The Next Frontier: Proteomics

By Unmesh Kher

Even before they finished decoding the human genome, scientists began the next and far more challenging step in explaining the molecular underpinnings of life. It's called proteomics--the cataloging and analysis of every protein in the human body.

Although proteins are the direct result of the instructions coded in our DNA, they are far more variegated and complex than DNA. They have to be. Every chemical reaction essential to life depends in one way or another on their services. Proteins are the beams and rafters of the cell and the glue that binds the body together; they're the hormones that course through our veins and the guided missiles that target infections; they're the enzymes that build up and break down our energy reserves and the circuits that power movement and thought.

What a protein does is largely determined by its shape. Proteins are stippled with pockets and grooves into which molecules fit as precisely as a key fits a lock. To fully understand how a protein works, you have to be familiar with every nook and cranny on its surface, which is why the National Institute of General Medical Sciences will spend $20 million this fall to establish a series of research centers dedicated to a branch of proteomics known as structural genomics. The centers will detail, over the next 10 years, the shapes of 10,000 proteins. That's a tiny fraction of all the proteins found in nature, but the NIGMS thinks that number will cover most of the structures relevant to biology and medicine.

Why not study all the proteins? For one thing, there are too many--50,000 to 2 million, depending how you count. For another, most of those millions of proteins are just variations on a handful of themes. Proteins with similar functions--be they in insect, worm or man--often share structural characteristics that are reflected in the genes that encode them. Structural biologist Stephen Burley of Rockefeller University estimates that the maximum number of distinct shapes may be as few as 5,000. The NIGMS hopes to construct a lexicon of shapes--barrels, doughnuts, globular spheres, molecular zippers and so on--that when mixed and matched will spell the shape of any gene's product. About 1,000 of these structures--and the genes that code for them--have already been cataloged.

But naming the shapes and knowing precisely how they are formed are two different things. Proteins twist and pleat themselves as they're synthesized. These basic forms are then further folded and linked to other proteins to create the uberstructures crucial to protein chemistry. Scientists traditionally dissect the atomic details of these folds by observing how crystallized proteins scatter X rays--experiments that can take years to complete. But robots and powerful X-ray generators have lately boosted the pace of discovery. Structures that two decades ago would have taken a couple of researchers 10 years to crack can now be solved by one in a matter of weeks. "By the end of the five-year pilot phase," predicts John Norvell, director of the NIGMS program, "each of the centers will be producing 100 to 200 protein structures a year."

That may not be fast enough. Already, a host of biotech firms--including PE Corp., the parent company of Craig Venter's Celera--have launched their own proteomics programs, some focused on protein structures, others trying to determine where in the body different proteins are produced and how each is controlled. The flurry of private activity raises the specter of intellectual-property disputes like those that plagued the Human Genome Project. Last spring the NIGMS co-sponsored the first international structural-genomics meeting, partly to nip those conflicts in the bud. The hope is that history will repeat itself a little differently this time around.

--By Unmesh Kher