"The race is on between sophisticated lab procedures and their computational analogues", stated team leader J. Andrew McCammon, a Howard Hughes Medical Institute investigator and UCSD professor of Pharmacology who holds the Joseph Mayer Chair of Theoretical Chemistry. "Our method both verifies and competes with costly lab methods that rely on thousands of trials of nearly similar molecules. And it does not require the synthesis and purification of proteins, nor is it limited by the sizes of the molecules." For their calculations, Dr. McCammon's group used computers at the San Diego Supercomputer Center (SDSC) and an SDSC "satellite site" at UCSD.
Drugs are typically small molecules, called "ligands", which bind tightly to some target receptor protein, either inhibiting or enhancing the protein's activity. Modern drug designers usually begin with the crystal structure of the receptor protein. "Designers then try to figure out where the likely sites for binding a ligand might be", Dr. McCammon explained. "But almost all the methods treat the crystal structure as a rigid object, when the structure is actually only a frozen snapshot of a molecule that is always in motion at room temperature."
The new method from UCSD samples the motions of the larger receptor, then sees how the ligand binds to an "ensemble" of receptor conformations. "Small ligands may bind most tightly while the receptor is in a stretched or curled position very different from the original crystal conformation", stated Jung-Hsin Lin, a postdoctoral researcher in Dr. McCammon's group who developed the method together with Dr. McCammon and two graduate students, Alex Perryman and Julie Schames.
The researchers tested the method on a receptor for the drug FK506, also called tacrolimus, a potent immune system suppressor used to prevent the rejection of transplanted organs. Natural FK506 was derived from a soil fungus and developed as a drug by the Fujisawa Company of Japan. But because of side effects, scientists are seeking less toxic analogues of FK506. Two laboratory methods of conducting such a search were published in 1996 and 2000 by scientists at pharmaceutical companies.
"The computational method of doing the same search is what we have developed", Dr. Lin stated. "We call this the relaxed-complex scheme because we first allow the receptor protein to relax into any possible conformation, then make computational complexes of receptor and ligand to test for binding affinity. This allows us to use a building-block approach to constructing the best ligands for a receptor."
The method could be applied to every conceivable type of drug for which there are three-dimensional structures or predicted structures for the target to which the drug would bind. While it is computationally intensive, Dr. McCammon pointed out that by avoiding a brute-force approach in the lab involving protein synthesis and purification, synthesis of the many ligands, and a large number of trials, drug designers may be able to save both time and money by using the new method.
The computers used for the study included Blue Horizon, an IBM supercomputer located at the San Diego Supercomputer Center, and the Keck II satellite cluster funded by the W.M. Keck Foundation of Los Angeles in Urey Hall. "The high-speed connections turn all these machines into a single grid, on which our calculations can run very quickly", Dr. McCammon stated.
The research was supported in part by the Howard Hughes Medical Institute which offered an investigatorship for Dr. McCammon and a predoctoral fellowship for Alex Perryman, the National Science Foundation (NSF), the National Institutes of Health, the W. M. Keck Foundation, and the National Biomedical Computation Resource at SDSC. Blue Horizon and SDSC are also supported by NSF through the National Partnership for Advanced Computational Infrastructure (NPACI).
The San Diego Supercomputer Center (SDSC) is an organised research unit of UCSD and the leading-edge site of NPACI. SDSC's mission is to develop and use technology to advance science, and SDSC provides leadership both nationally and internationally in computing, data management, biosciences, and other areas. As a national laboratory for computational science and engineering, SDSC is funded by the National Science Foundation through NPACI and other federal agencies, the State and University of California, and private organisations.