The P450 Puzzle

No matter how good a potential new drug looks in test-tube experiments, it won’t work if it can’t be broken down by the cells of the liver and removed from the bloodstream before it reaches toxic levels. The prime players in this process are a class of enzymes known as membrane-bound P450s. But these mysterious proteins’ structures are so poorly understood that scientists can’t predict whether new compounds will interact properly with them. That forces drug developers to guess whether it’s worth embarking on expensive and time-consuming animal studies.

Last year, UTMB researchers took a big step toward solving that problem: they produced the most detailed image ever made of a mammalian membrane-bound P450, revealing a previously unseen feature that provides hints about how the protein and others in its class break down foreign chemicals such as drugs. “This structure is very different from pretty much any P450 in certain regions of the protein,” says Emily Scott, a UTMB postdoctoral fellow in pharmacology and toxicology and lead author on the paper describing the discovery, which appeared in the November 11, 2003, issue of the Proceedings of the National Academy of Sciences. In previously determined P450 structures, the “active site”—the portion of the enzyme that works to break down drugs—was closed off behind what might be visualized as the bars of a molecular cage. It was hard to see how a substrate, the molecule to be broken down by the enzyme, could get to the active site in these structures.

“But this version of the protein has a very open cleft leading to the active site,” Scott says. “For the first time we’re seeing the protein really open up, and it’s very obvious how substrates can get to the business end of the protein.”

Scott collaborated on the project with Pharmacology and Toxicology Chairman James Halpert and senior lab associate You Ai He, as well as Christopher Chin and Mark White of the Sealy Center for Structural Biology. The structures of proteins that reside in cell membranes are extremely difficult to determine; the only previous membrane-bound P450 structures were obtained only by changing a number of important amino acids to make the structures easier to crystallize. Chin’s and White’s expertise was essential to finding a membrane-bound P450—in this case, derived from the membranes of rabbit liver cells—that could be crystallized with a smaller number of changes in less critical amino acids, and to show that membrane-bound P450 could yield detailed structural information.

“This new conformation raises a lot of interesting questions about how we model the interactions of these compounds with P450s,” Halpert says. “All the modeling that’s been done up to now has been based on the closed form of the enzyme. Now we have to ask, do we start from this thing and try to see how the enzyme might close up? I’m not sure we know the answer yet, because this is so brand new.”—Jim Kelly