Different effects of progesterone and estradiol on chimeric and wild type aldosterone synthase in vitro

Recently, we reported the unequal crossover break point in the CYP11B1/B2 gene [20]. The 50-bp crossover region contains segments of intron 3 of CYP11B1 (c.2937-40) and exon 4 of CYP11B2 (c.2937 + 10). Using the PCMV-CYP11B1 and PCMV-CYP11B2 vectors and based on restriction enzyme analysis, we selected a crossover point to create the fusion vector containing exons 1 to 3 of CYP11B1 (1-573 bp) and exon 4 to 9 of CYP11B2 (574-1512 bp). Exons 5 and 6 of CYP11B2 were maintained in the chimeric enzyme. PCMV-CYP11B1 and PCMV-CYP11B2 vectors were kindly provided by Dr. Walter L. Miller (University of California, San Francisco). Mutagenex Inc. (Hillsborough, NJ, USA) [21] performed the chimeragenesis of the CYP11B1/B2 gene. Restriction endonuclease digestion and Sanger sequencing confirmed the integrity of the plasmid constructs, and the amplification products were verified by sequencing at Macrogen (Rockvill, MD, USA) [22].

The human embryonic kidney cell line, HEK-293, was grown in high-glucose Dulbecco’s modified Eagle’s medium (DMEM HG, Life Technologies, Sao Paulo, Brazil) supplemented with 10% foetal bovine serum (FBS, Life Technologies, Sao Paulo, Brazil), 100 IU/mL penicillin and 100 μg/mL streptomycin. For the enzyme activity studies, DMEM HG-FBS medium was treated with activated charcoal to eliminate steroid contaminants. For the transfection experiments, HEK-293 cells were plated at 6×10⁵ cells per well in 6-well plates and then transfected using Turbofect™ in vitro transfection reagent (Fermentas, Thermo Scientific, Rockford, IL, USA), in accordance with the manufacturer's protocol. Briefly, 2 μg of pCMV4, pCMV4-CYP11B1, pCMV4-CYP11B2, or pCMV4-CYP11B1/B2 plasmid were added in DMEM. pZsGreen1-n1 (0.3 μg, Clontech, California, USA) was added as a marker of transfection efficiency. The transfected cells were visualised using a Panasonic–DMC-LC40 LUMIX and the Pro-QV7software. The efficiency of transfection was evaluated using an Olympus CKX41 inverted microscope coupled to an Olympus U-RFL-T. For this purpose, we took six photographs (40×) of plates containing HEK-293 cells (in each condition) under bright field and again with the same field of view under fluorescent light using a Micropublisher 3.3RTV camera and the Pro-QV7software. Then, the total number of cells was counted under both conditions using the Image J v1.46 program [18] in three independent trials.

The whole-cell extracts of untransfected HEK-293, PCMV-CYP11B1, PCMV-CYP11B1/B2 or PCMV-CYP11B2-transfected cells were obtained using 100 μL of lysis buffer containing protease inhibitors. The protein concentration was estimated using the BCA protein colorimetric assay kit (Thermo Scientific, Rockford, IL, USA) on an Infinite® 200 PRO NanoQuantmultimode reader (Tecan, Männedorf, Switzerland). The denatured protein samples (50 μg per well) were separated by 12% SDS-PAGE and were immobilised onto 0.45-μm-pore nitrocellulose membranes (Thermo Scientific, Rockford, IL, USA). The membranes were probed with a rabbit anti-human CYP11B2 antibody (aa 80–90 (RYNLGGPRMVC of CYP11B2) followed by a goat anti-rabbit peroxidase-conjugated antibody (Thermo Scientific, Rockford, IL, USA), as previously described [23]. The CYP11B2 protein was visualised using the enhanced chemiluminescence Super Signal Pico Chemiluminescent Substrate kit (Thermo Scientific, Rockford, IL, USA) and Agfa X-ray film (Agfa-Gevaert, N.V, Mortsel, Belgium). Protein extracts from a sample of human adrenal gland tissue were used as positive control. The loading control was β-actin. The CYP11B1/B2 has an amino acid sequence of 1–191 from the N-terminus of the CYP11B1 enzyme and was thus not detected by this antibody.

Expression of CYP11B2 and CYP11B1/B2 in transfected HEK-293 cells were evaluated by qRT-PCR. Total RNA was extracted from transfected HEK-293 cells treated or not with progesterone by TRIZOL® (Life Technologies, California, USA) then reverse transcribed using RevertAid H Minus Reverse Transcriptase (Thermo Scientific, California, USA) following the manufacture’s instruction. Quantitative real-time polymerase chain reaction was performed using Maxima SYBR (Thermo Scientific, California, USA). Primers were CYP11B2 forward: 5`- gga act tcc acc acg tgc cct tt-3` and CYP11B2 reverse: 5`- att gag gcc tgg cac gtc cc-3. GAPDH forward: 5-gaa cat cat ccc tgc ctc tac t −3`, and GAPDH reverse: 5 –cct gct tca cc acct tct tg −3. The mRNA expression was quantified by ΔΔCt method relative to that of GAPDH [24]. CYP11B2 primers are located in exons eight and nine.

To determine the kinetic constants under our assay conditions, the PCMV-CYP11B1, PCMV-CYP11B1/B2, and PCMV-CYP11B2 HEK-293 transfected cells at 18 h post-transfection were incubated for 24 h with increasing concentrations (ranging from 0.18 to 30 μmol/L) of deoxycorticosterone as substrate (DOC, Steraloids Inc., Andover, MA, USA). Resultant aldosterone production was quantified using HPLC-MS/MS (Agilent 1200, ABI Sciex API4000 Qtrap). The apparent kinetics parameters Km and Vmax were determined by plotting the aldosterone production versus the corresponding substrate concentrations and applying Michaelis–Menten kinetics using the Prism v5.03 program (GraphPad Software, Inc.).

All chemicals were purchased from Sigma (Sigma-Aldrich Quimica Ltda, Santiago, Chile). To determine any inhibition by sex steroid hormones, at 18 h post-transfection the cells were washed with PBS and 1.5 μmol/L DOC-DMEM steroid hormone-depleted FBS and exposed to 24 hours of increasing concentrations of progesterone or estradiol (0.625-10.0 μM). Ketoconazole (ranging from 0.625 to 5 μM) was used as an inhibitor control. The supernatant (1.0 mL) was collected, and aldosterone levels were measured by HPLC-MS/MS in four independent trials. The IC50 was determined by plotting the aldosterone production versus the corresponding log inhibitor concentrations and applying a dose–response inhibition analysis.

To determine the effect of progesterone on apparent kinetic constants under our assay conditions, the PCMV-CYP11B1/B2, and PCMV-CYP11B2 HEK-293 transfected cells, after 18 h post-transfection, were incubated for 24 h with increasing concentrations (ranging from 0.18 to 30 μM) of DOC (Steraloids Inc., Andover, MA, USA) and progesterone at the IC50 concentrations for each enzyme. The resultant aldosterone production in the supernatant was quantified using HPLC-MS/MS (Agilent 1200, ABI Sciex API4000 Qtrap). The apparent kinetics parameters Km and Vmax were determined by plotting the aldosterone production versus the corresponding substrate concentrations and applying Michaelis–Menten kinetics using the Prism v5.03 program (GraphPad Software, Inc.).

The CellTiter 96 AQueous One Solution Cell Proliferation Assay (Promega) kit containing the tetrazolium compound MTS ([3-(4,5-dimethyl-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt]) was used to monitor cell viability according to the manufacturer’s protocols. Briefly, human embryonic kidney (HEK-293) cells maintained in DMEM with 10% foetal bovine serum (FBS) under standard cell culture conditions (37°C, humidified, 5% CO2) were plated at a density of 1.5×10⁴ cells/well in a 96-well plate and incubated in growth media for 18 hours. Cells were treated with 0.8 to-50 μM of progesterone, estradiol or DOC in DMEM containing 10% FBS for 24 hours. After the indicated time of incubation with the appropriate medium, a 20 μL MTS/PMS (1:0.05) mixture was added per well, and cells were incubated for an additional hour. MTS is reduced by viable cells to formazan, which was monitored at 490 nm by an ELX-800 universal plate reader (BioTek, Winooski, VT). Formazan production is time dependent and proportional to the number of viable cells. Cells not incubated in DMEM were used as control condition. The percentage of cell death was calculated by the ratio of the optical density obtained in each treatment to that obtained for the controls. A cytotoxic concentration was noted when the optical density in each condition was less than the average of the control minus two standard deviations.

The amino acid sequence of the human CYP11B1 was retrieved from the Uniprot database (entry code P15538), and the chimeric CYP11B1/B2 sequence was obtained by performing DNA sequencing [25]. The comparative modelling was performed using the MODELLER program implemented in the Build Homology Models protocol in Discovery Studio v2.1 (Accelrys Inc., San Diego, USA) [26] using the recently reported crystal structure of human CYP11B2 in complex with DOC, which was identified as a suitable template for the modelling of both proteins (PDB id 4DVQ, resolution 2.49 Å) [27]. For modelling purposes, the first 33 residues of each protein were not included. The coordinates for DOC and the HEME group were modelled using the copy ligand parameter from the template structure, from which the Chain A was used. One hundred models for each protein were generated, and the top ranked by the MODELLER internal DOPE score energy minimized using the conjugate gradient algorithm until a RMS gradient of 0.001 kcal/mol Å was reached. The CHARMM22 force field with a dielectric constant of 4 and a distance-dependent dielectric implicit solvent model was used to mimic the membrane environment [28]. Model quality was assessed by Ramachandran plot analysis, PROSA and Verify3D structure validation [29‐31]. The electrostatic potential energy profiles were calculated using APBS [32]. Conformers of each docked compound were obtained with OMEGA v2.4.6 using default parameters [33,34]. Docking calculations were performed using FRED v3.0 (OpenEye Scientific Software, Santa Fe, New Mexico) [35], and the solutions ranked according to the Chemgauss4 scoring function [36].

Data are expressed as the mean +/- SEM. The kinetic parameters Vmax and Km were obtained by Prism v5.03. Differences between the means were analysed by repeated measures of an ANOVA and Tukey’s post hoc test. Differences of area under curve analyses were performed by the Mann Whitney test. Statistical analysis was performed using Prism v5.03 (GraphPad Software, Inc.). Differences were considered significant at p < 0.05.

As described in the methods section, the crossover point was used in the chimeragenesis to create the fusion vector CYP11B1 (1-573 bp)/CYP11B2 (574-1512 bp). A schematic representation of the FH-I crossover and the resultant ASCE product that was synthesised for assay studies is shown in Figure 1A.

A representative image of HEK-293 transfected with wild type or chimeric expression vectors are shown in Figure 1B. Morphological changes were not observed between the HEK-293 cells transfected with the constructs containing the CYP11B enzymes and the cells transfected with the PCMV vector or the non-transfected cells (NT). Comparable transfection efficiencies were observed by counting cells that express the green fluorescent protein, which correlated with the total number of cells for each assay condition (Figure 1C). The expression of the aldosterone synthase in HEK-293 cells transfected with the constructs containing the CYP11B2 was examined by Western blotting (Figure 1D). A human adenoma sample from the adrenal cortex was used as positive control (line 1), and the following were observed: no transfected HEK-293 (line 2); PCMV-CYP11B1 (line 3); PCMV-CYP11B1/B2 (line 4) and PCMV-CYP11B2 (line 5). An intense immunoreactive band with an apparent molecular weight of approximately 50 kDa was present in the human adenoma sample and in the HEK-293 cells transfected with PCMV-CYP11B2. Western blotting using a CYP11B1 antibody was performed for the same samples. Unfortunately, the immunoreactivity band was too weak. To probe the CYP11B1 activity of PCMV-CYP11B1 construction, we incubated PCMV-CYP11B1 transfected HEK-293 cells with increasing concentrations of corticosterone (0.6-1.5 μM), and we obtained the expected increasing levels of cortisol (data not shown).

The transfected HEK-293 cells supported CYP11B2- and CYP11B1/B2-dependent steroid conversion without the additional heterologous expression of the corresponding electron donor system, similar to previous experiments reported by Denner et al. [37]. Figure 2A shows a dose response curve performed for five independent HEK-293 transfections of PCMV-CYP11B1/B2, PCMV-CYP11B2 or PCMV-CYP11B1 (11β-hydroxylase gene, 11BH), incubated with increasing concentrations of DOC (0.18-30 μM) for 24 hours. The aldosterone production versus substrate concentration was plotted for ASCE (open circles), ASWT (closed circles) and 11BH (grey circles). 11β-hydroxylase did not produce aldosterone in any of the DOC concentrations that were probed. The apparent kinetic enzyme parameters obtained were Km = 1.191 μM and Vmax = 27.08 μM/24 h for ASCE and Km = 1.163 μM and Vmax = 36.98. The comparison of the area under the curve for aldosterone production by ASCE and ASWT did not show any significant differences (p = 0.3095 Mann–Whitney test). This finding suggests that both enzymes exhibited analogous affinity for the substrate and showed the same efficiency to produce aldosterone. DOC did not exhibit a toxic effect at any concentration probed (Figure 2B). Non-transfected HEK-293 cells incubated with DOC (1.5 μM) did not produce aldosterone (data not shown).

The average aldosterone production by ASCE was 13.6 μM/24 h and by ASWT was 15 μM/24 h when the enzymes were incubated with 1.5 μM of DOC. In the presence of DOC as substrate, progesterone inhibited the aldosterone production by both ASCE and ASWT in a dose-dependent manner with similar efficacies. Statistically significant inhibition of ASCE was achieved at 5 μM of progesterone (Figure 3A, white bars), and statistically significant inhibition of ASWT was achieved at 2.5 μM of progesterone (Figure 3A, black bars). For ASCE, the calculated IC50 value for progesterone was 3.907 μM, and for ASWT, it was 2.240 μM (Figure 3B). These inhibition values were not due to an inhibition in plasmid transcription efficiency (See Additional file 1: Figure S1). Estradiol did not affect ASCE or ASWT aldosterone synthase activity in the range of concentrations assayed (Figure 3C) or coincubated with progesterone (See Additional file 2: Figure S2). As expected, the known aldosterone synthase inhibitor ketoconazole inhibited both enzyme activities by 90% at all concentrations probed (Figure 3D). None of the steroids assayed nor ketoconazole demonstrated cytotoxic effects in HEK-293 cells in any of the concentrations that were probed (Figure 3E). As a control, non-transfected HEK-293 cells were treated with the same increasing concentrations of progesterone or estradiol. No aldosterone production was detected under these conditions (data not shown). The vectors PCMV-CYP11B2 or PCMV-CYP11B1/B2 used in the transfection experiments with HEK-293 cells have the same viral promoter. For that reason, the expression levels of chimeric or wild type aldosterone synthase should be affected by the same effectors. Data in Figure 3 are expressed as the mean +/− S.E.M. of four independent trials.

We developed 3D models of both proteins by comparative modelling using the human CYP11B2 (ASWT) crystal structure. In the final alignment used to model the proteins, the percentage of sequence identity was 93.6% and 97.7% for the modelled region, and 96.4% and 98.9% homologies were observed between the template structure and CYP11B1 and ASCE, respectively. For ASCE, the grey bar indicates the corresponding CYP11B1 portion, and the green bar represents the CYP11B2 limits for ASCE (See Additional file 3: Figure S3). A Ramachandran plot analysis indicated that the modelled proteins have more than 95% of the residues in the allowed region. The PROSA Z-scores and Verify3D profile scores also support the quality of the obtained models for further studies (Table 1). The obtained models exhibit a root-mean square deviation (RMSD) of alpha carbons of less than 0.4 Å when superimposed on the template structure. Figure 4 depicts the secondary structure of the modelled proteins compared to ASWT. The ASCE model (Figure 4B) is composed of the first 38% of the 11BH (residues 34 to 191), which contains the substrate (steroid) entry region of CYP11B1, and the last 62% of the ASWT (residues 192–503). All models have the common folding pattern of CYP450 enzymes, and according to the identified crossover, the shift will occur at the end of helix G. Electrostatic potential surface profiles of the proteins indicate that the surface potential of ASCE is more similar to that of ASWT, with the steroid binding site access channel in ASCE being more electronegative and more extended than that of ASWT, as determined by the size of the associated cavities, which are 471.87 and 553.62 Å³ for ASCE and ASWT, respectively (Table 2).

To gain insight into the inhibition profile of progesterone, we performed kinetic experiments in the presence of the corresponding IC50 concentrations of progesterone for each enzyme (Figures 5A and 5D). In progesterone presence, ASCE Km was very variable and lightly increase, Vmax was no statistically different (p = 0.0667). ASWT Km significantly increase (p = 0.0095) but Vmax was no statistically different (p = 0.1048). The new calculated kinetic enzyme parameters obtained in presence of progesterone were Km = 2.451 μM and Vmax = 37.91 μM/24 h for ASCE, and Km = 10.78 μM and Vmax = 51.61 μM/24 h for ASWT. Progesterone inhibits the aldosterone synthase wilde type in a competitive fashion.

To explore the putative binding mode of the steroid compounds in CYP11B1, ASCE and ASWT, we performed docking simulations within the binding site of all proteins. Figure 5B and C show the most favourable predicted binding mode obtained for DOC and progesterone within the ASCE, where progesterone binds with its 3-carboxy group facing the inner part of the binding pocket that is composed of the side chains of Arg120 and Glu310. The composition of this binding pocket is similar to that of DOC in ASCE and ASWT (Figure 5B and E, respectively), but progesterone penetrates deeper into the pocket with stabilisation via a hydrogen bond interaction with Arg120. The methyl groups at positions 18 and 19 face the HEME group and are surrounded by the side chains of Ala313 and Thr318. The beta side of the steroid scaffold faces the top of the aromatic cluster pocket composed by Trp116, Trp260, Phe231 and Phe487. Figure 5F represents the most favourable predicted binding mode obtained for progesterone with ASWT.

DOC and progesterone have a similar binding mode in CYP11B1 as that predicted for ASCE (See Additional file 4: Figure S4A and Figure S4B) and as the experimentally determined binding mode of DOC to ASWT (Figure 5B). Estradiol binds with its aromatic portion positioned inside the pocket and interacts with the pocket via its aromatic portion with the side chains of Phe130, Trp116 and Trp260 but fails to establish any H-bond interactions with the binding site (See Additional file 4: Figure S4C). Finally, ketoconazole binds with its imidazole moiety making direct contact with the iron atom in the HEME group, with its dichlorophenyl group making aromatic interactions with Phe130 and Trp116, and with its terminal amide in making H-bond interactions with Asn404 (See Additional file 4: Figure S4D). The binding energies displayed in Table 2 suggest that sex steroids can bind to both proteins, ASCE and ASWT, but with lower affinity than ketoconazole or DOC to ASWT, according to the ChemGauss4 scoring and Ludi2 scoring function-derived binding energies.