ACADEMIA

Nathan Baker’s research group uses Lonestar to explore medicines and membranes: It is impossible to talk about medicine without mentioning the membrane, the lipid bi-layers lining our cells, whose selective permeability keeps the cytoplasmic materials in and the foreign matter out. For Nathan Baker, associate professor of biochemistry at Washington University in St. Louis, the study of membranes presents an enormous opportunity to understand cellular mechanics. “When you think about any biological process, it’s happening in the context of a cell,” Baker explained. “And it’s hard to think about the cell without thinking about its membranes.” The membrane facilitates the transport of materials needed for survival and acts as a door through which drug molecules enter the cell. Yet, despite their significance, bio-membranes are under-explored, Baker believes. “Cells have membranes of all different compositions, but the basic physical chemistry of how they behave is still not very well understood.” To develop an atomic scale understanding of bio-membranes, the Baker group uses Lonestar, the high-performance computing system at the Texas Advanced Computing Center (TACC). Their explorations of bio-membranes interacting with salicylate, sterols, and nanoemulsions bridge the gap from the microscopic level to testable treatments. Collaborating with both laboratory and computational researchers, Baker’s research is helping to develop a basic understanding of membrane mechanics to explore the benefits and dangers of trans-membrane drug dispersal. A close-up look at a single salicylic acid molecule (ball-and-stick representation) interacting with nearby phospholipids. We observe strong hydrogen bonding between the salicylic acid's polar groups and the nearby phospholipid headgroups, shown in green. Understanding Aspirin Salicylate, the chemical compound on which aspirin is based, has been used in medicine for thousands of years. One would think everything about the drug had been discovered long ago, but essential aspects of the compound remain unexplained. Scientists believe salicylate reduces swelling and lowers body temperatures by altering enzyme production within the cell. But the compound can also affect the hairs of the inner ear, causing ringing or temporary hearing loss. Rob Raphael (Rice University) and William Brownell (Baylor College of Medicine), who study this phenomenon experimentally, suspected the side effect had something to do with salicylate’s effect on membranes and called Baker to help explore the system computationally. “Temporary hearing loss is associated with a whole class of drugs,” Baker explained, “and it appears that effect is due to the partitioning of these drugs in and out of membranes. The big question: Is there a molecular reason for the perturbation of the electromechanical coupling?” To get to the bottom of the question, Baker group members, Brett Olsen and Yuhua Song, simulated the actions of virtual salicylate molecules on a model lipid bilayer. They employed molecular dynamics techniques that reproduce the physical behavior of atoms and electrons in the interacting molecules on massive parallel computers, showing how the structure and electrostatics of salicylate can perturb the lipid bilayer. “Molecular dynamics simulations are a classic supercomputing application,” Baker said. “It’s the best way to look at the atomic scale properties of these systems. But you need a lot of simulations, you need a lot of statistics, a lot samples, and that means you need a lot of computational time.” Baker’s allocations on Lonestar have increased dramatically over the last three years. Growing from 30,000 computing hours in 2005 to 600,000 computing hours in 2008, the larger allocation has allowed Baker to add more complexity and realism to his models and produce simulations far faster. “On a single processor, a simulation would take between four and eight months. And we want to run them many times. That’s where the parallelism becomes very important,” Baker said. The research on salicylate-membrane interactions is far from complete. Based on discussions with the Raphael group, it was clear that when salicylate turns into an acid, its impact on the membrane changes. Brett Olsen is now working on new simulations to better model the titration state changes associated with salicylate-membrane interactions. The results of these simulations will be compared with experiments to generate new predictions about the salicylate-membrane interactions. Cholesterol’s Beneficial Cousin Baker's research into oxysterols — important components of many cellular membranes — was another case where experimentalists brought him in to help disentangle a membrane mystery. “I was telling them about our interest in the basic properties of membranes,” Baker recalled, “and they said, ‘You should do something useful. Why don’t you look at the properties of sterols?’” Working with a team of Washington University researchers, the Baker group undertook an investigation of an oxysterol relative of cholesterol with an extra hydroxyl (oxygen and hydrogen) group. Despite their low levels in the body, oxysterols are believed to be important mediators of cholesterol-induced effects. A cross-sectional view of a mixed phospholipid/oxysterol membrane shows us how the sterols fit into the membrane and alter its structure. Oxysterols, shown as ball-and-stick representation, can interact between leaflets using their polar hydroxyl groups, in red. “Some oxysterols have very different effects than cholesterol on lipid bilayers – and therefore the proteins in those bilayers,” Baker said. “We’re using our Lonestar time to understand the molecular basis for those effects, and so far, things look really good.” Through multiple simulations of the sterol-membrane systems, Baker group member Brett Olsen determined that the extra hydroxyl group changes the oxysterol’s orientation, depth and interactions with the lipids around it, inducing very different membrane properties. The results of these simulations are being prepared for an upcoming journal article. Nanoscale Cancer Drugs At Washington University’s Siteman Center of Cancer Nanotechnology Excellence, Baker is collaborating with renowned cancer researchers, Sam Wickline, Greg Lanza and Paul Schlesinger, on a third project simulating membrane interactions. This one explores an intriguing new drug delivery system: nanoemulsions. The tiny droplets — only 250 nanometers long — are composed of lipids, water and perfluorooctylbromide, a non-reactive, FDA-approved substance. The emulsions are remarkable both for their small size, which allows them to go anywhere in the body, and for their rare ability to stream materials directly into cells. Wickline and Lanza quickly realized that by “decorating” the outside of the nanoemulsion with binding agents, dyes and drugs, they could home in on specific targets, and create a new class of tools for imaging and therapy. “They showed that they could put imaging agents or therapeutic drugs on the nanoemulsions,” Baker said. “That’s the medical relevance and it’s why everyone is so excited about it from a cancer perspective. But there’s not a good understanding of how these nanoemulsions interact with cells once they get there.” Baker group member, Sunjoo Lee, is performing simulations to explore the atomic-level picture of how medicines embedded on nanoemulsions end up on a cellular membrane and eventually inside the cell. Such computational research enables testing and analysis on a scale not possible in the “wet lab” alone. Together with bench scientists, the group is piecing together a comprehensive theory of how nanoemulsions behave in relation to membranes. Understanding the molecular mechanisms of a medical platform such as nanoemulsions helps researchers improve design and anticipate potential side effects. “If the whole system is behaving as a black box, the best you can do is move some parameters around and hope for the best,” Baker explained. “But when you start to understand the molecular details of the system, you have a better chance of heading in the right direction.” Supercomputing a Cure Computer simulations that explore the underlying atomic basis of drug-molecule interactions represent only one of many ways that high-performance computing resources are driving novel medical treatments. Gene coding and therapy, advanced tumor diagnosis, drug modeling, and even computer-assisted surgery, all use supercomputers as a basic resource for understanding, treating and developing cures for illnesses. “With an exponential growth in data, and researchers seeking to understand biological phenomena in their full complexity, high performance computing has become a critical tool for life scientists,” said Michael Gonzales, TACC’s Life Sciences Program Director. “TACC’s expertise and world-class supercomputing facilities enable scientists to tackle even the most difficult computational biology problems.” For Baker, working with TACC has been a boon. “Without TACC, I don’t think we could’ve gotten it done,” Baker acknowledged. “The reason we’ve been so committed to TACC and continue to request entire allocations there is the infrastructure is perfect for what we’re doing and the support is great. When we have problems, they get answered.” The road is long but the rewards are great. New drugs are being developed faster and with greater understanding, thanks, in part, to the tireless work of collaborative scientists and high-performance computers. “Translational research, or the ‘bench-to-bedside paradigm’, doesn’t happen in one leap,” Baker reiterated. “It happens through collaborations and many important steps – starting with basic research – that address the problem at all scales.” ************************************************************************** Visit the Baker Group home page for a complete list of related publications. Aaron Dubrow Texas Advanced Computing Center Science and Technology Writer