Burge has focused his analyses on those alternative splicing events that are conserved between species, because he believes they are most likely to be functionally important. "It's a diverse collection of genes," he says, "but there are significant biases in what kinds of genes undergo conserved alternative splicing." These exons are more likely to be in genes that are involved in development, he says. "They're more likely to be transcription factors. They're more likely to be expressed in different brain regions."

Burge's group has also observed that alternatively spliced exons tend to fall between the segments of a gene that encode the functional units, or domains, of a protein. For most alternatively spliced transcripts, he says, "presumably both forms would produce stable, folded proteins that would have the same enzymatic activity. They would just differ in these regions that might affect other properties of the protein—maybe its localization or its regulation." By altering proteins' locations inside a cell or their sensitivity to activators or inhibitors, most conserved alternative splicing events, Burge expects, produce "subtle, but perhaps very important differences."

The goal of being able to predict splicing from a gene sequence is still tantalizingly out of reach.

While Burge explores the sequence clues to how transcripts are spliced, others are trying to unravel the convoluted system of regulatory proteins that contribute to splicing decisions. It's an extensive network of positive and negative regulators, some ubiquitous and some expressed only in specific cells, and Black thinks that easily hundreds of proteins could be involved.

Connecticut's Graveley, for example, focuses on the fruit fly Dscam gene, which he considers the perfect model for studying the intricate systems that regulate alternative splicing. Graveley saw Zipursky's first paper about Dscam the day it was published in the journal Cell (June 9, 2000); thinking "this is too good to be true," he started designing experiments that very day.

To begin the conversion of genetic information into proteins, cells produce a precise RNA copy of a gene using the DNA sequence as a template. In all but the simplest organisms, these transcripts are not functional mRNAs until they undergo a molecular editing process, when the cell snips out unintelligible or unwanted bits of genetic sequence. This splicing process often occurs while an RNA molecule is still being assembled, and for alternatively spliced transcripts, how quickly the cell's transcription machinery pieces together the RNA can influence which bits are discarded.

The spliceosome (the mass of proteins and RNA that controls genetic splicing) recognizes some splice sites more readily than others, ordinarily ignoring weaker sites in favor of stronger ones located nearby. But according to work by Alberto R. Kornblihtt, an HHMI international research scholar at the University of Buenos Aires, slowing down the transcription of certain genes can alter this selection process.

The effect is seen in genes where a weak splice site is transcribed before a stronger one. Ordinarily, the spliceosome would opt for the stronger site, but when transcription slows, there is a delay during which the weak site has been transcribed, but the stronger one does not yet exist. A spliceosome working on an incomplete transcript has no choice but to use the weaker splice site and consequently produces a different protein than is generated when transcription proceeds more rapidly.

Cells alter transcription rates to modulate the amount of proteins they produce. For genes whose splicing is coupled with transcription, this strategy may alter not just the quantity of those proteins but also some important aspect of their structure, Kornblihtt says. He also thinks this effect can explain why splicing is often coordinated between distant regions of the same gene. It's common for the inclusion of one exon to relate directly to the inclusion of another—which may mean that the splicing of each region depends on how quickly transcription proceeds. One transcription rate may mean that both exons are included, for example, while at a faster rate, both will be excluded.

	PAGE 4 OF 6	Go Back | Continue