Innovative Techniques Devised for Characterizing Proteins

Utilizing a combination of high-powered computers and sophisticated research magnetic resonance (MR) data, a U.S. biophysical chemist has developed techniques that enhance the manner in which scientists can examine and predict the structure and dynamics of proteins found in the human body. These developments could eventually shorten the time it takes researchers to develop new, more effective drugs and better understand biomedical processes that underlie a host of health disorders.

The new techniques "allow us to more accurately understand protein behavior and function at all levels, how enzymes work, and how to develop drugs that bind to certain proteins,” said Dr. Rafael Brüschweiler, a professor in Florida State University's (FSU; Tallahassee, USA) department of chemistry and biochemistry and associate director for biophysics at the U.S. National High Magnetic Field Laboratory at the FSU.

Given that there are hundreds of thousands of different proteins found in the human body, advances such as Dr. Brüschweiler's that can streamline their analysis and understanding are viewed as most desirable in the scientific community.

Over the past several years, Dr. Brüschweiler and his colleagues have incorporated a pair of complementary but powerful tools, both of which provide detailed data about the structure and dynamics of proteins at the atomic level. Nuclear magnetic resonance (NMR) data are first collected for a specific protein that is being analyzed. NMR is a research tool that utilizes high magnetic fields to measure the strengths, directions, and temporary fluctuations of magnetic interactions between the atoms in a protein fragment.

Subsequently, in a technique Dr. Brüschweiler has developed, high-powered computers are used to confirm the NMR data in terms of their realistic representation of protein structure and dynamics, as well as to make additional predictions of those characteristics.

The computational results significantly rely on the shape of the protein's "energy landscape,” the conformational space available to that protein under physiologic conditions. However, due to its intricacy, improving characterizations of the energy landscape is a difficult and time-consuming undertaking. In fact, until recently, a computer simulation of a single protein that represented just a microsecond took several months. Now, with the aid of the powerful computer array at Florida State's High Performance Computing Center, it takes Dr. Brüschweiler and his group only a fraction of the time it once did.

Working with a postdoctoral associate, Da-Wei Li, Brüschweiler has found a highly effective way to use directly the NMR information for improving the protein potential. The basic idea is to "recycle” an existing simulation of an intact protein, using methods taken from statistical physics, for many trial potentials until the one is found that yields the best agreement with experiment. This leads to an increase in speed by a factor of 100,000 or more over previous techniques. The approach is not only efficient but also permits the improvement of the protein potential directly on intact proteins, instead of on small fragments, as was previously the case.

"This has opened up a new way of becoming increasingly quantitative in our computations, which is key in developing a predictive understanding of the functions of proteins,” Dr. Brüschweiler said.

An article describing the research was published online August 16, 2010, in the journal Angewandte Chemie. "This is the culmination of a number of years of research on our part, so obviously we're excited about the progress we have made,” Dr. Brüschweiler said. "While this is fairly basic research designed to develop a greater understanding of life at a molecular level, it opens up a range of possibilities for future advances by scientists all over the world.”

Related Links:
Florida State University