It was almost 20 years ago when Loeb Professor of Cellular Biology Walter Gilbert first conceived his exon theory of genetics.

Gilbert hoped to convince fellow scientists that introns--inactive DNA segments--made genes susceptible to a type of DNA recombination that randomly reshuffled exons, or active DNA segments. Gilbert believed that random reshuffling weakens the gene and lessens its chance for survival.

Gilbert's evidence and his 1980 Nobel Prize in Chemistry, however, failed to convince doubters of his theory.

But when his latest report appears in the Proceedings of the National Academy of Science this week, Gilbert's skeptics may have to rethink their criticism.

Gilbert and his team of post-doctoral students, Sandro De Souza and Manyuan Long, have recently uncovered new evidence that draws a correlation between the structure of genes and the structure of proteins and supports Gilbert's original exon theory.

"The evidence concerns a specific group of proteins," said De Souza, a post-doctoral student who has worked with Gilbert for the past two years. "It's an important study because these implications can help us better understand protein structure."

De Souza explained that protein structure is essential to a gene's survival because when pieces of protein are thrown together, there is a slim chance these proteins will benefit the organism.

"When there is a correlation between the position of introns and protein structure, there is a better chance that the proteins will contain some advantagous characteristic," De Souza said. "Our research shows that such a correlation exists."

Gilbert's work also provides some insight concerning gene behavior at the time of the origin of life.

By looking at the structures and positions of exons and introns, Gilbert found that introns, and exons line up in a non-random way in the DNA of most living things, both primitive and modern.

"That means that ancient regions, which represent genes or portions of genes, have descended in a relatively unchanged manner from a common ancestor," Gilbert told reporters earlier this week. "This supports the theory that introns have facilitated the shuffling of exons from the time of the origin of genes."

Gilbert added that this evidence also suggests that genes are probably composed of fewer than 100 basic building blocks, not hundreds or thousands as was once believed.

But Gilbert's research does not yet conclusively prove that the common ancestor of modern life arose from the recombination of exons.

"Our theory demands certain consequences," Gilbert said. "If we make predictions based upon it and find that the predictions are correct, then we've gone as far as we can."

Projects like the Human Genome Project, a massive effort by scientists to map all the chemical bases of exons and introns, provided the information Gilbert needed to conduct his research.

"The amount of information available now is huge when compared to the information available before the Human Genome Project and other genome projects like it," De Souza said. "Because of projects like that, studies like our own can add to the body of information on genes."