RNA interference (RNAi) pathways mediate gene silencing by promoting translational inhibition or RNA degradation, and also act at the DNA level by repressing transposon activity and promoting the assembly of heterochromatin. Interestingly, however, the biogenesis of small interfering RNAs (siRNAs), which direct heterochromatin formation, is itself dependent on factors associated with heterochromatin. So which came first? Now, in Cell, Halic and Moazed show that a distinct class of small RNAs, called primal small RNAs (priRNAs), initiates the amplification of siRNAs in the absence of heterochromatin, which in turn induce heterochromatin formation.

The authors used Schizosaccharomyces pombe as a model for investigating RNAi and heterochromatin interactions at pericentromeric regions, which contain dg and dh repeat elements. The RNAi machinery core components are Argonaute 1 (Ago1), RNA-dependent RNA polymerase 1 (Rdp1) and Dicer 1 (Dcr1). Dcr1 processes double-stranded RNA (dsRNA) into siRNAs, which are loaded onto Ago1 and guide the RNA-induced transcriptional silencing (RITS) complex to target sequences through the interaction of siRNAs in Ago1 with nascent RNAs. The RITS complex associates with chromatin by binding histone H3 Lys9 (H3K9)-methylated nucleosomes and recruits the H3K9 methyltransferase Clr4 and the heterochromatin protein 1 homologue Swi6. The silencing signal is amplified by the action of Rdp1 on heterochromatin-bound transcripts and the processing of the resulting dsRNA into siRNAs.

Halic and Moazed found that in Clr4-deficient mutant strains that lack heterochromatin, siRNA levels were strongly reduced, but Ago1 was nevertheless able to bind small RNAs. Likewise, low levels of Ago1-bound small RNAs were present in Dcr1- and Rdp1-depleted cells. High-throughput sequencing experiments showed that many of these small RNAs derive from dg and dh repeats, thus showing that heterochromatin, Dcr1 and Rdp1 are not required for the initial accumulation of small RNAs — which have been termed priRNAs — at these repeat target sites.

Heterochromatin mutants with functional Dcr1 and Rdp1 accumulated high levels of dg repeat siRNAs. This amplification requires the RNA cleavage activity of Ago1, suggesting that an Ago-dependent pathway leads to siRNA amplification without the requirement of pre-existing heterochromatin, and that priRNAs might function as the initial trigger for siRNA amplification.

The authors also found that Ago1 promotes low levels of H3K9 methylation at centromeric DNA in the absence of siRNA amplification. This function is dependent on the ability of Ago1 to bind priRNAs, thus suggesting that priRNAs also trigger heterochromatin assembly.

The fact that priRNAs are enriched in the 3′ ends of annotated transcripts and depend on a partially functional exosome suggested that priRNAs might result from RNA processing events associated with transcription termination and that they are degradation products of single-stranded transcripts that bind Ago1.

This study defines a new class of Dicer-independent small RNAs that function in an Ago-dependent manner to trigger siRNA amplification and heterochromatin formation. The authors argue that priRNAs might associate randomly with Ago1 and function as a surveillance mechanism for abundant transcripts. However, the detailed mechanisms of priRNA biogenesis and whether this process is regulated remain to be determined.