79

Biochimica et Biophysics Acta, 529 (1978) 79-87 0 Elsevier/North-Holland Biomedical Press

BBA 57154

SOLUBILIZATION AND SEPARATION OF A5,3&HYDROXYSTEROID DEHYDROGENASE AND 3-OXOSTEROID-A’-A’ISOMERASE FROM BOVINE ADRENAL CORTEX MICROSOMES

J. GALLAY, M. VINCENT, C. DE PAILLERETS and A. ALFSEN

Labomtoire des Etats Lit% Molkulaires, Equipe de Recherche 64 du Centre National de la Recherche Scientifique, U.E.R. Biomkdicale des Saints P&-es, Universite Rend Descartes, 45, rue des Saints P&es, 75006 Paris (France)

(Received August 31st, 1977)

Summary

A physical separation of A5,3P-hydroxysteroid dehydrogenase and 3-0x0- steroid A4-A5-isomerase solubilized from bovine adrenocortical microsomes is described for the first time. The solubilization as well as the separation was car- ried out with a mixture of a detergent: a substituted betaine (Empigen BB/P) and sodium cholate. This latter detergent protects isomerase from complete inactivation by Empigen and is necessary for the recovery of a significant amount of soluble isomerase. Separation of dehydrogenase and isomerase was successfully accomplished by the use of a DEAE-Biogel A anion-exchanger. Dehydrogenase activity was eluted, while the isomerase was retained. Measure- ments of dehydrogenase activity with androst-5-en-3/3-ol-17-one, pregnen-3/3- ol-20-one and pregnd-en-(3fl,17a)diol-20-one and of isomerase activity with androst-&en-(3,17)dione and pregn-5-en-(3,20)dione suggested that more than one isomerase and more than one dehydrogenase form were present.

Introduction

In the adrenal cortex, steroid metabolism involves a sequence of enzymic reactions located either in the mitochondria or endoplasmic reticulum mem- brane [l]. In the mitochondria, the soluble non-heme iron protein adreno- doxin, adrenodoxin reductase and cytochrome P-450 [2-41 are the most studied.

The enzyme group in the microsomal membranes catalyses 17~ and 21-

Abbreviations: dehydroepiandrosterone, androat-5-en-3pol-17-one; S-androstenedione, androst-S-

en-(3,17)-dione; pregnenolone, pregn-S-en-3Q-ol-20-one; 17whydroxy pregnenolone. pregn-S-en-

(3&17a)diol-20-one; 5-pregnenedione. preen-S-en-(3,20)dione; dehydrogenase. A5,3@-hydroxy- steroid dehydrogenase (EC 1.1.1.51): Isomerase. 3-oxosterold-A4-A5-isomerase (EC 6.3.3.1); HEPES, N-2-hydroxyethylpiperazine-N’-2-ethanesulphonic acid: SDS, sodium dodecyl sulphate.

80

hydroxylations [5-71, Ci7-C& side-chain cleavage [8] and 3fl-hydroxy oxida- tion, A5-A4 isomerization of steroids [1,9-111. Other enzymes are also present but are less important [ 121.

With regard to the A5,3P-hydroxysteroid dehydrogenase, A4-A’-isomerase sequence, the question of the existence of one enzyme with both activities or at least two enzymes has often been discussed [13-161. Concerning the sub- strate specificity of each enzymic reaction, a number of hypotheses have been proposed on the basis of indirect kinetic data with inhibitors [13-171, or par- tial solubilization [13,14] or studies of pathological cases [l&19]. The only evidence at present is that the isomerase and dehydrogenase are located in the endoplasmic reticulum membranes; the activity associated with the mitochon- drial fraction is regarded as an artefact [20,21].

The main difficulty with the elucidation of these problems is the solubiliza- tion and purification of enzymes from the lipidic phase in which they are embedded. Ford and Engel [22], using a Triton X-100 extract of the micro- somal membrane from sheep adrenal cortex, claimed to have solubilized and purified to homogeneity a protein showing both activities. With the same method, other workers have succeeded in solubilizing the dehydrogenase from placental and adrenocortical microsomes 123,241. These authors were unable to purify the dehydrogenase in the presence of Triton X-100 at a concentration above the critical micelle concentration. Moreover, they did not mention the simultaneous detection of isomerase activity.

In this laboratory, the use of divalent cations at high concentration led to the fragmentation of the bovine adrenal cortex microsomes into smaller units bearing isomerase activity [25,26]. Later studies have shown that dehydro- genase activity was also present but to a lesser extent (solubilization yields 50 and 2.5% of the original microsomal activity for isomerase and dehydrogenase, respectively). The resulting particles are similar to the native membranes as shown by fluorescence techniques [27] and chemical analysis [28], but it is not possible to separate the individual enzymes using this method. The proteo- lipidic subunits are dependent on the presence of high levels of divalent cations

(Sr , 2+ Ca2+ or Mg2+) for their stability and enzymic activities. Furthermore, the precipitate of lipids and proteins obtained after elimination of the divalent salt remains almost insoluble, even in the presence of detergents. Studies with phospholipases have also failed to solubilize these enzymes but give clear indications about their phospholipid dependence [ 291.

Using native bovine adrenocortical microsomal membranes as starting mate- rial, various detergents have been tested for their ability to dissolve the proteo- lipidic structure of the membrane in relation to enzymatic activities. Many surfactants, including lysolecithins (solubilization yields 80% with La- lysophosphatidylcholine palmitoyl) and Triton X-100, are able to dissolve the dehydrogenase in an active form. In contrast, the isomerase activity is lost with Triton X-100, in disagreement with the recent results of Ford et al. [22].

This paper describes a method for solubilization of the adrenal cortex microsomes allowing for the first time the separation of the enzymes in the A5,3/3-hydroxysteroid oxidation, 3-oxosteroid-A4-A’-isomerization sequence.

81

Materials and Methods

Chemicals Steroids were kind gifts from Roussel-Uclaf (Romainville, France) and dis-

solved in ethanol. NAD+ was grade III from Sigma Chem. (St. Louis, U.S.A.). Empigen BB/P a gift from Albright and Wilson Ltd. (Whitehaven, U.K.) was an aqueous solution of dimethyl monofatty alkyl betaine based on C12/C14 fatty alkyl. The typical composition (w/v) as given by the manufacturer is as follows: CI#.& betaine 30 + 1; free amine 0.3 max.; NaCl 7.5; water content, balance to 100%. Sodium cholate and Triton X-100 were from Merck (Darmstadt, G.F.R.). All other chemicals were analytical reagents and were used as supplied.

Microsomes preparations The crude microsomes prepared as previously described [26] were washed in

0.1 M Tris - HCl (pH 7.5) by the addition of 10 ~01s. to give a final membrane suspension of approx. 1.8 mg - ml-* protein and centrifuged at 105 000 X g for 30 min in a Ti50 rotor of a Beckman Spinco model L50. The pellets were then homogenized in a Teflon glass homogenizer in the same buffer plus 0.1 M KC1 and centrifuged for 30 min at the same speed. The pellets were resuspended in 0.02 M HEPES buffer (pH 7.3)/20% (w/v) glycerol and stored in liquid Nz at a concentration of approx. 15 mg - ml-’ protein, as determined by the Lowry et al. method [30], with ovalbumin as standard (after dissolution of the mem- branes in SDS).

Enzyme assays Isomerase. Determination of the enzymic activity of the isomerase used the

ultraviolet absorption properties of the conjugated ketone product of the reac- tion as described before [26]. The substrate concentrations were either 117 PM 5androstenedione or 5 PM 5-pregnenedione and incubation at 30°C.

Dehydrogenase. Oxidation of the 3fi-hydroxyl group is followed with three physiological substrates (dehydroepiandrosterone, pregnenolone, and 17a- hydroxypregnenolone) and NAD’ as coenzyme. l-100 1.11 of the enzymatic solution were placed in a 1 X 1 cm Suprasil quartz cuvette containing 20 mM Tris - HCl (pH 8.5), 0.2 mM NAD+ and 0.5% 2-mercaptoethanol. Steroid con- centrations were 6 PM dehydroepiandrosterone, 3 PM pregnenolone and 3 PM 17a-hydroxypregnenolone. For the last two steroids, 5% (v/v) ethanol were added in the assay (final volume 1.5 ml). A sensitive fluorescence technique has been used to follow the appearance of reduced NAD+. The increase of the fluorescence signal at 470 nm (upon excitation at 340 nm), was recorded after addition of the enzyme. The fluorescence was recorded on a FICA55 spectro- fluorometer equipped with an Ifelec 3802 recorder at 36°C. Under these con- ditions, a zero-order rate is observed during 10 min. A solution of 1 mg/l reduced NAD’ in 20 mM Tris - HCl (pH 8.5) was used as standard.

Solubilization of the membranes 2% (w/v) sodium cholate/0.45% Empigen BB/P (w/v, final betaine concen-

tration)/20% glycerol (w/v)/100 PM dithioerythrito1/200 mM HEPES (pH 7.3) was added dropwise to an equal volume of membrane suspension (15 mg/ml

82

protein) with stirring at 0°C. After 15 min incubation at O”C, the suspension was centrifuged at 200 000 X g for 1 h at 4°C. The supernatant was stored at 0” C before use. Solubilization as a function of detergent concentration was car- ried out in the same way, but the membrane suspension was diluted 10 times in the solubilization solution. Solubilization with Triton X-100 was performed in the above described buffer with or without 0.2% (w/v) sodium cholate.

DEAE-chromatography The supernatant solution was loaded onto a Sephadex G-25 column for

equilibration with the DEAE-Biogel A buffer: 50 mM Tris * HCl (pH 7.8)/ 20% (w/v) glycerol/0.2% (w/v) sodium cholate/0.02% Empigen BB/P (w/v) final betaine concentration/100 PM dithioerythritol. The active proteins eluted in the void volume were layered onto a DEAE-Biogel A column (Biorad Lab., 2 X 16 cm). Proteins were eluted with a linear gradient of O-O.1 M KC1 in the same buffer. All operations were carried out at 4°C.

Results and Discussion

Solubilization with Empigen BB/P and Triton X-l 00 Effect of sodium cholate. The detergent : protein ratio is of primary impor-

tance in the solubilization of intact membrane proteins. This has been exten- sively discussed in recent reviews [31,32]. A more pronounced inhibition of isomerase (Fig. 1) than dehydrogenase (Fig. 2) was observed as a function of Empigen BB/P concentration. Likewise, a higher level of dehydrogenase activ- ity (18% of the original microsomal activity) in the concentration range of 0.04-0.06% (w/v) Empigen BB/P was found in the 200 000 X g supematant while this percentage was only 7-8s for the isomerase activity (Figs. 3 and 4). The rate-limiting exchange of the detergent-protein-lipid mixed micelles is likely of primary importance in the initial solubilization of the membranes. The general principle involved is that amphiphiles of low critical micelle concentra- tion, as is the case for Empigen BB/P, 2 and 0.15 mM for Cl* and Cl4 betaine, respectively, [33] are slow to dissociate from any particles with which they are associated. This environmental equilibration of slowly reacting detergents can be accelerated by the simultaneous presence of a faster reacting detergent [ 341. An example is provided by the ability of cholate to accelerate the exchange of vesicular phospholipids with endogenous lipids of sarcoplasmic reticulum mem- branes [ 351.

From these considerations, experiments with mixtures of Empigen BB/P and sodium cholate have been undertaken. Sodium cholate exhibits a protective effect on isomerase activity while the reverse is true for dehydrogenase (Figs. 1 and 2). This indicates that other phenomena are likely to occur, perhaps a spe- cific effect of cholate on the isomerase protein, since it has been observed that cholate is a competitive inhibitor of this enzyme. Nevertheless, both enzyme activities are increased in the 200 000 X g supernatant at cholate and Empigen BB/P concentrations of 0.2% (w/v) and 0.03-0.04% (w/v), respectively, (deter- gent : protein ratio of 1.5-1.6 : 1; Figs. 3 and 4). The results of solubilization at this ratio are summarized in Tables I and II. Two isomerase substrates and the three substrates of the dehydrogenase have been used for the determination

EMPIGEN 60/P (%) EMPIGEN B&P C X)

Fig. 1. Vaxiation of original isomerasc activtty as a function of Empigcn BB/P WV) concentration. e-----e, no sodium cholats added; 0 - - - - - 4,0.1% (w/v) sodium cholatc; *a - * - * ‘A, 0.296 (w/v) sodium cholate and A-. -. -A, 0,496 (w/v) sodium cholate. Activities determined with 6-sndr&.enedbone as sub- strate are exprewd i nlative units titb respect to the activity of the membrane without detswnt. Experimental conditions sre described in MaterUs and Methoda

Fig. 2. Variation of originsI dehydrogenase activity 88 a function of Empigen BBp (w/v) concentration. l -, no sodium cholate added; o - - - - - -0, 0.1% (wfv) sodium cholate; ** * * * * 4 0.2% (w/v) sodium aholate: b- * -. --d, 0.4% (w/v) sodium cholate. ActivWes determined with dehydxoeptandrosterone are expressed in rclativa units with respect to the activity detergent. Experimental conditions are described in Materials and Methods,

of enzyme activities. An average of 30 and 20% original microsomal activity for dehydrogenase and isomerase, respectively, were present in the 200 000 X g supematant with each of the substrates. Solubihzation experiments of the microsomes with Triton X-100 gave results somewhat different from those described by Ford et al. [22]: no isomerase activity could be measured in the 200 000 X g supematant (Fig. 5). For the dehydrogenase activity, the solubti- zation yield was 30% of the original microsomal activity (Fig. 6) which is com- parative with the Empigencholate solubilization yield. No protective effect of cholate against inhibition of the isomerase with Triton X-100 was observed.

Separation of dehydrogenase and isomerase Sepharose CL 6B chromatography was not suitable for the separation of the

enzymes although allowing a rough estimation of their sizes in the range of 9 - lo4 to 1.5 l lo5 daltons. A complete separation was achieved with the use of a DEAE-Biogel A anion exchanger. Dehydrogenase activity was eluted in the initial wash, while elution of the isomerase required a slight increase of the ionic strength (Fig. 6). No contamination of the dehydrogenase peaks with isomerase activity was observed; the same was shown for the isomerase. This

01 02

EMPIGEN BB/P ( %)

0’1 62

EMPIGEN BE/P (% )

Fig. 3. Variation of isomerase solubilixation yield ln the 200 000 X g supernatant asa function of Em&en BB/P concentration (96, w/v. final betaine concentration). Legends and experimental conditions are as in Fig. 1. Insert: specific isomerase activity in the 200 000 X g supernatant as a function of Empigen BB/P concentration (%, (w/v) final betalne concentration).

Fig. 4. Variation of dehydrogenass solublllzation yield in the 200 000 X g supernatant as a function of Em&en BB/P concentration. Legends and experimental conditions ss in Fig. 3. Insert: specific dehy- drogenase activity in the 200 000 X g supernatant as a function of Emplgen BB/P concentration (% (w/v) final betaine concentration).

separatjon of the activities and the fact that the behavior of each enzyme activ- ity is always different whatever the detergent used, strongly suggest that the dehydrogenase and isomerase moieties can be considered as distinct protein entities. Moreover, two dehydrogenase peaks are observed: in the first, dehy- drogenase activity towards dehydroepiandrosterone, pregnenolone and 17~ hydroxypr~~nenolone was present, in the second, the oxidation of dehydro-

TABLE I

SOLUBILIZATION OF MICROSOMES WITH

3-OXOSTEROID-A4A&ISOMERASE FROM BOVINE ADRENAL CORTEX A MIXTURE OF 0.23% (w/v) EMPIGEN BB/P AND 1% (w/v) SODIUM

CHOLATE

Final detergent: protein ratio, 1.6 : 1 (w/v). Experimental conditions are described in Materlals and Methods.

Substrates 5-Androstenedione bpregnenedione

Yield Specific activity Yield Specific activity

(%) (run01 * min-r . Pm (nmol * min-r * mP) mfC9

Microsome activity Microsome activity

after addition of detergents 200 00 x g supernatant

activity

100 * 480 100 ** 160

20-46 96-221 20-30 30-44

17-31 62-114 13-21 15-24

109% Activities are taken as * 8100 nmol - mm-r. ml membrane suspension; ** 2700 nmol . min-r ’ ml-r.

TA

BLE

II

SO

LU

BIL

IZA

TIO

N

OF

A

5,3

&Ii

YD

RO

XY

ST

ER

OID

D

EH

YD

RO

GE

NA

SE

FR

OM

A

DR

EN

AL

CO

RT

EX

M

ICR

OSO

ME

S

WlT

Ii

A

MIX

TU

RE

O

F

0.2

3%

(w

/v)

EM

PIG

EN

B

B/P A

ND

1%

(w

/v)

SO

DIU

M

CH

OLA

TE

Fin

al d

eter

gen

t:

Dm

tein

rat

io 1.6

: 1

(w

/v).

Exper

imen

tal

con

dit

ion

s er

e dee

cnie

d

in M

ater

ials

an

d M

ethods.

Subs

trat

es

Deh

ydro

epia

ndro

ster

on

e R

egn

enolo

ne

Yie

ld

f%)

spec

ific

ac

tivi

ty

(nm

ol

* m

icz

* mg’

) Y

&Id

w>

spec

ifk

acti

vity

(m

u01

. m

inx

- rn

g”)

Yie

ld

(%)

Spec

ific

ac

tivi

ty

(mnoI

* m

&-l

* rn

g9

Mic

roso

me

acti

vity

100 *

30

100 *

* 54

100 **

* 4.7

M

icro

som

e ac

tivi

ty

afte

r ad

dit

ion

of

det

erge

nts

2647

6.4

-14.1

20-4

0

11-2

2

30-6

0

1.4

-2.3

200 0

00 X

g s

upem

atan

t ac

tivi

ty

20-4

0

1144

16-3

5

11.4

-26

28-3

2

1.7

2-2

100%

act

ivit

ies

are

taken

as

* 760 m

nol

- m

ine1

. ml-

Lm

embr

ane

susl

pem

don

. **

980 m

nol*

min

-l*m

I-‘a

nd

***

86 m

nol-

m

ixrl

*mI-

?

--__‘-_-_---__. _____________________ a---------- .

_______-A--------a----- ___A___-----

32 0.4 0.6 OB

.I00

20

TRITON X OI f/.1

Fig. 5. Closed symbols: relative isomerase (0) and dehydrogenase (A) activities as a function of Triton

X-100 concentration (J, w/v). Open symbols: solubilization yield of isomerase (0) and dehydrogenase (A)

as a function of Triton X-100 concentration, Experimental conditions are described in Materials and

Methods. Isomerase and dehydrogenase activities were determined with bandrostenedione and dehy-

droepiandrosterone, respectively.

epiandrosterone and pregnenolone only was detectable. The isomerase activity towards 5-androstenedione and 5-pregnenedione was not eluted exactly in the same fractions. This should be compared to the recent findings of Orta-Flores et al. [ 361, who have suggested the existence of two 2l-hydroxylation systems, one specific for progesterone, the other for progesterone and 17a-hydroxy- progesterone. Since the experiments have been performed in the presence of a mixture of two detergents, a possible specific effect on each enzyme activity

oD280nm /“r

Iehydrogenose lsomemsa Actwty 4ct1v1ty

t

AOD248nml?4

0.6pM NADH,H+mr?l o,oo,

FRACTION NUMBER (4 ml)

Fig. 6. Elution profile from DEAE-Biogel A column. Fractions (4 ml) are assayed for isomerase and dehy-

drogenase activities (see Materials aud Methods for assaying procedure) =w&ivelY with &r&ost8ne-

dione (9- * -. -9); 6-pregnenedione (a- - - - - -@) and with dehydroepiandrosterone (o- . - * 4). Preg-

nenolone (o- - - - - -0) and 17whydroxy pregnenolone (*. . . . . ..I. Az80nm (-0).

87

cannot be excluded. The only way to demonstrate the existence of different enzyme systems for the As,3&hydroxysteroid dehydrogenase, 3-oxosteroid- A’-A’-isomerase will be the complete pu~fi~ation and ~h~acte~zation of these enzymes. This is in progress in our laboratory.

Acknowledgements

This work was supported by C.N.R.S. grant (E-R. 64), I.N.S.E.R.M. No. CA- 763-152 02 and D.G.R.S.T. grant No. A 659-1364.

References

1 Samuels, LT. and Uchiiawa, T. (1967) in The Adrenal Cortex (Eisenstein, A.B., ed.), p. 61, Little, Brown and Co., Boston

2 Lambeth, J.D., McCasIhr. D.R. and Kamin. H. (1976) J. Biol. Chem. 251.7545-7550 3 HannPii6. W. and Khnura. T. (1976) J. Biol. Chem. 261.6068-6074 4 PauI, D.P., Gallant. S., Orme-Johson. N.R., Orme-Johson. W.H. and Brownie, A.C. (1976) J. Biol.

Chem., 261,7120-7126 5 Masters, B.S.S., Taylor, W.E., Isaacson, E.L., Baron, J., Harkins. J.B., Nelson, E.B. and Bryan. G.T.

(1973) Ann. N.Y. Acad. Sci. 212.76-88 6 Cooper, D.Y., Estabrook. R.W. and Rosenthal, 0. (1963) J. BioL Chem. 238,1320-1323 7 Sweat, M.L., Dutcher, J.S., Young, R.B. and Bryson, M.J. (1969) Biochemistry 8.4956-4963 8 Yates, J., Deshpande. N. and Goldman, AS. (1975) J. Steroid Biochem. 6, 1325-1327 9 Samuds, L.T.. Hehnreich. M.L., Lasater. N.B. and Reich, H. (1951) Science 113, 49*491

10 Beyer, K.F. and SamueIs. L.T. (1956) J. Biol. Chem. 219.69-76

11 Goldman. A.S. (1968) J. CIin. Endocrinol. 18,1539-1546 12 Ghraf. R., Vetter, II. and Schrieffers, V. (1974) Hoppeseyler’s Z. Physiol. Chem. 355, 543-560 13 EwaId, W.. Werbin, H. and Chaikoff, I.L. (1964) Biochim. Biophys. Acta 81.199301 14 Cheatum, S.G., DouviIIe. A.W. and Warren. J.C. (1567) Biochim. Biophya. Acta 137‘172-178

15 NevIBe, A.M., Oir. J.C. and Engel. L.L. (1969) J. EndocrInoI. 43.599-608 16 McCune. R.W., Roberts, S. and Young, PL. (1970) J. BioL Chem. 245.3859-3867 17 Yates, J. and Deshpande. N. (1975) J. Endocrinol, 64,195-196 18 Wehcki, I., and Engel. L.L. (1963) J. Biol. Chem. 238,1302-1307 19 NeviIIe. A.M., Webb, L.J. and Symington. T. (1969) Steroids 13. 821-823 20 Bash, R.S.. Finegold, M.J. (1971) Biochem. J. 125, 983-989 21 Cowan, R.A., Giles, CA. and Grant, J.K. (1974) J. Steroid Biochem. 5, 607-608 22 Ford, H.C. and Engel. L.L. (1974) J. Biol. Chem. 249.1363-1368 23 GIbb, W. and Haperman, D.P. (1976) Steroids 28.31-41 24 Eastman, A.R. and NeviIIe, A.M. (1977) J. Endocrinol. 72.225-233 25 GaBau. J.. Vincent, M. and Alfsen, A. (1975) Bioohim. Biophys. Acta 397, 501-509 26 Geynet, P., Gallay. J. and Alfsen, A. (1972) Eur. J. Biochem. 31, 464-469 27 Gallay, J., Vincent, M. and A&en, A. (1975) Biochim. Biophys. Acta 397.489-500

28 Alfsan. A., GaIIay, J.. Vincent. M. and de PailIerets, C. (1976) in Water in BiologicaI Systems, ColIo- clus International du C.N.R.S. No. 246 (Alfsen, A. and Berteaud, J., eds.), pp. 161-168, Editions du C.N.R.S., Paris

29 Geneyt, P., de PaiBerets, C. and Alfsen. A. (1976) Eur. J. Bioehem. 71.607-fl12 30 Lowry, O.H., Rosebrouah, N.J.. Farr. L.W. and Randall, R.J. (1951) J. Biol. Chem. 193.265-275 31 Tanford. C. and Reynolds. J.A. (1976) Biochhn. Biophys. Acta 457. 133-170 32 Helenius, A. and Symons. K. (1975) Biochhn. Biophys. Acta 415. 29-79 33 Molyneux. P., Rhodes, C.T., Swarbick, J. (1965) Trans. Faraday Sot. 61, 1043 34 Racker, E.. Chien, T.F. and Kandraeh, A. (1975) FEBS Lett. 57.14-18 35 Warren G.B., Toon, P.A., Birds& N.J.M., Lee, A.G. and Metcalf+!, J.C. (1974) Biochemistry 13,

5501-5507

36 Orta-Ffores, 2.. Cantu, KM. and Dominauez. O.V. (1976) J. Steroid Biochem. 7.761-767