| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
Departments of Pathology,3 Molecular and Human Genetics,4 and Molecular and Cellular Biology,5 Baylor College of Medicine, Houston, Texas 77030
ABSTRACT
Growth differentiation factor 9 (GDF9) is preferentially expressed in oocytes and is essential for female fertility. To identify regulatory elements that confer high-level expression of GDF9 in the ovary but repression in other tissues, we generated transgenic mice in which regions of the Gdf9 locus were fused to reporter genes. Two transgenes (–10.7/+5.6mGdf9-GFP) and (–3.3/+5.6mGdf9-GFP) that contained sequences either 10.7 or 3.3 kb upstream and 5.6 kb downstream of the Gdf9 initiation codon demonstrated expression specifically in oocytes, thereby mimicking endogenous Gdf9 expression. In contrast, transgenes –10.7mGdf9-Luc and –3.3mGdf9-Luc, which lacked the downstream 5.6-kb region, demonstrated reporter expression not only in oocytes but also high expression in male germ cells. This suggests that the downstream 5.6-kb sequence contains a testis-specific repressor element and that 3.3 kb of 5'-flanking sequence contains all the cis-acting elements for directing high expression of Gdf9 to female (and male) germ cells. To define sequences responsible for oocyte expression of Gdf9, we analyzed sequences of Gdf9 genes from 16 mammalian species. The approximately 400 proximal base pairs upstream of these Gdf9 genes are highly conserved and contain a perfectly conserved E-box (CAGCTG) sequence. When this 400-bp region was placed upstream of a luciferase reporter (–0.4mGdf9-Luc), oocyte-specific expression was observed. However, a similar transgene construct (–0.4MUT-mGdf9-Luc) with a mutation in the E-box abolished oocyte expression. Likewise, the presence of an E-box mutation in a longer construct (–3.3MUT-mGdf9-Luc) abolished expression in the ovary but not in the testis. These observations indicate that the E-box is a key regulatory sequence for Gdf9 expression in the ovary.
evolutionary conservation, gene regulation, oocyte-specific
INTRODUCTION
The critical importance of the oocyte-secreted protein growth differentiation factor 9 (GDF9) in ovarian function has been established by both in vivo and in vitro studies. Female mice lacking GDF9 are infertile because of a block at the primary follicle stage in the ovary, but Gdf9 knockout males are normal [1]. Furthermore, our laboratory [2–4] and others [5–8] have shown that recombinant GDF9 enhances granulosa cell growth during folliculogenesis and regulates key factors and proteins (e.g., kit ligand, cyclooxygenase 2, hyaluranon synthase 2, LH receptor, urokinase plasminogen activator, pentraxin 3, prostaglandins, and progesterone) in early folliculogenesis and the periovulatory period. Thus, appropriate expression of GDF9 is critical for maintaining normal ovarian functions in vivo.
The Gdf9 mRNA and GDF9 protein are highly expressed in oocytes of the ovary [2, 9, 10]. Gdf9 was thought to be expressed exclusively in the ovary, until it was found to be expressed at much lower levels in the testis and hypothalamus as well [11]. The function of this low-level expression of Gdf9 in these other tissues is unclear, because Gdf9 knockout males are fertile and show no gross physical or behavioral defects [1]. In the mouse ovary, Gdf9 mRNA and its protein are not expressed in primordial (quiescent) follicles (types 1 and 2) but are expressed beginning at the early primary follicle (type 3a) stage through ovulation. Among the few identified oocyte-specific genes, Gdf9 shares an identical expression pattern in the ovary with the zona pellucida (Zp) genes. Mouse Gdf9 and mouse Zp mRNA are expressed in type 3a follicles and at high levels in larger growing oocytes [10, 12]. In both Zp3 and Gdf9 promoters, an E-box consensus sequence (CANNTG) is present [13]. The E-box sequences have been identified in several tissue-specific genes and shown to be involved in tissue-specific expression by binding heterodimers of helix-loop-helix transcription factors [14–17]. An in vitro study of the mouse Zp3 promoter identified an E-box (CACGTG) consensus sequence at 186 bp upstream of the transcription start site and showed that the E-box was necessary and sufficient for directing luciferase expression in microinjected oocytes [13]. In the studies herein, alignment of the proximal promoters of the mouse, rat, dog, sheep, chimpanzee, and human GDF9 genes revealed a perfectly conserved E-box sequence (CAGCTG). Given the similarity of the gene expression patterns in the ovary of Gdf9 and Zp genes and the similar E-box locations in the promoters of mouse Zp3 and Gdf9 genes, we hypothesized that the conserved E-box may play a role in the oocyte-specific expression of Gdf9 in the ovary.
Oocyte-expressed genes are difficult to study in vitro for several reasons. As a result of the unavailability of oocyte cell lines (because oocytes do not undergo mitosis), simple transfection into a cell line is not an option for studying oocyte-expressed genes in vitro. Additionally, the protective zona pellucida of freshly harvested oocytes prevents utilization of standard transfection modalities (e.g., lipofection and electroporation). Oocyte microinjection is an alternative for the transient expression of oocyte genes and has been used successfully [13]. However, because of the difficulty of maintaining the viability of oocytes in vitro and the limitation of the in vitro experiments performed in these single cells, such studies cannot address whether the gene expression is oocyte-specific (i.e., cannot define the sequences that restrict expression to the ovary versus other tissues). Therefore, we decided to pursue a transgenic approach. Although the generation of transgenic animals requires a significantly longer time, the transgene approach defines a broader scope of the expression at intraovarian and extraovarian sites. Here, we report successful use of the transgene strategy to study systematically the promoter regions of mouse Gdf9 in vivo.
MATERIALS AND METHODS
Chemicals and Reagents
Unless otherwise indicated, all chemicals and reagents were obtained from Sigma and Fisher. Unless otherwise specified, restriction enzymes were obtained from New England Biolabs.
Construction of Gdf9 Promoter-Reporter Gene Constructs
A full 16.3 kb of mGdf9 genomic sequence from a phage clone, including 10.7 kb of 5'-flanking region of mGdf9 and approximately 5.6 kb downstream of the transcription site of the gene, including exon 1 (397 bp), intron 1 (2.8 kb), exon 2 (1285 bp), and approximately 1 kb of the 3' downstream region, was subcloned into pBluescript SK vector (pB/S4). The initiation codon ATG of mGdf9 in pB/S4 was altered by site-directed mutagenesis into a BamHI site (pE22). The green fluorescent protein (GFP) plasmid was a kind gift from Dr. Roger Tsien (Department of Pharmacology, Department of Chemistry and Biochemistry, and Howard Hughes Medical Institute, University of California at San Diego). To generate the transgene –10.7/+5.6mGdf9-GFP, 714 bp of the coding region of GFP without its own polyadenylation site were inserted into the BamHI site of the pE22. With the convenient restriction enzyme digestion, sequences between –10.7 and –3.3 kb of the transgene –10.7/+5.6mGdf9-GFP were deleted and religated to generate new transgene, –3.3/+5.6mGdf9-GFP. Transgene –3.3/+5.6mGdf9-GFP contained 3.3 kb of 5'-flanking region of mGdf9 and approximately 5.6 kb downstream sequence linked to an enhanced version of GFP.
For mGdf9 promoter-luciferase constructs, pGL3-basic (Promega) containing firefly luciferase coding sequence was modified by inserting an oligonucleotide linker into its multiple cloning sites to generate appropriate restriction enzyme sites for subcloning. The 3.3 and 0.4 kb from the 5'-flanking region of mGdf9, with or without the E-box mutation, were products of convenient enzyme digestions of the existing subclones of mGdf9. The resulting different 5'-flanking sequences were ligated into appropriate restriction enzyme sites of the modified pGL3 vector.
The E-box mutation was generated by PCR-based, site-directed mutagenesis. Briefly, oligonucleotides containing the desired sequence changes were synthesized and utilized as primers in PCR. The synthesized oligonucleotides were as follows: OligoCY7, 5'-CTAGCCTCGAGCTGCAGATCTA-3', and oligoCY8, 5'-AGCTTAGATCTGCAGCTCGAGG-3'. The E-box consensus sequence CAGCTG was changed into an XbaI-site TCTAGA. Plasmid pE5, which contains 5'-flanking region of the Gdf9, was used as a template. Two separate PCRs were performed with primer pairs OligoCY7/T7 and OligoCY8/T3, respectively. The PCR was performed with Pfu DNA polymerase (Stratagene) under the PCR conditions described by the manufacturer. The purified PCR products from the two separated PCRs were mixed, denatured at 94°C, and then cooled to room temperature for annealing. The resulting product was used as template in another round of PCR with primers T7/T3. The PCR products were purified, digested with PstI and BamHI, and then cloned into the same sites of pBluescript SK vector (Stratagene). The DNA sequencing analysis was utilized to confirm the mutation.
Generation of Transgenic Mice
Linearized transgene fragments were purified, quantitated, and microinjected into the pronucleus of fertilized eggs of C57BL/6/C3H x ICR hybrids [2]. Microinjected eggs were then transferred into oviducts of pseudopregnant foster mothers. Transgenic mice were identified by Southern blot analysis or PCR. Briefly, mouse tail genomic DNA was isolated by overnight digestion with proteinase K followed by ethanol precipitation. The DNA was then dissolved in TE (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) buffer and subjected to overnight restriction enzyme digestion. The digestion was electrophoresed on 0.7% agarose gel and transferred to nylon membranes (Amersham). Probes were generated by random priming kit (Pharmacia) and hybridized with the blots. The [32P]GFP probe and enhanced [32P]EGFP probe were used, respectively, for detecting GFP transgenic and EGFP transgenic mice.
A primer pair, one from luciferase sequence 5'-CTAGCCTCGAGCTGCAGATCTA-3' and the other from the 5' sequence of mGdf9-5'-AGCTTAGATCTGCAGCTCGAGG-3', was synthesized for PCR genotyping. A small portion of tail genomic DNA was used as the PCR template. The PCR was performed with Hot-Start Taq DNA polymerase (Qiagen) under the conditions described by the manufacturer. The resulting PCR products were electrophoresed on ethidium bromide/agarose gel and visualized under ultraviolet (UV) light.
To confirm the presence of the E-box mutation in the transgenics, the PCR product was further digested with XbaI, electrophoresed on an ethidium bromide/agarose gel, and visualized under UV light.
All mice used in these studies were generated and maintained at Baylor College of Medicine. Mice were kept under standard laboratory conditions and maintained as per the National Institutes of Health guidelines and approved animal protocols of the Institutional Animal Care and Use Committee of Baylor College of Medicine.
Multitissue Northern Blot Analysis
Total RNA was extracted from the tissues of the transgene-positive mice using RNA STAT-60 according to the manufacturer's instruction (Leedo Medical Labortories). Poly(A)+ mRNA was prepared by mRNA purification kit (Stratagene). Fifteen micrograms of total RNA or 10 µg of mRNA were electrophoresed on 1.2% denaturing agarose gel, transferred to a Hybond-N nylon membrane (Amersham), and cross-linked by UV irradiation. Probes were generated using a Strip-EZ DNA probe synthesis and removal kit (Ambion). The RNA membrane was first hybridized with the [32P]GFP or [32P]EGFP coding region, then stripped and reprobed with 18S for the RNA loading control.
In Situ Hybridization
In situ hybridization was performed as described previously [18]. Freshly dissected ovaries from transgenic mice were fixed in 4% paraformaldehyde overnight. The ovaries were incubated sequentially in the following solutions: 1x PBS, 0.85% NaCl, 1:1 ethanol:saline, and 70% ethanol. Testes from transgenic mice were fixed in Bouin fixative at room temperature for 3 h. The testes were incubated in the multiple changes of 70% ethanol. The fixed ovaries and testes were then embedded in paraffin. Five-µm sections were cut, processed, and pretreated as described. The [
-35S-UTP]GFP or [-35S-UTP]EGFP antisense and sense probes were generated using T7 or T3 Riboprobe Systems (Promega). Hybridizations were carried out at 55°C for 12–16 h with 5 x 106 cpm of each riboprobe per slide in a solution containing 50% deionized formamide, 0.3 M NaCl, 20 mM Tris-HCl (pH 8.0), 5 mM EDTA, 10 mM NaPO4 (pH 8.0), 10% dextran sulfate, 1x Denhart solution, and 0.5 mg/ml of yeast tRNA. Slides were washed sequentially at 55°C in 5x SSC (1x: 0.15 M sodium chloride and 0.015 M sodium citrate) and 10 mM ß-mercaptoethanol and then in 50% formamide, 2x SSC, and 10 mM ß-mercaptoethanol; then, slides were treated with 20 µg/ml of RNase A in 1x TEN (10 mM Tris-HCl, 5 mM EDTA, 500 mM NaCl) buffer for 30 min at 37°C. A high-stringency wash was carried out at 65°C in a solution containing 2x SSC, 50% formamide, and 50 mM ß-mercaptoethanol for 20 min; in 2x SSC for 15 min; and then in 0.1x SSC for 10 min. Slides were then dehydrated and subjected to autoradiography with NTB-2 emulsion (Eastman Kodak). After developing and fixing, the slides were counterstained with hematoxylin and mounted for photography.
Luciferase Activity Assay
Tissues collected from transgene-positive and -negative mice were homogenized in 1x lysis buffer (Promega) with a Tissue Tearor electronic homogenizer (Biospec Products). The homogenate was centrifuged at 20 000 x g for 5 min at 4°C, and the supernatant was analyzed for luciferase activity by a Luciferase Assay kit as described by the manufacturer (Promega). Luciferase activity was measured with a luminometer. The luciferase final activity was normalized by the protein level in the tissues. Total protein was measured by the BCA protein assay kit (Pierce).
RESULTS
16.3 kb of Gdf9 Flanking Sequences Recapitulate Endogenous Gdf9 Expression
To maximize the likelihood of including all cis-acting elements necessary for oocyte-specific expression of Gdf9, our initial transgene (–10.7/+5.6mGdf9-GFP) (Fig. 1) contained all the genomic sequences between –10.7 and +5.6 kb (10.7 kb upstream and 5.6 kb downstream from the initiation codon of mouse Gdf9) [19]. Two independent transgenic lines expressing GFP mRNA were obtained. Northern blot analysis showed that the GFP reporter gene was expressed in the ovary and only minimally in the other tissues examined (Fig. 2A). Multiple transcript sizes are observed in the –10.7/+5.6mGdf9-GFP mice. These transcripts likely result from alternative polyadenylation sites, as observed in other testis and ovary transcripts [20, 21]. The GFP expression was only detectable in testis and brain when poly(A)+ RNA from these two transgenic lines were used (Fig. 2A). The GFP signal was further localized to the oocytes of the ovary by in situ hybridization analysis (Fig. 3, A and B). Thus, the 16.3 kb of genomic sequence flanking the Gdf9 gene contains all the elements required for efficient expression in oocytes and relative suppression of expression elsewhere. Identical to mouse Gdf9 expression in the ovary [2, 9, 10], the expression of transgene –10.7/+5.6mGdf9-GFP was not present in primordial oocytes, was first detected in oocytes of type 3a follicles, and was present in oocytes of all subsequent stages.
|
|
|
Compared to –10.7/+5.6mGdf9-GFP, the transgene –3.3/+5.6mGdf9-GFP contained the same sequences from –3.3 to +5.6 kb, but with a deletion of upstream sequences between –10.7 and –3.3 kb (Fig. 1). Four founder lines were transgene-positive, as demonstrated by Southern blot analysis, and three lines transmitted the transgene. Similar to the expression pattern of the 16.3-kb transgene, the expression of the 8.9-kb transgene was restricted to the oocytes of the ovary, as demonstrated by Northern blot analysis (Fig. 2, C and D) and in situ hybridization analysis (Fig. 3, C and D). Variable expression of the 8.9-kb transgene was seen in testis samples from males of these lines (Fig. 2, C and D). Higher expression of the 8.9-kb transgene was detected in the brain samples of multiple lines, although the size of the transcript was aberrantly large (Fig. 2, C and D). This suggests that these brain transcripts likely were derived from expression of other genes located 5' of the Gdf9 locus, possibly spliced to the GFP reporter coding region. When looking into the upstream region, an ubiquinol-cytochrome c reductase-binding protein gene (UniGene Mm.251621) was found approximately 2.5 kb upstream from Gdf9. Because this gene is expressed in multiple tissues and is transcribed in the opposite direction from Gdf9, it is unlikely to be responsible for the longer brain transcript. However, one seemingly aberrant cDNA (GenBank accession no. AK078302) from an olfactory brain library is transcribed toward Gdf9, showing partial overlap with that ubiquinol-cytochrome c reductase-binding protein gene. It is likely that putting the transgene in a different genomic context may have enhanced transcription, extending into the GFP:Gdf9 locus, similar to that of the aberrant cDNA. Thus, these two constructs showed that sequences between –3.3 and +5.6 kb contain cis-acting elements that are sufficient both for directing Gdf9 expression specifically to oocytes in vivo and for suppressing the expression of Gdf9 in testis and other tissues.
3.3 kb of 5'-Flanking Region of mGdf9 Is Sufficient to Drive Expression Specifically in Germ Cells
We produced an additional transgene (–3.3mGdf9-EGFP) that contained only 3.3 kb of the 5'-flanking sequences of Gdf9 fused to the cDNA of EGFP (EGFP) but that lacked the +5.6 kb downstream sequences of mouse Gdf9 (i.e., the exons, intron, and some 3' noncoding regions). Among the five founder lines that were generated, only two lines transmitted the –3.3mGdf9-EGFP transgene. As demonstrated by multitissue Northern blot analysis, the EGFP mRNA was detected not only in the ovary but also at a much higher level in the testis in both lines (Fig. 2B and data not shown). The other positive line also had detectable expression of EGFP mRNA in the lung (data not shown), which likely was an artifact of the genome insertion site.
To quantitate and confirm the high expression level in the testis, we took advantage of the high sensitivity and quantitative features of the luciferase reporter gene. The above 3.3-kb sequences were fused to the coding region of the firefly luciferase gene (–3.3mGdf9-Luc) (Fig. 4). The new transgene –3.3mGdf9-Luc in two independent transgenic lines showed luciferase expression in the gonads in an identical expression pattern to that of transgene –3.3mGdf9-EGFP (data not shown). Luciferase activity in the testis of the –3.3mGdf9-Luc transgenic mice was approximately fivefold higher than that in the ovary (Fig. 4).
|
To examine the cell populations expressing the reporters in the ovary and testis, we performed in situ hybridization. In the ovary, the signal of EGFP mRNA was only detected in the oocytes of the ovary (Fig. 3, E–H). In the testis, the expression was restricted to germ cells and, specifically, primary spermatocytes through round spermatid stages (Fig. 3, I–L). Thus, the upstream 3.3-kb genomic sequences of mouse Gdf9 are sufficient to drive the expression in both oocytes in the ovary and germ cells in the testis.
To refine the region conferring high expression in the testis, we first tested the distal 5'-flanking region, specifically between –10.7 and –3.3 kb. Similar to transgene –3.3mGdf9-Luc, transgene –10.7mGdf9-Luc was constructed by fusing the entire 10.7 kb of 5'-flanking sequences of Gdf9 to the cDNA of the luciferase gene. The resulting two transgenic lines were analyzed for the luciferase activity in comparison with the –3.3mGdf9-Luc transgenic line. As shown in Figure 4, in both ovary and testis, the luciferase activity was significantly higher in the –10.7mGdf9-Luc transgenic than in the –3.3mGdf9-Luc transgenic line. Thus, these results suggested the presence of enhancer activity in the upstream region between –3.3 and –10.7 kb and eliminated the possibility of the presence of repressor activity within the upstream 10.7-kb region.
Testis-Specific Repressor Element Was Located Within the 5.6-kb Sequences Downstream of the Start of Translation
We have shown that transgene –10.7mGdf9-Luc, transgene –3.3mGdf9-EGFP, and transgene –3.3mGdf9-Luc, all lacking the +5.6-kb sequences, were capable of driving high-level reporter expressions in testis (Figs. 1, 2, and 4). In contrast, we have demonstrated that the transgenes carrying the 5.6-kb sequences, –10.7/+5.6mGdf9-GFP and –3.3/+5.6mGdf9-GFP, specifically directed the reporter expression to the oocytes in the ovary, but not to the testis (Fig. 1). Together, these data indicate that the downstream 5.6-kb sequences of Gdf9 must contain a testis-specific repressor element.
E-box Is Required for Expression of Gdf9 Only in Ovary, Not in Testis
The studies described above show that the oocyte expression of Gdf9 requires only the 3.3-kb upstream region. We therefore wished to define further the sequences that contributed to this expression. Alignment of the proximal promoter sequences from Gdf9 genes of 16 mammalian species revealed that a perfectly conserved E-box element (CAGCTG) is present (Fig. 5A). In the mouse, this E-box lies at –182 bp upstream of the start of translation. Interestingly, the approximately 400-bp region upstream of the start of translation was well conserved between mouse and human compared to the less conserved 3'-untranslated region (UTR) and downstream region (Fig. 5B). The conservation was still evident when the corresponding upstream regions from these 16 mammalian species were aligned and examined (Fig. 5A). We therefore explored the possibility of whether a minimal promoter containing only 0.4 kb of the mouse Gdf9 5'-flanking region and including this conserved E-box was sufficient to direct expression to oocytes. We linked the 0.4-kb sequences to the cDNA of the luciferase reporter gene to generate the transgene –0.4mGdf9-Luc. In all three transgenic lines, the luciferase expression in ovary was at levels comparable to those in the –3.3mGdf9-Luc transgenic mice. Additionally, the expression of the luciferase was restricted to the ovary (Fig. 4 and data not shown).
|
To examine further the role of the conserved E-box in the mouse Gdf9 promoter, the E-box consensus sequence CAGCTG in the above –0.4mGdf9-Luc transgene was mutated into an XbaI site (TCTAGA). The mutation was confirmed by DNA sequencing. Three founder mice carrying the –0.4Mut-mGdf9-Luc transgene were identified by PCR genotyping. The mutation in the transgene was further confirmed by XbaI digestion of the PCR product (data not shown). As analyzed by a luciferase assay, mutation of the E-box in these transgenic mice abolished the luciferase expression in the ovary, in contrast to the observation in the –0.4mGdf9-Luc transgenic mice (Fig. 4). However, neither –0.4mGdf9-Luc nor –0.4Mut-mGdf9-Luc transgenics had detectable luciferase activity in testis (Fig. 4) or other examined tissues (data not shown). These results suggest that the 0.4-kb sequences contain promoter activity sufficient for expression in the ovary, but not for expression in the testis or other nonovarian tissues.
To determine further whether the function of the E-box was ovary-specific, we made the same E-box mutation in transgene –3.3mGdf9-Luc, which contains the longer 3.3-kb promoter sequences of Gdf9 (–3.3Mut-mGdf9-Luc). The E-box mutation (i.e., XbaI site) in all four generated transgenic lines was confirmed (data not shown). Three of the four transgenic lines consistently showed no detectable luciferase expression in the ovaries but high expression in the testes, at levels comparable to those in –10.7mGdf9-Luc transgenics (Fig. 4). In fact, one of these transgenic lines expressed the highest luciferase activity in testis among all of the examined transgenic lines in the present study. Only one line showed luciferase expression in both ovary and testis at a level similar to the transgene carrying wild-type E-box (data not shown). One likely explanation for this E-box-independent ovarian aberrant expression is an influence of enhancer sequences at a different chromosomal transgene integration site. Thus, these observations are in line with the findings from the –0.4Mut-mGdf9-Luc transgenics, suggesting that the conserved E-box plays an essential role for the expression in the ovary but is not required for the expression in the testis.
DISCUSSION
In the present study, we used transgenic mouse technology to study the promoter of mouse Gdf9. Transgenes harboring different regions of Gdf9 were fused to reporter genes. We demonstrated that transgenes –10.7/+5.6mGdf9-GFP, –3.3/+5.6mGdf9-GFP, and –3.3mGdf9-GFP, were capable of driving the reporter gene expression either specifically to the oocyte or to both oocytes in females and spermatocytes and round spermatids in males. The results suggest that 3.3 kb of the 5'-flanking region of mouse Gdf9 contains all the cis-acting elements that are sufficient for targeting the expression to male and female germ cells.
Comparison of the expression pattern among the transgenes generated in the present study reveals that the 5.6-kb sequences downstream of the initiation codon of mouse Gdf9 are critical for suppressing Gdf9 expression in testis. The deletion of this region causes the initiation of high expression in the testis, indicating that a testis-specific repressor element is located within the 5.6-kb region. An unknown testis-specific repressor therefore must interact with the cis-acting sequences within this 5.6-kb region to suppress mouse Gdf9 expression specifically in testis. Many studies have shown that the presence of repressors is one mechanism to control transcription in a tissue-specific manner. This 5.6-kb testis-repressing region of Gdf9 includes the coding sequences, intron, and 3'-UTR. The 3'-UTR has been demonstrated to function as a repressor binding region in several genes [22, 23]. Regulatory sequences in the 3'-UTR likely are conserved between species [24], and regions showing 50% or more identity were evident when comparing the 3'-UTR and the downstream regions of human Gdf9 and mouse Gdf9 (Fig. 5B). Further deletion analysis would help to define these testis-specific repressor elements. A more expedient way to pursue identification of the testis-repressor element would be in vitro transfection into commercially available male germ cell lines. For example, germ cell line-1 (GC-1), derived from testicular germ cells, has been successfully used for in vitro transfection [25]. We cannot detect Gdf9 mRNA in these GC-1 cells (unpublished data). Thus, this cell line may be useful for further delineating possible repressor elements through an in vitro transfection approach.
An E-box motif (CANNTG) has been found in the promoter of many tissue-specific genes, including the Zp genes, Cyp19, and P2X1. The consensus sequence CANNTG can bind to proteins of a basic helix-loop-helix (bHLH) family to regulate transcription in a tissue-specific manner. Mutational analysis of the conserved E-box at (–182 to –177 bp) upstream of the Gdf9 initiation codon has demonstrated that the E-box is essential for the expression of Gdf9 in the ovary. However, the intact E-box is not required for expression in the testis, suggesting that the E-box functions in an ovary-specific manner. A similar phenomenon has been demonstrated in the mouse Zp3 gene. An E-box CACGTG present at –210 to –205 bp upstream of the Zp3 initiation codon is necessary and sufficient for Zp3 expression in the oocyte, but it is not sufficient for ectopic expression in 10T1/2 cells. Additionally, an ovary-specific bHLH transcription factor, FIG1a , binds to the Zp3 E-box through heterodimerization with protein E12, a ubiquitous bHLH [13]. The FIG1a is expressed at high levels in primordial oocytes and persists in growing oocytes. The expression window of FIG1a precedes and is coincident with the Gdf9 expression. Thus, FIG1a is a likely candidate transcription factor for germ cell-specific regulation of Gdf9. However, the E-box (CAGCTG) identified in the Gdf9 promoter was similar but not identical to the E-box (CACGTG) in Zp3 gene. It has been observed that the central two nucleotides contribute to discriminatory binding among the different bHLH family members [16, 26]. It will be interesting to determine whether the inversion of the central G and C nucleotides of the E-box permits regulation by FIG1a or requires a novel transcription factor that specifically regulates Gdf9 expression in the ovary. This question remains to be addressed by further analysis, such as gel mobility shift assays.
In addition to the conserved E-box, several other conserved transcriptional binding sites also were identified by aligning the mammalian Gdf9 promoter sequences. For example, transcription factor-binding sites conserved in both mouse and human were found by the Match program (http://www.gene-regulation.com/) searching the TRANSFAC 6.0 database. These sites include HNF-3ß, FOXD3, GATA-3, GAGA-X, Lmo2 complex, NF-Y, CCAAT box, Pax-4, Brn-2, HLF, and Nkx2–5. The significance of these putative transcription factor-binding sites in regulating Gdf9 expression remains to be demonstrated.
The 3.3 kb of 5'-flanking region of the mouse Gdf9 gene has been used to drive Cre recombinase, and the Cre recombinase activity recapitulated Gdf9 expression in growing oocytes both temporally and spatially [10, 27], allowing the creation of conditional oocyte-specific knockouts beginning from the primordial follicle stage. In addition to Zp3Cre [28] and Msx2Cre [29], in which the oocyte-specific Cre activity was observed starting from the primary follicle stage and the secondary follicle stage, respectively, Gdf9-iCre in mice appears to result in efficient oocyte-specific conditional knockout from the primordial follicle stage. This is particularly important for studying the postnatal functions of germ cell-expressed genes that cause early embryonic lethal or gonad formation failure in conventional knockouts. One candidate for this type of tissue-specific and, in fact, cell type-specific knockout approach is the oocyte-expressed KIT tyrosine receptor. When lost constitutively, absence of KIT leads to defects in embryonic germ cell migration and proliferation, resulting in an absence of germ cells in both the male and female gonad postnatally [30, 31], preventing the study of KIT function during folliculogenesis.
Interestingly, medium to high levels of reporter gene expression were observed in the testis of transgenic mice with all our transgene constructs that utilized the 3.3 kb of 5'-flanking region of the mouse Gdf9 as promoter, but two of the three transgenic Gdf9-iCre lines generated by Lan et al. [27] that utilized the same 5'-flanking region as promoter showed only oocyte-specific Cre recombinase activity. Because transgenes are integrated randomly into the genome and usually in tandem repeats of variable copy numbers, it is difficult to assess the causes of the different expression levels in the testis observed in these two studies. One possible solution is to target a single copy of each of these transgenes to a specific locus, such as Hprt [32], for a more consistent genomic milieu for promoter analysis.
In summary, the present study provides an example of transcriptional regulation in germ cells. We have provided evidence to demonstrate that cis-acting sequences conferring oocyte-specific expression are present in the 3.3 kb of 5'-flanking sequences of mGdf9. Our results also suggest that a testis-specific repressor element is located within a region 5.6 kb downstream of the Gdf9 translation initiation site. In addition, we identified an evolutionarily conserved E-box at –182 to –177 bp upstream from the ATG site in mouse Gdf9 that is essential for Gdf9 expression in ovary but not in testis. Furthermore, we identified a promoter region of Gdf9 directing high expression in testis. The strong promoter activity in the testis resulting from the 3.3 kb region is of great value for researchers interested in overexpressing genes in spermatocytes and spermatids and for targeting gene expression to the male germ cells at the particular stages. Germ cells contribute to the blueprint of the next generation. Appropriate regulation of genes in sex-differentiated germ cells is of critical importance for reproductive function in males and females.
ACKNOWLEDGMENTS
We thank Shirley Baker for help with manuscript formatting.
FOOTNOTES
1 Supported in part by National Institutes of Health (NIH) grant HD33438. 
2 Correspondence: Martin M. Matzuk, Department of Pathology, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030. FAX: 713 798 5833; mmatzuk{at}bcm.tmc.edu
Received: 3 December 2005.
First decision: 25 December 2005.
Accepted: 22 February 2006.
REFERENCES
, a germ cell-specific transcription factor involved in the coordinate expression of the zona pellucida genes. Development 1997 124:4939-4947[Abstract]This article has been cited by other articles:
|
|
 |
|
  T. Hamatani, M. Yamada, H. Akutsu, N. Kuji, Y. Mochimaru, M. Takano, M. Toyoda, K. Miyado, A. Umezawa, and Y. Yoshimura What can we learn from gene expression profiling of mouse oocytes? Reproduction, May 1, 2008; 135(5): 581 - 592. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Bettegowda, J. Yao, A. Sen, Q. Li, K.-B. Lee, Y. Kobayashi, O. V. Patel, P. M. Coussens, J. J. Ireland, and G. W. Smith JY-1, an oocyte-specific gene, regulates granulosa cell function and early embryonic development in cattle PNAS, November 6, 2007; 104(45): 17602 - 17607. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
