Eur. J. Biochem. 64, 573-581 (1976)

Hepatic 15-Hydroxylation of Corticosteroids in the Rat Substrate Specificity Studied in the Isolated Perfused Liver Hdkan ERIKSSON Kemiska Institutionen I, Karolinska Institutet, Stockholm (Received July 29. 1975/January 21, 1976)

The substrate specificity of a sex-specific hepatic 15-hydroxylase active on different C,,O, and C,,O, steroids was studied in the isolated perfused liver from female rats. Liquid-chromatographic separation methods in combination with computerized gas chromatography-mass spectrometry was employed to identify the metabolites formed. The majority (between 75 - 90 %) of 15-hydroxylated compounds isolated were present as monosulphate conjugates while smaller amounts of disulphates were also detected. Hydroxylation was found to take place exclusively at position 158. A certain number of 11-deoxy-21 -hydroxy, 11-0xo-21 -hydroxy and 11,21-dihydroxy steroids with a 3-keto-d4-, 3-keto-Sa(SB)-, 3a,5 a- or 38,5P-structure were readily converted to 1SB-hydroxylated metabolites. Depending on the structure of the substrate, between 20 and 87% of the total metabolites formed were 15 b-hydroxylated. 5 a-Reduced steroids were better substrates for the hydroxylase than the corresponding 3-keto-d4- or SP-reduced compounds. The configuration of the hydroxylgroup at C-3 did not affect the degree of 15-hydroxylation. 11 8-Hydroxylated steroids served as better substrates than the corresponding 11-dehydro-, 11-deoxy- or 1 1 a-hydroxy compounds. 5 r-Dihydrocorticosterone and 3 a,5 a-tetrahydrocorticosterone were the best substrates for the 15b-hydroxylase

Corticosterone, the predominant glucocorticoid hormone in the rat undergoes reductive metabolism in the liver yielding 3 a, 118,21 -trihydroxy-5 a-pregnan20-one and 3 ~,11~,21-trihydroxy-5 a-pregnan-20-one as the quantitatively most important biliary C,,O, steroids in the female animal whereas 3 8 , l l p,21-trihydroxy-5 a-pregnan-20-one and 5 a-pregnane-3 P,11/3, 20P,21-tetrol constitute the major metabolites in bile from the male rat [1,2]. However, the major biliary corticosterone metabolite in the female animal is 3 a, 3 1 8,21-trihydroxy-5 a-pregnan-20-one hydroxylated in position C-15 [1,3]. This hydroxylation, which is sex-specific, is known to occur in the liver [4]. The 21-monosulphate of corticosterone (or of a corticosterone metabolite) probably serves as substrate for the hydroxylase [ S ] . Enzyme. Steroid hydroxylase (EC 1.14.-.-). Note. Retention times fR are calculated relative to 5 wcholestane. Triciul Names. Corticosterone, 1 3 p.21-dihydroxy-4-pregnene3,20-dione; 5 a-dihydrocorticosterone, 118,21 -dihydroxy-5 cc-pregnane-3,20-dione; 5 B-dihydrocorticosterone, 11 P,21-dihydroxy-5 ppregnane-3,20-dione; 11-dehydrocorticosterone, 21 -hydroxy-4-pregnene-3,11,20-trione; deoxycorticosterone, 21-hydroxy-4-pregnene3,20-dione; progesterone, 4-pregnene-3,20-dione; silyl, trimethylsilyl; oxime silyl, 0-methyloxime trimethylsilyl ether.

The present investigation was undertaken to investigate the substrate specificity of this hepatic 15-hydroxylase system. Since the enzyme probably utilizes steroid sulphates as substrates, this study was carried out using the isolated perfused liver, a system where sulphatation, oxido-reduction and hydroxylation of steroids take place to almost the same extent as in uiuo [4].

EXPERIMENTAL PROCEDURE

Reference Compounds Corticosterone, 11-dehydrocorticosterone and deoxycorticosterone were purchased from Sigma (Sigma Co., U.S.A.). Sa-(and 5 P)-Dihydrocorticosterone, 3 a,lI~,21-trihydroxy-Sa-pregnan-20-one, 3 a(and 3 p), 11 8,21-trihydroxy-5 fi-pregnan-20-one, 3 421 -dihydroxy-5 a-pregnan-20-one and 3 a,21-dihydroxy-5 tlpregnane-I 1,2O-dione were obtained from Ikapharm (Ramat-Gan, POB 31, Israel). Professor W. Klyne donated 3 /3,118,21-trihydroxy-5 a-pregnan-20-one and 11 ~,21-dihydroxy-4-pregnane-3,20-dione from the Medical Research Council Steroid Reference Collection

574

(London). Dr J. Babcock (Upjohn Co., Kalamazoo, Mich., U.S.A.) kindly supplied 15 a,21-dihydroxy-4pregnene-3,11,20-trione and 15 a-hydroxy-4-pregnene3,20-dione. Dr R , Neher gave 15a,21-dihydroxy- and 15 p,21 -dihydroxy-4-pregnene-3,20-dione,while 1 5 bhydroxy-4-pregnene-3,20-dionewas a gift from Dr J. Fried. 3 b,15 421 -Trihydroxy-5 a-pregnan-2O-one, 3 b, 15b,21-trihydroxy-5 a-pregnan-20-one, 30,15 a,21 -trihydroxy-5 a-pregnan-I 1,20-dione and mixtures of the 3 a,5 ~ ,3 a ,5 B an d3 8 3 8 isomers of 3,15a(and 158),21trihydroxy-5-pregnan-20-one and 3,15 a,21 -trihydroxy5-pregnan-I 1,20-dione were prepared by hydrogenation with palladium on charcoal in ethanol of 15=(and 15p),21-dihydroxy-4-pregnene-3,20-dione and 15 a,21dihydroxy-4-pregnene-3,I1,20-trione, respectively. 3 M (and 3/3),11fl,I5a,21-Tetrahydroxy-5a-pregnan-20-one were obtained by reduction of 15 a,21-dihydroxy-4pregnene-3,11,20-trione. [4-14C]Corticosterone (specific activity 56.7 mCi/ mmol) and [ 1 ,2-3H]deoxycorticosterone (specific activity 55.0 Ci/mmol) were purchased from the Radiochemical Centre (Amersham, England). The purity of the compounds was assayed as previously described [61. Digestive juice from Helix pomatia was obtained from Industrie Biologique FranCaise (Gennevilliers, France).

Animal Experiments Female rats of the Sprague-Dawley strain, weighing between 230 - 270 g, were used. Prior to operation the animals were kept at 20 'C and had free access to food and water. Isolated liver perfusions were carried out as described in detail elsewhere [6,7]. 12 ml of homologous rat blood, diluted with 48 mi Krebs-Henseleit bicarbonate buffer [8] was used as perfusion medium. Oxygenation was accomplished by a membrane oxygenator using O,/CO, (95/5). The pH of the perfusion medium was measured continuously and maintained at 7.40. In each perfusion 5 mg of steroid mixed with [4-'4C]corticosterone (0.5 x 1O6 counts/min) was used as substrate (except when deoxycorticosterone was the substrate, when [I ,2-3H]deoxycorticosterone was added). The steroid was dissolved in 50 pl of ethanol which was added to 3.0 ml of 1% (w/v) serum albumin and the mixture given by constant infusion during 120 min at a rate of approximately 25 pl/min, Bile was collected over a period of 180 min. To isolate a large amount of endogenous 3 a,l1/3, 15,21-tetrahydroxy-5 a-pregnan-20-one from rat bile, twenty female animals were supplied with bile fistulas and bile was collected in 6-hourly portions for one week. By this means, approximately 3 1 of bile was obtained.

15-Hydroxylation of Corticosteroids

Extraction und Purification oj Biliury Steroids The bile samples from all perfusions (1.5-2.8 ml) were analyzed essentially as described in previous publications [ I , 61. After extraction from bile the steroids were separated according to conjugate class. Steroid mono- and disulphates were hydrolysed with enzymes of the digestive juice from Helix pomatiu followed by solvolysis in acidified ethyl acetate [9,10]. The steroids liberated after hydrolysis were purified on Unisil '' [l] and used for derivative formation.

Prrpurcition of 3cn,110, i5.21- Tetrahydrroxy-5x-pregnan20-one from Bile The 3-1 pool of rat bile was extracted with 10 volumes of acetone/ethanol ( l / l , v/v) for 24 h at 39 'C. The extract was evaporated to dryness, redissolved in 70% (v/v) ethanol and passed through an Amberlyst-I 5 column in the sodium form. The eluate was evaporated and chromatographed on a 100 g LH-20 column prepared in and eluted with chloroform/methanol ( l / l , v/v), 0.01 M with respect to NaCl [l].The monosulphate fraction was obtained between 650 - 1 000 ml of effluent. Hydrolysis of the sulphates was carried out as described above and the liberated steroids were separated on a 20 g LH-20 column prepared in and eluted with chloroform/heptane/ethanol (5/5/1, by vol.), saturated with water [I 11. The 3 a , l l b,15,21tetrahydroxy-5 a-pregnan-20-one was obtained between 280-360 ml of effluent. A second chromatography of the steroid in this fraction using the same system yielded about 10 mg of product, with a purity of 9504 (as judged by gas chromatography).

Gas-Liquid Chromutography und Gas ChromatographyMass Spec tronzet ry The steroids obtained from bile were analyzed as their trimethylsilyl ethers [12] and O-methyloximetrimethylsilyl ethers [I 31. Gas-liquid chromatography was carried out using SE-30, HiEff 8 BP and QF-1 as stationary phases while gas chromatography-mass spectrometry was performed with an LKB 9000 instrument equipped with a 1.5 SE-30, or a 1 % HiEff 8 BP column. Mass spectra were obtained by repetitive scanning and stored on magnetic tape for further treatment in an IBM 1800 computer [14].

x,

Radioactivity Mrasuremrnts Radioactivity was measured in a Packard liquid scintillation counter (model 3003), using Instagel '' as scintillator.

575

H.Eriksson Table 1 . M ~ f u h n l i /,joi.mrd i ~ . ~ during prrfuxion oJ isoluted lir!ers,/iom.fcmcrlr r u / s w i / h di/jkrunr ,strroid substrates Substrates

Met it bol i (es Formed

C,’D-ringand side chain configuration cH20n

AIB-ring configuration

3-Keto-d4-

(1)

I

I

I

C=o

I

substrate 1 1 1

3-Keto-A4(2) 3a-Hydroxy-5a- (3)

1V V

3a.21 -dihydroxy-5r-pregnane-l1.20-dione 3a, 15/L21-trihydroxy-5n-pregnane-l 1,20-dione

2.3

3a.I la,21-trihydroxy-5a-pregnan-20-one 3r,l1r.l5/m -tetrahydroxy-5a-pregnan-20-one

4 4

3r,lIfl,21-trihydroxy-5a-pregnan-20-one

5,6,8

CH20H

3-Keto-d4-

(4)

VI VII

V20H

3-Keto-d43-Keto-5a3-Keto-5b3a-Hydroxy-5%3p-Hydroxy-SB-

(5) (6) (7) (8) (9)

V111

I

3a,21-dihydroxy-5a-pregnan-20-one

11 3a,l5b.21-trihydroxy-5n-pregnan-20-one 111 3/?,15fl,21-lrihydroxy-5r-pregnan-20-one

C=O

CH20H

corn- n a m e pound

2,3

1X 3B.1 1~,21-trihydroxy-5a-pregnan-20-one X 3 x 1 1/~,21-trihydroxy-5[~-pregnan-20-one

I

XI

7.9

38.1 lfl.21 -trihydroxy-5fl-pregnan-2O-one

5

XI1 3a.l1~,15/?,21-tetrahydroxy-5r-pregnan-20-one 5.6,s XI 1I XI V XV

3/!,1 l{f,l5~,21-tetrahydroxy-5a-pregnan-20-one 5 3a,lI~.l5~.2l-tetrahydroxy-5fl-pregnan-20-one7 3/j,lI/j.l 5/1,2l-tetrahydroxy-5/1’-pregnan-20-one 9

RESULTS

Identlfication of Mrtrtbolites of Deoxycorticosterone

The compounds secreted into bile during perfusion of female rat livers with different steroid substrates are summarized in Table 1. Between 75 -90% of the metabolites were present as monosulphate conjugates as judged by their chromatographic behaviour on Sephadex LH-20. Two main groups of steroids were identified: ring-A-reduced metabolites, and ring-A-reduced metabolites hydroxylated in position 15. The proportion of these two groups of metabolites varied depending on the substrate. The relative amounts of 15-hydroxylated metabolites formed during perfusion of different substrates are shown in Table 2.

Three metabolites of deoxycorticosterone were isolated both as their mono- and disulphates (compounds 1 - 111). The mass spectrum of the silyl ether of compound I was identical with that of the silyl ether of authentic 3 a,21-dihydroxy-5 a-pregnan-20one, showing a molecular ion at mje 478, a base peak at mJr 257 and prominent peaks at mje 375 (M-103) and 285 (M-(90+103)). Retention time data, as well as the mass spectrum of the oxime silyl ether of compound I confirmed its identity as 3 a,21 -dihydroxy-5 TXpregnan-20-one. The silyl ethers of compounds I1 and 111 yielded very similar mass spectra, with molecular ions at

1 5-Hydroxylation of Corticosteroids

576 Table 2. Amounts of ring-A-reduuced metabolites hydroxylated in position I5 obtuined in bile after perfusion oJ isolutedJemale rat her.r with different substrates Results are expressed as a percentage of total metabolites formed Substrate

Ring-A-reduced metabolites hydroxylated in position I5

% 2 1-Hydroxy-4-pregnene-3,20-dione 48 21 -Hydroxy-4-pregnene-3,11,20-trione 66 3a.21 -Dihydroxy-5 a-pregnane-I 1,2071 dione 1 1 a,21-Dihydroxy-4-pregnene-3,2020 dione 11 p.21-Dihydroxy-4-pregnene-3.2060 dione 11 P.21-Dihydroxy-5 a-pregnane-3,2086 dione 11 P,2l-Dihydroxy-SP-pregnane-3,2042 dione 3 a,l1 P.21-Trihydroxy-5a-pregnan20-one 87 3/J,11 fl,2l-Trihydroxy-Sp-pregnan44 20-one

~~-

the latter compound and those of the corresponding derivatives of compound V were 1.07 and 1.09, respectively. The same ratios were obtained for the corresponding derivatives of 15a- and 15 P-hydroxyprogesterone and for 15a,21- and 158,21-dihydroxy-4pregnene-3,20-dione. In a previous paper it has been shown that acetylation of 15-hydroxylated steroids under mild conditions gives rise to complete derivatization of the 1 5 ~ hydroxyl group, although the 15 P-hydroxyl group is only partially acetylated [5]. When compound V was acetylated under the same conditions, about 60% was derivatized in position 15. Based on these data, compound V was identified as 3 a,l5p,21-trihydroxy-5 a-pregnane-ll,20-dione.

Identijication o j Metabolites of Corticosterone, lla,21Dihydroxy-4-pregnene-3.20-dione,5a (and 5P) -Dihydrocovticosterone, 3a,l1/3,21- Tvihydroxy-Sa-pregnan20-one and 38,ll P , Z l - Trihydroxy-5p-pregnan-20-one

The steroids recovered from the mono- and disulphate fractions of bile after perfusion with corticosterone or reduced corticosterone analogues were m/e 566 and prominent peaks at mje 463 (M-103), identified as different isomers of trihydroxypregnan373 (M-(90+ 103)), 283 ( M - ( 2x 90+ 103)) and 255 20-one or 1 5-hydroxy-trihydroxypregnan-20-one (M-(2 x 90 + 131)) (Fig. 1). These peaks are typical (Table 1). The configurations at C-3 and C-5 of the of a tris (trimethylsily1oxy)mono-oxo pregnane structrihydroxypregnan-20-one isomers (compounds Vl, ture with a primary trimethylsilyloxy group [15]. VllI -XI) were determined by gas chromatographyRetention time analysis and mass spectrometric commass spectrometry as described previously [6]. All parison with reference steroids (Fig. 1) showed that substrates used were partially converted to ring-Acompounds I1 and 111 were identical with 3a- and reduced metabolites with a 15-hydroxyl group (see 3 p,15 8,21-trihydroxy-5 a-pregnan-20-one, respectiTable 1, compounds VII and XI1 -XV, and Table 2). vely. Compound XI1 constituted about 60% of the metabolites in bile after perfusion of corticosterone and Identification of Metabolites of 11-Dehydrocorticoabout 86 when 5 a-dihydrocorticosterone and 3 a, sterone and 3a,21-Dihydroxy-5ci-pregnane-ll,20-dione 11 P,2l-trihydroxy-5 a-pregnan-20-one were perfused (Table 2). In order to determine the exact configuration About 50% of administered 11-dehydrocorticoat C-15 of this quantitatively important corticosterone steroiie was recovered in bile as the mono- and dimetabolite, about 10 mg of compound XI1 was isosulphates of 3 a,21 -dihydroxy-5 a-pregnane-1 1,20-dilated from bile of female rats. The purified steroid one (compound IV), as judged from gas chromatowas analyzed by gas chromatography and by gas graphic-mass spectrometric analysis. The remaining chromatography-mass spectrometry as its silyl- and 50% was converted into a steroid (compound V), the oxime silyl ether derivatives. The mass spectrum of mass spectrum of which is illustrated in Fig. 2. The the silyl ether of compound XI1 (Fig. 3B) displayed same compound accounted for about 77% of the prominent ions atmle 551 (M-l03), 461 (M-(90+ 103)), metabolites of perfused 3 421 -dihydroxy-5 a-pregnane371 (M-(2 x 90 + 103)) and 260 while that of the oxime 11,20-dione. The mass spectrum of the silyl ether of silyl ether derivative showed ions at m/e 175, 188 and compound V showed a molecular ion at mle 580, a 289, indicating a 15,21-dihydroxy-20-keto structure base peak at m/e 387 (M-(90+103)) and prominent [15]. The mass spectra of the silyl ethers of 3a,11 b, ions at 297 (M-(2x90+103)), 269 (M-(2x90+131)) and 260. The same set of ions, but with different rela15a,21 -tetrahydroxy-5 a-pregnan-20-one and comtive intensities, were also seen in the mass spectrum pound XI1 gave the same fragment ions, although of the silyl ether of 3a,15a,21-trihydroxy-5a-pregnan- with different relative intensities (Fig. 3 A, B). Furthermore, gas-liquid chromatographic retention time data 21,20-dione (Fig. 2). The ratios between the retention for these two steroids on three stationary phases inditimes on SE-30 of the silyl- and oxime silyl ethers of

511

H. Eriksson

200

100

300

283

200

300

200

300

400 m Ie

500

100

I373

400 rn Ie

500

400

500

mle Fig. 1. Purtiul n-~assspectra of the .rilyl rthrss o/' 3p,I S x , 2 l - t r i l i ~ ~ r o . ~ ~ - S ~ x - ~ ~ r ~ ~ (~An)~, i38, n - ISli,2I-trihy~iroxy-5n-pre~nun-2O-~~ne 20-onr (B) und compound I I I

cated that they differed only in the configuration of the 15-hydroxyl group. As described in a separate communication [ 5 ] 3 a,l 1 p,15,21-tetrahydroxy-5 a-pregnan-20-one (compound XII) is only partially acetylated in position 15 under mild reaction conditions, a finding which also supports a 15j3-configuration. Eriksson et al. have previously described 15hydroxylated 21 -deoxy-C,,O, metabolites from corticosterone [18], and presented evidence which indicated that these steroids had the unusual 14fi(H) configuration. In order to determine the stereoisomerism at the C/D ring-junction in compound XII, the purified sample was subjected to I3C nuclear magnetic resonance analysis (kindly performed by Dr Alma Burlingame). In this analysis, the chemical shifts of the quaternary carbon atoms (C-10 and C-13)

confirmed that the C/D rings in compound XI1 were trans-fused (14 a (H) ). Based on the above data, compound XI1 was identified as 3 a , l l P,lSfi,21-tetrahydroxy-5a-pregnan20-one.

DISCUSSION Previous investigations have clearly demonstrated that hepatic 15-hydroxylation is of great quantitative importance in the metabolism of C,, steroids in female rats, as shown by analyses of the steroid composition in bile [1,3], faeces [15] and urine [16] of female rats. This paper presents a detailed study of the substrateand stereo-specificities of hepatic 15-hydroxylation in the rat.

578

8

1 5-Hydroxylation of Corticosteroids 0

i

3 1 $32

8

N

%

m

H. Eriksson

Substrate Specficity All steroid substrates used in this study were converted into 15-hydroxylated derivatives, which accounted for between 20 and 87% of the total metabolites formed. 5 a-Reduced steroids were better substrates for the 15-hydroxylase than the correspond-

579

ing 3-keto-A4 or 5p-reduced compounds (Table 2). The configuration of the hydroxyl group at C-3 did not affect the degree of 15-hydroxylation. 11 pHydroxylated steroids served as better substrates than did the corresponding 11-dehydro-, 11-deoxy or 11a-hydroxy compounds.

580

All 15-hydroxylated compounds identified in bile after the liver perfusions had the chromatographic properties of mono- and disulphates. About 75 - 90 % were monosulphate conjugates. In a recent study it has been shown that the monosulphates of 3 a(and 3 B), 11fi,15,21-tetrahydroxy-5 a-pregnan-20-one in bile from female rats are sulphated exclusively in position 21 151. These findings are of interest in view of a recent report by Gustafsson and Ingelman-Sundberg on the substrate specificity of a steroid sulphate-specific hydroxylase system in female rat liver rnicrosonies 1171. These authors found an absolute requirement of a 21-sulphate group for the 15-hydroxylation of deoxycorticosterone in this system in vitro. Configuration o j the 15-Hydroxyl Group in 3a,llfi,15,21-trtrahydroxy-5a-pregnan-20-one 3 a , l 1 fi,15,21-Tetrahydroxy-5 a-pregnan-20-one was first isolated from pooled faeces from male and female germ-free rats [15]. Later it was found in urine from female rats [16] and as the predominant corticosterone metabolite in bile from female animals [1,3]. In these studies, the compound was tentatively identified as 3 a,l 1B,15 a,21-tetrahydroxy-5 a-pregnan-20-one. In a separate study, incubations of this compound, isolated from faeces of germ-free animals, with caecal contents from conventional rats yielded 15-hydroxylated C,,O, metabolites [I81 which appeared to possess the unusual 14P(H) configuration. This finding suggested that 3 a , l l P,15,21 -tetrahydroxy-5 cr-pregnan-20-one might also have the 14B(H) structure. In order to determine the configuration of the C/D ring junction and the orientation of the 15-hydroxyl group, 10 mg of 3 a , l l fi,15,21-tetrahydroxy-5 a-pregnan-20-one was isolated from bile of female rats. The purified compound was analyzed by gas chromatography-mass spectrometry as its silyl- and oxime silyl ethers, and by 13C nuclear magnetic resonance. The mass spectrum of the silyl ether of 3 a , l l fi,15,21-tetrahydroxy-5 a-pregnan-20-one (Fig. 3 B) showed prominent losses of 103, (90 + 103), (2 x 90 + 103) and ( 3 x 90 + 103) mass units from the molecular ion, while the mass spectrum of 3 a,llfi,15 a,21 -tetrahydroxy-5 apregnan-20-one silyl ether (Fig. 3A) gave strong ions correspondingtolossesof131,(90+131),(2 x90+131) and (3 x 90+131) mass units. Loss of 103 mass units results in these compounds from cleavage of the C,,-C,, bond (loss of the CH,O(Me),Si group), whereas the fragment of mass 131 corresponds to loss of the whole side chain [ 151. Comparison of the mass spectra of the silyl ethers of pairs of 15 a-and 15B,21-dihydroxy20-ketosteroids (Fig. 1 - 3) showed that the 15a-isomerS had a greater tendency to lose the whole side whereas those with a l 5B-group rather lost the - CH2O(Me)3Si fragment. This difference may be partly due to steric effects. Thus, the bulky (pseudo-

15-Hydroxylation of Corticosteroids

/

a *.

Fig. 4. Illus~rutiono/ the stereochenzistry of the C / D ring unrl side chain in the silyl ether ?f 3a,I I~,lS~,2I-tetrahydroxy-5~-pre~nan20-one. (Me),Si, trimethylsilyl

axial) 15fi-silyl group will interact to a greater extent with the side-chain at C-17 than would be the case with the corresponding 15 a-isomer (Fig. 4). On this basis, loss of 103 mass units would be favoured in silyl ethers of 158,21-dihydroxy-20-keto-steroids. Recently it was shown that isomers of 15-hydroxytetrahydrocorticosterone isolated from bile were only acetylated at C-15 to an extent of 60% [5]. Under the same conditions, 15 fi-hydroxyprogesterone behaved similarly towards acetylation, whereas 3a,15 421trihydroxy-5 a-pregnane-1 1,20-dione was completely derivatized at C-15. These results further support a 15B-hydroxy structure for the biliary 3a,llfi,15,21tetrahydroxy-5 a-pregnan-20-one. As mentioned above, microsomal preparations from female rat livers catalyze the 15-hydroxylation of deoxycorticosterone-21 -sulphate [17]. This hydroxylation is stereospecific and takes place in position 15 8. In conclusion, the data obtained in the present study together with those reported by other authors, permit positive identification of the 15-hydroxy-tetrahydrocorticosterone in bile from female rats as 3 a,l 1 fi,ISfi,21-tetrahydroxy-5 a-pregnan-20-one. The formation of 15-hydroxylated metabolites from corticosterone is strictly confined to the female rat [l,31. Induction of 15-hydroxylating enzymes in the liver from adult male animals cannot be achieved by castration or by treatment with estrogens 161. Neonatal castration of male rats, however, leads to a “female” pattern of corticosterone metabolism with excretion of 15-hydroxylated corticosterone metabolites in bile [19]. Furthermore, a 15-hydroxylase active on corticosterone (or corticosterone metabolites) can be induced by estradiol in the regenerating liver from adult castrated male rats [6]. These data indicate a possible influence of androgens and estrogens on the regulation of hepatic 15-hydroxylating enzymes in the rat. I am greatly indebted to Dr Alma Burlingame for the I3C: nuclear magnetic resonance analysis and to Dr Thomas Baillie for valuable discussions concerning mass spectrometric data and linguistic help. The technical assistance of Mrs Ulla Thalin and Mrs Kathleen Brown is gratefully acknowledged. This study was supported by grants from the Swedish Medical Research Council (Project no. 03X-3972), Karolinska Institutets Fonder and the World Health Organization.

H. Eriksson

REFERENCES 1. Cronholm, T., Eriksson, H. & Gustafsson, J.-A. (1971) Eur. J . Biochem. 19, 424-432. 2. Cronholm. T., Eriksson, H. & Gustafsson, J.-A. (1972) Steroids, 19,455-470. 3. Begue, R. J., Gustafsson, J.-A. & Gustafsson, S. (1973) J . Steroid Biochem. 4, 393 - 400. 4. Eriksson, H. & Gustafsson, J.-A. (1971) Eur. J . Biochem. 20, 231 -236. 5. Baillie, T. A. & Eriksson, H. (1975) FEBS Lett. 58, 170- 174. 6. Eriksson, H. (1974) Eur. J. Biochem. 46, 603-611. 7. Eriksson, H. (1975) Acta Physiol. Scand. 93, 206-219. 8. Krebs, H. A. & Henseleit, K. (1932) Hoppe-Seyler’s Z. Physiol. Chem. 210.33 - 66. 9. Adlercreutz, H. (1962) Acia Endocrinol. Suppl. 72, 111 - 139.

H. Eriksson Kemiska Institutionen I, Karolinska Institutet, Solnnvagen 1, S-104 01 Stockholm, Sweden

58 1 10. Burstein, S. & Lieberman, S. (1958) J. B i d . Chem. 233, 331 335. 11. Nystrom, E. & Sjovall, J. (1968) Ark. Kemi, 29, 107- 115. 12. Makita, M. & Wells, W. E. (1963) Anal. Biochem. 5, 523-530. 13. Thenoy, J.-P. & Horning, E. C. (1972) Anal. Lett. 5, 21 -23. 14. Reimendal, R. & Sjovall, J. (1972) Anal. Chem. 44, 21 -29. 15. Gustafsson, J.-A. & Sjovall, J. (1968) Eur. J. Biochern. 6, 236241. 16. Gustafsson, J.-p\. (1970) Eur. J. Biochem. 14, 560-566. 17. Gustafsson, J.-A. & Ingelman-Sundberg, M. (1975) J . B i d . Chem. 250, 3451 -3458. 18. Eriksson, H., Gustafsson, J.-A. & Sjovall, J. (1971) Eur. J. Biochem. 19, 433-441. 19. Begue, R.-J., Gustafsson, J.-A. & Gustafsson, S. (1973) Eur. J. Biochem. 40, 361 -366