Ricochet transcription of pachytene piRNA promoters in mammals

The origin and stability of germline noncoding RNAs remain largely uncharacterized in mammals. Here, we demonstrate that the vast majority of pachytene piRNAs originate from less than 100 bi- and unidirectionally transcribed loci in the mouse genome. These loci show features characteristic of transcribed protein-coding genes, including primary transcription by RNA polymerase II, H3K4 trimethylation marking the transcription initiation site, and H3K36 trimethylation present at elongation regions in a highly tissue-specific manner. We identified MYBL1, a spermatocyte-enriched transcription factor, as a potential driver of piRNA precursor transcripts. By similarly mapping piRNA clusters in testes of dog and opossum, we revealed that the regulatory architectures generating piRNAs appear to be highly conserved in mammals. Finally, we discovered that apparently neutral chromosomal rearrangements in evolution can act as an unexpected mechanism to generate novel piRNA clusters in diverse mammals. Thus, our results demonstrate how an integrated, multi-species approach can help define the genomic features and evolutionary mechanisms underlying mammalian germline noncoding RNAs.