Thursday, November 3, 2011

CTCF, new role in mediating alternative splicing

CTCF-promoted RNA polymerase II pausing links DNA methylation to splicing

What the authors did is to compare transcriptome of wildtype with that of CTCF-depletion one in two cell lines; they found an alternative splicing event on exon 5 of gene CD45, which is included (or more likely to be included) in the wildtype comparing to the depletion. And the reduced inclusion is likely correlated with the binding of CTCF downstream of the exon (instead of upstream). They argued that the CTCF binding can promote the RNA Pol II pausing at the exon when it's not methylated. When it's methylated, the CTCF can not bind and therefore cannot pause the pol II at the exons, which result in exclusion of the exon.  That's the story. 

Several questions:
  1. I know the anti-correlation of DNA methylation and CTCF binding, which has been shown by other papers. But how the methylation in the EXON can prohibit the CTCF binding DOWNSTREAM? I meant when they are not that close.
  2. Even though they observed reduced inclusion in the CTCF-depletion transcriptome, how do they know it's caused by the DOWNSTREAM binding, not binding elsewhere? Can they mutate the binding site of CTCF in the wildtype and observe the same exon exclusion?
If it's true that CTCF is (and there will be a long list of) location-specific DNA-binding ‘splicing factors’, I would wonder how this is linked with the ChIA-PET data, which can track the long-range interaction points (physically) in the chromosome, like the DNA looping structure. A very straightforward question is: whether the alternative splicing events is also mediated by the looping event? This might be other side of the picture, where CTCF binding can form a loop which can 'loop' out exons.

Tech Note:
The mixture of isoforms (MISO) model was applied to RNA-seq data (Supplementary Table 3) to identify exons with a high probability of differential expression in response to CTCF depletion, as assessed by the Bayes factor confidence index39.

No comments: