Remember the genomics gold rush? Back about five or six years ago? Sure you do! People were lining up to throw money at companies that could deliver human gene sequences, as part of the never-ending search for new drug targets. (OK, it's not quite never-ending, but for the time horizon we have in the industry, we'll stick with that adjective.) Well, a good part of the reasoning behind all those sequencing deals may have just taken another hit.
Even while the genomic craze was at its peak, there were doubters. Gene sequences should, in principle, read out directly into protein sequences. But there are already some complications, since DNA and RNA sequences can, at various points, be spliced and recombined. We think we know the signs of that happening, but it's still another thing to worry about.
But even if there's no funny business, it's not easy getting useful information from a raw sequence. We still can't predict protein structures de novo, not to the degree that medicinal chemistry needs. You can learn a few things - homology statistics can place your unknown protein into a known family, or (failing that) at least tell you stuff like whether it's likely to be membrane-bound. But knowing that something's, say, a G-protein-coupled receptor sure isn't enough to tell you what it does in vivo, or if it's a valid drug target (and if so, for what disease.)
And there's always been a lot more to the cohort of proteins than just the corresponding genomic sequences. That's where the doubting voices got louder. Proteins get modified in all kinds of ways. They get phosphorylated and glycosylated around their outsides, for example, which can profoundly change their function. And they can get sliced up into smaller proteins, too. Happens all the time - plenty of bioactive proteins are produced from a larger percursor, carved off as needed like sandwich meat at the deli counter. (The enzymes that do the carving can be very good drug targets indeed.)
Enter the latest craziness, from J. C. Yang's lab at the National Cancer Institute. There's an exhilarating (or alarming, depending on your point of view) paper in the ">latest issue of Nature (427, 252), whose authors have seen something that no one had ever seen in higher organisms. They've shown that not only can proteins be chopped up in the cell, but that the various fragments can be spliced back together in new combinations. In their case, they showed that cells could produce a nine-amino-acid peptide from a 49-amino-acid precursor. The middle 40 got snipped out, and the two ends were spliced together to make the nine-mer. You're never going to be able to read off the sequence for that one, now, are you?
This sort of thing goes on all the time in single-celled creatures, and is known all the way up to, oh, bean plants. But it had sure never been seen in mammalian cells. How does this process happen, and how important is it? Who knows! It might turn out to be a rare curiosity, or it might turn out to be something really important that we've completely missed seeing all these years. To be sure, no one's reporting coming across a lot of important proteins whose sequences couldn't be matched in the genome somewhere. But there are an awful lot of proteins whose sequence we don't know, so the upper and lower bounds of this new phenomenon are fuzzy.
This paper, you can bet, has already set off a flurry of research. Perhaps there are some unexplained proteomic problems out there which this will turn out to answer. And here's a prediction: it wouldn't surprise me if protein splicing had already been seen by someone else, who looked at the data, thought about it, and said "Naaaah. That can't happen. Must have messed something up somewhere. . ." If this turns out to be physiologically important, it's Nobel material for sure. Listen closely, and you may hear the sound of someone kicking themselves.