GNN - Genome News Network  
  Home | About | Topics
Watching Genomes Work
A technology called TraSH asks, What do all these genes do?
Edward R. Winstead

Featured article.

In the future every genome will have a genome project of its own. Hundreds already do. The number of genome or chromosome projects launched since 1995 is about 480; the challenge now is to figure out what all these newly sequenced genes and genomes do. Two teams of Harvard scientists, working independently, have developed a tool for the task.

The tool is known as TraSH, which stands for Transposon Site Hybridization. It is a genomic version of the 'knockout' studies researchers have done in mice for decades: Mutate a gene and see what happens to the mouse. If the animal has health problems, researchers can hypothesize about what the gene does in healthy animals.

‘The discovery of this technology was probably inevitable.’

TraSH operates on the same principle in colonies of single-celled organisms like bacteria. Instead of creating one mutation per mouse, researchers create one mutation per cell. But there are so many cells per colony (about 100,000), that the whole genome, in effect, is mutated. After colonies of mutant cells have grown for several weeks in different pools, the pools are analyzed using DNA microarrays, or gene chips.

The chips reveal which cells lived and which died. Bacteria that died had mutations in genes they could not live without—under those conditions. It is possible to predict biological roles for genes by knowing the growth conditions under which the genes are essential. For example, genes that synthesize amino acids are essential in pools that lack amino acids.

"TraSH asks which genes are required for an organism to survive under two growth conditions," says Eric J. Rubin, of Harvard School of Public Health, who led the project. The tool was created as part of an ongoing project to develop vaccines against strains of the tuberculosis bacterium, M. tuberculosis, that are resistant to the current vaccine.

The researchers are using TraSH to identify M. tuberculosis genes that, when deleted, will make a strain of the bacterium harmless to humans but still effective as a vaccine. In the process, they are also classifying genes in functional categories based on what they learn from the experiments. The development of TraSH was described in the October 23 issue of Proceedings of the National Academy of Sciences.

One of the reviewers of that paper commented that as a name for the new method, TraSH was perhaps a little "juvenile." But there were no catchy alternatives. This may be why, when George M. Church, of Harvard Medical School, and colleagues came up their own, very similar version of TraSH, they did not name it. Their research appears this month in Nature Biotechnology.

Biologists around the world are trying to turn genomic data into knowledge, and both new studies might have come from researchers on different continents. Instead, they were done at the same university in buildings separated by a parking lot. "I can see the windows of the Church laboratory from my office," says Rubin.

"The most striking difference between the two studies is the organisms used in each," says Church. His group used E. coli, while Rubin's used a strain of the tuberculosis bacterium. Both studies were done primarily to see if the technology would work.

The similarities between the projects, including the publication of scientific papers ten days apart, are coincidental. The teams started and completed the projects separately, and only recently did they exchange detailed notes. But the researchers point out that TraSH represents a logical next step for the field.

"As with some other important technologies, such as PCR, the discovery of this technology was probably inevitable," says Church, referring to a technique for amplifying DNA. "It was simply a question of who would stumble upon it first. In this case, we both stumbled upon it at the same time."

TraSH is the latest incarnation of the versatile microarray. Best known as a way to profile gene expression, gene chips have also been used to compare the content of genomes. In the new technology, they serve as sensors, detecting the number of cell survivors and the locations of each mutation in the genome.

The technique of inserting DNA to cause mutations, known as transposon insertions, has been around for years. But determining the locations of mutations was done one mutation at a time. In TraSH, the mutation itself generates a marker that can be read by the microarray, and the entire genome is analyzed through parallel steps. The insertion of DNA to mutate the bacteria, the identification of mutations, and the assessment of the effects of mutations are done for the entire genome simultaneously.

Schematic of TraSH. View larger

"In one experiment, we defined the complete set of essential genes for growth in the minimal environment but not the nutrient-rich environment," says Rubin. The bacterial genes that are essential to cause human disease are the ones they hope to discover next.

Taking a step toward this goal, Rubin's colleague Christopher M. Sassetti has infected mice with colonies of mutant bacteria, while at the same time growing a control group of bacteria in a dish. In several weeks, he will remove the colonies and use microarrays to assess the results.

"Identifying essential genes for infection in mice puts us a step closer to understanding how the bacterium causes disease," says Sassetti. "We are primarily interested in the organism, and for us, the technology is a means to an end."

A second reason to develop the technology was to learn about M. tuberculosis, which has been sequenced but is still very much a mysterious pathogen. The Sanger Centre in Great Britain sequenced the H37Rv strain of M. tuberculosis in 1998. A second sequenced strain is available from The Institute for Genomic Research (TIGR) in Rockville, Maryland.

The new method is particularly useful for investigating M. tuberculosis because it limits contact with the organism, which is tough to handle in the laboratory and which grows slowly. But Rubin says that TraSH should work well in most microorganisms, which comprise the largest portion of sequenced genomes.

"The genome projects have generated an enormous amount of data, and much of this information has never been annotated," says Rubin. "The new technology—and others like it—can help us figure out what all these unknown genes do."

Church had the same idea when his team began developing their version of TraSH a few years ago. "We have been interested in functional genomics since before the term was coined, and it struck me that insertional mutagenesis and microarrays could be components of a new strategy."

For some perspective on the efficiency and scale of the new technology, he says, consider the experience of yeast researchers. After the organism was sequenced, the community divided up the genome and created individual mutants for every gene possible. This took several years.

"To repeat this process over and over for every organism" is no longer practical, Church argues, particularly given the number of pathogens being sequenced and the need to identify gene targets for vaccines and therapies. Furthermore, the new approach peppers the entire genome with mutations, allowing researchers to see the effects of disrupting different parts of genes.

Dozens of cells will have mutations in the same gene, but each mutation falls in a different part of the gene. Importantly, regions that regulate genes will be disrupted. Church's laboratory is particularly interested in how genes are regulated. "This lets you do a fine structural analysis of mutations throughout the entire genome," says Vasudeo Badarinarayana, who led their analysis.

Microarray used to create TraSH.

TraSH also provides a perspective on existing microarray data. In Church's laboratory, Jeremy Edwards compiled a database of all available information on E. coli, and this allowed the researchers to cross-reference findings and check the accuracy of predications. They classified several previously uncharacterized genes and learned that some highly expressed genes were not essential for growth.

Church's and Rubin's team did not communicate much until recently, but they shared notes earlier in the year. "We were reassured by the fact that both groups were making progress," says Church. "We knew we were on the right track."

. . .

Sassetti, C.M. et al. Comprehensive identification of conditionally essential genes in mycobacteria. Proc Natl Acad Sci USA 98, 12712-12717 (October 23, 2001).
Badarinarayana, V. et al. Selection analyses of insertional mutants using subgenic-resolution arrays. Nat Biotechnol 19, 1060-1065 (November 19, 2001).

Back to GNN Home Page