By Dean Haycock

Special To BioWorld Today

If test-tube experiments produce in vitro results, and animal studies yield in vivo results, then computer analysis of genomic data may be said to yield "in silico" results. The availability and potential of computer databases, stuffed with byte after byte of gene sequences, have made the computer as familiar a scientific tool as the petri dish. In a paper in the September issue of Nature Biotechnology, information obtained "in silico" is combined with information provided by bacteria growing on petri dishes. The results may lead to novel drug targets and new insights into the nature of conserved and essential genes in all organisms.

It seems reasonable to assume that evolution would favor the conservation of essential genes for use in different species. The assumption satisfies common sense: If everybody has a gene, the reasoning goes, it must be really important, something that no one can do without. Essential genes are those whose products are necessary to sustain life. It makes sense to assume that many or most conserved genes would be essential.

Many bettors would have wagered on that proposition before Hannes Loferer and his colleagues then working at the Geneva Biomedical Research Institute, Glaxo Wellcome Research and Development S.A., in Geneva, Switzerland, published "A genome-based approach for the identification of essential bacterial genes."

They present data indicating only a small percentage of conserved genes is essential for bacterial survival. But, by picking out the truly essential genes, the biotechnology scientists hope to identify novel targets for future antibiotics.

Could Be A 'Dream Way' To Find Antibiotics

Identifying likely gene targets depends equally on effective computer searching techniques and traditional bacterial culturing techniques. The first step was to access the already available genome of the smallest, genetically least complicated, self-replicating organism on the planet, Mycoplasma genitalium. This bacterium, which lives in close association with its host, has a mere 450 or so genes in its entire genome. The codes contained in these genes were then compared to the more complex genomes of other bacteria such as Escherichia coli, which has 4,200 genes. Any genes in common between the living minimalist M. genitalium and E. coli, Haemophilus influenzae or other complex bacteria would be conserved genes and, most would assume, essential to bacterial life. The authors further refined the list of genes they were interested in by eliminating any that had a known function.

"The reason to start off this pilot experiment by comparing E. coli genes of unknown function with Mycoplasma was that Mycoplasma had been described in the literature as being rather close to having what we might call a "minimum genome, the minimum set of genes that are required to sustain basic replication of a cell," said Loferer, now head of microbiology at Genome Pharmaceuticals Corp., of Munich, Germany.

Stepping away from the computer, the researchers returned to the bench to determine the ability of complex bacteria to grow after each of the essential genes remaining on their list had been knocked out or eliminated. Surprisingly, only six out of 26 of the conserved genes common to M. genitalium and E. coli turned out to be essential.

The authors speculate other genes in the E. coli genome may be able to "cover" or substitute for the function of what appear to be conserved but not essential genes in the knockout stains.

M. genitalium and E. coli are, of course, both members of the same kingdom, known as Monera to taxonomists. When the authors turned their attention to organisms from a different kingdom, Mycetae or Fungus, they found that even fewer of the six essential genes identified in bacteria were essential for growth in yeast.

Only two of the six genes essential to E. coli were essential to the yeast Saccharomyces cerevisiae. The researchers had thus identified four genes that would make promising targets for new broad spectrum antibiotics to counter bacterial infection but which might not be expected to affect organisms belonging to other kingdoms.

George Church, professor of genetics at the Harvard Medical School, in Boston, said the method is, "in a way, a number of companies' and individual scientists' dream way for finding broad spectrum antibiotics."

Loferer emphasizes, however, that proteins encoded by genes not identified as essential in this study might still have potential as therapeutic targets under different sets of conditions. Loferer and Church are collaborating to explore this possibility.

"It is clear that deep understanding of what all these unknowns are doing is going to pay off, " Church said.

The principle of using automatic genome comparisons to pre-select a list of potential targets and validating them by genetics could also be applied to select, specific targets for chronic infections, Loferer indicated.

Hints Sought For Human Disease Models

Since publication of the Nature Biotechnology paper, the group has made an additional 100 or so knockout mutants in E. coli and identified additional targets. Ongoing research is directed at uncovering the function of some of the genes indicated by the approach described in the paper.

Despite years of searching for essential genes in organisms such as E. coli, the genes identified as essential by Loferer and his coworkers had not been identified as such before now.

"This mean that the classic techniques [for identifying essential genes] don't pick up every gene," Loferer told BioWorld Today. "On the other hand, it means that we don't know everything about the basic cell processes that are required for division and growth of E. coli."

In the near future, when the complete genomes of organisms such as the nematode worm and the fruit fly are known, it will be possible to compare them with all the available human sequences. This might provide hints of where to look first to identify good "knockout" gene candidates to provide models of human diseases, according to Loferer. *