But in mammals, Black says, "the regulation gets very complicated." His lab has been teasing out the precise ways in which specific splicing factors, and the sequences they bind to, orchestrate that regulation. "For most exons that have been analyzed, what we've found is there are multiple regulators," he says. Some of these are so globally expressed that additional regulators must be around simply to counteract them when they're not needed. In some cases, a single molecule can activate splicing in one cell type while repressing it in another. The finely tuned balance of these splicing factors determines how each molecule behaves and how a transcript is ultimately spliced. "There's this combinatorial system of regulation similar to transcriptional regulation," Black says.

The average mammalian gene has eight or nine exons, and with most human genes undergoing some form of alternative splicing, virtually all of these are candidates for elaborate control. Most of the molecules and their interactions remain to be elucidated. In that light, the goal of being able to predict splicing from a gene sequence seems attainable, but that knowledge is still tantalizingly out of reach. "We're not very good at predicting what splicing patterns are going to be, let alone how those splicing patterns will be regulated," Black says. "People are certainly making progress, but we don't know enough about the mechanism."

Cells invest significant resources in carrying out and regulating splicing—both constitutive and alternative. In mammals particularly, the machinery for identifying splice sites, snipping out introns, and reconnecting the severed transcript has long been known to be a large complex of regulatory and catalytic components, both protein and RNA.

Melissa J. Moore, an HHMI investigator at Brandeis University, says that not long ago, she estimated that 50 to 70 proteins participated in the splicing process. So it came as a surprise when, in 2002, her lab and others purified the spliceosome in various stages of the process and found about 100 proteins at each phase. The proteins known to come and go during splicing total about 300.

These studies, Moore says, focused on constitutively spliced exons, which are more efficiently recognized and removed than those that are alternatively spliced; many of the molecules regulating the latter process have probably been left out of recent structural models. "My guess is that the list of proteins affecting splicing is not complete yet. We haven't even begun to scratch the surface of the proteins that affect alternative splicing," she says.

Scientists are still debating how the components of the massive complex arrive at a splice site, but according to HHMI investigator Michael Rosbash, also at Brandeis, "It's rather difficult to imagine how alternative splicing would take place if the spliceosome was preassembled." While some researchers argue for a model in which the spliceosome is at least partially preassembled, recent studies from Rosbash's lab and another have demonstrated that, at least in yeast, loading of the spliceosome occurs in a stepwise fashion.

Rosbash has found that the five small nuclear ribonuclear particles (snRNPs) at the core of the splicing complex load sequentially onto the transcript, presenting an opportunity for regulation. "A lot of alternative splicing regulation could take place at the level of which snRNPs jump on where, and in what order," he notes. "It just gives you many more degrees of freedom." Although there's no evidence yet of this stepwise assembly in other organisms, "the core mechanisms are so similar that it would be shocking if something as fundamental as this were not conserved between humans and yeast," Rosbash says.

	PAGE 6 OF 6	Go Back