Cellular differentiation entails loss of pluripotency and gain of lineage- and cell-type-specific characteristics. Using a murine system that progresses from stem cells to lineage-committed progenitors to terminally differentiated neurons, we analyzed DNA methylation and Polycomb-mediated histone H3 methylation (H3K27me3). We show that several hundred promoters, including pluripotency and germline-specific genes, become DNA methylated in lineage-committed progenitor cells, suggesting that DNA methylation may already repress pluripotency in progenitor cells. Conversely, we detect loss and acquisition of H3K27me3 at additional targets in both progenitor and terminal states. Surprisingly, many neuron-specific genes that become activated upon terminal differentiation are Polycomb targets only in progenitor cells. Moreover, promoters marked by H3K27me3 in stem cells frequently become DNA methylated during differentiation, suggesting context-dependent crosstalk between Polycomb and DNA methylation. These data suggest a model how de novo DNA methylation and dynamic switches in Polycomb targets restrict pluripotency and define the developmental potential of progenitor cells.

Argonaute proteins of the PIWI clade are central to transposon silencing in animal gonads. Their target specificity is defined by 23-30 nt PIWI interacting RNAs (piRNAs), which mostly originate from discrete genomic loci termed piRNA clusters. Here, we show that a complex composed of Rhino, Deadlock, and Cutoff (RDC) defines dual-strand piRNA clusters genome-wide in Drosophila ovaries. The RDC is anchored to H3K9me3-marked chromatin in part via Rhino's chromodomain. Depletion of Piwi results in loss of the RDC and small RNAs at a subset of piRNA clusters, demonstrating a feedback loop between Piwi and piRNA source loci. Intriguingly, profiles of RNA polymerase II occupancy, nascent transcription, and steady-state RNA levels reveal that the RDC licenses noncanonical transcription of dual-strand piRNA clusters. Likely, this process involves 5' end protection of nascent RNAs and suppression of transcription termination. Our data provide key insight into the regulation and evolution of piRNA clusters.

In animal gonads PIWI-clade Argonaute proteins repress transposons sequence-specifically via bound piRNAs. These are processed from single-stranded precursor RNAs by largely unknown mechanisms. Here we show that primary piRNA biogenesis is a 3′ directed and phased process that, in the Drosophila germline, is initiated by secondary piRNA-guided transcript cleavage.