Unistrand piRNA clusters are an evolutionarily conserved mechanism to suppress endogenous retroviruses across theDrosophilagenus

Abstract

The PIWI-interacting RNA (piRNA) pathway plays a crucial role in preventing endogenous genomic parasites, transposable elements (TEs), from damaging the genetic material of animal gonadal cells. Specific regions in the genome, called piRNA clusters, define each species' piRNA repertoire and therefore its capacity to recognize and silence transposons. In the somatic cells of the Drosophila melanogaster ovary, the flamenco (flam) unistrand cluster is the main source of piRNAs and primarily regulates Gypsy family TEs that are able to form virus-like particles and infect neighbouring germ cells. Disruption of the flam locus or failure to process flam precursor transcripts into piRNAs results in sterility, yet it remains unknown whether this silencing mechanism is employed widely across Drosophilidae. Here, using both synteny-based analyses and de novo TE annotation, we identify candidate loci sharing both their organisation and TE targeting repertoire with flam in widely divergent Drosophila species groups. Small RNA-sequencing validated these loci as bona-fide unistrand piRNA clusters and revealed their predominant expression in somatic cells of the ovary, likely to counter TE mobilisation in this tissue. This study provides compelling evidence of co-evolution between virus-like Gypsy family transposons and a host defence mechanism in the form of soma-expressed, unistrand piRNA clusters.

In-cell structure and snapshots ofcopiaretrotransposons in intact tissue by cryo-electron tomography

“…S2A), arguing against a non-canonical, env-independent soma-to-germline transmission process. These data are consistent with the absence of an anti-copia defense program by the host piRNA pathway in follicle cells that efficiently silences the expression of Errantiviruses [45][46][47]: piRNAs targeting copia in somatic ovarian cells are approximately 10-to 100-fold lower than those targeting the different Errantiviruses (Fig. 5A).…”

A dual histone code specifies the binding of heterochromatin protein Rhino to a subset of piRNA source loci

“…5a). We compared our CUT&RUN data for H3K9me3 and H3K27me3 marks with small RNA-seq of ovaries from these species (Lopik et al 2023). We divided the genomes into 10kb bins that were then classified onto 15 groups according to piRNA levels present in each species.…”

FlaHMM: unistrandflamenco-like piRNA cluster prediction inDrosophilaspecies using hidden Markov models

“…Each genomic bin was assigned one of three possible states: none (0), flam-like cluster (1), and centromere-like region (2). As ground truth we used previously reported cluster coordinates (van Lopik et al, 2023). Centromeric regions were defined at chr2 and chr3 as described in Supplementary Methods S1.…”

“…For a long time, flam -syntenic clusters were identified only in species closely related to D. melanogaster (Chirn et al ., 2015; Malone et al ., 2009). We recently reported that flam is evolutionarily conserved beyond the melanogaster subgroup and that flam -like loci exist in species that diverged from D. melanogaster at least 33 million years ago (van Lopik et al ., 2023). This raises the possibility that unistrand piRNA clusters may control Gypsy -family TEs across the whole Drosophila genus.…”