Microbial Synthesis and Biological Activity of 20β-Hydroxylated Progestins: Ovarian and Neural Action of 17α,20β,21α-Trihydroxy-4-Pregnen-3-One in Danio rerio

Abstract

In this study, the biocatalytic activity of four steroid-transforming strains isolated from the African frog Xenopus laevis and identified as Streptomyces rochei towards pregnane steroids has been investigated. All the isolated strains facilitated the reduction of the C20-carbonyl group and the structures of the metabolites were confirmed by mass spectrometric (MS) and ¹H NMR spectroscopic analyses. Hydrocortisone and progesterone were poorly transformed by the streptomycete strains, whereas cortexolone (Reichstein’s substance S) was effectively biotransformed, yielding more than 90% of 17α,20β,21α-trihydroxy-4-pregnen-3-one (20β-S). Primarily, 20α-reduction was detected when the microbial isolates were incubated with 17α-hydroxyprogesterone with the yield of 17α,20α-dihydroxy-4-pregnen-3-one (17,20α-P) reaching 70%. The biological activity of 20β-S was evaluated in Danio rerio. The results demonstrated that 20β-S modulated stress- and anxiety-related behavioral responses and activated Pgr-dependent transcriptional pathways in the brain and ovarian tissues. These observations support the potential relevance of the synthesized progestin as a functional regulator in teleost physiology. The findings enhance our understanding of the biodiversity of steroid-transforming actinomycetes inhabiting amphibians and can be successfully employed for the effective microbiological synthesis of biologically active 20-hydroxylated progestins that serve as bioregulators in teleosts.

1. Introduction

Actinomycetes (representatives of the phylum Actinomycetota) are widespread in terrestrial and aquatic ecosystems [1] and are known for their ability to produce secondary metabolites, including antibiotics; approximately 70% of all known antibiotics are produced by actinomycetes, mainly representatives of the genus Streptomyces [2,3]. Furthermore, streptomycetes are known to produce many other valuable biomolecules, including biotechnologically important enzymes (lipases, pectinases, cellulases, amylases, proteases) [4], as well as antitumor and immunosuppressive agents [1,5,6].

Actinobacteria are of particular interest in the field of structural modification of steroids [7,8,9]. The ability to catalyze the oxidation of the side chain of sterols resulting in the formation of androstane steroids [7,8,9,10], as well as hydroxylation reactions (at positions 7α, 9α, 6β, 11(α/β), 14α, 16α, etc.) [11,12], dehydrogenation and hydrogenation (for instance, the introduction of a double bond at C1(2) [13,14,15] or its reduction, respectively [16,17], and oxidation of hydroxyl groups (3β-OH, 17β-OH) [18,19] or reduction of keto groups (17C=O, 20C=O) [14,20,21,22] has been demonstrated for various representatives of actinomycetes. It is worth noting that 20β-reduction is a key reaction in the synthesis of progestin 20β-alcohols which can act as bioregulators of reproductive function in teleosts and are in demand by many fisheries worldwide [23].

One of these compounds, 17α,20β-dihydroxy-4-pregnen-3-one (17,20β-P), is considered an important regulator of the final stages of gamete maturation in various teleost fish species [24]. The mechanisms of their action are generally understood, but different studies emphasize various aspects and highlight different links in the signaling pathway. At the same time, most authors agree that the nuclear progesterone receptor (Pgr) is the primary target of these steroids. Through Pgr activation, 17,20β-P affects the expression of several genes, including pla2g4a, ptger4a, and ptger4b, which are sequentially involved in the formation of prostaglandin E₂. Further action of this pathway leads to a series of events culminating in follicle rupture and ovulation in Danio rerio and other teleost fish species [24]. However, some effects of 17,20β-P do not fit the classical genomic model. Several studies have described very rapid physiological responses that are unlikely to be linked to changes in transcription. Such responses are primarily associated with membrane progestin receptors. These receptors are thought to play a role in both the control of oocyte maturation and the regulation of certain aspects of reproductive behavior [25]. The contribution of these membrane mechanisms appears to vary significantly between species and depends on experimental conditions, but their very existence is beyond question.

Structural analogues of 17,20β-P, particularly 17α,20β,21-trihydroxy-4-pregnen-3-one (20β-S), may exhibit comparable biological activity. Such compounds have been shown to stimulate the final stages of oocyte maturation in several teleost fish. This effect has been demonstrated most thoroughly for Danio rerio, although similar effects have also been reported in European eel, catfish, and other species [26]. In male gonads, progestins of the same group are associated with the support of spermatogenesis and the enhancement of sperm motility, probably due to a combination of nuclear signaling pathways and local paracrine effects [27].

In recent years, researchers’ focus has gradually shifted to the action of progestins outside the gonads. Accumulating evidence indicates that the brain of teleost fish contains receptors that respond to these steroids, as well as enzymes that mediate local steroidogenesis. This has led to the development of the concept whereby 20β-reduced progestins can also function as neuroactive steroids. Their effects are most often associated with the stress response, anxiety-like behavior, and motor characteristics. Some studies also note their role in regulating social interaction. However, the nature and strength of the effects can vary significantly depending on the behavioral model used [28,29,30]. Accumulating evidence suggests that the key mechanism here may involve Pgr-dependent regulation of the hypothalamic–pituitary adrenal axis and local neurosteroid production in the brain.

All these observations indicate that 20β-reduced progestins hold a unique position at the intersection of reproductive and behavioral regulation. Consequently, 20β-S is potentially capable of impacting both ovarian Pgr-dependent prostaglandin pathways and central mechanisms underlying stress- and anxiety-like behavior in Danio rerio.

The chemical synthesis of 20β-reduced progestins is complex, multi-stage, and often accompanied by the formation of impurity steroids. Concurrently, the possibility of their microbiological synthesis remains largely unexplored, limited to only a few studies that demonstrated low selectivity of the target reaction and the use of low concentrations of steroid substrates [31,32,33]. This scenario underscores the necessity to search for new microorganisms that can catalyze the 20β-reduction reaction of pregnane steroids facilitating the selective accumulation of valuable 20β-reduced progestin metabolites.

Four bacterial cultures (St1–St4) were previously isolated from the dermal mucosa of the African clawed frog Xenopus laevis. Microscopic and morphological studies allowed us to preliminarily classify the microbial isolates as actinomycetes, specifically streptomycetes, due to their ability to form highly differentiated mycelium (Figure 1). Based on 16S rRNA gene sequencing, full-genome sequencing, DDH analysis, and ANI, the isolates were classified as Streptomyces rochei species [34].

Figure 1.

Phase-contrast micrographs (24 h), ×200 (optical microscopy) (a) and colony morphology (12 days) (b) of S. rochei MTOC St1–St4 isolates.

In this study, the biocatalytic activity of the isolated strains of Streptomyces rochei towards pregnane steroids was investigated, including cortexolone (Reichstein’s substance S) and 17α-hydroxyprogesterone, and the main metabolites were identified. The isolates were found to exhibit high biocatalytic activity in reducing the C20-carbonyl group with one isolate efficiently producing 17α,20β,21α-trihydroxy-4-pregnen-3-one (20β-S) as a major metabolite. The effects of microbiologically produced 20β-S on zebrafish were studied with particular attention paid to embryotoxicity, stress- and anxiety-related behavioral changes in adult fish, and the expression of Pgr-dependent genes in the brain and gonadal tissues.

2. Materials and Methods

2.1. Chemicals

Steroids: 17α,21-dihydroxypregn-4-ene-3,20-dione (syn. cortexolone, Reichstein’s substance S), 11β,17α,21-trihydroxypregn-4-ene-3,20-dione (syn. hydrocortisone, compound F by Edward Kendall), pregn-4-ene-3,20-dione (progesterone), 17α-hydroxypregn-4-ene-3,20-dione (17α-hydroxyprogesterone, 17α-OH-progesterone) were purchased from Symbiotech (Pigdamber, Rau, India). Yeast extract was purchased from Difco (Becton Drive Franklin Lakes, NJ, USA); corn extract from Sigma-Aldrich (St. Louis, MO, USA). Other reagents were of the best purity grade from domestic commercial suppliers.

2.2. Microorganisms

The strains of Streptomyces rochei MTOC St1–St4 were previously isolated from the dermal mucosa of African clawed frog Xenopus laevis following the described protocol [35] and were kindly provided by Dr. Suzina (Pushchino Center for Biological Research of the Russian Academy of Sciences). The strains were maintained on starch-malt extract-yeast extract (SMY) agar medium: potato starch—10 g/L, malt extract—10 g/L, yeast extract—4 g/L, agar—15 g/L, pH 7.0.

2.3. Cultivation and Steroid Bioconversion

The spore suspension of Streptomyces rochei isolates from one SMY agar slant was inoculated into 50 mL of liquid SMY medium (without adding agar) and incubated in a 750 mL Erlenmeyer flask aerobically on a rotary shaker (200 rpm) at 28 °C for 24 h. An aliquot (10% v/v) of the first-generation cells was then transferred to 45 mL of fresh liquid SMY medium. The steroid substrate was added to a final concentration of 0.1 g/L as a hot dimethyl sulfoxide (DMSO) solution (the final DMSO concentration in the medium did not exceed 1% v/v). Bioconversion was carried out for 24–96 h under the conditions described above.

2.4. Thin Layer Chromatography (TLC)

The samples of cultivation broth (5 mL) were extracted with 2 mL of ethyl acetate EtOAc. The extracts (20 µL) were applied to ALUGRAM SIL G/UV254 TLC sheets (Macherey-Nagel GmbH & Co. KG, Düren, Germany) and benzene–acetone (3:1.5, v/v) was used as a mobile phase. To visualize steroids, UV detection (254 nm) with CN-15MC UV Darkroom (Vilber Lourmat, Torcy Z.I. Sud Marne-la-Vallée, France) was used.

2.5. High-Performance Liquid Chromatography (HPLC)

The analyses were performed on the Agilent 1200 instrument (Agilent Technologies, Waldbronn, Germany) and Symmetry C18 column (5 µm, 4.6 × 250 mm) with Symmetry C18 precolumn (5 µm, 3.9 × 20 mm) (Waters, Milford, MA, USA). Detection was carried out by ultraviolet absorption at λ = 240 nm or 254 nm. Eluent (mobile phase): acetonitrile (MeCN):H₂O:trifluoroacetic acid (TFA) in a ratio of 40:60:0.02 (v/v) (isocratic) (for cortexolone and its derivatives); 22% tetrahydrofuran (THF) aqueous solution (isocratic) (for hydrocortisone and its derivatives); MeCN:H₂O:TFA in a ratio of 40:60:0.02 (A) and MeCN:TFA in a ratio of 100:0.02 (B) (gradient: 0–2′—A 100%; 2–15′—A 100 ≥ 50%; 15–17′—A 50%) for progesterone, 17α-OH-progesterone, and their derivatives. Flow rate—1 mL/min; column temperature—50 °C; concentrations were calculated using peak area values. Retention time (R_t): cortexolone, 8.53 min; 20β-OH-cortexolone (20β-S), 6.03 min; hydrocortisone, 14.53 min; 20β-OH-hydrocortisone (20β-F), 8.89 min; progesterone (Pr), 17.26 min; 20β-OH-progesterone (20β-OH-Pr), 17.32 min; 17α-OH-progesterone (17α-OH-Pr), 12.43 min; 17α,20β-diOH-progesterone (17α,20β-diOH-Pr), 9.86 min; 17α,20α-diOH-progesterone (17α,20α-diOH-Pr), 11.08 min.

2.6. Isolation of Steroids

Steroids were extracted from the culture broth (50 mL) with three volumes of EtOAc. The solvent was evaporated under vacuum, obtaining the crude residue.

Individual steroid metabolites were obtained by preparative column chromatography as described previously [36] and then analyzed by mass spectrometer and ¹H NMR methods.

2.7. Mass Spectrometry (MS), 1H NMR Spectroscopy

MS spectra of steroid metabolites were recorded with a Bruker Maxis Impact spectrometer. ¹H and ¹³C NMR spectra were recorded at 400 MHz with a Bruker Avance 400 spectrometer (Bruker, Billerica, MA, USA). Chemical shifts were measured relative to a solvent signal.

2.8. Preparation of the Crystalline Product of 20β-OH-Cortexolone

Steroids from the 48-h fermentation broth (200 mL) of cortexolone (0.5 g/L) transformation using the most active S. rochei MTOC St3 isolate were extracted with a double volume of EtOAc, which was then evaporated under vacuum. The resulting oily crude residue was washed with approximately 5 mL of hexane after which 2 mL of hot EtOAc was added and mixed until the crude residue was completely dissolved. The extract was left at room temperature for several hours until precipitation occurred. The precipitate was separated from the solvent and washed twice with 5 mL of EtOAc. The washed precipitate was dissolved in 2 mL of ethanol and left at room temperature. After a few days, a crystalline precipitate was observed. The residual ethanol was separated from the crystalline precipitate, which was then washed again with 1.5 mL of ethanol. The washed crystalline precipitate was dried at 60 °C for 72 h, yielding 50 mg of the product.

2.9. Animals

Adult zebrafish (Danio rerio, 3–6 months old, mixed sex) were maintained in a recirculating aquarium system (28 ± 0.5 °C; 14 h light/10 h dark photoperiod; dissolved oxygen > 6 mg L⁻¹; pH 7.0–7.5; conductivity 450–500 µS cm⁻¹). The fish were fed twice daily with Artemia nauplii and a commercial flake diet (TetraMin, Melle, Germany). Prior to the experiments, fish were acclimatized for 7 days in 3 L glass tanks (10 fish per tank).

2.10. Test Compound

Highly purified 20β-OH-cortexolone (20β-S) was dissolved in 96% ethanol to obtain a stock solution which was subsequently diluted with saline to the required nominal concentrations. The final ethanol concentration in all test solutions did not exceed 0.05% (v/v). Control groups received only the vehicle.

To conduct tests on embryos, a solution of 20β-S of the required concentration was prepared in advance and added to each well of the incubation plate. In experiments on adult fish, 20β-S was administered intraperitoneally (i.p.) at doses of 0 (control), 0.1, 1, and 10 mg/kg body weight (n = 10 fish per group). Before injection, the fish were briefly anaesthetized by immersion in 10 °C water for no longer than 20–30 s.

2.11. Embryotoxicity and Teratogenicity Assay

Embryos (<6 h postfertilization) were distributed into six-well plates (25 embryos per well, 5 mL solution) and exposed to 0 (vehicle), 0.1, 1, or 10 mg L⁻¹ 20β-S under static-renewal conditions at 28 °C for 96 h. Mortality and developmental endpoints (egg coagulation, somite formation, tail detachment, curvature, heartbeat, hatching time, yolk-sac resorption) were evaluated by stereomicroscopy.

2.12. Behavioral Assays in Danio rerio

Behavioral testing began 4 h after i.p. injection of the compound. The testing sequence started with the novel tank diving test (NTT), followed immediately by the light–dark box test (LDB). The behavioral apparatus was manufactured by Open Science LCC (Krasnogorsk, Russia). All tests were performed during the light phase.

The NTT was conducted following Maximino et al. (2011) [37] with minor modifications. Individual fish were placed in a novel tank (29 × 9 × 19 cm; 3 L water; light intensity ~400 lx) and recorded for 5 min. Behavior was video recorded, and the data were analyzed using EthoVision XT15 (Noldus, Wageningen, The Netherlands).

The following parameters were quantified: total distance traveled, swimming speed, number of visits to the bottom, middle, and upper thirds of the tank, and the time spent in each of these zones during the 5-min session. Reduced time spent in the upper part of the tank is interpreted as decreased exploratory activity or increased avoidance behavior [38].

The LDB test was based on the protocol of Maximino et al. (2011) [37]. A rectangular tank (45 × 10 × 15 cm, light intensity ~400 lx) was divided into light and dark compartments by an opaque partition. Fish were introduced into a neutral central zone for 1 min and were then allowed to explore freely for 5 min. Behavior was video recorded and analyzed using RealTimer software v.1.30 (Open Science LCC, Krasnogorsk, Russia).

An increased duration spent in the dark compartment (scototaxis) is considered an indicator of elevated anxiety levels in zebrafish [37].

Immediately after behavioral testing, fish were euthanized by rapid chilling (≤4 °C) and subsequently decapitated. The brain was dissected from the head for subsequent gene expression analysis, while the remaining body was used for whole-body cortisol determination.

2.13. Sample Preparation and Cortisol Measurement

Whole-body cortisol was extracted using a modified protocol based on [39]. Whole-body samples (excluding the head) were thawed on ice and weighed individually. Each sample was cut into small pieces with a scalpel and transferred into 1 mL of ice-cold phosphate-buffered saline (PBS, 25 mM). The tissue was homogenized for 30–60 s and kept on ice throughout the procedure.

Diethyl ether (5 mL) was added to each homogenate, followed by vigorous vortexing and centrifugation at 2500× g for 15 min. The organic phase, containing the extracted cortisol, was carefully transferred into a separate glass tube. The extraction step with ether, including vortexing and centrifugation, was repeated two additional times to maximize cortisol recovery.

The collected organic phases were left to evaporate completely under a fume hood at room temperature until only a thin yellow residue remained.

After evaporation, 1 mL of ice-cold PBS was added to each tube to reconstitute the extracted cortisol. Cortisol concentrations were quantified using a competitive ELISA kit (Cayman Chemical, Ann Arbor, MI, USA) according to the manufacturer’s instructions. Cross-reactivity between 20β-S and the cortisol antibody was assessed and found to be negligible.

2.14. Gene Expression

For gene expression analysis, brain samples were collected immediately after behavioral testing. Ovarian tissue was obtained in a separate experiment. Females received an intraperitoneal injection of the test compound (single dose of 1 mg/kg, n = 5 in each group). Four hours after administration, the fish were euthanized, and the ovaries were carefully dissected using fine scissors and a scalpel.

Total RNA was extracted from zebrafish brain and ovarian tissues using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. Approximately 20–30 mg of tissue was rinsed in ice-cold phosphate-buffered saline (PBS, pH 7.4), transferred to a sterile microcentrifuge tube containing 1 mL of TRIzol, and homogenized on ice with a disposable RNase-free pestle until complete disruption.

cDNA was synthesized from 1 µg of total RNA using M-MuLV reverse transcriptase and oligo-dT primers from the MMLV RT Kit (Evrogen, Moscow, Russia). Quantitative RT-PCR was performed on an Applied Biosystems 7500 real-time PCR amplifier using a reaction mix containing Rox and SYBR Green (Evrogen, Moscow, Russia) and specific primers designed to amplify fragments of zebrafish β-actin, pla2g4a, ptger4a, and ptger4b genes (Table S1). The PCR program included a preheating stage at 95 °C for 10 min and 40 cycles of amplification at 95 °C for 15 s, 62 °C for 20 s, and 72 °C for 20 s. Melting curve analysis was performed to assess amplification specificity. Expression levels were normalized to β-actin.

2.15. Statistical Data Processing

All experiments and measurements prior to the in vivo assays were performed in triplicate, and standard deviations for each sample are presented as error bars in the corresponding graphs. Statistical analyses of behavioral and physiological data obtained from Danio rerio were conducted using STATISTICA 12 (TIBCO Software, Palo Alto, CA, USA). The normality of data distribution was assessed with the Shapiro–Wilk test. For normally distributed datasets, one-way ANOVA followed by Fisher’s LSD post hoc test was applied. For datasets that did not meet normality assumptions, a Mann–Whitney U test or the Kruskal–Wallis test followed by Dunn’s post hoc comparisons was used.

3. Results and Discussion

3.1. Bioconversion of Steroid Substrates by Streptomyces Rochei Isolates

Biotransformations of cortexolone, hydrocortisone, progesterone, and 17α-hydroxyprogesterone were conducted to assess the biocatalytic potential of St1–St4 isolates in the structural modification of pregnane steroids.

3.1.1. Transformation of Cortexolone

Incubation of S. rochei MTOC St1–St4 isolates with cortexolone led to the significant accumulation of one major metabolite (I) (over 90% w/w after 24 h of transformation in all variants) (Figure 2).

Figure 2.

Time course of cortexolone bioconversion by the S. rochei MTOC St1 (a), St2 (b), St3 (c), and St4 (d) isolates; TLC profile of the culture broth samples (24 h) (~10 mkg of steroids in the spot; S, a mixture of reference steroids in the spot (from top to bottom)): cortexolone (3 mkg), hydrocortisone (3 mkg) (e).

According to HPLC analysis, the maximum yield of metabolite I (94.28% w/w) was noted for the St3 isolate (Figure 2c). It is important to mention that, during further incubation (48–72 h), the yield of metabolite I (93–95% w/w) and the amount of steroid substrate cortexolone (4–6% w/w) remained unchanged across all the studied variants, indicating the absence of any steroid-transforming activity in bacterial cells towards metabolite I accumulated within the first day of bioconversion (Figure 2a–d).

Mass spectrometric analysis (MS) of metabolite I enabled us to determine its molecular weight (M 348), which increased by 2 units compared to the steroid substrate cortexolone (M 346), suggesting a possible reduction of the C=O group (Figure S1a).

Based on ¹H and ¹³C NMR analysis metabolite I was identified as 17α,20β,21-trihydroxypregne-4-en-3-one (20β-reduced derivative of cortexolone, 20β-S) (Figure S2a), further corroborating the MS data that indicated the possible presence of a reduced keto group in the structure of steroid metabolite I.

The results obtained indicate the ability of the studied strains to selectively catalyze the 20β-reduction reaction towards cortexolone, resulting in the formation of 20β-S as the major product (Figure 3a).

Figure 3.

Schemes of cortexolone (a), hydrocortisone (b), progesterone (c), and 17α-OH-progesterone (d) bioconversion by the S. rochei MTOC St1–St4 isolates.

One of the first identified actinomycete 20β-HSD or 3α,20β-HSD enzymes was a short-chain dehydrogenase isolated from Streptomyces hydrogenans, capable of catalyzing reversible oxidation/reduction reactions at positions 3α and 20β of the pregnane and androstane steroids [40]. The ability to catalyze 20β-reduction towards prednisolone has been demonstrated for Streptomyces roseochromogenes TS79, previously known as a biocatalyst for the 16α-hydroxylation reaction [22]. Optimal conditions for the transformation of prednisolone into 20β-hydroxyprednisolone were established; however, the activity of the strain towards other steroid substrates, including cortexolone, has not been examined. The bacterial strain Bacillus megaterium was able to catalyze the 20β-reduction of cortexolone; however, the yield of 20β-S was significantly lower compared to that obtained in this study using S. rochei MTOC St1–St4 isolates, which almost exclusively produced this product (over 95%). Notably, in the case of Bacillus megaterium, the primary product of cortexolone transformation was identified as 20α-S and to obtain 20β-S, knockouts of four genes associated with the synthesis of enzymes responsible for 20α-reduction of cortexolone were utilized [31].

3.1.2. Transformation of Hydrocortisone

Extremely low biocatalytic activity of the studied bacterial isolates was observed when incubating with hydrocortisone (11β-hydroxylated derivative of cortexolone) (Figure 4). After 24 h of bioconversion, only minor amounts (3–6%) of one product (metabolite II) were detected in the culture broth samples. The following day, the yield of metabolite II increased to varying degrees depending on the strain used: the maximum accumulation of the product (up to 18.3% w/w) was observed in the variant with S. rochei MTOC St4 (Figure 4d) while the lowest yield (less than 7% w/w) was recorded with St3 isolate (Figure 4c). Subsequent incubation (72–96 h) did not result in any change in the content of the residual steroid substrate hydrocortisone and metabolite II across all tested variants (Figure 4a–d).

Figure 4.

Time course of hydrocortisone bioconversion by the S. rochei MTOC St1 (a), St2 (b), St3 (c), and St4 (d) isolates; TLC profile of the culture broth samples (48 h) (~10 mkg of steroids in the spot; S, a mixture of reference steroids in the spot (from top to bottom)): cortexolone (3 mkg), hydrocortisone (3 mkg) (e).

Based on the MS analysis, the molecular weight of metabolite II was determined as M 364, which was 2 units greater than that of the steroid substrate hydrocortisone (M 362), thus indicating a possible reduction of the carbonyl group (Figure S1b).

¹H NMR analysis allowed us to identify the metabolite II as 11β,17α,20β,21-tetrahydroxypregn-4-en-3-one (20β-reduced derivative of hydrocortisone, 20β-F) (Figure S2b).

It is noteworthy that the lowest yield of 20β-F (<7%) was detected in the variant with St3 isolate (Figure 4c), which at the same time was most active towards cortexolone providing maximum accumulation of 20β-S (>94%) (Figure 2c). Obviously, the presence of an additional hydroxyl group at the C11-Hβ position in the structure of the hydrocortisone in comparison with cortexolone hinders binding of the steroid substrate molecule to the active center of the steroid-transforming enzyme catalyzing the reaction of 20β-reduction (Figure 3b).

Similar results were observed when studying the functional activity of two identified frog 20β-HSD enzymes: Hsd20b2.L and Hsd20b2.S, which were shown to actively catalyze the 20β-reduction of C21 steroids such as cortexolone, progesterone, and 17α-OH-progesterone but exhibited extremely weak activity towards hydrocortisone [41].

Among bacteria, the ability to catalyze 20β-reduction towards hydrocortisone has previously been demonstrated for certain intestinal anaerobic bifidobacteria, including Butyricicoccus desmolans, Clostridium cadaveris, and Bifidobacterium adolescentis. These microorganisms were shown to synthesize 20β-HSD, which is involved in the bacterial steroid 17,20-desmolase pathway of cortisol metabolism to proandrogens, the formation of which may, in turn, influence the physiology of the host organism [32,33]. The ability to reduce the 20-keto group of cortisone and hydrocortisone was also noted for several representatives of the genus Rhodococcus, which also actively catalyzed the Δ1-dehydrogenation reaction of steroid substrates [14]. The synthesis of small amounts of 20β-reduced derivatives was observed during the transformation of hydrocortisone by Nocardioides simplex, resulting in the undesirable formation of 20β-alcohols during the production of prednisolone [42].

3.1.3. Transformation of Progesterone

Bioconversion of progesterone by the isolates St1–St4 proceeded with low efficiency, providing the accumulation of one product (metabolite III) (from 12 to 25%) while over 70% of progesterone remained unconverted even after 48 h of incubation (Figure 5a–d). The maximum yield of derivative III was observed in the variant with S. rochei MTOC St1 (up to 25% by 48 h) (Figure 5a).

Figure 5.

Time course of progesterone bioconversion by the S. rochei MTOC St1 (a), St2 (b), St3 (c), and St4 (d) isolates; TLC profile of the culture broth samples (48 h) (~10 mkg of steroids in the spot; S, a mixture of reference steroids in the spot (from top to bottom)): progesterone (3 mkg), 17α-OH-progesterone (3 mkg) (e).

Molecular weight (M 316) of metabolite III was 2 units higher than that of the steroid substrate progesterone (M 314), indicating a possible reduction of the carbonyl group (Figure S1c).

Based on ¹H NMR analysis metabolite III was identified as 20β-hydroxypregn-4-en-3-one (20β-reduced derivative of progesterone, 20β-Pr) (Figure S2c).

It should be noted that, during further incubation (72 h), the amount of accumulated metabolite III and the residual progesterone in the culture broth remained unchanged in all the studied variants (Figure 5a–d).

The structural features of the progesterone, which differs from cortexolone by the presence of a methyl group at the C21 position instead of a hydroxyl group, as well as the absence of a hydroxyl group at the C17α position, seem to be the reason for a decrease in accessibility of the steroid substrate molecule for binding to the active center of the enzyme that catalyzes the reduction of the C20 carbonyl group in the studied bacterial cultures (Figure 3c).

Predominantly hydroxylase activity has been shown among the representatives of the Streptomyces genus towards progesterone, while the ability to catalyze 20β-reduction of steroid substrate has not been mentioned. Thus, Streptomyces roseochromogenes NCIB 10984 catalyzed the reaction of 16α-hydroxylation to form 16α-hydroxyprogesterone, which was additionally hydroxylated at position 2β to form the corresponding dihydroxylated derivative 2β,16α-dihydroxyprogesterone [43]. A comprehensive study of progesterone bioconversion by various representatives of the genus Streptomyces made it possible to divide the tested isolates into several groups according to the type of reaction being catalyzed, among which hydroxylation at positions 6β, 11α, and 16α and the C1-C2 dehydrogenation reaction were identified [44].

3.1.4. Transformation of 17α-OH-Progesterone

More intense biotransformation of the steroid substrate was observed while incubating S. rochei isolates with 17α-OH-progesterone. Notably, the formation of two metabolites (IV and V) was revealed during the bioconversion (Figure 6). S. rochei MTOC St1 and St3 were shown to be the most active in accumulation of these products, while the maximum yield of metabolites IV and V (totally more than 84% w/w) was detected in the variant with St1 isolate after 48 h of biotransformation (Figure 6a,e).

Figure 6.

Time course of 17α-OH-progesterone bioconversion by the S. rochei MTOC St1 (a), St2 (b), St3 (c), and St4 (d) isolates; TLC profile of the culture broth samples (48 h) (~10 mkg of steroids in the spot; S, a mixture of reference steroids in the spot (from top to bottom)): progesterone (3 mkg), 17α-OH-progesterone (3 mkg) (e).

MS analysis of metabolites IV and V revealed the same molecular weight (M 332) for both compounds, which was 2 units higher than that of the steroid substrate 17α-OH-progesterone (M 330), indicating, as in previous cases, the possible presence of a reduced carbonyl group in the structure of the formed metabolites (Figure S1d,e).

Based on ¹H NMR analysis metabolite IV was identified as 17α,20β-dihydroxypregne-4-en-3-one (20β-reduced derivative of 17α-OH-progesterone, 17α,20β-Pr) (Figure S2d). At the same time, metabolite V based on ¹H NMR analysis was identified as 17α,20α-dihydroxypregne-4-en-3-one (20α-reduced derivative of 17α-OH-progesterone, 17α,20α-Pr) (Figure S2e).

It should be noted that there were some differences in the dynamics of the accumulation of metabolites IV and V by the studied microbial isolates: the amount of metabolite IV (17α20β-Pr) accumulated by the first day of transformation in the variants with St1, St3, and St4 isolates slightly decreased (by 4–6% w/w) during the subsequent incubation (48 h) (Figure 6a,c,d) or increased (by 8–10% w/w) in the variant with St2 isolate (Figure 6b). On the other hand, the amount of metabolite V (17α,20α-Pr) accumulated by isolates St1, St3, and St4 during 24 h of transformation was significantly increased (approximately 2-fold) during the following incubation day: from 25–35% to 55–75% in the variants with St1 and St3 isolates (Figure 6a,c), respectively, or from 12% to 22% in the variant with the St4 strain (Figure 6d). It should be noted that, in the case of the St2 isolate, the amount of 17α,20α-Pr, in contrast, decreased from 30 to 20% (w/w) by 48 h of incubation (Figure 6b).

The obtained results also suggest the presence, in the studied microbial isolates, in addition to the enzyme catalyzing the reaction of 20β-reduction, the enzyme responsible for the 20α-reduction, whose activity is defined only towards the molecule of 17α-OH-progesterone and not other pregnane steroid substrates, namely cortexolone, hydrocortisone, and progesterone. It is also impossible to exclude the possibility of catalyzing the reactions of 20β- and 20α-reduction by a single enzyme. In such a case, the combination of hydroxyl groups at C21 and C17α positions in the structure of 17α-OH-progesterone seems to be a key factor contributing to an increase in the availability of the steroid substrate molecule for predominantly 20α-reduction (Figure 3d).

In any case, it becomes obvious that the presence of a hydroxyl group at C17α in the structure of pregnane substrates promotes the binding of the steroid molecule to the active center of the enzyme(s) that catalyze the reduction of the C20 carbonyl group in the cells of the studied streptomyces isolates.

It should be noted that the most active 20α-reduction of 17α-OH-progesterone was detected for the isolate S. rochei MTOC St1, which produced over 80% of the 20α/β-reduced isomers, with the major one (more than 70% w/w) identified as 17α,20α-diOH-progesterone (Figure 6a).

Similar activity was previously observed for Bacillus megaterium, which was able to catalyze both 20α- and 20β-reductase activity towards 17α-OH-progesterone, with accumulation of 17α,20α-diOH-progesterone as the main bioconversion product. In that case, it was proved that the 20α- and 20β-reduction reactions are catalyzed by different enzymes: four enzymes were identified as 20αHSDs, whereas the 3-oxoacyl-(acyl-carrier-protein) reductase FabG was found to exhibit 20βHSD activity [31]. 20α-Hydroxysteroid dehydrogenase was also revealed in the human gut microbe Clostridium scindens with activity tightly regulated along with steroid-17,20-desmolase by the presence of host steroids cortisol and cortisone [45,46].

The results of the bioconversion of pregnane steroids by S. rochei MTOC St1–St4 isolates, including the data on retention factor (R_F) and relative retention time (R_T) from TLC and HPLC analysis, respectively, of steroid metabolite products I–V, along with their structures determined through MS and ¹H NMR methods are summarized in Table 1.

Table 1.

The results of the bioconversion of pregnane steroids by S. rochei MTOC St1–St4 isolates.

Steroid Substrate/Product	RF (TLC)	RT (HPLC)	Structure (MS/NMR)	Molecular Weight/NMR Data
Cortexolone Metabolite I (20β-S)	0.46 0.19	8.53 6.03		(M 346) (M 348)/1H-NMR (CDCl3) δ: 5.73 (s, 1H, H-4), 3.85–3.70 (m, 3H, H-21, H-20α), 1.19 (s, 3H, H-19), 0.85 (s, 3H, H-18). 13C-NMR (CDCl3) δ: 199.9 (C-3), 171.7 (C-5), 123.7 (C-4), 85.2 (C-17), 73.2 (C-20), 64.7 (C-21), 53.4, 49.0, 47.4, 38.6, 35.7, 35.6, 33.9, 33.2, 32.9, 32.0, 31.7, 23.8, 20.6, 17.3 (C-19), 15.1 (C-18)
Hydrocortisone Metabolite II (20β-F)	0.23 0.05	14.52 8.88		(M 362) (M 364)/1H NMR (CD3OD) δ: 5.64 (s, 1H, H-4), 4.35 (m, 1H, H-11α), 3.65–3.57 (m, 3H, H-20α, H-21) 1.47 (s, 3H, H-19), 1.09 (s, 3H, H-19)
Progesterone Metabolite III (20β-Pr)	0.80 0.63	17.26 17.32		(M 314) M 316)/1H NMR (CD3OD) δ: 5.70 (s, 1H, H-4), 3.65 (q, J = 6.1 Hz), 1.23 (s, 3H, H-19), 1.10 (d, J = 6.1 Hz, 3H, H-21), 0.81 (s, 3H, H-19)
17α-OH-progesterone Metabolite IV (17α,20β-Pr) Metabolite V (17α,20α-Pr)	0.64 0.47 0.39	12.43 11.10 9.89		(M 330) (M 332)/1H NMR (CD3OD) δ: 5.70 (s, 1H, H-4), 3.93 (q, J = 6.3 Hz, 1H, H-20), 1.23 (s, 3H, H-19), 1.13 (d, J = 6.3 Hz, 3H, H-21), 0.87 (s, 3H, H-19) (M 332)/1H NMR (CD3OD) δ: 5.70 (s, 1H, H-4), 3.78 (q, J = 6.3 Hz, 1H, H-20), 1.23 (s, 3H, H-19), 1.16 (d, J = 6.3 Hz, 3H, H-21), 0.79 (s, 3H, H-19)

3.2. Candidate Genes Related to 20-Carbonyl Reduction

Earlier, we performed a genome sequence of the Streptomyces rochei MTOC St3 strain (the draft genome assembly has been deposited in GenBank under the accession no. JBIENY000000000.1) [34].

As is known from the literature, 20-ketoreduction of steroids in certain bacteria, such as Bacillus megaterium [31] and Mycobacterium tuberculosis [47], can be catalyzed by either 3-oxoacyl-(acyl-carrier-protein) reductase FabG or by 3α,20β-hydroxysteroid dehydrogen-ase (3α,20β-HSD), as observed in Streptomyces hydrogenanse [39]. In this context, to search for candidate genes in the assembled genome of Streptomyces rochei MTOC St3 that may be linked to the reduction of 20-ketoreduction, the known genes encoding FabG and 3α,20β-HSD from various microorganisms were used as reference sequences.

Consequently, three fabG genes (locus_tags “ACGU38_15900,” “ACGU38_19100,” and “ACGU38_30040”) and one glucose-1-dehydrogenase gene (locus_tag “AC-GU38_08065”) were identified in the St3 genome, which may encode enzymes related to the 20-reduction of pregnane steroids. Our forthcoming studies will concentrate on the functional identification of the revealed candidate genes. Additionally, transcriptomic analysis will be conducted to identify the differentially expressed genes in response to pregnane steroids.

3.3. Biological Activity of Microbiologically Synthesized Progestin 20β-S in Danio rerio

The highly purified crystalline product of 20β-S was obtained from the fermentation broth of cortexolone bioconversion using the most active St3 isolate (refer to the Section 2) (Figure S3a). According to HPLC analysis, the purity of the crystalline product exceeded 98% (Figure S3b). The biological activity of the purified 20β-S was then evaluated in Danio rerio.

3.3.1. Embryotoxicity

During the analysis of the effect of microbiologically obtained 20β-S on the early development of Danio rerio, a dose-dependent effect was noted (Figure 7g) Specifically, incubation in a solution with a concentration of up to 1 mg/L did not result in any deviations from the norm: no signs of teratogenicity, developmental delays, or obvious morphological abnormalities were observed. However, at a higher concentration (10 mg/L), embryo mortality sharply increased. The first signs of embryonic lethality, including the absence of a heartbeat and cessation of blood circulation, were noted approximately 48 h after fertilization and persisted for up to 72 h, until the last larva died.

Figure 7.

External morphology of Danio rerio embryos and larvae at representative developmental stages: 6 hpf (a); 24 hpf (b); 48 hpf (c); hatching stage (d); 72 hpf (e); 96 hpf (f). Mortality of embryos/larvae after incubation in 20β-S solution; n = 25 per group at the start of the experiment (g).

Interestingly, embryos that survived the exposure showed no obvious abnormalities in body proportions or organ structure, indicating that the increased mortality was not accompanied by discernible structural defects.

3.3.2. Behavioral Responses (Adult Fish)

In experiments involving adult zebrafish, the effect of 20β-S on behavior was assessed using standard tests for anxiety-like reactions. After a single intraperitoneal administration of 20β-S at a dose of 1 mg/kg, the fish exhibited significant behavioral changes in the novel tank test: specifically, an increase in the total time spent in the upper water layers compared to the control group (Figure 8c). A similar pattern was observed in the light–dark box test, where 20β-S-treated fish made more transitions between the light and dark compartments and spent more time in the illuminated zone (Figure 9c,d). Notably, at this dosage, no reduction in general locomotor activity was revealed, allowing us to interpret the changes as an anxiolytic-like effect rather than sedation. However, at 10 mg/kg, the behavior differed: the fish demonstrated reduced overall activity and remained motionless for longer periods. These changes may indicate the emergence of a sedative component at higher doses (Figure 8d). Dose–response curves of this nature have been reported for several neurosteroids in teleosts and mammalian systems [28]. In teleost models, rapid neuromodulatory effects of brain-derived steroids have been documented [29]. Overall, the behaviors recorded in the tests suggest that 20β-S acts in a dose-dependent manner, with the most pronounced shifts appearing in anxiety-related readouts.

Figure 8.

Novel tank diving test (NTT). External view of the experimental setup (a); Zones used for behavioral analysis (b); Time spent in the upper zone, s [F(3,36) = 3.397, p = 0.0281] (c); Total distance moved, cm [F(3,36) = 3.002, p = 0.0431]; * p < 0.05 vs. vehicle group (one-way ANOVA with post hoc Fisher’s LSD test) (d). All values are expressed as mean ± SEM.

Figure 9.

Light–dark box test (LDB). External view of the experimental setup (a); Light zone used for behavioral analysis (b); Time spent in the light zone, s [F(3,36) = 3.306, p = 0.0309] (c); The number of transitions between the light and dark compartments [F(3,36) = 6.294, p = 0.00152]; * p < 0.05 vs. vehicle group (one-way ANOVA with post hoc Fisher’s LSD test) (d). All values are expressed as mean ± SEM.

3.3.3. Cortisol Level

The analysis of whole-body cortisol content in zebrafish following administration of 20β-S revealed that the hormonal response also changes in accordance with the behavioral data. Fish exposed to 1 mg/kg 20β-S exhibited lower cortisol concentrations than the vehicle controls (Figure 10). This decrease in cortisol levels suggests that, during periods of stress, activation of the hypothalamic–pituitary–adrenal axis may have been diminished under these conditions. The HPA system’s response to the stress stimulus was partially suppressed, which aligns with behavioral data. Several studies have indicated that other endogenous progestins and neuroactive steroids can reduce stress-induced activation of the HPA system and consequently mitigate both physiological and behavioral manifestations of the stress response [25,30]. Therefore, 20β-S demonstrates characteristics of a compound that not only alters behavior but also diminishes physiological stress reactions at the endocrine level.

Figure 10.

Cortisol level, ng/mg, total homogenate [F(3,36) = 3.754, p = 0.0192]; * p < 0.05 vs. vehicle group (one-way ANOVA with post hoc Fisher’s LSD test). All values are expressed as mean ± SEM.

3.3.4. Transcriptional Regulation of Target Genes

To evaluate potential molecular mechanisms underlying the observed effects, the expression of several Pgr-dependent genes was analyzed in the brain and ovaries of zebrafish after 20β-S administration. According to qRT-PCR data, an increase in mRNA levels of ptger4a and ptger4b was observed in both brain and ovarian tissues (Figure 11a,b and Figure 12a,b), while upregulation of pla2g4a was detected only in the ovary (Figure 12c). This combination of changes corresponds to the current concept of the Pgr-PGE-EP4 signaling pathway, which plays a crucial role in the regulation of ovulation and related processes in teleosts [24]. At the same time, the activation of ptger4a/b in the brain further supports the notion of a functional interaction between neuroendocrine and reproductive regulatory axes in fish. Overall, the transcriptional data obtained indirectly confirm that 20β-S is capable of acting through Pgr-dependent pathways in both ovarian and neural tissues.

Figure 11.

mRNA levels of ptger4a (a), ptger4b (b), and pla2g4a (c) genes in brain tissue homogenates. n = 10 per group. * p < 0.05 vs. vehicle group (Kruskal–Wallis test followed by Dunn’s post hoc comparisons). Data represent the results of three independent experiments.

Figure 12.

mRNA levels of ptger4a (a), ptger4b (b), and pla2g4a (c) genes in ovarian tissue homogenates. n = 5 per group. For this experiment, the most effective dose identified in our previous studies: 1 mg/kg. * p < 0.05 vs. vehicle group (Mann–Whitney U test) was selected. Data represent the results of three independent experiments.

3.3.5. Mechanistic Considerations

Taking into account the steroid nature and sufficient lipophilicity of 20β-S, it is reasonable to assume that this compound can cross the blood–brain barrier, as has been shown for several other neurosteroids, including those that are effective at very low concentrations [48,49]. In this context, it becomes possible to explain the simultaneous influence of 20β-S on behavior, cortisol levels, and expression of Pgr-dependent genes in the brain. On the one hand, in the ovary, 20β-S likely activates the classical Pgr-pla2g4a-PGE-ptger4 pathway, which is involved in the final stages of oocyte maturation and ovulation. On the other hand, in the brain, binding to Pgr and, potentially, to other steroid-sensitive targets may lead to modulation of stress-associated signaling pathways, resulting in a reduction of stress-induced cortisol production. Thus, 20β-reduced progestins such as 20β-S can be regarded as important regulators involved in reproductive and stress-related processes in teleost fish.

4. Conclusions

In the present study, the biocatalytic potential of the novel steroid-transforming species Streptomyces rochei towards pregnane steroids was investigated. Selective 20β- and 20α-reductions of cortexolone and 17α-hydroxyprogesterone, respectively, were observed when the strains were incubated with steroid substrates, leading to the accumulation of over 90% of 20β-S and up to 70% of 17,20α-P. It was shown that microbially synthesized 20β-S represents a biologically active progestin for Danio rerio, exhibiting a complex profile of action. At concentrations effective in behavioral tests, the compound displays relatively low embryotoxicity, while simultaneously altering anxiety-like behavior, reducing whole-body cortisol, and inducing expression of several Pgr-dependent genes in the brain and ovaries. Taken together, these effects suggest ovarian and neural responses to 20β-S and support the role of 20β-reduced progestins as important modulators of reproductive and stress-related processes in fish. In the future, it appears promising to conduct more detailed studies on the receptor-binding characteristics of 20β-S, its tissue distribution, and activity in other teleost models. In addition, Pgr protein localization could be investigated using immunohistochemical approaches. The resulting data may be of interest not only from a fundamental perspective but also for potential applications in biotechnology and aquaculture.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/biom16020196/s1, Figure S1: Mass spectra of steroid metabolites produced by the S. rochei MTOC St1-St4 isolates during the bioconversion of cortexolone (metabolite I) (a); hydrocortisone (metabolite II) (b); progesterone (metabolite III) (c); 17α-OH-progesterone (metabolite IV) (d); and metabolite V (e); Figure S2: ¹H NMR spectra of the steroid products formed from cortexolone (metabolite I) (a), hydrocortisone (metabolite II) (b), progesterone (metabolite III) (c), and 17α-OH-progesterone (metabolites IV (d) and V (e)) by the S. rochei MTOC St1-St4 isolates; Figure S3: The crystalline product of 20β-OH-cortexolone (20β-S) (a) and the data of its HPLC analysis (b); Table S1: Oligonucleotide primers used in the work.

Author Contributions

Conceptualization and design of experiments, V.V.K. and V.V.P.; conducting experiments, V.V.K., V.V.P., A.A.V., N.A.M., E.O.M., O.V.F. and M.L.L.; HPLC analysis of steroid metabolites, A.A.S.; MS and 1H NMR analysis of steroid products, A.V.K.; supervision of the research, M.V.D.; writing the manuscript, V.V.K., V.V.P. and M.V.D. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

The animal study was reviewed and approved by Bioethics Commission LLC Collagel (protocol No. 01/0625; approval date: 3 June 2025).

Informed Consent Statement

Not applicable.

Data Availability Statement

The datasets obtained during the present study are available from the corresponding author on reasonable request.

Conflicts of Interest

Author Nikita A. Mitkin was employed by the company Collagel LLC. Author Olga V. Fadeeva was employed by the company Institute of Mitoengineering MSU LLC. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Funding Statement

This work was supported by the Ministry of Science and Higher Education of the Russian Federation No. 125041005029-5 on the theme FMRM-2025-0032 and RUDN University Scientific Projects Grant System, project № 202789-2-000.