UC Berkeley

Genomes have evolved multiple ways to regulate gene expression. Alternative splicing of genes is a strategy by which more than one transcript and/or protein can be encoded by a single gene. For cases in which expression is harmful, genomes have evolved strategies to repress expression. Epigenetic modifications, and in particular repressive heterochromatin, can repress expression. RNAi, which uses sequence homology to target transcripts for degradation, is another strategy that has evolved to control expression.

I characterized alternative splicing (“AS") within and between Drosophila species, sexes, tissues, and developmental stages. Alternative pre-mRNA splicing (“AS”) greatly expands proteome diversity, but little is known about the evolutionary landscape of AS in Drosophila and how it differs between embryonic and adult stages or males and females. I studied the transcriptomes from several tissues and developmental stages in males and females from four species across the Drosophila genus. I found that 20–37% of multi-exon genes are alternatively spliced. While males generally express a larger number of genes, AS is more prevalent in females, suggesting that the sexes adopt different expression strategies for their specialized function. While the number of total genes expressed increases during early embryonic development, the proportion of expressed genes that are alternatively spliced is highest in the very early embryo, before the onset of zygotic transcription. This indicates that females deposit a diversity of isoforms into the egg, consistent with abundant AS found in ovary. Cluster analysis by gene expression levels shows mostly stage-specific clustering in embryonic samples, and tissue-specific clustering in adult tissues. Clustering embryonic stages and adult tissues based on AS profiles results in stronger species-specific clustering, suggesting that diversification of splicing contributes to lineage-specific evolution in Drosophila. Most sex-biased AS found in flies is due to AS in gonads, with little sex-specific splicing in somatic tissues.

Large portions of eukaryotic genomes consist of repetitive DNA, including transposable elements (“TEs”), whose mobilization can have deleterious effects on the host genome. The establishment of transcription-repressing heterochromatin at repetitive regions during early development safeguards genome integrity. Heterochromatin formation coincides with genome-wide activation of zygotic expression. Males often contain substantially more heterochromatic DNA than females, due to the presence of a large, repeat-rich Y chromosome. I characterized repeat expression over embryonic development in two Drosophila species and found that repeats are expressed at a higher level in males than in females during early embryogenesis. I carried out chromatin immunoprecipitation (ChIP-seq) on male and female early embryos targeting a repressive histone mark associated with transposable elements (H3K9me3), and showed that heterochromatin formation is delayed in early male Drosophila embryos. This coincides with the increase in repeat expression in male versus female embryos, and more de novo insertions of repeats in males. Thus, the Y chromosome may indirectly create a mutational burden in males by postponing the establishment of silencing chromatin marks genome-wide, and permitting expression and mobilization of transposable elements during early embryogenesis.

Piwi-interacting RNAs (piRNAs) are short (23-29 nt) RNAs that bind to PIWI proteins and direct post-transcriptional and transcriptional silencing of target transposons. Many piRNAs are sense or antisense to repeats, and are produced by loci called piRNA clusters. These clusters often contain fragments of many different repeats and act as traps for transposable elements, and piRNAs produced from these clusters use sequence homology to target transcripts of transposable elements for degradation, and induce transcriptional silencing of repeats through heterochromatin formation. Repeat content evolves quickly across species, but less is known how piRNA clusters co-evolve. I sequenced piRNAs from D. melanogaster, D. pseudoobscura, and D. miranda ovaries, testes, and 0-1 hour embryos, to identify and characterize piRNAs and piRNA clusters in these three species. Comparison of piRNA clusters reveals a dynamic picture of their evolution. We identify many piRNA clusters that are unique to a species, or that have newly emerged on the repeat-rich neo-Y chromosome of D. miranda. We also find several homologous clusters that are syntenic between D. miranda and D. pseudoobscura, but whose sequence composition and repeat content has dramatically changed between species. Thus, piRNA clusters co-evolve with the rapidly changing repeat content of species, both by de novo formation of clusters, and by turn-over of the repeats contained within a cluster.