| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
Department of Molecular and Cellular Physiology,3 University of Massachusetts Medical School, Worcester, Massachusetts 01655
Department of Cell Biology, School of Medicine,4 University of South Carolina, Columbia, South Carolina 29208
ABSTRACT
Spermatogeniccells elaborate a highly specialized differentiation program that is mediated in part by germ cell-enriched transcription factors. This includes a novel member of the sterol response element-binding factor family, SREBF2_v1/SREBP2gc. Somatic SREBFs are predominantly synthesized as precursor proteins and are critical regulators of cholesterol and fatty acid synthesis. In contrast, SREBF2_v1 bypasses the precursor pathway and has been directly implicated in spermatogenic cell-specific gene expression. During spermatogenesis, SREBF2 precursor transcripts predominate in premeiotic stages, while SREBF2_v1 is highly upregulated specifically in pachytene spermatocytes and round spermatids. In the present study, we demonstrate thatSrebf2_v1mRNAs are present in the testis of several mammalian species, including humans. The basis for the stage-dependent transition in SREBF2 isoforms was also investigated. A 3' rapid amplification of cDNA ends (RACE)-PCR analysis of the rat and human revealed thatSrebf2_v1transcripts are generated by alternative pre-mRNA cleavage/polyadenylation. This involves the use of an intronic, A(A/U)UAAA-independent poly(A) signal within intron 7 of theSrebf2gene. Developmentally regulated competition between germ cell factors that control RNA splicing and pre-mRNA cleavage/polyadenylation may underlie this process. These results define an important role for alternative polyadenylation in male germ cell gene expression and development by controlling a stage-dependent switch in transcription factor structure and function during spermatogenesis. TheSrebf2gene thus provides a useful model to explore the role of alternative polyadenylation in regulating stage-dependent functions of important protein regulators in spermatogenic cells.
developmental biology, gametogenesis, gene regulation, spermatogenesis, testis
INTRODUCTION
The sterol regulatory element-binding factor (SREBF) family of transcription factors plays a key role in controlling cholesterol and lipid metabolism [1]. In somatic cells, the SREBF protein family consists of three members: SREBF1 and SREBF1_v1, which are generated from a single gene by alternative promoter usage and RNA splicing, and SREBF2, which is derived from a separate gene. Somatic SREBFs are synthesized as membrane-bound precursor proteins that contain an N-terminal transactivation domain and conserved basic helix-loop-helix leucine-zip (bHLH-Zip), transmembrane, and C-terminal SREBF cleavage-activating protein (SCAP)-binding domains. SREBF precursors are proteolytically processed to generate mature, transcriptionally active SREBF proteins via a SCAP-dependent mechanism. During this process, the SREBF precursor is initially transported from the endoplasmic reticulum to the Golgi in association with the SCAP. The precursor is cleaved within the Golgi by site 1 and 2 proteases to release the mature SREBF protein containing the N-terminal transactivation and bHLH-Zip DNA-binding domains. The mature SREBF is then transported into the nucleus via the ß-importin pathway [2], where it regulates its response element (SRE)-containing target promoters.
Spermatozoa are highly specialized, motile cells that are formed in the testis through a complex series of proliferative and differentiative stages: mitotic spermatogonia, meiotic spermatocytes, and spermiogenic spermatids. Their proper development requires the elaboration of a unique gene expression program that is distinguished in several respects from somatic cells. For example, spermatogenic cells synthesize a large number of cell-specific and stage-dependent mRNAs that are generated by multiple mechanisms, including alternative transcriptional initiation, splicing, and polyadenylation [3, 4]. Expression of germ cell-specific transcription and RNA processing factors appears to play a central role in these events [5, 6].
We recently identified a spermatogenic cell-specific SREBF isoform, SREBF2_v1/SREBP2gc, that is highly upregulated in pachytene spermatocytes and round spermatids [7]. This transcription factor appears to have a unique function as a determinant of germ cell-specific gene expression. SREBF2_v1 was recently shown to regulate the promoter for the spermatogenic cell-specific proacrosin gene, which is exclusively expressed in meiotic spermatocytes and haploid spermatids [8]. Also, its SREs are important for proacrosin promoter expression in spermatogenic cells in vivo [8]. SREBF2_v1 is the first germ cell-enriched transregulator shown to regulate a cell-specific gene expressed in meiotic spermatocytes.
The synthesis of mature SREBF2 in somatic cells is under homeostatic control via a sterol-feedback inhibition mechanism [9]. When cellular cholesterol levels are adequate, sterols block the SCAP-dependent transport of SREBF2 precursor proteins from the endoplasmic reticulum to the Golgi, thus blunting the generation of the mature transfactor. Conversely, depletion of cholesterol allows SREBF2 precursor transport and processing to occur, leading to increased formation of the mature SREBF2 transcription factor and activation of cholesterol synthesis genes. In contrast, SREBF2_v1 is synthesized in a sterol-independent fashion during spermatogenesis. Its mRNA encodes a truncated protein consisting of the N-terminal transactivation and bHLH-Zip DNA-binding domains but missing the membrane-spanning and C-terminal regulatory domains [7]. Thus, SREBF2_v1 is a soluble, constitutively active protein able to directly activate target genes in spermatocytes and spermatids. Further, its lack of dependence on precursor processing ensures that SREBF2_v1 protein levels in spermatogenic cells are insensitive to intracellular cholesterol levels [7].
Formation of mRNA 3' ends in metazoans requires the interaction of the pre-mRNA cleavage/polyadenylation complex with a bipartite poly(A) recognition signal (reviewed in Proudfoot et al. [10]). For the majority of somatic mRNAs, a canonical A(A/U)UAAA hexamer sequence is located upstream of the RNA cleavage site, while a more poorly defined U- or GU-rich region is located 20–30 nucleotides downstream. The downstream U/GU-rich element is bound by cleavage stimulatory factor (CstF), which is specifically required for the pre-mRNA cleavage step. The upstream hexamer sequence binds cleavage/polyadenylation specificity factor (CPSF), which directly interacts with poly(A) polymerase. Interestingly, the 3' ends of many male germ cell-specific transcripts are generated by one of several non-A(A/U)UAAA poly(A) signals [11]. The precise mechanisms responsible for this cell-specific polyadenylation in the male germ line remain uncertain. However, several components of the pre-mRNA cleavage/polyadenylation complex exhibit testis- or spermatogenic cell-specific expression [11–13].
SREBF2 isoforms are expressed in distinct stage-dependent patterns during spermatogenesis. Mouse spermatogonial stages express mainly Srebf2 precursor mRNAs, while the smaller Srebf2_v1 transcript and protein predominate in late meiosis and spermatogenesis [7]. This indicates distinct, stage-dependent requirements and functions for the sterol-sensitive and -insensitive SREBF2 isoforms in the male germ line. A key question raised by these observations is how the cell- and stage-dependent formation of SREBF2 isoforms is controlled during sperm cell development. In the present study, we demonstrate that this transition in Srebf2 mRNA and SREBF2 protein isoforms is determined by a developmental switch in mRNA 3'-end formation.
MATERIALS AND METHODS
Spermatogenic Cell Preparations
Adult male germ cells enriched in pachytene spermatocytes and round and condensing spermatids were prepared from the testes of Sprague-Dawley rats and CD-1 mice by enzymatic digestion, essentially according to Bellve et al. [14]. Purified spermatogenic cell types were prepared from either immature or mature mouse testes by unit gravity sedimentation as previously described [14, 15]. Purities of various germ cell types ranged from 90% to 96%. All studies reported that involved animals were conducted according to the Guide for Care and Use of Laboratory Animals by the National Academy of Science.
RNA Extraction, RT-PCR, and 3' RACE-PCR
Total RNAs from spermatogenic cells and tissues were isolated with Tri-Reagent (Sigma, St. Louis, MO). For RT-PCR, RNAs were analyzed by the Titan One-Step RT-PCR kit (Roche Applied Science, Indianapolis, IN). Total RNAs from human tissues were purchased from Clontech (Mountain View, CA). For 3' RACE-PCR, total RNA was amplified by a Gene Racer kit (Invitrogen, Carlsbad, CA) with upstream and nested gene-specific sense strand primers. PCR products were subcloned into pCR4-TOPO vectors for sequencing by the TOPO TA cloning kit (Invitrogen). Primer sequences used in RT- and RACE-PCR are available upon request.
Northern Blots
Total RNA from cells and tissues was run on formaldehyde-containing agarose gels and transferred to Gene Screen Plus membranes (NEN Life Science, Boston, MA). Equivalency of loading and transfer was verified by ethidium bromide staining. A cDNA probe for rat Srebf2_v1 [7] was labeled with 
-32P-dCTP (deoxycytidine 5'-triphosphate) by the Prime-a-Gene Labeling System (Promega, Madison, WI). Membranes were hybridized with a probe (2 x 106 cpm/ml), washed, and subjected to autoradiography as described previously [16]. Ratiometry of Srebf2_v1/Srebf2 mRNAs was performed by ImageJ software (National Institutes of Health).
RESULTS
A Transition in the Pattern of Srebf2 mRNAs During Early Meiosis
It was previously shown that the ratio of Srebf2_v1 to Srebf2 precursor transcripts markedly increased between spermatogonial and late meiotic (pachytene) spermatocyte stages in the mouse [7]. Preleptotene spermatocytes are an early transitional stage between spermatogonia, from which they arise, and later meiotic prophase spermatocytes. There was a significant decrease in the abundance of the 5.2/4.2-kb Srebf2 precursor transcripts in preleptotene cells relative to spermatogonial stages (Fig. 1A). The relative proportion of the 2.5-kb mouse Srebf2_v1 mRNA is enhanced during the subsequent progression to early (prepuberal) pachytene spermatocytes, and this mRNA predominates in late (adult) pachytene spermatocytes and spermatids (Fig. 1A). Thus, there is a gradual shift in the relative amounts of the Srebf2_v1 and precursor mRNAs as meiosis progresses.
|
A similar developmental pattern was observed for rat male germ cells. Previous studies demonstrated that enriched adult rat spermatogenic cells, consisting mainly of pachytene spermatocytes and round, elongating, and condensing spermatids, express a 3-kb Srebf2_v1 mRNA in addition to the larger 5.2- and 4.2-kb Srebf2 precursor transcripts [7]. Srebf2 precursor mRNAs were the major transcripts, with very little Srebf2_v1 mRNA detected, in postnatal Day 18 rat testes (Fig. 1, B and C) when spermatogonia and earlier spermatocyte stages predominated [17, 18]. However, the ratio of Srebf2_v1 to Srebf2 precursor mRNAs steadily increased on Days 24 through 31, as late pachytene spermatocytes gradually become abundant and round spermatids appear, and was markedly elevated (>2-fold) on Day 40 in adult rat testes when haploid spermatids are the major rat testicular cell types [17].
Srebf2_v1 Transcripts in Different Species
The rat Srebf2_v1 and Srebf2 precursor protein mRNAs share common 5'-untranslated regions (UTRs) as well as the first 460 codons. However, beyond this point, the rat Srebf2_v1 mRNA contains unique sequences that truncate the SREBF2_v1 coding region [7]. Rat SREBF2_v1 is highly similar to another truncated SREBF2 protein identified in a hamster cell line specifically selected for sterol resistance [19]. Since the novel 3'-UTR of the hamster transcript was previously shown to arise from an unspliced intron, it is possible that rat Srebf2_v1 mRNA was also generated by alternative RNA splicing [7]. To explore these mechanisms and their universality in greater detail, we further examined the 3'-UTRs for Srebf2_v1 transcripts for the human, rat, hamster, and mouse.
Initially, we used RT-PCR to examine the predicted unspliced exon-intron junction neighboring the 3'-UTR for mouse, hamster, and human Srebf2_v1 transcripts, as previously performed for the rat mRNA [7]. With testis RNA, single PCR bands were generated with lengths of 
400 bp (mouse and hamster) and 550 bp (human), while no PCR products were detected in somatic tissues such as the liver (Fig. 2A). Sequencing of these PCR products showed that all three were identical to the predicted coding region and unspliced Srebf2 intron sequences (Fig. 2B). On the basis of GenBank data for the mouse, human, and rat Srebf2 genes, these sequences occur at the junction of exon 7 and intron 7 for the Srebf2 precursor transcript. Because of the retained intronic region, the encoded SREBF2_v1 protein extends 1 (rat and hamster), 8 (human), and 18 (mouse) amino acid residues beyond the preceding Srebf2 exon 7 sequence (Fig. 2C).
|
Comparison of Srebf2_v1 mRNA 3' Ends
To further establish how Srebf2_v1 mRNAs are formed in male germ cells, 3'-UTR sequences were determined by 3' RACE-PCR. Two PCR products were generated for the human SREBF2_V1 transcripts, one being more predominant (data not shown). Sequencing of these two products showed that both were derived entirely from intron 7 of the human SREBF2 gene (Fig. 3). However, they terminated at different positions within this intron. The 3'-UTR for the previously characterized rat Srebf2_v1 cDNA also was wholly contained within intron 7 of the rat Srebf2 gene (Fig. 3). Interestingly, intron 7 for the rat and human Srebf2 genes lacked canonical A(A/U)UAAA hexamer poly(A) signals upstream of the Srebf2_v1 mRNA 3' ends. However, multiple upstream noncanonical poly(A) signals were present in each case that highly resembled those identified for other germ cell-enriched mRNAs [11] (Fig. 3). Furthermore, potential U/GU-rich binding sites for CstF were present at variable distances downstream of the identified 3' ends for both the human and rat Srebf2_v1 mRNAs (Fig. 3). Together, these findings indicate that the human and rat Srebf2_v1 mRNAs arise by the selection of alternative, A(A/U)UAAA-independent poly(A) signals (Fig. 4A). This results in an extended terminal exon for the Srebf2_v1 transcript (exon 7gc) and the formation of a truncated mRNA and protein. Comparison of genomic sequences also revealed that the exon-intron organization of the Srebf2 gene is generally conserved in the rat, human, and mouse (Fig. 4A).
|
|
RACE-PCR studies to identify 3'-end sequences for mouse Srebf2_v1 transcripts did not yield discrete products, apparently because of the presence of a 30-bp GC-rich region that was approximately 222-bp downstream from the start of exon 7gc (data not shown). Attempts to overcome this problem under modified RACE-PCR conditions (including thermostable reverse transcriptase), as well as primers to regions downstream of the GC-rich domain within intron 7, also proved unsuccessful. The location of the alternative 3' end for Srebf2_v1 mRNA in mouse germ cells was not further examined.
The conservation of alternative polyadenylation sites within intron 7 of the Srebf2 precursor among rats and humans suggests that truncation of Srebf2_v1 mRNA specifically within this region is functionally important. For example, the upstream intron 6 lacks putative polyadenylation signals, either canonical or noncanonical (data not shown). This may be explained by the exon distribution of different protein domains for SREBF2_v1 (Fig. 4B). Domains critical for transcriptional activity (transactivation and DNA binding) are wholly present in exons 1–6. While sequences required for dimerization and nuclear import are largely contained within the HLH-Zip DNA-binding domains within exon 6, they extend beyond this into exon 7 as well [20] (Fig. 4B). Further, exons located 3' of intron 7 encode transmembrane and related domains that tether the SREBF2 precursor to the membrane and mediate its interactions with SCAP. On the basis of this information, transcript termination within intron 7 specifically provides the means for generating a soluble, fully active SREBF2_v1 transregulator (Fig. 4B).
DISCUSSION
SREBF2 isoforms are expressed in differing developmental patterns during spermatogenesis, suggesting that they have distinct, stage-dependent functions. Precursor-derived mature SREBF2, which predominates in premeiotic stages, has a critical function in controlling cholesterol metabolism. However, SREBF2_v1 was recently shown to regulate the spermatogenic cell-specific proacrosin promoter that is expressed in a stage-dependent manner in both spermatocytes and spermatids [8]. These observations suggest that SREBF2_v1 has been adapted to regulate germ cell-specific gene expression. This hypothesis is also supported by the differing responsiveness of SREBF2 isoforms to cellular cholesterol concentrations: mature SREBF2 is subject to direct feedback inhibition of its proteolytic formation, whereas constitutively active SREBF2_v1 in meiotic and haploid germ cells is not [7]. Sterol-independent synthesis of SREBF2_v1 indicated that mechanisms other than variation in cellular cholesterol levels determined its developmental expression during spermatogenesis. The present findings show that cell-specific, stage-dependent alternative 3'-end formation governs the switch in SREBF2 isoforms and that these mechanisms are conserved in several mammalian species, including humans. In part, this switch in isoforms may be important as a means to developmentally upregulate the expression of cell-specific SRE-dependent genes in later germ cell stages, which would otherwise be difficult to effect via the sterol-dependent negative feedback mechanism.
The proximal polyadenylation signal within intron 7 of the Srebf2 gene consists of a noncanonical, A(A/U)UAAA-independent signal characteristic of many spermatogenic cell-specific transcripts [11, 21]. These transcripts are frequently enriched in pachytene spermatocytes, spermatid stages, or both. Cell- or tissue-specific 3'-end formation also has been implicated in numerous examples of testis-enriched versus somatic mRNA isoforms [22–25]. Alternative polyadenylation also occurs for the germ cell-enriched transcription factors CREM (cAMP-responsive element modulator), active tau variant, and NR6A1, resulting in the differential stage-dependent expression of mRNA isoforms having 3'-UTRs of distinct length [26, 27]. In general, the physiological significance of alternative polyadenylation during spermatogenesis has not been fully investigated. In several instances, increased mRNA stability has been implicated for shorter germ cell isoforms [24, 26]. Alternative poly(A) site usage also may provide an important mechanism for controlling the cellular specificity and/or stage dependence of mRNA synthesis. Importantly, the present study demonstrates that alternative polyadenylation has a central role in controlling the stage-dependent expression of distinct transcription factor isoforms during spermatogenesis. It is likely that additional germ cell proteins undergo similar developmental alterations in structure/function due to alternative 3'-end formation.
Several components of the pre-mRNA cleavage/polyadenylation complex are uniquely expressed in the testis and male germ line. For example, testis-specific CSTF2T markedly increases in pachytene spermatocytes and round spermatids, coincident with the increased usage of noncanonical poly(A) signals during spermatogenesis [11]. The ubiquitous form of CSTF2 is also expressed at elevated levels in meiotic (mouse) or meiotic and postmeiotic (rat) stages [12, 28]. PAPOLB is a testis-specific, cytoplasmic poly(A) polymerase that is essential for proper mRNA polyadenylation in spermatids and for the completion of spermatogenesis [13]. These observations suggest that the nature and amounts of cleavage/polyadenylation complexes differ considerably in the meiotic and haploid stages of spermatogenesis relative to somatic tissues. How these RNA processing proteins interact in male germ cells with direct cell-specific mRNA 3'-end formation remains to be determined.
An interesting parallel to the developmental control of SREBF2 formation in spermatogenic cells is the regulation of immunoglobulin heavy chain M (IgM) synthesis during B-lymphoid maturation [29, 30]. Alternative pre-mRNA processing results in the formation of either secreted or membrane-associated IgM proteins (µs and µm, respectively). The µs form is generated by a canonical cleavage/polyadenylation signal within an intron, while the µm protein is produced by splicing out the intronic µs poly(A) site and using a downstream canonical poly(A) signal. In undifferentiated B cells, the amounts of the two mRNAs are similar, while the µs mRNA increases [31], and enhanced accumulation of CSTF2 protein in differentiated B cells has been directly implicated in the shift toward the secretory form due to alternative polyadenylation [32]. In addition, developmentally regulated expression of splicing-associated factors appears to contribute to alternative polyadenylation by interfering with CSTF2 binding to IgM pre-mRNA and RNA cleavage activity [33, 34]. Thus, competition between cleavage/polyadenylation and splicing activities is apparently important for the developmental expression of alternative IgM mRNAs during B-cell development [35].
A similar competition between splicing and cleavage/polyadenylation pathways also may determine the developmental switch in Srebf2 mRNAs. As noted earlier, CSTF2 gene expression is 250-fold higher in the testis relative to the somatic tissues in the mouse [12]. Thus, increased expression of CSTF2 and CSTF2T proteins in late meiotic and/or haploid germ cells may promote a similar shift toward the formation of Srebf2_v1 transcripts at the expense of the precursor isoforms. Further, several general RNA splicing components, including Rbm (RNA-binding motif) and multiple hnRNPs (heterogenous nuclear ribonucleoproteins), are downregulated in the later stages of mouse spermatogenesis [36–38]. This also may contribute to the increased ratio of Srebf2_v1:precursor mRNAs during late meiosis and spermatogenesis.
The present studies define a highly useful model for defining the molecular components that control alternative polyadenylation during spermatogenesis. Preliminary studies with a human SREBF2 minigene indicate that overexpression of CSTF2T alone is not sufficient to shift the ratio of splicing:polyadenylation in transfected cell cultures (data not shown). This suggests the need for additional cleavage/polyadenylation components, alterations in splicing-related factors, or both in this process. Analyses of the cis-elements required for stage-dependent expression of the two transcript forms will shed light on the mechanisms involved.
ACKNOWLEDGMENTS
We thank Mr. George Gagnon for his invaluable technical assistance in the present study.
FOOTNOTES
1 Supported by Public Service grant R01 HD045723. 
2 Correspondence. FAX: 508 856 5997; daniel.kilpatrick{at}umassmed.edu
Received: 15 March 2006.
First decision: 30 March 2006.
Accepted: 8 May 2006.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  B. L. Sartini, H. Wang, W. Wang, C. F. Millette, and D. L. Kilpatrick Pre-Messenger RNA Cleavage Factor I (CFIm): Potential Role in Alternative Polyadenylation During Spermatogenesis Biol Reprod, March 1, 2008; 78(3): 472 - 482. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
