We have mapped the transcriptome of mouse germ cells at three critical stages of spermatogenesis, namely type A spermatogonia at mitosis, pachytene spermatocytes at meiosis, and round spermatids at post-meiotic differentiation. During this past year, we applied various computational and biochemical analyses to annotate functional elements and to identify the interactions among genes and related molecules. By applying unsupervised hierarchical clustering, genes represented by SAGE tags preferentially expressed at different stages of germ cell development were identified. The promoter elements of the clustered genes and the biological processes represented in each cluster were analyzed. Thirty six biological processes found to be over-represented in spermatogonia based on the ontology of clustered genes included ribosome biogenesis and assembly, integrin signaling pathway, carbohydrate metabolism, protein biosynthesis and RNA processing. Twenty one gene ontology categories found to be over-represented in pachytene spermatocytes included chromosome segregation, cell cycle, cytoskeleton and ubiquitin regulation while 26 gene ontology categories enriched in round spermatids were mainly related to cellular physiological process and metabolic process related to the ubiquitin cycle, proteolysis and peptidolysis. Aside from the known transcription regulators such as NF-fUB, SP1, AP-1, and EGR, novel promoter elements such as E-box in spermatogonia genes, GATA in spermatocyte genes, and Gut-enriched Krueppel like binding factor site in spermatid genes were also observed. To visualize the relationship among the clustered genes, potential biological networks were constructed linking the gene candidates with neighboring genes, proteins, transcription factors and small molecules. This exercise allowed the identification of in silico targets for regulating spermatogenesis and the generation of representative """"""""!signature networks!"""""""" essential for specific stages of spermatogenesis. These results will provide a foundation for further in vitro and in vivo studies on germ cell development. There are numerous reports of the identification of novel transcript variants in germ cells. However, none of the reports addressed the issue of stage-specificity of alternative splicing and how this is achieved during development. Other than solely providing expression information, we examined the idea of discovering alternative splicing variants on a genome-wide scale by extracting information from the germ cell SAGE libraries. A systematic search for transcript variants with alternative 3!| end usage was performed. Seventy four genes were found to have 3'end alternative splicing variants (3'ASV) in spermatogonia. Similarly, 58 and 62 genes had 3'ASV in spermatocytes and spermatids, respectively. There were also genes with different 3'ASV in two cell stages; 207 genes with 3'ASV expressed in spermatogonia and spermatocytes, 249 genes with 3'ASV in spermatogonia and spermatids, and 158 genes with 3'ASV in spermatocytes and spermatids. Seventy three genes were found to produce different 3'ASV specific for each cell stage. Among these were novel variants of genes involved in developmental and transcriptional controls, such as those of heat shock protein 4 (Hsp4), H3 histone, family 3B (H3f3b) and ubiquitin protein ligase E3A (Ube3a). Studies are underway to determine the functional elements that regulate stage specific splicing. Potential structural modifications on protein products will also be studied. Even though antisense transcription has been known to occur in prokaryotes for many years, the widespread occurrence of antisense transcripts in humans and mice has only been documented recently. A number of processes in spermatogenesis such as genomic imprinting, translation repression, stage-specific alternative splicing, etc., are frequently associated with antisense transcripts. However, a systematic search for antisense transcripts in spermatogenic cells has not been reported. Mapping of the transcriptome of mouse germ cells by SAGE offers a unique opportunity to examine the occurrence of antisense transcription during spermatogenesis. Among 64 differentially expressed genes 19 were shown to have antisense transcripts with orientation specific RT-PCR and nucleotide sequencing. These antisense transcripts arose through a variety of mechanisms, including transcription of the sense mRNA in the cytoplasm, transcription of the opposite strand of the sense gene locus, transcription of a pseudogene, or transcription of neighbouring genes and the intergenic sequence. Expression studies of the sense and anti-sense transcript of 9 genes showed the testicular levels of the sense transcripts to be higher than that of the antisense transcripts while the relative distribution of these transcripts in non-testicular tissues was variable. This study showed that antisense transcription was prominent among germ cell genes. To facilitate the study of the interaction between the sense and the antisense transcripts, we had established several mouse cell lines as models for studying the functional activities of the cloned anti-sense transcripts. A defective Lin28 results in precocious larval stage progression. In contrast, over-expression of Lin28 leads to retarded development owing to reiteration of the earlier larval stage. We hypothesize that the mouse Lin28 homolog (mLin28) would have a similar role in regulating the decision between germ cell proliferation and differentiation. To this end, we are delineating the mode of regulation of mLin28 expression at both transcriptional and post-transcriptional levels. By performing RACE experiments, we observed a differential use of transcription start sites of mLin28 transcripts as germ cells differentiate. Using promoter element prediction algorithms, we identified specific promoter elements and modules which were known to elicit a transcriptional activation effect. In similar experiments, we analyzed the 3'end of mLin28 transcripts and found that they were heterogeneous in nature: we had isolated alternative 3!| ends of the transcripts which displayed stage-specific expression patterns. We also observed the use of both canonical and unconventional polyadenylation signals. We hypothesize that specific fragments of the mLin28 3'UTR are involved in modulating the translation of the encoded protein. Specifically, we predicted the presence of potential binding sites for a subset of microRNAs and other regulatory factors on the 3'UTR. The relationship between the elongation/shortening of the 3'ends and the gain/loss of modulation is now being tested. We expect the elucidation of these regulation pathways will explain at the genetic level the testis-specific, as well as, the spermatogenic stage-specific expression pattern of mLin28. We cloned a novel mArd1 homolog, which we named mArd2 that demonstrated testis-specificity and elevated expression in pachytene spermatocytes. Earlier studies in yeasts have identified a diverse role for ARD1 from cell cycle regulation to DNA repair and recombination. We found that the expression of mArd2 protein was delayed with respect to the transcript expression pattern. Interestingly, both mLin28 and mArd2 transcripts caried very long 3'UTRs. We hypothesize that the repression of mArd2 protein would be related to th related to the presence of regulatory elements on the 3' UTR of the transcripts. We identified several regulatory elements known known to mediate translational repression effect on the 3' UTR of mArd2. We also identified potential microRNA binding sites present on the mArd2 3'opUTR. The mechanism of regulation of mArd2 expression will be tested in a similar fashion as that of mLin28.
Showing the most recent 10 out of 14 publications
