Genome-wide and cell-specific epigenetic analysis challenges the role of polycomb in Drosophila spermatogenesis.

The Drosophila spermatogenesis cell differentiation pathway involves the activation of a large set of genes in primary spermatocytes. Most of these genes are activated by testis-specific TATA-binding protein associated factors (tTAFs). In the current model for the activation mechanism, Polycomb plays a key role silencing these genes in the germline precursors, and tTAF-dependent activation in primary spermatocytes involves the displacement of Polycomb from gene promoters. We investigated the genome-wide binding of Polycomb in wild type and tTAF mutant testes. According to the model we expected to see a clear enhancement in Polycomb binding at tTAF-dependent spermatogenesis genes in tTAF mutant testes. However, we find little evidence for such an enhancement in tTAF mutant testes compared to wild type. To avoid problems arising from cellular heterogeneity in whole testis analysis, we further tested the model by analysing Polycomb binding in purified germline precursors, representing cells before tTAF-dependent gene activation. Although we find Polycomb associated with its canonical targets, we find little or no evidence of Polycomb at spermatogenesis genes. The lack of Polycomb at tTAF-dependent spermatogenesis genes in precursor cells argues against a model where Polycomb displacement is the mechanism of spermatogenesis gene activation.

Co-transcriptional architecture in a Y loop in Drosophila melanogaster.

The Y loops of Drosophila spermatocytes are formed by the expression of huge individual transcription units on the Y chromosome and their large size provides a unique system for the investigation of the organisation of transcription in intact nuclei. By labelling ribonucleo-protein (RNP) components, the loop chromatin and nascent transcripts in Y loop C, we reveal a highly structured organisation of RNP domains associated with nascent transcripts. We distinguish two types of RNP domain, a proximal domain that runs alongside the chromatin of loop C and a distal RNP domain that wraps around the proximal domain and the loop chromatin. The proximal domain is marked by the Pasilla protein, and separate distal subdomains are marked by the S5 antigen and Boule. We discuss the implications of this highly structured co-transcriptional architecture for the organisation of the process of transcription.