Structures May Not Solve Everything
This week's C&EN cover story is about how X-ray crystal structures of G-protein coupled receptors (GPCRs) help the hunt for new drugs. GPCRs are already a major target for drugs (if not the most popular drug target), but until recently, researchers knew little about the finer points of their structures.
As I mentioned in that story, those high-resolution protein pictures aren't a panacea, and they won't replace established drug-discovery technology so much as complement it. I didn't have room to flesh out that idea in print, so I'm posting a few researchers' thoughts on this area here today.
Some scientists thought that GPCR X-ray structures are so far of limited utility for discovering allosteric drugs, a class of GPCR-targeted drugs that can dial activity up or down rather than turning it on or off. Some GPCR-targeted drugs on the market already work this way, such as the HIV medication Maraviroc, and many more are in development. (As an aside, I feel as though every time I attend an ACS meeting talk about GPCRs, the room is packed).
“It’s the chicken and the egg story,” says Robert Lutjens, head of core biology at Addex Pharmaceuticals, which specializes in GPCR drug discovery. To get an X-ray structure of an allosteric molecule binding to a GPCR, which would be useful for developing virtual screens, one would first need to find just the right allosteric molecule—one that stabilizes the GPCR sufficiently to enable it to be crystallized. That’s difficult to do, so powerful biological assays are still critical for finding molecules that act at allosteric sites, Lutjens says.
Allosteric drugs aside, Rockefeller University biochemist Thomas P. Sakmar told me that GPCR structures don’t show the detailed dynamics of GPCRs in complexes with protein partners and in the environment of the cell membrane—information he says will be critical for developing drugs. Researchers should further develop nuclear magnetic resonance methods and computational methods to study the molecular motions of GPCR complexes, he says. One computational technique, molecular dynamics, is designed to simulate atoms’ motions over time like a virtual microscope, but it is in its infancy in the GPCR field, he adds.
Finally, GPCR structures don’t take decades to solve any more, but the structural biology is still a bit slower than what pharmaceutical drug development projects typically require, says Brian K. Kobilka, a Stanford biochemist who specializes in solving GPCR structures. It might be possible to get a new GPCR target’s structure in the first year of a drug discovery program, but it’s not something to bank on just yet, he adds. And without a structure or a reliable model, virtual screening isn’t an option, so “traditional screening is still going to be a major contributor to identifying hits for the time being,” he says.