Since then, Graveley and his colleagues have found 47 proteins that alter Dscam splicing, and they expect to turn up more. Their strategy has been to systematically eliminate each of the 250 or so proteins in the fruit fly that bind to RNA—good candidates for splicing regulators—and examine the effects of such knockouts. Some of the molecules they've identified in their screen are generic splicing factors, thought to be required for the splicing of many genes in a variety of cell types; others are novel, perhaps controlling a more defined set of spliced exons. Most of the factors regulate the inclusion of a single exon within Dscam.

"When we knock these out, they don't seem to just globally screw up splicing—they're very specific alternative splicing events, which was a little bit surprising," Graveley says. "So we definitely know that there are some exons that are highly regulated."

And this, Zipursky says, is a system that is in some ways random. Since his initial discovery of the gene, Zipursky has found that Dscam's role is to enable the extending processes of a neuron's dendrites and axons to distinguish between themselves and their neighbors—and he thinks that by producing so many forms of the protein, the fruit fly has actually minimized the need for elaborate control.

Each neuron expresses 10 to 50 forms of Dscam at any given time, and Zipursky's group has shown that each form has unique recognition properties. What is important, he says, is that cells express forms of Dscam that are different from one another. With more than 38,000 to choose from, odds are—even if they're randomly selected—one cell's Dscam forms won't be the same as those expressed by a neighboring cell.

"The fly has invested in making many, many different types of forms," Zipursky concludes, "but it hasn't invested in the detailed control that would be necessary to make specific forms in specific neurons. That would require a tremendous investment in genetic regulation."

		Fundamentally, alternative splicing is little different from routine, or constitutive, splicing. View Illustration Scientists have known since the 1970s that, in eukaryotic organisms, gene sequences that encode proteins are interrupted by segments of noncoding DNA known as introns. These regions, which can make up more than 90 percent of a gene, are transcribed into RNA along with the interspersed coding segments, or exons. Because they cannot be used during the translation of RNA to protein, they must be removed. A complex of proteins and RNA known as the spliceosome is responsible for identifying splice sites on an RNA molecule, snipping out the appropriate segments, and reconnecting the severed transcript. For some transcripts, this process is always the same: Introns are removed and the remaining exons are pieced back together. But for alternatively spliced transcripts, things are more complicated. A subset of exons is often removed along with the introns (the most prevalent form of alternative splicing in mammals); in other cases (most commonly in plants and lower animals), introns are retained in the final transcript. The inclusion or exclusion of each segment determines the structure of the resulting protein, and whether to remove an alternatively spliced exon is a decision the spliceosome must make each time a gene is transcribed, with the assistance of a sizable collection of regulatory molecules.

	PAGE 5 OF 6	Go Back | Continue