Genetics: Level-headed Stardust Knows Which Way Is Up Chronic Diseases: Sickle Cell Disease Cured in Mouse Model Disease Profiling: Diagnosis by Database Shows Promise Genomics: Technique Enables Quick Accounting of Gene Function Medical Ethics: Panelists Frame Ethics of Stem Cell Debate Primary Care: Summers on Patient Care Pain and Pleasure Activate Same Brain Structures Microbial Master of Disguise is Unmasked Risk of Mad Cow Disease in U.S. Called Low Animal Model for Obesity Developed Center for Educational Technology Opens News Brief In Memoriam: John Brooks Thomas Durant W. Morton Grant Francis Moore A Better Way to Care for Teen Moms Front Page

GENOMICS Technique Enables Quick Accounting of Gene Function They are the new classifiers. But rather than scarabs and stink bugs pinned in iridescent arrays, their obsession is a DNA microchip with the coloration of a butterfly wing and the size of a fish scale. Now that whole genomes have been sequenced, and the machinery to generate endless strings of GATCs is running 24/7, a group of scientists has geared up for the next phase: identification and classification of newly discovered coding regions. Eric Rubin and postdoc Christopher Sassetti (l to r, photo at left) and George Church (photo at right) have independently developed a method to screen genes for function on a genomewide scale, greatly extending the range of the already versatile DNA microchip. Photos by Graham Ramsay It is a daunting task. According to George Church, HMS professor of genetics, approximately 600,000 bacterial genes will be sequenced in the coming two years. And while the naturalists of yesteryear may have been interested in morphological description, today's biologists want to define function. "We need to have a variety of ways to figure out not just what the genes are doing, but we also want to investigate functional domains and control regions." This can be accomplished through mutagenesis and a judicious selection of growth conditions, but today's challenge is to interpret the sequence data without "handcrafting" every single mutant. Eric Rubin, HSPH assistant professor of immunology and infectious disease, and Church have each independently developed a new method that marries the best of microchip analysis and another technique, transposon mutagenesis, to evaluate the function of newly sequenced genes. Equally important, they can make this assessment on a genomewide scale, sorting thousands of coding regions at once. Double Take: A Look at Parallel Discovery "Ideas do not belong to anyone," the Bishop said to the Marquis, in Gabriel García Márquez's Of Love and Other Demons. "They fly around up there like the angels." Indeed, it is as if George Church, from HMS, and Eric Rubin, from HSPH, independently plucked the same idea from the heavens (see main story). The history of science is littered with coincidental, independent discoveries. In a letter to Sir Charles Lyell on a paper he just received from Alfred Russel Wallace (and that he would later submit to Nature together with his own theory of natural selection), Charles Darwin wrote: "I never saw a more striking coincidence.... If Wallace had my manuscript sketch written out in 1842, he could not have made a better short abstract! Even his terms now stand as heads of my chapters." In truth, scientific thought is a progression and ideas are not born in a vacuum. Wallace and Darwin were strongly influenced by the Essay on Population by Thomas Malthus. Likewise, scientists today work in an intellectual crucible. "The discovery of this technology was probably inevitable," said Church. But he who stumbles upon an idea first may need more than luck. Rubin lab postdoc Christopher Sassetti and Church lab postdoc Vasudeo Badarinarayana both ran up a lot of blind alleys before they found the conditions that made their experiments work. Yet in the end, both methods are very similar, to the smallest details. "It's uncanny," said Rubin, who has since heard that he and Church were not the only ones working on the problem. "The fact that our methodology is so incredibly similar probably means that there may not be an infinite number of experimental solutions." Whatever else came into play, geography seems irrelevant. Wallace was on the other side of the world from Darwin, starving and feverish in Ternate, when inspiration struck. Rubin can see into Church's lab when he looks out his window. The DNA microchip, developed just a few years ago, has already become a standard tool in the geneticist's repertoire. With genomic sequence in hand, the researcher can synthesize all the genes in an organism--or even just fragments of them--and then dot them in a regular pattern on a glass slide. For many organisms, up to 40 percent of the DNA sequences in such a microarray have no known function. The challenge is to design probes that make a particular spot stand out. One approach has been to evaluate gene transcription. But whole families of genes, those that are conditionally essential, for example, will not show up when assaying for differential expression. As Church explained, "Microarrays have typically been used to quantitate mRNAs in a mixture; but we wanted to measure DNA." Rubin has another way of looking at it: "Earlier studies looked at what's turned on, not what is necessary." Fitness Test Essential genes, those necessary for robust growth or survival, can be conditional depending on environment. The classic example is a gene that synthesizes an amino acid in minimal medium. To find these conditionally essential genes, Christopher Sassetti, working in the Rubin lab, and Vasudeo Badarinarayana, in Church's lab, introduced mutations into bacteria using specially constructed transposon vectors. "This allowed us to make many mutations per gene for all the genes in the genome and assess their relative contribution to any phenomenon for which we could devise a selection," said Rubin. His lab mutated the M. tuberculosis genome while the Church group studied E. coli. Their results appear in the Oct. 23 Proceedings of the National Academy of Sciences and the November Nature Biotechnology, respectively. Unlike point mutations or deletions, the transposon introduces a tag into the DNA. It is this element that the researchers latch onto when searching for mutants. Thus, each individual gene can be assayed for an insertion by demanding the presence of a transposon landmark in or near the gene of interest. If the insertion compromises growth, that particular mutant will be underrepresented or absent from the pool of bacteria, many of which harbor mutations in nonessential genes. The absence of the mutant bacterium by death or retardation of growth then fails to produce a signal upon amplification by PCR. This absence can be detected by hybridization to the microarray. "It's like taking a snapshot," said Church. "Now, if you change the conditions, where some things grow well and others less so, you get another image." If DNA from the two cultures is labeled with color-coded dyes, then mixed and allowed to seek out its target DNA, the differences are rendered as a color display on the chip. Traditionally, insertion deletions would be analyzed using an involved method called footprint analysis, in which the mutants are investigated one by one using gel electrophoresis. "I don't want to dismiss footprinting," said Rubin. "You get much finer discrimination of each area of the chromosome with it. But our new technique is a lot faster and easier." He backs up this statement with a quick calculation: for a genome the size of M. tuberculosis, the output from one microchip is equivalent to 200 gels. The two studies, published simultaneously, provide the proof of principle that this new technology works. They identified genes both known and unknown that are conditionally essential in a minimal environment. But both groups have bigger fish to fry. Biologists with an Attitude Rubin and Church are smack in the middle of functional genomics, a moniker that even researchers in the field find hard to define. "Some say, oh, we've just returned to biology," said Church. "But it's more like we've made a spiral. Because we bring with us the clues from genomics that help us to identify RNAs and proteins, and we've got attitude--the comprehensive attitude--we want to model every single gene, and we've got automation." Rubin is committed to tuberculosis. It seems reasonable that the complement of essential genes will differ depending upon whether a pathogen grows in a petri dish or in the lungs of a patient. That is why he is growing the M. tuberculosis transposon library in a mouse. "Using these mutants, we can investigate virulence and genes involved in fighting the host immune response." And experiments are under way to identify those genes that when mutated, produce attenuated bacteria that cannot cause disease anymore. "Each one represents a candidate for vaccine," Rubin said. --Anne Mahon