Properties and Hormonal Regulation of Two Structurally Related cAMP Phosphodiesterases from the Rat Sertoli Cell*

upon exposure to 10 ng of FSH/ml, ratPDE3 levels could be further with higher FSH concentrations. The intensity of an immunoreactive band with characteristics identical to a purified CAMP-PDE, correlated with the increased cAMP hydrolytic activity after FSH or dibutyryl treatment, demonstrating that changes in CAMP-PDE protein levels are involved in this regulation. These data provide evidence that multiple CAMP- PDE forms are expressed in the rat Sertoli cell. Although differences in the pattern of activation of these forms were observed, these data show, that in the rat Sertoli cell, the CAMP-PDE activity is regulated by hormones via a novel mechanism that involves a CAMP-dependent activation of transcription of a PDE gene. related CAMP-PDEs is under the control of hormones. We show that differences in expression occur at least in part at the level of gene transcription and that an increase in PDE activity follows changes in PDE protein concentration.

Upon exposure to follicle-stimulating hormone (FSH), the gonadotropin-responsive Sertoli cell expresses increased rolipram-sensitive CAMP-specific phosphodiesterase (CAMP-PDE) activity. To understand the mechanisms leading to this activation, the CAMP-PDEs present in the Sertoli cell were characterized and their regulation studied. Comparison of the conceptual translates of two groups of PDE cDNA clones isolated from a Sertoli cell cDNA library (ratPDE3 and ratPDE4) showed that the encoded proteins were structurally similar, containing a core region of 455 amino acids with a sequence identity of 87%. The amino and carboxyl termini were divergent. Expression of these cDNAs in Escherichia coli and monkey COS-7 cells demonstrated that the encoded CAMP-PDEs had similar affinities for the cAMP substrate and were equally sensitive to a number of PDE inhibitors (rolipram > Ro 20-1724 > cilostamide). FSH stimulation of the Sertoli cell produced an increased rate of transcription of the ratPDE3 gene and elevated mRNA levels for ratPDE3 and to a lesser extent of ratPDE4. The increase in mRNA levels was detected after 1 h of stimulation. Forskolin, cholera toxin, and M,02'-dibutyryl cAMP produced a similar increase in rate of transcription and elevated mRNA levels, indicating that this activation is mediated by an increase in intracellular CAMP. RatPDE4 mRNA levels were maximal upon exposure to 10 ng of FSH/ml, whereas ratPDE3 mRNA levels could be further elevated with higher FSH concentrations. The intensity of an immunoreactive band with characteristics identical to a purified CAMP-PDE, correlated with the increased cAMP hydrolytic activity after FSH or dibutyryl cAMP treatment, demonstrating that changes in CAMP-PDE protein levels are involved in this regulation. These data provide evidence that multiple CAMP-PDE forms are expressed in the rat Sertoli cell. Although differences in the pattern of activation of these forms were observed, these data show, that in the rat Sertoli cell, the CAMP-PDE activity is regulated by hormones via a novel mechanism that involves a CAMP-dependent activation of transcription of a PDE gene.
* This work was supported by National Institutes of Health Grant HD20788. 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

MZ35349-M25350.
The responsiveness of the target cell to hormones is continuously adjusted in synchrony with changes in the extracellular environment. The mechanisms involved in this adaptation are complex and involve changes in the components of the signal transduction machinery at the plasma membrane level as well as changes in the enzymes involved in intracellular signalling. Phosphodiesterases (PDEs),' the enzymes that degrade cyclic nucleotides might play an important role in these processes. In most cells, including the Sertoli cell (l), degradation of cyclic nucleotides is carried out by multiple PDEs with different properties in terms of affinity for cAMP and cGMP, regulation of hydrolytic activity, and sensitivity to inhibitors (2). Although several types of PDEs are able to hydrolyze CAMP, it is now recognized that two families of PDEs have a high affinity for this nucleotide. One of these forms, referred to as cGMP-inhibited PDE is sensitive to the selective inhibitor cilostamide and the cardiotonic milrinone (2). In several cell types, this enzyme is membrane bound and rapidly activated by insulin and P-adrenergic agonists (3). The other type, termed CAMP-specific PDE (CAMP-PDE) is insensitive to cGMP and is selectively inhibited by the antidepressant rolipram (2). Also this latter form of PDE has been shown to be regulated by hormones. In the Sertoli cell, FSH via an increase in cAMP causes the activation of a PDE with the properties of a CAMP-PDE (1,4). Similar hormonedependent regulation of cAMP hydrolyzing activity has also been demonstrated in glioma and astrocytoma cell lines (5-8), in fibroblasts (9, lo), granulosa (11), and lymphoma cells (12). The CAMP-PDE which in the rat Sertoli cell is stimulated by FSH, is thought to be involved in the refractoriness that follows hormone stimulation (13-15). It is now clear that the family of the CAMP-PDEs is composed of multiple related forms. Molecular cloning data (16)(17)(18) have demonstrated that in the rat there are at least four genes encoding cAMP phosphodiesterases which have been highly conserved through evolution (16,17). It has been shown that the steady state mRNA level for one of these PDEs (ratPDE3) is regulated by hormones and cAMP (19). Here we have studied in detail the mechanisms involved in this regulation. In addition, the finding that another CAMP-PDE gene (ratPDE4) is expressed in the Sertoli cell prompted us to study the properties and the regulation of this latter PDE in response to FSH. Here we report that the expression of both these structurally related CAMP-PDEs is under the control of hormones. We show that differences in expression occur at least in part at the level of gene transcription and that an increase in PDE activity follows changes in PDE protein concentration.

Sequence Analysis
The cloning of PDEs from a rat Sertoli cell cDNA library has been described (16). The complete ratPDE3 DNA sequence (19) and also a partial cDNA sequence of two ratPDE4 clones have been published (16). These clones have been completely sequenced and also two additional ratPDE4 clones (ratPDE4.3 and ratPDE4 .4) have been partially characterized.
The sequences were aligned using the MicroGenie software package (Beckmann).

Prokaryotic Expression
To express ratPDE3 and ratPDE4 cDNAs in bacteria, the prokaryotic expression vector pRC23 (20) (kindly provided by Dr. R. Crowl) was used. This vector contains the h phage PL promoter followed by a ribosome-binding site and an EcoRI cloning site. The distance between the ribosome-binding site and the AUG codon is critical for efficient translation (21). Because the EcoRI site in cDNA clone ratPDE3.1 precedes the putative initiation AUG codon by 150 bases (19), a new EcoRI site was created 3 bases 5' of the AUG by the polymerase chain reaction. cDNA clone ratPDE4.2 starts with an GcoRI site 5 bases upstream of the putative AUG and was ligated into the EcoRI site of pRC23 without further manipulation. Escherichia coli DH5a bacteria were transformed (22) with the pRC23 compatible plasmid pRK248cIts (provided by Dr. R. Crowl) which encodes a temperature-sensitive hcIAt2 repressor (23) and confers tetracyclin resistance to the cells. Tetracyclin resistant pRK248cIts containing DH5a bacteria were transformed with the pRC23-rat-PDE3 and pRC23-ratPDE4 constructs and grown in the presence of ampicillin (the selectable marker of pRC23) a t 30 "C. Expression of the ratPDEs by shifting the temperature from 30 to 42 "C for 3 h was monitored by polyacrylamide gel electrophoresis followed by Coomassie Brilliant Blue staining or Western blot analysis (with K l l l antiserum, see below) and PDE assays. For PDE assay, the bacteria were pelleted, resuspended in 20 mM Tris and 1 mM EDTA, and lysed with 5 mg/ml of lysozyme. The solution was adjusted to 0.2 mM EGTA, 50 mM benzamidine, 0.5 pg/ml leupeptin, 0.7 pg/ml pepstatin, and 2 mM phenylmethylsulfonyl fluoride.

Eukaryotic Expression
The pCMV5 expression in vector (24) drives the expression of cDNAs cloned into the multiple cloning site in between the human cytomegalovirus major immediate early gene promoter and enhancers and the transcription termination and polyadenylation signals of the human growth hormone. The ratPDE3.1 and ratPDE4.1 cDNAs were inserted into the EcoRI site of the plasmid. After transformation, ampicillin-resistant DH5a bacteria were tested for the presence and the orientation of the ratPDE cDNAs. Plasmids were prepared from 1 liter of bacterial cultures using p2523 columns according to the protocol of the manufacturer. Monkey kidney cells (COS-7) were grown in high glucose Dulbecco's modified Eagle's medium containing 5% (v/v) fetal calf serum. When the cells reached a density of approximately 40% of confluency, the medium was aspirated and medium without serum was added. The cells were transfected using the DEAE-dextran method (25) with 10 pg of plasmid DNA. Four hours after the addition of the DNA the cells were treated with 10% dimethyl sulfoxide in phosphate-buffered saline (PBS), rinsed with PBS, and incubated in complete medium for 48 h. "Mock" transfected cells were treated exactly the same way but did not receive DNA. The cells were rinsed with PBS and lysed for RNA extraction, or collected with a rubber policeman and homogenized in an all-glass Dounce homogenizer in a buffer containing 20 mM Tris.Cl (pH 8.01, 10 mM EDTA, 0.2 mM EGTA, 50 mM benzamidine, 0.5 pg/ml leupeptin, 0.7 pg/ml pepstatin, and 2 mM phenylmethylsulfonyl fluoride (homogenization buffer). After centrifugation for 15 min a t 14,000 X g, the supernatant and the pellet fractions were subjected to Western blot analysis, immunoprecipitation, or PDE assay.

Sertoli Cell Cultures
Sertoli cell cultures from 15-to 17-day-old Sprague-Dawley rats were prepared as described (26) and maintained in medium without serum for 4 days. On the fourth day of culture the cells were rinsed with Hank's balanced salt solution, the medium was replaced and FSH (NIDDK ov-FSH S16), N6,O""dibutyryl CAMP, forskolin, cholera toxin, or vehicle were added to the medium. At the indicated time after treatment, the cells were rinsed twice with PBS and lysed for RNA extraction. For immunoprecipitation, Western blot analysis, or PDE assay, the cells were scraped in homogenization buffer with a rubber policeman and homogenized in an all-glass Dounce homogenizer. Cellular fractions were prepared by centrifugation at 14,000 X g for 15 min.

P D E Assay
PDE assays were performed basically as described (27) and as detailed previously (19). Cell extracts in homogenization buffer were adjusted to 40 mM Tris.Cl (pH 8.0), 10 mM MgC12, 1.25 mM 2mercaptoethanol, and 1 p~ ["HICAMP (0.1 pCi/reaction). After incubation a t 34 "C for 3-10 min the reactions were terminated by addition of 1 volume of a 40 mM Tris. C1 (pH 7.5) solution containing 10 mM EDTA and heat denaturation. To each reaction tube 50 pg of C. atrox snake venom was added. After incubation a t 34 "C for 20 min the reaction products were separated by anion exchange chromatography on AG 1-X8 resin and the amount of released radiola-

RNA Extraction and Northern Blot Analysis
Total RNA was prepared as described (29). Sertoli or COS-7 cells were lysed in a buffer containing 50% (w/v) guanidine thiocyanate, 16.7 mM Na citrate, 0.5% (w/v) Na lauryl sarcosine, 80 mM 2mercaptoethanol, and 0.1% (v/v) Sigma antifoam A. This cell lysate was layered on top of a solution containing 5.7 M cesium chloride and 167 mM Na acetate (pH 5.0) (30). The RNA was pelleted by centrifugation for 16 h a t 100,000 X g. After centrifugation the RNA pellet was redissolved in ddH,O and ethanol precipitated. Poly(A)' RNA was prepared by oligo(dT)-cellulose chromatography as described (31). Five micrograms of poly(A)+ RNA or 10-20 pg of total RNA was denatured by incubation a t 50 "C for 1 h in a buffer containing 50% (v/v) dimethyl sulfoxide, 20% (v/v) glyoxal, and 20 mM sodium phosphate (pH 6.8) (31). After separation on a 1% agarose gel in 10 mM sodium phosphate (pH 6.8), the gel was blotted onto a Biotrans nylon membrane (32) and baked a t 80 "C for 1 h. Blots were prehybridized in 50% v/v) formamide, 50 mM sodium phosphate (pH 6.5), 5 X SSC (20 X SSC, 3 M NaCl and 0.3 M Na:, citrate), 5 X Denhardt's (100 X Denhardt's, 2% (w/v) Ficoll 400, 2% (w/v) polyvinylpyrrolidone, 2% (w/v) bovine serum albumin, fraction V), 250 pg/ml of boiled sonicated salmon sperm DNA, 0.5% (w/v) SDS, and 1% (w/v) glycine at 42 "C for 4 h. RatPDE3, ratPDE4, and actin cDNA clones were radiolabeled (33) using a random primed labeling kit and follow- Nuclear Run-off Nuclear run-off assays were performed basically as described (34). Isolation of Nuclei-Sertoli cells from 17-day-old male Sprague-Dawley rats were cultured in medium without serum. On the fourth day of culture the medium was changed and the cells were treated with 500 ng/ml of FSH for different lengths of time (0 min, 30 min, 3 h, 24 h) or with 500 ng/ml of FSH, 1 mM B2cAMP, 100 p M forskolin, or 100 ng/ml of cholera toxin for 3 h. After the appropriate incubation time the medium was removed and the cells were rinsed twice with ice-cold PBS. The cells were collected by scraping with a rubber policeman and centrifugation for 5 min at 500 X g. For each nuclear run-off reaction nuclei from approximately 5 X lo7 cells were prepared by Dounce homogenization as described (34).
Binding Plasmids to Nitrocellulose-Plasmids containing cDNA inserts of c-fos, ubiquitin, ratPDE3, or ratPDE4 were linearized with the appropriate restriction enzymes, phenol and chloroform extracted, and ethanol precipitated. The linearized plasmids were denatured in 0.1 M NaOH for 30 min at room temperature and neutralized by addition of 10 volumes of 6 X SSC (20 X SSC, 3 M NaCl, 0.3 M Na3 citrate). Five micrograms of each plasmid was spotted onto nitrocellulose filters using a slot blot apparatus. The filters were baked for 2 h at 80 "C under vacuum.
Nuclear Run-off Reactions-These reactions were performed as described (34) and the total 3zP incorporation into newly transcribed RNA was estimated by scintillation counting. After hybridization to the nitrocellulose strips containing the linearized immobilized plasmids, the filters were washed twice in 2 X SSC for 1 h at 65 "C and single stranded unhybridized RNA was digested by RNase A. After one more wash in 2 X SSC at 37 "C for 1 h the filters were air-dried and exposed to x-ray film. The radioactive spots were excised from the membranes and counted.
Immunoprecipitation After three washes with PBS the Pansorbin-antibody-antigen complexes were pelleted at 14,000 X g. The pellets were resuspended in 45 p1 of PBS containing 1% SDS and incubated at room temperature for 10 min to release the antibodyantigen complexes from the Pansorbin. After centrifugation at 14,000 X g the supernatant was collected, 2 X sample buffer was added (1 X sample buffer, 62.5 mM Tris.Cl, pH 6.8, 10% (v/v) glycerol, 2% SDS (w/v), 715 mM 2-mercaptoethanol, 0.0025% (w/v) bromphenol blue), and the samples were subjected to SDS-polyacrylamide gel electrophoresis.

Western Blot Analysis
To cellular fractions or immunoprecipitates from Sertoli cells, 1 volume of 2 X sample buffer was added. The samples were heated to 100 "C for 5 min and subjected to electrophoresis in a 10% SDSpolyacrylamide gel. After electrophoresis the proteins were blotted onto an immobilon membrane which was blocked in a PBS solution containing 0.05% (v/v) Tween 20 and 5% (w/v) dry nonfat milk. K l l l antiserum at a dilution of 1/100 was incubated with the membrane in blocking buffer and afterwards rinsed three times with a PBS washing solution containing 1% (v/v) Nonidet P-40 and 0.2% (v/v) Tween 20. Finally, the membrane was incubated with '''1labeled protein A (0.5 pCi/ml) in blocking buffer, rinsed 5 times with PBS washing solution, and autoradiographed.

RESULTS
Sequence Comparison of RatPDE3 and RatPDE4"As previously reported (16), two groups of cDNA clones (ratPDE3 and ratPDE4) were isolated from a cDNA library derived from cultured 15-day-old Sprague-Dawley rat Sertoli cells that had been treated with 500 PM Bt,cAMP (19). The DNA sequence of ratPDE4 (data not shown, submitted to Gen-Bank) is compiled of two independent cDNA clones from the unamplified library. No in-frame termination codon 5' of the first ATG was found in the available ratPDE4 sequence. Therefore, it is not clear whether this first ATG is the correct start site. The sequence surrounding this ATG in ratPDE4.1 fits the Kozak consensus sequence since it has a purine in position -3 (35). The putative ratPDE4 major open reading frame therefore encodes a protein with a calculated molecular mass of 64.273 kDa or larger. The first three codons following the most 5' ATG encode the same amino acids around the putative initiation codon in the ratPDE3.1 cDNA (Fig. 1). After a short stretch of unrelated sequences at the amino terminus, the similarity between ratPDE3 and ratPDE4 resumes and continues for about 454 amino acids (453 for ratPDE3 and 455 for ratPDE4). In this region the sequence identity is 87%. The carboxyl termini of both putative proteins are less homologous although both are highly negatively charged ( Fig. 1) and contain sequences that fit the consensus sequence for phosphorylation by casein kinase I1 (36). Both putative proteins also contain a sequence of 4 consecutive basic amino acids, similar to the nuclear localization signal of FIG. 1. Comparison of the deduced amino acid sequences of ratPDE3 and ratPDE4 from the rat Sertoli cell. Amino acid sequences of the major open reading frames deduced from ratPDE3 (19) and ratPDE4 cDNAs were aligned. Amino acid residues are represented by the single letter code. Gaps have been introduced in the sequences to optimize the alignment. Regions of sequence identity have been boxed. Vertical arrowheads delineate the highly homologous region of 455 amino acids. The underlined amino acid residues represent the sequence of a synthetic peptide (K111) that was used to raise polyclonal antibodies. A putative nuclear translocation signal sequence is indicated by small arrows. Horizontal arrowheads delineate the region of ratPDE4 shown in Fig. 2. Putative sites for phosphorylation by casein kinase I1 are indicated by asterisks. the SV40 large T antigen (37) and other nuclear proteins (38)(39)(40)(41)(42).
A cDNA clone (DPD) similar to the ratPDE4 clones has also been isolated from rat brain (18). 525 codons of the open reading frame and the available 3"untranslated region of the ratPDE4 and DPD clones are identical (Fig. 2). The 5' portions of the putative open reading frames are unrelated (Fig.  2). The 5' portion of our ratPDE4 sequence was derived from two independent clones (ratPDE4.1 and ratPDE4.2) from a Sertoli cell unamplified library. In addition, the same sequence was present in a genomic clone derived from a rat cosmid library (data not shown), ruling out the possibility of cloning artifacts.
Characterization of the Recombinant PDE Activity Encoded by RatPDE4 and Comparison with RatPDE3-The high degree of homology between the ratPDE3 and the ratPDE4 sequences suggests that both enzymes have similar kinetics and physicochemical characteristics. T o verify this possibility, both PDEs were expressed in prokaryotic as well as in eukaryotic cells.
T o express the two cDNAs in E. coli, cDNAs were modified t o arrange the putative initiation methionine immediately after the ribosomal-binding site of the pRC23 vector. Transformation of E. coli DH5a with pRC23-ratPDE3 and pRC23-ratPDE4 led to the appearance of major immunoreactive bands (with the K l l l antibody, see "Experimental Procedures") of 71 and 68 kDa, respectively (data not shown). Less intense immunoreactive bands of lower molecular weight were often recovered, suggesting degradation of the recombinant protein. The immunoreactivity was associated with the appearance of high affinity PDE activity in the bacterial lysates. cAMP hydrolysis by both recombinant enzymes followed Michaelis-Menten kinetics and Lineweaver-Burke plots were linear with substrate concentrations ranging between 0.1 and 50 PM. The K , values for cAMP (Table I)    both PDEs (1.9 and 2.5 p~ cAMP for ratPDE3 and ratPDE4, respectively). Velocity determined in crude extracts ranged between 0.5 and 1 nmol/min/mg of protein (Table I). In addition, both recombinant enzymes displayed similar sensitivities to a variety of PDE inhibitors, rolipram and Ro 20-1724 being more effective inhibitors than cilostamide or milrinone (Table I).
Since expression in prokaryotic cells does not allow the same post-translational modifications as in eukaryotic cells, the two cDNAs were also expressed in monkey kidney (COS-7) cells. Affinities for cAMP of the two PDEs recovered from COS-7 cell transfection were very similar to those determined from the bacterial expression (ratPDE3 K,, = 1.81; ratPDE4 K,, = 2.85). However, although Northern and Western blot analysis indicated that transfection efficiency, transcription, and translation of the two constructs in COS-7 cells was comparable, differences were observed in the amount and in intracellular distribution of the CAMP-PDE activity recovered (PDE activity in picomoles/min . mg of protein ? S. E

Regulation of Steady State mRNA Levels for RatPDI33 and RatPDE4 in Sertoli Cells-Previous biochemical characterization of the Sertoli cell PDEs indicated that rolipram-sensitive
PDEs were present in this cell, mostly after hormonal stimulation (13). The cloning data suggested that mainly ratPDE3 and ratPDE4 were expressed in the Sertoli cell (16). T o further assess this possibility, Sertoli cell mRNA was hybridized to the PDE cDNA clones available. These Northern blot analyses indicated that ratPDE3 and ratPDE4 are the two dominant ratPDEs expressed in these cells, and that the cAMP analog treatment has major effects on mRNA levels. Sertoli cells from 15-day-old rats cultured for 4 days in defined medium expressed ratPDE3 mRNA levels that were at the limit of detection by Northern blot analysis (Figs. 3 and 4). After 24 h treatment with the cAMP analog Bt, CAMP, however, the ratPDE3 steady state transcript level is increased a t least 100-fold (19). To test if this is also the case After removal of the probe, the membrane was rehyhridized with ratPDE4 cDNA and consecutively with actin cDNA. Size markers were radiolabeled HindIII-digested X DNA fragments. FIG. 4. Dose dependence of the FSH stimulation of ratPDE3 a n d r a t P D E 4 s t e a d y state mRNA levels in the rat Sertoli cell. On the fourth day of culture in defined medium, primary Sertoli cells were rinsed and incubated for 24 h with increasing concentrations of FSH. Total RNA was prepared, separated on a 1% agarose gel, blotted onto a nylon membrane, and hybridized with :"P-labeled ratPDE3 cDNA and autoradiographed. After removal of the probe, the membrane was rehybridized with ratPDE4 cDNA and consecutively with actin cDNA. The mRNA abundance as determined by densitometric scanning of the autoradiograms are expressed in arbitrary units. for ratPDE4, the same Northern blots were hybridized with radiolabeled ratPDE4 cDNAs of similar specific activities as the ratPDE3 cDNAs and autoradiographed. The ratPDE4 steady state transcript level was readily detectable under basal conditions and was increased only about 5-fold as determined by spectrophotometric scanning (Fig. 3). These differences were confirmed in several other Northern blot experiments from independent Sertoli cell cultures and were also evident after FSH or forskolin treatment (Figs. 4 and 5, and data not shown). The increase in steady state levels of ratPDE3 and ratPDE4 mRNA by FSH is dose-dependent and occurs a t physiological concentrations of FSH ranging between 1 and 100 ng/ml. Whereas ratPDE4 mRNA levels were maximal upon exposure of the cells to 10 ng of FSH/ml of medium, ratPDE3 mRNA levels could be further elevated with higher FSH concentrations (Fig. 4).
When Sertoli cells were exposed to FSH, the AMP-PDE activity increases after a time lag of 1 h and reaches a maximum a t 18-24 h (15). The time course of changes in the ratPDE steady state mRNA levels in primary Sertoli cell cultures treated with 100 ng/ml FSH is reported in Fig. 5. After a time lag of about 1 h the level of transcripts increased, reached a maximum between 3 and 12 h, and then gradually decreased to near basal levels in a time span of 3 days. Confirming the above reported data, ratPDE4 transcripts were present under basal conditions, while ratPDE3 transcripts were not detectable under these experimental conditions.
Marked differences in hormone dependence of the two PDEs were also present when rats were treated with FSH in vivo. Whereas ratPDE3 mRNA levels undergo a large increase after FSH injection, no significant changes in ratPDE4 mRNA could be detected under these conditions (data not shown).
Regulation of Transcription of RatPDE3 and RatPDE4-To test whether the differences in regulation of steady state mRNA levels for ratPDE3 and ratPDE4 are caused by differences in mRNA stability or by differential regulation of gene transcription, nuclear run-off assays were performed. The rate of transcription of c-fos and ubiquitin was measured as positive and negative control for the FSH effects. In the three different experiments performed, Sertoli cell nuclei isolated prior to FSH treatment showed low levels of ratPDE3 and ratPDE4 transcription ( Table 11). The transcription rate of ratPDE3 increased earlier than 30 min after FSH treatment, reached a maximum a t 3 h, and returned towards basal within 24 h. The rate of transcription of ratPDE4 did not significantly change during the time period studied. Furthermore, dibutyryl CAMP, forskolin, and cholera toxin produced an increase in the ratPDE3 rate of transcription similar to FSH (Table 111). This indicates that hormone effects on PDE gene

TABLE I1
Time course of FSH regulation of ratPDE3 and ratPDE4 rate of transcription Nuclear run-off assays were performed as described under "Experimental Procedures." Five micrograms of plasmids containing rat-PDE3, ratPDE4, c-fos, or ubiquitin cDNAs were immobilized on nylon filters. Sertoli cells from 17-day-old Sprague-Dawley rats were treated with 500 ng/ml of FSH for the indicated lengths of time, nuclei were prepared, and incubated with [n-'"P]UTP. The total :'T incorporation into newly transcribed RNA was estimated by scintillation counting. After hybridization to the nylon strips containing the plasmids, the filters were washed and exposed to x-ray film. The radioactive spots were excised from the membranes and counted. These counts were divided by the total number of counts in the hybridization solution to express the rate of transcription in parts per million. The rate of transcription of c-fos and ubiquitin was measured as positive and negative control for the FSH effects.
Specific RNA concentration after FSH " ND, not detectable.

TABLE I11
Regulation of ratPDE3 and ratPDE4 rate of transcription by FSH, &CAMP, forskolin, and cholera toxin Nuclear run-off assays were performed as described in Table 11. Sertoli cells from 17-day-old Sprague-Dawley rats were incuhated in the presence or absence of 500 ng/ml FSH, 1 mM Bt,cAMP, 100 p M forskolin, or 100 ng/ml cholera toxin for 3 h.   6. Immunoblot analysis of soluble fractions ( A ) and partially purified CAMP-PDE preparations ( B ) from the rat Sertoli cell with K l l l antiserum. On the fourt,h day of culture in defined medium, primary Sertoli cells were incubated in the presence or absence of 1 mM Bt2cAMP for 24 h. The cells were scraped with a rubber policeman and homogenized in an all-glass Dounce homogenizer. After fractionation of the cell extract by centrifugation a t 14,000 X g for 15 min, the supernatant proteins ( A ) and a CAMP-PDE protein preparation, partially purified from the rat Sertoli celp ( R ) , were subjected to SDS-polyacrylamide gel electrophoresis. After blotting onto nylon membranes, duplicate blots were incubated with the K l l l antiserum (raised against a synthetic ratPDE peptide) or with the K l l l antiserum preincubated with excess peptide used for immunizations. Immunoreactive proteins were visualized by incubating the blots with "'1-labeled protein A and autoradiography. transcription are mediated by the activation of the CAMPdependent pathway.

Immobilized
Regulation of the RatPDE Protein Levels-To determine whether the increase in ratPDE mRNA results in an increase in PDE protein content, Sertoli cell extracts were subjected t o immunoblot analysis. A polyclonal antiserum (K111) was raised against a synthetic peptide based on the sequence of ratPDE3. This antiserum recognizes both recombinant rat-PDE3 and ratPDE4 proteins (data not shown).
Although immunoblot analysis of crude Sertoli cell extracts showed several bands, only the signals of a 67 and a 54-kDa polypeptide were abolished by competing the antibody with the peptide used for immunization (Fig. 6A). The 67-kDa protein level was markedly affected by Bt'cAMP treatment, being undetectable in extracts from unstimulated Sertoli cells (Fig.  6A). T o confirm that the 67-kDa protein corresponds to the CAMP-PDE, extracts from stimulated Sertoli cells were partially purified by high performance liquid chromatography' and the peak fraction of CAMP-PDE activity was subjected t o immunoblot analysis. Under these conditions only the immunoreactivity of a 67-kDa polypeptide was present and the signal was competed by the peptide (Fig. 6B). Using an alternative approach to increase the immunoblot signal, soluble extracts from control, FSH-, or dibutyryl CAMP-stimulated Sertoli cells were first immunoprecipitated with the K l l l antiserum, and the immunoprecipitated proteins were analyzed by immunoblotting. Even after enriching the PDE protein by immunoprecipitation, no detectable band could be observed under basal conditions (Fig. 7A). Conversely, an immunoreactive 67-kDa band appeared after FSH and dibutyryl cAMP treatment (Fig. 7A). The intensity of these bands correlated with the levels of CAMP-PDE activity measured in the supernatants before immunoprecipitation (Fig. 7B). Finally, the size of this band was identical to that of a purified CAMP-PDE from the Sertoli cell (Fig. 7A).'

DISCUSSION
Biochemical characterization and molecular cloning of the CAMP-PDEs has uncovered an unexpected complexity of this family of enzymes. In the rat there are at least four genes encoding cGMP-insensitive, rolipram-and Ro 20-1724-inhibited CAMP-PDEs (16), and data suggesting alternate splicing open the possibility to expression of an even larger number of PDE proteins (16,17,19). Transcripts derived from two of these genes (ratPDE3 and ratPDE4) are present in the rat Sertoli cell in culture. Although differences in basal levels and in the degree of hormone stimulation are present, the expression of both ratPDE3 and ratPDE4 is under the control of follicle-stimulating hormone and CAMP. The data presented here demonstrate that the increase in cAMP hydrolyzing activity observed in this cell is the result of an increase in rate of CAMP-PDE gene transcription, accumulation of mRNA, and neosynthesis of the CAMP-PDE protein. These On the fourth day of culture the cells were incubated in the presence or absence of 500 ng/ml FSH or 1 mM Bt2cAMP. Twenty-four hours after the additions the cells were scraped with a rubber policeman and homogenized in an all-glass Dounce homogenizer. This cell extract was fractionated by centrifugation a t 14,000 X g for 15 min. The supernatant fractions were subjected to PDE assays and immunoprecipitation and the anti-CAMP-PDE antibody K111. The immunoprecipitated proteins were solubilized and separated on a 10% polyacrylamide gel. CAMP-PDE protein purified from the rat Sertoli cell2 was loaded on the same gel. After blotting onto a nylon membrane, the blot was incubated with the K l l l antibody and ""I-labeled protein A and autoradiographed. data, then, provide evidence for a novel mechanism of regulation of PDEs that involves hormone-dependent CAMPmediated changes in the rate of a PDE gene transcription.
A molecular cloning approach has demonstrated that the Sertoli cell expresses ratPDE3 and ratPDE4. As previously shown (19), ratPDE3 encodes a protein with a calculated molecular mass of 66.2 kDa, while conceptual translation of the ratPDE4 cDNAs is consistent with a PDE protein of 64.2 kDa or higher molecular mass. Although no in-frame upstream termination codon was found in the cDNA clones that we have available, preliminary sequence available from a genomic clone demonstrates the presence of a stop codon 108 bases upstream from the putative initiation methionine. Nevertheless, further studies are needed to unequivocally prove that the ATG that we have selected codes for the initiation methionine also in uiuo. On the other hand, these CAMP-PDEs might have multiple initiation methionines (17). The ratPDE4 cDNAs isolated from the rat Sertoli cell cDNA library have sequences identical to the DPD clone isolated by complementation from a rat brain library except for the first 40 amino acids of the putative open reading frames (18). Although cloning artifacts cannot be completely excluded, it is possible that alternate splicing of a single ratPDE4 gene occurs in brain and testis. Evidence suggesting alternate splicing or multiple start sites have also been reported for other CAMP-PDE encoding cDNAs (17,19). In addition, most cell types studied express multiple transcripts of the ratPDE genes, some of which are tissue specific (16). There is, then, potential for the expression of a large number of rolipramsensitive CAMP-PDEs. That alternate splicing might be a property of PDEs is also suggested by data on other PDE families. It has been reported that two distinct amino termini have been found for a Ca2+/calmodulin-dependent PDE (43).
Based on the sequence similarity in the central portion of the proteins, it was expected that the kinetic properties of both enzymes are very similar. This was confirmed here by prokaryotic expression of the two PDEs. The two enzymes have a similar K,,, for cAMP and similar sensitivity to different inhibitors. Therefore, the catalytic centers are structurally and functionally very similar. It is possible that the amino and carboxyl termini, being different in the two proteins, serve different functions in the two CAMP-PDEs.
Both ratPDE3 and ratPDE4 transcript steady state levels were increased by FSH after a lag of about 1 h and remained elevated for at least 2 days. This is in agreement with the appearance of the cAMP hydrolyzing activity in the Sertoli cell supernatant (15). The increase in ratPDE3 mRNA levels in response to Bt,cAMP or FSH was consistently larger than that of ratPDE4. Nuclear run-off experiments showed that a 10-fold increase in the rate of ratPDE3 gene transcription followed FSH treatment in uitro. Conversely, the basal transcription rate of ratPDE4 was at the limit of detection and FSH activation of transcription was small and not significant. I t is, then, possible that also mRNA stabilization plays a role in the regulation of the steady state mRNA levels of these genes. The FSH effect on the ratPDE gene transcription is mimicked by cAMP analogs and other cAMP increasing agents, indicating that this hormonal activation of transcription is mediated by CAMP. Transcriptional activation of cAMP responsive genes such as c-fos is commonly thought to occur via CAMP-dependent phosphorylation of cAMP response element-binding proteins (44) or by regulating the levels of trans-acting factors that bind to the AP1 and/or AP2 elements (45,46). Similar regulatory sequences can therefore also be expected in the 5'-flanking region of the ratPDE3 gene. It is worth noting, however, that the time course of ratPDE3 gene transcription is slower than that observed for c-fos. The transcription of c-fos reaches a maximum at the 30-min time point, while ratPDE3 transcription peaks only after 3 h. It is then to be expected that the FSH-dependent CAMP-mediated transcription of the two genes involves different steps. Although the basic mechanism might be the same, e.g. phosphorylation of a cAMP response elementbinding protein, additional regulatory steps either delay the activation of ratPDE3 gene transcription, or cause a more rapid termination of c-fos transcription. The possibility should also be considered that the c-fos gene product might be involved in the modulation of transcription of the ratPDE3 gene. In fact, the c-fosljun complex is a transcription activator that binds to an AP-1 regulatory element (45), and hormonal activation of the Sertoli cell leads to increased levels of both fos andjun mRNAs (47).
The finding that cAMP regulates the expression of CAMP-PDEs explains earlier reports on cell lines with diminished protein kinase A and phosphodiesterase activity. The kinand PPD mutant strains of S49 lymphoma cells, for instance, have a decreased protein kinase A activity in comparison with the wild type cells. The kinase defect is associated with a decrease in cAMP hydrolytic activity, indicating that protein kinase A mediates, probably by regulating transcription factors, the expression of certain PDE forms (48,49).
The hypothesis that CAMP-PDE activation is dependent on new synthesis of the PDE protein is here proven by the use of polyclonal antibodies. It is, however, still possible that additional modifications such as phosphorylation are necessary to express the full activity of the PDE. One puzzling finding is that although ratPDE4 mRNA is expressed in the Sertoli cell under basal conditions, no immunoreactive PDE protein could be detected in the Sertoli cell soluble fraction. It is possible that ratPDE4 protein levels are below the sensitivity of the immunodetection employed in our study. An alternative possibility is that the ratPDE4 protein is not present in the soluble cell fraction, but compartmentalizes in other subcellular districts. This latter possibility would also reconcile differences found in the stimulated Sertoli cell. While high levels of both ratPDE3 and ratPDE4 mRNA are present under these culture conditions, only one immunoreactive band could be detected even after immunoprecipitation. Other reports have demonstrated the presence of a low K,,, CAMP-PDE in the particulate fraction of the brain (50). Consistent with this view also is the finding that rolipram binding to membrane-bound sites was detected in brain (51). Also in the rat liver, a membrane-bound PDE is sensitive to rolipram (52,53).
The time course of this CAMP-mediated activation of these CAMP-PDEs in the Sertoli cell is very different from that described for the CAMP-dependent activation of a cGMPinhibited PDE from human platelets (54)(55)(56) and rat adipocytes (57,58). This latter enzyme is membrane bound and very rapidly activated by phosphorylation. In a cell where the two modes of regulation coexist, there will be a complex pattern of PDE activation, with early and delayed increases in PDE activity. The exact significance of this complexity is unclear, but there is evidence that the long term PDE activation is involved in the regulation of the hormone responsiveness of the Sertoli cell (15). The homologous and heterologous desensitization of the rat Sertoli cell is in part or completely reverted after inhibition of the PDEs by Ro 20-1724 (13). This indicates that these CAMP-PDEs serve to control the responsiveness of the Sertoli cell to hormones. On the other hand, experiments with in uitro model systems containing hormone-sensitive adenylate cyclase, CAMP-de-pendent protein kinase 11, and cyclic nucleotide phosphodiesterase support the hypothesis that rapid cAMP turnover may function as a mechanism for signal amplification by the CAMP-dependent protein kinase (59). These hypotheses can now be tested by studying other hormone-responsive cell systems that express a different subset of CAMP-PDEs and by stable transfection of hormone-responsive cell lines with the available CAMP-PDE cDNAs.