10,000 proteins fit the width of a single strand of hair—too small for the human eye to even see under a microscope. Proteins are constantly moving, changing shape millions of times each second. To understand the complex world of protein motion, graduate student Logan Ahlstrom and the team working under the direction of Dr. Osamu Miyashita in the University of Arizona’s Department of Chemistry and Biochemistry are turning to supercomputers and their computer simulations to study how and why proteins change shape in response to their biological environment. But the experimental methods commonly used to study protein shape yield just a single snapshot in time rather than revealing the multitudes of shapes proteins assume every nanosecond.
Logan says the main purpose of the research is to learn more about protein shapes and how the shape changes dictate function and the roles these molecules could carry out in our bodies. The team also wants to provide further understanding of abnormal protein shape changes that occur during diseases such as neurodegeneration. Importantly, predictions from their research can inspire new experiments, forming part of a combined effort to solve the puzzle of how proteins work.
Because proteins are so small and move so fast, they are impossible to directly see. But now researchers are able to create simulations of protein motion using the UA Research Computing’s High Performance and High Throughput (HPC/HTC) systems. These simulations provide a closer look at protein motions by producing many snapshots of their theoretical shapes over time. “All proteins are dynamic, so the way they move is very important for their functions. Simulation allows us to study these dynamics in atomic detail. Protein shape is also important for pharmaceutical development when companies design drugs that are meant to interact with proteins,” Ahlstrom says.
But how does Logan convert the proteins to the numbers, which will be used in the supercomputers? He says to think of it as a “connect the dots” game. “When a protein structure is solved, it is really just a collection of points or dots in 3-dimensional space. Each dot represents an atom and we know how the atoms are connected to one another because of the rules of chemistry. Since proteins obey the laws of physics, each atom exerts a force on every other atom,” says Ahlstrom. In the end, the computers are given three important pieces of information. “X, Y, and Z coordinates for each dot, information about how the dots are connected, and some equations describing how the atoms feel one another,” Ahlstrom explains.
Due to the immense number of snapshots that must be obtained, these simulations of protein motion are only possible with the power of the HPC/HTC systems. Still it takes Ahlstrom several weeks to recreate what Mother Nature does in less than the blink of an eye. “For the protein we are currently studying, its shape changes are important for interacting with DNA to regulate when a gene is turned on and off. We needed to see these motions occur many times in order to properly characterize them, and we can only do that by using the HPC/HTC systems. Without those resources, that would not have happened,” Logan says.