miRNA and RNA interference is so hot right now ...
I'm not in the field, but I do keep an eye on it. Right now the there is quite a few papers on how miRNAs regulate translation. There is some data that indicate that miRNA and the associated RISC complex (RNA Interference Silencing Complex) inhibits the ability of ribosomes from engaging targeted mRNAs. The initiation of protein synthesis would thus be blocked. Other data indicate that miRNAs inhibit the activity of ribosomes that are bound onto transcripts. In this second model miRNAs prevent the elongation step of protein synthesis. These two models may not be mutually exclusive as miRNAs may inactivate translation in multiple ways and that miRNAs may have various effects depending on the target transcript. Don't forget that if the miRNA forms a perfect base pair with the target transcript, the mRNA will be cleaved by RISC. To make things more complicated, a recent paper from the Steitz lab provides some evidence that a miRNA associated proteins can even activate the translation of certain mRNA (see below). So it seems like anything is possible.
So how are miRNAs and the RISC complex turning off the translation? It is clear that miRNAs promote the segregation of target transcripts to P-bodies where the mRNAs are either stored or degraded. Another important clue came out a couple of weeks ago, and it has to do with the cap.
Some background: all mRNAs have a cap structure. This funny structure is basically an inverted guanosine ribose that is 5' link to a triphosphate group that is inturn linked to the very first nucleotide of the mRNA transcript.
In the nucleus the cap recruits the cap binding complex (CBC) which potentiates splicing and export (see this post). Once the mRNA is exported it is thought that CBC initiates the pioneer round of translation, this is where the ribosome translates the mRNA into protein. If something is wrong with the mRNA then the transcript is shuttled to the mRNA degradation machinery, if the mRNA is fine the CBC is replaced by the cytosolic cap binding protein eIF4E.
eIF4E is a subunit of the eIF4F translation initiation complex. I've blogged and complained about eIF4E before. This protein helps recruit the ribosome to the mRNA (although apparently you don't really need eIF4E to do this ... more on this some other time). In addition the eIF4F complex contains RNA helicases that can unwind RNA structures found at the start of many mRNAs which may inhibit the ability of the ribosome to scan the beginning of an mRNA as it attempts to locate the start of the open reading frame (i.e. the portion of the mRNA that actually codes for protein). eIF4E is also found in all these funny mRNA aggregates such as stress-granules and P-bodies. On top of all this, eIF4E is thought to be limiting in most cells ... in other words with all the ribosomes and all the translation ina typical cell, there just isn't enough eIF4E to go around. Now it turns out that argonaute proteins, which are one of the key components of the RISC complex, have a domain (the "MC" domain) that shares some similarity with the cap binding domain of eIF4E.
The authors also demonstrate that argonaute proteins bind to cap structures in vitro and that mutant forms of the argonaute protein Ago2 that can't bind to the cap also fail to inhibit translation of target transcripts. Interestingly Ago2 that is artificially tethered to an mRNA inhibits that mRNA's cap depedent, but not cap independent, translation. Now my only complaint is that the authors should have shown that argonautes and eIF4E compete for cap binding, but this is a minor point.
Interestingly, in certain circumstance argonaute proteins may potentiate translation (perhaps by binding to the cap?) Remember the paper that I mentioned above? It turns out that in some cases Ago2 can associate with an RNA element (an AU-rich element or ARE, for you RNA aficionados) in the TNF-alpha transcript. It's not clear whether Ago2 is loaded onto this transcript by a miRNA, but it does associate with fragile-X-mental-retardation-related protein 1 (FXR1), another one of these P-body associated proteins. When either Ago2 or FXR1 levels were lowered by siRNA, translation of the TNF-alpha gene was inhibited.
Now it is unclear how Ago2 and FXR1 are binding to the ARE of this transcript. AREs also attract other strange mRNA binding proteins such as HuR and TTP (both of which are also found in various RNA aggregates) and so there may be other interactions between all these funny RNA binding proteins. Due to their strange nucleotide composition, AREs may also attract some small non-coding RNA that is associated with Ago2.
Lots of questions ... it's clear we have a lot to learn.
Kiriakidou M, Tan GS, Lamprinaki S, De Planell-Saguer M, Nelson PT, Mourelatos Z.
An mRNA m(7)G Cap Binding-like Motif within Human Ago2 Represses Translation.
Cell (2007) 129:1141-51 doi:10.1016/j.cell.2007.05.016
So what about Ago1? I got the impression from the Kiriakidou et al paper that they thought Ago1 should also interact with the cap and repress translation (based on homology) but they didn't test it (yet?). Ago1, of course, interacts with fragile x mental retardation protein (FMRP) based on data from the Warren lab (Nat Neuroscience, 2004). Then again, maybe FMRP also interacts with Ago2, or all of these things will turn out to be true. The question, at least for me, is, in neurons, what is the switch to go from repression to active translation? Fascinating stuff happening all around!
It seems like they could have tested quite a few things including
- other Agos
- competition between Agos and eIF4E
- ensuring that eIF4E and Agos don't bind to the same transcripts in vivo or show that when Ago climbs on to a transcript in vivo due to miRNA targeting, eIF4E falls off
- show that Ago inhibits cap but not not IRES dependent translation of an endogenous transcript targeted by a miRNA
"show that Ago does not inhibit IRES dependent translation"
Although Ago may bind to the cap and thus prevent eIF4E dependent translation it may inhibit IRES dependent translation by segregating the mRNA to processing-bodies.
funny that you mention the eIF4E paper. It contains a very blatant case of junk bioinformatics. My blog will get to that once I am done with GO :-)