Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Dev Cell. Author manuscript; available in PMC 2010 August 23.
Published in final edited form as:
PMCID: PMC2925128

PIWI Goes Solo in the Soma


The recently discovered piRNA pathway serves to protect the genome from transposon activity in the germ-line. Now Li et al. and Malone et al. in a recent issue of Cell show that the piRNAs are made by more than one means and that their defensive function extends into the germline’s circumjacent soma.

Since the realization in 1993 that short, noncoding RNA molecules can regulate gene expression through a process termed RNAi (RNA interference), the number of known short RNA regulatory pathways has dramatically multiplied, and our understanding of their biological functions has greatly improved. These short RNAs now include many classes: microRNAs, casiRNAs, tasiRNAs, natsiRNAs, exo-siRNAs, piRNAs, and endosiRNAs. Each short RNA provides target recognition to silencing effectors in the Argonaute family of proteins, many of which have endonucleolytic activity. piRNAs are named for the Argonautes they bind, members of the PIWI family. In the fly, the PIWI family proteins are Argonaute3 (Ago3), Aubergine (Aub), and PIWI (Ghildiyal and Zamore, 2009). In a recent issue of Cell, the Zamore and Hannon labs add to the cacophony of silencing by identifying three subgroups of piRNAs, two in the germline, and, surprisingly, one in the adjacent soma. Each group is produced by different means, but all belong to a small-RNA-based immune system that suppresses mobile elements in the Drosophila genome (Li et al., 2009; Malone et al., 2009).

Both siRNAs and miRNAs are cleaved from longer transcripts by RNAse III enzymes called Dicers. piRNAs, by contrast, are produced by Dicer independent means, are longer than either siRNAs or miRNAs (26–30 rather than 21 nucleotides), are mostly antisense, and, like siRNAs, but unlike miRNAs, have 2′O methylation on the 3′ termini, (Vagin et al., 2006). The less-abundant sense piRNAs generally associate with Ago3 and are often complementary to PIWI-bound antisense piRNAs in their first ten nucleotides. Furthermore, the Aub-bound antisense piRNAs typically have a 5′ uracil corresponding to the adenine commonly found at the tenth nucleotide of the Ago3-bound sense piRNAs. These observations led to the so-called “ping-pong” model for piRNA biogenesis. In this model, sense oriented transcripts of genomic repeats with sequence similarity to transposons, or transcripts of transposons themselves, provide a template that is bound by Ago3. The sense templates guide Ago3 to cleave antisense transcripts, producing antisense piRNAs, which then bind Aub, guiding them to produce more sense piRNA (Brennecke et al., 2007). PIWI, however, binds predominately antisense piRNAs that do not exhibit a strong ping-pong signature.

In the fly genome, piRNAs map to transposons or to repetitive sequences resembling transposons. Mutations in piRNA pathway components lead to elevated levels of transposon transcripts in the germline and to developmental abnormalities. These defects are relieved by mutations in DNA damage checkpoint genes, suggesting that piRNAs normally repress transposon activity that would otherwise promote clastogenesis in the germline and consequently activation of checkpoint mechanisms and developmental aberrations (Klattenhoff et al., 2007). The germline hasa greater need for transposon suppression than somatic tissues since only the germline passes the genome to the next generation. In fact, both the piRNAs themselves and the machinery that produces them are passed to the next generation through the female germline to ensure the continued suppression of transposons (Brennecke et al., 2008).

Both the Zamore and Hannon groups provide support for the ping-pong model in piRNA biogenesis in the germline using deep sequencing of piRNAs from ovaries and early embryos of Ago3 and Aub mutants. They identify two piRNA subgroups whose production depends on Ago3 or Aub and requires ping-pong amplification. Group I is predominantly antisense while group II is predominantly sense.

They also identified a third group of antisense PIWI-binding piRNAs, produced independently of Ago3 or Aub. Two facts suggested to the Zamore lab that group III piRNAs may be soma specific. First, PIWI is found in the nuclei of somatic and germ cells while Aub and Ago3 are found only in the germline, and second, group III overwhelmingly consists of sequences from the flamenco locus. The flamenco locus contains gypsy class retrotransposons that are active in the somatic tissues around the germline and are known to invade the germline as encapsulated retroviral particles. Consistent with a somatic origin for flamenco-derived piRNAs, the Hannon lab detected the piRNAs in the ovaries, but not in young embryos. Since the Drosophila egg chamber loses its somatic envelope late in development, this finding is consistent with the idea that flamenco derived piRNAs originate in the soma. While it makes intuitive sense for Drosophila to have a system for fighting the gypsy retrotransposons in the soma, before they invade the germline, conclusive evidence for this piRNA function will require further experiments.

One intriguing difference between the germline and putative somatic piRNA systems is a perinuclear organelle found only in the germline, nuage. Nuage is posited to serve as a clearing house for mRNAs exported from the nucleus, and the protective function of piRNAs against unwanted transposon transcripts is thus consistent with the presence of piRNA pathway components (including Aub and Ago3) in the nuage (Klattenhoff et al., 2007). PIWI, in contrast, is localized to the nucleus in both the germline and the soma. So it seems likely that this game of ping-pong is played by Aub and Ago3 in the nuage.

To gain insight into the function of piRNA pathway components, Malone et al. (2009) examined the effects of mutating various genes previously implicated in the piRNA pathway, as well as mutation of the flamenco locus. Four of the genes are essential for localization of Ago3 to the nuage—Aub, krimper, spindle-E, and vasa. Ago3, conversely, is required for Aub’s placement in the nuage. Interestingly, previous work showed interdependence among Maelstrom (a putative piRNA pathway component), Aub, and spindle-E for nuage localization, suggesting that nuage may be dedicated to RNA processing via short RNA mechanisms (Findley et al., 2003). Three genes, in addition to Ago3 and Aub, were found to play a role in ping-pong amplification: spindle-E, krimper, and zucchini. The flamenco locus mutation did not affect piRNAs that exhibit a ping-pong signature, confirming the Ago3 and Aub independent function of somatic piRNA biogenesis.

The combination of mutational analysis and deep sequencing in these papers thus provides evidence for three subgroups of piRNAs: two in the germline, produced by Ago3/Aub dependent ping-pong amplification, and one in the soma, produced without the ping-pong amplification step. Group I is predominantly antisense, group II is predominantly sense, and the somatic group III is antisense and independent of Ago3 and Aub.

Important questions about piRNAs remain unanswered. First, it is unclear whether we can generalize from flies to mammals. Although PIWI-related proteins are found in mammals, piRNA loci are devoid of transposons, and the structure of mammalian repetitive elements is different from that in flies. Sequences producing piRNAs in mammals are reported to be the fastest evolving loci in the mammalian genome, suggesting that powerful selection is acting on these loci (Assis and Kondrashov, 2009). We also do not know how piRNAs are directed to specific PIWI family proteins. Nor do we understand whether transcription of the antisense strand, on which the ping-pong model depends, is constitutive, stochastic, or actively regulated. Finally, we still do not know how piRNA suppress transposons: does the degradation of transposon transcripts prevent their proliferation, or do piRNAs specify silencing by chromatin modifications, or perhaps both? We anticipate surprises.


  • Assis R, Kondrashov AS. Proc. Natl. Acad. Sci. USA. 2009 in press. Published online April 8, 2009. 10.1073/pnas.0900523106. [PubMed]
  • Brennecke J, Aravin AA, Stark A, Dus M, Kellis M, Sachidanandam R, Hannon GJ. Cell. 2007;128:1089–1103. [PubMed]
  • Brennecke J, Malone CD, Aravin AA, Sachidanandam R, Stark A, Hannon GJ. Science. 2008;322:1387–1392. [PMC free article] [PubMed]
  • Findley SD, Tamanaha M, Clegg NJ, Ruohola-Baker H. Development. 2003;130:859–871. [PubMed]
  • Ghildiyal M, Zamore PD. Nat. Rev. Genet. 2009;10:94–108. [PMC free article] [PubMed]
  • Klattenhoff C, Bratu DP, McGinnis-Schultz N, Koppetsch BS, Cook HA, Theurkauf WE. Dev. Cell. 2007;12:45–55. [PubMed]
  • Li C, Vagin VV, Lee S, Xu J, Ma S, Xi H, Seitz H, Horwich MD, Syrzycka M, Honda BM, et al. Cell. 2009;137:509–521. [PMC free article] [PubMed]
  • Malone CD, Brennecke J, Dus M, Stark A, McCombie WR, Sachidanandam R, Hannon GJ. Cell. 2009;137:522–535. [PMC free article] [PubMed]
  • Vagin VV, Sigova A, Li C, Seitz H, Gvozdev V, Zamore PD. Science. 2006;313:320–324. [PubMed]