Comprehensive annotation of Physcomitrella patens small RNA loci reveals that the heterochromatic short interfering rna pathway is largely conserved in land plants

Many plant small RNAs are sequence-specific negative regulators of target mRNAs and/or chromatin. In angiosperms, the two most abundant endogenous small RNA populations are usually 21-nucleotide microRNAs (miRNAs) and 24-nucleotide heterochromatic short interfering RNAs (siRNAs). Heterochromatic siRNAs are derived from repetitive regions and reinforce DNA methylation at targeted loci. The existence and extent of heterochromatic siRNAs in other land plant lineages has been unclear. Using small RNA-sequencing (RNA-seq) of the moss Physcomitrella patens, we identified 1090 loci that produce mostly 23- to 24-nucleotide siRNAs. These loci are mostly in intergenic regions with dense DNA methylation. Accumulation of siRNAs from these loci depends upon P. patens homologs of DICER-LIKE3 (DCL3), RNA-DEPENDENT RNA POLYMERASE2, and the largest subunit of DNA-DEPENDENT RNA POLYMERASE IV, with the largest subunit of a Pol V homolog contributing to expression at a smaller subset of the loci. A MINIMAL DICER-LIKE (mDCL) gene, which lacks the N-terminal helicase domain typical of DCL proteins, is specifically required for 23-nucleotide siRNA accumulation. We conclude that heterochromatic siRNAs, and their biogenesis pathways, are largely identical between angiosperms and P. patens, with the notable exception of the P. patens-specific use of mDCL to produce 23-nucleotide siRNAs.