Island Time

Extended Simulation on Anton Shows How Cell-Surface Molecules Cluster

Why It’s Important: Virtually every process in human health and disease relies on signals getting across cell membranes—the flexible “bags” that enclose the contents of all our cells. 

The immune system’s response to infection, the spread of cancer cells and the communication between nerve cells that underlies brain function and mental illness are among the many life processes that involve such signals.

Many biochemists believe cholesterol plays a role in cross-membrane signals by limiting how proteins and other molecules in the membrane can move, causing them to clump into “islands.” Other experts have pointed out that we have very little direct evidence for this membrane clustering. Understanding whether this phenomenon actually happens and how it changes signals can give us clues about how to intervene in a way that reverses diseases involving cell-surface signaling.


When you think of cholesterol, most people think of heart attacks, statins, and so on. But it turns out your [healthy] cell membranes are composed mostly of cholesterol, by percentage of molecules. What’s all this cholesterol doing there? —Edward Lyman, University of Delaware

How PSC Helped: Edward Lyman and colleagues at the University of Delaware used the D.E. Shaw Research Anton supercomputer hosted at PSC to see how changing the composition of cholesterol and other fatty molecules, or lipids, in a simulated membrane affects the movement of these molecules. They simulated the membrane for close to 100 microseconds—a very long time in such molecular dynamics simulations, and only possible thanks to Anton’s specialized architecture. In the simulations, adding cholesterol did in fact cause the other lipids in the membrane to sort and form islands that would likely limit protein movement. This result, featured on the cover of the Biophysical Journal, explained previous experimental findings and will help guide new experiments aimed at understanding how clustering affects signal proteins.


One of the most notable accomplishments that scientists have achieved using PSC’s resources was the early-2000s simulation of the aquaporin protein by Klaus Schulten and colleagues at the University of Illinois, using PSC’s LeMieux supercomputer. The group solved a mystery about how aquaporin, a channel in the cell membrane, could speed water across the membrane at a blistering pace without allowing water’s usual consort, the hydrogen ion, through as well. A pirouette at the center of the channel, they found, allowed water molecules to ditch their steady date, giving aquaporin its remarkable specificity. The work made it into the prestigious journal Science and was cited by Peter Agre, the discoverer of aquaporin, in his 2003 Nobel Prize acceptance speech.