield shifts of C-1 (Δ 4.4 ppm), C-5 (Δ 8.1 ppm), C-7 (Δ 3.5 ppm), C-12 (Δ 4.1 ppm), and C-14 (Δ 6.8 ppm) signals in comparison with the spectrum of metabolite 1 indicated that 9α-hydroxylation had occurred. The proposed structure, 9α-hydroxy-oxandrolone (3), was supported by the downfield shift of C-19 methyl signal by 0.10 ppm with respect to the substrate. The spectroscopic data of this metabolite were consistent with those described in the literature [21] and 9α-hydroxyoxandrolone (3) was previously reported as one of three products of a transformation using Rhizopus stolonifer with an isolated yield of 8% after 10 days of cultivation [21]. In the current study with M. isabellina AM212, the yield was improved to 71%. In comparison to the transformation of epiandrosterone (3β-hydroxy-5α-androstan-17-one) by M. isabellina, which resulted in a mixture of 7α-, 9α-, and 11α-hydroxy derivatives [9], there is another example that the introduction of a lactone moiety into the A-ring of 5α-saturated substrate enhances the regioselectivity of the hydroxylation.
Two others metabolites (4 and 5 at a 1:4.2 ratio) were detected after ten days of incubation of 1 with Laetiporus sulphureus AM498 (Table 1, Figure 3).   The MS spectra of both metabolites showed an [M] + at m/z 323.2, corresponding to the formula C 19 H 30 O 4 and indicating the incorporation of an oxygen atom into the structure of the substrate. The 1 H NMR spectrum of the less polar product revealed a new broad signal at δ H 3.45-3.49 ppm suggesting hydroxylation at an equatorial proton. The significant shift of the 4-H signals (by 0.46 ppm and 0.12 ppm) and lack of a shift of both the methyl signals pointed towards substitution at 6α-H. The appearance of a new methine carbon signal at δ C 69.8 ppm, in combination with the downfield shift of the C-5 (∆ 6.7 ppm) and C-7 (∆ 10.0 ppm), and γ-carbon upfield signal shift of C-4 (∆ 3.1 ppm) and C-8 (∆ 0.9 ppm) confirmed 6α-hydroxylation. Additionally, the NOESY spectrum showed a correlation between 6β-H and 4β-(δH 2.34 ppm), 8β (δH 1.61 ppm) proton, and C-19 (δH 1.02 ppm) methyl signals. These observations fully supported the structure of 4 as 6α-hydroxyoxandrolone and were consistent with the spectroscopic data available in the literature [21]. It was previously reported that 6α-hydroxyoxandrolone (4) was one of three metabolites obtained after 10 days transformation of 1 by Rhizopus stolonifer with a yield of 5% [21]. The 1 H NMR spectrum of the main reaction product, metabolite 5, showed a new broad hydroxyl-bearing methine proton signal at δ H 4.07 ppm. The characteristic shape and multiplicity of this signal (td, J = 3.4 Hz; J = 9.3 Hz) suggested 15α-hydroxylation. This was supported by an oxygen-bearing methine carbon signal at δ C 72.3 ppm, downfield shifts of the β-carbons C-14 (∆ 7.7 ppm) and C-16 (∆ 11.6 ppm), and γ-carbon upfield signal shifts of C-8 (∆ 0.5 ppm) and C-17 (∆ 2.5 ppm). Further supporting evidence of the stereochemistry of the new hydroxyl group was provided by NOESY spectrum analysis, in which the correlation signals between 15β-H and H-8 (δ H 1.68 ppm), and H-16 (δ H 2.42 ppm) and C-18 methyl (δ H 0.87 ppm) signals were visible (Table 2 and Figure 4). These results confirmed the identification of 5 as 15α-hydroxyoxandrolone. This compound is a new derivative of oxandrolone.  To the best of our knowledge, this is the first report confirmin oxandrolone by fungi. The use of L. sulphureus AM498 allowed us ylactone with a high yield (70% mol.). Moreover, the use of myce ypore in the transformation of steroids compounds is described he

Substrate and Microorganisms
The substrate oxandrolone (1) was purchased from Biosynth C United Kingdom). The fungal strains Fusarium culmorum AM28 AM212 and Laetiporus sulphureus AM498 were obtained from the c ment of Pharmaceutical Biology and Botany of the Wrocław Wrocław, Poland. Fungi were maintained on Sabouraud 4% dex freshly sub-cultured before use in the transformation experiments

Conditions of Cultivation and Transformation
All strains were incubated in 300 mL Erlenmeyer flasks with composed of 3% of glucose and 1% of proteose peptone (Biocorp, W were grown at 26 °C on a rotary shaker (130 rpm). After a specif that depended on the growth rate of the strain (2 days for Laetip and Fusarium culmorum AM282, 3 days for Mortierella isabellina AM 1 in the form of an acetone suspension was equally distributed am fungal cultures (final concentration 0.2 gl −1 ). Transformations wer same conditions until the substrate was metabolized. To study th gress of the bioconversion, 5 mL samples of the transformation m at regular intervals (48 h) from the reaction flasks and processed 3.3. To the best of our knowledge, this is the first report confirming 15α-hydroxylation of oxandrolone by fungi. The use of L. sulphureus AM498 allowed us to obtain this hydroxylactone with a high yield (70% mol.). Moreover, the use of mycelium of this edible polypore in the transformation of steroids compounds is described here for the first time.

Substrate and Microorganisms
The substrate oxandrolone (1) was purchased from Biosynth Carbosynth ® (Berkshire, United Kingdom). The fungal strains Fusarium culmorum AM282, Mortierella isabellina AM212 and Laetiporus sulphureus AM498 were obtained from the collection of the Department of Pharmaceutical Biology and Botany of the Wrocław Medical University, Wrocław, Poland. Fungi were maintained on Sabouraud 4% dextrose agar slopes and freshly subcultured before use in the transformation experiments.

Conditions of Cultivation and Transformation
All strains were incubated in 300 mL Erlenmeyer flasks with 100 mL of a medium composed of 3% of glucose and 1% of proteose peptone (Biocorp, Warsaw Poland). Fungi were grown at 26 • C on a rotary shaker (130 rpm). After a specified time of cultivation that depended on the growth rate of the strain (2 days for Laetiporus sulphureus AM498 and Fusarium culmorum AM282, 3 days for Mortierella isabellina AM212), 0.06 g of substrate 1 in the form of an acetone suspension was equally distributed among three flasks with fungal cultures (final concentration 0.2 gl −1 ). Transformations were carried out under the same conditions until the substrate was metabolized. To study the time-dependent progress of the bioconversion, 5 mL samples of the transformation medium were taken out at regular intervals (48 h) from the reaction flasks and processed as described in Section 3.3.

Isolation and Identification of Products
After transformations were performed, the mixture was extracted three times with chloroform (3 × 100 mL). The organic extract was dried over anhydrous magnesium sulphate, concentrated in vacuo and analyzed by TLC and GC. Merck Silica gel 60 F 254 plates and eluent chloroform/methanol (70:6 v/v) were used for TLC analysis. To detect the steroids, the TLC plates were sprayed with a methanol/concentrated sulfuric acid (1:1 v/v) mixture, heated at 120 • C until the colors developed, and observed under UV light (254 nm). GC analysis was performed on a Hewlett Packard 5890 Series II GC instrument (FID, carrier gas H 2 ) with the DB-5MS capillary column (30 m × 0.32 mm, film 0.25 µm; Agilent Technologies, Inc., Warsaw, Poland). The temperature program for the GC analysis was as follows: 220 • C for 1 min increasing by increments of 4 • C/min to 270 • C, then at a gradient of 30 • C/min to 300 • C, held at 300 • C for 3 min; the injector and detector temperature was taken at 300 • C. MS analysis was performed on Varian CP-3800/Saturn 2000 apparatus (Varian, Walnut Creek, CA, USA) with a Zebron ZB-5 MSI (30 m × 0.25 mm × 0.25 µm) column (Phenomenex, Torrance, CA, USA). Temperature conditions were as follows: 220 • C for 1 min, increasing by 5 • C/min to 300 • C, held at Catalysts 2021, 11, 16 7 of 9 300 • C for 7 min. The injector temperature was 270 • C and the detector temperature was 220 • C. Biotransformation products were separated by column chromatography using Merck Silica gel 60 (0.040-0.063 mm) with chloroform/methanol (70:6 v/v) as eluent. After evaporation of the eluent, the separated metabolites were weighed. The products' structures were defined in accordance with 1 H NMR and 13 C NMR analysis. The spectra were recorded on a Bruker Avance TM 600 MHz spectrometer in CDCl 3 . Jasco P-2000 digital polarimeter (the version with an iRM controller, Mary's Court, Easton, MD, USA) was used to measure optical rotation. Melting points (uncorrected) were determined on the Boetius apparatus.

Conclusions
Here we reported the relative resistance of oxandrolone (1) to extensive transformations by the tested strains of fungi. The comparison of the transformation results between compound 1 and other steroidal substrates [9,12] indicated that the structure of the A ring of the steroid molecule significantly influences the position of the introduced hydroxyl group, making these processes highly regioselective. The obtained results extend our knowledge of the steroid-transforming activities of microbial fungi. The high substrate conversion rates and achieved product yields indicate that these strains may be promising biocatalysts in the synthesis of new bioactive steroids which might be of interest in academia and clinical medicine in the future.