Research
Introduction
To put it mildly, computers have changed many aspects of our lives. Nowhere is this more true than in the realm of science, where for over 60 years, investigators have been using computers to solve complex problems.
Although computers and science have been intertwined for decades now, their relationship has enjoyed a stimulating growth particularly over the past 10 years. With the advent of fast processors, larger memory storage, and feasible inter-hardware communication, today, even a modest desktop computer can perform many scientific calculations. The user no longer requires a warehouse full of machines for simple scientific tasks. Moreover, parallel programming codes and architecture make it possible for one to link up two or more desktop machines (or simply their processors), for even faster computing. In such a parallel fashion, even a small cluster can perform at the level of yesterday's supercomputer.
The other dramatic change that has occurred over the past 10 years is in the demographic of the computing community. What used to be a body of only "hard" scientists (physicists, engineers, mathematicians) now consists of investigators from many disciplines (even non-traditional sciences, like epidemiology, geography, and finance). Many subjects now reap the benefits of the predictions and modeling afforded by computational science. In due time, every discipline will have a significant computational component to it.
As evidenced by the 1998 Nobel Prize in Chemistry, computational science propels chemical thinking in the 21st century. Around the world, chemists harness various computational techniques to understand phenomena such as protein-ligand binding, macromolecular association, chemical reactivity, and molecular dynamics.
The McCammon Lab here at UCSD is at the forefront of this revolution in chemistry. Using the most advanced computational techniques, we have shed light on a diverse set of chemical topics, such as electrostatic steering, diffusion, ligand docking, free-energy calculations, and molecular dynamics. The following is some of the work that I have done here in the McCammon lab. Thanks for reading, and please feel free to email me with any comments, questions, or requests for prints.
Computational investigation of pressure profiles in lipid bilayers with embedded proteins.
Justin Gullingsrud, Arneh Babakhani, and J. Andrew McCammon. Molecular Simulation 2006, 32, 831838.
The distribution of surface tension within a lipid bilayer, also referred to as the lateral pressure profile, has been the subject of theoretical scrutiny recently due to its potential to radically alter the function of biomedically important membrane proteins. Experimental measurements of the pressure profile are still hard to come by, leaving first-principles all-atom calculations of the profile as an important investigative tool. We describe and validate an efficient implementation of pressure profile calculations in the molecular dynamics package NAMD, capable of distinguishing between internal, bonded and nonbonded contributions as well as those of selected atom groups. The new implementation can also be used in conjunction with Ewald summation for long-range electrostatics, improving the accuracy and reproducibility of the calculated profiles. We then describe results of the calculation of a pressure profile for a simple protein–lipid system consisting of melittin embedded in a DMPC bilayer. While the lateral pressure in the protein–lipid system is nearly the same as that of the bilayer alone, partitioning of the lateral pressure by atom type revealed substantial perturbation of the pressure profile and surface tension in an asymmetric manner.
Peptide insertion, positioning, and stabilization in a membrane: Insight from an all-atom molecular dynamics simulation
Arneh Babakhani, Alemayehu A. Gorfe, Justin Gullingsrud, Judy E. Kim, J. Andrew McCammon. Biopolymers 2007, Accepted.
Peptide insertion, positioning, and stabilization in a model membrane are probed via an all-atom molecular dynamics simulation. One peptide (WL5) is simulated in each leaflet of a solvated dimyristoylglycero-3-phosphate (DMPC) membrane. Within the first 5 ns, the peptides spontaneously insert into the membrane and then stabilize during the remaining 70 ns of simulation time. In both leaflets, the peptides localize to the membrane interface, and this localization is attributed to the formation of peptide-lipid hydrogen bonds. We show that the single tryptophan residue in each peptide contributes significantly to these hydrogen bonds; specifically, the nitrogen heteroatom of the indole ring plays a critical role. The tilt angles of the indole rings relative to the membrane normal in the upper and lower leaflets are approximately 26° and 54°, respectively. The tilt angles of the entire peptide chain are 62° and 74°. The membrane induces conformations of the peptide that are characteristic of -sheets, and the peptide enhances the lipid ordering in the membrane. Finally, the diffusion rate of the peptides in the membrane plane is calculated (based on experimental peptide concentrations) to be approximately 6 Å2/ns, thus suggesting a 500 ns time scale for intermolecular interactions.