| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
research-article |
Tanaka Project,2 Center for Advanced Science and Innovation, and Research Collaboration Center on Emerging and Re-emerging Infections,3 Osaka University, Osaka 565-0871, Japan
Bioscience Research-Education Center,4 Akita University, Akita 101-8543, Japan
Center for Developmental Biology,5 RIKEN Kobe, Chuo-ku, Kobe 650-0047, Japan
ABSTRACT
Haspin is a unique protein kinase expressed predominantly in haploid male germ cells. The genomic structure of haspin (Gsg2) has revealed it to be intronless, and the entire transcription unit is in an intron of the integrin alphaE (Itgae) gene. Transcription occurs from a bidirectional promoter that also generates an alternatively spliced integrin alphaE-derived mRNA (Aed). In mice, the testis-specific alternative splicing of Aed is expressed bidirectionally downstream from the Gsg2 transcription initiation site, and a segment consisting of 26 bp transcribes both genomic DNA strands between Gsg2 and the Aed transcription initiation sites. To investigate the mechanisms for this unique gene regulation, we cloned and characterized the Gsg2 promoter region. The 193-bp genomic fragment from the 5' end of the Gsg2 and Aed genes, fused with EGFP and DsRed genes, drove the expression of both proteins in haploid germ cells of transgenic mice. This promoter element contained only a GC-rich sequence, and not the previously reported DNA sequences known to bind various transcription factors—with the exception of E2F1, TCFAP2A1 (AP2), and SP1. Here, we show that the 193-bp DNA sequence is sufficient for the specific, bidirectional, and synchronous expression in germ cells in the testis. We also demonstrate the existence of germ cell nuclear factors specifically bound to the promoter sequence. This activity may be regulated by binding to the promoter sequence with germ cell-specific nuclear complex(es) without regulation via DNA methylation.
gamete biology, GC-rich, gene regulation, integrin, intronless, methylation, spermatid, spermatogenesis, testis, transcription
Germ cells are essential for the maintenance of species. In mammalian males, germ cell differentiation, called spermatogenesis, is a complex process that includes the proliferation and differentiation of diploid stem cells that undergo meiosis and dramatic morphological changes from spermatid to sperm without cell division [1]. To elucidate the mechanism of spermatogenesis, we isolated many cDNA clones specifically expressed in germ cells using a subtracted cDNA library prepared from supporting cells of germ cell-less mutant (W/Wv) testes and wild-type testes [2]. One of these, Gsg2, was identified as a gene predominantly expressed in the germ cell [2]. Afterward, the full-length cDNA of Gsg2 was isolated and subsequently characterized as a haploid-specific unique protein kinase, haploid germ cell-specific nuclear protein kinase (haspin), which is localized in nuclei of round spermatids, bound to DNA, and has Ser/Thr protein kinase activity [3]. Because transfection of Gsg2 (haspin) cDNA into cultured somatic cells causes the cessation of cell proliferation, GSG2 protein may be involved in the regulation of proliferative activity, as well as in specific gene expression in haploid germ cells [3]. GSG2 has several interesting domains, including a basic region, a region homologous to murine myocyte enhancer factor 2B (MEF2B) [4], and protein kinase consensus domains [5, 6]. Higgins [7] identified the detailed expression profile of mRNA of Gsg2 (haspin) and integrin 
E (Itgae), and found that signals of mRNA of Gsg2 and alternatively spliced integrin alphaE-derived mRNA (Aed) are strongly detected in the testis. Furthermore, he found that Gsg2 mRNA is detected in adult thymus and bone marrow, and is weakly detected in various organs [7]. Recently, it was shown that GSG2 phosphorylates H3 at Thr 3 in vitro, and is required for H3 Thr 3 phosphorylation in mitotic cells [8, 9]. Gsg2 is found at the centrosomes and spindles during mitosis, and integrates the regulation of chromosome and spindle function during mitosis and meiosis [8, 9].
Analysis of these germ cell-specific genes has revealed that they contain a large number of intronless genes [10]. Isolation of the mouse and human Gsg2 genome has shown that the genomic constructs of both Gsg2s are nearly identical [11]. The upstream region of the initiation position of the Gsg2 genome DNA contains neither a TATA box nor cAMP-responsive promoter sequences [11, 12]. The whole transcription unit is located in an intronless gene within an intron of the Itgae gene, and the directions of transcription for these two genes are opposite [7]. Repetitive sequences (GAGCC[A/T]T) have also been found at both ends of the Gsg2 genome conserved in human and mouse Gsg2 genomic DNA [11]. These results indicate that Gsg2 may have been generated by reverse transcriptase activity involving a transposition in a common ancestor prior to the split between humans and mice [11].
With the completion of the human genome sequence, it was revealed that head-to-head gene constructs are a common occurrence [13]. For example, 42% of the genes responsible for DNA repair are transcribed by bidirectional promoters. Most genes constructed head to head include a CpG island between each gene [13]; however, the functions of the CpG island are unclear. Methylation of the CpG island on genomic DNA is a central mechanism for regulating tissue-specific transcription [14]. Methylation induces heterochromatin condensation through histone acetylation, as well as the inactivation of the X chromosome [14] and imprinting of chromosomes [14] during germ cell differentiation. De novo methylation and demethylation of chromosomal DNA are essential for these processes [15]. Genes expressed during spermiogenesis avoid methylation in male germ cells, although they are methylated in somatic tissues. The promoter region of Gtf2a1lf, a germ cell-specific counterpart of the large (/ß) subunit of the general transcription factor IIA,1 [16], is methylated in somatic cells, non-germ cells of the testis, and epididymal spermatozoa. Gtf2a1lf is demethylated, but not yet expressed, in spermatogonia, in which specific transcription factors are required to initiate expression; however, the mechanism of gene regulation associated with methylation in spermiogenesis is not clear.
In this study, we identified the 193-bp genomic fragment from the 5' end of the Gsg2 and Aed genes that specifically, bidirectionally, and synchronously induces the transcription of both Gsg2 and Aed in haploid germ cells in testes of transgenic mice. Methylation analysis has shown that the promoter sequence is not methylated in somatic tissue and germ cells, and gel retardation assays indicated that testis-specific nuclear protein complex(es) binding are available for this element. The promoter sequence introduced bidirectional transcription in 16 of 17 independently transgenic mice, with reproducibility. The bidirectional promoter of Gsg2 and Aed could have arisen from introduced transcriptions without the need for regulation via methylation and germ cell-specific nuclear protein complex(es).
Construction of Reporter Plasmids and Microinjection
The pDsRed-N1 plasmids (Clontech, Palo Alto, CA) were reconstructed for promoter assays. The promoter region of pDsRed-N1 was removed using AseI and NheI restriction enzymes. The EGFP gene fragment cut out from pEGFP-N1 (Clontech) by AflI and AgeI was cloned into pDsRed-N1 plasmids treated with AseI and NheI after blunt-end treatment. The prepared plasmid was named pBiPRO. The 193-bp SacI-KpnI fragment of the genomic Gsg2 (haspin) gene containing the 114-bp 5' upstream sequence from the transcription initiation site, together with the 79-bp 5' end of Gsg2 cDNA, was cloned into the SacI/KpnI restriction sites of the pBiPRO vector (pBiPRO-1; see Figs. 1 and 2). To change the direction of the 193-bp promoter sequence, it was recovered by BglII/KpnI restriction enzymes and recloned into the SacI–BamHI restriction sites of the pBiPRO vector (pBiPRO-2) after blunt-end treatment. The 1073-bp fragment between the KpnI linker-primer P1 and the KpnI restriction enzyme was cloned into the KpnI restriction sites of the pBiPRO vector (pBiPROs 3 and 4; see Figs. 1 and 2). The KpnI linker-primer P1 and the specific primer downstream of the KpnI restriction enzyme were used for PCR of the Gsg2 genomic clone. PCR products were digested with KpnI restriction enzyme and recovered using a SUPREC PCR spin column (Takara, Shiga, Japan). The purified fragments were cloned into the KpnI restriction sites of the pBiPRO vectors bidirectionally. DNA sequences were confirmed by direct sequencing using DsRed- and EGFP-specific primers. To examine the promoter activity of these elements, pBiPROs 1 and 2 were transfected into HEK 293 cells using Lipofectamine Plus reagent (Gibco BRL, Grand Island, NY) following the manufacturer's protocol. After 72 h, the cells were observed under a fluorescence microscope.
|
|
The constructed pBiPRO vectors described above were digested with ApaLI and DraIII, and a DNA fragment was resolved by agarose gel electrophoresis and recovered from the gel using glass beads (Qiaex; Qiagen K.K., Tokyo, Japan). The excised DNA fragment was injected into the pronuclei of BCF2 eggs. The offspring from the microinjected eggs were identified by PCR with a set of EGFP-specific primers (385 primer 5'-TGAACCGCATCGAGCTGAAGGG-3' and 386 primer 5'-TCCAGCAGGACCATGTGATCGC-3'). DNA from mouse tails was added to 15 µl of reaction mixture containing 0.1 mM of each primer, 8 mM of dNTP mixture, 1x Ex Taq buffer, and 0.3 U of Ex Taq (Takara). The cycling conditions were as follows: 35 cycles of denaturation at 96°C for 45 sec, annealing at 60°C for 45 sec, and extension at 72°C for 45 sec. Transgenic mice were subsequently mated to C57BL/6 mice, and positive male offspring from this cross were used for experiments. F2 and F3 mice were used for the analysis of gene expression.
All mice were bred and maintained in our laboratory animal facilities and used in accordance with Japanese Association for Laboratory Animal Science guidelines for care and use of laboratory animals. Mice were kept under controlled temperature and lighting conditions during experiments and were provided food and water ad libitum.
Various organs, freshly removed from transgenic mice, were observed under ultraviolet (UV) excitation light and then homogenized on ice with lysis buffer (10 mmol/L Na2 HPO4 [pH 7.2], 160 mmol/L NaCl, 1% Triton X-100, 1% deoxycholic acid, 0.3% SDS, 2 mmol/L phenylmethylsulfonyl fluoride [Sigma, St. Louis, MO]). After centrifugation, protein concentrations of each supernatant were estimated using a Bradford Protein Assay (Bio-Rad, Richmond, CA). Then, 50 µg of protein from each extract was subjected to SDS-PAGE, followed by electroblotting to polyvinylidenedifluoride (PVDF) membrane filters (Millipore, Bedford, MA). The filters were blocked with 5% nonfat dry milk and washed for 15 min with Tris-buffered saline (TBS; 100 mmol/L Tris–HCl [pH 7.5], 150 mmol/L NaCl). The filters were then reacted with anti-EGFP rat monoclonal antibody [3] or anti-DsRed rabbit serum (Clontech) in TBS for 1 h at 25°C and washed in TBS, once for 3 min, then three times for 5 min each. Finally, the filters were incubated with anti-rat or anti-rabbit immunoglobulins conjugated with peroxidase (1:500; Amersham Pharmacia Biotech, Tokyo, Japan) for 1 h at 25°C. After further washing, reactive bands were visualized by development with the POD staining kit (Wako, Osaka, Japan).
Freshly removed organs of transgenic mice were homogenized in RNA zolTM B (Invitrogen, Tokyo, Japan). Total RNA was extracted according to the manufacturer's instructions and quantified by optical density measurement. RNA samples containing 2.2 M formaldehyde were subjected to electrophoresis in a 1.1% agarose gel containing 0.66 M formaldehyde. RNA was transferred to a nitrocellulose filter in 20x SSC solution. Hybridization was performed with 32P-labeled cDNA prepared using the BcaBest random primer kit (Takara) at 42°C for 16 h in a solution containing 4x SSC, 5x Denhardt solution, 0.2% SDS, 12 µg/ml denatured salmon sperm DNA, and 50% formamide. As probes, cDNAs of mouse Gsg2 and integrin E (Itgae) were prepared as previously reported [3]. Whole cDNA of EGFP and DsRed was prepared from pEGFP-N1 and pDsRed-N1 plasmids (Clontech) using appropriate restriction enzymes. Filters were washed twice in 0.1x SSC and 0.1% SDS at 60°C. Signals of the bands were detected using Image Analyzer (Fuji Film, Tokyo, Japan).
Fractionation of Testicular Cells
Testicular germ cells were fractionated as described previously [17]. Briefly, seminiferous tubules were placed in PBS-E (1x PBS and 1 mM EDTA) and dispersed by gentle pipetting a few times to remove interstitial cells. The interstitial cells were collected by centrifugation (Leydig cell-enriched fraction). The tubules then were transferred to a plastic Petri dish, cut into small fragments with a knife, transferred to a conical tube, and washed by pipetting in PBS-E. The conical tube was left standing for 5 min. Subsequently, the supernatant was filtered twice through a nylon mesh (NBC Industries, Tokyo, Japan), and the filtrate was used as the germ cell-enriched fraction. The cells on the nylon mesh were washed with PBS-E and filtered twice through a second nylon mesh. The cells remaining on the second nylon mesh were the Sertoli cells.
The testes of 2-mo-old transgenic mice were sampled. The testes were fixed in 4% paraformaldehyde at 4°C for 12 h and then embedded in glycol methacrylate (Technovit 8100; Heraeus Kulzer GmbH, Hanau, Germany). Sections 5-µm thick were prepared using a cryomicrotome (HM325; Microm, Woodstock, CT). The sections were then examined under a fluorescence microscope (BX50; Olympus, Tokyo, Japan). Because the fluorescence intensity of the DsRed protein is rather weak relative to that of EGFP, particularly in the observation of cross sections, freshly prepared live seminiferous tubules or separated testicular cells were observed on glass slides under a fluorescence microscope.
Bisulfite sequencing was carried out using a CpGenome DNA Modification kit according to the manufacturer's recommendations (Chemicon, Temecula, CA) [18]. Then, 5 µg of genomic DNA was incubated with a bisulfite-hydroquinone solution for 16 h at 50°C; 0.1 µg of the purified DNA was subjected to PCR. Appropriate primers for PCR were selected based on the putative nonmethylation region using MethPrimer (available at http://www.urogene.org/methprimer/index1.html) [19]. Two sets of primers in exon 27 of Itgae and +138 to +161 of Gsg2 (Fig. 1B) were used for methylation analysis. One set of primers was haspin pr 5'-1 TTTTGATTTTGGGTGTATTTAGTGT and haspin pr 3'-1 ACTAAAACCATTACTCAACCAACC; the other was haspin pr 3'-2 TCTTAACTCTAAATACACTCAATAT and haspin pr 5'-2GTTGGAATTATTGTTTAGTTAGTT. PCR products 334 bp in length were cloned into PCRII vector (Invitrogen), and 10 colonies for each tissue genome DNA were selected. Inserted fragments of each plasmid were sequenced using M13 reverse and forward primers.
Germ cell fractions of the testes and nuclei of germ cells were prepared as described in our previous report [17, 20]. Nuclear extracts were prepared from testicular germ cells and liver in a stock solution of 10 mM Tris-HCl (pH 7.5), 1 mM EDTA, and Brij-58 at –70°C [21]. The reaction mixture (19 µl), containing nuclear extract, 1 µg poly dIdC, 10 mM Tris-HCl (pH 7.6), 50 mM NaCl, and 1 mM EDTA, was incubated at 4°C for 30 min. The 193-bp SacI/KpnI fragment of Gsg2 genomic DNA was then added to the reaction mixture. After 10 min, the reaction mixture was electrophoresed on a 4% nondenaturing polyacrylamide gel. The gel was fixed in 10% acetic acid and methanol, dried, and autoradiographed.
Expression Units of Gsg2 (Haspin) and Integrin E
Transcription of the alternative variant of integrin E mRNA (Aed) occurs in the bidirectional Gsg2 (haspin) promoter region [7]. Approximately 0.7 kb of Aed was expressed predominantly in haploid germ cells in mouse testes. Northern blot analysis using a 324-bp fragment of integrin E (Itgae) cDNA showed alternative variants of 0.7-kb Aed strongly expressed in testes, as previously reported [7], in comparison to the conventional 4.5-kb Itgae mRNA widely expressed in various tissues [22]. To identify the transcription start site of the alternative variants of Aed in mouse testes, two kinds of capped libraries of the mouse testis were used (Wako and Invitrogen). The results showed that the longest 5' terminal of Aed is located in the transcriptional unit (+26) of mouse Gsg2 (Fig. 1).
Detection of Bidirectional Promoter Activity in Germ Cells Using a Transgenic Animal Technique
To identify the promoter sequence responsible for inducing germ cell-specific bidirectional transcription in the testis, we constructed four reporter plasmids (Fig. 2), which were differentiated by having either the shorter (193 nucleotides; pBiPROs 1 and 2) or the longer (1073 nucleotides; pBiPROs 3 and 4) genomic DNA insert in one of two directions between the two reporter genes of EGFP and DsRed, arranged in a head-to-head conformation (Fig. 2). We examined the promoter activity of these promoter elements through transfection into cultured cells (HEK 293) [3]. After 3 days of transfection, the expression of EGFP and DsRed protein was weakly detected under a fluorescent microscope with UV light excitation (data not shown). Next, we constructed various lines of transgenic mice and examined the expression of EGFP and DsRed reporter genes in vivo. At least six independent transgenic mouse lines of each construct were raised (Table 1). Most organs of transgenic mice were observed under UV light to examine the expression of fluorescent proteins, and expression profiles were examined using Western blotting for protein extractions of testis and liver from 2- and 3-wk-old transgenic mice (Table 1). All transgenic mice predominantly expressed DsRed protein in haploid germ cells. The expression of EGFP protein was not detected by Western blotting in one line of the mice that possessed the pBiPRO-2 vector (Table 1). To better understand the expression of reporter genes of EGFP and DsRed, we performed Northern blot and Western blot analyses of various organs of four representative transgenic mice having each transgene. Both reporter genes were detected predominantly in the testes by Northern blotting (Fig. 3). Western blotting, using each specific antibody, showed the positive immunoreaction predominantly in germ cells of the testis (Fig. 3). In addition, Northern blotting showed that the EGFP and DsRed mRNA expressed in the 16-day-old to adult testis were the same as endogenous Gsg2 and Aed mRNA (Fig. 4). Higgins showed that the mRNAs of Gsg2 and Aed were expressed predominantly in the testis and weakly in various tissues [7]. We examined whether EGFP protein is expressed in specific cells in various organs or embryos using microscopy. No EGFP signal was detected in tissues (see Supplemental Figs. 1 and 2 available at www.biolreprod.org). These results suggest that Gsg2 and Aed are expressed in some mouse tissues very weakly.
|
|
|
Histological observations of the testis showed that EGFP fluorescence was predominantly expressed from round spermatids to elongated spermatids, and weakly expressed in pachytene spermatocytes (Fig. 5 and Supplemental Fig. 4 available at www.biolreprod.org). These observations were in good agreement with the results of age-based Northern blotting. Fluorescent signals of DsRed protein in testis sections, however, were weakly detected by histological observation. To demonstrate the synchronous expression of two fluorescent proteins of EGFP and DsRed in single haploid germ cells, the germ cells were spread onto glass slides, and fluorescent signals were observed under UV light (Fig. 6). Although the expression of each of the EGFP and DsRed proteins was not directly proportional, EGFP and DsRed proteins were detected in each single cell in testicular germ cells. Some lines of transgenic mice of pBiPROs 1, 2, 3, and 4 showed mosaic EGFP expression in the testicular tubules (Fig. 7); however, the expression of reporter genes was detected predominantly in haploid germ cells (Fig. 7). Background noise signals were observed as yellow signatures in cells between tubules. Green fluorescence was detected predominantly in haploid germ cells of the testes having both the shorter 193-nucleotide fragment (pBiPROs 1 and 2; Fig. 3) and longer 1073-nucleotide fragment (pBiPros 3 and 4; data not shown) of the genomic DNA of transcriptional elements (Table 1). This indicated that the 193-nucleotide fragment controlled the testis-predominant bidirectional transcription.
|
|
|
The reporter gene construct showed that the 193-bp genomic fragment from the 5' end of the Gsg2 and Aed genes was indeed sufficient for the germ cell expression of both genes, and the genes were clearly and predominantly regulated simultaneously in germ cells. These results indicate that Gsg2 and the alternative variant of Aed were mainly expressed in haploid germ cells. They also indicate that the testis-specific Aed gene may be expressed from the opposite strand, starting at a transcription initiation site downstream from the Gsg2 initiation site, resulting in a 26-bp genomic DNA overlap in the transcription units of the Gsg2 and Aed genes in mice.
Methylation of the Promoter Region
The promoter region did not include consensus DNA sequences for germ cell expression as a cAMP response element (CRE) consensus sequence without TCFAP2A1 (AP2), SP1, E2F1, and GC-rich sequences for recognition sites of common transcription factors. To examine whether demethylation introduced bidirectional transcription in somatic cells, liver and fibroblast cells from transgenic animals were cultured in methylcytosine analog 5-azacytidine [23] or trichostatin A [24]. Western blotting did not detect expression of EGFP and DsRed protein (data not shown). To determine whether silencing of the gene expression of Gsg2/Aed in somatic cells was associated with DNA methylation of the core promoter sequence, we examined CpG methylation around the intrinsic 193-bp core promoter sequence of genomic DNA in liver, kidney, and germ cells using a bisulfite sequencing technique. Interestingly, no methylation of the core promoter sequence was observed in either Gsg2/Aed-expressing or nonexpressing tissue (Fig. 1B). One methylated cytosine was detected independently in positions –53, –8, –6, and –113 in 10 screened liver samples. One methylated cytosine was detected in position –94 in 10 screened kidney samples. One methylated cytosine was detected in position +136 in 10 screened germ cell samples. These signals of methylated cytosine may have been background noise. We also examined the core promoter region of EGFP-DsRed transgenes, and methylation was not generally added compared with the intrinsic core promoter sequence (data not shown). These results indicate that the core promoter sequence was not methylated in any tissue, regardless of Gsg2/Aed expression, and may require transcriptional factors not contained in somatic cells.
The bidirectional promoter sequence did not contain consensus DNA sequences recognized by known transcription factors, with the exception of E2F1, AP2, and SP1. To examine whether DNA-binding proteins binding to the Gsg2 promoter sequence occur in germ cell nuclei, the 193-bp (SacI-KpnI) DNA fragment was subjected to a gel retardation assay using testicular germ cells and liver nuclear lysates. Two retardation bands were detected in the nuclear lysates of the germ cells. One of these bands was observed in the lysates of both the germ cells and the liver, whereas the other, in a position indicating high molecular weight, was observed exclusively in germ cell lysate. The gel retardation assay strongly suggested that testicular germ cell nuclear lysate contains specific factors that bind to the Gsg2 promoter sequence (Fig. 8).
|
The Gsg2 cDNA has been identified in many cDNA clones specifically expressed in germ cells using a subtracted cDNA library prepared from supporting cells of germ cell-less mutant (W/Wv) testes and wild-type testes [2]. We identified the nucleotide sequence of full-length Gsg2 cDNA, which encodes Ser/Thr protein kinase, and named it haspin [3]. We showed that Gsg2 (haspin) is predominantly expressed in testis, and specifically in germ cells in mouse testis [3]. Higgins [7], on the other hand, showed that Gsg2 (haspin) is also expressed in adult thymus, bone marrow, fetal tissues, and other tissues. To investigate the mechanisms for germ cell-specific gene regulation, we constructed transgenic mice with the 193-bp genomic fragment from the 5' end of the Gsg2 and Aed genes, fused with EGFP and DsRed genes. Using UV light to examine the expression of fluorescent proteins, we did not detect signals of EGFP or DsRed in any organs of the transgenic mice without haploid germ cells. These results indicate that the 193-bp genomic fragment from the 5' end of the Gsg2 and Aed genes predominantly introduce the germ cell-specific gene expression in the testis. The mRNA and proteins of EGFP and DsRed were detected predominantly in haploid germ cells of adult transgenic mice using Northern blots, Western blots, and histochemical analyses. Although we did not examine detailed expression in the thymus, bone marrow, and so on, we did demonstrate that the 193-bp genomic fragment from the 5' end of the Gsg2 and Aed genes introduces the gene expression of specific cells in tissues, including haploid germ cells. Higgins [7] showed that Gsg2 is weakly expressed in assorted tissues. This suggests that Gsg2 and Aed are expressed very weakly in some mouse tissues or in a limited number of cells. Otherwise, other DNA elements might be needed for expression in other tissues. We have demonstrated that the 193-bp DNA sequence is sufficient for specific, bidirectional, and synchronous expression in germ cells in the testis. These results are important for the study of gene expression in germ cells.
It has recently been reported that many bidirectional promoters occur in the human genome [13]; furthermore, it has been reported that expression of genes expressed in specific tissues [25] and bidirectional genes interfere with each other [26]. Genomic sequencing has revealed that Gsg2 is intronless and integrated into an intron of integrin E (Itgae) [11, 12]. In spermiogenesis, some transcripts are known to contain non-open reading frames (ORFs) [27], and the alternative transcription of integrin E (Aed) does not include any ORFs for translation [7]. Although the physiological meaning of these non-ORF transcripts and the creation of new transcription units still remains unclear, creation of both the Gsg2 gene and these non-ORF transcripts may play an important role in spermiogenesis. In our analyses of germ cell-specific genes, we found that many intronless genes are expressed in spermatogenesis [10]. It was recently reported that the promoter regions of spermiogenesis-specific genes contain CpG islands. Gsg2 is an example of an intronless gene expressed specifically in germ cells, and the promoter region of Gsg2 consists of a GC-rich DNA sequence. Genes bidirectionally transcribed, however, occur head to head on the chromosome, and the promoter regions of these genes have the common feature of a CpG island between the genes. The GC-rich DNA sequence of the Gsg2 promoter region might regulate testis-specific expression of Gsg2. With the transgenic mouse, it was shown that the 193-bp DNA sequence, which contains a GC-rich DNA sequence, is sufficient to predominantly, bidirectionally, and synchronously control the expression of both Gsg2 and Aed mRNA in haploid germ cells. The 1073-bp DNA sequence also controls gene expression similarly to the 193-bp DNA sequences. This 193-bp core promoter sequence also induces mosaic expression of reporter genes in haploid germ cells. This mosaic expression may be induced by the location of reporter genes on chromosomal DNA.
CpG islands of germ cell-specific genes are methylated in somatic cells and demethylated in germ cells [28]. The expression of these genes and Gtf2a1lf is regulated by DNA methylation and is tightly repressed in nontesticular somatic tissue [29]. Tight repression of these genes is regulated by chromatin condensation via the modification of histones. To examine whether demethylation induces bidirectional transcription, liver and fibroblast cells from transgenic animals were cultured in methylcytosine analog 5-azacytidine and/or trichostatin A of a histone deacetylase inhibitor. We did not detect expression of EGFP and DsRed protein using Western blotting. The GC-rich DNA sequence of the Gsg2 promoter region in genomic DNA was not methylated in testis, kidney, or liver. These results indicate that the Gsg2/Aed promoter sequence expression is not regulated by demethylation. Gene groups were identified that are expressed in testes and in cancerous tissues [30]. We found weak activity of these promoter sequences through transfection to cultured cells (HEK 293). These results may indicate that unidentified transcriptional factors may play important roles in spermiogenesis.
Currently, only a small number of short promoters for haploid- or meiosis-specific expression have been characterized [31]. The important consensus sequence appears to be CRE for binding CRE modulator (CREM) 
transcription factor [32]. When the Crem gene is deleted by gene targeting, a large number of postmeiotic genes with the CRE sequence in the upstream control region are not expressed, and spermatogenesis is arrested in early spermiogenesis. Some of the haploid germ cell-specific genes, however, do not necessarily have a CRE motif. Another common feature is a high GC content on the promoter region [31]. When aligned with the promoter sequences that have been tested in transgenic assays, approximately half did not contain a TATA sequence [31], and no consensus sequences were found. The Gsg2/Aed promoter sequence had similar features to the promoters that did not contain DNA sequences known to bind transcription factors, with the exception of common transcription regulatory factors E2F1, AP2, and SP1. Some transcription may occur in nonhistone chromatin during late spermiogenesis. Although the detailed mechanism of transcription regulation during this period is not known, we sought to uncover the mechanism of transcription in nonhistone chromatin during spermiogenesis.
Using the gel shift assay, we examined whether a testicular protein would bind to the Gsg2/Aed promoter sequence. Although the nuclei of spermiogenic cells have chromatin contents very different from that of somatic nuclei, the results indicated that a specific protein complex(es) binds to the Gsg2/Aed promoter sequence.
It was previously reported that human endogenous retrovirus (HERV)-K long terminal repeats (LTRs) have bidirectional promoter activity, according to a luciferase reporter assay in a teratocarcinoma cell line [33]. A large number of solitary HERV LTRs have been introduced during human evolution. At least 8% of the human genome is occupied by ERV sequences [33]. Most of the HERVs sequenced to date represent defective proviruses. Gsg2 may have been constructed in an intron of Itgae by a retrotransposon before the divergence of humans and mice. The transcriptional activity of HERV-K LTR may suggest that the ancient retrovirus Gsg2 integrated into an intron of Itgae with an LTR having bidirectional promoter activity.
The transcript unit of Gsg2 contains features of germ cell-specific genes, including the lack of introns and a GC-rich DNA sequence. Further studies on the regulation of Gsg2 will help detail the transcriptional control of germ cells and their evolutionary history, such as the diversification of the genes by retrotransposons. The bidirectional, specific promoter we identified would have been introduced long ago in our evolutionary history, before the split between humans and mice. Despite the length of time since its introduction, this promoter site is relatively unchanged, and is still actively used in both humans and mice. We do not know whether we can split the 193-bp bidirectional control element into 2 independent promoters for different transcriptional directions in cell-specific expression. We are currently investigating the interacting proteins that control the function of the bidirectional, synchronous promoter activity specific to germ cells.
ACKNOWLEDGMENTS
We thank H. Nishimura, K. Tsunekawa, and M. Ikenishi for technical assistance.
Correspondence: 1Hiromitsu Tanaka, Tanaka Project, Center for Advanced Science and Innovation, Osaka University, 3-1 Yamadaoka, Suita, Osaka 565-0871, Japan. FAX: 81 6 6879 4857; e-mail: tanaka{at}biken.osaka-u.ac.jp
Received: 3 July 2006.
First decision: 4 August 2006.
Accepted: 8 November 2006.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  H. Tang, A. Kung, and E. Goldberg Regulation of Murine Lactate Dehydrogenase C (Ldhc) Gene Expression Biol Reprod, March 1, 2008; 78(3): 455 - 461. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
