In eukaryotic cells, precursors of messenger RNAs (mRNAs) are synthesized and processed in the nucleus. During processing, noncoding introns are spliced out, and a cap and poly-adenosine sequence are added to the beginning and end of the transcript, respectively. The resulting mature mRNA is exported from the nucleus to the cytoplasm by crossing the nuclear pore. Both the introns and the cap help to recruit factors that are necessary for nuclear export of an mRNA. Here we provide evidence for a novel mRNA export pathway that is specific for transcripts coding for secretory proteins. These proteins contain signal sequences that target them for translocation across the endoplasmic reticulum membrane. We made the surprising observation that the signal sequence coding region (SSCR) can serve as a nuclear export signal of an mRNA that lacks an intron or functional cap. Even the export of an intron-containing natural mRNA was enhanced by its SSCR. The SSCR export signal appears to be characterized in vertebrates by a low content of adenines. Our discovery of an SSCR-mediated pathway explains the previously noted amino acid bias in signal sequences, and suggests a link between nuclear export and membrane targeting of mRNAs.
Basically I've discovered a new nuclear mRNA export pathway that operates on transcripts that encode secreted proteins. To speculate a bit (this is a blog after all), my work suggests that whole classes of genes (for example those encoding secreted proteins) maybe regulated as a group at the level of mRNA export from the nucleus to the cytoplasm. If you want to take a look at the paper, click here. If you would like to read a synopsis of the paper by Kira E. O'Day, click here.
Way to self advertize...
There's one part of the paper that seems a little iffy, though, which is little bit on amino-acid bias; the codon-bias is much more convincing, whereas amino-acid bias could very well be much more influenced by the structural recognition of the signal sequence. Isoleucine is much more structurally stiff than leucine, for example, even though they seem to be quite similar, and arginine is obviously chemically different from lysine, though they're both positively charged.
Congratulations on the paper, Alex! I have a background in secretory biology myself (and, curiously, find myself meandering back to it in my current work), so ever since you described the findings I've been eager to read about it in detail.
It could be that the signal recognition particle (SRP) has some altered specificity for leu vs ile, but it would seem that the biggest determinant for SS specificity is hydrophobicity. In this respect, ile and leu have very similar hydrophobicities. We also saw other disparities in a.a. usage between a.a. with similar hydrophobicities as described by Hessa et. al., Nature (05) 433: 377381. For example cys, which is encoded by UGU and UGC, is enriched over met (AUG) in a pattern very similar to the leu vs ile and arg vs lys biases. It could be that the SRP co-evolved with the SSCR system, to better recognize these subtle changes imposed by SSCR binding proteins, but SRP also recognizes transmembrane domains which don't have these biases.
Also, if the second author is the Mike Springer I know from UCSF, say hi from me and ask him to tell you about his hot dog eating skills. Or better yet, to demonstrate them.
Just read the paper -- loved it. Very clear and logical.
Have you looked the export of a non-signal-sequence-encoding mRNA with an adenine-deficient leader sequence (the 69 ntd after the initiator codon)? The results with fs-ftz-delta-i and UUG-ftz-delta-i (which technically don't encode a signal sequence either) suggest that such a mRNA would also be exported via the same splicing-independent pathway as t-ftz-delta-i, but then translated by cytosolic ribosomes instead of targeted to the ER.
Following up on that idea, are there non-secretory proteins encoded by genes with adenine-deficient leader sequences; do they get exported by this novel pathway; and do their functions fall into any immediately recognizable category? I'm sure this is just one of many things you've already thought of; you guys have your choice of future directions.
In any case -- a great paper. Congratulations again.
Funny how interconnected we all are, I'll have to ask Mike about his dog! Also I looked you up - it's funny that you were in Peter Walter's lab - the world really is small! (BTW Peter was the scientific monitoring editor for the paper).
Incredibly, I haven't looked to see if other transcripts with long "no-A-tracts" are exported - we think that there must be something more then the lack of adenines ... I'm working on it. As for other transcripts that may use this pathway, I'm guessing that there must be, but it is unclear what they might be. What seems strange is that transcripts that encode membrane proteins but have no SSCR don't seem to have anything that looks like an SSCR. I've transfected various genes encoding Sec61 components (all membrane proteins with no SS) and the mRNAs are all exported poorly. I don't know what to make of that.
Cool mechanism! Congrats again!
-- Just a thought...
Do you think that it is important to find out whether different SSCR sequences promote mRNA export through a common machinery, or different ones? If you do, have you tried injecting cells with a large amount of the SSCR RNA oligo to "hijack" the machinery responsible for the SSCR-dependent export pathway, and then see if the endogenous SSCR-containing messages get exported less efficiently. In case this strategy works, then by changing the SSCR oligo sequences you inject into cells, you might be able to test the above possibilities.
Doesn't the first TMD act like a signal sequence in proteins that don't have a hydrophobic leader? Not in the sense of being cleaved, but in the sense of being recognized by SRP. I think that's how it works in yeast, but I can't remember whether I ever even knew how the mammalian system works.
Regardless -- that's interesting about Sec61 et al., and makes me wonder whether there's some stress condition under which those messages might be efficiently exported. It sounds from the paper like you're already exploring UPR/other ER stress pathways in this context -- we know in yeast that the UPR transcriptionally induces genes encoding translocon components; wouldn't it be cool if the UPR induces mRNA export of the same genes in mammals?
Best of luck with the follow-up experiments. Can't wait for chapter 2. Keep us posted.
Yeah that's a great experiment. I haven't tried, but maybe I should ... it would be a good way to probe the ability of SSCRs to promote export.
"Doesn't the first TMD act like a signal sequence in proteins that don't have a hydrophobic leader? "
Yes exactly - and it would be reasonable to think that the first transmembrane domains (at the nucleotide level) would promote nuclear export. But as far as we can tell, it doesn't seem like they have this activity. Moreover they do no have the anti-adenine bias that characterizes SSCRs (at the a.a. level or at the codon level). I'm puzzled as to why this would be. Why does the cell have a distinct nuclear export machinery for transcripts that are targeted to the ER by SSs but not by their first TMD? We looked hard but couldn't find anything. We also analyzed the A-content of 5'UTRs, but we couldn't find any differences between different classes of transcripts. So this issue remains unclear.
In comparing the similarities and differences between SSCRs and the nucleotide sequences that encodes TMDs demonstrates that the leu vs ile bias in SSs is not a prerequisite for SRP recognition. The leu vs ile bias correlates with nuclear export and not SRP recognition or translocon insertion.
Re UPR. Yeah I haven't totally given up on that angle. We were encouraged by the fact that prion protein, which is very prone to being misfolded and is not efficiently inserted into the ER under stress, has lots of adenines in its SSCR. We then analyzed the A-content of transcripts that are upregulated and downregulated by UPR. Mike has the data and tells me that there isn't much there. I need to walk across the quad and look at the data with him.
I personally find it shocking that there's anything the UPR can't do.
Congratulations! That is a great accomplishment.
Hi Alex! Congratulations on your paper. Only now found the time to read it. Cool stuff.
Congratulations. Great paper. I read it with great interest before seeing your blog. Are the mRNAs for this category of proteins also localized at the membranes?
In COS-7 cells, most of the mRNA that contain SSCRs localize to the membrane of the endoplasmic reticulum (see figure 3A). This localization, as far as we know, seems to depend on active translation. A good example is shown in figure 5C where t-ftz-delta-i fails to target to the ER when cells are treated with a translation inhibitor.