GOVERNMENT

By Karla Harby for NCSA -- Alliance scientists have their eyes on cholesterol, all right, but not just for their health. They're set to make discoveries about how cells communicate and take some of the first steps toward nanodevices based on living cells. CHAMPAIGN, IL--Cholesterol. The word itself evokes images of arteries narrowed by atherosclerotic plaque, the gunk that sticks to vessels and impedes the passage of blood the way a kink in a garden hose slows the flow of water. Or perhaps we think of cholesterol as something our favorite foods—ice cream, two-crust pies, steaks, French fries—are rich in, a culinary spoiler that makes us choose the fish yet again. (Hold the Béarnaise, please.) But the drumbeat of public health messages about the hazards of too much dietary cholesterol obscures a more complex reality, namely that cholesterol is essential to animal life. In fact, much of the cholesterol in us is manufactured by our own bodies, not obtained from our foods. It serves as a precursor to the sex hormones estradiol and testosterone and to vitamin D, which is necessary for the formation of bone. Cholesterol is also needed to produce the bile acids that digest fats. Cholesterol is only dangerous when the body's regulation of it goes awry, owing to genetic or environmental causes. For a group of scientists using Alliance resources, the most fascinating aspect of cholesterol is its role in regulating membrane fluidity in animal cells. This group—Eric Jakobsson of the Beckman Institute for Advanced Science and Technology, Urbana, Illinois; H. Larry Scott, chair of biological, chemical, and physical science at the Illinois Institute of Technology in Chicago; and R. Jay Mashl and See-Wing Chiu, also of Beckman, the University of Illinois at Urbana-Champaign, and NCSA—use the prototype 64-processor, Itanium-based Linux cluster at NCSA to perform computational studies that could lead to a deeper understanding of how cells communicate with each other. What's more, their work may contribute to the development of nanodevices based on living cells. Far from being simple walls, cell membranes are instead complicated liquid crystals. About 50 percent of a cell membrane is composed of lipids (oily organic compounds like fats) and most of the rest is made of proteins. The whole lot is characterized by randomness, chaos, and subtlety. "It is amazing," Scott says. "Membranes are not rigid little sheets. They fluctuate a lot, they're dynamical systems, and it really is remarkable." The normal fluidity of cell membranes is now thought essential to life functions, and disruptions can have dire consequences. Just this year scientists reported that cholesterol's regulation of membrane fluidity may be involved in the destruction of brain cells that leads to Alzheimer's disease; that reduced fluidity in the membranes of red blood cells may be related to psoriasis outbreaks; that abnormalities in membrane lipid content may cause resistance to leptin, the hormone that regulates appetite to maintain normal body weight; and that chromium, a carcinogen, causes tumors by reducing membrane fluidity. "Membrane fluidity is probably involved in all cell processes associated with communication of cells with each other and with the outside world," Jacobsson says. Until recently, researchers studying cell membranes had to make do with "toy" systems, little patches of assemblies that everyone hoped would faithfully represent the complete system, but whose chief virtues were that they were simple enough for computation. "It's really just now that [supercomputing] is kicking in," Jakobsson notes. Similarly, while many experimental studies suggest that membrane fluidity and cellular communication are related, up until now they have been handled separately in computational studies owing to limited computing power. "With the new NCSA Linux superclusters Platinum (now up and running) and Titan (coming on line later in the year), we have the exciting prospect of finally having enough computer power to consider these two aspects of membrane biology together," Jakobsson notes. Over the past year, the group's pace of discovery has quickened. Jakobsson attributes this not only to the use of the Linux cluster, but also to Gromacs, the open-source molecular dynamics software. "Gromacs is exceptionally efficient on Linux clusters, as well as being highly flexible for adaptation to particular biomolecular calculations," Jakobsson explains. Scott has also used Monte Carlo techniques, which help ensure the efficient sampling of a system's important configurations, to write additional computer code. This code hones in on states that are important to the study and will eventually occur in nature. These configurations would probably not be accessible to the simulated system over the nanosecond timescale of the molecular dynamics simulations. These tools have enabled the group to sample enough different concentrations of cholesterol in cell membranes to characterize both the individual molecular interactions of cholesterol and lipid as well as the collective effects. For example, the group has shown that membranes consisting of the lipid diphenoylphosphatidylcholine and cholesterol undergo a clear phase transition between a gel and a fluid state at a lipid/cholesterol ratio of about eight. At this and higher concentrations of cholesterol, the lipid is held in a gel-like state by the cholesterol. At lower concentrations of cholesterol, the lipid melts into a fluid state. This observation is significant, because phase change is associated with a significant change in the character of the cell membrane. The high-cholesterol gel phase is much less fluid than the low-cholesterol phase. The cholesterol-induced phase change observed in a membrane patch in the computer shows the mechanism for the experimentally observed cholesterol-induced fluidity changes, presumably those associated with the normal processes of cell fusion and perhaps even the pathology of Alzheimer's disease. "We also think the observed phase change is cool computational surface chemistry," Jakobsson adds. The group also looks at the ion channel proteins embedded in cell membranes. These little living batteries convert the chemical potential of ions into electrical currents. They are essential to intercellular signaling and to transporting material between the inside and the outside of a cell. Jay Mashl notes that if scientists can understand these biomolecules on the atomic level—understand exactly how a given protein structure results in an observed current—then it should be possible to engineer ion channels into the workhorses of nanodevices built partly of biological materials. Jakobsson predicts that, in time, the role of computer simulation will expand from understanding natural cell membranes to designing artificial membrane complexes for nanodevices. In the relatively new field of nanobiotechnology scientists attempt to integrate life forms with nonliving materials on a nanometer (one billionth of a meter) scale. These living machines might accomplish, in a precisely controlled way, the functions, such as signaling and energy transduction, of biological membranes. This possibility brought the Scott-Jakobsson group into collaboration with a group from another scientific discipline altogether, the Beckman Institute Computational Electronics Group that includes Alliance users Karl Hess, Umberto Ravaioli, and Narayan Aluru. Eventually the components of highly sophisticated nanodevices might be embedded into such engineered cell membranes. Such nanodevices would live in the body much like any other cell, taking over functions lost to disease or injury. Although scientists have already demonstrated the feasibility of building such hybrid nanodevices, much more work is needed before such devices can be successfully installed and function in a living organism. Relevant URLs --Access story: http://access.ncsa.uiuc.edu/Stories/cholesterol/ --NCSA Computational Biology Group: http://glycine.ncsa.uiuc.edu/ --UIUC Beckman Institute Computational Electronics Group: http://www.beckman.uiuc.edu/research/compe.html