Introduction

ABSTRACT

The ovulatory process in mammals involves a substantial increase in the metabolism of steroids and eicosanoids in response to a surge in LH, or to an injection of hCG into experimental animals. This study provides evidence that the ovulatory stimulus causes induction of the gene for 3a-hydroxysteroid dehydrogenase (3a-HSD), an enzyme that belongs to several oxidoreductase superfamilies that affect steroid and eicosanoid metabolism. Immature Wistar rats were primed with 10 IU eCG s.c., and 48 h later the 12-h ovulatory process was initiated by 10 IU hCG s.c.. Ovarian RNA was extracted at 0, 2, 4, 8, 12, and 24 h after injecting the animals with hCG. The RNA extracts were used for RT-PCR differential display to detect gene expression in the stimulated ovarian tissue. One of the PCR primer sets differentially amplified a cDNA fragment that is 52.3% homologous with a 3a-HSD gene in rat liver. Northern analyses revealed that maximum transcription was at 8 h after the animals had been treated with hCG. The Northerns also indicated that the 3a-HSD cDNA probe cross-hybridized with as many as six different bands of mRNA on the blots. In situ hybridization localized 3a-HSD mRNA in the granulosa and thecal layers of mature follicles and in newly formed corpora lutea at 24 h after the ovulatory stimulus. In conclusion, gene(s) for 3a-HSD are transcribed in ovarian follicles in response to an ovulatory dose of gonadotropin. A possible function of the oxidoreductase enzyme that is translated from the 3a-HSD mRNA may be to reduce the toxic aldehyde and ketone components of the steroids and eicosanoids that accumulate in the mammalian ovary at the time of ovulation.

INTRODUCTION

Mammalian 3a-hydroxysteroid dehydrogenase (3a-HSD) operates in conjunction with NADPH to catalyze the oxidoreduction of steroids and other compounds that possess aldehyde and ketone functional groups [1-3]. It is a rather ubiquitous reducing agent that belongs to the aldo-keto reductase and the short-chain dehydrogenase reductase superfamilies that encompass more than 60 enzymes in mammals, amphibians, insects, and bacteria [3-5]. Genetic expression of the enzyme has been reported in a wide variety of tissues, with steady-state levels in liver and intestine, and limited expression in stomach, lung, testis and ovary [1].

The amino acid sequence of 3a-HSD is quite different form 3b-HSD [6], which is well known for its anabolic role in the synthesis of ovarian progesterone at the beginning of the ovulatory process in mammals [7]. In contrast to 3b-HSD, 3a-HSD has been associated with the catabolism of progesterone—converting this ovulation-dependent steroid into biologically inactive, or less active, products [3, 8]. In addition to its action on progesterone and other steroids, this multifunctional carbonyl reductase is known to catalyze polycyclic aromatic hydrocarbons and prostaglandins [1, 9].

Evidence for 3a-HSD activity in ovarian tissue was first detected two decades ago by assessing the conversion of dihydrotestosterone to other steroids in the rat ovary [10]. Subsequently, there have been several brief reports of 3a-HSD activity in avian [8] and rat [11] granulosa cells. More recently, there has been a study of ovarian 3a-HSD activity at various stages of the estrous cycle of the adult rat [4]. However, the latter report concluded that the enzyme activity was minimal at the time of ovulation. Thus, none of these ovarian studies associated 3a-HSD with the biochemical events of ovulation.

In the present RT-PCR differential display study, expression of mRNA for 3a-HSD was discovered in the ovaries of immature rats that were primed with gonadotropins to induce ovulation. The cDNA fragment that was differentially expressed on an acrylamide gel hybridized at multiple sites on a Northern blot, suggesting that additional forms (or lengths) of the 3a-HSD transcript reported here are expressed in the ovary at the time of ovulation. The results from in situ hybridization show that ovarian expression of 3a-HSD mRNA occurs primarily in the granulosa layer and to a lesser extent in the thecal tissue. This transcription is not affected by ovulation-inhibiting doses of epostane or indomethacin, indicating that ovarian 3a-HSD expression is not dependent on the ovulatory increase in ovarian progesterone or prostanoid synthesis. The oxidoreductase enzyme that is translated from the 3a-HSD mRNA may function to reduce the toxic aldehyde and ketone components of the steroids and eicosanoids that accumulate in the mammalian ovary at the time of ovulation.

MATERIALS AND METHODS

Animal Tissue and Animal Injections, Ovulation Rate, and Prostaglandin Assay

Immature Wistar rats were induced to superovulate by treatment with eCG and hCG as described previously [12]. Ovarian RNA was extracted at the peri-ovulatory intervals of 0, 2, 4, 8, 12, and 24 h after hCG. These nucleic acid extracts were used for differential display and for Northern blotting. Epostane and indomethacin were injected s.c., also as described previously [12]. These anti-ovulatory agents were administered at 3 h after hCG in doses of 5.0 mg and 1.0 mg, respectively. The ovulation rate in the various experimental animals was determined by a procedure that also has been described previously [12]. Ovarian prostaglandin E₂ (PGE₂) was assayed by a procedure that has also been described previously [13]. For the determination of ovulation rate, the assay of ovarian PGE₂, and the extraction of ovarian RNA, rats were killed by exposure to CO₂. The animals were acquired, retained, and used in compliance with the NIH Guide and with the approval of the institutional animal care review committee.

Differential Display Protocols That Lead to Detection of 3a-HSD

The steps of the differential display were carried out as described previously [12]. In brief, RNA was extracted by a standard guanidine isothiocyanate/cesium chloride procedure. RT- PCR was performed using primers from an RNAimage Kit (G503, GenHunter Corporation, Nashville, TN). The primer set that yielded differentially expressed cDNA for 3a-HSD was comprised of the poly-T primer 5'-HTTTTTTTTTC-3' and the random primer 5'-HGGCTATG-3', with "H" representing the point of attachment of a HindIII restriction site to each primer. After extraction and re-amplification of the differentially expressed cDNA, standard Northern analyses were performed as described previously [12] to confirm the ovulation-specific expression of the parent mRNA for 3a-HSD. Lengths of the 3a-HSD mRNAs were determined by including a lane in the Northern blots that contained RNA markers (G3191, Promega Corporation, Madison, WI). Subsequently, the unique cDNA fragment was subcloned using a pCR-TRAP Cloning System (P404, GenHunter), and cloning colonies containing the 3a-HSD cDNA were identified by secondary Northern analyses. Manual sequencing of the cDNA was performed using a Sequenase Version 2.0 DNA Sequencing Kit (US70140, Amersham Pharmacia Biotech, Inc., Piscataway, NJ). In situ hybridization was performed as described previously [12].

Statistical Analysis

The intensity of the signals from the Northern blots was analyzed by the NIH-image densitometry program, as described previously [12]. Numerical data are presented as means ± SEM. The significance of the differences among the six principal time points of 0, 2, 4, 8, 12, and 24 h after hCG and the four groups in the epostane and indomethacin study were determined by Duncan's multiple range tests after a completely randomized one-way analysis of variance of the means of the groups. The probability value used as the cutoff between "significant" and "not significant" was P = 0.05.

RESULTS

Temporal Pattern of Ova Accumulation in the Oviduct During Ovulation

The established convention to tabulate ovulation rate is to count the number of ova in the oviduct. In this study, the oviducts were examined from 10 rats each at 0, 2, 4, 6, 8, 10, 12, 14, 16, 20, and 24 h after hCG administration to the animals. In these groups of animals, the number of ova were 0, 0, 0, 0, 0, 0, 0.7 ± 0.4, 5.3 ± 2.3, 28.3 ± 7.2, 64.6 ± 6.9, and 72.3 ± 3.2, respectively (Figure 1). Thus, ova first began to appear at 12 h after hCG administration, and they continued to accumulate during the next 8 h. Therefore, the data show that mature follicles begin to rupture at approximately 12 h after induction of the ovulatory process, but some follicles do not release ova until some hours later.

Differential Display of 3a-HSD cDNA During the Ovulatory Process

Following RT-PCR, the sub-populations of radioactively labeled cDNAs that were generated from RNA extracts at each of the six stages of the peri-ovulatory period were separated from one another by electrophoresis on polyacrylamide gels. The autoradiograph of these PAGE results revealed a relatively weak signal for a differentially expressed cDNA band that was most evident at 8 h after hCG, but was negligible in the 0-h control lane that represented the amount of gene expression just prior to initiation of the ovulatory process by hCG (Figure 2). Therefore, this most intense cDNA band (i.e., the band in the 8-h lane) was excised from the acrylamide gel and re-amplified for use as a probe in Northern analyses.

Northern Analysis of 3a-HSD mRNA Expression During Ovulation

The Northerns revealed an expression of mRNA (eventually identified as the message for 3a-HSD) during ovulation that was comparable to the pattern of expression of cDNA on the differential display autoradiograph (Figure 3). In order to compare the intensity of the signals from the Northern blots with other data on gene expression during ovulation, the amount of signal from the 8-h lane was arbitrarily set at 100%, and the densities at the other times during the peri-ovulatory period were expressed as fractions of that maximum. Accordingly, the NIH-image program was used to digitize all of the bands on the Northerns, and the ratio of the density of each experimental band to its corresponding b-actin control band was calculated for each lane. Means (± SEM) of the signal densities at 0, 2, 4, 8, 12, and 24 h after hCG were 4.7% ± 1.0%, 18.3% ± 5.8%, 55.1% ± 11.3, 100%, 68.5% ± 2.9%, and 20.7% ± 4.6%, respectively. Thus, 3a-HSD gene expression increased 21-fold within 8 h after initiation of the ovulatory process by injecting hCG into the animals. Subsequently, by 12 h after hCG (i.e., at the approximate time that follicles begin to rupture in this experimental animal), 3a-HSD gene expression had declined significantly.

Temporal Pattern of Ovarian PGE₂ Production During Ovulation

Ovarian PGE₂ was assayed at 0, 2, 4, 8, 10, 12, 14, 16, 20, and 24 h after hCG treatment of the animals in order to correlate the pattern of synthesis of this major prostanoid with the pattern of expression of 3a-HSD mRNA. The mean values of PGE₂ in groups of 10 rats at each of the experimental intervals after hCG were 0.53 ± 0.10, 0.65 ± 0.11, 16.58 ± 3.39, 29.95 ± 5.05, 30.13 ± 5.38, 30.56 ± 3.21, 20.80 ± 3.57, 11.64 ± 1.18, 8.70 ± 1.55, 5.33 ± 0.85, and 2.43 ± 0.30 ng PGE₂ per mg protein in the ovarian extracts (Figure 4). Thus, the results show that the biphasic pattern of PGE₂ production during ovulation is similar to the pattern of 3a-HSD mRNA expression (Figure 3).

Expression of Multiple mRNAs for 3a-HSD

The cDNA for 3a-HSD actually hybridized to multiple bands of mRNA on the Northern blots. These multiple bands were elucidated more vividly by performing an NIHimage densitometric analysis of the 0-h control lane and the 8-h lane (Figure 5). This optical density analysis was performed on one of the Northerns from the epostane and indomethacin experiment (see Figure 6) because this particular Northern exhibited the finest resolution of the bands. The results indicated that the 3a-HSD cDNA hybridized with two ovarian mRNAs of approximately 3.6 and 1.8 kb that were constitutively expressed before and during the ovulatory process. However, by 8 h after hCG, there were an additional four hybridization bands of approximately 5.0, 3.0, 2.5, and 1.4 kb, representing mRNAs that were uniquely expressed during ovulation.

Effects of Epostane and Indomethacin on 3a-HSD Gene Expression

For these tests, Northern blots were prepared from RNA that was extracted from control ovaries at 0 and 8 h into the ovulatory process, or extracted from experimental ovaries that were taken at 8 h after hCG from rats that had been treated 5 h earlier with ovulation-inhibiting doses of epostane or indomethacin. These experimental intervals were selected in order to compare the present data with related studies using epostane and indomethacin [12, 14, 15]. The signal density (normalized against the b-actin control) of the 8-h control lane was arbitrarily set at 100% (Figure 6). There was minimal expression of 3a-HSD mRNA at 0 h, but substantial expression at 8 h. In animals treated with the anti-ovulatory agent epostane, which blocks progesterone synthesis [15-17], the signal density of 101.3% ± 6.3% was not significantly different from the 8-h control value. Animals treated with the anti-ovulatory agent indomethacin, which blocks prostanoid synthesis [15, 18], had a signal density that was 105.1 ± 5.4% of the 8-h control value and was not significantly different from the 8-h control value. On the other hand, the ovulation rates in parallel groups of animals treated with epostane and indomethacin were significantly inhibited (Figure 6). These results indicate that the ovulation-related increase in 3a-HSD gene expression is not dependent on ovarian progesterone or ovarian prostaglandin levels.

Sequence of the cDNA Fragment for 3a-HSD

After the hCG-induced expression of the 3a-HSD gene was confirmed by Northern analysis, the cDNA fragment of this gene was subcloned and sequenced. The NCBI accession number for this 501-bp fragment is #AF159099. This fragment of cDNA is homologous to segments of a 3a-HSD gene that has been cloned from rat liver (NCBI accession #M64393 [1] and NCBI accession #D17310 [9]). In comparison, the 501-bp ovarian cDNA that was isolated in the present study using RT-PCR differential display is 52.3% homologous (at the nucleotide level) with one of the rat liver clones (i.e., NCBI accession #M64393). The ovarian sequence contains four different segments, comprising 286 of the 501-bp total, that collectively are 91.6% homologous to complementary segments of the rat liver gene for 3a-HSD (Figure 7). (The sequence which was identified above as NCBI accession #D17310 is essentially identical to #M64393, except the segment of this liver mRNA that was submitted to the database did not include the last 21 bp at the poly-A end of the transcript.)

Localization of 3a-HSD mRNA Expression by In Situ Hybridization

In situ hybridization confirmed the temporal pattern of 3a-HSD mRNA expression that was observed in the differential display autoradiographs and the Northern analyses. There was minimal signal from the 0-h control ovaries, a substantial increase in signal between 4-8 h after hCG treatment, a declining signal between 8-12 h, and limited signal at 24 h (Figure 8). Hybridization was localized primarily in the granulosa of the larger follicles in ovaries staged at 4-12 h. In addition, a noticeable amount of 3a-HSD mRNA was visible in the thecal tissue, and this expression was the most obvious at 4 h after hCG was administered to the animals (Figure 9). Also, there was conspicuous signal from a limited number of luteal follicles in ovaries taken at 24 h after hCG treatment (Figure 9).

DISCUSSION

There are more than 60 related enzymes that exhibit aldo-keto reductase activity [5]. In the present study, the cDNA fragment of a gene that encodes such an enzyme is most homologous to rat liver 3a-HSD [1]. This gene for rat liver 3a-HSD transcribes a 2.4 kb message that encodes a 322 residue protein [1]. Northern analyses reveal that transcripts of this and related genes cross-hybridize with several different sizes of mRNA on Northern blots prepared from diverse tissues, including the ovary [1, 9]. However, these Northern studies report inconsistent sizes for the mRNAs that hybridize with hepatic 3a-HSD. In one study, two bands of 3a-HSD were identified at 3.6 and 1.8 kb on Northern blots of total RNA from rat ovary [1], while in another investigation four bands at 5.0, 3.0, 2.5, and 1.4 kb were detected on Northern blots from rat ovary [9]. It is difficult to explain the discrepancy between these reports. However, in both cases the experimental tissues were taken at random from adult rats, and therefore the ovaries could have been at different stages of the sexual cycle. It is also possible that, in the former report, the two bands represented constitutively expressed 3a-HSD mRNA that is transcribed at steady state levels in the ovary, while in the latter study the four bands may have resulted from transcription in response to gonadotropin action during the peri-ovulatory period. Such an explanation would be congruous with the present observation of what appears to be two "housekeeping" mRNAs that are approximately 2.4 and 1.4 kb in length, and four gonadotropin-induced mRNAs of approximately 3.8, 2.1, 1.8, and 0.8 kb that were expressed at 4-12 hours after treatment of the immature rats with hCG to induce ovulation. The limited expression of two "housekeeping" genes might explain the modest mount of 3a-HSD mRNA that was detected in 0-h control tissue. In any case, the data indicate that multiple mRNAs for 3a-HSD-like activity are transcribed in the rat following treatment of the animals with gonadotropin doses that induce ovulation.

It is relevant that members of the oxidoreductase families to which 3a-HSD belongs are homologous to a number of other enzymes that exhibit dehydrogenase, or reductase, activity on prostanoids, steroids, and other polycyclic aromatic hydrocarbons. The ensemble of related enzymes includes 20a-HSD [2], aldose reductase [1-3], chlordecone reductase [2], rho-crystallin [3], and prostaglandin F₂_a synthase [1-3, 6]. Prostaglandin F₂_a synthase is of particular interest in a study of ovarian function because this enzyme regulates the metabolism of prostaglandins, which have been frequently implicated in the mechanism of ovulation during the past three decades [7, 19]. With regard to its action on steroid substrates, 3a-HSD is especially known for its reduction of progesterone to 4-pregnen-3a-ol,20-one [3, 8], and for its interconversion of 5a-dihydrotestosterone or 5a-androstane-3,17-dione to 5a-androstane-3a,17b-diol or 5a-androstane-3a-ol,17-one [4]. Yet, its principal substrate(s) and product(s) in the ovary at the time of ovulation remain to be determined.

Regarding prostaglandin metabolism, 3a-HSD (which is also known as dihydrodiol dehydrogenase) can function as a 9-, 11-, and 15-hydroxyprostaglandin dehydrogenase, while prostaglandin F₂_a synthase reduces prostaglandin E₂ to prostaglandin F₂_a [6]. It is especially interesting to note that both of these enzymes are considered to be targets for nonsteroidal anti-inflammatory drugs (NSAIDs) [6]. Furthermore, NSAIDs are well known for their ability to block mammalian ovulation, and this inhibitory action has been attributed previously to interference with ovarian cyclo-oxygenase activity [7, 19, 20] and/or to inhibition of ovarian lipoxygenase activity [18] during the ovulatory process. Now, in view of the present evidence of transcription of 3a-HSD mRNA in ovulatory follicles, it is important in the future to consider the possibility that at least part of the anti-ovulatory action of NSAIDs may be the result of inhibition of ovarian 3a-HSD. Furthermore, it may be relevant that the temporal pattern of ovarian 3a-HSD expression coincides with the transient increase in ovarian prostaglandin E₂ and F₂_a synthesis [16, 17], along with 12- and 15-hydroxyeicosatetraenoic acids [17], during the ovulatory process. This correlation between ovarian a-HSD mRNA expression and prostanoid synthesis can be observed by comparing the data in Figures 3 and 4. Collectively, this information suggests that 3a-HSD may be another factor in the metabolism of ovarian eicosanoids in response to gonadotropic stimulation of the ovary.

As indicated above, 3a-HSD enzyme activity is inhibited by NSAIDs [6]. However, based on the present findings, gonadotropin-induced transcription of 3a-HSD mRNA is not affected by the potent NSAID, indomethacin. Similarly, epostane, which is a potent inhibitor of progesterone synthesis, did not affect the increase in ovarian 3a-HSD. These results show that the ovulatory increase in 3a-HSD mRNA occurs independently of ovarian prostanoid and progestin synthesis.

Several earlier studies on ovarian 3a-HSD did not relate this oxidoreductase activity to the process of ovulation [4, 8, 10, 11]. It should be noted that the initial study assessed 3a-HSD activity in adult rat ovaries by measuring the conversion of dihydrotestosterone to 5a-androstane-3a, 17b-diol [10]. In this case, it was reported that peak ovarian activity was on the day of proestrus, while the nadir was on the day of estrus. However, the precise time of day that the ovaries were extracted for 3a-HSD activity was not recorded. In contrast, the other major study of ovarian 3a-HSD activity was based on immunohistochemical localization of the enzyme protein [4]. In this latter report, the intensity of immunostaining (in granulosa and luteal cells) was highest at 0800 h on the day of proestrus and at noon on the day of estrus. Thus, the former study concluded that ovarian 3a-HSD activity was lowest at estrus, while the latter study found the most activity on this same day. Quite oppositely, the current investigation of mRNA shows that expression of the transcript is minimal in the 0-h control ovaries of gonadotropin-treated immature rats, and such a 0-h control is equivalent to an ovary that is taken from an adult rat at 0800-1200 h on the day of proestrus (i.e., before the usual endogenous surge of gonadotropin at approximately 1300-1400 h). Furthermore, the time of 2000 h on the day of proestrus (i.e., the reported time of the nadir for immunostaining of ovarian 3a-HSD [4]) is comparable to 6-7 hours after hCG injection to eCG-primed immature rats (i.e., the approximate time of the peak for 3a-HSD mRNA expression that was observed in the present study). Therefore, since mRNA transcription and enzyme translation usually coincide with one another, the present findings on the temporal pattern of ovarian 3a-HSD mRNA expression are not compatible with previous estimates of ovarian 3a-HSD enzyme levels. The reason for this incongruity is not clear. However, it should be emphasized that the temporal data from the differential display autoradiograph, the Northern analysis, and the in situ hybridization experiments of the present study all show that ovarian 3a-HSD mRNA expression is minimal at the beginning of the ovulatory process and maximal at 8 hours after inducing the process by an injection of hCG. Future studies on the temporal pattern of expression of the 3a-HSD protein and/or its enzymatic activity should clarify the discrepancy between the previous reports on 3a-HSD enzymatic activity and the current data on the mRNA for this enzymatic activity.

Although the protein that is translated from ovarian 3a-HSD mRNA is most likely a reductase, the nucleotide sequence of the parent mRNA is not homologous to the polymorphic sequence of an ovarian carbonyl reductase gene that we have recently reported as being up-regulated following gonadotropic stimulation [12]. Furthermore, the message for carbonyl reductase is expressed in the thecal layers of ovarian interstitial tissue, whereas 3a-HSD mRNA is expressed primarily in the granulosa layer of mature follicles and in the differentiated granulosa of some corpora lutea. It was noted that 3a-HSD mRNA is not uniformly expressed among the numerous corpora lutea on the super-ovulated ovaries. This irregularity in the luteal distribution of 3a-HSD could be related to the fact that all of the follicles on a given rat ovary do not ovulate at the same time. The first follicles begin to rupture at approximately 12 h after hCG administration, but they continue to be released for the next 8 hours. Therefore, the follicles that are slower to rupture may be the ones that contain the greater amount of lingering 3a-HSD mRNA that was observed by in situ hybridization at 24 h after hCG administration.

In summary, 3a-HSD mRNA is expressed in substantial amounts in ovarian follicles that have been stimulated by an ovulatory dose of gonadotropin. The expression of this oxidoreductase message occurs simultaneously with the well-known increase in ovarian eicosanoids and progesterone. Therefore, since it has been established that 3a-HSD reduces prostanoids, steroids, and a variety of other polycyclic aromatic hydrocarbons (3, 9), it has been suggested that one possible function of such reductase enzymes is to "detoxify" aldehydes and ketones that sometimes accumulate in tissues that are undergoing considerable metabolic differentiation during processes such as follicular-luteal transition [8, 12]. However, it should be noted that the reductase gene fragment that was isolated in the present study does not share any significant homology with a mouse vas deferens protein gene that has been reported recently as having "a detoxifying role" in murine ovaries that have been stimulated by LH/hCG [21]. It has also been suggested that oxidoreductase enzymes like 3a-HSD naturally influence the life-span of bioactive eicosanoids and steroids by moderating the duration of their interaction with nuclear and membrane-bound receptors [3]. In any event, the significant increase in 3a-HSD mRNA in the granulosa layer of ovulatory follicles suggests that the translated enzyme has some relevant purpose in the biochemical events of ovulation.

ACKNOWLEDGEMENT

We appreciate the excellent assistance of Mrs. Claire Lo in preparing the in situ hybridization data.

REFERENCES

1. Stolz A, Rahimi-Kiani M, Ameis D, Chan E, Ronk M, Shively JE. Molecular structure of rat hepatic 3a-hydroxysteroid dehydrogenase. A member of the oxidoreductase gene family. J Biol Chem 1991; 15253-15257.

2. Miura R, Shiota K, Noda K, Yagi S, Ogawa T, Takahashi M. Molecular cloning of cDNA for rat ovarian 20a-hydroxysteroid dehydrogenase (HSD1). Biochem J 1994; 299:561-567.

3. Penning TM, Pawlowski JE, Schlegel BP, Jez JM, Lin HK, Hoog SS, Bennett MJ, Lewis M. Mammalian 3a-hydroxysteroid dehydrogenases. Steroids 1996; 61:508-523.

4. Takahashi M, Iwata N, Hara S, Mukai T, Takayama M, Endo T. Cyclic changes in 3a-hydroxysteroid dehydrogenase in rat ovary during the estrous cycle. Biol Reprod 1995; 53:1265-1270.

5. Duax WL, Ghosh D, Pletnev V. Steroid dehydrogenase structures, mechanism of action, and disease. Vitam Horm 2000; 58:121-148.

6. Pawlowski JE, Huizinga M, Penning TM. Cloning and sequencing of the cDNA for rat liver 3a-hydroxysteroid/dihydrodiol dehydrogenase. J Biol Chem 1991; 266:8820-8825.

7. Espey LL, Lipner H. Ovulation. In: Knobil E, Neill JD, (eds.), The Physiology of Reproduction. New York: Raven Press; 1994: 725-780.

8. Wiebe JP, Buckingham KD, Zobell RL, Hertelendy F. Metabolism of progesterone by avian granulosa cells in culture. J Steroid Biochem Mol Biol 1990; 37:113-120.

9. Usui E, Okuda K, Kato Y, Noshiro M. Rat hepatic 3a-hydroxysteroid dehydrogenase: expression of cDNA and physiological function in bile acid biosynthetic pathway. J Biochem (Tokyo) 1994; 115:230-237.

10. Jarrell J, Robaire B. On the presence and regulation of 3a-hydroxysteroid dehydrogenase in the adult rat ovary. J Steroid Biochem 1982; 16:725-730.

11. Tedeschi C, Hazum E, Kokia E, Ricciarelli E, Adashi EY, Payne DW. Endothelin-1 as a luteinization inhibitor: inhibition of rat granulosa cell progesterone accumulation via selective modulation of key steroidogenic steps affecting both progesterone formation and degradation. Endocrinology 1992; 131:2476-2478.

12. Espey LL, Yoshioka S, Russell D, Ujioka T, Vladu B, Skelsey M, Fujii S, Okamura H, Richards JS. Characterization of ovarian carbonyl reductase gene expression during ovulation in the gonadotropin-primed immature rat. Biol Reprod 2000; 62:390-397.

13. Espey LL, Adams RF, Tanaka N, Okamura H. Effects of epostane on ovarian levels of progesterone, 17b-estradiol, prostaglandin E₂, and prostaglandin F₂_a during ovulation in the gonadotropin-primed immature rat. Endocrinology 1990; 127:259-263.

14. Espey LL, Yoshioka S, Russell DL, Robker RL, Fujii S, Richards JS. Ovarian expression of a disintegrin and metalloproteinase with thrombospondin motifs during ovulation in the gonadotropin-primed immature rat. Biol Reprod 2000; 62:1090-1095.

15. Espey LL, Ujioka T, Russell DL, Skelsey M, Vladu B, Robker RL, Okamura H, Richards JS. Induction of early growth response protein-1 gene expression in the rat ovary in response to an ovulatory dose of human chorionic gonadotropin. Endocrinology 2000; 141:2385-2391.

16. Espey LL, Adams RF, Tanaka N, Okamura H. Effects of epostane on ovarian levels of progesterone, 17b-estradiol, prostaglandin E₂, and prostaglandin F₂_a during ovulation in the gonadotropin-primed immature rat. Endocrinology 1990; 127:259-263.

17. Espey LL, Tanaka N, Adams RF, Okamura H. Ovarian hydroxyeicosatetraenoic acids compared with prostanoids and steroids during ovulation in rats. Am J Physiol 1991; 260:E163-E169.

18. Tanaka N, Espey LL, Kawano T, Okamura H. Comparison of inhibitory actions of indomethacin and epostane on ovulation in rats. Am J Physiol 1991; 260:E170-E174.

19. Richards JS. Hormonal control of gene expression in the ovary. Endocr Rev 1994; 15:725-751.

20. Espey LL, Stein VI, Dumitrescu J. Survey of antiinflammatory agents and related drugs as inhibitors of ovulation in the rabbit. Fertil Steril 1982; 38:238-247.

21. Brockstedt E, Peters-Kottig M, Badock V, Hegele-Hartung C, Lessl M. Luteinizing hormone induces mouse vasdeferens protein expression in the murine ovary. Endocrinology 2000; 141:2574-2581.

LEGENDS

Figure 1. Onset of ovulation, based on the accumulation of ova in the oviduct following treatment of immature rats with hCG. Each mean is based on the number of ova in the oviducts of 10 rats.

Figure 2. Autoradiograph of differentially displayed 3a-HSD cDNA (arrows). Note that the cDNA for 3a-HSD is not visible in the lane containing the 0-h PCR product.

Figure 3. Intensity of Northern blot signals at the six intervals of the peri-ovulatory period following hCG administration. The signal density at 8 h was arbitrarily set at 100%, and the other points on the graph represent the mean values from probing six Northern blots. Two each of the six Northern blots were prepared from three different ovarian RNA extracts, each obtained from groups of 6-12 ovaries that were extracted at each of the peri-ovulatory intervals. Therefore, each mean value is based on at least 18 ovaries, but the statistical analysis was performed on the densitometric data from six Northern blots. The actual Northern signal for one radio-labeled 3a-HSD cDNA probe, along with the b-actin control for that Northern, are shown below the linear graph. Note that the greatest intensity is at 8 h after hCG.

Figure 4. Temporal pattern of ovarian PGE₂ production following the induction of ovulation with an injection of hCG. Each mean is based on the RIA of PGE₂ in individual ovarian extracts from 10 rats at each interval following gonadotropin treatment.

Figure 5. Simultaneous plots of the signal densities of the 0-h control lane and the 8-h post-hCG lane of one Northern. Note that the 0-h plot reveals two signal peaks that have corresponding (albeit smaller) peaks that are superimposed among the four peaks on the 8-h plot. This NIH-image analysis was performed on the Northern data illustrated in Figure 6, below.

Figure 6. Comparison of the % of signal from Northern blots containing RNA extracted at 8 h after hCG from animals that were also treated with either 5 mg of epostane (Epo), or 1 mg of indomethacin (Indo) administered at 3 h after hCG. Bar graphs are based on NIH-image analyses of four different Northerns probed with 3a-HSD cDNA. The signal from the 8-h control lane (Ctrl) was arbitrarily set at 100% OD. In parallel groups of rats, the ovulation rate was determined at 24 h after hCG.

Figure 7. Comparison of the 501-bp sequence of ovarian 3a-HSD mRNA with the corresponding segment of hepatic 3a-HSD mRNA (NCBI accession #M64393). Ovarian sequence is in normal letters. The four segments of homologous hepatic sequence are represented by underlined, bold letters below the ovarian sequence. The beginning of the ovarian sequence is the 5'-end of the fragment.

Figure 8. Change in intensity of the in situ hybridization signal during the six peri-ovulatory intervals following hCG administration. Light-field micrographs on the left show the histology of ovarian sections stained with hematoxylin and eosin (H & E), while the dark-field micrographs of the same sections show the localization of 3a-HSD mRNA as detected by hybridization of a ³⁵S-labeled anti-sense probe derived from the 3a-HSD cDNA. (Magnification ~7X)

Figure 9. Closer view of the distribution of probe in the thecal tissue of follicles in an ovary at 4 h after hCG injection (white arrows), in the granulosa layer (white arrow) of a follicle in an ovary at 8 h after hCG, and in a corpus luteum (white arrow) of an ovary at 24 h after hCG. Note that several other luteal masses in the same area emitted less signal. (Magnification ~35X)