THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 263, No. 9, Issue of March 25, pp. 4166-4171, 1988 Printed in U.S.A.

Reciprocal Regulation of Sex-dependent Expression of Testosterone 15a-Hydroxylase (P-45OI5J in Liver and Kidney of Male Mice by Androgen EVIDENCE FOR A SINGLE GENE*

(Received for publication, August 4, 1987) E. James Squires$ and Masahiko Negishig From the Laboratory of Pkrmmobgy, National Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park, North Carolina 27709

Testosterone 15a-hydroxylase activities and its mRNA levels are higher in kidneys than in livers from male 129/J mice. Castration of 129/J male mice re- sulted in repression of P-45Ol5, in kidney, but in- creased it in liver. Two types of cDNA (p16a-29 (Type I) and -15 (Type 11)) encodingP-45OIa, were previously cloned from 129/J female livers (Burkhart, B. A., Ha- rada, N., and Negishi, M. (1985) J. BioZ. Chern. 260, 15357-15361). With the use of p15a-29 as a probe, Type I and I1 P-46015, cDNAs were isolated from li- braries of 129/J kidney poly(A)+ RNA. The nucleotide sequences of the cDNAs showed that Type I and I1 cDNAs from liver and kidney were identical and shared 98.3% similarity. The deduced amino acid se- quence from a full-length Type I cDNA indicated that Type I P-45Ola, consists of 494 amino acids with a molecular weight of 56,594. Nine amino acid substi- tutions were found in the Type I1 clone in 432 amino acids overlapping Type I.

Type I cDNA clones accounted for approximately 90% P-45015, clones isolated from a male kidney li- brary, whereas approximately 90% of cDNA clones in a female kidney library were Type 11. Liver cDNA libraries from males and females contained similar ratios of Type I and 11. Effects of castration on Type I and I1 mRNAs were determined by Southern hybridi- zation of a “P-labeled CZaI-CZuI fragment from p15a- 29 to cDNAs synthesized from kidney and liver poly(A)+ RNAs prepared from sham-operated, cas- trated 129/J mice. The double-stranded cDNAs were digested with CZuI and PstI prior to gel electrophoresis to create the diagnostic restriction fragments specific for Type I or 11. Castration resulted in decreased levels of Type I mRNA in male kidney. In male liver, only Type I mRNA rose significantly in response to castra- tion. Testosterone administration returned the Type I mRNA to normal levels in castrated mice. It therefore appears that the high levels of P-45OIs, in male kidney were due to androgen-dependent induction of Type I mRNA. Both Types I and I1 were repressed in male

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

to the GenBankTM/EMBL Data Bank with accession number($ The nucleotide sequence(s) reported in thispaper has been submitted

503549. $ Recipient of a postdoctoral fellowship from the Natural Sciences

and Engineering Research Council of Canada. Present address: Dept.

tario, Canada NIG 2W1. of Animal and Poultry Science, University of Guelph, Guelph, On-

§ Present address: Pharmacogenetics Section, Laboratory of Re- productive and Development Toxicology, NIEHS/NIH, P. 0. Box 12233, Research Triangle Park, NC 27709.

liver, which results in decreased levels of P-4501~~. Androgen was responsible for the repression and ex- pression of Type I in liver and kidney, but not Type 11.

Many steroid hormones such as androgens and estrogens are metabolized by the membrane-bound monooxygenase sys- tem which is found in microsomes of many tissues, particu- larly liver and kidney. The monooxygenase system consists of NADPH-P-450 reductase, NADH cytochrome 65, P-450, and cytochrome b5. P-450 represents a group of hemoproteins which serve as the terminal oxidase for this system. Depend- ing upon the form of P-450, a variety of steroids are metabo- lized specifically to hydroxylated products by the monooxy- genase system. It appears that sex-specific metabolism of steroids by microsomes is due to sexual dimorphism of P-450 forms specific for these metabolic activities.

P-45015, is a form isolated from 129/J female mouse liver, which has high activity and extreme regio- and stereospecific- ity for the hydroxylation of testosterone at position 15a (1). In liver microsomes, expression of P-45015, is a female char- acteristic with respect to both the activity and mRNA levels (2, 3), whereas the sexual dimorphism in p-45ol5, expression in kidney is reversed, with higher levels found in males than in females (2). Recently, it has been demonstrated that spe- cific repression of P-45015, in livers of male mice is regulated by episodic secretion of growth hormone (4). It is not known, however, whether other hormones such as sex hormones are involved in the regulation of P-45Ol5,.

Androgens are known to dramatically affect gene expres- sion in a wide variety of tissues, in addition to those directly involved with reproduction (5). In kidney, androgens cause an increase in organ size and stimulate the expression of alcohol dehydrogenase, glucuronidase, and ornithine decar- boxylase (6). In liver, androgens are known to regulate en- zymes involved in drug and steroid metabolism and a secretory protein called major urinary protein (7).

In this paper, the experiments were designed to understand how P-45015, expression is differentially regulated in kidney and liver by androgens and whether or not a single gene product of P-45015a is involved in the induction and repression of P-45Ol5, in both tissues. In a previous report from this laboratory (3), two types of cDNA clones for P-45Ol5, were isolated from a library constructed from liver poly(A)+ RNA. The corresponding cDNAs here were isolated from libraries of male and female kidney poly(A)+ RNAs. All of the cDNAs from kidney and liver libraries were sequenced, and the dis- tribution of two types of P-45Ol6, mRNA was determined in kidneys and livers of 12915 mice. These experiments indicate that androgens reciprocally regulate expression of one P- 45015a gene in kidney and liver.

4166

Sexual Dimorphism of P-45015, in Liver and Kidney 4167

EXPERIMENTAL PROCEDURES

Animals-Inbred 129/J mice (5-11 weeks old) were obtained from Jackson Laboratory (Bar Harbor, ME). Mice were castrated or sham- operated a t 8-20 weeks of age under pentobarbital anesthesia (25 mg/kg). After 7 days of recovery, mice were injected either with testosterone propionate (20 mg/kg) once a day for 7-10 days (0.1 ml in corn oil vehicle) or with corn oil alone. Materi~ls-[4-~'C]Testosterone was obtained from Du Pont-New

England Nuclear, and [cY-~'P]~CTP (3000 Ci/mmol), "S-dATP (>lo00 Ci/mmol), and a nick translation kit were obtained from Amersham Corp. Oligo(dT)-cellulose (type 2) was supplied by Collab- orative Research (Lexington, MA). Nitrocellulose (BA85) and modi- fied nylon-66 membrane (NYTRANTM) were purchased from Schleicher & Schuell. Silica Gel GHL plates were obtained from Analtech (Newark, DE), and Sigma supplied testosterone propionate. Restriction enzymes were purchased from New England Biolabs (Beverly, MA). Reverse transcriptase (Maloney leukemia virus) was supplied by Bethesda Research Laboratories. Escherichia coli DNA polymerase I, T 4 DNA ligase, EcoRI methylase, E. coli NAD' ligase, linker d(CCGAATTCGG) used for cDNA cloning, plasmid and M13 vectors, and reagents for dideoxy sequencing were obtained from Pharmacia LKB Biotechnology Inc. Promega Biotec (Madison, WI) supplied the EcoRI-cut Xgtll vector and packaging solution. Boehrin- ger Mannheim supplied the linker d(AATTCCGGAATT) used for creating a cDNA library without EcoRI digestion. The Cyclone Sys- tem was purchased from International Biotechnologies, Inc. (New Haven, CT). All chemicals used were of the highest quality available from suppliers.

Preparation of Microsomes and Assay of P-45OIh-dependent Activ- ity-Mice were killed by cervical dislocation, and livers and kidneys were homogenized in 10 mM Tris-HCI buffer, pH 7.5, containing 150 mM KC1 and 10 mM EDTA using a Polytron homogenizer. Micro- somes were pelleted from the 10,000 X g supernatant by centrifugation at 105,000 X g for 60 min, and pellets were resuspended in 0.25 M sucrose containing 100 mM potassium phosphate, pH 7.5, 1 mM EDTA, and 2 pg/ml leupeptin. Testosterone 15a-hydroxylase activity was measured as previously described (1). P-45OIh-dependent testos- terone 15a-hydroxylase activity in microsomes was determined by inhibition with anti-P-45OI6, antibody raised against the purified enzyme (3). Protein concentrations were determined with Coomassie Blue stain by using serum albumin as a standard (8).

Isolation of Poly(A)+ RNA and cDNA Cloning-Total RNA was extracted from liver with guanidine HCI (9) and from kidney with guanidine isothiocyanate, followed by CsCl centrifugation (10). Poly(A)' RNA was enriched by oligo(dT)-cellulose chromatography (11). T o prepare a cDNA library, poly(A)' RNA was treated with reverse transcriptase to make RNA-DNA hybrids, which were then converted to double-stranded cDNA with DNA polymerase I and RNase H and blunt-ended with T 4 DNA polymerase (12). After methylation of EcoRI sites using EcoRI methylase and S-adenosyl- methionine, EcoRI d(CCGAATTCGG) linkers were ligated to cDNA using T 4 DNA ligase. cDNA was digested with EcoRI and fractionated on Sepharose 4B. The cDNAs were precipitated from pooled eluates from the initial portion of the first peak from a Sepharose 4B column with ethanol. The average size of cDNA was checked by agarose gel electrophoresis to be 1.5-2.0 kilobase pairs. The size-selected cDNAs were ligated into EcoRI-cut Xgtll and packaged (13). One cDNA library was constructed using d(AATTCCGGAATT) linkers, which produces EcoRI "sticky ends" without EcoRI digestion (14); and for this procedure, methylation of the EcoRI sites and EcoRI digestion of the cDNA were omitted. cDNA libraries (-6 X lo6 recombinants) were screened using nick-translated 32P-labeled p15a-29 cDNA, and positive clones were purified by repetitive plating and subcloned into pUC plasmids.

M I 3 Subcloning and Nucleotide Sequencing-Fragments of cDNA were generated by digestion with appropriate restriction enzymes and subcloned into M13 vectors digested to give complementary cohesive ends. Single-stranded recombinant M13 clones were prepared (15) and subjected to dideoxy sequencing analysis using ? W A T P (16). In some cases, deletion mutants of M13 subclones were generated by cycloning (17). Sufficient recombinant M13 subclones were sequenced to generate overlapping sequence data along the entire cDNA clone. In most cases, sequence data were obtained in both directions, and each fragment was sequenced at least twice.

Northern and Southern Blot Analyses-Northern blots of poly(A)+ RNA were prepared after electrophoresis in 1% agarose gels contain- ing 5 mM methyl mercury (18) by blotting to NYTRAN. Blots were

KIDNEY LIVER "

$ 9 8 9 "_.

FIG. 1. Northern hybridization of p15a-29 with poly(A)+ RNA from kidneys and livers of 129/J mice. Total RNA was extracted and enriched for poly(A)+ RNA as described under "Exper- imental Procedures." Five pg of poly(A)+ RNA was denatured by 5 mM methyl mercury hydroxide and electrophoresed on 1% agarose gel containing 5 mM methyl mercury hydroxide and then transferred to diazobenzyloxymethyl paper. The cDNA insert from p15a-29 was nick-translated with ["PIdCTP and hybridized with the diazoben- zyloxymethyl paper as described under "Experimental Procedures." The hybridized paper was exposed to x-ray films.

prehybridized in buffer containing 0.9 M NaCI, 90 mM sodium citrate, 0.1% Ficoll, 0.1% polyvinylpyrrolidone, 0.1% bovine serum albumin, 0.1% sodium dodecyl sulfate, and 0.1 mg/ml sonicated herring sperm DNA and hybridized overnight in the same solution containing nick- translated ["PIcDNA probe (1 X lo7 cpm). The blots were then washed a t 42 "C in solution containing 15 mM NaCI, 1.5 mM sodium citrate, and 0.1% sodium dodecyl sulfate and autoradiographed. Southern blots of DNA fragments were prepared after electrophoresis in agarose gels by transfer to NYTRAN (19). Blots were hybridized to nick-translated [32P]cDNA probes under the conditions described above for Northern blots and autoradiographed. The relative inten- sities of bands were estimated by laser scanning densitometry.

RESULTS

Effect of Castration on P-45OI5, Expression in Kidney and Liver-Previous reports from this laboratory (1, 3) have shown that P-45OI5, is expressed in female liver a t approxi- mately 6-10 times higher levels than in male liver of 129/J mice. On the contrary, in 129/J kidney, this sexual dimorph- ism is reversed P-45OI,,-dependent testosterone 15a-hydrox- ylase activity is approximately 10 times higher in male than in female 129/J mice (2). Fig. 1 shows Northern hybridization of nick-translated 32P-labeled p15a-29 cDNA with poly(A)+ RNA isolated from kidneys and livers of 129/J males and females. An amount of hybridizable P-45015, mRNA was higher in female than in male livers, which was consistent with our previous finding (2, 3); whereas in male kidneys, poly(A)+ RNA contained about 5 times more hybridizable P- 45015, mRNA than in female kidneys, which agreed with the levels of P-45OI5,-dependent activity in kidney microsomes from these mice (2). Liver and kidney microsomes were pre- pared from sham-operated, castrated, and castrated plus tes- tosterone propionate-treated 129/J males. The P-45OI5,-de- pendent testosterone 15a-hydroxylase activities in these mi- crosomes are summarized in Fig. 2, indicating that after castration, levels of P-450'5, in kidney fall to about 40% of that in sham-operated mice, whereas P-45OI5, levels rise about 10-fold in livers. The levels of P-45OI5, in castrated mice were restored to near-normal values by administration of testos- terone propionate. These changes also took place at P-45OI5, mRNA and protein levels not only in 129/J mice, but also in BALB/cJ mice.' The reciprocal regulation of P-45OI5, in liver and kidney by androgens raised a question of whether only one P-45OI5, gene product is involved or different P-45Ol5, gene products are expressed in the two tissues. We have

E. J. Squires and M. Negishi, manuscript in preparation.

4168 Sexual Dimorphism of P-45OI5, in Liver and Kidney

KIDNEY LIVER FIG. 2. Effect of castration and subsequent administration

of testosterone on P-450,,, in kidney and liver. Microsomes were prepared from kidneys and livers of sham-operated, castrated, and castrated plus testosterone propionate-treated 129/J male mice. P-450,~dependent testosterone l5a-hydroxylase activity in micro- somes was determined as the activity inhibited with anti-P-45OI5, antibody as previously described (3). 8, sham-operated male mice injected with corn oil vehicle; , castrated males injected with corn oil; TP testosterone propionate-treated males which have been cas- trated. The activities in liver and kidney microsomes from each group of mice were measured at the same time in order to minimize intrasample variations. The same experiment was repeated with another set of animals, and the values presented were averaged from the two experiments. The variation was -&lo%.

constructed cDNA libraries from kidney poly(A)+ RNA and isolated P-450x5, cDNAs by using p15a-29 as probe. The isolated kidney P-45Ol5, cDNAs and liver counterparts were sequenced for comparison.

Isolation and Sequencing of P-45Ol6, cDNAs from Kidney and Liver-The cDNA libraries were screened for P-45Ox5,- positive clones using p15a-29 cDNA previously isolated from 129/J female liver (3). First, p15a-40 isolated from the female kidney library was found to be a cDNA clone containing more sequence toward the 5‘-end than did p15a-29. Due to the presence of an internal EcoRI site in p-15a-40, the X g t l l was digested with PvuII, which cut at 4 and 350 bp2 from the EcoRI cloning site. The insert was then ligated into the SmaI site of pUC13. This 5”EcoRI fragment of p15a-40 was used as a probe to screen more cDNA libraries, resulting in p15a- 16 from the male kidney library, which turned out to be a full-length cDNA for P-45Ol5,. Restriction maps of P-45015, cDNAs from kidney and liver are shown in Fig. 3. As previ- ously reported for 129/J female liver (3), two types of P-45O1,, cDNA were found in cDNA libraries of 129/J kidney and were categorized as Types I and 11. Type I1 cDNA (representing p15a-15) was distinguished from Type I cDNA (representing p15a-29) by extra diagnostic HindIII and PstI sites in the common 950-bp ClaI-ClaI fragment from these clones. Each of these clones plus p15a-67 (Type I1 from female kidney) and p15a-3 (Type I from male kidney) were sequenced ac- cording to the general strategy outlined in Fig. 3.

The nucleotide sequences of Type I and I1 cDNAs and the deduced amino acid sequences are shown in Fig. 4. Sequence analysis confirmed that Type I cDNAs from kidney and liver are identical and similarly that Type I1 cDNAs from both tissues are identical. The full-length Type I cDNA (p15a-16) has an open reading frame of 148 bp beginning 98 bp from the 5’-end and is 1802 bp in total length. The longest Type

The abbreviations used are: bp, base pairs; SDS, sodium dodecyl sulfate.

5’- E C H,,HB R, H H R C R i- ‘ :pl50-16

I \ I 1 I / I I I ’ : ,,0.2,) T y p e ’

a 1 ~ : ~ : ~ } T Y P ~ I I Hind Ill Pstl 100 bp - - - - - - -

c”--o - - - - - FIG. 3. Restriction maps and sequencing strategy of P-450

cDNAs from kidney and liver libraries. The restriction maps of these clones were constructed by estimating the sizes of restriction fragments by agarose gel electrophoresis and confirmed by nucleotide sequences. B, BamHI; C, ClaI; E, EcoRI; H, HaeIII; R, RsaI. p15a-29 and p15a-15 were previously isolated from a cDNA library of 129/J female liver (3). p15a-16 and p15ty-40 were cloned from kidney poly(A)+ RNAs of male and female 129/J mice, respectively. p15a-29 and p15a-16 represent Type I cDNA, and p15a-15 and p15a-40 are Type I1 cDNA with extra HindIII and PstI sites. The arrows indicate the direction and extent of sequencing of DNA fragments.

I1 cDNA (p15a-40) is missing 245 bp of the 5’-coding region. Type I and I1 cDNAs differed in only 24 base substitutions, making them 98.3% homologous in nucleotide sequence. Two diagnostic restriction sites (PstI and HindIII) in Type I1 cDNA are created by single nucleotide substitution at these sites. Interestingly, two AATAAA polyadenylation signals are tandemly repeated in both cDNAs, which end 19 bp upstream from the beginning of the poly(A)+ tract.

Comparison of Amino Acid Sequence of P-45O1,, to Others- Type I P-45Ol5, protein consisted of 494 amino acids with a calculated molecular weight of 56,594. Type I and I1 P-45OI5, differed in only nine amino acids in the overlapping 432 amino acids, three of which should contribute a higher net negative charge to the Type I1 protein. Comparison of the total amino acid sequence of Type I P-45Ol5, was made with P-450a (20),

(25), and mouse C-P-45ol6, (26). Among those compared, P- 45015, showed the highest similarity to P-450a (69%) and the lowest to P-45Op (24%). Three regions appeared to be well conserved in some of these proteins (Fig. 4). In the first conserved region, from to Leu4’, P-45ol5, had 87% sequence similarity to P-450a, but only 37% similarity to P- 450p. In the second conserved region, originally proposed by Ozols et al. (27), from Met349 to P-450x5, had the highest similarity to P-450b (84%) and then P-450a (73%), but only 26% to P-45Op. In the third region, which contained the cysteine residue for the fifth ligand for heme binding (28), P- 4501,, had 77% similarity to P-450f and 68% to P-450a and only 37% to P-45Op. Of these forms of P-450, P-45Ol5, would be judged closest to P-450a. It should be added that rat P- 450a catalyzes exclusively ’la-hydroxylation of testosterone, but not 15a-hydroxylation (29).

Differential Expression of Type I and II P-4501,, mRNAs in Kidney and Liver-It is of interest t o know whether or not Type I and I1 mRNAs are cooperatively or differentially regulated in kidneys and livers in male and female mice. In order to investigate this, all available P-45Ol5, cDNAs from the libraries were first classified as Type I1 by the presence of a PstI site in the 950-bp ClaI-ClaI fragment or as Type I by the absence of this diagnostic site. Table I summarizes the number of Type I and I1 clones isolated from cDNA libraries from male and female livers and kidneys. The data indicated that approximately 90% of the clones from female kidney were Type I1 cDNA; whereas in male kidney, approximately 90% were Type I cDNA. In a female liver cDNA library, 67% of the clones isolated were Type I1 and 33% were Type I. In a male liver cDNA library, only four clones were examined, and three of these were Type 11. The preferential expressions

P-450f (21), P-450b (22), P-450M1 (23), P -450~ (24), P - 4 5 0 ~

Sexual Dimorphism of P-45OI5, in Liver and Kidney 4169 34

P-4501% : P P C P T P L P F V C N F L Q L P-45Oal I Y

100

P-450f: 87

I 1 I 81

49 (I similarity)

- P-45Ob1 P-45OMl I P-45Oc:

R L L L 75

W C I H V T 68 62

I 1 T I Q

p-450p1 I K F T V N Y 56 V W V L L v 62

~~

C-P-45016. 1

349 367 H P Y T E A V I H E I Q R F A D L I P 100

Q N S I A 73 : D M V I N F V 63

D s v 84 D V Y I v 73

L L F L T F H S S F V 41 E L D H L N T L L Y P I C N 26

N V C I V 73

415 459 : K S D A n m F S 1 C K R Y C F C U ; L I M Q N 100 t NA L T F L D K L T L 68

Y L A A V 0 T L 77 E M T I L I N- F T L 71 Y M A I A A T F T L 71

L EKVIL CL K I TIC L V AILL Q 42 D P Y W L C N P N I H R F L N M K A KVL 37 PE M A R S L A F CLL R 61

FIG. 5. Comparison of conserved regions in amino acid se- quence. Three regions were picked by the sequence homology pro- gram Microgenie, with a criteria of 60% minimum homology. Type I P-45016, is shown with the differences in the sequences of P-450a, -f, -b, "1, -c, and -p and mouse C-P-45016, indicated. The numbers show the amino acid positions of each region of the sequence com- pared.

TABLE I Distribution of Type I and 11 cDNA clones in kidney and liver cDNA

libraries P-45OI6, cDNA clones in Xgtll, which are long enough to contain

the diagnostic region, were digested with PstI and ChI. The digested DNAs were electrophoresed on agarose gel and transferred to NY- TRAN paper. The paper was hybridized with the nick-translated 32P- labeled ChI-ChI fragment from p15a-29. Type I and I1 cDNAs were judged from intensities of 950 and 700 bp, respectively, as described for Fig. 6A.

Total clones Type I Type I1

No. % No. %

Kidney Female 21 3 14 18 86 Male 15 13 87 2 13

Female 21 7 33 14 67 Male 4 1 25 3 75

Liver

of Type I in male kidney and of Type I1 in female kidney and somewhat mixed expression of both types in liver were further confirmed by the following methods.

The basis of the experimental strategy is illustrated in Fig. 6A. When double-stranded cDNA is synthesized from poly(A)+ RNA, digested by ChI and PstI, and hybridized with the 32P-labeled ChI-ChI fragment, only a 950-bp band should be visualized for Type I P-45OI6, cDNA. On the other hand, the same digestion for Type I1 cDNA should generate two bands (750 and 250 bp) which hybridize to the probe. Double- stranded cDNA were synthesized from total poly(A)+ RNA, digested with ChI and PstI, and then electrophoresed on agarose gel. The DNA was transferred to NYTRAN paper and hybridized with the 32P-labeled ChI-ChI fragment. The relative intensities of bands at 950 and 700 bp measured by

beginning of the initiation codon, which is boxed. The termination codon is indicated by triple asterisks. The three conserved regions are underlined with broken lines, and the diagnostic PstI and Hind111 sites are shown by wavy lines. The tandemly arranged poly(A) signals are also marked by solid underlines. The beginning of Type I1 cDNA (p15a-40) is shown by the arrowhead.

4170 Sexual Dimorphism of P-4501s, in Liver and Kidney

1 digest with Pst I and Cla I

c c c c "U "- C C P P c c

t agarose gel electrophoresis

950 bp 700 bp

250 bp + blot and probe ~ i t h ~ ~ P - l a b e l e d Cla I-Cla I fragment

9 5 O b p I - I

250 bp "I/ autoradiography

B

950 bp

700 bp

250 bp L FIG. 6. Strategy used to measure relative amounts of Type

I and I1 cDNAs. A, the outlined scheme was used to classify P- 45OIs, cDNA as either Type I or I1 and to estimate relative amounts of both types in total poly(A)+ RNA from tissues. cDNA clones were digested with PstI and ClaI, and then the digest was separated by agarose gel electrophoresis and blotted on NYTRAN paper as de- scribed under "Experimental Procedures." The blotted paper was prehybridized and hybridized with the nick-translated '*P-labeled ClaI-ClaI fragment. Type I cDNA shows a single band a t 950 bp, whereas Type I1 shows two bands a t 700 and 250 bp due to the presence of extra PstI sites in the ClaI-ClaI fragment. R, the method outlined above was tested to estimate ratios of Type I and I1 mRNA in total poly(A)+ RNA from 129/J female liver. Double-stranded cDNAs were synthesized from 5 pg of poly(A)' RNA as described under "Experimental Procedures." A part (-5 pg) of the obtained cDNA was digested with Pstl and ClaI and electrophoresed on agarose gel. The cDNA was then transferred to NYTRAN paper and hybrid- ized with the radioactive ClaI-ClaI fragment as outlined above. The relative intensities of hybridization to the bands at 950 and 700 bp were representative of the relative amounts of both types in the total poly(A)+ RNA and were 39 and 60% of the total, respectively.

aser scanning densitometry were taken as a measure of the relative amounts of Type I and I1 mRNAs in total poly(A)' RNA. Fib. 6B is a demonstration of an experiment with poly(A)+ RNA from female liver, where the P-45015, mRNA level is the highest. The relative intensities of the 950- and 700-bp bands were 39 and 60%, respectively. This agreed perfectly with the ratio of Type I and I1 cDNAs estimated from the number of Type I and I1 clones from female liver (Table I), indicating the accuracy of this method. It should be noted that these restriction analysis alone would suggest, but not prove, that all cDNAs categorized to either type are identical a t nucleotide base levels. However, since the identi- cal nucleotide sequences were obtained from a t least three independent cDNAs for each type, it was thought to be reasonable that the majority of cDNAs, if not all, in each type are identical. By use of this method, therefore, the relative distributions of Type I and I1 P-45OI5, proteins in male kidney and liver and the effects of androgens were measured.

Effect of Androgen on Type Z and ZZ P-45OI5, mRNA-The results are shown in Fig. 7. In sham-operated animals, Type I predominated in kidney; whereas in liver, both Types I and I1 accounted for 70 and 30% of the mRNA, respectively. Following castration, Type I mRNA levels fell in kidney, which coincided with a decrease of total p-45ol5, mRNA. Type I1 mRNA levels were so low that the charges of Type I1 would not effect an overall regulation of kidney P-45015, in males. The already repressed levels of Type I1 mRNA in kidney were slightly decreased in castrated mice. Thus, an- drogen was not a repressor of Type I1 P-45015, in kidney. However, in liver, castration caused Type I mRNA levels to rise, to levels a t least 10 times higher than those in sham- operated male mice; whereas the increase in levels of Type I1 was not significant and probably was in the range of experi- mental deviation. Administration of testosterone propionate returned the Type I mRNA levels to near normal in castrated mice. The differences in intensities between the four samples (female, sham-operated male, castrated male, and androgen- treated male) were reasonable reflections of relative amounts of P-45OI5, mRNA in each group. It therefore appears that the induction and repression of P-45OI5, mRNA are due to androgen-dependent modulation of Type I P-45Ol5, in kidneys and livers of 129/J mice.

FIG. 7. Effect of castration and testosterone administration on levels of Type I and I1 P-4501., mRNAs in kidney and liver. Double-stranded cDNAs were prepared from total poly(A)' RNAs isolated from kidneys and livers of 129/J male mice as de- scribed under "Experimental Procedures." These were digested with PstI and ClaI, and Type I and I1 cDNA levels were determined by the method outlined and tested in Fig. 6, ( A and B). Mice used were sham-operated and injected with corn oil vehicle (cL,,,), castrated and injected with corn oil ( 3). or castrated and injected with 20 mg/ kg testosterone propionate in corn oil ( 8Tp).

Sexual Dimorphism of P-45015, in Liver and Kidney 4171

DISCUSSION

In this paper, we have presented nucleotide sequences for two types of cDNA encoding P-45OI5, from kidney and liver libraries of male and female 129/J mice. These two types, designated Types I and 11, were approximately 98.3% homol- ogous in nucleotide and deduced amino acid sequences. The sequence of cDNAs for kidney and liver P-45015, proteins were identical, indicating that the same genes were expressed in both kidney and liver. Type I (p15a-16) cDNA was full- length and coded for 494 amino acids with a calculated mo- lecular weight of 56,500. In a previous report from this labo- ratory describing the isolation of P-45OI5, (3), the molecular weight was estimated as 48,000 by SDS-polyacrylamide gel electrophoresis. This discrepancy is not surprising and has occurred for P-450. For instance, mouse C-P-45OI5, had a molecular weight of approximately 56,000 deduced from the nucleotide sequence of cDNA and 49,500 determined by from SDS-polyacrylamide gel electrophoresis (7, 26). A large dis- crepancy between the calculated molecular weights and the values obtained from SDS-polyacrylamide gels was also re- ported for rabbit P-4501 (31, 32).

P-45OI5, has a high degree of sequence similarity to rat P- 450a. In comparison with the three conserved areas among the P-450s, P-45OI5, is more closely related to P-450a, -b, and -f than to P-45Op. This is particularly apparent for the second conserved region examined in Fig. 3 between amino acids 349 and 367, and it is of interest that studies on the expression of chimeric forms of P-45Oc and P-450d (33) have suggested that this area may be important in substrate binding.

Based on the nucleotide sequences of the cDNAs, we have shown that identical forms of P-45OI5, are expressed in kid- neys and livers from l29/J mice and encoded by two highly (98.5%) homologous mRNA species. Burkhart et al. (34) have previously found that only two genes exist for P-45015, (one for Type I and one for Type II), and the two genes are organized tandemly within about 35 kilobase pairs of genomic DNA on mouse chromosome 7. From the presence of only two genes for P-45OI5, and the identical nucleotide sequences of cDNAs from liver and kidney, it can be concluded that P- 45OI5, in both tissues are the same gene product.

Although the two genes are highly homologous, expressions of these genes in kidneys and livers of male and female mice are differentially regulated. Type I P-45OI5, mRNA, which is expressed in female liver a t levels as high as Type 11, was repressed in male liver to the same levels as Type 11. The repression of Type I, but not Type 11, P-45OI5, in male liver was shown to be androgen-dependent; in addition, Type I was expressed in male kidney and was also shown to be androgen- dependent. Type I1 P-45Ol5, was female-specific in liver, and its repression in male liver was independent of androgen in adult 129/J mice. It should therefore be concluded that a high P-45OI5, level in male kidney was due to high levels of Type I and that, in female liver, the high P-45Ol5, level results from high levels of both Types I and 11.

A mechanism by which expression and repression of a single P-45015, gene (Type I) are regulated by androgen in two different tissue remains to be determined. In a recent study, Saltzman et al. (30) have shown that androgens can enhance the expression of a 29-kDa protein in liver while repressing this protein in ventral prostates of rats; however, a lack of nucleotide sequences for 29-kDa cDNAs from liver and pros- tate makes it impossible to determine whether only one 29- kDa gene is involved in this regulation.

Type I1 p-45ol5, mRNA is expressed predominantly in female liver. Type I1 in male livers are expressed a t at least a 10-fold lower level than in female liver. In kidney, the expres- sion of Type I1 is also a female characteristic, although the

degree of expression is lower than in female liver. In a previous report (2), we demonstrated that P-45OI5,-dependent testos- terone l5a-hydroxylase activity in kidney microsomes is higher in 129/J than in BALB/cJ female mice and that this higher activity in 129/J is inherited as an autosomal dominant trait. Rsh was chosen to represent this locus. It is therefore likely that the Rsh locus is a transacting genetic element for expression of the Type I1 P-45OI5, gene in female kidney.

In summary, we would propose that the expression of P- 45015, in adult mice is a particularly good system in which to study sex- and tissue-dependent gene regulation for the fol- lowing reasons. 1) The expression of the Type I gene is a female characteristic in livers, but is a male characteristic in kidneys of 129/J mice. Androgens are major inducers of the Type I gene in kidneys, but are repressors in livers of 129/J males. 2) The Type I1 gene, which produces a gene product with 98% sequence similarity to Type I, is female-specific in both tissues, but its repression is not primarily regulated by androgens in male mice. Furthermore, the P-45Ol5, system has the advantage of having a small number of genes (most likely, two). Finally, this system provides an opportunity to study how growth hormone and androgens repress P-45OI5, in livers of male mice.

Acknowledgments-We wish to thank Barbara Burkhart for her useful discussion throughout this research and Debbie Garner for typing this manuscript.

REFERENCES

1. Harada, N., and Negishi, M. (1984) J. Biol. Chem. 259 , 1265-1271

3. Burkhart, B. A., Harada, N., and Negishi, M. (1985) J. Biol. Chem. 260 , 2. Squires, E. J., and Negishi, M. (1986) Biochemistry 25,4913-4918

4. Noshiro, M., and Negishi, M. (1986) J. Biol. Chem. 2 6 1 , 15923-15927 5. Bardin, C. W., and Catterall, J. F. (1981) Science 211 , 1285-1294 6. Roy, A. K., and Chatterjee, B. (1983) Annu. Rea Physiol. 45,37-50 7. Harada, N., and Negwhl, M. (1984) J. Biol. Chem. 259 , 12285-12290 8. Bradford, M. M. (1976) Anal. Biochem. 72 , 248-254 9. Cox, R. A. (1968) Methods Enzymol. 12B, 120-129

15357-15361

10. Chirgwin, J. M., Przyhyla, A. E., MacDonald, R. J.. and Rutten. W. J.

11. Avlv, A., and Leder, P. (1972) Proc. Natl. Acad. Sci. U. S. A. 6 9 , 1408-

12. Watson, C. J., and Jackson, J. F. (1985) in DNA Cloning A Practical

13. Young, R. A., and Davis, R. W. (1983) Proc. Natl. Acad. Sci. U. S. A. 80,

14. Seth. A. (1984) Gene Anal. Tech. 1. 99-103

(J979) Biochemistry 18,5294-5299

1412

Approach (Glover, D. M., ed) Vol. 1, pp. 79-88, IRL Press Ltd., Oxford

1194-1198

15. Messing,'J. (1983) Methods Enzymol. 101: 20-78 16. Biggin, M. D., Gibson, T. J., and Hong, G. F. (1983) Proc. Natl. Acad. Sci.

17. Dale, R. M. K., McCure, B. A., and Honchins, J. P. (1985) Plosmid 13.31- U. S. A. 80,3963-3965

40 18. Bailey, J. M., and Davidson, N. (1976) Anal. Biochem. 7 0 , 75-85 19. Southern, E. M. (1975) J. Mol. Biol. 9 8 , 503-517 20. Nagata, K., Matsunaga, T., Gillette, J., Gelboin, H. V., and Gonzalez, F. J.

21. Gonzalez, F. J., Kimura, S., Song, B.-J., Pastewhea, J., Gelboin, H., and (1987) J. Biol. Chem. 262 , 2787-2793

22. Suwa, Y., Mizukami, Y., Sogawa, K., and Fujii-Kuriyama, Y. (1985) J. Bid. Hardwick, J. P. (1986) J. Biol. Chem. 261 , 10667-10672

23. Yoshioka, A., Morohashi, K., Sogawa, K., Miyata, Z., Kawajiri, K., Hiroseo, Chem. 260 , 7980-7984

T., Inayama, S., Fuji-Kuriyama, Y., and Omura, T. (1987) J. Biol. Chem. 262,1706-1711

24. Yabusaki, T., Shimezir, M., Murakami, H., Nakamura, K., Oeda, K., and Ohkawa, H. (1984) Nucleic Acids Res. 12,2929-2938

25. Gonzalez, F. J., Nebert, D. W., Hardwick, J. P., and Kasper, C. B. (1985) J. Biol. Chem. 260 , 7435-7441

26. Wong, G., Kawajiri, K., and Negishi, M. (1987) Biochemistry 2 6 , 8683- 8690

27. Ozols, J., Heinemann, F. S., and Johnson, E. F. (1981) J. Biol. Chem. 256 , 11405-1 140R

28. Fuji-Kariyama, Y . , Mizakami, Y., Kawajiri, K., Sogawa, K., and Mura-

29. Wood, A. W., Ryan, D. E., Thomas, P. E., and Levin, W. (1983) J. Biol.

30. Saltzman, A. G., Hiipakka, R. A,, Chang, C., and Liao, S. (1987) J. Biol.

31. Dieter, H. H., Muller-Eherhard, U., and Johnson, E. F. (1982) Biochem.

32. Tukey, H. R., Okino, S., Barnes, H., Griffin, K. J., and Johnson, E. F.

33. Sakaki, T., Shibata, M., Yakusaki, Y., and Ohkawa, H. (1987) DNA (N .

"~.. "."

matsu, M. (1982) Proc. Natl. Acad. Sci. U. S. A. 79 , 2793-2797

Chem. 258,8839-8847

Chem. 262,432-437

Biophys. Res. Commun. 106,515-520

(1985) J. Biol. Chem. 260, 13347-13354

Y. 1 6 . 3 1-39 34. Burkhak. B., Skow, L., Lalley, P., and Negishi, M. (1986) Fed, Proc. 4 5 ,

1853