Characterization of human adrenal cytochrome P450 11B2 products of progesterone and androstenedione oxidation

Abstract

Cytochrome P450 (P450) 11B1 and 11B2 both catalyze the 11β-hydroxylation of 11-deoxycorticosterone and the subsequent 18-hydroxylation of the product. P450 11B2, but not P450 11B1, catalyzes a further C-18 oxidation to yield aldosterone. 11-Oxygenated androgens are of interest, and 11-hydroxy progesterone has been reported to be a precursor of these. Oxidation of progesterone by purified recombinant P450 11B2 yielded a mono-hydroxy derivative as the major product, and co-chromatography with commercial standards and 2-D NMR spectroscopy indicated 11β-hydroxylation. 18Hydroxyprogesterone and a dihydroxyprogesterone were also formed. Similarly, oxidation of androstenedione by P450 11B2 yielded 11β-hydroxyandrostenedione, 18-hydroxyandrostenedione, and a dihydroxyandrostenedione. The steady-state kinetic parameters for androstenedione and progesterone 11β-hydroxylation were similar to those reported for the classic substrate 11-deoxycorticosterone. The source of 11α-hydroxyprogesterone in humans remains unresolved.

1. Introduction

11-Oxygenated steroids circulate in the body, with serum levels varying considerably among species. Humans and non-human primates have the highest levels [1]. These compounds include the 11α- and 11β-hydroxy (OH) and 11-keto derivatives of androstenedione, testosterone, and progesterone [2–18]. These compounds have varying biological activities. For instance, 11-keto testosterone is a potent androgen [19]. 11α-OH progesterone has been reported to have anti-androgenic properties and also to potentiate the anti-inflammatory properties of cortisol [15]. 11β-OH androstenedione has a potential role in prostate cancer [7]. 11-Oxygenated steroids are elevated in patients with P450 21A2 deficiency [8, 10, 11]. The 11-oxygenated steroids have been indicated in alternative routes to 5α-dihydro androgens in castration-resistant prostate cancer [7, 14].

The biosynthetic pathways involving 11-oxygenated steroids are still not all clear [15]. The 11α-hydroxylase enzyme has not been identified in human tissue [15]. Strushkevich et al. [20] reported that progesterone, testosterone, and androstenedione were all hydroxylated by human cytochromes P450 (P450, CYP) 11B1 and 11B2 but did not report the identities of any products. Gent et al. [15] reported that 11α-OH progesterone is not a substrate for 11β-OH steroid dehydrogenase (11β-HSD) Type II but acts as a competitive inhibitor of the enzyme. Both 11α- and 11β-OH progesterone have been reported to be inhibitors of both 11β-HSD Types I and II [22]. 11β-OH progesterone is a substrate for 11β-HSD Type II, however [15, 21]. Interestingly, 11-ketoprogesterone was found to be a substrate for reduction by 11β-HSD Type I [13]. 11α-OH and 11β-OH progesterone are both substrates for P450 17A1 and are converted to 11α,17α-diOH- and 11β,17α-diOH- progesterone and on to 11α-OH and 11β-OH androstenedione, respectively [15, 16]. These reactions could be demonstrated in PC3 and LNCaP prostate cancer cells as well as with recombinant enzymes. Human microsomal P450 3A4, generally recognized as a drug-metabolizing enzyme [23], has also been reported to catalyze trace 11β-hydroxylation of testosterone [4].

In the course of our own work on the kinetics and processivity of human P450 11B2 in the 3-step oxidation of deoxycorticosterone to aldosterone [24], we examined the ability of this enzyme to catalyze the oxidation of progesterone and androstenedione (Fig. 1). We identified the major product of progesterone as 11β-OH progesterone using mass spectrometry (MS), NMR, and co-chromatography with commercial standards and another product was identified as 18-OH progesterone. The conversion of androstenedione to 11β-OH androstenedione was established by co-chromatography and mass spectrometry. Steady-state kinetic parameters were also measured, which indicate relatively high rates (~20–30 min⁻¹) for these P450 11B2 11β-hydroxylation reactions.

Fig. 1.

Reactions catalyzed by human P450 11B2 reported earlier [20, 24, 25] and demonstrated in this work.

2. Materials and methods

2.1. Materials

Androstenedione, progesterone, and 11β-OH androstenedione were purchased from SigmaAldrich and used without further purification. 18-OH progesterone and 11-ketoprogesterone were obtained from Steraloids (Newport, RI). 11β-OH progesterone was purchased from both ChemScene (Monmouth Junction, NJ) and from Steraloids. 11α-OH progesterone was obtained from Tokyo Chemical Industry (TCI) America, (Portland, OR).

2.2. Enzymes

A modified recombinant version of human P450 11B2 was expressed in Escherichia coli cells and purified as described previously (first 24 N-terminal amino acids removed, following 6 changed to MATKAAR, C-terminal (His)₉ tag) [24]. Bovine adrenodoxin (Adx, lacking the mitochondrial targeting sequence, N-terminal Met modification) and bovine NADPH-Adx reductase (AdR, C-terminal (His)₆ tag modification) were expressed in E. coli and purified as described in detail elsewhere [26–29].

2.3. Enzyme reaction conditions

Typical incubations included 5 nM P450 11B2, 1 μM Adx, 0.5 μM AdR, 30 μM L-α−1,2-dilauroyl-sn-glycero-3-phosphocholine, and 0.5–100 μM substrate (androstenedione, progesterone, or 11β-OH androstenedione, dissolved in ethanol and diluted to ≤1% final ethanol concentration in each reaction) in 50 mM potassium phosphate buffer (pH 7.4). After a 5-minute pre-incubation at 37 °C, reactions were initiated by the addition of 1 mM NADPH. Incubations used for LC-MS analysis were prepared as described above, with 10 μM substrate. Additional incubations with OH-steroids were prepared as described above with 11β-OH progesterone, 11α-OH progesterone, 18-OH progesterone (10 μM) and 11β-OH androstenedione (1 μM).

For steady-state kinetic measurements, the final volume was 1 mL and the reaction time was 8 minutes. Incubations with OH-steroids had a reaction time of 20 minutes. Reactions were terminated by the addition of a 4× volume of ethyl acetate (4 mL) and mixing (vortex device). The layers were separated by centrifugation (2000 × g, 5 minutes), and a 3.8 mL aliquot of the organic phase (upper phase) was dried under an N₂ stream. The residues were dissolved in 150 μL of 9:1 A:B UPLC mobile phase (v/v, see below) and transferred to autosampler vials with inserts for UPLC.

An aliquot of each sample (20 μL) was analyzed by UPLC, done with a Waters Acquity system using an Acquity BEH octadecylsilane (C₁₈) column (1.7 μm, 2.1 mm × 100 mm) at 35 °C with a flow rate of 0.35 mL min⁻¹. Solvent A was 95% H₂O, 5% CH₃CN, 0.1% HCO₂H and Solvent B was 99% CH₃CN, 1% H₂O, 0.1% HCO₂H (all v/v/v). The solvent gradient used was: 0 min, 0% B; 7.5 min, 62.5% B; 8 min, 62.5% B; 8.25 min, 0% B; 10 min, 0% B.

For preparative reactions with the substrate progesterone, the reaction volume was 12.5 mL and the reaction time was 4 hours, with additional NADPH (200 μM) added every hour. The reaction was terminated by the addition of a 4× volume of CH₂Cl₂ (50 mL) and mixing (vortex device). An aliquot of the organic phase (lower phase) was reduced to dryness in vacuo and resuspended for preparative HPLC.

Preparative HPLC was done under similar conditions using a Beckman Ultrasphere octadecylsilane (C₁₈) column (5 μm; 4.6 mm × 250 mm) and a linear gradient consisting of increasing CH₃CN from 5% to 95% in 0.1% aqueous HCO₂H (all v/v) over a period of 20 min. The flow rate was 1.5 mL min⁻¹ and UV detection was at 245 nm. The peak of interest was collected, and the CH₃CN was removed under an N₂ stream. The product was separated from the remaining aqueous solution by extraction with dichloromethane and concentrated by removal of the solvent under a stream of N₂.

2.4. Mass spectrometry

LC-MS analysis was performed on an Acquity UPLC system (Waters) coupled with a Thermo-Finnigan LTQ-Orbitrap or LTQ XL-Orbitrap mass spectrometer (Thermo Scientific, San Jose, CA) equipped with an atmospheric pressure chemical ionization (APCI) source. Samples were separated by LC as described above. MS data were acquired in the positive ion mode using Xcalibur software (Thermo) after tuning with progesterone. Total ion scans were taken in the FTMS mode with 60,000 resolution from m/z 200 to 1500. Settings on the LTQ-Orbitrap were as follows: capillary temperature: 300°C; APCI vaporizer temperature 350 °C; sheath gas flow: 44; auxiliary gas flow 13; sweep gas flow 0; source current 4.5 μV; capillary voltage: 45 V, tube lens voltage 96 V. Settings on the LTQ XL-Orbitrap were as follows: capillary temperature: 275°C; APCI vaporizer temperature 450 °C; sheath gas flow: 59; auxiliary gas flow 10; sweep gas flow 5; source current 5 μV; capillary voltage: 34 V, tube lens voltage 90 V.

2.5. NMR spectroscopy

NMR experiments were acquired using a 14.0 T Bruker magnet equipped with a Bruker AV-III console operating at 600.13 MHz. All spectra were acquired in 3 mm NMR tubes using a Bruker 5 mm TCI cryogenically-cooled NMR probe. Chemical shifts were referenced internally to CDCl₃ (7.26 ppm) which also served as the ²H lock solvent. For 1D ¹H NMR, typical experimental conditions included 32 K data points, 13 ppm sweep width, a recycle delay of 1.5 s and 64 scans. For 2D ¹H-¹H COSY experiments, experimental conditions included 2048 × 512 data matrix, 13 ppm sweep width, recycle delay of 1.5 s and 8 scans per increment. The data were processed using a squared sinebell window function, symmetrized, and displayed in magnitude mode. Nuclear Overhauser effect correlated spectroscopy (NOESY) experiments were acquired using a 2048 × 512 data matrix with a 600 ms mixing time, 2 s recycle delay, and 8 scans per increment. The data was processed using a π/2 shifted squared sine window function and displayed in absorption mode. Multiplicity-edited HSQC experiments were acquired using a 1024 × 128 data matrix, a J(C-H) value of 145 Hz which resulted in a multiplicity selection delay of 34 ms, a recycle delay of 1.5 s, and 128 scans per increment along with GARP decoupling on ¹³C during the acquisition time (150 ms). The data was processed using a π/2 shifted squared sine window function and displayed with CH/CH₃ signals phased positive and CH₂ signals phased negative. J₁(C-H) filtered HMBC experiments were acquired using a 2048 × 128 data matrix, a J(C-H) value of 9 Hz for detection of long range couplings resulting in an evolution delay of 55 ms, J₁(C-H) filter delay of 145 Hz (34 ms) for the suppression of one-bond couplings, a recycle delay of 1.5 s, and 128 scans per increment. The HMBC data were processed using a π/2 shifted squared sine window function and displayed in magnitude mode.

2.6. Steady-state kinetic analysis

Rates of product formation were calculated from integration of the A₂₄₅ peaks in the UPLC chromatograms (all Δ⁴ steroids have very similar extinction coefficients [30, 31]). The product yields were converted to rates (v) (nmol product formed min⁻¹ (nmol P450)⁻¹). The concentrations of minor products formed from progesterone and androstenedione were calculated based on quantification of the corresponding 11β-OH steroid standard curve. Plots of v for each individual product vs. substrate concentration (S) were fit to hyperbolae (Michaelis-Menten kinetics) in Prism (GraphPad, San Diego, CA), fitting to k_cat and k_sp (i.e. k_cat/K_m) as suggested by Johnson [32], then calculating K_m from these parameters. The rates from the assay with 11β-OH androstenedione as the starting substrate were fit by linear regression to estimate k_cat/K_m because the rate of the reaction did not saturate.

3. Results

3.1. Oxidation of androstenedione

HPLC and LC-MS of the reaction products generated by the incubation of P450 11B2 with androstenedione yielded a major peak, judged to be a mono-OH product by MS (Fig. 2, Table 1, Sup Fig. 1) (calc. For C₁₉H₂₆O₃H⁺, 303.1960; found, 303.1964; Δ1.25 ppm). The compound co-eluted with a commercial sample of 11β-OH androstenedione (Sup Fig. 2B). Steady-state kinetic analysis yielded k_cat = 24.8 min⁻¹, K_m = 1.9 μM, and k_cat/K_m = 13 μM⁻¹ min⁻¹ (2.2 × 10⁵ M⁻¹ s⁻¹) (Fig. 3, Table 1). The reported values of k_cat and k_cat/K_m are not completely accurate due to the simultaneous formation of other products.

Fig. 2.

UPLC analysis of products of androstenedione (Andro) oxidation by P450 11B2. (A) LC-UV (A₂₄₅); (B) LC-MS. Reactions were performed with 10 μM substrate in the presence (red) or absence (black) of NADPH.

Table 1. Steady-state kinetics of P450 11B2 reactions.

See Figs. 3, 8, and Supplemental Fig. 6.

Substrate	Product	kcat, min−1	Km, μM	kcat/Km, min−1 μM−1
Progesterone	11β-OHa (tR 6.6 min, m/z 331.2 [M+H]+)	31 ± 1	5.7 ± 0.5	5.4 ± 0.4
	18-OHa (tR 6.3 min, m/z 331.2 [M+H]+)	5.2 ± 0.1	6.8 ± 0.5	0.77 ± 0.06
	di-OHa (tR 5.2 min, m/z 347.2 [M+H]+)	8.2 ± 0.8	13 ± 2	0.63 ± 0.09
Androstenedione	11β-OHa (tR 5.5 min, m/z 303.2 [M+H]+)	24.8 ± 0.5	1.9 ± 0.2	13 ± 1
	OH-a (tR 4.7 min, m/z 303.2 [M+H]+)	2.4 ± 0.1	1.9 ± 0.3	1.3 ± 0.2
	di-OHa (tR 4.25 min, m/z 319.2 [M+H] +)	5.0 ± 0.4	3.8 ± 0.9	1.3 ± 0.3
11β-OH Androstenedioneb	di-OH	-	-	0.0083 ± 0.0003
11-Deoxycorticosteronec	11β-OH	33 ± 1	2.7 ± 0.3	13 ± 1
Corticosteronec	18-OHa	14.9 ± 0.4	31 ± 1	0.49 ± 0.02
	Aldosteronea	2.8 ± 0.1	28 ± 2	0.10 ± 0.01

^aMultiple products are formed in these reactions. The presented values are not true steady-state parameters because the data are reports of simultaneous multiple reactions.

^bOnly k_cat/K_m was determined, see details in “Steady-state kinetic analysis” methods.

^cFrom [24].

Fig. 3.

Steady-state kinetics of P450 11B2 11β-hydroxylation of androstenedione. See Table 1. Reactions were completed in duplicate, and both points are shown.

Two other products were formed, at levels an order of magnitude lower (Fig. 2, Table 1). The peaks eluting at t_R 4.25 and 4.7 minutes had MH⁺ ions at m/z 319.2, and 303.2, respectively, indicating that they are putative di-OH (calc. For C₁₉H₂₆O₄H⁺, 319.1909; found, 319.1914; Δ1.5 ppm) and mono-OH products (calc. For C₁₉H₂₆O₃H⁺, 303.1960; found, 303.1965; Δ1.6 ppm) (Sup Fig. 1). The t_R 4.7 min peak is assigned as 18-OH androstenedione, which is known to be present in adrenals [33]. Although we did not have a standard reference compound and the amount of product we obtained did not permit NMR analysis, the strong MH⁺−30 peak in the mass spectrum (Sup Fig. 1A) was not seen in progesterone or other methylene-hydroxylated progesterone derivatives. It appears to be indicative of hydroxylation products of steroid methyl groups, e.g. as seen in 19-OH androstenedione [34] (The MH⁺−30 peak was confirmed to be due to loss of the elements of HCHO (calc. for C₁₈H₂₅O₂⁺ 273.1855, found 273.1857, Δ 0.9 ppm), and a proposed mechanism for the neutral loss is shown in Fig. S3). We tentatively conclude that the product is 18-OH androstenedione and not 19-OH androstenedione because it lacks the strong MH⁺−18 peak also seen with that compound [27]. The MH⁺−30 peak is also observed in the mass spectrum corresponding to di-OH androstenedione (Sup Fig. 1A), suggesting that it may be oxidized in the same position.

Incubation of 11β-OH androstenedione with P450 11B2 yielded a minor product at t_R 4.5 (Sup Fig. 4D). Based on the similar t_R and mass spectrum (same MH⁺−30 peak observed, data not shown), this is proposed to be the same di-OH androstenedione product formed by P450 11B2 from androstenedione (Sup Fig. 1). Steady-state kinetic analysis of this reaction yielded a k_cat/K_m of 0.0083 μM⁻¹ min⁻¹, three orders of magnitude lower than the reaction with androstenedione as the starting substrate.

3.2. Oxidation of progesterone and characterization of products

Incubation of progesterone with P450 11B2 yielded a major peak upon HPLC analysis, with m/z 331.2272 (calc for C₂₁H₃₀O₃H⁺ 331.2273, Δ 0.36 ppm), indicating mono-hydroxylation (Fig. 4, Sup Fig. 5).

Fig. 4.

UPLC of products of oxidation of progesterone (Prog) by P450 11B2. (A) UV (A₂₄₅); (B) LC-MS. Reactions were performed with 10 μM substrate in the presence (red) or absence (black) of NADPH.

Attempts at identification by co-chromatography were ambiguous due to differences between samples of commercial 11β-OH progesterone obtained from two different suppliers. Accordingly, we incubated a larger preparation of progesterone and P450 11B2 and collected the product for NMR analysis (Fig. 5). The NMR data confirmed the hydroxylation of this compound. The position of the hydroxylation at position C-11 is supported by the 2D NMR spectroscopy. 2D ¹H-¹³C HSQC revealed the presence of nine CH/CH₃ signals. Since the structure is known to contain three CH₃ groups that leaves six peaks that can be assigned to CH groups. In the absence of hydroxylation, the number of observable CH groups should only be five. The additional CH signal is evidence of a hydroxylation. Hydroxylation assigned to H-11 position (4.44 ppm) is supported by ¹H-¹H COSY correlations to a CH group (H9, 1.009 ppm) and a CH₂ group (H12–13, 1.67–2.20 ppm). The assignment for H9 was further confirmed by COSY with correlation to another CH (H8, 1.17 ppm) while the assignment of H12–13 was supported by 2D ¹H-¹³C HMBC which revealed a 3-bond correlation to a CH₃ (H-18, 0.91 ppm).

Fig. 5.

NMR spectra of major product of oxidation of progesterone. Spectra were acquired in CDCl₃ at 600 Mz. (A) 1-D ¹H spectrum; (B) 2-D COSY spectrum; (C) 2-D HSQC spectrum; (D) 2-D HMBC spectrum.

A sample of commercial 11α-OH progesterone (TCI) did not co-elute with the P450 11B2-generated product (Sup Fig. 2A). Standard 11β-OH progesterone (from Steraloids) co-eluted with the sample (Sup Fig. 2A). Analysis of the NMR spectra of the commercial samples (Fig. 6) indicated that the TCI 11α-OH progesterone and Steraloids 11β-OH progesterone have the correct structures. The assigned stereochemistry of the 11β-OH product (Figure 7B) was supported by 2D Nuclear Overhauser effect (NOE) correlated spectroscopy, which showed correlations for the α-CH (C11, 4.44 pm) to CH (H9, 1.009 ppm), CH (H8, 1.17 ppm), CH₂, (H12–13,1.67–2.20 ppm) and CH₃ (H19, 1.44ppm). While the stereochemistry for the 11α-OH group (Fig. 7A) was supported by observed NOE correlations from the β-CH (C11, 4.00 ppm) to CH₃ (H18, 0.68 ppm), CH₃ (H19, 1.31 ppm), and CH₂ (H12-H13, 1.51–2.34 ppm).

Fig. 6.

1D NMR spectra of commercial 11-OH progesterone samples. (A) 11α-OH progesterone (TCI); (B) 11β-OH progesterone (Steraloids).

Fig. 7.

NOESY NMR spectra of commercial 11-OH progesterone samples. (A) 11α-OH progesterone (TCI); (B) 11β-OH progesterone (Steraloids). Assigned connectivity patterns are shown.

The conclusion that P450 11B2 catalyzes 11β-hydroxylation of progesterone is consistent with the published results of van Rooyen et al. [13]. In steady-state kinetics, the k_cat was 31 min⁻¹ and K_m was 5.7 μM (k_cat/K_m 5.4 min⁻¹ μM⁻¹, i.e. 9 × 10⁴ M⁻¹ s⁻¹) (Fig. 8). As in the reactions with androstenedione, the reported values of k_cat and k_cat/K_m are not completely accurate due to the formation of other products in the reaction. Two minor products were formed (Fig. 4), at levels an order of magnitude less than the 11β-OH product. The peaks eluting at t_R 5.2 and 6.3 minutes had MH⁺ ions at m/z 347.2227 (calc. for C₂₁H₃₀O₄H⁺ 347.2222, Δ1.3 ppm) and 331.2280 (calc for C₂₁H₃₀O₃H⁺ 331.2273, Δ 2.0 ppm), respectively, indicating that they are putative di-OH and mono-OH products (Sup Fig. S5), similar to what was observed with androstenedione. The other mono-OH product co-eluted with 18-OH progesterone (Sup Fig. 2A). HPLC analysis indicated that the unknown minor hydroxylation product was not 11 keto-, 6β-, 11α-, 16α-, 17α-, or 21-OH progesterone (results not shown).

Fig. 8.

Steady-state kinetics of P450 11B2 11β-hydroxylation of progesterone. See Table 1. Reactions were completed in duplicate, and both points are shown.

Incubation of 11β-OH progesterone with P450 11B2 yielded a minor product (Sup Fig. 4A), which had a t_R of 5.48 (Sup Fig. 5) and mass spectrum (data not shown) similar to the di-OH progesterone identified in reactions with progesterone. Additionally, incubation of 11α-OH progesterone with P450 11B2 yielded a minor product with t_R 4.96 (Sup Fig. 4B). This product elutes before the proposed di-OH progesterone product that is formed in incubations with 11β-OH progesterone (Sup Fig. 3A). Given the earlier t_R of 11α-OH progesterone in comparison with 11β-OH progesterone, the t_R 4.96 product may be a di-OH progesterone with an 11α-OH instead of an 11β-OH. Incubation of 18-OH progesterone with a reconstituted system and NADPH did not yield any new products (Sup Fig. 4C).

4. Discussion

11β-OH androstenedione and 11β-OH testosterone have been known for many years, and these and other 11-oxygenated androgens have several biological properties. We confirm that P450 11B2 is one source of 11β-OH androstenedione. Alternatively, 11-OH progesterone can be converted to 11β-OH androstenedione by P450 17A1 via 17α-hydroxylation and lyase activity [16] and androstenedione can be hydroxylated to 11β-OH androstenedione by P450 11B1 in the adrenal [17]. Minor hydroxylation products and further oxidation of 11β-OH androstenedione to a putative di-OH androstenedione were found (Fig. 2, Sup Fig. 4D), but the positions of oxidation were not identified due to the small amounts of material available. Given the mass spectra of the products (Sup. Fig. 1A), we tentatively conclude that the minor hydroxylation product is 18-OH androstenedione and the di-OH product is 11β,18 di-OH androstenedione.

Although the 11β-hydroxylation of progesterone by P450 11B1 and 11B2 has been reported [5, 13], comprehensive characterization of progesterone oxidations by P450 11B2 had not been done. A primary product of progesterone oxidation was obtained, with the m/z corresponding to monooxidation (Fig. 2). Two-dimensional NMR implicated C-11 as the site of hydroxylation (Fig. 5), and we considered the possibility that P450 11B2 might be the source of 11α-OH progesterone but a standard sample did not co-elute. Careful NOESY NMR characterization of all of the commercial 11-OH progesterone samples was done (Fig. 7). The 11-H/19-CH cross peaks were key to resolving the issue (Fig. 7), and we conclude that only the β isomer of 11-OH progesterone is formed. Further oxidation of 11β-OH progesterone and 11α-OH progesterone was minor, though incubations with both produce products with t_R of what you would expect for di-OH progesterone.

We note that the specificity constants (k_cat/K_m) for 11β-hydroxylations of progesterone and androstenedione were nearly the same that we previously reported for 11-deoxycorticosterone (Table 1, Figs. 3, 8) [24]. This finding is important in considering the physiological roles of these reactions, although we are not implying that the reactions with progesterone and androstenedione are necessarily as physiologically important as the 11β-hydroxylation of 11-deoxycorticosterone, which leads to aldosterone. As mammalian P450s go, the specificity constants (~ 10⁵ M⁻¹ s⁻¹) are relatively high [35], and these reactions with progesterone and androstenedione should not be considered artifacts. On the other hand, the specificity constant (k_cat/K_m) for hydroxylation of 11β-OH androstenedione is much lower than what we previously reported for 18-hydroxylation of corticosterone (Table 1, Sup Fig. 6E). While we did not directly investigate the processivity of the reactions presented in this study as we did with the formation of aldosterone [24], the steady-state kinetic results suggest that the reactions with androstenedione may occur via a processive mechanism (i.e. since the production of di-OH androstenedione is >100 times more favored when androstenedione is the substrate compared with 11β-OH androstenedione).

While we have demonstrated that P450 11B2 can perform these reactions in vitro, the localization of steroids and differential expression of steroid-metabolizing enzymes within the adrenal gland is important to consider when addressing the potential importance in vivo. In normal conditions P450 11B2 is expressed in the zona glomerulosa of the adrenal cortex [36]. Progesterone can be synthesized from pregnenolone in the zona glomerulosa or zona fasciculata by 3β-HSD2. Androgens, including androstenedione, are normally secreted by the zona reticularis (pathways in each adrenal zone reviewed in [37]). Because of this, P450 11B2 may encounter progesterone in normal tissue, but likely not androstenedione. P450 11B2 is also expressed in aldosterone-producing adenomas (APAs) [36] and in some cases, P450 17A1 can be co-expressed [38]. The expression of P450 17A1 could allow for the production of dehydroepiandrosterone (DHEA) within the APA, which could then be converted to androstenedione by 3β-HSD2 (which has also been shown to be strongly expressed in APAs [39]). The co-expression of P450 17A1 and P450 11B2 in APAs has already been proposed to lead to the production of the hybrid steroids 18-OH cortisol and 18-oxocortisol [18]. A similar principle may apply to 11-oxygenated androstenedione.

We did not prepare P450 11B1 and compare kinetic parameters for the 11β-hydroxylation reactions. van Rooyen et al. [13] found a higher conversion of progesterone to 11β-OH progesterone by P450 11B1 than 11B2, but the results are for single time points and are not normalized for levels of expression in the cell culture system (and AdR was not over-expressed in the system). No other products (other than 11β-OH) were reported, while we identified additional minor products. In our work we measured the oxidation of only one androgen, androstenedione, and not testosterone. Swart et al. [5], using a cell system, reported that P450 11B1 hydroxylated androstenedione but P450 11B2 did so only at a very low rate, and both P450 11B1 and 11B2 catalyzed the hydroxylation of testosterone. However, our own rates (Table 1) are as high as for the substrate 11-deoxycorticosterone and higher than for progesterone.

18-Hydroxylation of progesterone and androstenedione was not as efficient as 11β-hydroxylation. Similar to these reactions, the rate of 18-hydroxylation of the classical P450 11B2 substrate, 11-deoxycorticosterone, to 18-hydroxy-11-deoxycoricosterone is also lower than the 11β-hydroxylation to corticosterone [20]. The 18-hydroxylation of C19 and C18 steroids, as well as C21, has been recognized in several systems, for some time [33, 40–44], and may not seem surprising in light of the activity of P450 11B2 on corticosterone. Some 18-oxygenated steroids are biologically active [2, 3], although exactly what their physiological roles are remains unknown.

Finally, the source of 11α-OH steroids remains unknown. It is of interest that the 11α-hydroxylation of progesterone by a fungal (Rhizopus) P450 was one of the first examples of the use of an enzyme in a chemical synthesis in industry [45].To our knowledge the enzyme that forms this in humans has not been identified [15]. 11β-OH progesterone is oxidized to the 11-keto product by 11β-HSD Type 2, and the reverse reaction is catalyzed by 11β-HSD Type 1 but apparently these reactions are stereospecific and do not yield 11α-OH progesterone [15, 21].

Highlights.

• Human cytochrome P450 11B2 catalyzes the 11β-hydroxylation of progesterone and androstenedione.

• The rates and specificity constants are similar to those for 11β-hydroxylation of the classic substrate 11-deoxycorticosterone by the same enzyme.

• Minor products of both androstenedione and progesterone were also identified, including 18-hydroxy products.

Abbreviations:

Adx

adrenodoxin

AdR

NADPH-adrenodoxin reductase

APA

aldosterone-producing adenoma

APCI

atmospheric pressure chemical ionization

HSD

hydroxy steroid dehydrogenase

MS

mass spectrometry

P450 or CYP

cytochrome P450

POR

NADPH-cytochrome P450 reductase

COSY

correlated (NMR) spectroscopy

HMBC

heteronuclear multiple bond correlation (NMR) spectroscopy

HSQC

heteronuclear single quantum coherence (NMR) spectroscopy

NOESY

nuclear Overhauser effect correlated (NMR) spectroscopy