Thursday, January 25, 2007

Tag! You're It!

The protein kinases are a sizable chunk of the human proteome (>500 members) and the subject of intense research. Some estimates peg about a quarter of all current small molecule drug discovery efforts as targeting kinases, and a lot of basic research remains to be done on the class.

A key question for a given kinase is what are the proteins it phosphorylates, its substrates. It turns out that this is a decidedly difficult problem. With mass spectrometry one can now read out lots of phosphorylation sites on proteins, but figuring out which kinase is phosphorylating which site remains difficult. A lot of methods have been proposed and successfully used, but while the mass spec folks are getting adept at churning out thousands of phosphorylation sites (here is another new one from PNAS), papers linking kinases to substrates typically have very small numbers of substrates (such as uno!) in them -- though there are some notable exceptions [not claiming this is an exhaustive list]. Which is too bad, as quite a few kinases have no known substrates (or none besides themselves; most kinases will trans-phosphorylate amongst themselves). Indeed, one hint that a breakthrough is needed is the fact that the number of methods keep proliferating; there's no one (or a few) good method to settle on yet.

A new report in Journal of the American Chemical Society is pretty intriguing in this light. I'll confess I haven't gotten past the abstract because my overdue ACS renewal is still riding in my backpack, but the abstract is tantalizing enough. The claim is that a biotin-ATP conglomerate will be accepted by a kinase and result in the kinase biotinylating the substrate. Since pulling down biotin-tagged proteins or peptides is truly old hat (heck, I did it as an undergraduate in the 80's -- and it was my all thumbs experience there that helped make me a computational biologist!), this is pretty exciting. If it proves to work with many (can we dare hope for most?) kinases, this could really revolutionize things.

Now, some previous tagging schemes had been developed, but they all required additional steps and perhaps even genetically modifying the kinase first. None of these methods has seemed to have caught on, at least from the standpoint of a growing pile of papers.

Clearly, the proof will be in the replication. If the next couple of years sees a flurry of papers reading out kinase substrates by biotinylation, that will be validation. If not, then let the next round of scheming begin!

No comments: