RNA Journal Club 8/4/11
Svetlana Lebedeva, Marvin Jens, Kathrin Theil, Björn Schwanhäusser, Matthias Selbach, Markus Landthaler, and Nikolaus Rajewsky
Molecular Cell 43, 1–13, 5 August 2011.
This week’s crystal clear summary and analysis by David Koppstein:
Two papers were recently and concomitantly published in Molecular Cell (Lebedeva et al, 2011; Mukherjee et al, 2011) describing the transcriptome-wide binding sites of the RBP HuR by employing the second-generation high-throughput technique PAR-CLIP (photoactivable ribonucleoside crosslinking and immunoprecipitation). HuR is a member of the Hu/ELAV family of proteins, which each have three RRM domains and whose knockouts are embryonic lethal. Unlike the rest of its family members, which are restricted to the nervous system and have been reported to affect alternative splicing, HuR is ubiquitously expressed and has been previously implicated in stabilizing messages with AU-Rich Elements (AREs) in their 3’-UTRs (Fan and Steitz, 1998). The mechanism of stabilization is unclear, although it is thought to sterically interfere with binding of other ARE destabilization factors such as hnRNP D and TTP. There have also been conflicting reports about crosstalk between HuR and miRISC. One group found that binding of HuR by CAT-1 relieves miRNA repression (Bhattacharya et al, 2006), while another found that HuR recruits let-7 to repress c-myc (Kim et al, 2009).
Consistent with previous results, the majority of HuR binding sites were in the 3’-UTR. Unexpectedly, both groups found a large fraction of reads mapping to intronic regions. Upon closer inspection, it was found that there is a distinct peak of HuR clusters ~20bp upstream of the 3’ splice site of introns. This is probably not an experimental artifact of the 4SU that was used, since the Rajewsky group also used 6SG to crosslink and recovered many of the same intronic polypyrimidine motifs. The role of binding to the 3’ splice site has yet to be elucidated, however, since both groups were unable to find significant correlations with exon inclusion or exclusion upon HuR knockdown.
Both groups showed that, upon knocking down HuR with siRNAs, messages with HuR clusters mapping to them are destabilized, confirming the role of HuR in stabilizing messages. Interestingly, the Keene group provided evidence that more HuR binding sites correlated with more destabilization upon HuR knockdown. Furthermore, messages with purely intronic binding sites were just as destabilized, if not more so, than messages with purely 3’-UTR sites. The Rajewsky group also performed pulsed SILAC, and found that protein synthesis levels essentially mirrored mRNA levels, in contrast to previous suggestions that HuR modulates translation (Mazan-Mamczarz et al, 2003).
The Rajewsky group also compared small RNA profiles of mock- and anti-HuR siRNA-transfected cells, perhaps looking for HuR regulation of miRNA levels. Although most miRNAs remained unperturbed, the authors found that miR-7 was strongly upregulated upon HuR knockdown. Although the mechanism of this regulation is unclear, the presence of clusters adjacent to the mirtron precursor of miR-7 in hnRNP K is suggestive of direct regulation, as opposed to some secondary effect.
In contrast, the Keene group utilized Tuschl’s cocktail of anti-miR 2’-OMe oligonucleotides to block the efficacy of several microRNAs, and then made mRNA-seq libraries. They found that messages with miRNA sites overlapping the HuR binding site were less affected by blocking the miRNA than those without, consistent with the idea that HuR can sterically inhibit binding of miRISC. The Rajewsky group also observed that miRNA seed sites tend not to overlap with HuR binding sites.
These two papers made several important and novel findings, including prevalent intronic binding of HuR, the increased efficacy of multiple binding sites, and the regulation of miR-7. Further studies will be needed to elucidate the role of binding upstream of the 3’ splice site, and the mechanism and biological relevance of miR-7 regulation. Neither paper examined the role of HuR in stress conditions, which has been established as a role of HuR previously (Meisner et al, 2010). Altogether, these papers are an interesting read about an important RBP and an exercise in interpreting high-throughput experimental data.