Multiple molecular forms of cytochrome P-450SCC purified from bovine corpus luteum mitochondria

Biochimica et Biophysica Acta, 994 (1989) 235-245 Elsevier

235

BBA 33308

Multiple molecular forms of cytochrome P-450sc c purified from bovine corpus luteum mitochondfia Sachiko Sugano 1.,, Mitsuhiro Okamoto 3, Hodaka Ikeda ~, Naosada Takizawa 2 and Shigeo Horie Departments of I Biochemistry and z Physiology,, School of Medicine, Kitasato University, Kanagawa and 3 Departw~.entof Molecular Physiological Chemistry, Osaka University Medical School (Japan)

(Received 26 July 1988)

Key words: Cytochrome P.450scc; Purification; Multiple molecular form; (Bovine cc~us luteum); (Mitochondrion)

Cytochrome P-450 related to side.chain cleavage of cholesterol (P-450sc c) was isolated from bovine corpus luteum mitochondria in the form of its stable cholesterol complex. The isolation procedure included ammonium sulfate |ractionafion and chromatography on ¢~-aminohexyl-Sepharose (AH-Sepharose). Corpus iuteum P-450sc c was resolved into one minor (AH-|) and two major (AH-|| and AH-|||) fractions by the chromatography. Results of re-chromatography suggested the possibility that AH-|H Fraction was originally complexed with lipidic material. The two major fractions purified by the re-chromatography (AH-IIR and AH-IHR Fractions) showed essentially a single band on sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis and their absorption spectra were indistinguishable from each other. Both fractions were further resolved into two major and some minor bands of P-450sc c by isoelectric focusing on polyacrylamlde gel in the presence of a non-ionic detergent, as detected by protein staining, berne staining and immunoblot analysis with anti-bovine P-450sc c monoclonal antibody. Both AH-IIR and AH-HIR Fractions were further resolved by high-performance liquid chromatography (HPLC) on SP-TSK gel column into two fractions, SP-| and SP-ll. These fractions had the same N-terminal amino acid sequence, showed similar catalytic activity and resolved into one major and a few minor hands on isoelectr|c focusing on polyacrylamide gel. Much more heterogeneity was observed in purified P-450sc c preparations from bovine adrenal cortex mitochondria. These results indicated the presence of multiple molecular forms of corpus luteun P-450sc c as well as adrenal cortex P-450sc c. Computer simulation studies were carried out in order to analyze the mechanism of formation of multiple bands on isoelectric focusing. The multiple bands of corpus luteum P-450~scc could be explained by postulating the presence of two isozymes (or molecular forms) having a pair of sites each with or without a charged group.

Introduction Corpus luteum mitochondria contain P-450 [1] and the P-450 has been solubilized [2] and purified to near homogeneity in the cholesterol-free low-spin form [3]. The only P-450 species present in mitochondria of the corpus luteum is P-450sc c related to side-chain cleavage of cholesterol [4,5]. Although considerable evi-

* On leave from Department of Biochemistry, Osaka University Medical School. Present address: School of Nursing, Kitasto Univezsity, Sagamihara, Kanagawa 228, Japan. Abbreviations: SDS, sodium dodecyl sulfate; HPLC, high-preformance liquid chromatography; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-l-propanesidfonate; PAS, periodic acid-Schiff. Correspondence: S. Horie, Department of Biochemistry, School of Medicine, Kitasato University, Sagamihara, Kanagawa 228, Japan.

dence has been presented by several authors [6-11] for the presence of more than one molecular form of P450scc in adrenal cortex mitochondria, a systematic study has not been carried out on the multiplicity of molecular forms of corpus luteum P-450scc. In the present paper, a simple procedure is described for the purification of corpus luteum P-450scc in the form its stable cholesterol complex. The presence of multiple molecular forms of the P-450scc is also reported. Materials and Methods

Preparation of mitochondria from corpus luteum Corpora lutea were collected from bovine ovaries, supplied by a local slaughter house, and stored at - 1 5 ° C . Mitochondria were prepared from the frozen tissue in the same way as described previously for the

0167-4838/89/$03.50 © 1989 Elsevier Science Publishers B.V. (Biomedical Division)

236

preparation from bovine adrenal cortex [6,12]. All manipulations were carried out at below 5°C. The nfitochondria were sonicated and the resulting particulate material was collected by centrifugation at 100000 × g for 60 min [6].

Extraction and ammonium sulfate fractionation of corpus luteum P-450scc The sonicated mitochondria obtained from 2160 g of the tissue were combined and suspended in 0.1 M sodium phosphate buffer (pH 7.0) containing 0.1 mM sodium-EDTA at a protein concentration of 30 mg per ml. 9 ml of 10% (w/v) sodium cholate solution containing 0.05% (w/v) cholesterol were added per 100 ml of the suspension. The mixture was stirred for I h, centrifuged at I00000 × g for 90 min, and the clear middle supematant [12] was collected. To remove excess cholate, the cholate extract was passed through a Sephadex G-25 column (5 × 95 cm) equilibrated with the same phosphate buffer. To the semi-turbid fraction, 0.861 vol. of 80% saturated ammonium sulfate solution (pH 8.0) was added to obtain a fmal concentration of 37% saturation [12]. The precipitate (37% S Fraction) was collected by centrifugation at 11000 x g for 30 rain, suspended in 150 ml of the phosphate buffer, and the protein concentration was adjusted to 10 mg/ml. This fraction was then subjected to ammonium sulfate fractionation as described previously [6], except that the fractionation was carried out in the presence of 0.3 mg of sodium cholate and 0.0015 m$ of cholesterol per mg of protein, and also that the finally obtained fraction was that precipitating between 33 and 58% saturation of ammonium sulfate (33-58% S Fraction).

Preparation of bovine adrenocortical P-450scc Two kinds of adrenocortical P-450sc c were prepared. One preparation (AH-IIG Fraction) was prepared in the same way as described for corpus luteum P-450sco except that the second fraction eluted from the AH-Sepharose column with buffer A was collected (elution position: 120-140 ml) and, instead of the rechromatography, was further purified by gel filtration on Toyopearl HW55 superfine column (1.6 x 95 cm) equilibrated with buffer A. The other preparation, referred to as standard P-450scc preparation, was prepared by the method of Takikawa et al. [13].

High-performance liquid chromatography (HPLC) of P450scc on sulfopropyl-TgK gel HPLC of P-450scc (AH-IIR or AH-IIIR Fraction) was carried out on a freshly purchased column of suifopropyi-TSK gel (SP-5PW, 0.8 × 7.5 cm, supplied by TOSO Company, Tokyo) equilibrated with 10 mM potassium phosphate buffer, (pH 7.4) containing 20% glycerol, 0.1% Emulgen 913 and 0.1 mM EDTA. Less than 20 nmol of the sample dissolved in the same buffer was applied. The adsorbed P-450sc c was eluted by a linear gradient to 0.5 M NaCI in the same buffer. The flow rate was 0.4 mi/min, and the effluent was monitored at a wavelength of 400 nm. P-450scc was eluted in the low-spin form. Repeated use of the column resulted in poor adsorption of the sample. The apparatus used was an LKB Ultrochrome GTi system equipped with a type 2150 HPLC pump and a type 2151 variable wavelength monitor.

Chromatography of P-450scc on AH.Sepharose

SDS-polyacrylamide gel electrophoresis and isoelectric focusing on polyacrylamide gel

33-58% S Fraction was dissolved in a minimum volume of 0.05 M potassium phosphate buffer (pH 7.4) containing 0.3% (w/v) sodium cholate and applied to an AH-Sepharose column (!.6 × 33 cm) equilibrated with 0.05 M potassium phosphate buffer (pH 7.4) containin8 20% (v/v) glycerol, 0.05 mM EDTA, 0.3% (w/v) sodium cholate and 0.0015% (w/v) cholesterol (buffer A). The preparation was eluted first with 200 ml of buffer A and then by a linear gradient to 0.16% (w/v) Emulgen 913. The mixer initia._lly contained 300 ml of buffer A and the reservoir initially contained 300 ml of buffer A to which 0.16% Emulgen 913 was added. The flow rate was about 10 m l / h and the absorbance of the effluent was monitored at 400 and 280 nm. The resultin8 fractions of P-450scc were concentrated by membrane filtration and purified by re-chromatography under the same conditions as the first chromatography. If necessary, the preparations were further purified by gel filtration in the same way as described in the next paragraph. When desired, exce~ cholesterol in the preparations was removed by gel filtration in the presence of cholate.

SDS-polyacrylamide gel electrophoresis was carried out with the use of ready-made 10% gel plates (0.1 x 8 × 8 cm) supplied by TEFCO Company (Tokyo). The resulting protein bands were fixed and stained with 12.5% (w/v) trichloroacetic acid containing Coomassie blue (3250. Isoelectric focusing was carried out at 4 4- 0.5 ° C on home-made plates (0.1 × 11 × 11 cm) of 5% polyacrylamide gel containing 2% (w/v) Ampholine (pH 3.5-10), 20% (v/v) glycerol and 0.1% (w/v) Emulgen 913. The apparatus used was a Bio-Rad Horizontal Isoelectric Focusing System equipped with a temperature control unit. Differently from the usuel method of application, samples dissolved in buffer A were applied on the acidic side of the gel, 1.5 to 2 cm apart from the anode strip. The electric power supply was initially set to a constant power of 11 W and the upper limit of voltage was set to 1800 V. After the focusing was run for 3 h, protein bands were fixed by immersing the gel in 12.5% trichloroacetic acid solution containing 2.5% (w/v) sulfosalicylic acid for an hour with gentle shaking. The gel was washed by immersing in 12.5% trichlo-

237 roacetic acid overnight with gentle shaking to remove Ampholine and then subjected to protein staining with Coomassie blue G250. In some experiments, protein was detected by the silver staining [14]. For the staining of heine, the protein bands were fixed as above and the gel was washed with 1~o (v/v) acetic acid for an hour with gentle shaking. The heine staining was then carried out by immersing the gel in the o-dianisidine-H202 reagent described by Owen et al. [15] at 37°C for 30

plied by Kao-Atlas Chemicals Company (Tokyo). CHAPS (an amphlytic detergent, 3-[(3-cholamidoprop y l ) d i m e t h y l a m m o n i o ] - l - p r o p a n e s u l f o n a t e ) was purchased from Wako Pure Chemical Industries (Tokyo). Other chemicals used were of the available highest grade of purity.

min.

Purification of corpus iuteum P-450sc c The results of the purification of corpus luteum P-450sc c are summarized in Table I. The P-450 extracted with cholate was exclusively P-450scc; no P450n/~was found upon analysis by SDS-polyacrylamide gel electrophoresis followed by immunoblot analysis [17] with anti-bovine P-450sc c and anti-bovine P-450n~ monoclonai antibodies. This finding agreed with the results of activity measurements described by previous authors [4,5]. The yield at the end of ammonium sul[ate fractionation was less than half, but the purification was more than 4-times, those reported by Kashiwagi et al. [3]. This was probably because higher purity was pursued in the present study at the expense of the yield. As illustrated in Fig. 1A, the ammonium sulfate fraction was resolved by chromatography on AH-Sepharose into three P-450sc c fractions, AH-I, AH-II and AH-III (solid line). Although the linear gradient elution technique was employed in this chromatography, Emulgen 913 did not appear in the effluent until the column became equilibrated with this non-ionic deter-

Immunoblot analysis (Western blotting) After the isoelectric focusing, the gel was carefully immersed in 0.7~$ (v/v) acetic acid [16] at 0 °C for I to 2 h without shaking and then the protein bands were transferred to nitrocellulose membrane [16]. The membrane was incubated, as described [17], with mouse anti-bovine adrenocortical P-450scc monoclonal antibody and then with peroxidase-bound anti-mouse lg (F(ab')2 fragment) supplied by Amersham Japan Company (Tokyo). The bound peroxidase was treated with the 4-chloro-l-naphthol-H202 reagent [17] three to five times each for 60-90 s at room temperatures. The membrane was immediately washed with water between these treatments. The developed color was stable for at least a week when stored in cold water in the dark. Other analytical methods Absorption spectra of P-450 were measured at 22 _+ 1°C with a Hitachi 557 spectrophotometer connected to an NEC PC-9801 computer and the concentration of P-450 was calculated from the CO-difference spectrum using the value of 91 r a M - i , cm-1 for the differences in millimolar absorption coefficient increment between 447 and 490 nm [18]. The concentration of protein was measured by the biuret method as described previously [6]. Spectrodensitometry of stained gel plates was carded out with a computerized Shimadzu CS-910 DualWavelength Chromatoscanner. The enzymatic activities of the P-450 preparations were measured according to the method of Kashiwagi et al. [3] with some modifications. N-terminal amino acid sequence was determined with the use of The Applied Biosystems Model 470A Gas-Phase Protein/Peptide Sequencer. Chemicals Sephadex G-25, AH-Sepharose 4B, standard proteins for the estimation of molecular weight and those for the estimation of isoelectric point (pl) were purchased from Pharmacia Japan Company (Tokyo). Toyopearl HW55 superfine and DEAE-Toyopearl were products of TOSO Company (Tokyo). Ampholine (pH 3.5-10) was obtained from LKB Japan Company (Tokyo). Nitrocellulose membrane (Membrane Filter BA85) was the product of Schleicher and Schull (F.R.G.). Emulgen 913 (polyoxyethylene nonylphenol ether) was kindly sup-

Results and Discussion

TABLE I

Purification of P.450scc from corpus iuteum mitochondria Mitochondria were obtained from 2160 g of corpus luteum tissue. The purification procedure is described in Materials and Methods. The purity (~) was calculated either on the basis of the analytical data (concentrations of total protein and P-450) and molecular weight of P-450sc c (55000) or from the results of SDS.polyacrylamide gel electrophoresis. Prepara'ion

Protein (rag)

P-450 (nmol)

Yield (%)

Purity (~)

Mitochondria (sonicated, washed)

12100

2180

(100)

0.99

Cholate extract

4980

1740

80

1.9

Ammonium sulfate 37~ S Fraction 33-58% S Fraction

2540 350

1220 680

56 31

2.6 11

AH-Sepharose (1st) AH-I Fraction AH-II Fraction AH-III Fraction

28 338 214

580

27

17 95 70

AH-Sepharose (2nd) AH-IIR Fraction AH-IIIR Fraction

93 118

211

10

98 96

238 .....

gent. Thus Emulgen was not necessary for the elution of AH-I and AH-II Fractions. The amount of AH-I Fraction was small and this fraction was contaminated with P-t20 and many extraneous proteins. AH-II Fraction was mostly in the high-spin state and already had a relatively high purity of about 9570 at this stage, as estimated by SDS-gel electrophoresis followed by densitometry. On re-chromatography, AH-II Fraction appeared at about the same elution position (Fig. 1A, dashed line) and the purity of the resulting fraction (AH-IIR Fraction) was as high as 9870. AH-III Fraction, on the other hand, was eluted in the presence of Emulgen 913 and the elution coincided with the incipient rise of the Emulgen concentration in the effluent, indicating the greater hydrophobicity of this fraction than AH-II Fraction. AH-III Fraction was also mostly in the high-spin state. On re.chromatography, AH-III Fraction was eluted much earlier, even slightly earlier than AH-II Fraction (Fig. 1B, solid line), and the resulting fraction (AH-IIIR-I) showed a purity of 9670. On further re.chromatography, this fraction appeared at about the same elufion position as that of AH-IIR (Fig. 1B, AH-IIlR-2, dashed lined). It seems possible, therefore, that AH-III Fraction was originally complexed with lipidic material and that this material was either removed or replaced by Emulgen on the first chromatography. Presumably, the Emulgen complex was eluted earlier on the second chromatography and the bound Emulgen was removed by the third chromatography. However, the exact reason for the changes in the elufion position was not clarified in the present study. After completion of the elution of P.450scc in the first chromatography, some colored material was still bound at the top of the column. Further treatment of the column with buffer A containing 0.5 M KCI and 0.5~ (w/v) Tween 20 yielded a fraction containing a small amount of P420 (less than one-tenth of P.450) and many extraneous proteins. When the 33-58~ S ammonium sulfate Fraction was treated in a similar manner by hydrophobic chromatography on octyl-Sepharose instead of AH-Sepharose chromatography, at least half of the P-450 was converted to P-420, and the remaining P450scc was in the low.spin (cholesterolfree) form. The purification .procedure reported by Kashiwagi et al. [3] included successive chromatography on columns of DEAE-cellulose, hydroxyapatite, cellulose-phosphate and heptyI-Sepharose, and the final purity achieved was 9970. Thus the present fractionation with the repeated use of AH-Sepharose chromatography could provide a much simpler method of purification. The AH-Sepharose chromatography could also resolve adrenal cortex P450sc c (33-58~ S Fraction) into four to five fractions, and the second fraction further purified by gel fdtration (AH-IIG) was used as a reference sample in the present study. AH-IIR, AH-IIIR and AH-IIG Fractions dissolved in buffer A could be stored

0.25

i

A

1'"

i

i

i

gr

0.20 0.15 0.10 0.05 0 0.15

-/

I

[3

4/

I

I....

"'~

.....

l..

~..

-. - -

lllR-I

0.10~

0'05t O!

'"' i '"'"'" ~---,-,-t

__L . . . . . . . .

0 100 200 300 400 500 600 700 Effluent, Votume(rnt)

Fig, 1. Chromatographyof corpus luteum P.4$Oscc on an AH-Sepharose column (l.6x 33 cm). The gradient elution with Emulgen was started at the position indicated by the arrow (gr &). The horizontal bars represent the ranges of the fractions which were concentrated and appfied to re-chromatography.The details of the procedure are givenin the text. (A) The result of chromatographyof 33-58~ S AmmoniumSulfateFractionis shownby , and that of re-chromatographyof AH-il Fractionis shownby . . . . . . . (B) The result of re.chromatographyof AH-Ill Fractionis shownby , and that of furtherre-chromatographyof AH-IiiR Fractionis shown by----..-, at - 2 0 ° C for at least several months without any change in their spectra and catalytic activity. The absorption spectra of AH-IIR Fraction are illustrated in Figs. 2 and 3. The P-450sc c as prepared was in the ferric high-spin (cbolesterol-bound) state as evidenced by the Sorer maximum at 392 nm. The cholesterol-bound form was stable, and allowing the preparation to stand at room temperature for 1-2 days did not cause any appreciable spectral change. AH-IIIR Fraction showed spectra indistinguishable from those of AH-IIR Fraction. Neither fraction showed spectral evidence for the presence of /'-420 (Fig. 3). Under the conditions employed, AH-IIR and AH-IIIR Fractions showed cholesterol side-chain cleavage activity of 8 to 14 nmol of pregnenolone formed per win per nmol of P-450. The results of SDS-polyacrylamide gel electrophoresis of corpus luteum P-450scc (AH-IIR and AH-IIIR Fractions) and adrenocortical P-450scc (AH-IIG Fraction) are shown in Fig. 4. It is apparent that these preparations were practically free of contamination with other proteins. The results of computer-assisted spectrodensitometry indicated that the purities of AH-IIR, AH-IIIR and AH-IIG Fractions were about 98, 96 and 9970, respectively. A standard P-450scc preparation prepared from adrenocortical mitochondria by the

239 i

.u 150 279' I

.

|

392

/4/4'7

!

.

!

410

.

.

.

549 I

..,.

512 '

!i

•

'..

/

/

20

.-,

~.~

-......':

/\',,-

/

o

"6

30

637

,

I

,

......'...,., ....~ .

E

.= ' 300

' 500 600 Wavelength (nm)

'

700

4/~nw

Fig. 2. Absorption spectra of corpus luteum P-450scc (AH-IIR Fraction). ., oxidized form. - . . . . . , reduced form obtained by t h e addition of dithionite. • . . . . . , carbon monoxide-bound form. C

100

|

,

|

10

!

.u_ 50 .=-8 578

0 -50 0

E

-loo

~E

l

I

.

,

500 600 W(zveleng th (nm)

'

1 -10

700

Fig. 3. Carbon monoxide-differencespectra of corpus luteum P-450scc (AH-IIR Fraction). method of Takikawa et al. [13] showed a purity of 99.9% under the same conditions. In order to exclude the possibility of contamination

Fig. 4. SDS-polyacrylamide gel electrophoresis of corpus luteum P-450scc and adrenal cortex P-450scc preparations. The positive electrode was to the bottom as the photograph is shown. Other conditions are described in the text. Lanes A and E, the molecular weight standards (phosphorylase b (Mr 94000), bovine serum albumin (Mr 68000), ovalbumin (Mr 43000), carbonic anhydrase (M, 30000), soybean trypsin inhibitor (Mr 20100) and a-lactaibumin (Mr 14400), from the top to the bottom). Lane B, adrenal cortex AH-IIG Fraction. Lane C, corpus luteum AH-IIR Fraction. Lane D, corpus luteum AH-IIIR Fraction.

with microsomal P-450, adrenocortical microsomes were treated in the same way as mitochondria, and the resulting 33-58% S Fraction was subjected to the AH-Sepharose chromatography. Although a trace amount of P-450 was found around the elution position corresponding to AH-III Fraction, this P-450 had the same molecular weight as P-450sc c and reacted with anti-P450sc c monoclonal antibody. Therefore, the trace amount of P-450 was probably originated from mitochondrial fragments contaminated in the microsomal fraction. Thus, significant contamination of AH-II and AH-III Fractions with microsomal P-450 was unlikely to occur.

Isoelectric focusing of P-450scc on polyacrylamide gel The purified P-450sc c pleparations, each showing a single band on SDS-polyacrylamide gel electrophoresis, could be resolved into multiple bands by isoelectric focusing on polyacrylamide gel in the presence of 0.1% Emulgen 913. Addition of Emulgen to the gel and application of the samples on the acidic side of the gel plate were essential for the clear separation of the bands. The results of protein staining with Coomassie blue, shown in Fig. 5, indicated that corpus luteum A H - I I R and IIIR Fractions were composed of two major and several minor subfractions and that the adrenal cortex A I I - I I G Fraction was composed of four major and more than 10 minor subfractions. The major subfractions showed p l values between about 7.9 and 8.3, and the minor bands showed values between about 6 and 7.5 (Fig. 6). It must be noted, however, that these values were calculated on the assumption that the presence of 0.1% Emulgen 913 did not affect the p l values of the proteins. On closer observation, it was found that the major bands were further composed of two closely associated bands. There was a slight difference in the relative intensity of the two major bands between corpus luteum A H - I I R and A H - I I I R Fractions. Although not shown in the figure, the highly purified standard P450scc preparation from adrenal cortex was also resolved into four major and several minor subfractions in the same fashion as adrenal cortex AH-IIG Fraction, except that a smaller number of minor bands was observed. As shown in Fig. 7, most of the protein bands could be stained by heme staining. Since the P-450s¢c samples were spectrally pure, all of the multiple heme bands were probably those of P-450scc. Formation of P-420 during the isoeleetric focusing was not likely to occur in view of the stability of the cholesterol complex of P-450sc c and also in view of the mild conditions of the isoelectric focusing. Among the p l standards used, two of the myoglobins were hemoproteins, and therefore, only two bands are seen in the lanes of the p l standards in this figure. '

240

Fill. $. lsoelectric focusing of corpus luteum P450scc and adrenal cortex P-450scc preparations on polyacrylamide gel (stained with Coonumie blue). The positive electrode was to the bottom as the photowaph is shown. Other conditions are described in the text. Lanes A and E, the pl standards (lentil lectins (pl: 8.65, 8.45 and 8.15). myoglobim (7.35 and 6.85). human carbonic anhydrasc B (6.55). bovine carbonic anhydrase B (5.85), Jg-lactoglobulin A (5.20), soybean trypsin inhibitor (4.55). from the top to the bottom). Lane B, adrenal cortex AH-IIG Fraction. Lane C, corpus luteum AH-IIR Fraction. Lane D, corpus luteum AH-lllR Fraction.

A

6.55 7.35

4.sss.z0

8A5

ls..s 8. s le.ss

C - -

s,,

|

,

0

E

6.55 7.35

0.45

4.s5 uo s.es lus: o.ls l u s

Fig. 6. DensitoMam of the isoelectric focusing of corpus luteum P-450scc and adrenal cortex P.450scc preparations on polyacryl8mide gel (stained with Coomassie blue and measured at 610 nm). The positive electrode was to the left as the tracing is shown. The arrows (,I,) indic=re the positions where samples were applied. Other conditions are described in the texL (A and E) The p l standards. (B) Adrenal cortex AH-llG Fraction. (C) Corpus luteum AH-IIR Fraction. (D) Corpus luteum AH-IlIR Fraction.

Fig. 7. Heme staining with o-dianisidine-H202 reagent after the isoclectric focusing of corpus lutenm P-450scc and adrenal cortex P4$0scc preparations on polyacrylamide gel The conditions are described in the text. Lanes A and E, the p l standards (only two kinds of myoglobin with p l values of 7.35 and 6.85 are stained). Lane B, adrenal cortex AH-IlG Fraction. Lane C, corpus luteum AH-IIR Fraction. Lane D, corpus iuteum AH-IIIR Fraction.

In order to confirm that the observed multiple bands were those of P-450sco immunoblot analysis was carried out with the use of anti-bovine adrenocortical P-450scc monoclonai antibody. There was some difficulty in performing the analysis, in that Ampholine caused non-specific development of colored bands. This difficulty could be considerably reduced, however, by immersing the gel carefully in cold 0.7~ acetic acid for 1 to 2 h and by staining the membrane repeatedly with the peroxidase reagent for 60 to 90 s. The results shown in Fig. 8 indicate that most of the protein bands were those of P-450scc. The exceptions are one or two faint bands which were observed, especially when larger amounts of the samples were applied, in the alkaline side of the major bands after both protein staining and immunoblot analysis but not after heme staining. These faint bands were, therefore, probably those of ape-protein of P-450scc. Essentially the same isoelectric focusing pattern as those of the purified preparations was observed in the chelate extract from corpus luteum mitochondria when it was examined by immunoblot analysis~ A low-spin type (cholesterol-free) P-450scc preparation prepared from the chelate extract of corpus luteum mitochondria by hydrophobic chromatography on octyl-Sepharose also showed the same isoelectfic f~u~ing pattern as the high-spin type (cholesterol-bound) preparations. Isoelectric focusing followed by immunoblot analysis of the chelate extract from bovine adrenal cortex

241

A

B

C

D

E

Fig. 8. Immunoblotanalysis with anti-bovine P-450scc monoclonal antibody after isoelectric focusing of corpus luteum P-450scc and adrenal cortex P-4$0scc preparations on polyacrylamidegel. The details are described in the text. Lanes A and E, the pl standards. Lane B, adrenal cortex AH-IIG Fraction. Lane C, corpus luteum AH-IIR Fraction. LaneD, corpus luteumAH-IIIR Fraction.

mitochondria obtained from a single animal gave essentially the. same pattern as the purified preparations. These results of isoelectric focusing followed by specific staining indicate that multiple molecular forms of P-450scc exist in mitochondria of the corpus luteum as well as those of the adrenal cortex. However, the problem of whether the observed multiplicity indicates multiple isozymes or post-translational modifications remains to be solved. Staining of the isoelectric focusing gel with periodic acid-Schiff's (PAS) reagent gave only faint ambiguous spots under the conditions where glycoprotein standards gave intense purple spots. Isoelectric focusing in the presence of 1 mM EDTA or 1 mM CaCl 2 did not affect the pattern. Extensive treatment with sialidases also did not affect the pattern. Vilgrain et al. [19] have suggested that P-450scc can be phosphorylated in vitro up to 4 mol of phosphate per mol. We failed to detect any essential change in the focusing pattern of main bands after treating our preparations with phosphatases from various sources or with an ATP-protein kinase system. However, the intensity of some minor bands at pH 6.7 to 7.2 decreased slightly after treating adrenocortical AH-IIG Fraction with the catalytic subunit of protein phosphokinase from bovine heart (Sigma) ir~ the presence of ATP. Concomitantly, three to four broad bands at around pH 6 appeared after the treatment. Results of gel filtration experiments indicated that P-450scc was in dimeric (to tetrameric) state m the presence of 0.1~ Emulgen 913 and in monomeric state in the presence of either 0.3~ sodium cholate or 0.3~ CHAPS. But the isoelectric focusing pattern did not

show any essential change when Emulgen concentration was increased or Emulgen was replaced by CHAPS. In order to analyze N-terminal amino acid sequence, AH-IIR and AH-IIIR Fractions were combined and further highly purified by gel filtration on Toyopearl HW55S column followed by passing through a DEAEToyopearl column. The resulting preparation, giving essentially the same isoelectric focusing pattern as AHfIR and AH-IIIR Fractions, had a single N-terminal amino acid sequence of Ile-Ser-Thr-Lys-Thr-Pro-ArgPro-Tyr-Ser-Glu-Ile-Pro-Ser-Pro-Gly. . . . . This sequence was identical with the known N-terminal sequence of adrenocortical P-450sc c [11]. The analysis of a purified preparation of AH-IIIR also gave the same result. Tsubald et al. [11] already resolved their purified adrenocortical P-450scc preparation into three fractions (a, b and c) by chromatofocusing and precisely studied their properties. They analyzed N-terminal and C-terminal amino acid sequences of these fractions and found that N-terminal isoleucine of the fraction c was deleted from the above sequence, although all these fractions had identical C-terminal sequence. We could, however, find only one kind of N-terminal sequence with our preparations of corpus luteum P-450scc. As shown in Fig. 9, AH-IIR and AH-IIIR Fractions could be further resolved by HPLC on SP-TSK gel into two fractions, SP-I and SP=II. These SP Fractions were subjected to isoelectric focusing and the results are shown in Fig. 10. Although a few minor bands were still detected by sensitive silver staining, SP-I Fraction was composed mostly of the main molecular form with a less alkaline isoelectric point, and SP-II Fraction was

!

!

,

i

i

,

|

l

SP-2

0.2{;

E 0.IE

C O ¢D O

®

0.t2

).5

u C

o

,',

).4

SP-

'- 0.0~

"-

0

,n

).3~ ..

&3

)2 z

0

,

'

0

20

40

i ~

0

60

Minutes

Fig. 9. HPLC of corpus luteum P-450scc (AH-IIIR Fraction) on Sulfopropyl-TSK gel (SP-5PW). The conditions are described in Materials and Methods.

242 TABLE II

Catalytic activity of corpus luteum P-450scc ($P-I and SP-H Fractions) Conditions for the measurement are described in Materials and Methods. nmoles of preanenolone formed per rain per nmol P-4$0scc

Substrate

cholesterol Corpus luteum SP-i Fraction SP-II Fraction Adrenal cortex AH-IlG Fraction

Fi& 10. lseelectric focusing of SP-! and SP-I! Fractions of corpus luteum P-450sc c on polyacrylamide 8el (stained with silver reagent). The conditions are described in the text. Lane A, adrenocortical P4S0scc AH-ilG Fraction. Lane B, SP-! Fraction. Lane C, SP-ll Fraction.

composed mostly of the other main molecular form with a more alkaline isoelectric point. The analysis of N-terminal amino acid sequence revealed that both SP-I and SP-II Fractions had the identical sequence. It seems possible that the peptide structure other than N-terminal region of the two main molecular forms at~ different, because Tsubaki et al. [11] have reported some difference in the peptide maps of the molecular forms of adrencgortical P-450scc. They have found a small but distinct difference in the HPLC profiles between the tryptic digests of their fractions a and b, and suggested that the difference might be ascribed to the post-translational modification(s) in the middle portion of the polypeptide. But we could not

22( R )-hydroxycholesterol

7.8 8.3

19.6 20.3

11.6

47.2

carry out peptide mapping because supply of corpus luteum was limited, the concentration of P-450sc c in corpus luteum mitochondria was low (about one-tenth [1]) compared with adrenal cortex mitochondria and the yield of SP-I and SP-II Fractions was small. Thus so far examined we could not find any clear evidence for the structural basis of the observed multiplicity of corpus luteum P-450scc. The enzymatic activities of SP-I and SP-II Fractions were then measured with cholesterol and 22(R)-hydroxycholesterol as the substrates. As shown in Table I!, both fractions showed essentially the same activity. In view of the complex isoelectric focusing pattern observed in the present study, two or more factors are probably responsible for the formation of multiple molecular forms of P-450scc. Results of our computersimulation studies indicated that, besides a difference in the middle portion of the polypeptide chain like that suggested by Tsubaki et al. [11], a pair of sites (each site with or without a charged group) on the molecular surface had to be postulated to explain the multiple isoelectric focusing bands of corpus luteum P-450sc c, and at least one more additional site to explain the multiple bands of adrenocortical P-450sc c. The simulation of the densitogram of corpus luteum A~I-IIR Fraction was carried out as the summation of the Gaussian distributions, and a simple simulation was

TABLE !il

Resldts of free simulation for tile densitogram of corpus iuteum AH.llR Fraction Peak No,

Peak position (Pl)

Relative peak height

Peak width (pH span)

Relative peak strength

Corresponding postulated isozymes, with or without charge

i' 2 3 4 5 6 7 8

8.28 8.19 8.08 7.98 7.90 7.82 7.71 7.60

85.0 60.0 107.0 76.0 35.0 38.0 26.0 22.0

0.09 0.09 0.09 0.07 0.07 0.06 0.11 0.09

776 547 1025 607 258 249 299 211

aa bo ap bp aT b~r a6 ba

243 s

20

5.85

6.55 6.85

8.15

8.

Fig. 11. "Fitting densitogram' of corpus luteum AH-IIR Fraction.

carried out first without any assumption. The results of the simple simulation are shown in Table III and Fig. 11, which is comparable with Fig. 6C. It is apparent that the simulated densitogram fitted very well with the observed densitogram. Therefore, instead of the observed densitogram, we used this 'fitting densitogram' as the reference of out further theoretical analysis. In analyzing the origin of the multiple bands, we considered the iollowing assumptions: (A) Two 'isozymes', a and b, exist and they are different in the structure (either replacement or post-translational modification of an amino acid residue) of the middle portion of the polypeptide. They have thus slightly different conformation and slightly different pl. (B) There is a pair of other sites (01, 02) and these sites can be charged positively or negatively. Therefore, we can express eight combinations of the charge states of the sites. For example, two combinations among the eight are as follows: [a(0,0), fi(0,- ), ~,(-,0) and 8 ( - , - )] or [a(+,0), fi(0,0), V ( + , - ) and ~(0,-)]. (C) The differences in peak positions in p l (in pH span) of a and tho~e of b are equal, ,~,, - aj --/',,~ -/,j

(i

or

j = ,~,/3,v,,~)

(1)

(D) The assumption (B) leads the following relations: ao-

aij = av - a a

a n d b,. - bl~ = bv - b a

(2)

where either a~ or b~ is the peak position in p l of the same 'isozyme' with different charge. And also, eo/ep

=

e,/e6

(3)

where P~ is the probability (the relative concentration) of charged form i. The observed differences in p l of the postulated 'isozymes' a and b in the 'fitting densitogram' were relatively small and did not vary appreciably (0.09, 0.10, 0.08 and 0.11). Therefore, the assumption (A) is consistent with the observed values. The observed differences in p l of the same postulated 'isozymes' with different charge were as follows: a ~ - a2 = 0.20, a 2 a 3 = 0.18, a 3 - a 4 = 0.19, b I - b 2 = 0.21, b2 - b 3 = 0.16

and b a - b4 = 0.22. These values approximately satisfy Eqn. 1 in the assumption (C) and Eqn. 2 in (D). These values are also compatible with the fact that ~he addition or removal of a single charge to or from a usual protein molecule ~sults in a change in p l by about 0.15 (to 0.5) in pH span. The ratios P,~/PI~ and Pv/Ps were 0.79 and 1.35 for 'isozyme' a, and 0.79 and 1.73 for 'isozyme', b, respectively. These values seemed not to satisfy Eqn. 3. But we think that these values are acceptable in this kind of calculation. Thus, the assumptions are compatible with the "fitting densitogram' and, conseqt~ently, with the observed densitogram. The theoretically simulated densitogram obtained on the basis of the assumptions naturally agreed well with the observed densitogram, although a good fitting does not necessarily mean for the postulated mechanism actually to occur. The ratio of the amount of postulated 'isozymes' a:b calculated from the peak heights and widths was around 1.5 : 1. Although there was a possibility of presence of another isozyme, 'isozyme' c, its amount was likely to be small, if any, and therefore it was not taken into consideration in the present study. Since the above assumptions seemed to be appropriate, a similar but more complex simulation study was also made for the densitogram of adrenal cortex AH-IIG Fraction by extending the assumptions. Initially, we further postulated four states (named r, ~, # and v) of an additional pair of sites (each site with or without a charged group). This time, the simulation was carded out from the first on the basis of the assumptions and the theoretically calculated densitogram was directly compared with the observed densitogram of adrenal cortex AH-IIG Fraction. In carrying out the simulation, however, we found that Eqn. 3 had to be abandoned because the peak heights of the calculated densitogram were very different from those of the observed densitogram. By disregarding Eqn. 3 we could obtain a moderately acceptable simulation (data not shown). This fact indicated that postulation of another kind of additional charging site was required for a better simulation of the densitogram of adrenal cortex AH-IIG Fraction. However, the fact that we could obtain an acceptable simulation indicated at least that multiple minor bands could appear in the acidic side of the main bands if P-450sc c molecules had some kind of additional charging site(s). From the results of further theoretical examinations, we feel, at present that either one to three negatively charged small molecules can be bound cumulatively or one of the oligomers of low molecular weight compound with one to three negative charges can be bound to this additionally postulated site (assumption (E)). As shown in Table IV and Fig. 12, the theoretical densitogram obtained on the basis of the last assumption (E) agreed well with the observed densitogram in both peak positions and peak heights (the dots in Fig. 12 represent the peak positions and

244 TABLE IV

Res~ls o~ theoretical simulation for the denatosram of adrenal cortex AH-II6 Fraction A peak width of 0.09 in pH span was uniformly adopted,

No.

position

heisht

(concentration of is¢~ymes)

(in p l )

1

s.o

2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19 20 21 22 23 24 2S 26 27 28 29 30 31 32

8.19 8.08 7.99 7.91 7.82 7.70 7.61 7,83 7.74 7,63 7.54 7.46 7.37 7.2S 7.16 7.18 7.09 6.98 6.89 6.81 6.72 6.60 6.51 6.35 6.26 6.15 6.06 S.98 5.89 5.77 S.68

814 611 1025 769 527 395 383 287 244 183 308 231 158 119 115 86

63.8 107.0 80.3 55.0 41.3 40.0 30.0 28.$ 19.1 32.1 24.1 16.5 12.4 12,0 9.0 25,5 19.1 32.1 24.1 16.5 12.4 12.0 9.0 11.1 8.3 13.9 10.4 7.2 5.4 5.2 3.9

5.85 ,

j

6.55 6.85 I

j

aa-0

bo-0 a#-O

b~-0 ay-O

bT-O aa - 0 ba-O aa-1

b.-I aa - 1

bp-1 a T- 1 bv-t ae -1 ba-1

aa~2 183 308 231 158 119 I15 86 106 79 133 100 68 51 SO 37

heights of the obs~,ed densito~ram). Presumably, the probability of the formation in vivo (in adrenal cortex) of the f~rst charge is lower (about 30~$) than those of the second and the third charge (about $0~). Thus even the complex electrofocusing pattern of adrenal cortex AH-IIG Fraction may be simulated by

0

Posturated isozymes (with or without charge)

8.15 8.4 I

,,,

@Q F'q~.12.11mmvticallycalculated densitolgmn of adrenal cortex AH-IIG Fraction. Clhe dots represa~t the peak positions and h e i s t s of the obsaved densitognun.)

/,-2 aB - 2 bAs-2 ay-2

b7-2 ae-2

ha-2 ao-3

/.-3 -3 bp-3 av -3

aB

by-3

ae-3 b,I-3

postulating two 'isozymes' each having a pair of sites (each site with or without a charged group) and also another kind of site with or without charge(s). Ogishima et al. [20] reported that P-4$0scc contained about 42 (Ash + Asp) residues and 51 (Gin + Glu) residues. Therefore, the extent of deamidation (or amidation) is one of the possible structural basis of the formation of multiple bands observed on isoelectric focusing. The phosphorylation [19] or dephosphorylation may also be one of the possible structural basis of the formation of multiple minor bands with acidic pl. Most of the minor fractions with relatively acidic pl seem to have been removed during the course of usual methods of purification, especially by ion-exchange chromatography. Since different patterns were obtained in the present study with P-450scc fractions from the corpus luteum and the adrenal cortex, the multiplicity of the molecular form seems to be tissue specific and is likely to be controlled somehow by gonadotropic hormones and adrenocorticotropic hormone, respectively.

245 Conclud|ng remarks A simple and efficient procedure with the use of AH-Sepharose chromatography was developed for the purification of P-450sc ¢ from corpus luteum mitochondria. Evidence for the presence of multiple molecular forms of P-450sc c was obtained by isoelectric focusing on polyacrylamide gel. Acknowledgement We express our thanks to Dr. H. Suzuki of the Department of Biophysical Chemistry, School of Medicine, Kitasato University for his help in performing N-terminal amino acid sequence analysis. References 1 Yohro, T. and Horie, S. (1967) J. Biochem. 61,515-517. 2 Mclntosh, E.N., Mitani, F., Uzgiris, V.I., Alonso, C. and Salhanick, H.A. (1973) Ann. NY Acad. Sci. 212, 392-403. 3 Kashiwagi, K., Dafeldecker, W.P. and Salhanick, H.A. (1980) J. Biol. Chem. 255, 2606-2611. 4 Ichii, S., Forchielli, E. and Dorfman, R.I. (1963) Steroids 2, 631-656.

5 Mclntosh, E.N., Uzgiris, V.I., Alonso, C. and Salhanick, H.A. (1971) Biochemistry 10, 2909-2916. 6 Horie, S. and Watanabe, T. (1975) J. Steroid Biochem. 6, 401-409. 7 Jefcoate, C.R. (1977) J. Biol. Chem. 252, 8788-8796. 8 Wolfson, A.J. and Lieberman, S. (1979) J. Biol. Chem. 254, 4096-4100. 9 Greenfield, N.J., Gerolimatos, B., Szwergold, B.S., Wolfson, A.J., Prasad, V.K. and Lieberman, S. (1981) J. Biol. Chem. 256, 4407-4417. 10 Greenfield, N.J. and Gerolimatos, B. (1985) J. Steroid Biochem. 22, 809-816. 11 Tsubaki, M., Ohkubo, H., Tsuneoka, Y., Tomita, S., Hiwatashi, A. and Ichikawa, Y. (1987) Biochim. Biophys. Acta 914, 246-258. 12 Mitani, F. and Horie, S. (1969) J. Biochem. 65, 269-280. 13 Takikawa, O., Gomi, T., Suhara, K., Itagaki, E., Takemori, S. and Katagiri, M. (1978) Arch. Biochem. Biophys. 190, 300-306. 14 Toyoshima, I. (1982) Brain Nerve 34, 973-979 (in Japane~se). 15 Owen, J.A., Silberman, C. and Got, C. (1958) Nature 182, 1373. 16 Towbin, H., Staehelin, T. and Gordon, J. (1979) Proc. t,~atl. Acad. Sci. USA 76, 4350-4354. 17 Sugano, S., Ohnishi, T., Hatae, N., Ishimura, K., Fujita, H., Yamano, T. and Okamoto, M. (1985) J. Steroid Biochem. 23, 1013-1021. 18 Omura, T. and Sato, R. (1964) J. Biol. Chem. 239, 2370-2378. 19 Vilgrain, I., Defaye, G. and Chambaz, E.M. (1984) Biochem. Biophys. Research Commun. 125, 554-561. 20 Ogishima, T., Okada, Y., Kominami, S., Takemori, S. and Omura, T. (1983) J. Biochem. 94, 1711-1714.