APPLICATIONS

Using X-ray data and advanced computer simulations carried out on TeraGrid systems at the National Center for Supercomputing Applications (NCSA), the Pittsburgh Supercomputing Center (PSC), and Indiana University, researchers have modeled a critical part of a mechanism by which bacteria take up large molecules. Their findings provide a rare window on the complex interplay of proteins involved in the active transport of materials across cell membranes. The study appears online in the Biophysical Journal and was described in the May 25, 2007 edition of Science. Transporting large molecule -- such as vitamin B12, citric acid or other vital nutrients -- across the outer membranes of Gram-negative bacteria is not a simple task. The cells must be selective in what substances they take up, and the outer membrane contains no energy-generating machinery to power the job of hauling large molecules inside. The new study examined an outer membrane transport system that depends on an energy-generating inner membrane protein, TonB. This TonB-dependent transporter (TBDT) contains a beta-barrel domain: a series of parallel sheets that form a tunnel through which large molecules can pass. Another region of the protein, the luminal domain, clogs this barrel until the cell is ready to allow large molecules to pass through. Crystallographic studies had shown that TonB binds to one end of the luminal domain. Researchers had hypothesized that TonB somehow draws the luminal domain out of the barrel or changes its conformation to make way for the large molecules. Previous studies had been inconclusive, however. Molecular biologists have difficulty studying systems that involve complex interactions between proteins, particularly when one domain moves into and out of a structure like a beta-barrel, said University of Illinois biochemistry professor Emad Tajkhorshid, principal investigator on the study. "It's very difficult to assess this experimentally," Tajkhorshid said. Instead, Tajkhorshid and graduate student James Gumbart used NAMD and VMD (programs developed in the University of Illinois' National Institutes of Health Resource for Macromolecular Modeling and Bioinformatics) to simulate and visualize complex molecular interactions. By entering detailed data about the position and characteristics of every atom in the system, the researchers ran simulations of various scenarios to test which hypotheses were most feasible. Their work relied on detailed crystallographic studies of the molecules provided by University of Virginia researcher Michael C. Wiener. "The good thing about simulations is that you can monitor the position of every atom," Tajkhorshid said. The task was enormous, however. "These simulations were computationally extremely demanding, and were made possible only because of computational resources provided at our NSF-supported national centers," Tajkhorshid said. "The computer time and great support and cooperation of the staff at these centers to optimize our program and facilitate our research is highly appreciated." The researchers addressed two key questions: First, could the bond between TonB and the luminal domain withstand the force needed to pull the luminal domain downward, away from the barrel? Second, how does the luminal domain respond to force in order to expose a permeation pathway through the barrel? In the first simulation the researchers applied a force to TonB and showed that the multiple hydrogen bonds between TonB and the luminal domain were strong enough to remain intact while TonB pulled the end of the luminal domain away from the barrel. The simulation also showed the luminal domain gradually unfolding, changing its conformation in ways that would open up enough space for a molecule of vitamin B12 to pass through the barrel. A second simulation exerted a force near the center of gravity of the luminal domain, pulling it out of the barrel in one piece. This required an enormous input of energy, however. "The force that we applied was about 4,000 picoNewtons," Tajkhorshid said. "This is an order of magnitude higher than that required to induce the unfolding. This clearly indicates that the unplugging mechanism is very unlikely." Funding for the research was provided by the National Institutes of Health.