3β-Hydroxysteroid dehydrogenase in human placental microsomes and mitochondria: Co-solubilization of androstene and pregnene activities

3~~DRO~STEROID DEHYDROGENASE IN HUMAN PLACENTAL M~CROSOMES AND MITOCHONDR~ CU-SOL~HILIZATION OF ANDROSTENE AND PREGNENE ACTIVITIES Ebeneeer Asibey-Becko, James L. Thomas, and Ronald C. Strickler* Department of Obstetrics and Gynecology, Washington University School of Medicine at Jewish Hospital 216 S. Kingshighway, St. Louis1 ~~souri 63110 Received

May 27, 1986

Revised February

6, 1987

ABSTRACT a&Hydroxysteroid dehydrogenase (a@-HSD) was solubilised from human termplacental mierosomes and mitochondria using the non-ionic detergent, polyoxyethylene 20 cetyl ether (BrijB-58). Electron photomic~graphs showed microsomes and mitochondria well disrupted by the detergent. The pregnene AC-21) and androstene (C-19) activities co-solubilired over a range (0.04 - 0.44) of Brij -58/prote~n (B/P) concentration ratios (w/w). Optimal solubi~sati~n of the C-19 and C-21 activities were 63.3 rt 2.6% (mean f SEM) from mitoe~o~dr~a (B/P ratio 0.37) and 71.8 A 2.1% from ~~rosomes (B/P ratio 0.34). Detergent treatment of microsomes and mitochondria - varying time (5-90 min, pH 7.4) or varying pH (6.6-7.8, 90 min) - yielded C-19 activities identical with C-21 BCtivities. The C-21/C-19 specific activity ratios of 3/?-HSD in particulate, solubilised and chromatographed preparationswere2.28 i 0.16 (mean i_ SEM) for mitoehondria and 1.97 f 0.07 for microsomes. 3@-HSD molecular weight estimates were 208,960 (microsomes) and 220,000 (mitochoudr~a). These studies argue that a single protein is responsible for both the C-19 and C-21 activities of 3&HSD and that this protein is the same in microsomes and mitochondria.

INTRODUCTION 3~~yd~oxys~ero~d dehydrogen~~/~is~mer~e ~3~~~D~, 8 ratelimiting enzyme complex (1,Z) found in most steroid-metabo~i~i~g tissues, is present in both mierosomes and mitochondria from term human placenta (3-5). Although 3/S-HSD accepts both pregnene (C-21) and androstene (C-19) substrates, there is confusion M to whether the active site(s) for pregnenolone and dehydroepiandroster~ne exist on one or separate proteins. Reports by Ferre et ~1 on the human placenta (3), Gibb and Hagerman on bovine ovaries (6) and Gallay et al (7) on bovine adrenal cortex STEROIDS 47/6

June

1986 (351-363)

351

352

Asibey-Berko

et al

suggested multiple B/3-HSD enzymes. by Gibb (11) 1ocalized both activities

Later work (8-10) and a repeat on one protein.

study

The localization of C-21 and C-19 oxidoreductaae activity on one or In term pregnant animals which two loci has physiologic significance. “withdraw progesterone” to initiate parturition (12), the fetus may precipitate this hormonal change with a flood of dehydroepiandrosterone sulfate (DHEA-S) which competitively inhibits pregnenolone oxidation at a single placental 3@-HSD active site. We solubilized and partially purified 3P_HSD from mitochondrial and microsomal fractions of human placental villous tissue. The C-21 and C-19 activities co-solubilized and remained together during purification experiments.

MATERIALS Materi&.

AND METHODS

Steroids (chromatographically-pure

pregnenolone

and DHEA), NAB+, EDTA,

bovine serum albumin, 2-mercaptoethanol, DL-dithiothreitol, Brij*-58 (polyoxyethylene 20 cetyl ether) and digitonin were purchased from Sigma Chemical Company. Reagent-

grade salts, analytical-grade solvents, Brilliant Blue G, ScintiVerse liquid scintillation counting fiuid, Eastman thin-layer chromatography sheets and buffer solutions (pH meter calibration) were obtained from Fisher Scientific Company. Hollow-fibre bundle systems and Spectrapor dialysis membranes were ordered from Spectrum Medical Industries. Sephacryl S-300 superfine gels and gel-filtration columns came from Pharmacia Fine Chemicals Company. 3H-Pregnenolone (specific activity 13.0 Ci/mmol) and ‘H-DHEA (specific activity 55.0 Ci/mmol) from New England Nuclear were checked with thin-layer chromatography and found pure. Aqueous solutions were prepared with glass-distilled de-ionired water. Bufletr. (i) Glycerol buffer: 0.01 M potassium phosphate buffer, pH 7.4, containing 20% glycerol (v/v). (ii) Glycerol/EDTA buffer: 2 mM EDTA in glycerol buffer, pH 7.4. (iii) Glysuc buffer: 0.25 M sucrose in glycerol buffer, pH 7.4. (iv) Glysuc/EDTA buffer: 0.25 M sucrose, 2 mM EDTA in glycerol buffer, pH 7.4. (v) Glycerol/Mercaptan buffer: 2 mM mercaptoethanol in glycerol buffer, pH 7.4. (vi) Glycerol/Mercaptan/EDTA buffer: 2 mM mercaptoethanol, 2 mM EDTA in glycerol buffer, pH 7.4. (vii) Glycerol/DTT buffer: 1 mM dithiothieitol in glycerol buffer, pH 7.4. (viii) Glycerol/DTT/EDTA buffer: 1 mM DTT and 2 mM EDTA in glycerol buffer, pH 7.4. (ii) Modified Tyrode’s buffer: 0.60% NaCl, 0.02% KCl, 0.02% CaCl , 0.01% NaH,PO,, 0.01% MgCl,, 0.01% NaHCO,, 0.10% sucrose (all w/v in H,O), pH 7.f. Mieroromol and mitoehondrial preparations. Human term-placentas from normal gestations were chilled on ice immediately after delivery. Placental villi were dissected (approximately 400 g) and homogenired in a blender with 1 mL of glysuc buffer, pH 7.4, for 1 g of tissue. The homogenate was centrifuged for 20 min at 1500 x g in a Beckman preparative ultracentrifuge model LS-65. The residue was diicarded and the eupernatant

HUMAN

3,&HYDROXYSTEROlD

DEHYDROGENASE

353

apun for 20 min at 12,000 x g. The mitochondrial pellet was washed twice by gentle manual homogenization in glysuc buffer, pH 7.4, and centrifugation at 12,000 x g. The final mitochondrial pellet was again suspended in 54 mL glysuc ‘buffer, pH 7.4. The microsomes were pelleted at 105,000 x g for 1 h, washed twice and resuspended in 60 mL glysuc buffer, pH 7.4. Protein concentrations of both preparations were determined with the method of Bradford (13). Detergent treatment od mitoehondrial and mierosomal preparatione for aetivity ratio study. A volume of 1% Brij -58 detergent in glysuc buffer, pH 7.4, was added to the mi.$chondrial (5.0 mg/mL) or microsomal (5.0 mg/mL) preparations to yield 0.1% finaIRBrrJ concentration (14). The mixture was gently stirred at 4 C. The removal of Brlj from the retentate was verified quantitatively by thin-layer chromatography on silica-gel plates developed in CHC$:ethyl acetate (9:l) (BrijR-58 R, = alO, visualized by iodine vapor). The mitochondrial retentate was centrifuged at 110,000 x g for 1 h. The microsomal retentate was centrifuged at 110,000 x g for 1.5 h. The supernatants (50 mL) were carefully pipetted and the pellets reconstituted in 50 mL glycerol/EDTA buffer, pH 7.4. Aliquots (5 mL) of microsomal and mitochondrial preparations were treated exactly like the samples, replacing BrijR detergent with buffer. This yielded “zero-BrijR” supernatants and reconstituted “zero-BrijR” pellet controls. Protein and enzyme assays. Protein concentrations were determined by the method of Bradford (13) using a Varian Cary 219 recording spectrophotometer. 3PHSD activity in each aliquot was measured in duplicate, using the conversion of [7-?-I]pregnenolone or [1,2-3H]DHBA to tritiated progesterone or androstenedione respectively. Reaction vessels contained not more than 0.5 mg protein plus 50 pM 3H-pregnenolone (or 3H-DHBA) and 0.4 pM NAD+ in pH 7.4 buffer (total vol 1.0 mL). Nonspecific conversion of substrates was accounted for by both “zero-time” and “zero-enzyme” blanks. The reaction mixtures were incubated in air in a shaking bath at 37 C for 8 min for pregnenolone assays and 10 min for DHBA assays. Time course and protein concentration studies demonstrated that these conditions yielded rates of pregnenolone and DI-IEA conversions that were linear with respect to time and enzyme protein concentration. Less than 20% conversion of pregnenolone or DIIEA (15% average in our laboratory) occurred under these conditions, limiting product feedback inhibition (3,s). The digitonin precipitation method of Philpott and Peron (.15) was used to separate unreacted pregnenolone (or DHBA) from the reaction products. This yielded a supernatant of reaction products containing 3H-progesterone (or 3H-androstenedione). After liquid scintillation counts, enzyme activities could be calculated. A Beckman model LS7500 liquid scintillation counter was used to measure radioactivity (tritium counting efficiency 48%). Electron

mieroaeopy. Microsomal and mitochondrial pellets were examined with the electron microscope. Controls (non-detergent-treated) were examined along with detergenttreated samples. Detergent/protein ratios of 0.37 and 0.34 (w/w) were used for solubilizing mitochondrial and microsomal preparations respectively. Pellets were fixed in 2% gluteraldehyde in modified Tyrode’s buffer. After washing in buffer and treating with 1% buffered osmium tetroxide, pellets were dehydrated and embedded in Polybed 812 (Polysciences, Inc., Warrington, PA). Sections were examined in a Phillips 201 microscope.

354

Asibey-Berko

et al

Figure 1. Electron microscopy of detergent (BrijR-58~treated microeomes and mitochondria. A) Microeome control. B) Detergent-treated microsomea. C) Mitochondria control. D) Detergent-treated mitochondria. MagniGcation: approximately 32,000.

HUMAN

3IFHYDROXYSTEROID

DEHYDROGENASE

355

TABLE 1 STABILITY OF 3&HSD IN DIFFERENT BUFFER SYSTEMS

Stand at 4 C/Stir 3 h/ Stand at 23 C

Stand at 4c

Specific Activity (nm/nWmg)

Buffer Systems

Controlb Glycerol buffer Glycerol/EDTA G lysuc Glysuc/EDTA Glycerol/Mercaptan Glycerol/Mercaptan/ EDTA Glycerol/DTT Glycerol/DTT/EDTA

% Denatured’

Specific Activity (nm/min/mg)

3.30 3.14 3.26 2.71 2.67 2.86

4.63 1.05 17.89 13.06 19.31

3.30 2.85 2.98 1.67 2.79 2.62

2.81 2.55 2.70

14.90 22.75 18.28

2.36 2.59

0.00

a % Denaturation = (l-Specific Activity/Control b Glycerol buffer; sample frozen at -20 C.

Specific Activity) x 100.

PREGNENE/ANDROSTENE 3PHSD SPECIFIC ACTMTY RATIOS IN PARTICULATE, SOLUBILIZED AND CHROMATOGRAPHED PREPARATIONS Ratios for Each Preparation

Ratio* from Each Source

Microsomes Solubilired Microsomes After Sephacryl S-300

2.03 =t 0.33 1.82 it 0.31 1.97 & 0.36

1.94 f 0.07

Mitochondria Solubilired Mitochondria After Sephacryl S-300

2.34 f 0.29 1.98 f 0.02 2.52 rt 0.18

2.28 f 0.16

a Mean & SEM; N=3.

0.00

13.61 9.58 49.32 15.29 20.71

__

TABLE 2

Enryme Preparation

% Denatured

27.95

21.45

356

et al

Asibey-Berko

A 80

0-0

’

A--A

Preg DHEA

60

20

0’

I

1,

0.10 0.20 mg Brij@-58 /mq

I

0.30 PROTEIN

I

0.40

Figure 2. The effect of ditferent concentrations of detergent (BrijR-58) on solubilization of BP_HSD. A) Microsomal and B) mitochondrial extracts were assayed for both pregnenolone (Preg) and DHEA activities.

HUVAN 3&HMIROXYSTEROJD

DEHYDROGENASE

357

RESULTS Electron

photomicrographs. Controls show very pure microsomal and mitochondrial preparations with intact vesicles (Fig. 1A and 1C) Detergent treatment fractured and disrupted the vesicles (Fig. 1B and 1D). Enzyme

stability in different buffers. Stability of mitochondrial 3@-HSD under two denaturing conditions was tested in the eight buffer systems. Enzyme solutions stood overnight at 4 C, or were further stirred 3 h, followed by standing overnight at room temperature (23 C). Results of enzyme assays (Table 1) indicate that EDTA protected S,&HSD from denaturation in each buffer system tested. The buffer offering most protection was the glycerol/EDTA buffer, pH 7.4 (0.01 M potassium phosphate buffer containing 20% glycerol (v/v) and 2 mM EDTA). These stability studies do not preclude accelerated enzyme denaturation by certain buffer constituents. Co-solubilization

studies.

a) Specific activity ratios. Three microsomal and mitochondrial preparaorganelles, detergent-solubilized enzyme and tions - isolated particulate Sephacryl S-300 gel-chromatographed enzyme - were assayed for 3@-HSD activity. The ratios of 3PHSD activities with pregnenolone and DHEA remained unchanged through each partial purification step (Table 2). b) Effect of BrijR-58/protein ratios (w/w) on solubilization.To 3 mL rnitx chondrial (or microsomal) samples (5.55 mg/mL protein), 0.333 mL Brlj solutions (ranging from 0.50 to 2.20% in glycerol/EDTA buffer, pH 7.4) were added, yielding a range of BrijR/protein ratios (w/w) from 0.04 to 0.44. Control consisted of 0.333 mL buffer added to 3 mb sample. Samples were shaken on ice for 1.5 h, centrifuged at 110,000 x g and both supernatants and reconstituted pellets assayed. Solubilization is reported as percent maximal activity (supernatant activity as percent of maximal observed pellet plus supernatant activity). Both pregnenolone and DHEA activities of 3,8-HSD peak at the same BrijR/protein ratios of 0.37 in the mitochondria and 0.34 in the microsomes (Fig. 2). c) Eflect of pH on solubilization. Placental villi were processed, yielding microsomal and mitochondrial preparations in buffers ranging from pH 6.0 to 7.8. The pre arations were solubilized, using the same pH values and g the optimal Brij /protein ratios found for microsomes and mitochondria. Both pregnenolone and DHEA activities of 3@-HSD in the mitochondria

358

Asibey-Eterko et al

I

I

70

6.5

I

24

I

ZE

PH

Figure 3. Effect of pH on solubilization of B&HSD. A) Microsomal and B) mitochondrial extracts were assayed for pregnenolone (Preg) and DHEA activities.

HUMAN 3/3-HYDROXYSTEROID

DEHYDROGENASE

1

60-A

40;y== + 3 : is 2

ii

20

e3

f

0

6-A

Preg DHEA

I

I

I

I

20

40

60

80

60 - B

x Qr 40 \o 0 20 0

MINUTES

Figure 4. The effect of extraction time on the solubilieation of I@HSD. A) Microsomal and B) mitochondrial extracts were assayed for pregnenolone

(Preg) and DHEA activities.

359

360

Asibey-Berko et al

and microsomes showed identical pH-dependent solu~lization profiles {Fig. 31. d) Effect of extraction time on solubilization. Microsomal and mitochondrial preparations were detergent-extracted with shaking in a Dubnoff metabolic incubator over ice (0 C). Extraction times ranged from 5 to 90 min. In comparison to ‘zero-time and detergent-free controls, maximum 3@-HSD was extracted in 5-10 min. Pregnenolone and DHEA activities showed an identical solubilization profile over the entire time range (Fig. 4). Criteria of solub~lization - Get ~ltration pur~~cation. Solubilized samples were concentrated, using hollow-fibre bundle ultra-filtration systems. Samples (10 mL) were ap lied to Sephacryl S-300 columns (2.6 x 69 cm, B exclusion limit - 1.5 x 10 daltons) and eluted as 3-4 mL fractions with glycerol/EDTA buffer, pH ‘7.4. In both mitochondrial and microsomal preparations, a clear enzyme peak eluted 1.34 void volumes beyond the void volume. Enzyme protein molecular weight - Gel filtration. Detergent extracts of samples were dialyzed against 300-fold their volume of glycerol/EDTA buffer, pH 7.4, to remove the BrijR-58; 2.2-r& sample retentates were ap plied to Sephacryl S-300 columns (1.6 x 69 cm; bed volume 138 I&) previously characterized with blue dextran and calibrated with protein standards of known molecular weights. Molecular weight estimates were 208,000 daltons and 219,585 daltons in the microsomal and mitochondrial preparations respectively (Fig. 5). If the proteins retain a coating of detergent despite dialysis, the moiecular weights may be overestimates.

DISCUSSION Evidence for a single protein with C-19 and C-21 actiuities. Gibb showed that the apparent Km values for DHEA and pregnenolone as substrates for 3/3-HSD of placental microsomes were 15 nM and 40 nM respectively (11). Mixed substrate experiments showed that pregnenolone and DHEA competitively inhibited each other. The Ki of pregnenoione inhibition of DHEA was 38 nM - about the same as the pregnenolone Km of 40 nM. The Ki of DHEA inhibition of pregnenolone was 17 nM - statistically the same as the DHEA Km of 15 nM. The oxidation of both substrates by S/3-HSD also preferred IUD’ as cofactor. Gibb (11) concluded that a single enzyme oxidized both substrates. Byrne et al (lo), working with crude human adrenal microsomes, drew the same conclusions.

0.20

>

270.10

A cl

C 100

200 MOLECULAR

300 WEIGHT

500

700

f x 103)

Figure 5. Est~atio~ of the molecniar weight of microeomal and mitochoadriai 3&HSD by gel filtration chromatography.

362

Asibey-Berko et al

Our data extend these observations. Throughout our purification steps involving crude sample preparation, 90 min detergent extraction and Sephacryl S-300 gel chromatography, the pregnenolone/D~A specific activity ratio remained unchanged. A two-fold higher oxidation of pregnenolone compared with DHEA is reflected in the specific activity ratios. Since the enzyme has a higher affinity (lower Km) for DHEA (ll), this may suggest a higher turnover number for the pregnenolone reaction. Both pregnene and androstene activities showed an optima at the same detergent/protein ratios of 0.37 and 0.34 in the mitochondria and microsomes, respectively. This points to the extraction of a single protein with activity towards pregnenolone and DHEA, Gel filtration data demonstrated that the enzyme activity eluted well after the void volume, indicating true solubilization of the enzyme from mierosomal and mitoehondriai membranes. In both organelles, pregnenolone and DHEA activities showed identical solubilization profiles with change of pH or extraction time. These argue for a single protein responding to changes in physico-chemical conditions. Among a11 buffers tested, glycerolJEDTA conferred most stability on the 3,L?-HSDcomplex. EDTA increased stability in all buffer systems. The chelating effect of EDTA may account for this. Hiwatashi et al (16) found bovine adrenocortical microsome B/3-HSD stable for at least 3 months at -80 C. Byrne et af (10) f ound human adrenal microsomal 3@HSD stable for over a year in liquid nitrogen. Rabe et al (5) found the activity in human placental mitochondria stable for at least 6 months at -20 C. They showed, however, that Cu++, Zn++ and Cd+’ severely inhibit it. Mn’+, Fe++, Co++ and Nif+ inhibit it moderately. Some proteases (e.g., carboxypeptidase A and thermolysin) are metalloenzymes (17). By chelating Znf+ and other bivalent cations, EDTA may inactivate these proteases and protect 3#?-HSD. In addition to arguing for a single 3/3HSD protein, our data suggest that the same enzyme protein exists in the microsomes and mitochondria. We found pregnenolone/DHEA activity ratios of 1.94 f 0.07 in microsomes and 2.28 it: 0.16 in the mitochondria. Detergent/protein ratios (w/w) giving optimal 3@-HSD percent solu~lization were 0.34 in the microsomes and 0.37. in the mitochondria. The differences between these figures are insignificant. In both subcellular organelles, optimum enzyme solubilization occurred within 10 min in our system. Most importantly,

HtvJ&fAN3~~~~~ST~~~~

DE~~~~E~AS~

363

the enzyme protein molecular weight estimated in the microsomes (208,000 daltons) is concordant with the m.olecular weight (220,000 daltons) in mitocbondria. We ale continuing our experiments to purify mitochortdrial and microsomal 3pHSD in preparation for affinity alkylation studies which will further characterize this enzyme’s substrate specificities.

This work was ,supported by the National Institutes of Health Award I-ID20055 (RCS). We wish to thank Barbara Knalhoff for preparation of the manuscript and Robert Henry for the electron photomicrography. NOTES *To whom correspondence

should be addressed.

REFERENCES 1. 2. 3. 4. 5. 8. 7. 8. 9. 10. 11, 12. 13, 14. 15. 13. 17.

Ward, M-G. and Engel, L.L., J. BfUL. CBEM, 219,BS flQ@). Inane, if,. and Tamaoki, B., ~N~~~R~~ULOGY Q*, 1074 (IS@). Ferre, F,, Breuviller, M., Cedard, L., Duchesne, M.J., Saintot, M,, Rescomps, B. and Crastes de Paulet, A,, STEROIDS Zb, 551 (1975). Gibb, W., STEROIDS 83, 459 (1979). Rabe, T., Brandstetter, K., Kellermann, J. and Runnebaum, B., J, STEROID BIOCBEM. x2,427 (1982). Gibb, W, and Hagerman, D.D., STEROIDS 28, 131 (lQ78$ G&lay, J., Yincent, M., DePaiiterets, C. and Alfsen, A., B~O~~~, BIOPHYS. ACTA 619, 79 (1978). Ford, H.C. and Engel, L.L,, J. BIOL. CHEM. 149, 1363 (1974). Eastman, AR. and Neviile, A,M., J.,ENDOCRINOL. 72, 225 (1977). Byrne, C.C,, Ferry, Y.C. and Winter, J.S.D., J. CLIN. ENDQCRINOL. METAB. 511, 934 (1985). Gibb, W., STEREOS Q7, 23 (1981). Csapo, A.[., CIBA FOUND, SYMP. 47, 159 (1977). Bradford, MM., ANAL. BIOCHEM. 72, 248 (1976). Inano, H., Hayashiyama, J, and Tamaoki, B., J. STEROID BIOCHEM. 16, 587 (1982). Phiipott, JR. and Peron, F.G., EN~OCRrNOLOGY 88, 1082 (lQ?l). Hiwatashi, A., Hamamoto, I, and Echikawa, Y., J. BIUCHEM, 98, lSl9 (1985). Metrier, DE. in: ~~Q~ke~~8tr~. The Ckemied Beuctiolbct of Liviszg Ceifa, Academic Press, New York, p 378 (1977).