BERKELEY, Calif.--Using a pair of radical bioinformatics strategies for headlights, a team of university and private-sector researchers have peered into the dark terrain of the Mycobacterium tuberculosis genome and identified a host of potential drug targets.
David Eisenberg and colleagues at the University of California, Los Angeles/Department of Energy Laboratory of Structural Biology and Molecular Medicine introduced their search strategies, called phylogenetic profiling and Rosetta-Stone analysis, in a series of papers last year. To exploit the technologies, the scientists formed Protein Pathways, a company that the US National Institutes of Health has awarded $100,000 to analyze the M. tuberculosis genome.
Eisenberg recently presented the group's findings in a talk on the University of California campus here. Genes identified by their methods may yield targets for the next generation of drugs against the increasingly resistant and deadly microbe.
What's more, the findings bolster evidence that the strategies can discern functional relationships among sequences that resemble none previously characterized. A computational strategy published in 1998 by Terry Gaasterland and Mark Ragan while they were at Argonne National Laboratory, did not suggest such broad promise. Gaasterland told BioInform that their implementation of an approach similar to one of Eisenberg's failed more often than it succeeded to group proteins that made up the same biochemical pathways.
But many regard with awe the techniques that Eisenberg described. Structural biologist Tom Alber of University of California, Berkeley, called them "the next step beyond sequence homology for assigning function."
Simple principles underlie the two approaches. The premise behind phylogenetic profiling, Eisenberg said, is that proteins work in functional groups--such as the assembly lines that synthesize small biomolecules and the tag-teams that react to particular cell signals. Proteins in such groups could be expected to appear together in every organism in which they exist, Eisenberg argued.
Eisenberg and his colleagues created a "profile" for each protein in Mycobacterium tuberculosis, indicating which other organisms did and which did not express a homolog. They classified proteins with like profiles as likely to be associated functionally, and they assigned a broad role to whole groups if at least one member shared homology with a protein whose function bench scientists had characterized.
The researchers identified members of two types of candidate drug target groups: Those containing a protein targeted by a current antibiotic and those with a protein homologous to any that had been shown to be essential in other microbes but not in humans. They then expanded these groupings by the Rosetta-Stone method, which Eisenberg said identifies proteins that directly couple to one another by binding.
Such relationships appear often to evolve through an incubator stage, in which precursors of the two proteins appear as domains within a single polypeptide chain, Eisenberg explained.
To identify physically interacting proteins in M. tuberculosis, the researchers combed other completed genomes for relic sequences--so-called rosetta stones--in which regions homologous to two tuberculosis genes exist within a single open reading frame.
Eisenberg said that by applying the two techniques together, his team identified 59 candidate drug targets. He noted that strains of the tuberculosis pathogen have emerged that resist multiple drugs, making the need for new drugs increasingly urgent.
The analyses turned up two proteins targeted already by existing antibiotics. The teams' search methods grouped these with proteins that resembled essential proteins in other microbes. Alber called this "the proof of the pudding" and a validation of the researchers' bioinformatics approach.
According to Kenneth Goodwill, Protein Pathways' business development director, the company has an exclusive license to use phylogenetic profiling and Rosetta-Stone analysis from the University of California, which he said is seeking to patent the inventions.