Dual-layer transposon repression in heads of

piRNA-mediated repression of transposable elements (TE) in the germline limits the accumulation of mutations caused by their transposition. It is not clear whether the piRNA pathway plays a role in adult, nongonadal tissues in Drosophila melanogaster. To address this question, we analyzed the small RNA content of adult Drosophila melanogaster heads. We found that the varying amount of piRNA-sized, ping-pong positive molecules in heads correlates with contamination by gonadal tissue during RNA extraction, suggesting that most of the piRNAs detected in heads originate from gonads. We next sequenced the heads of wild-type and piwi mutants to address whether piwi loss of function would affect the low amount of piRNA-sized, ping-pong negative molecules that are still detected in heads hand-checked to avoid gonadal contamination. We find that loss of piwi does not significantly affect these 24-28 nt RNAs. Instead, we observe increased siRNA levels against the majority of Drosophila TE families. To determine the effect of this siRNA level change on transposon expression, we sequenced the transcriptome of wild-type, piwi, dicer-2 and piwi, dicer-2 double-mutant heads. We find that RNA expression levels of the majority of TE in piwi or dicer-2 mutants remain unchanged and that TE transcripts increase only in piwi, dicer-2 double-mutants. These results lead us to suggest a dual-layer model for TE repression in adult somatic tissues. Piwi-mediated gene silencing established during embryogenesis constitutes the first layer of TE repression whereas Dicer-2-dependent siRNA-mediated silencing provides a backup mechanism to repress TEs that escape silencing by Piwi.