Functional and Molecular Characterization of the Transcriptional Regulatory Region of the Proacrosin Gene*

F’roacrosin, the zymogen form of the serine protease acrosin, is located within the acrosomal vesicle of mam- malian spermatozoa and has been suggested to be involved in the fertilization process. In mouse and rat, expression of the promrosin gene starts in pachytene spermatocytes and continues through the early stages of spermiogenesis. We have shown recently that 2.3 kilo-base pairs of the 5‘-flanking region of the rat proacrosin gene is sufficient to direct chloramphenicol acetyltransferase gene expression in a germ cell-specific and devel- opmental stage-specific manner in the mouse. Addi- tional transgenic lines have been generated which include two deletions in the 5”flanking region and a tyrosinase minigene as marker for gene expression. Transgenic mice bearing these two truncated fragments showed different patterns of reporter gene expression. Transgenic lines (BM, B2) harboring the 397-base pair (bp) fragment (from 45 to 442 bp upstream of ATG) showed no chloramphenicol acetyltransferase (CAT) ac- tivity in either testis but analysis via reverse transcription polymerase chain reaction con- firmed low levels of reporter gene transcription in testis. line longer 877 (from 45 922 showed a reporter gene expression and chloramphenicol acetyl- transferase for 25 min on ice. The protein-DNAcomplexes were separated from the free DNAprobe by electrophoresis on a nondenatured polyacrylamide gel. The gel was dried and autoradiographed. Indirect Immunofluorescence Method-Testis cell smears of mature transgenic mice of the line TC were prepared according to Florke et al. (1983) and fixed in acetone/methanol (1:l). Indirect immunofluorescence was performed using the reagents and suggested protocols of an immunofluorescence kit from ABCR (Karlsruhe, Germany). A rabbit polyclonal antibody raised against proacrosidacrosin from boar was generously provided by Dr. Topfer-Petersen (Andrological Department, University of Hannover).

F'roacrosin, the zymogen form of the serine protease acrosin, is located within the acrosomal vesicle of mammalian spermatozoa and has been suggested to be involved in the fertilization process. In mouse and rat, expression of the promrosin gene starts in pachytene spermatocytes and continues through the early stages of spermiogenesis. W e have shown recently that 2.3 kilobase pairs of the 5'-flanking region of the rat proacrosin gene is sufficient to direct chloramphenicol acetyltransferase gene expression in a germ cell-specific and developmental stage-specific manner in the mouse. Additional transgenic lines have been generated which include two deletions in the 5"flanking region and a tyrosinase minigene as marker for gene expression. Transgenic mice bearing these two truncated fragments showed different patterns of reporter gene expression. Transgenic lines (BM, B3, B2) harboring the 397-base pair (bp) fragment (from 45 to 442 bp upstream of ATG) showed no chloramphenicol acetyltransferase (CAT) activity in either testis or other tissues, but analysis via reverse transcription polymerase chain reaction confirmed low levels of reporter gene transcription in testis. Transgenic line TC bearing a longer fragment of 877 bp (from 45 to 922 bp upstream of ATG) showed a reporter gene expression and chloramphenicol acetyltransferase enzyme activity which was identical to that found in mice harboring the 2.3-kilobase pair 5"flanking region. The analysis of the CAT gene expression during testicular development showed diploid transcription and haploid translation. It can be concluded that all sequences required for a basic level of testis-specific transcription of transgene are present within the 397-bp fragment, and other DNA sequences located outside of the 397-bp fragment but present within the 877-bp fragment can function as enhancer elements. Two fragments within the 877-bp region were identified by gel retardation assays as binding exclusively to nuclear factor(s) from testis protein extracts. In both fragments we identified sequence elements which are present in the promoter region of the germ cell-specific genes for histone H2B and protamine I , respectively.
During spermatogenesis, mitotically dividing spermatogonia are first transformed into spermatocytes undergoing meiosis and subsequently into spermatozoa carrying a haploid nuclear content (Handel, 1987). The analysis of genes expressed in a restricted temporal and spatial manner during spermatogenesis (En 84/20-1). The costs of publication of this article were defrayed in * This work was supported by the Deutsche Forschungsgemeinschaft part by the payment of page charges. This article must therefore be hereby marked "uduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
j: To whom correspondence should be addressed: Institut fur Humangenetik, Goplerstrape 12d,  has given some insight into different gene-regulatory mechanisms active in male germ cells (Willison and Ashworth, 1987). Most studies for understanding the mechanisms regulating and coordinating gene expression during spermatogenesis resort to the use of transgenic mice or DNA-protein-interaction experiments. Studies of the promoter region of the protamine 1 (Zambrowicz et al., 1993), protamine 2 (Stewart et al., 19881, the phosphoglycerate kinase-2 (Robinson et al., 1989;Gebara and McCarrey, 1992) and the testis-specificH2B histone genes (Choi and Chae, 1991) have revealed DNAsequence motifs controlling transcriptional activity during spermatogenesis.
An important gene specifically expressed during mammalian spermatogenesis is that for the serine protease acrosin (EC 3.4.21.10). The enzyme is synthesized in a zymogen form, proacrosin, which is activated to the mature enzyme during the acrosome reaction. Acrosin has been implicated in the recognition, binding, and penetration of the zona pellucida of the ovum (Klemm et al., 1991). The expression of the proacrosin gene is under translational control. While transcription of the proacrosin gene occurs in diploid spermatogenic cells (Kashiwabara et al., 1990;Kremling et al., 1991a), translation only occurs in haploid germ cells (Florke et al., 1983;Kallajoki et al., 1986). Earlier, we reported that 2.3 kb' of the 5"flanking region of the rat proacrosin gene was sufficient to direct transcription of the bacterial chloramphenicol acetyltransferase (CAT) reporter gene in pachytene spermatocytes and translation in haploid spermatids (Nayernia et al., 1992). In this report we describe experiments that used chimeric constructs containing shorter 5"flanking regions of the rat proacrosin gene (397 and 877 bp upstream fragments) in order to define more narrowly the sequences required for the regulation of proacrosin gene expression. Two DNA fragments in the 877-bp flanking region were found to bind nuclear proteins of germ cells in gel retardation assay. This work is a step toward the identification and characterization of transcription factors critical for male germ cellspecific expression of the proacrosin gene.
EXPERIMENTAL PROCEDURES Construction of the Fusion Genes-The fusion gene constructs used for microinjection are shown in Fig. 1. The construct prACRIII-CAT was cleaved from plasmid prACRII-pBLCAT3 (Nayernia et ul., 1992). This construct contains the rat proacrosin 5"flanking region from 45 to 922 bp upstream of the translation start site, the bacterial gene for CAT, and SV40 splice and polyadenylation sequences (Luckow and Schiitz, 1987). The 5"truncated construct prACRIV-CAT was obtained by cleaving the plasmid prACRII-pBLCAT3 with BstEIUSstI restriction enzymes. This construct contains the rat proacrosin 5"flanking region from 45 to 442 bp upstream of the translation start site. Both constructs were purified with the Geneclean Kit (BIO 101 Inc., La Jolla, CA) before injection. The constructs were coinjected with a mouse tyrosinase gene construct ptrTYFL5. The tyrosinase minigene was used as a marker and leads to pigmentation in the eyes and skin in the transgenic albino mouse strain NMRI (Beermann et al., 1991).
The abbreviations used are: kb, kilobase(s); bp, base paids); CAT, chloramphenicol acetyltransferase; RT-PCR, reverse transcriptase-polymerase chain reaction; CRE, CAMP responsive element. Production and Characterization of Dansgenic Mice-Fertilized oocytes were obtained from superovulated NMRI females mated with NMRI males. 3-pg ml-I aliquots of a DNA solution in 10 mM Tris-HC1, pH 8.0, 1.5 mM EDTA were injected into male pronuclei using an Eppendorf micromanipulator (Eppendorf, Hamburg, Germany). The microinjected zygotes were then transferred into oviducts of NMRI pseudopregnant foster mothers according to published procedures (Hogan et al., 1986). For preliminary identification of transgenic animals, the dot or Southern blot analyses were performed using 10 pg of DNA prepared from the tail tissues of the pups at the age of 3-4 weeks and :"P-labeled 1.6-kb CAT and tyrosinase gene fragments as probes. Copy number of the proacrosin-CAT transgene was estimated by comparing the blot intensity of CAT hybridization with that of known standards.

32181
Cell Separation-Testes from mature transgenic mice were dissected and used for the preparation of testicular cell suspension by the collagenase/trypsin method of Romrell et al. (1976). Cell suspension from the testes of 7 animals was loaded onto a 1000-ml 2-496 bovine serum albumin gradient in a CELSEP chamber (Depart Inc., Washington, D.C.) with a 10% bovine serum albumin cushion at the bottom. 50-ml fractions were collected. Purity of the 16 collected fractions was analyzed by phase-contrast microscopy. The highly enriched fractions with pachytene spermatocytes, round spermatids, and elongating spermatids were used for the experiments.
RNA Isolation and Analysis-Total RNA was isolated from different tissues or cell populations enriched with specific spermatogenic cell types using RNaid-Kit, according to the manufacturer's instructions (Dianova, Hamburg, Germany).
Genomic DNA contamination was eliminated from RNAby DNase I treatment. RT-PCR was carried out by the rTth Kit from Perkin-Elmer (Vatterstetten, Germany) and run in a Perkin Elmer Thermal Cycler. "No template" and "no reverse transcriptase'' controls were run for each experiment to rule out template contamination in reaction components or DNA contamination in RNA samples, respectively, as a source of positive amplification signals. 200-400 ng of total RNA was reverse transcribed into cDNA at 72 "C for 10 min. The amplification profile involved 2 min at 95 "C for 1 cycle, 1 min a t 95 "C, and 1 min at 60 "C each for 25-35 cycles, and 7 min a t 60 "C for 1 cycle. To test the RNAin each sample, the products of each reverse transcriptase reaction were divided into two aliquots. One aliquot was then subjected to 25 cycles of PCR to amplify a 230-bp fragment of mouse p-actin cDNA using 22-mer primers (upstream primer: 5'-GGACGACATGGAGAAGATCTGG-3'; downstream primer: 5"CTCCG-GAGTCCATCACAATGCC-3'). The other aliquot was subjected to 35 cycles of PCR to amplify a 380-bp fragment of the CAT coding region using 20-mer primers (upstream primer: 5"CGTTCAGCTGGATAT-TACGG-3'; downstream primer: 5'-GTTGTCCATATTGGCCACGT-3').
The PCR reaction products were run on a 2% agarose gel, blotted to nitrocellulose filter, and hybridized with a 32P-labeled CAT DNA probe. For Northern blot analysis 20-pg RNA samples were denatured in a formaldehyde/formamide buffer and electrophoresed on 1.5% agarose gels. RNA was then transferred to nitrocellulose and probed with a 32P-labeled CAT probe. For the proof of intact RNA in each lane, Northern blots were stripped and rehybridized to actin cDNAprobe (Hanauer et al., 1983).
CAT Enzyme Assay-Protein extracts for CAT assays were made by freeze-thawing tissue homogenates or spermatogenic cells in 0.25 M Tris-HC1, pH 7.8. Protein concentrations of extracts were measured using the Bio-Rad protein assay kit (Bio-Rad). After heat inactivation for 10 min at 65 "C, CAT enzymatic activity was assayed according to Gorman et al. (1982). To quantitate the CAT activity in individual samples, signals on the autoradiograph corresponding to the acetylated forms of chloramphenicol were scraped from the silica gel plate and counted by liquid scintillation counter.
Gel Retardation Assay-Nuclear extracts were prepared from rat testis and brain (Tamura et al., 1989) and from highly enriched spermatocytes and spermatids (Dignam et al., 1983). 5-10 pg of each protein extract was incubated in a total volume of 20 p1 with 2-10 pg of poly(d1,dC) as a nonspecific inhibitor, 10 mM Tris-HC1, pH 7.5, 50 mM NaCl, 1 mM EDTA, 5 mM MgCl,, 5% glycerol for 15 min on ice. A 1.7-kb CAT DNA fragment was used as a nonspecific competitor. Different nonlabeled double-stranded oligonucleotides were used as specific competitor for precise identification of the binding site(s) of nuclear factor(s). 10,000 cpm of a 32P-end labeled DNA fragment (A and B in Fig.   1) was added to the reaction mixture and incubated for 25 min on ice. The protein-DNAcomplexes were separated from the free DNAprobe by electrophoresis on a nondenatured polyacrylamide gel. The gel was dried and autoradiographed.
Indirect Immunofluorescence Method-Testis cell smears of mature transgenic mice of the line TC were prepared according to Florke et al. (1983) and fixed in acetone/methanol (1:l). Indirect immunofluorescence was performed using the reagents and suggested protocols of a n immunofluorescence kit from ABCR (Karlsruhe, Germany). A rabbit polyclonal antibody raised against proacrosidacrosin from boar was generously provided by Dr. Topfer-Petersen (Andrological Department, University of Hannover).

Generation of Transgenic Mice Containing prACRIII-CAT and prACRIV-CAT Fusion
Genes-The hybrid constructs were coinjected with a 11.2-kb tyrosinase minigene ptrTYR5 into fertilized eggs of the albino mouse strain NMRI. We have used the tyrosinase minigene as a marker for visual identification of transgenic mice and as a control of the position effect on transgene expression (Beermann et al., 1991). Three transgenic mice (All, A32, TC) for the prACRIII-CAT construct (877-bp 5' region) were identified from a total of 36 offspring using a CAT DNAprobe. These transgenic mice contained 2,8, and 20 copies of the transgene, respectively. One of the founder animals (All) produced no transgenic F1 offspring, suggesting that this founder was mosaic with respect to germ line integration. DNA analyses of F1 animals of founder mice A32 and TC showed that the coinjected DNAs (prACRII1-CAT and tyrosinase minigene ptrTYR5) are integrated in the same chromosomal sites. Neither the transgene prACRIII-CAT nor ptrTYR5 were expressed in transgenic mouse line A32. One possible explanation for this result could be integration into an inactive region of the genome. The transgenic mouse TC was pigmented, indicating that this mouse expresses the ptrTYR5 minigene. Founder mouse TC was bred to nontransgenic NMRI mice and the offspring screened for cotransmission of the injected DNAs. These screenings indicated that all pigmented F1 offspring were transgenic for both ptrTYR5 minigene and prACRIII-CAT.
Of 24 animals generated by coinjection of prACRIV-CAT (397-bp 5' region) and ptrTYR5, dot blot analysis of tail DNA revealed that three pigmented mice (B2, BM, B3) contained 3, 8, and 14 copies of prACRIV-CAT, respectively. Progeny analysis of transgenic mice B3 and BM indicated that both transgenes are integrated in the same chromosomal sites. The transgenic mouse B2 was infertile.
Expression Pattern of the prACRIII-CAT Fusion Gene-CAT assays were conducted on extracts of different organs of adult TC transgenic mice and CAT activity was found only in testis extract (Fig. 2). To determine if the testis-specific expression of the CAT gene is regulated at the transcriptional level, we performed a Northern blot analysis with RNA from a variety of tissues. A CAT transcript was detected only in testicular tissue, indicating apparent testis-specific transcription of the CAT gene (Fig. 3).
To identify the spermatogenic cell type in which translation of the transgene mRNA occurs, we have performed CAT assays with extracts from highly enriched populations of specific spermatogenic cell types. We first detected the CAT enzymatic activity in round spermatids with increasing of the CAT activity during spermatid differentiation (Fig. 4). The developmental regulation of the prACRII1-CAT fusion gene was assayed by monitoring the appearance of CAT transcript and CAT enzyme activity in testes of transgenic line TC between days 11 and 40 of postnatal development. The first CAT transcript was detected in RNA from testes a t day 15-17 ( Fig. 5) with increasing transcriptional activity during testicular development. CAT activity was first found in extracts from testes a t day 20 (Fig. 6).
The CAT activity remains constant a t a low level in testicular extract of 20-, 21-, and 23-day-old mice. Thereafter, from day 25, an increasing of CAT enzyme activity was observed. Day 15-17, 20, and 25 coincide with the appearance of pachytene spermatocytes, round spermatids, and elongating spermatids, respectively (BellvB, 1979). These results demonstrate that the transgene prACRII1-CAT is transcriptionaly active and that mRNA is subjected to post-transcriptional regulation.
We have used the indirect immunofluorescence procedure to localize the proacrosidacrosin and CAT protein in spermatogenic cells of the transgenic line TC. Proacrosidacrosin could be localized in the acrosome of spermatids and sperms (Fig.  7A). Using the antibody against CAT protein, the whole cytoplasm of the spermatids and cytoplasmic droplets of testicular sperms stained positive (Fig. 7C).
Expression Pattern of the prACRN-CAT Fusion Gene-We also performed a CAT assay on extracts from different tissues of the transgenic lines BM, B3, and B2 which contain the 397-bp fragment construct prACRIV-CAT. None of the three lines exhibited CAT activity in any tissue assayed. To increase the specificity of the CAT assay we have measured the CAT enzyme activity in extracts from specific spermatogenic cells of the transgenic lines BM and B3. No CAT activity was found in any cell type assayed. We observed no CAT hybridization in any of the three transgenic mice with Northern blot analysis (Fig.  3). By contrast, a low level of specific transcription of the CAT gene was obtained by RT-PCR analysis in testicular RNA and in RNA isolated from spermatocytes and spermatids (Figs. 8 and 91, with an increase of transcriptional activity in spermatids. These results indicate that fusion gene prACRIV-CAT lacks the regulatory elements necessary for the efficient expression of the CAT gene.

A
Sequences in the Promoter of Proacrosin Gene That Bind Nuclear Factor(s)-Expression analysis of transgenic mice showed that the 877-bp 5"flanking region contains the sequence information necessary for efficient tissue-specific expression of the transgene. To search for potential regulatory elements which bind specific nuclear factors, gel retardation assays were performed. Various 5' fragments of the 877-bp 5"flanking region of the proacrosin gene were assayed. Specific binding with testicular extracts and extracts from spermatocytes-and spermatid-enriched fractions were observed with 200-bp fragment BstELL-StyZ (A in Fig. 1)) and 150-bp fragment XhoI-BstEII ( B in Fig. 1). The BstEII-Sty1 fragment (Fig.  1OA) formed one complex with proteins from testis. The same complex was identified after incubation of this DNA fragment with nuclear extracts from separated germ cells. This retarded band could be eliminated only by a 50-fold excess of the unlabeled oligonucleotide OAl (Fig. 1OA). OAl contains the sequence element ACGTCA which is found in the promoter region of testis-specific histone H2B (Choi and Chae, 1991). This complex could not be obtained with nuclear extract from the so- matic tissue. The gel retardation assay with the XhoI-BstEII fragment yielded four retarded bands (a, b, c, and d) with nuclear protein from testis and nuclear extracts from separated pachytene spermatocytes and spermatids, but not with nuclear proteins from brain (Fig. 10B). The retarded bands could be eliminated completely or partly by a 50-fold excess of the unlabeled oligonucleotide OBl,OB2,OB3, and OB5 (Fig. 1OB). In oligonucleotides OB1 and OB2 we have found the sequence AACTTCAAAA, which has an 80% homology to the binding site of testis-specific transcription factor Tet-1 (Tamura et al., 1992). We have not found these DNA-protein complexes with nuclear extracts from cultured Sertoli and HeLa cells (data not shown). These results suggest that these complexes represent DNA-protein interactions with germ cell-specific nuclear factors. At present, it is unknown whether the multiple retarded bands formed by the XhoI-BstEII fragment represent the binding of different proteins or binding of the same protein as monomeric and multimeric units.

DISCUSSION
In previous studies we have generated transgenic mice bearing the fusion gene prACRII-CAT. This fusion gene contained 2.3 kb of the 5"flanking region of the rat proacrosin gene and was capable of targeting CAT reporter gene expression exclusively in mouse testicular germ cells (Nayernia et al., 1992). CAT gene was first transcribed in pachytene spermatocytes while enzyme activity was first detected in isolated round spermatids. The mRNAs for the proacrosin-CAT transgene and the endogenous mouse proacrosin gene were found for the first time in the testis of 17-day-old mice but the CAT protein was first observed in testis of 21-day-old mice. This correlates with the appearance of pachytene spermatocytes and early spermatids (Bellve, 19791, respectively, which is in accordance with the reported onset of transcription and translation of proacrosin gene during spermatogenesis (Florke et al., 1983;Kashiwabara et al., 1990). The results demonstrate that the spatial and temporal expression of the prACRI1-CAT fusion gene mimics the expression of the endogenous mouse proacrosin gene.
The studies described here represent a detailed dissection of the proacrosin promoter region. Two truncated fusion genes, prACRII1-CAT and prACRIV-CAT (Fig. l), were generated in consideration of the transcription start points of rat and mouse proacrosin gene. Kremling et al. (1991aKremling et al. ( , 1991b identified one transcription initiation site in rat and mouse a t positions 564 and 581 bp upstream of ATG, respectively, by primer extension analysis. In contrast, Watanabe et al. (1991) have reported that the mouse proacrosin gene has five heterogeneous transcription initiation sites (6, 11, 16,21, and 30 bp) upstream from the ATG start codon. Using the more sensitive method of rapid amplification of cDNA ends-RCR we have found two transcription initiation sites about 100 and 700 bp upstream from ATG in mouse and rat.2 The fusion genes prACRII1-CAT and prACRIV-CAT contain 877 and 397 bp of the 5"flanking region of the rat proacrosin gene, respectively. Therefore the construct prACRIV-CAT lacks the putative transcription start point which is located about 700 bp upstream of ATG and the construct prACRII1-CAT contains all suggested transcription start points, except those reported by Watanabe et al. (1991).
We have generated three transgenic mice (All, A32, and TC) with the construct prACRII1-CAT and three transgenic mice (B2, BM, and B3) with the construct prACRIV-CAT. These chimeric genes were coinjected with a tyrosinase minigene ptrTYR5. The tyrosinase minigene, when introduced into an albino mouse strain leads to pigmentation in eyes and skin with high penetrance. DNA analyses of F1 transgenic mice revealed that the tyrosinase minigene and the proacrosin-CAT fusion genes are integrated in the same chromosomal site in transgenic linesA32, TC, BM, B2, and B3. The transgenic mice TC, BM, B2, and B3 are pigmented. In spite of cointegration of ptrTYR5 and prACRII1-CAT in the genome of transgenic mouse A32, this mouse was not pigmented. This is an indication for the integration of both transgenes in an inactive chromosomal site.
Expression analysis of the fusion genes in transgenic mice serves to define the extent of the sequences that encode transcriptional control elements. The expression pattern of the prACRIII-CAT fusion gene in transgenic mouse line TC revealed that the 877-bp fragment contains information essential for directing male germ cell-specific and developmentally regulated expression. The expression pattern of prACRII1-CAT (877-bp 5' region) compared with that of prACRII-CAT (2.3-kb 5' region) in transgenic mice suggests that sequences located upstream of the 922-bp fragment are not essential for proacrosin gene expression. None of the three pigmented transgenic mice containing prACRIV-CAT (397-bp 5' region) fusion gene exhibited CAT activity in any tissue assayed. We only found a testis-specific transcription detectable by RT-PCR. These results indicate that the 387-bp fragment contains sequence elements that are probably essential for basal expression of the CAT gene in testis of transgenic mice. However, these elements clearly require the sequences located between 442 and 992 bp upstream ofATG for efficient expression of the CAT gene. Interestingly in mouse protamine 1 gene the region between 150 and -37 bp is sufficient to direct gene transcription while sequences 5' of 150 bp are required for high level transcription (Zambrowicz et al., 1993).
It is known that the sequences proximal to the transcription initiation site normally control the position and frequency of initiation by RNA polymerase I1 in eukaryotic promoters. One of the sequences necessary and sufficient for proper initiation by RNApolymerase I1 is the TATAbox (Latchman, 1992). Kremling et al. (1991aKremling et al. ( , 1991b) detected a putative TATA box at position 607 bp upstream of the ATG translation start codon in mouse and 588 bp in the rat. In our studies with transgenic mice we have found a basal level of transcription with the construct prACRIV-CAT which contains only a 387-bp 5"flanking region and lacks any typical TATA box. Many genes regulated developmentally or during cell differentiation do not contain a TATA box and transcription initiation occurs at any one of a number of start sites. The testis-specific genes lacking TATA sequences include those for farnesyl pyrophosphate synthetase, metallothionein (Salehi-Ashtiani et al., 1993), mouse pgk-2 (Boer et al., 19871, cytochrome ct (Virbasius and Scarpulla, 19881, and rat proenkephalin (Zinn et al., 1991). As is the case in these genes, it is possible that an initiator element present in the 387-bp fragment regulates the transcriptional initiation of the CAT gene in transgenic mice lines BM, B3, and B2. It can be suggested that genes expressed in male germ cells are activated by a common regulatory or signaling mechanism, possibly involving identical transcriptions factors. If this is the case, germ cell-specific genes would be expected to share common DNAbinding sites for such factors. The genes for Zfp-35 (CunlifFe et al., 1990), PGk-2 (Robinson et al., 1989), Tcp-lObt (Ewulonu et al., 1993), histone H2B (Choi and Chae, 1991), and histone H l t (Kremer and Kistler, 1992) have similar expression patterns as that of the rat and mouse proacrosin gene. The transcription of these genes begins in pachytene spermatocytes. The 5"flanking region of these genes and the rat proacrosin gene share similar sequence elements. One of these sequence elements that is highly conserved between many testis-specific genes is the 8-bp motif TGAGGTCA.
This motif is homologous to the CRE ( C A M P responsive element) consensus sequence TGACGTCA (Roesler et al., 1988). Recently, a novel CREM (CRE modulator) isoform, C R E q , was discovered in adult testis (Foulkes et al., 1992). C R E q activates transcription in response to CAMP. Premeiotic cells express low levels of the CREM gene in the antagonist form. During the pachytene stage, a switch in splicing pattern results in the exclusive production of high levels of C R E q . I t is possible that this factor is involved in the activation of gene expression in pachytene spermatocytes and also in proacrosin gene regulation. A reverse complementary sequence to the motif TGAGGTCA is present in fragment A. Fragment A which is located in the 442-bp 5"flanking region and fragment B which is located in the region between 442 and 922 bp upstream of ATG were found to bind specifically germ cell nuclear proteins in an electrophoretic mobility shift assay (Fig. 10). This results together with the observation that only basal expression of the construct prACRIV-CAT indicates that the cooperation of sequence elements in both regions (A and B fragment in Fig. 1) is required for appropriate temporal and germ cell-specific expression of rat proacrosin gene. The promoter studies in transgenic mice, together with sequence comparisons of promoter regions of genes that have a similar pattern of expression to the proacrosin gene, led to the identification of potential regulatory sequences. Fragment A was found to form a DNA-protein complex with nuclear extracts from germ cells, which could be eliminated with the oligonucleotide OAl (Fig. 1OA). The oligonucleotide OAl contains the sequence ACGTCA, which is present as the binding site of a testis-specific transcription factor in the promoter region of histone H2B (Choi and Chae, 1991). Both, the proacrosin and the histone H2B genes are first transcribed in pachytene spermatocytes. In fragment B we have found the sequence AACTTCAAAA which is identical in 8 of 10 positions to the sequence GACTTCATAA. This sequence was found in the promoter of the mouse protamine 1 gene as the binding site for the testis-specific trans-acting factor Tet-1 (Tamura et al., 1992). Further detailed deletion analysis using a transgenic animal system is necessary to functionally dissect the proacrosin gene promoter structure and identify discrete elements and factors involved in germ cell-specific regulation.