Alternative splicing events in the coding region of the cytochrome ...

305

Alternative splicing events in the coding region of the cytochrome P450 aromatase gene in male rat germ cells J Levallet, H Mittre1, B Delarue1 and S Carreau Laboratoire de Biochimie-IRBA, 1Laboratoire de Biochime, CHRU Cle´menceau, CNRS EP 9, Universite´, Esplanade de la Paix, 14032 Caen-Ce´dex, France (Requests for offprints should be addressed to S Carreau)

ABSTRACT

Expression of cytochrome P450 mRNA in rat germ cells was characterized by reverse transcription PCR with various primers located at the 3 -end of the coding region. At least two unusual isoforms (Ex10-S and INT) of P450 aromatase (P450arom) mRNA were expressed. Analysis of their sequences demonstrated that an alternative splicing event occurred first at the exon–intron boundary of the GT consensus sequence of the last coding exon, and second in the internal 5 donor inside exon 9 used as a minor cryptic splicing site. These isoforms lacked the last coding exon which contained the heme-binding domain; in addition, for the Ex10-S transcript, the catalytic domain was also absent

because of a frameshift in the open reading frame. The deduced amino acid sequences led to truncated P450arom polypeptides without the heme-binding domain, which were probably unable to convert androgens into estrogens. Adult rat germ cells are able to express P450arom mRNA, which is then translated into a biologically active enzyme which is involved in estrogen production. Moreover, for the first time, we report the existence of alternative splicing events of P450arom mRNA in pachytene spermatocytes and round spermatids, which probably cannot encode functional aromatase molecules.

INTRODUCTION

multiple transcriptional start sites and alternative first exons have also been demonstrated for tissuespecific expression of the bovine aromatase gene (Fu¨rbass et al. 1997). Corbin et al. (1995) described a functional ovarian and placental isoform of the aromatase cDNA, whereas, in contrast, Fu¨rbass & Vanselow (1995) demonstrated that the bovine genome contains a non-functional copy (pseudogene) of bCYP19. In the rat, alternative splicing in the coding region of P450arom mRNA is responsible for three different sizes of mRNA in the rat ovary or R2C cells (rat Leydig tumor cells): one transcript contains exon 10, and the two others represent smaller P450arom mRNA transcripts with an intronic sequence instead of the last coding exon (which contains the heme-binding domain) and thus cannot encode a functional enzyme (Lephart et al. 1990). The precise localization of P450arom within the testicular cells has been the subject of much debate: in the rat there is an age-related change in the cellular localization of the aromatization site, being mainly in Sertoli cells in immature animals but located in Leydig cells in adults (Papadopoulos et al. 1986, Levallet &

The synthesis of estrogens (C18) from androgens (C19) is catalyzed by three successive hydroxylations. The microsomal enzymatic complex involved in this transformation is named aromatase and is composed of two proteins, a specific heme glycoprotein, cytochrome P450 aromatase (P450arom), and a ubiquitous non-specific flavoprotein, NADPH-cytochrome P450 reductase (Simpson et al. 1994). Human P450arom is the most extensively characterized complex at both the protein and nucleic acid levels; two mRNAs arising from alternative use of two polyadenylation sites appear to be functional (Corbin et al. 1988). Analysis of transcript and genomic sequences indicates that the tissuespecific expression of P450arom in placenta, gonads, brain, liver and adipose tissue is regulated in part by the use of an alternatively spliced untranslated first exon (Means et al. 1991, Simpson et al. 1994, 1997, Agarwal et al. 1997). However, the protein encoded by these transcripts is always the same, as there is only one human P450arom enzyme encoded by a single-copy gene (Simpson et al. 1997). Recently, Journal of Molecular Endocrinology (1998) 20, 305–312 0952–5041/98/020–305 $08.00/0

Journal of Molecular Endocrinology (1998) 20, 305–312

 1998 Society for Endocrinology Printed in Great Britain

306

  and others

· P450 aromatase transcripts in male rat germ cells

Carreau 1997). Recently, in the adult rat, the P450arom (associated with a biologically active protein) was immunolocalized not only in Leydig cells but also in pachytene spermatocytes, round spermatids and spermatozoa (Levallet et al. 1998). In the present work, it was our aim to investigate by reverse transcription PCR (RT-PCR) the possibility of purified adult rat germ cells expressing aromatase mRNA transcripts and to characterize the 3 -end of these transcript(s). MATERIALS AND METHODS

Purification of germ cells Germ cells were prepared from 80-day-old Sprague–Dawley rat testes by the trypsin–DNase method (Meistrich et al. 1981). Briefly, the testes were decapsulated, after mechanical disruption in PBS (pH 7·2), and, after elimination of the interstitial cells by four successive decantations, the seminiferous tubules were submitted to enzymatic digestion (trypsin 0·025% and DNase 0·1% in PBS) for 45 min at 32 C. The dispersed cells were washed in PBS supplemented with 1·5 mM glucose and 1·2 mM pyruvate, and then the aggregated Sertoli cells and spermatozoa were removed by filtration. The separation of germ cells was then accomplished by unit gravity sedimentation on a 0·2–2·75% BSA gradient (Bellve´ et al. 1977). The fractions containing pachytene spermatocytes and round spermatids were identified under a phase contrast microscope, and the highest homogeneous fractions were pooled and washed. The purity of each germ cell fraction was first evaluated by morphological characterization; it was shown that the major contaminant (2–5%) in the cell suspensions was other germ cells but no Leydig cells, visible after staining for the specific 3âhydroxysteroid dehydrogenase (3âHSD). The cells were counted and used immediately for RNA isolation. Purification of Leydig and Sertoli cells The testes of mature rats were coarsely minced and then subjected to enzymatic treatment with collagenase/dispase (0·05%), soybean trypsin inhibitor (0·005%) and DNase (0·001%) in Ham’s F-12/Dulbecco’s modified Eagle’s medium (1:1, v/v) for 10 min at 32 C. The Leydig cells were purified on discontinuous Percoll gradients (Papadopoulos et al. 1985), and then fractions assumed to contain Leydig cells were incubated for 60 min at 32 C under air/CO2 in a phosphate buffer solution containing 1·5 mg nitroblue tetrazolium, Journal of Molecular Endocrinology (1998) 20, 305–312

6 mg NAD, 1 mg niacinamide and 10 µl dehydroepiandrosterone (10 mg/ml); coloured cells were identified as Leydig cells (3â-HSD positive). Only the most enriched fraction containing more than 98% of 3â-HSD-positive cells was used. After an additional enzymatic treatment with the above cocktail and a third enzymatic digestion in the presence of hyaluronidase (0·01%), soybean trypsin inhibitor (0·005%) and DNase type I (0·001%) (Tung et al. 1984, Foucault et al. 1994), the cells were filtered, washed and plated for 48 h under an air/CO2 (5%) atmosphere at 37 C in Ham’s F12/ Dulbecco’s modified Eagle’s medium supplemented with rat serum (5%) in order to attach the Sertoli cells. On day 2, the culture medium containing germ cells was removed and replaced with a fresh one without serum. On day 4 the cells were submitted to hypotonic treatment to remove the remaining germ cells, and the culture dish was washed with fresh medium and then incubated for 24 h. RT-PCR studies Total RNA from freshly isolated cells was prepared using a modification of the guanidinium isothiocyanate method (RNAXEL; Eurobio, Les Ulis, France). mRNA present in the total RNA was reverse-transcribed into cDNA as follows: 30 min at 37 C with 200 IU Moloney murine leukemia virus reverse transcriptase (Life Technology, Eragny, France), 500 µM dNTP, 0·2 µg oligodT (12–18mer) and 24 IU RNasin. The cDNAs obtained were further amplified by PCR using selected oligonucleotides. PCR was performed using 35 cycles (30 s at 94 C, 30 s at 60 C and 1 min with 2 s delay for each subsequent cycle at 72 C) in the presence of 10 nmol dNTPs, 50 pmol primer and 1·5 IU Taq polymerase in a final volume of 50 µl (Eurobio). To selectively detect transcripts with different 3 -end coding regions, specific PCR primer sets (Table 1) were used (Eurobio). Sequencing of P450arom RT-PCR transcripts The RT-PCR products were extracted from agarose gel by the Sephaglas Band Prep Kit (Pharmacia Biotech, Orsay, France). The DNA was then amplified by a second PCR round for an additional 30 cycles with the same conditions as described above using the same set of primers. The new PCR products were purified by agarose gel electrophoresis and sequenced using the fluorescent dideoxychain terminating method (Genome Express, Grenoble, France).

P450 aromatase transcripts in male rat germ cells ·

  and others

 1. Primers used for RT-PCR analysis

Primer Aro-Ex8 Aro-Ex9 Aro-Ex10 Aro-Int Aro-Ex9Tr

Sequence

Orientation

GCTTCTCATCGCAGAGTATCCGG CAAGGGTAAATTCATTGGGCTTGG GGAATCGTTTCAAAAGTGTAACCAG AAGTCTTTGCCAAATTAAGGACGC GGCTGAAAATACCTGTAGGGAACTCG

Sense (1555–1577)* Antisense (1821–1844)* Antisense (1949–1973)* Antisense† Antisense‡

Sense and localization of each primer used have been indicated in the figures. *Numbered as described by Hickey et al. (1990) (GenBank, accession no. M33986). †Located in intronic sequence described by Lephart et al. (1990). ‡Degenerated primer.

RESULTS AND DISCUSSION

Expression of P450arom by germ cells has been demonstrated in the testis of several species. RT-PCR showed P450arom mRNA in pachytene spermatocytes and round spermatids of the rat (Janulis et al. 1996a, Levallet et al. 1998); furthermore, a protein of 55 kDa, able to transform androgens into estrogens, has been immunolocalized in rat and mouse germ cells (Nitta et al. 1993, Janulis et al. 1996b, Levallet et al. 1998). Besides rodents, P450arom has also been reported in germ cells of rooster (Kwon et al. 1995) and brown bear (Tsubota et al. 1993). In this work, we used specific primers to show P450arom expression in adult rat Leydig cells, Sertoli cells and germ cells (pachytene spermatocytes and round spermatids). The primer (Aro-Ex8) and the antisense primer complementary to the exon 10 sequence (Aro-Ex10) were used (Fig. 1A). The expected 418 bp signal (Ex10-L) obtained with this set of primers was detected in somatic cells (Leydig and Sertoli cells) as well as in pachytene spermatocytes and round spermatids. Also, another band of smaller size (Ex10-S) was seen mainly in pachytene spermatocytes and at a lower level in round spermatids (Fig. 1B). In 1990, Lephart et al. demonstrated differential processing of the aromatase mRNA in rat ovary and R2C cells, and two mRNA transcripts without the hemebinding domain (characteristic of all species of cytochrome P450) were described. The nucleotide sequence analysis suggests the occurrence of an alternative splicing event during the processing of the aromatase mRNA molecule. These transcripts contained an unspliced intron instead of the last coding exon (Lephart et al. 1990). To prove the potential of rat testicular cells to express such transcripts, we have used other antisense primers (Aro-Int) located in the intronic sequence described by Lephart et al. (1990) (Fig. 1A). With these primers, a 567 bp amplified product (INT) was

visible in both germ cells and Leydig cells. In order to sequence the spermatocyte transcripts, each amplified product (INT, Ex10-L and Ex10-S) was excised from the agarose gel, purified and submitted to another PCR with the same set of primers (Fig. 1C). The nucleotide sequence of each PCR product was compared with the corresponding region (Fig. 2) of the ovary sequence (Hickey et al. 1990). The largest product (Ex10-L) obtained with the Aro-Ex10 primer shows 100% homology with the corresponding fragment of the rat ovary, whereas Ex10-S has the same sequence until nucleotide 1813 in exon 9, the remaining 87 bp showing only 18% homology. However, this non-homologous 87 bp sequence could be perfectly matched to the 5 -end of exon 10. The Ex10-S amplified product was the result of a deletion of the last 49 bp of the 3 -end of exon 9 and had a predictive size of 369 bp. It is noteworthy that the position at which the homology ceased is centered on a GT dinucleotide and thus suggests that dinucleotide 1813–1814 is a site of an intron–exon boundary. In the human P450arom gene, the consensus sequence for RNA splicing is -gta- (Toda & Shizuta 1993). Denhez & Lafyatis (1994) described in hSWAP mRNA (human vertebrate homologs to a Drosophila splicing regulator) an internal 5 donor inside an exonic sequence. The first six nucleotides of this splicing donor have the same hexameric sequence (GTATTT) as that of aromatase exon 9 internal donor. Consequently, it is tempting to speculate that this sequence, which is also present in P450arom at position 1813–1818, could be used as a minor cryptic splicing site. The 567 bp product amplified by Aro-Ex8 and Aro-Int showed 100% homolgy with the ovary sequence until exon 10 (Hickey et al. 1990), then the sequence was mismatched but had more than 98% homology with the intronic region in place of the last coding exon described by Lephart et al. (1990). To distinguish the exon 9 truncated isoform from the full-length exon 9 product, a new antisense Journal of Molecular Endocrinology (1998) 20, 305–312

307

308

  and others

· P450 aromatase transcripts in male rat germ cells

 1. Amplification of various P450arom mRNAs from rat germ cells. A: schematic representation of 3 -end region of P450arom. Position of the various primers and functional domains: I, the helical region; II, the aromatic region; III, the heme-binding region (the nucleotide positions are numbered as decribed by Hickey et al. (1990)). The intronic sequence instead of exon 10 described by Lephart is also represented. B: cDNA product obtained after RT-PCR with the Aro-Ex8 and Aro-Ex10 primer set with 250 ng of total RNA from purified rat Leydig cells (LC), Sertoli cells (SC), pachytene spermatocytes (PS) and round spermatids (RS). C: electrophoresis of cDNA after purification and reamplification: lane 1, 418 bp transcript amplified with Aro-Ex10; lane 2, additional shorter transcript obtained with Aro-Ex10 (369 bp); lane 3, 567 bp transcript amplified with Aro-Int primer. All the amplified products were separated on 4% agarose gel stained with ethidium bromide.

primer was synthesized. Aro-Ex9Tr is a degenerated primer consisting of the three nucleotides before the 49 bp lacking region and the 23 nucleotides at the beginning of exon 10 (Fig. 3A). The reverse transcription and co-amplification of Leydig cells, Sertoli cells, pachytene spermatocytes and round spermatid RNAs with Aro-Ex8 and the two antisense primers Aro-Ex10 and Aro-Ex9Tr are shown in Fig. 3B. Three amplified products were produced in germ cells, but mainly in pachytene spermatocytes, two at the expected size 418 bp (Ex10-L) and 369 bp (Ex10-S) (corresponding to the Aro-Ex10 antisense primers), and a 280 bp product resulting from the lack of the last 49 nucleotides of exon 9. In the Leydig cells, the 418 bp was amplified well but the two smaller transcripts are quite undetectable. Moreover, in Sertoli cells only the 418 bp transcript was visible. To confirm the expression of a transcript containing the aromatic region, an antisense primer (Aro-Ex9) located in this sequence was used with Aro-Ex8 (Fig. 3A). A 289 bp product that contained Journal of Molecular Endocrinology (1998) 20, 305–312

the aromatic region was amplified well in somatic cells as well as in pachytene spermatocytes and round spermatids (Fig. 3C). These amplified products have 100% homology with Hickey’s ovary sequence. The deduced amino acid sequences of Ex10-S and INT were compared with the corresponding ovary sequences. For Ex10-S, the 49 bp deletion leads to a switch of the reading frame and to the appearance of a stop codon; in addition, a stop codon in the INT transcript appears five amino acids after the end of exon 9. These two transcripts encode a putative protein that lacks the hemebinding domain; the predictive sequences of Ex10-S and INT contained 428 and 426 amino acids respectively (Fig. 4). Moreover, Ex10-S had lost the region (nucleotides 1817–1852) involved in the interaction between rings C and D of the steroid substrate (Chen & Zhou 1992) and named the aromatic region by Hickey et al. (1990). Recently two cDNA clones of P450arom were isolated from a medaka (Oryzia latipes) ovarian

P450 aromatase transcripts in male rat germ cells ·   and others

 2. Aligned nucleotide sequences of the P450arom cDNAs amplified after RT-PCR with rat pachytene spermatocyte RNA. The amplified product (418 bp) with Aro-Ex10 primer is labeled Ex10-L (long) and the additional transcript (369 bp) named Ex10-S. The cDNA obtained with Aro-Int (567 bp) is labeled INT. These sequences were compared with the rat ovary sequence (Hickey et al. 1990). The primer sequences used for PCR are shaded, positions of identity are shown as asterisks, and gaps, introduced when the respective sequences are aligned, noted as dashes. Functional domains described by Hickey et al. (1990) are as follows: the helical region (broken line), the aromatic region (underlined) and heme-binding region (thick underlined).

follicle cDNA. One of them has an additional A residue at the 3 -end of exon 9, which causes a frameshift in the open reading frame, resulting in a truncated P450arom polypeptide without the hemebinding region which is unable to convert androgen into estrogen (Fukada et al. 1996). Furthermore, in various rabbit tissues, an aromatase cDNA clone

without the heme-binding region is expressed at a level similar to that of the full-length transcript (Delarue et al. 1998). Alternative splicing events in the coding sequences of aromatase mRNA are obvious in various species (rodent, porcine, bovine, fish) and tissues (testis, ovary, placenta), but the mechanisms involved are not yet clarified (Vanselow Journal of Molecular Endocrinology (1998) 20, 305–312

309

310

  and others

· P450 aromatase transcripts in male rat germ cells

 3. Amplification of the wild-type and truncated exon 9 from the germ cell P450arom transcripts. A: schematic representation of 3 -end region of P450arom and localisation of various primers. B: coamplification of 250 ng total RNA from purified Leydig cells (LC), Sertoli cells (SC), pachytene spermatocytes (PS) and round spermatids (RS) with Aro-Ex8 and the two antisense primers Aro-Ex10 and Aro-Ex9Tr. C: cDNA product obtained after amplification of 250 ng total RNA from each purified cells with Aro-Ex8 and Aro-Ex9. All the amplified products were electrophoresed on 2% agarose gel stained with ethidium bromide.

& Fu¨rbass 1995, Fukada et al. 1996, Choi et al. 1997, Fu¨rbass et al. 1997). Expression of CYP19 isoforms is also developmentally regulated; the two smaller transcripts containing an unspliced intron in place of the last coding exon described in the ovary were not detected in perinatal rat brain, whereas the large functional transcript was expressed (Lephart et al. 1992). The two unusual transcripts (Ex10-S and INT) were amplified in adult rat Leydig cells as well as in ovary (data not shown), but at a lower intensity than in rat germ cells. Expression of exon 9 truncated transcripts (ratio 280 bp/418 bp; Fig. 3B), evaluated after densitometric analysis, represents 35% of the total amplified products in pachytene spermatocytes, 15% in round spermatids but only 4% in Leydig cells. The translation of germ cell truncated transcripts needs to be established as well as the control of their expression. Furthermore, the involvement of these putative inactive aromatase molecules in the regulation of estrogen production in the testis is Journal of Molecular Endocrinology (1998) 20, 305–312

completely unknown. The exact role and mechanism of action of estrogens in the reproductive organs of the male remain to be clarified, as does the regulation of aromatase gene expression, especially in germ cells during rat testicular development (Levallet et al. 1998). There is evidence to suggest first that estradiol improves rat gonocyte development (Li et al. 1997) and second that estrogens are required for complete maturation of spermatozoa and for the fertility of male mice (Lubahn et al. 1993, Eddy et al. 1996). In conclusion, together with Leydig cells, rat germ cells are able to synthesize P450arom mRNA, which is translated in a biologically active enzyme involved in estrogen production. For the first time we demonstrate alternative splicing events in pachytene spermatocyte and round spermatid P450arom mRNAs leading to the presence of a putative inactive protein. Alternative splicing events may be another necessary step, like paracrine and endocrine factors for Leydig and Sertoli cells, in the

P450 aromatase transcripts in male rat germ cells ·

  and others

 4. Predictive nucleotide sequence of Ex10-S and INT transcripts compared with the rat ovary sequence (Hickey et al. 1990). Positions of identity are shown as asterisks. Functional domains: the helical region (broken line), the aromatic region (underlined) and heme-binding region (double underlined).

control of aromatase expression (and consequently the modulation of estrogen production) in germ cells of rat testis. ACKNOWLEDGEMENTS

The authors wish to thank Dr T N Ledger for revision of the English in the manuscript and Mrs Edine for excellent technical assistance. REFERENCES Agarwal VR, Ashanullah CI, Simpson ER & Bulun SE 1997 Alternatively spliced transcripts of the aromatase cytochrome P 450 (CYP 19) gene in adipose tissue of women. Journal of Clinical Endocrinology and Metabolism 82 70–74. Bellve´ AR, Cavicchia JC, Millette CF, O’Brien DA, Bhatnagar YM & Dym M 1977 Spermatogenic cells of the prepubertal mouse. Journal of Cellular Biology 74 68–85. Chen S & Zhou D 1992 Functional domains of aromatase cytochrome P 450 inferred from comparative analyses of amino acid sequences and substantiated by site-directed mutagenesis experiments. Journal of Biological Chemistry 267 22587–22594. Choi I, Collante WR, Simmen RCM & Simmen FA 1997 A developmental switch in expression from blastocyst to endometrial/placental-type cytochrome P450 aromatase genes in the pig and horse. Biology of Reproduction 55 688–696.

Corbin CJ, Graham-Lorence S, McPhaul MJ, Mason JL, Mendelson CR & Simpson ER 1988 Isolation of a full-length cDNA insert encoding human aromatase system cytochrome P450 and its expression in non-steroidogenic cells. Proceedings of the National Academy of Sciences of the USA 85 8948–8953. Corbin CJ, Khalil MW & Conley AJ 1995 Functional ovarian and placental isoforms of porcine aromatase. Molecular and Cellular Endocrinology 113 29–37. Delarue B, Bre´ard E, Mittre H & Leymarie P 1998 Expression of two aromatase cDNAs in various rabbit tissues. Journal of Steroid Biochemistry and Molecular Biology (In Press). Denhez F & Lafyatis R 1994 Conservation of regulated alternative splicing and identification of functional domains in vertebrate homologs to the Drosophila splicing regulator, suppressor-of-white-apricot. Journal of Biological Chemistry 269 16170–16179. Eddy EM, Washburn TF, Bunch DO, Goulding EH, Gladen BC, Lubahn DB & Korach KS 1996 Targeted disruption of the estrogen receptor gene in male mice causes alteration of spermatogenesis and infertility. Endocrinology 137 4796–4805. Foucault P, Drosdowsky MA & Carreau S 1994 Germ cell and Sertoli cell interactions in human testis: evidence for stimulatory and inhibitory effects. Human Reproduction 9 2062–2068. Fukada S, Tanaka M, Matsuyama M, Kobayashi D & Nagahama Y 1996 Isolation, characterization, and expression of cDNAs encoding the Medaka (Oryzias latipes) ovarian follicle cytochrome P-450 aromatase. Molecular Reproduction and Development 45 285–290. Fu¨rbass R & Vanselow J 1995 An aromatase pseudogene is transcribed in the bovine placenta. Gene 194 287–291. Journal of Molecular Endocrinology (1998) 20, 305–312

311

312

  and others

· P450 aromatase transcripts in male rat germ cells

Fu¨rbass R, Kalbe C & Vanselow J 1997 Tissue-specific expression of the bovine-aromatase-encoding gene uses multiple transcriptional start sites and alternative first exons. Endocrinology 138 2813–2819. Hickey GJ, Krasnow JS, Beattie WG & Richards JS 1990 Aromatase cytochrome P450 in rat ovarian granulosa cells before and after luteinization: adenosine 3 ,5 -monophosphatedependent and independent regulation. Cloning and sequencing of rat aromatase cDNA and 5 genomic DNA. Molecular Endocrinology 4 1–12. Janulis L, Bahr JM, Hess RA & Bunick D 1996a P450 aromatase messenger ribonucleic acid expression in male germ cells: detection by reverse transcription-polymerase chain reaction amplification. Journal of Andrology 17 651–658. Janulis L, Hess RA, Bunick D, Nitta H, Janssen Y, Osawa Y & Bahr JM 1996b Mouse epididymal sperm contains active P450 aromatase which decreases as sperm traverse the epididymis. Journal of Andrology 17 111–116. Kwon S, Hess RA, Bunick D, Nitta H, Janulis L, Osawa Y & Bahr JM 1995 Rooster testicular germ cells and epididymal sperm contain P450 aromatase. Biology of Reproduction 53 1259–1264. Lephart ED, Peterson KG, Noble JF, George FW & McPhaul MJ 1990 The structure of cDNA clones encoding the aromatase P-450 isolated from a rat Leydig cell tumor line demonstrates differential processing of aromatase mRNA in a rat ovary and a neoplastic cell line. Molecular and Cellular Endocrinology 70 31–40. Lephart ED, Simpson ER, McPhaul MJ, Kilgore MW, Wilson JD & Ojeda SR 1992 Brain aromatase cytochrome P-450 messenger RNA levels and enzyme activity during prenatal and perinatal development in the rat. Molecular Brain Research 16 187–192. Levallet J & Carreau S 1997 Aromatase gene expression in rat testicular cells in vitro. Comptes-Rendus de l’Academie des Sciences de Paris 320 123–129. Levallet J, Bilinska B, Mittre H, Genissel C, Fresnel J & Carreau S 1998 Expression and immunolocalization of functional cytochrome P450 aromatase in mature rat testicular cells. Biology of Reproduction 58 (In Press). Li H, Papadopoulos V, Vidic B, Dym M & Culty M 1997 Regulation of rat testis gonocyte proliferation by platelet-derived growth factor and estradiol: identification of signaling mechanisms involved. Endocrinology 138 1289–1298. Lubahn DB, Moyer JS, Golding TS, Couse JF, Korach KS & Smithies O 1993 Alteration of reproductive function but not prenatal sexual development after insertional disruption of

Journal of Molecular Endocrinology (1998) 20, 305–312

the mouse estrogen receptor gene. Proceedings of the National Academy of Sciences of the USA 90 11162–11166. Means GD, Kilgore MW, Mahendroo MS, Mendelson CR & Simpson ER 1991 Tissue-specific promoters regulate aromatase cytochrome P450 gene expression in human ovary and fetal tissues. Molecular Endocrinology 5 2005–2013. Meistrich ML, Longtin J, Brock WA, Grimes SR & Mace ML 1981 Purification of rat spermatogenic cells and preliminary biochemical analysis of these cells. Biology of Reproduction 25 1065–1077. Nitta H, Bunick D, Hess RA, Janulis L, Newton SC, Millette CF, Osawa Y, Shizuta Y, Toda K & Bahr JM 1993 Germ cells of the mouse testis express P450 aromatase. Endocrinology 132 1396–1401. Papadopoulos V, Carreau S & Drosdowsky MA 1985 Effects of phorbol ester and phospholipase C on LH-stimulated steroidogenesis in purified rat Leydig cells. FEBS Letters 188 312–316. Papadopoulos V, Carreau S, Szerman-Joly E, Drosdowsky MA, Dehennin L & Scholler R 1986 Rat testis 17â-estradiol: identification by gas chromatography–mass spectrometry and age related cellular distribution. Journal of Steroid Biochemistry 24 1211–1216. Simpson ER, Mahendroo MS, Means GD, Kilgore MW, Hinshelwood MM, Graham-Lorence S Amarneh B, It Y, Fisher CR, Michael DM, Mendelson CR & Bulun SE 1994 Aromatase cytochrome P450, the enzyme responsible for estrogen biosynthesis. Endocrine Reviews 15 342–355. Simpson ER, Michael MD, Agarwal VR, Hinshelwood MM, Bulun SE & Zhao Y 1997 Expresssion of the CYP 19 (aromatase) gene: an unusual case of alternative promoter usage. FASEB Journal 11 29–36. Toda K & Shizuta Y 1993 Molecular cloning of a cDNA showing alternative splicing of the 5 -untranslated sequence of mRNA for human aromatase P-450. European Journal of Biochemistry 213 383–389. Tsubota T, Nitta H, Osawa Y, Mason IJ, Kita I, Tiba T & Bahr JM 1993 Immunolocalization of steroidogenic enzymes, P450 scc, 3â-HSD, P450c17, and P450arom in the Hokkaido brown bear (Ursus arctos yesoensis) testis. General and Comparative Endocrinology 92 439–444. Tung PS, Skinner MK & Fritz IB 1984 Fibronectin synthesis is a marker for peritubular cell contaminants in Sertoli cell-enriched cultures. Biology of Reproduction 30 192–211. Vanselow J & Fu¨rbass R 1995 Novel aromatase transcripts from bovine placenta contain repeated sequence motifs. Gene 154 281–286.

 15 September 1997