PerspectiveMolecular Biology

A New RNA Dimension to Genome Control

See allHide authors and affiliations

Science  21 Jul 2006:
Vol. 313, Issue 5785, pp. 305-306
DOI: 10.1126/science.1131186

Micro RNAs (miRNAs) and small interfering RNAs (siRNAs) are 21- to 25-nucleotide RNA molecules that influence their much bigger relatives, the messenger RNAs (mRNAs). Over the past few years, these small RNA species have captivated the study of gene regulation and modified our notions about how gene expression is controlled. A recent clutch of papers describe for the first time a class of small RNA cousins that are distinct from miRNAs and siRNAs (16). They promise to yield fascinating new insights into genome control.

The genesis of the discovery of a third type of small RNAs is linked to the Argonaute family of proteins. Certain Argonaute proteins such as Ago1 and Ago2 associate with miRNAs and siRNAs to form ribonucleoprotein complexes that associate and repress the expression of target mRNAs. Sometimes, mRNA targets are cleaved by a mechanism that is catalyzed by the particular associated Ago protein. However, a subclade of Argonaute proteins are phylogenetically distinct from the Ago1/Ago2 subclade and do not appear to associate with siRNAs or miRNAs (7). Led by its founding member, the Piwi protein of Drosophila melanogaster, this subfamily appears to play an important role in germline development. For example, genetic analysis of Piwi and its mouse orthologs (Miwi, Mili, Miwi2) indicates that they are essential for spermatogenesis (79). How the Piwi subfamily regulates the germline has remained for the most part elusive.

A recent breakthrough regarding this question has emerged from the identification of the small RNA partners that associate with Piwi proteins. In research reported on page 363 of this issue, Lau et al. (1) partially purified a ribonucleoprotein complex from extracts of rat testis and found testis-specific RNAs of 25 to 31 nucleotides, with a dominant subpopulation of 29- to 30-nucleotide RNAs. These RNAs are distinct in size from miRNAs and are associated with distinct protein complexes. Lau et al. purified the RNA-protein complex by conventional chromatography and identified the rat homologs to Piwi (Riwi) and the human RecQ1 protein as subunits of the complex. On this basis, the RNAs have been named Piwi-interacting RNAs (piRNAs) and the complex is called the Piwi-interacting RNA complex (piRC).

piRC exhibits adenosine triphosphate-dependent DNA helicase activity, which is likely attributable to RecQ1. Interestingly, the RecQ1 homolog in Neurospora crassa has been implicated in gene silencing (10). Lau et al. (1) also found that piRC will cleave RNA targets in a manner dependent on piRNA complementarity, much like Ago2 cleavage of siRNA targets. This might imply that piRC has some posttranscriptional role in gene silencing. Indeed, piRNAs are associated with polysomes fractionated from mouse testis extract (4). However, genetic studies have implicated Piwi proteins in transcriptional gene silencing by altering chromatin conformation (7). Consistent with the idea that piRC plays a role in transcriptional silencing, RNAs that are longer than 24 nucleotides (such as piRNAs) have been associated with this mode of gene silencing in a wide variety of species (11).

One of the real surprises has been the nature of piRNAs themselves. Deep sequencing of complementary DNAs derived from piRNAs revealed that they correspond to regions of the genome previously thought not to be transcribed (13). These regions are limited in number to about 100 clusters ranging in size from 1 to 100 kilobases and are distributed across the genome. Very few clusters contain repetitive DNA; in fact, repetitive DNA is underrepresented. A greater surprise is that piRNAs in a typical cluster exclusively map to either one or the other strand of genomic DNA (15). A minority of clusters generates piRNAs from both strands, but plus-strand piRNAs are segregated from minus-strand piRNAs into distinct regions that are separated by a gap of a few hundred base pairs (see the figure). The paucity of evidence for overlapping complementary RNAs or potential foldback RNA precursors suggests that piRNAs are not derived from double-stranded RNA precursors. This would suggest a biogenesis mechanism distinct from that of siRNA and miRNA, both of which are derived from dsRNA through enzymatic cleavage by Dicer.

The known and unknown features of piRNAs.

Shown is a genomic region that generates a cluster of piRNAs. The left side of the region generates antisense RNA transcripts (blue) and the right side generates sense transcripts (green); a short gap in between likely acts as the promoter for transcription of both sides (divergent red arrows). An RNA polymerase of unknown identity is shown in active transcription. These transcripts are processed into 25- to 31-nucleotide piRNAs by an unknown mechanism. piRNAs then associate with Piwi and RecQ1 homologs to form piRC complexes. These complexes might regulate the genome at the level of DNA or histones, or at a posttranscriptional level. The events that are under direct control of the piRC mechanism within developing sperm cells are currently unknown.

piRNAs and piRC complexes are not restricted to rats; they have been detected in testes of other mammals, including mouse and human (15). The organization of piRNA genomic clusters is conserved in other mammalian species as well. Most clusters in the rat, mouse, and human are homologous or syntenic, even extending to strand specificity. Nonetheless, piRNA sequences are not conserved between species. Sequence heterogeneity is consistent with the idea that the genomic clusters are subject to neutral selection pressure.

What testis cells express piRNAs? In the mouse, Mili is expressed in male germ cells from primordial to pachytene stages, whereas Miwi is expressed from pachytene to spermatid stages (8, 9). Both Mili and Miwi associate with piRNAs, which suggests that piRNAs are produced within the developing male germ cells. Consistent with a germline-specific expression pattern, piRNAs are not detectable in WV mutant mice, which are missing differentiating germ cells (2, 5), and piRNAs are reduced in Miwi mutants (4). piRNAs are detected throughout sperm development but appear to peak in abundance at the round-spermatid stage (15). The abundance of piRNAs in spermatids is staggering; about 1 million molecules are estimated per round-spermatid cell (3).

Although mammalian piRNAs are not associated with repetitive DNA, the situation might be different in Drosophila. On page 320 of this issue, Vagin et al. (6) describe repeat-associated siRNAs (rasiRNAs) in the fly germline as 24- to 29-nucleotide species that arise primarily from the antisense strand of repetitive sequences such as retrotransposons. These RNAs are associated with Piwi and another member of the Piwi subclade, and mutations in the Piwi class of genes cause derepressed retrotransposon silencing coupled with altered levels of rasiRNA abundance. Interestingly, these effects are not restricted to the male germline but also apply to the female germline. Perhaps rasiRNAs in Drosophila use a molecular mechanism similar to that of mammalian piRNAs to silence portions of the genome.

Many questions ensue from these studies. Are testis-specific piRNAs found in species other than mammals? Does piRC regulate male meiosis by regulating genome organization, or is it a surveillance mechanism to ensure genome integrity during germ cell maturation, including suppression of selfish elements? Is piRC male-specific, or are other classes of RNAs associated with Piwi in the female germline? How are piRNAs produced? Their structures might suggest a ribonuclease III-independent origin. Indeed, neither of the two Dicers from Drosophila is essential for rasiRNA biogenesis and repeat DNA silencing (6), although it is possible that each is redundant for the other or that a third enzyme, Drosha, carries out processing. Further investigation should reveal how piRCs regulate the genome.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
View Abstract

Stay Connected to Science

Navigate This Article