Identification of Outer Mitochondrial Membrane Cytochrome b 5 as a Modulator for Androgen Synthesis in Leydig Cells*

Outer mitochondrial membrane cytochrome b 5 is an isoform of microsomal membrane cytochrome b 5 . In rat testes the outer mitochondrial membrane cytochrome b 5 is present in both mitochondria and microsomes, whereas microsomal membrane cytochrome b 5 is undetectable. Outer mitochondrial membrane cytochrome b 5 present in the testis was localized in Leydig cells with cytochrome P-450 17 (cid:1) , which catalyzes androgenesis therein. We there- fore analyzed the functions of outer mitochondrial membrane cytochrome b 5 in rat testis microsomes by using a proteoliposome system. In a low but physiological concentration of NADPH-cytochrome P-450 reductase and excess amount of progesterone, outer mitochondrial membrane cytochrome b 5 stimulated the cytochrome P-450 17 (cid:1) -cata- lyzed reactions, 17 (cid:1) -hydroxylation and C17-C20 bond cleavage. The effects were different from those by microsomal membrane cytochrome b 5 as follows: preferential elevation of the 17 (cid:1) -hydroxylase activity by outer mitochondrial membrane cytochrome b 5 in an amount- dependent manner versus that of the lyase activity by microsomal membrane cytochrome b 5 at the low concen- tration, and the inhibition of both activities at the high concentration. At a low concentration of progesterone reflecting a physiological cholesterol supply, outer mitochondrial pmol of purified proteins were elec-trophoresed. B, amounts of P-450 reductase in rat testis microsomes were analyzed by SDS-PAGE (7.5%) and immunoblotting using anti-rat P-450 reductase-IgG and goat anti-rabbit IgG-horseradish peroxidase complex. The chemiluminescence of the immunoreactive bands was quantified with purified P-450 reductase as a standard. The integrated values of the luminescence of the microsomal fractions from two indi-vidual rats ( lanes 3–5 and 6–8 , 50, 25, and 12.5 (cid:3) g of protein, respec- tively) were calibrated with those of standard bands ( lanes 1, 2, and 9–11 , 0.25 0.125, 0.0125, 0.025, and 0.1 pmol, respectively). The inte- gration was performed with CS Analyzer (version 2.0, Atto Instruments).

hydrophobic, and (c) carboxyl-terminal hydrophilic domains. A protoheme is bound to the first domain, which is highly conserved between b 5 and OMb with 70% identity (4,5). The hydrophobic domain, consisting of about 20 amino acid residues, is embedded in the lipid bilayer and functions for the insertion of proteins into the membranes as tail-anchored proteins (6). The carboxyl-terminal 10 amino acid residues of b 5 are exposed to the luminal side of the ER cisterna (7,8) and are required for the targeting of the cytochrome to the organelle (9). The visible spectroscopic properties of the reduced forms are characteristic of the b 5 -type hemoproteins. OMb and b 5 are spectrographically indistinguishable from each other due to the highly conserved heme-binding portion (1,2). OMb and b 5 have a similar mobility on SDS-PAGE and are co-purified unless specific precaution was taken in the purification procedures. Antibodies directed against b 5 that had been purified without the removal step of OMb cross-react with OMb. In some case, cross-reactions are observed even if a highly purified form of OMb or b 5 was immunized in rabbits. These facts made the discrimination between OMb and b 5 difficult. The most convincing way of the discrimination is use of the specific antibodies. We have such antibodies in a large collection of anti-OMb and anti-b 5 antibodies.
Because b 5 is a hemoprotein with sixth co-ordination, it is incapable of activating an oxygen molecule as cytochrome P-450 does. One of its physiological functions is an electron transfer to terminal oxidases such as stearyl-CoA desaturase (10). Some evidence also suggests that b 5 functions as a modifier for some cytochrome P-450-catalyzed reactions although its mechanism is not clear (11,12). For example, b 5 stimulates the 6␤-hydroxylation of testosterone and nifedipine oxidation by recombinant CYP3A4 (13). It also augments C17-C20 lyase activity of pig, guinea pig, and human P-450 17␣ leading to predominant formation of androgens (androstenedione and dehydroepiandrosterone) over 17␣-hydroxylated steroids (progesterone and pregnenolone) (14 -16). Without involvement of added b 5 , the lyase activity of the P-450 17␣ is too low to account for physiological production of androgens in vivo.
The physiological functions of OMb are not well understood. The sole experimental evidence for the function is that specific antibodies against rat OMb inhibited semidehydroascorbate reductase, which catalyzes regeneration of half-oxidized ascorbic acid, in outer mitochondrial membrane fractions of rat liver cells (17,18). In a study on the mechanism for subcellular localization of OMb, we have recently observed that it moves from the outer mitochondrial membrane to the ER membrane in a few hours after administration of dexamethasone, pregnenolone-16␣-carbonitrile, or phenobarbital to rats. 2 In tissues other than liver, such as kidney and adrenal glands of rats, OMb was present in ER in addition to mitochondria. These results indicate that OMb has the potential to be localized either in outer mitochondrial membranes or in ER membranes, although the localization signal of OMb seems to reside in the carboxyl-terminal 10 amino acid residues favorable for mitochondrial targeting (19). In guinea pig, OMb is abundant in adrenal glands, where its localization is exclusively ER. Surprisingly, b 5 is scarcely detected in rat testicular cells, where OMb is distributed almost equally between the outer mitochondrial membranes and ER membranes in considerably high amounts. This suggests that OMb and not b 5 could regulate testicular androgen synthesis by modifying the lyase activity of P-450 17␣ . The indistinguishable properties between OMb and b 5 as described above propose a possibility that the stimulatory effects on P-450 that have been believed to be done by b 5 are exerted by OMb. Thus, in this study we analyzed the effects of OMb on rat P-450 17␣ , and we found that this cytochrome is a genuine modulator for testicular androgen synthesis in rats.

EXPERIMENTAL PROCEDURES
Animals-Male Sprague-Dawley rats weighing 120 -150 g and male Hartley guinea pigs weighing 450 -500 g were purchased from Kyudo Co., Ltd. (Kumamoto, Japan). All surgical and experimental procedures were approved and conducted in accordance with the policies of Kyushu University's Animal Care and Use Committee.
Cellular Fractionation and Detection of OMb and b 5 -Male Sprague-Dawley rats weighing 120 -150 g were intraperitoneally injected with sodium phenobarbital at a dose of 80 mg/kg and killed after 24 h. Liver and testes were removed and homogenized using a Teflon glass homogenizer in ice-cold 10 mM Tris-HCl buffer (pH 7.6) containing 0.25 M sucrose, 0.1 mM EDTA, 2 g/ml leupeptin, and 2 g/ml pepstatin A. After the homogenates were centrifuged at 600 ϫ g for 10 min, the resultant supernatants were recentrifuged at 6,000 ϫ g for 10 min to obtain the mitochondrial fractions. Microsomal fractions were obtained by centrifugation of the post-mitochondrial supernatants at 105,000 ϫ g for 60 min. All procedures were done at 4°C. For analysis of the distribution of OMb and b 5 in the tissues, the subcellular fractions form the tissues and were separated by Tricine/SDS-PAGE (20) and subsequently transferred to polyvinylidene difluoride membranes. The membrane was incubated first with specific anti-rat OMb or anti-rat b 5 IgG and second with goat anti-rabbit IgG-horseradish peroxidase complex, and the immunoreactive bands were finally visualized by chemiluminescence using Western Lighting TM (PerkinElmer Life Sciences).
Enzyme Preparations-All enzyme preparations except as described otherwise were conducted with K-P i buffer (pH 7.4) containing 20% glycerol, 0.1 mM EDTA, and 0.1 mM dithiothreitol. Rat P-450 17␣ and OMb were solubilized from rat testis microsomes with the 10 mM K-P i buffer containing 1.4% (w/v) Emulgen 913 (a kind gift from Kao Chemicals, Osaka, Japan) and 2 g/ml each of leupeptin and pepstatin and after dilution of the detergent to 0.2% absorbed to -aminohexyl-Toyopearl column (1.2 ϫ 8.0 cm) equilibrated with 10 mM K-P i buffer containing 0.2% Emulgen 913. After washing with the 10 mM K-P i buffer containing 0.2% Emulgen 913, the column was eluted with the 10 mM K-P i buffer containing 0.2% Emulgen 913 and 0.2% sodium cholate. The eluate contained P-450 17␣ . The column was then washed with the 10 mM K-P i buffer containing 0.4% Emulgen 913 and eluted with the 10 mM K-P i buffer containing 0.5 M KCl and 0.4% Emulgen 913. The eluate contained OMb. Both enzymes were further purified by high pressure liquid chromatography on a TSK DEAE-5PW column (7.5 ϫ 75 mm, TOSOH, Tokyo, Japan) with a linear gradient of KCl (0 -500 mM/60 min) in the 10 mM K-P i buffer containing 0.4% Emulgen 913. Purified P-450 17␣ was applied to a hydroxylapatite column, and the column was washed extensively with the 10 mM K-P i buffer without Emulgen 913. The enzyme was eluted from the gel batch with the 400 mM K-P i buffer containing 10% (w/v) L-␣-phosphatidylcholine from egg yolk (Wako Pure Chemical Co. Ltd., Osaka, Japan) and 1% sodium cholate. Proteoliposomes embedded with P-450 17␣ were prepared by a cholate dialysis method as described (21). By using a hydroxylapatite column, detergent-free OMb was also prepared. Rat b 5 was solubilized from liver microsomes and purified with the same method as OMb. NADPH-cytochrome P-450 reductase (P-450 reductase) was solubilized from liver microsomes of phenobarbital-treated rats and purified as described (22). Guinea pig P-450 17␣ , OMb, and b 5 were purified from the microsomal fractions of the adrenal glands with the same method as the rat enzyme. The OMb and b 5 were separated from each other by the TSK DEAE-5PW and hydroxylapatite column chromatographies.
Construct of Recombinant OMb and Its Expression in E. coli-An amino-terminal hexahistidine-tagged OMb construct was created by a PCR. The amplified cDNA was inserted into pET-3d vectors at the NdeI and BamHI sites to yield pET-hisOMb. BL21 (DE3) strain transformed with pET-hisOMb was cultured for 40 h at 25°C, and then the transformant was harvested by centrifugation at 1,000 ϫ g for 15 min. After sonication the suspension was centrifuged at 177,100 ϫ g for 1 h. The pellets were solubilized in 10 mM Hepes-KOH (pH 7.5) buffer containing 20% glycerol and 2% Triton X-100 for 1 h at 4°C. After centrifugation at 177,100 ϫ g for 1 h, the resultant supernatant was applied to a Ni 2ϩ -preloaded HisTrap chelating column (1 ml, Amersham Biosciences AB) equilibrated with 10 mM Hepes-KOH (pH 7.5) buffer containing 20% glycerol, 0.1% Triton X-100, 10 mM imidazole, and 0.5 M NaCl. After washed with the equilibrating buffer, the column was eluted with the same buffer except containing 0.5 M imidazole. Because effects of the recombinant OMb on the P-450 17␣ -catalyzed reaction were indistinguishable from those of the native forms that were purified from testes and livers, the recombinant OMb was mainly used in this study.
Measurement of 17␣-Hydroxylase and C17-C20 Lyase Activities of P-450 17␣ -P-450 17␣ containing proteoliposomes were preincubated with P-450 reductase and either OMb or b 5 in a volume of 0.05 ml of 50 mM K-P i buffer (pH 7.4) containing 20% glycerol, 0.1 mM EDTA, and 0.1 mM dithiothreitol for 60 min at 0°C. After diluting 10-fold with 20 mM Hepes-NaOH buffer (pH 7.4), progesterone in propylene glycol was added to the reaction mixture. Reaction was started with addition of NADPH at a final concentration of 1 mM and proceeded for 10 -20 min at 30°C. The reaction was stopped by addition of 2 ml of n-hexane, and the products were extracted twice with the solvent. After drying under a stream of N 2 gas, the steroids were separated by high pressure liquid chromatography using a COSMOSIL 5SL-II column (4.6 ϫ 150 mm, Nacalai Tesque, Kyoto, Japan) equilibrated and eluted with a solvent (n-hexane:2-propanol:acetic acid ϭ 95:5:1). Androstenedione (lyase activity) and 17␣-hydroxyprogesterone (17␣-hydroxylase activity) were quantified by measuring their absorbance at 240 nm. In some experiments, 17␣-hydroxyprogesterone was used as the substrate.
Enzyme and Protein Concentrations-The concentrations of P-450 17␣ were determined spectrophotometrically using a molecular coefficient of 91 mM Ϫ1 ⅐cm Ϫ1 for the CO difference spectrum of the reduced form (23). The concentrations of OMb and b 5 were determined from the reduced minus oxidized difference spectrum between 424 and 409 nm by using a molecular coefficient of 185 mM Ϫ1 ⅐cm Ϫ1 of b 5 assuming that spectrographic properties of detergent-solubilized forms of OMb and b 5 were indistinguishable from that of a trypsin-solubilized form of b 5 at a Soret band (23,24). The concentration of P-450 reductase was determined from the absorbance at 455 nm by using a molecular coefficient of 21.4 mM Ϫ1 ⅐cm Ϫ1 (25). Amounts of P-450 reductase in rat testis microsomes were analyzed by immunoblotting using anti-rat P-450 reductase-IgG and goat anti-rabbit IgG-horseradish peroxidase complex. The chemiluminescence of the immunoreactive bands was photographed with Light Capture (model AE-6969N, Atto Instruments, Tokyo, Japan) and quantified using CS Analyzer (version 2.0, Atto Instruments). Proteins were determined by a bicinchoninic acid (BCA) method (26).
Immmunohistochemical Localization of OMb and P-450 17␣ in Rat Testis-Testes were excised from adult male rats (Sprague-Dawley) and immediately fixed in 4% paraformaldehyde solution buffered with 10 mM phosphate, pH 7.4, at 4°C overnight. Fixed tissues were then dehydrated and embedded in paraffin. Three-micron sections were deparaffinized and hydrated using graded alcohol concentrations for standard indirect peroxidase immunohistochemistry. In brief, the preparations were first treated with 0.5% hydrogen peroxide in Na-P ibuffered saline (PBS), pH 7.4, for 20 min at room temperature to block the endogenous peroxidases and then with 5% of normal goat serum in PBS for 20 min at room temperature to block the nonspecific binding. This was followed by overnight incubation at 4°C with the first antibodies against rat OMb or guinea pig P-450 17␣ that had been incubated with 10% normal rat serum for 60 min at 0°C. The preparations were then incubated for 2 h at room temperature with horseradish peroxidase-conjugated goat FabЈ fragment to rabbit immunoglobulin G (ICN Pharmaceuticals, Inc., Aurora, OH). The immunoreactive proteins were visualized with 3,3Ј-diaminobenzidine tetrahydrochloride (Sigma) and hydrogen peroxide as described before (27). Nuclear counterstaining was performed using methyl green (Lab Vision Corp., Fremont, CA). For controls, the first antibody, anti-OMb or anti-P-450 17a IgG, was omitted from all procedures described above. The antibody against P-450 17␣ was a generous gift from Dr. S. Kominami of Hiroshima University.

Localization of OMb in Rat Testis Cells-
The antibodies used in this study are very specific. No cross-reaction between OMb and anti-b 5 IgG nor between b 5 and anti-OMb IgG was observed if appropriate application of the samples and appropriate dilution of the antibodies were employed. Fig. 1 shows that OMb was distributed between mitochondrial and microsomal fractions almost at a ratio of 1:1 on a protein concentration basis in rat testicular cells, where b 5 was not detected. A large amount of the sample loading on the SDS-PAGE scarcely detected the b 5 band, which is less than 1/100 that of liver cells. Amounts of OMb present in testes were comparable with those in liver per protein. Dual localization (mitochondria and microsomes) of OMb was also observed in the kidneys and adrenal glands (data not shown). Although localization of OMb in the liver cells had been believed to be strictly the outer mitochondrial membrane, translocation of OMb from there to the microsomal membranes was observed at 12 h after administration of phenobarbital to rats. Because this reagent induces not only CYP2A but also CYP3A families, the amount of CYP3A2 was increased in the microsomal fraction. Translocation of OMb to the microsomes was reproduced with other P-450-inducible agents such as dexamethasone, and pregnenolone-16␣-carbonitrile (data not shown). Rat OMb thus has a capability to reside in both the outer mitochondrial and microsomal membranes. In the testicular cells, translocation of OMb as manifested in the liver cells was not observed as tested so far by administration of phenobarbital or dexamethasone. In contrast to OMb, b 5 did not change its localization in liver and other cells. Precise results and discussion about the translocation of OMb in liver cells are to be published elsewhere. 2 Purification of Enzymes Used in This Study- Fig. 2A shows the purity of the enzyme preparations used in this study. We purified P-450 17␣ from rat testis microsomes and both b 5 and P-450 reductase from rat liver microsomes. We mainly used a recombinant OMb because this form had the same effects on rat P-450 17␣ with the native form that was purified from the testis microsomes. Use of the recombinant form allowed us to perform the experiments with a constant and sufficient supply of OMb. The data presented herein were collected by using the recombinant OMb instead of the purified OMb. Guinea pig enzymes were purified from the adrenal glands because the P-450 17␣ , b 5 , and OMb were probably most abundant in the tissue. The specific content of P-450 17␣ in rat testis microsomes was 0.1 nmol/mg protein as determined from the CO difference spectrum. Contribution of P-450 scc , a mitochondrial protein, to the spectrum is negligible for the microsomal preparation as verified from immunoblotting with anti-P-450 scc IgG (data not shown). The specific content of P-450 reductase was determined from the results of Fig. 2B to be 1.2 Ϯ 0.3 pmol/mg protein. OMb was present in the testis microsomes at an amount of 0.02-0.05 nmol/mg protein as determined from the reduced minus oxidized difference spectrum. The measurement is, however, not accurate in a crude microsomal fraction with limited amounts of b 5 -type protein.
Efficient Electron Transfer Enhances Lyase Activity of P-450 17␣ -By using excess amounts of progesterone as the substrate, we can measure both 17␣-hydroxylase and lyase activities simultaneously because P-450 17␣ first converts the steroid to the 17␣-hydroxylated intermediate, 17␣-hydroxyprogesterone, and then the second P-450-catalyzed reaction cleaves the C17-C20 bond to give androstenedione without releasing the intermediate. Kinetic studies employing a rapid quenching method proved that androgen formation from progesterone by guinea pig P-450 17␣ and dehydroepiandrosterone formation from pregnenolone by bovine P-450 17␣ proceeded successively (28,29). In the successive reaction, rates of release of the intermediate from the substrate-binding pocket in the enzyme and the C-C bond cleavage together with release of the C17 ketosteroid determine yields of both the 17␣-hydroxyl and 17-ketosteroids. Based on this assumption, efficient supply of electrons from P-450 reductase to P-450 17␣ should facilitate the cleavage reaction. This assumption is valid as shown in Fig. 3. Production of androstenedione was increased in a linear fashion with the increase of P-450 reductase added (Fig. 3A). The amount of 17␣-hydroxyprogesterone produced, however, was constant in the range tested. These indicate that the first hydroxylation step is relatively fast and that the following steps (i.e. the release of the intermediate and C-C bond cleavage) are rate-limiting as shown previously (28). The increase of the electron transfer should enhance the second P-450-catalyzed reaction (i.e. C17-C20 bond cleavage) leading to androstenedione formation without affecting the release of the intermediate. The lyase activity was about 1/10th the 17␣hydroxylase activity at a low ratio of P-450 reductase to P-450 17␣ (reductase:P-450 17␣ of 1:100), and it was more than 2.5 times the 17␣-hydroxylase activity at a ratio of P-450 reductase to P-450 17␣ of 0.5. Obviously, rat P-450 17␣ has a lyase activity that is higher than the 17␣-hydroxylase activity compared with the guinea pig enzyme (see Refs. 28 and 30 and see the text below).
Effects of OMb and b 5 on 17␣-Hydroxylase and Lyase of Rat P-450 17␣ -Addition of OMb to the P-450 17␣ -dependent reaction system enhanced both the 17␣-hydroxylase and lyase activities over the range of ratios of P-450 reductase to P-450 17␣ tested. The stimulatory effects, however, tended to be satu-rated at a low ratio of OMb to P-450 17␣ when the electron transfer was slow due to a low concentration of P-450 reductase (compare the profiles in Fig. 3, B-D). Because the amount of the reductase is very low in the testis microsomes as revealed by immunoblotting with anti-P-450 reductase IgG (Fig. 2B), we hereafter studied the effects of OMb at a ratio of P-450 reductase:P-450 17␣ of 1:50 unless otherwise mentioned. OMb stimulated the 17␣-hydroxylase activity linearly at ratios of OMb to P-450 17␣ below 0.1 and continued to increase it gradually up to the ratio of 1.0 to an extent of 7-8-fold (Fig. 4A). Stimulation of the lyase activity was also observed but reached a plateau at a ratio of OMb to P-450 17␣ of about 0.1. Above the molar ratio of OMb:P-450 17␣ of 1.0, the lyase activity was almost constant up to the ratio of 2.0, whereas the 17␣-hydroxylase activity continued to increase gradually (data not shown). Maximum stimulation of the lyase activity by OMb was about 3-fold. Addition of b 5 to the reaction system containing P-450 17␣ and P-450 reductase also augmented both lyase and 17␣-hydroxylase activities but with a manner different than that of OMb (Fig. 4B). It stimulated the lyase activity to an extent similar to that of OMb (3-fold) at ratios of b 5 to P-450 17␣ of 0.2-0.5, whereas it enhanced the 17␣-hydroxylase activity about twice at the most. At the high ratio, the 17␣-hydroxylase and lyase activities dropped to the same value and one-third of the original (without b 5 ) activity, respectively (Fig. 4B), and further decreased above the ratio of 1.0 (data not shown). The differences in the stimulatory effects between OMb and b 5 were apparent when the 17␣-hydroxylase:lyase activity ratio was plotted in the y axis (Fig. 4C). OMb had a tendency to elevate preferentially the 17␣-hydroxylase, whereas b 5 stimulated the lyase preferentially. A rise of the 17␣-hydroxylase:lyase at a ratio of b 5 to P-450 17a of 1.0 was due to the strong inhibitory effect on the lyase in a high concentration of b 5 relative to P-450 17␣ .
Effects of OMb and b 5 on 17␣-Hydroxylase and Lyase of Guinea Pig P-450 17␣ -In general, P-450 17␣ of most animals has strong 17␣-hydroxylase and weak lyase activities in the absence of b 5 except murine P-450 17␣ . In the guinea pig P-450 17␣catalyzed reaction, b 5 has been reported to stimulate the lyase activity without significant elevation of the 17␣-hydroxylase activity (30). We purified b 5 and OMb from microsomes of guinea pig adrenals and analyzed the effects of these proteins toward the P-450 17␣ that had also been purified from the same membrane preparations and embedded in proteoliposomes. Rat b 5 greatly stimulated the lyase activity up to a ratio of b 5 to guinea pig P-450 17␣ of 0.5 together with slight activation of the 17␣-hydroxylase. However, it reduced both activities at the high ratio (Fig. 5B). The stimulatory profile of rat b 5 on guinea pig P-450 17␣ was essentially the same on rat P-450 17␣ except that activation of the lyase was more dominant than that of the 17␣-hydroxylase and that the maximum stimulation of both activities occurred at lower ratios, b 5 :P-450 17␣ of 0.05-0.1. At a ratio of rat b 5 to guinea pig P-450 17␣ of 0.5, the increase in the lyase activity was about 6-fold and that of the 17␣-hydroxylase was 1.5-fold. Guinea pig b 5 also greatly stimulated the lyase activity up to the ratio of 0.5, where the activity was 6-fold and the effect reached a plateau (Fig. 5E). The 17␣-hydroxylase activity continued to increase up to a ratio of guinea pig b 5 to guinea pig P-450 17␣ of 1.0. In contrast, the stimulatory effect of rat OMb on guinea pig P-450 17␣ was not observed over the range measured (Fig. 5A). Only activation of the 17␣-hydroxylase was observed in an OMb-dependent manner up to a ratio of rat OMb to guinea pig P-450 17␣ of 0.5, where stimulation was about 3.5-fold. The effect of guinea pig OMb on guinea pig P-450 17␣ was intermediate between those of rat OMb and guinea pig b 5 (Fig. 5D). Guinea pig OMb stimulated the 17␣hydroxylase and lyase activities about 3-and 4.5-fold, respec-

FIG. 3. Effects of P-450 reductase and OMb on the rat P-450 17␣catalyzed reactions.
A, purified rat P-450 17␣ was incorporated into egg yolk phosphatidylcholine liposomes. The proteoliposomes were incubated with various amounts of P-450 reductase for 60 min at 0°C. The enzyme reaction was performed, and the products were quantified as described under "Experimental Procedures." B-D, the proteoliposomes were incubated with a fixed amount of P-450 reductase and various amounts of OMb for 60 min at 0°C. The enzyme reaction was performed, and the products were quantified as described under "Experimental Procedures." Molar ratio of P-450 reductase:P-450 17␣ , 0.01 (B), 0.1 (C), and 1.0 (D). Ⅺ, androstenedione; q, 17␣-hydroxyprogesterone. tively, at a ratio of OMb:P-450 17␣ of 1:1. There is a more distinct difference in the 17␣-hydroxylase:lyase ratio between the stimulatory effects on guinea pig P-450 17␣ by rat OMb and b 5 (Fig. 5C).
Effects of OMb and b 5 on Lyase Activity toward 17␣-Hydroxyprogesterone-We then directly analyzed the C17-C20 lyase activities toward 17␣-hydroxyprogesterone using the steroid as the substrate (Fig. 6). Rat b 5 augmented the lyase activity of rat P-450 17␣ up to a ratio of b 5 to P-450 17␣ of 0.5 about 4-fold, but the activity drastically dropped at the ratio of 1.0. Such a stimulatory profile was essentially the same on guinea pig P-450 17␣ by rat b 5 . Rat OMb stimulated the lyase activity as the ratio to rat P-450 17␣ increased, and the activation reached 5-fold at the ratio of 1.0. As shown in the successive reaction with an excess amount of progesterone (Fig. 5A), the inability of rat OMb to exert a stimulatory effect on the lyase activity of guinea pig P-450 17␣ toward 17␣-hydroxyprogesterone was reproduced.
Effect of OMb on P-450 17␣ -catalyzed Reaction at a Low Concentration of Progesterone-The experiments on the effects of OMb and b 5 , which were described above, were performed with a saturating concentration of progesterone (0.1 mM). Under this condition, we measured initial velocities of both 17␣-hydroxylase activity and successive 17␣-hydroxylase and C17-C20 lyase activities of P-450 17␣ at the same time. We then conducted the P-450 17␣ -catalyzed reaction at a low concentration of progesterone. Under such a condition, a mixture of the successive reaction and two-step reaction would proceed, i.e. at the first step 17␣-hydroxylation of progesterone and subsequent release of 17␣-hydroxyprogesterone from the enzyme pocket and at the second step reincorporation of the intermediate and  6. Effects of rat OMb and b 5 on lyase activity of P-450 17␣ toward 17␣-hydroxyprogesterone. The rat P-450 17␣ -containing proteoliposomes were incubated with P-450 reductase and rat OMb (q) or rat b 5 (E) for 60 min at 0°C. The guinea pig P-450 17␣ -containing proteoliposomes were incubated with P-450 reductase and rat OMb (f) or rat b 5 (Ⅺ) for 60 min at 0°C. The enzyme reaction was performed with 0.1 mM of 17␣-hydroxyprogesterone for 20 min, and the products were quantified as described under "Experimental Procedures." its C17-C20 cleavage. At 1.5 M progesterone, rat P-450 17a alone produced a small amount of androstenedione together with a little more 17␣-hydroxyprogesterone (Fig. 7). In the presence of rat OMb (0.5 eq), P-450 17a rapidly converted progesterone mainly to 17␣-hydroxyprogesterone and secondly to androstenedione reflecting the product ratio under an excess amount of progesterone as shown in Fig. 4A. The amount of 17␣Ϫhydroxyprogesterone was constant between 30 and 60 min, during which period the input and output were equilibrated and then started to decline, and the intermediate was finally almost consumed. The rate of decrease of progesterone became slow after 60 min and seemed to have stopped at 120 min. On the other hand, production of androstenedione accelerated after 30 min and persisted up to 120 min, at which time point most of the substrate and intermediate were consumed. Addition of rat b 5 to the P-450 17␣ -catalyzed reaction system stimulated productions of 17␣-hydroxyprogesterone as much as that of OMb during the first 30-min reaction and facilitated a rapid consumption of the intermediate after 30 min. It enhanced androstenedione formation more than that of OMb at the first 30-min reaction, and the formation accelerated to 60 min and had almost stopped by that time probably because of exhaustion of usable substrates (progesterone and 17␣-hydroxyprogesterone). The reason for a significant amount of progesterone remaining even after 180 min of incubation could be attributed to an inhibitory effect of androstenedione on the P-450 17␣ -catalyzed reaction. Although androstenedione formation facilitated by b 5 was faster than that by OMb, amounts of the product at 120 min were almost equal. After the reaction for 180 min, the androstenedione formation stimulated by OMb was more than five times that by P-450 17␣ alone. Effects of rat OMb and b 5 on guinea pig P-450 17␣ were also examined in the reactions at a low concentration of progesterone. Rat b 5 stimulated androstenedione formation by guinea pig P-450 17␣ in a fashion similar to that by rat P-450 17␣ . In contrast, rat OMb did not exert any stimulatory effects on the production despite a significant activation of 17␣-hydroxyprogesterone production (data not shown).
Localization OMb in Leydig Cells of Rat Testis-Because testicular cells are heterogeneous, the following question arises: Is OMb really present in the steroidogenic cells, i.e. Leydig cells? To solve this, we conducted immunohistochemical analysis using anti-rat OMb IgG on the rat testis. OMb as well as P-450 17␣ was only detected in Leydig cells; no positive reactions were observed in Sertoli cells, spermatogonia, spermatocytes, spermatids, and spermatozoa (Fig. 8). In contrast, cells were not significantly stained with anti-rat b 5 IgG (data not shown). Localization of OMb in Leydig cells was also confirmed with the guinea pig testis (data not shown). DISCUSSION Although the roles of OMb have not been known, those of b 5 are considerably well known. They are as follows: (a) transfer of electrons from NADH to desaturases such as stearoyl-CoA desaturase (10) and ⌬7-sterol 5-desaturase (31); (b) NADH-dependent reduction of methemoglobin to regenerate hemoglobin (32), and (c) stimulation of some P-450-dependent oxygenation. The intracellular localization of b 5 is absolutely ER, whereas that of OMb is variable between outer mitochondrial and ER membranes from cells to cells or animals to animals. Furthermore, OMb changes its localization at least in hepatic cells under different conditions (Fig. 1). In Leydig cells of rat testes, OMb is distributed between outer mitochondrial and ER membranes. In contrast, b 5 , which is believed to be responsible for stimulation of the P-450 17␣ of several animals other than rats in efficient production of androgen, is scarcely detectable in rat testicular cells. In the present study, we analyzed the effects of rat OMb on P-450 17␣ purified from rat testis microsomes using a reconstituted system with proteoliposomes. This system is believed to provide membrane proteins with a reaction field comparable with the biomembrane environment and allowed us to perform quantitative analyses in such a way as controlling the enzyme, modulator, and substrate concentrations. Androgen syntheses from progesterone by guinea pig P-450 17␣ (28) and from pregnenolone by bovine P-450 17␣ (29) are proven to proceed by a successive reaction in the presence of excess amounts of progesterone, in which efficient electron supplies from P-450 reductase to P-450 facilitate the C17-C20 cleavage reaction of the 17␣-hydroxylated intermediates bound to the enzyme pocket. The stimulatory effect of OMb on rat P-450 17␣catalyzed reactions was also observed (Fig. 3). In the rat testis, the amount of P-450 reductase, however, is limited, i.e. from 1/50 to 1/100 of P-450 17␣ (Fig. 2). We thus analyzed the effects of b 5 and OMb at a ratio of reductase:P-450 of 0.02:1.0 assuming that the physiological electron transfer between the reductase and P-450 in the rat testis ER operates at about this ratio. Although different from the P-450 17␣ of other animals, the lyase activity is higher than the 17␣-hydroxylase activity in rat P-450 17␣ if sufficient electrons are supplied, it becomes drastically reduced under the physiological concentration of P-450 reductase. In such a low concentration of P-450 reductase, androgen production is very slow probably due to the dominant release of the 17␣-hydroxylated intermediate from the enzyme pocket, which becomes relatively faster than the second twoelectron transfers needed for the C17-C20 cleavage reaction. Androstenedione produced at an initial rate from a saturation amount of progesterone by rat P-450 17␣ in the presence of the physiological concentration of P-450 reductase is very small in the absence of b 5 or OMb (Fig. 4). Moreover, the majority of products from progesterone is 17␣-hydroxyprogesterone, which is about twice of androstenedione (Fig. 4) and assumed not to be converted to androstenedione in the presence of an excess amount of progesterone in the successive reaction model (29). Rat b 5 from the liver virtually acted on the P-450 17␣ by compensating for the limited reduction. It elevated the lyase activity to a level corresponding to that at a ratio of P-450 reductase: P-450 of 0.1:1.0 in the absence of b 5 . This effect is, however, saturated at a low ratio of b 5 to P-450 17␣ and even inhibitory to the P-450 17␣ -catalyzed reaction at the high ratios if the amount of P-450 reductase in the reaction system is limited. The reason for such an inhibitory action of b 5 is unclear but could be attributed to some interaction of b 5 with the reductase or even with P-450 17␣ in a complicated manner. Rat OMb also showed stimulatory effects on rat P-450 17␣ with some different manners. The lyase activity was increased linearly to the amount of OMb at low ratios of OMb to P-450 17␣ and reached a constant around the ratio of 0.2 without a significant decrease at the high ratios. Increase in the 17␣-hydroxylase activity was larger than that in the lyase activity, and the increase curve also tended to saturate around the ratio of 0.2 but proceeded gradually to the ratio of 1.0 and more. Differences of the stimulation profiles for P-450 17␣ between b 5 and OMb was more obvious when they were evaluated for effects on guinea pig P-450 17␣ , which has a very high 17␣-hydroxylase activity relative to the lyase activity. The effects of rat b 5 were essentially the same for rat P-450 17␣ , whereas rat OMb only activated the 17␣-hydroxylase activity without elevation of the lyase activity. Although a reaction system of guinea pig P-450 17␣ with rat b 5 or rat OMb is non-physiological, it disclosed differences of the effects of two similar hemoproteins clearly.
In rat testicular cells, OMb is present nearly equally in both the mitochondria and microsomes with a concentration comparable with that in liver mitochondria. By immunohistochemical analyses, the rat testicular OMb was localized in Leydig cells, where P-450 17␣ coexisted. In contrast, we scarcely detected b 5 in the rat testis microsomes, not to mention its presence in the mitochondria, by cell fractionation. Immunohistochemical analysis using anti-rat b 5 IgG did not detect significant staining of b 5 in Leydig cells.
There have been numerous reports on stimulation of b 5 on P-450 17␣ except rat enzymes employing various assay systems (14 -16, 33). Most of them confirmed the function as a stimulator for the lyase. There is no study, however, on involvement of OMb in the androgen synthesis except a recent one on human OMb (type 2 cyt-b 5 ) and P-450 17␣ both expressed transiently in HEK-293c cells (34). In the present study employing enzymatic and cell biological analyses, we were able to demonstrate the role of OMb and not b 5 as a physiological modulator for rat testicular P-450 17␣ . Although rat OMb preferentially stimulates 17␣-hydroxyprogesterone production from progesterone, it should be an activator for androgen synthesis by P-450 17␣ . In an excess amount of progesterone, rat P-450 17␣ produces androstenedione by the successive reaction from the substrate. This situation, however, could not be feasible under physiological conditions. In steroidogenic cells, cholesterol, the precursor of steroidogenesis, is supplied from a pool in the outer mitochondrial membrane to the inner membrane where P-450 SCC is present through action of steroidogenic acute regulatory (StAR) protein (35). Activation of StAR protein is triggered by the stimulus of pituitary hormones such as luteinizing hormone and ACTH. The cholesterol is rapidly converted to pregnenolone by the rate-limiting enzyme, P-450 SCC , and the pregnenolone is then oxidized by 3␤-hydroxysteroid dehydrogenase to give a transient increase of progesterone. The produced progesterone is promptly consumed to the lower stream products because steroidogenetic enzyme systems in general have no rate-limiting step other then the side chain cleavage. Such a flow of steroids would be reflected by the reaction condition with a low amount of progesterone as conducted in the experiment of Fig. 7. At a low concentration of progesterone, OMb stimulates rat P-450 17␣ to produce both 17␣-hydroxyprogesterone and androstenedione, in which 17␣-hydroxyprogesterone production is preferential. As the result, 17␣-hydroxyprogesterone accumulates and then serves as the second but main substrate. The intermediate steroid is then converted to androstenedione, which process is also stimulated by OMb. Although the rate of androgen production stimulated by b 5 is about twice as fast as that by OMb, the amount at 120 -180 min was nearly the same for both hemoproteins. A question then arises as to why OMb and not b 5 acts as the modulator for androgen production in the rat testis. One reason could be the inhibitory effects of b 5 on P-450 17␣ at a high ratio of b 5 per P-450 17␣ . In this respect, OMb does not inhibit P-450 17␣ activity at the ratio of 1.0. The early study by Onoda and Hall (14) pointed out difference of the effects between pig liver and newborn testis b 5 on androgen production in adult pig testis microsomes. The testicular b 5 stimulated concentrationdependently the lyase activity, whereas the liver b 5 had an optimum concentration for the maximum stimulation. Rat OMb has a tendency to stimulate preferentially the 17␣-hydroxylase activity of P-450 17␣ . In fact, rat P-450 17␣ has a week 17␣-hydroxylase activity relative to the P-450 17␣ of other animals. A high 17␣-hydroxylase activity is necessary for cortisol production by other animals than murines, which do not produce cortisol due to the absence of P-450 17␣ in the adrenal glands. Preferential stimulation of the 17␣-hydroxylase activity by rat OMb suggests that the considerably high 17␣-hydroxylase should be needed for an unknown reason in Leydig cells. Finally, the mobile property of OMb in rat liver cells (Fig. 1) could also have a physiological role as a modulator that transfers between mitochondria (inactive) and ER (active) in Leydig cells, although we have not analyzed translocation of OMb in rat testicular cells.
There are essentially two possible explanations for the stimulatory effects of b 5 on some P-450-catalyzed reactions. One is that b 5 is involved in the second electron transfer during the P-450 reaction cycle. The participation in the first electron transfer is not practical because the redox potential of the ferric state of P-450 is lower than that of the ferrous state of b 5 . For the electron transfer hypothesis, the stimulatory effect of apo-b 5 on testosterone 6␤-hydroxylation and nifedipine oxida- tion by CYP3A4, however, disapproved (16). A second explanation is to give P-450 conformational changes through complex formation between the P-450 and b 5 . Thus, the P-450 obtains a stimulated catalytic activity in the complex. The stimulatory profile of P-450 17␣ activities in the presence of b 5 thus suggests some interaction between these proteins but with complexity ( Figs. 4B and 5B). The effect of the amount of P-450 reductase on the lyase activity of P-450 17␣ in the present study (Fig. 3A) does not seem in conflict with the first hypothesis. The more efficiently the electrons are supplied, the more rapidly the second P-450-catalyzed reaction (C17-C20 cleavage) proceeds, and thereby production of androstenedione overtakes the release of 17␣-hydroxyprogesterone from the enzyme pocket. The effects of OMb in the preferential elevation of 17␣-hydroxylation, however, are not accountable with the first hypothesis because the efficient electron supply would not facilitate the release of 17␣-hydroxyprogesterone. Yamazaki et al. (29) reported that the release of the intermediate (17␣-hydroxyprogesterone), the C17-C20 cleavage reaction, and the release of the final product (androstenedione) are all very slow relative to the first 17␣-hydroxylation. This means that 17␣-hydroxylase and lyase activities are mainly determined by the rate of the intermediate release and the rates of the C-C bond cleavage and final product release, respectively. If the first hydroxylation step (17␣-hydroxylation) is fast, the overall rate does not change even if the first step is influenced by OMb or b 5 . Thus, at least the increase of the rate for the release of 17␣-hydroxyprogesterone is impractical by activation of the electron transfer. It is likely that rat OMb has abilities to facilitate the release of 17␣-hydroxyprogesterone and C17-C20 cleavage reaction and/or release of androstenedione. This is the first demonstration of the physiological role of OMb. The functions in other tissues as well as the mechanism and function of the translocation between mitochondria and ER should be elucidated.