Turnover and lineage-specific broadening of the transcription start site in a testis-specific retrogene

Relocation facilitates the acquisition of short cis-regulatory regions that drive the expression of retrogenes during spermatogenesis in Drosophila

Retrogenes are functional processed copies of genes that originate via the retrotranscription of an mRNA intermediate and often exhibit testis-specific expression. Although this expression pattern appears to be favored by selection, the origin of such expression bias remains unexplained. Here, we study the regulation of two young testis-specific Drosophila retrogenes, Dntf-2r and Pros28.1A, using genetic transformation and the enhanced green fluorescent protein reporter gene in Drosophila melanogaster. We show that two different short (<24 bp) regions upstream of the transcription start sites (TSSs) act as testis-specific regulatory motifs in these genes. The Dntf-2r regulatory region is similar to the known β2 tubulin 14-bp testis motif (β2-tubulin gene upstream element 1 [β2-UE1]). Comparative sequence analyses reveal that this motif was already present before the Dntf-2r insertion and was likely driving the transcription of a noncoding RNA. We also show that the β2-UE1 occurs in the regulatory regions of other testis-specific retrogenes, and is functional in either orientation. In contrast, the Pros28.1A testes regulatory region in D. melanogaster appears to be novel. Only Pros28.1B, an older paralog of the Pros28.1 gene family, seems to carry a similar regulatory sequence. It is unclear how the Pros28.1A regulatory region was acquired in D. melanogaster, but it might have evolved de novo from within a region that may have been preprimed for testes expression. We conclude that relocation is critical for the evolutionary origin of male germline-specific cis-regulatory regions of retrogenes because expression depends on either the site of the retrogene insertion or the sequence changes close to the TSS thereafter. As a consequence we infer that positive selection will play a role in the evolution of these regulatory regions and can often act from the moment of the retrocopy insertion.

The RNA export factor, Nxt1, is required for tissue specific transcriptional regulation

The highly conserved, Nxf/Nxt (TAP/p15) RNA nuclear export pathway is important for export of most mRNAs from the nucleus, by interacting with mRNAs and promoting their passage through nuclear pores. Nxt1 is essential for viability; using a partial loss of function allele, we reveal a role for this gene in tissue specific transcription. We show that many Drosophila melanogaster testis-specific mRNAs require Nxt1 for their accumulation. The transcripts that require Nxt1 also depend on a testis-specific transcription complex, tMAC. We show that loss of Nxt1 leads to reduced transcription of tMAC targets. A reporter transcript from a tMAC-dependent promoter is under-expressed in Nxt1 mutants, however the same transcript accumulates in mutants if driven by a tMAC-independent promoter. Thus, in Drosophila primary spermatocytes, the transcription factor used to activate expression of a transcript, rather than the RNA sequence itself or the core transcription machinery, determines whether this expression requires Nxt1. We additionally find that transcripts from intron-less genes are more sensitive to loss of Nxt1 function than those from intron-containing genes and propose a mechanism in which transcript processing feeds back to increase activity of a tissue specific transcription complex.

Transcription start site evolution in Drosophila

Transcription start site (TSS) evolution remains largely undescribed in Drosophila, likely due to limited annotations in non-melanogaster species. In this study, we introduce a concise new method that selectively sequences from the 5′-end of mRNA and used it to identify TSS in four Drosophila species, including Drosophila melanogaster, D. simulans, D. sechellia, and D. pseudoobscura. For verification, we compared our results in D. melanogaster with known annotations, published 5′-rapid amplification of cDNA ends data, and with RNAseq from the same mRNA pool. Then, we paired 2,849 D. melanogaster TSS with its closest equivalent TSS in each species (likely to be its true ortholog) using the available multiple sequence alignments. Most of the D. melanogaster TSSs were successfully paired with an ortholog in each species (83%, 86%, and 55% for D. simulans, D. sechellia, and D. pseudoobscura, respectively). On the basis of the number and distribution of reads mapped at each TSS, we also estimated promoter-specific expression (PSE) and TSS peak shape, respectively. Among paired TSS orthologs, the location and promoter activity were largely conserved. TSS location appears important as PSE, and TSS peak shape was more frequently divergent among TSS that had moved. Unpaired TSS were surprisingly common in D. pseudoobscura. An increased mutation rate upstream of TSS might explain this pattern. We found an enrichment of ribosomal protein genes among diverged TSS, suggesting that TSS evolution is not uniform across the genome.