As any marksman will confess, the first step to hitting a target is seeing it. Similarly, scientists who aim to find drugs that target disease-related proteins develop sophisticated computing methods to set their sights and improve their odds of a direct hit.

Nagarajan Vaidehi, second from left, and her team predicted the structures of key proteins targeted for new drug development. (Photo by Allen Mao)

Researchers led by Nagarajan Vaidehi, Ph.D., professor in the Department of Immunology, recently gained high marks in an international contest designed to test their methods of predicting the three-dimensional shape of proteins and where drugs might bind to them.

The contest, called GPCR Dock 2010, sponsored by The Scripps Research Institute in La Jolla, Calif., invites research teams from around the world to predict the structures of key proteins that affect brain function and immune response and that play other important roles in the body. The proteins belong to a large class called G-protein coupled receptors, or GPCRs.

“These GPCRs are primary targets for new drugs to treat cancer, diabetes and cardiovascular disease,” said Vaidehi.

GPCRs and other proteins are dynamic molecules, constantly changing shape in the fluid environment of the cell. Vaidehi’s team is developing advanced, computer-based mathematical methods and software that can predict the shape of GPCRs. They also aim to foresee where small drug-like molecules will bind to GPCRs. The methods allow researchers to screen potential drugs to see if they might effectively target a disease-related protein.

For the contest, researchers at the Scripps Institute determined the structures of two GPCRs, called CXCR4 and dopamine receptor D3, using an established method called X-ray crystallography. Without revealing the proteins’ shapes, the Scripps scientists next asked contestants to submit their predictions of what the proteins’ structures would be with drug-like molecules bound to them.

“GPCR Dock asked us to submit our best five predicted models for each of the proteins they crystallized and rank these five models from what we thought was our best fit to the one we liked least,” Vaidehi said.

Out of more than 32 groups aiming to predict CXCR4’s structure, the City of Hope team placed second. For the dopamine receptor, they placed third out of 25 groups.

Goutham Balaraman helped with protein modelling. (Photo by Allen Mao)

Equally important, they correctly ranked their predictions for both proteins, a feat no other group matched.

Overall, the City of Hope team was the only group with predicted structures landing in the top 10 for all proteins.

The team learned of their results during a seminar presented by Scripps faculty.

“We were sitting at the edge of our seats waiting for the structures to be shown. It was extremely joyous when we saw the structure that looked like our model, and we had to restrain ourselves from shouting out in the seminar hall,” Vaidehi said.

The results showed not only that the team can predict protein shape well, but also that they can rank models with confidence, which makes their software useful for pharmaceutical industry drug design, Vaidehi said.

“Drug discovery normally is very expensive and time-consuming, so lowering costs and speeding the process ultimately will benefit patients,” she added.

Researchers have linked CXCR4 to the growth and spread of more than 20 types of human cancer, and dopamine receptor D3 is implicated in Parkinson’s disease and schizophrenia.

Using its models, Vaidehi’s team currently collaborates with Joseph Kim, M.D., associate professor of surgery, and Richard Yip, Ph.D., director of City of Hope’s High Throughput Screening Core facility, to target CXCR4 and similar receptors for pancreatic cancer. They also work with John Shively, Ph.D., professor and chair of immunology, to design a positron emission tomography imaging agent for early diagnosis of the disease.

The City of Hope GPCR Dock 2010 team included Alfonso Lam, Ph.D., Supriyo Bhattacharya, Ph.D., graduate student Hubert Li, Goutham Balaraman, Ph.D., and Michiel Niesen, M.S.