"Reading the History of Life in the Text of Modern Genes"
By Philip J. Hilts

Is it possible to recreate the origin of genes, the first stirrings of life in the mists of time more than three billion years ago? One of the country's leading biologists is certain that with the power of thought and statistical technique those mists can be burned away. He is Dr. Walter Gilbert and he has been working on the problem for nearly 15 years.

 In a basement room at Harvard University's science center one morning a few weeks ago, Dr. Gilbert was at the blackboard, describing the puzzle to nonscientists. He noted that the genes in modern animals were stored in the DNA not as nice readable texts but as choppy fragments that the cell's machinery must stitch together. That curious arrangement is one major clue. Another is that some 95 percent of a person's DNA does not code for genes at all. It seems to be just gibberish, or junk DNA.

 Genes in pieces and huge stretches of junk? "At first glance, it is peculiar," Dr. Gilbert told the bleary-eyed students. "Then, at second glance, it is even more peculiar. I think it's one of the major unsolved mysteries of the basis of life."

 Dr. Gilbert and his colleagues have sketched out a theory of how it all may have happened and what the first genes were like. It was he who coined the terms for the interrupted pattern in which genes are stored on the DNA; the working parts he called "exons" and the regions in between, which the cell has to splice out, he termed "introns."

 After a period of controversy, then of acceptance, his ideas have come under new challenge. An electronic debate about primordial exons and introns is taking place this week and next on the Internet at http://hmsbeagle.com/scripts/web.exe. So, now, Dr. Gilbert and his colleagues have done new experiments and written new papers, using extensive computer and statistical analysis, to support his exon theory of genes.

 As Dr. Philip Sharp, a molecular biologist at the Massachusetts Institute of Technology, puts it, few scientists are willing to spend time on so vexing a puzzle. "Yes," Dr. Sharp said, "it is very difficult. It may even be unsolvable. That won't stop Wally Gilbert, of course." Dr. Sharp, who shared a Nobel Prize in Physiology or Medicine in 1993 for finding introns, one of the discoveries that sent Dr. Gilbert off on this track, said the ideas that Dr. Gilbert put forward "captured the imagination of the field, and still has it, I think."

 If Dr. Gilbert's approach is valid, in principle of the entire history of life on earth could be inferred from the DNA of modern genes. His theory is an effort to figure out how, in the primordial waters where life began, the earliest genes were assembled. A modern gene is a chemical test with a thousand of more letters. For such a structure to evolve at random would be a lengthy process.

 Instead, Dr. Gilbert suggests, the first genetic elements were simple modules, the forerunners of today's exons, and the exons were then mixed and matched to build up the lengthy chains that make longer and longer genes. By analyzing the structure of contemporary genes, it should be possible to discern the ancient modules within.

 For the modules to mix and match, Dr. Gilbert believes, they would have needed strips of extra DNA, like the leader on a reel of film, so that two modules could be spliced together without risk of cutting into their genetic message in the process. The introns, the elements interspersed between the coding regions of genes, are in this view related to the ancient leaders that enabled the modules to be assembled. Critics say the theory leaves some awkward issues unexplained. Introns are found in the DNA of plants and animals but not in that of bacteria. So did bacteria lose their introns or never acquire them?

 Dr. Jeffrey D. Palmer at Indiana University and a few other experts in molecular evolution suggest that Dr. Gilbert's theory is wrong. They say that the introns and the shuffling of genetic modules were mechanisms that did not arise until two billion years after life first began. And, they regard the introns not as clever Tinker Toy construction elements, but as just what they appear to be, pieces of junk DNA left over from billions of years of evolution.

 Noting that bacteria, presumably more primitive than plants and animals, do not have introns in their genes, they say it is more reasonable to assume bacteria never had introns than that they had and later lost them.

 But bacteria may have streamlined their genetic material by shedding their introns leading ot the "pleasantly perverse" conclusion that humans can be considered as "less evolved" than bacteria, Dr. Gilbert said with a grin. "maybe we are not in all ways at the pinnacle of evolution, as people like to think," he said. And worse, maybe nature will evolve right past humans and their junk-laden genomes in the future.

 Dr. Gilbert grew up in Boston. His father, Richard Gilbert, was a Harvard economist who belonged to Franklin D. Roosevelt's New Deal "brains trust."

 Science has always come naturally for Dr. Gilbert. As Dr. Mark Ptashne, a Harvard colleague, put it: "He would be a scientist no matter what. Put Wally in a room and give him a pencil and a pipecleaner and he will do an experiment."

 He won a Westinghouse science fellowship in high school, went to Harvard for a master's in physics, then to Cambridge University in England for a Ph.D. in mathematics. He worked in theoretical physics until 1964. Then he switched to the awakening field of molecular biology, just as did other physicists, like Max Delbruck, Leo Szilard, and Francis Crick, who became leading lights of the new subject.

 In 1966, Dr. Gilbert and Dr. Ptashne, working down the hall from each other, independently isolated and described the first two "repressor" molecules, chemical switches that keep genes turned off. In 1977, he and Dr. Frederick Sanger of England independently devised methods for rapidly reading the "letter code" of DNA. This work later won him a Nobel Prize in Chemistry. At the same time, biology had become political issue as critics protested the first uses of genetically engineered organisms. The City Council of Cambridge, Mass., threatened to shut down the biological laboratories and the buildings were cleared periodically because of bomb threats.

 Dr. Gilbert, now 64, was an advocate of the new genetic techniques and was one of the first scientist to found a biotechnology company. He studied new patent laws and management issues with the same energy he brought to pursuits like Chinese-cooking, a late-life interest in playing the piano and what is perhaps his current chief passion outside science: the study and collection of ancient, mostly Greek, sculpture.

 Following his business interest, Dr. Gilbert left academia to become chief executive officer of Biogen Inc. He ran the company for several years, but in a disagreement over management style, was dismissed by the company's board.

 He later founded the Genome Corporation, a company whose purpose was to read all of the genes in the human genome. But the company failed after the stock market crash of 1987 and by 1988, perhaps spurred by Dr. Gilbert's challenge, the Federal Government began its Human Genome Project.

 After a long career with an array of accomplishments behind him, Dr. Gilbert is now working at Harvard on the most difficult questions in biology. In his laboratory, young scientists like Sandra De Souz and Mayuan Long are carrying out experiments to probe the origin of the genes. If the genes are indeed composed of smaller modules formed at the dawn of life, it may be possible to figure out when each gene evolved. The theory may hold a practical consequence for drug design, because it may allow scientists to recognize and thus to manipulate, the different working parts within proteins.

 Dr. Gilbert an dhis colleagues have developed interesting evidence in support of the "exon theory of genes." If genes were originally assembled Tinker Toy-style, some signs of the connectors and joints, as well as of the modules, should still be visible. One such sign is that since the genetic information is processed by the cell in groups of three letters at a time, the parts being mixed and matched needed to be in phase, since if the processing machinery gets out of step the genetic message will be garbled.

 In a paper published last year, Dr. Gilbert and his colleagues reported finding just such evidence. In looking at 11,000 introns, they found that their ends were still most often "in phase" -- so they could snap together just as in ancient days.

 Dr. Gilbert believes that he has recently found another faint signature of ancient exons in the structure of modern proteins. Many proteins have a complex architecture in which each region or module performs a different biochemical operation. If these modules are exons that have been repeatedly shuffled around in the course of evolution, then in the protein's gene the introns should still tend to be found at the points between the modules.

 In a paper to be published next month in The Proceedings of the National Academy of Sciences, Dr. Gilbert, Dr. Long and Dr. De Souza say they have found that introns do indeed occur more often at the boundaries between modules, and that these modules seem to have come in four different sizes. The size pattern is what might be expected if the first exons were small, coding for 15 to 20 of the amino acid units of proteins, and if some later fused into double, triple or quadruple exons.  

Critics are skeptical about the conclusions, noting that they rely on statistical arguments, and that these may not hold up as more data emerge from genome-sequencing projects. 

Dr. Palmer of Indiana University says that genes do get shuffled but that it is not yet known when the shuffling mechanism first appeared. "New genes do arise by mixing and matching parts," Dr. Palmer said. "That part of Gilbert's theory has been borne out wonderfully. But the question is whether this kind of shuffling occurred in the primordial times to assemble the first sets of genes. For that, the evidence is not there."

Dr. Gilbert acknowledges that the data are not overwhelming, but notes that the flood of new data from genome sequencing projects will soon allow the theory to be put to a more exact test. "We are going to know the answers," he said tentatively. Then, with more conviction, he declared: "Yes. We are definitely going to know the answers."