You'd Prefer An Argonaute

Many X-linked microRNAs escape meiotic sex chromosome inactivation

Rui Song, Seungil Ro, Jason D Michaels, Chanjae Park, John R McCarrey and Wei Yan

Nature Genetics 41 (4): 488-93, April 2009.
doi:10.1038/ng.338

This week’s succinct summary and analysis provided by Anonymous:

Meiotic sex chromosome inactivation (MSCI) is a process that silences unsynapsed chromosomes during meiosis, specifically the X and Y chromosomes in males.  While the reasons for this silencing are not understood, it has been postulated that MSCI prevents genomic instability.  Following up on previously published observations, Song et al. investigated whether X-linked miRNA genes escape silencing.   Carrying out a real-time PCR-based survey of X-linked miRNAs, the authors found that numerous miRNAs were upregulated transcriptionally during stages of spermatogenesis in which MSCI is expected to function, suggesting a major distinction between transcriptional regulation of miRNAs and mRNAs.

This report systematically investigates all X-linked miRNAs in various stages of spermatogenesis, checks their transcriptional regulation to verify that increased levels are not due simply to increased stability of miRNAs transcribed prior to MSCI, and follows up on a few miRNAs in more detail using in situ hybridization.  One weakness of the study, however, is that absolute miRNA expression levels were not determined. Therefore, although a miRNA may be upregulated, it is unclear whether this upregulation results in  (1) full or only a fractional escape from MSCI, and (2) a biologically active concentration of miRNA.   Nevertheless, this study provides an interesting starting point to investigate differential regulation during MSCI.

Splicing Factors Facilitate RNAi-Directed Silencing in Fission Yeast

Elizabeth H. Bayne, Manuela Portoso, Alexander Kagansky, Isabelle C. Kos-Braun, Takeshi Urano, Karl Ekwall, Flavia Alves, Juri Rappsilber, Robin C. Allshire

Science 322 (5901): 602-606, October 2008.
doi: 10.1126/science.1164029

This week’s paper dissection by David Weinberg:

Centromeres in fission yeast (Schizosaccharomyces pombe) consist of a central kinetochore domain flanked by heterochromatic outer repeats. These outer repeats are transcribed by RNA polymerase II and maintained in a heterochromatic state by the RNAi pathway. According to the current model of S. pombe RNAi, the RDRC complex (Rdp1, Hrr1, and Cid12) converts nascent outer repeat transcripts into dsRNA, which serves as a substrate for Dcr1. siRNA-loaded Ago1, in complex with Tas3 and Chp1, is directed to complementary nascent transcripts and stimulates H3K9-dimethylation of histones by the Clr4 histone methyltranferase.

A previous forward genetic screen for ts lethal defects in centromere silencing had identified two components of the spliceosome, Cwf10 and Prp39. Based on this evidence, the authors hypothesized that there may be a link between splicing and centromere silencing in S. pombe. In this paper, they take a reverse genetics approach to further investigate this potential connection.

The paper begins by characterizing the extent of centromere silencing in a collection of ts lethal splicing mutants. They find that only specific splicing mutants show a defect in centromere silencing, i.e. there are many splicing mutants that show no effect. Among their mutant collection, they find a strong correlation between silencing of a centromere-embedded reporter gene, transcript levels of endogenous outer repeats, and quantities of outer repeat-derived siRNAs.

At this point, the authors point out that there are many possible “mundane” explanations for the effects they see, namely that splicing itself may affect mRNAs encoding proteins that are directly/indirectly involved in RNAi (including potentially Ago1 itself). Their numerous attempts to disprove this potential artifact are commendable – thorough, clever, cautious. While their data alone cannot entirely rule out the mundane, they did as much as could be done to convince the reader (and reviewers, presumably) that there must be interesting (i.e. splicing-independent) science at work here.

Figure 3 contains, by far, the least novel/interesting experiments of the paper. Basically the authors demonstrate that the centromere silencing defect in splicing mutants is due to a minor disruption in the maintenance of RNAi-dependent heterochromatin. The effect they see is surprisingly weak in comparison to a dcr1-null strain, suggesting that the role of the spliceosome is not essential for RNAi-directed heterochromatin.

Luckily, the paper ends on a high note with Figure 4. They use mass spec to show that Cid12 interacts with the RDRC complex, and chromatin immunoprecipitation to demonstrate an association between the spliceosome and centromere repeat DNA. In this way, the authors convert a nebulous genetic interaction between splicing and RNAi into a physical interaction between the spliceosome and the RNAi machinery. This interaction is also consistent with Figure 4B, which suggest that splicing factors act downstream of RITS recruitment (e.g. at the level of RDRC-dependent siRNA amplification).

This final set of experiments hint at mechanism and lead the authors to speculate that the spliceosome may provide a platform that facilitates RDRC recruitment/action. While this model is completely consistent with the data, there is little direct support for this model over a variety of other consistent models (which they entirely ignore). Along these lines, I was most intrigued by their initial observation that only specific splicing mutants showed the centromere silencing defect…but there was not even a mention of how this relates to their model. Still, the novel aspects of this paper – namely, the link between the spliceosome, but not splicing, and RNAi in the form of a physical interaction between the spliceosome and RDRC – make this an important paper in the field. Given the absence of RdRP machinery in metazoans, it will be interesting to see if a similar splicing-RNAi interaction is at work in higher eukaryotes.

nhl-2 Modulates MicroRNA Activity in Caenorhabditis elegans

Christopher M. Hammell, Isabella Lubin, Peter R. Boag, T. Keith Blackwell, and Victor Ambros

Cell 136 (5): 926–938, March 2009.
doi:10.1016/j.cell.2009.01.053

This week’s careful summary and analysis by Joel Neilson:

Lin-41 is a founding member of the heterochronic pathway and a member of the TRIM-NHL family of proteins.  The phenotype of the lin-41 loss- of-function is precocious development, but is not fully penetrant.  The authors were examining whether this phenotype could be accentuated by crossing in additional mutant alleles for other TRIM-NHL family  members.  In contrast to the accentuated phenotype they were expecting, they found that one of these loss-of-function mutants,  nhl-2, resulted in a mildly retarded heterochronic phenotype and rescued the defects in the lin-41 mutant.

To briefly summarize this study in the wrong order and a completely oversimplified manner, they then demonstrate that:

(1)  loss of nhl-2 gene function enhances the phenotype of individually non-penetrant LOF alleles for miRNAs in the let-7 family

(2)  loss of nhl-2 gene function enhances the phenotype of a weak LOF allele of a miRNA in a second family (lsy-6)

(3)  loss of nhl-2 gene function offsets phenotypes observed in animals ectopically expressing a let-7 family member

(4)  loss of nhl-2 gene function accentuates heterochronic defects in worms with mutations in core miRNA machinery components

(5)  all of this happens without modulation of the levels of mature miRNAs and is through previously characterized miRNA targets

(6)  NHL2 exhibits broad temporal-spatial expression

(7)  NHL2 co-localizes with and in fact touches CGH1.  They also genetically interact.

(8)  NHL-2 and CGH-1 physically interact with the core miRNA machinery in an
RNA-dependent fashion

(9)  CGH-1 still interacts with the core miRNA machinery in nhl-2 mutants.

This is a one of the best papers I have chosen for this forum and I got particularly excited upon reading the following in the introduction: “Current models do not adequately account for the facts that some miRNA targets appear to be regulated primarily at a translational level while others are regulated by mRNA turnover, or that a particular miRNA can have dramatically different potencies on distinct miRNA target reporters (Eulalio et al., 2007) therefore, it is likely that additional proteins can interact with miRISC to modulate the nature and efficacy of miRISC activity.”  Looking at the last clause of that sentence, they did in fact demonstrate that additional factors can modulate miRISC activity.  But that’s not what I got excited about in reading the introduction. To really nail down  the parts that current models do not adequately account for, someone really does need to show that a defined target, which sometimes (in a spatial or temporal manner) is affected one way by miRNA/RISC recognition. . .say, translational repression. . . and sometimes is  affected another way. . .say, deadenylation. . .by the same miRNA/RISC, and show what dictates this specificity.  This study did not directly address this issue but is definitely moving us in the right direction.

A distinct class of small RNAs arises from pre-miRNA–proximal regions in a simple chordate

Weiyang Shi, David Hendrix, Mike Levine & Benjamin Haley

Nature Structural & Molecular Biology 16 (2): 183-189, February 2009.
doi:10.1038/nsmb.1536

This week’s forthright summary and analysis comes from Anna Drinnenberg:

Using high-through-put sequencing the authors of this paper describe the identification of a new class of small RNAs encoded in the Ciona intestinalis genome that arise from positions adjacent to pre-miRNAs in predicted hairpins. In total, these small RNAs are in very low abundance in the sequencing libraries (~1500 reads compared to more than 1 million reads for mature miRNAs) and have about half the number of reads compared to miRNA* sequences. They state that half of the miRNA genes in Ciona give rise to such miRNA-offset RNAs. In three examples shown, the total number of reads mapping to these offset RNAs add up to about the total number of reads of miRNA-offset RNAs in all libraries together which raises the question whether a significant amount of miRNA-offset RNA is generated from the other miRNA genes that were not shown. They suggest a model in which a sequential or parallel processing step by Drosha could explain the biogenesis of these RNAs. Based on their data in which complementary sequences to miRNA-offset RNAs are detectable in only a few cases at low abundance, and considering the highly unpaired structure of shown miRNA-offset RNA duplexes, it seems unlikely that an RNaseIII enzyme is involved in processing. The experimental procedure that included expression of a Drosophila miRNA hairpin to induce miRNA-offset RNA production in the Ciona tadpole is a very interesting and creative follow-up to their sequencing results, however, the assumption that these RNAs constitute a new class of small RNA and aren’t just by-products of (exonucleolytic degradation) miRNA processing is not convincing.