First, scientists learned to crack entire genomes. Now, they're ready for the next leap: determining the three-dimensional structure of each and every protein that an organism produces. The results of a pilot study show that such a mass approach to structural biology is feasible, provided researchers develop new techniques to speed up the process.
Traditionally, structural biologists have picked specific, interesting proteins and spent months or years solving their atomic structures. But just as DNA sequencing was scaled up dramatically in the 1990s, many researchers now hope to determine the structure of many proteins in parallel--a field dubbed "structural proteomics." Knowing the shape of every protein would yield new clues to their function and how they work, and ultimately provide thousands of potential new drug targets. Earlier this week, the National Institute of General Medical Sciences (NIGMS) announced that it would award nearly $150 million dollars to seven consortiums for a 5-year structural proteomics effort (ScienceNOW, 26 September).
In the October issue of Nature Structural Biology, a group headed by Cheryl Arrowsmith and Aled Edwards of the University of Toronto reports the results of a first quick-and-dirty attempt. They took a simple bacterium named Methanobacterium thermoautotrophicum and looked at a sample of 424 of its 1871 genes. For each, they tried to clone the gene and express it in a bacterium, purify and crystallize the protein, and determine its structure through x-ray crystallography or nuclear magnetic resonance spectroscopy.
For each of these steps, every protein has its own ideal set of conditions; but in its wholesale approach, the team used only a set of standard conditions, thus dropping many proteins in the process. Even so, they were able to determine the structure of 10 proteins in the time than it often takes to solve a single structure; for several of them, the structures provided clues to their function that were not apparent from the gene sequence alone. Many more structures could be determined, says Arrowsmith, if researchers develop automated techniques to determine and apply the best conditions for each protein. But the number of structures solved will be limited by access to expensive equipment such as high-energy x-ray sources, she says.
"They've shown the potential for solving a lot of structures very quickly," says Tom Terwilliger of Los Alamos National Laboratory in New Mexico. Now the question is how to prevent more proteins from dropping out along the way, he adds. Structural biologist Gaetano Montelione of Rutgers University in New Jersey agrees: "The results are valuable as a benchmark of what's possible," he says.