Winners. Nobel chemistry computer wizards Arieh Warshel, Martin Karplus, and Michael Levitt (left to right).

USC, Martin Karplus, Stanford University

Computer Modelers Earn Chemistry Prize

How does that work? The question is the starting point for anyone who tinkers with an engine, a watch, or in the case of this year’s winners of the Nobel Prize in chemistry, the proteins inside our bodies. Today, three U.S.-based chemists shared this year’s prize for developing computer models that reveal how proteins and other compounds undergo chemical reactions.

All three of this year’s chemistry laureates are naturalized U.S. citizens. Martin Karplus of Harvard University and the University of Strasbourg in France was born in 1930 in Vienna and moved to the United States just before the outbreak of World War II. Michael Levitt of Stanford University’s School of Medicine in Palo Alto, California, was born in Pretoria in 1947, and today is a British, U.S., and Israeli citizen. And Arieh Warshel of the University of Southern California in Los Angeles was born in 1940 in Kibbutz Sde-Nahum, Israel, and still holds Israeli citizenship.

The three pioneered new tools for studying chemistry in motion. By the 1970s, researchers had a good idea of what many molecules, including large proteins, looked like. Tools such as x-ray crystallography and nuclear magnetic resonance spectroscopy provided images of molecules in atomic detail. But those pictures were static. They didn’t reveal the intricate dance of electrons and atoms involved in the making and breaking of chemical bonds.

To explore how reactions take place, researchers in the middle of the 20th century built simulations of molecular motion. The simulations took mathematical models for the forces between atoms and used them to calculate how molecules hold together, move, and react. Later, chemists incorporated the models into computer codes that could run more complex simulations. By the late 1960s, there were two such approaches. One used equations of classical Newtonian physics to model the motions of atoms and bonds in molecules like balls connected by springs. Because this approach was mathematically tractable for large numbers of atoms, it enabled researchers to simulate proteins and other large molecules. In 1969, Levitt and Warshel, then both at the Weizmann Institute of Science in Rehovot, Israel, designed just such a program that could track the movements of proteins and other large biomolecules. But it couldn’t calculate the changes in energy involved when chemicals react and form new molecules.

Meanwhile, at Harvard, Karplus was deeply enmeshed in the second approach to simulation, called quantum chemistry. This approach was far better at simulating the motion of electrons and atomic nuclei involved in chemical reactions. But it was so computationally demanding that it was useful only in solving the behavior of small molecules.

The effort to bring the quantum and classical worldviews together “was a fairly natural progression,” Levitt says. In 1970, Warshel visited Karplus’s lab and brought his classical program with him. The two soon constructed a program that welded their approaches, treating mobile electrons—called pi electrons—with a quantum chemical treatment and atomic nuclei with a classical approach. They then used this to calculate the behavior of linear organic molecules. Although the molecules were far smaller than proteins, the simulation marked the first successful construction of a hybrid model.

Karplus and Warshel’s program worked only for flat “planar” molecules. But in 1976, Warshel and Levitt followed it up with a more general approach, and showed that it worked for simulating the behavior of the protein lysozyme, an antibacterial compound that was the first enzyme to have its structure solved by x-ray diffraction. “The prize recognizes developments that started over 40 years ago that still reverberate today through much of chemistry and biology,” says Klaus Schulten, a computer modeling expert at the University of Illinois, Urbana-Champaign.

Today, those reverberations include related hybrid “multiscale models” capable of simulating more than 4 million atoms, which researchers are using to reveal the complex chemistry involved in everything from the translation of genes into proteins to the conversion of sunlight into chemical fuel that is at the heart of photosynthesis. Such processes are so complex, Schulten and others say, that they are impossible to study without computers. Now, thousands of researchers are hard at work in the field. Still, Schulten says, the Nobel Committee got it right: “If you had to single somebody out, I think there’s a good case to be made that these three were the right choice.”

Even so, Warshel says that he was wary when he reached for the ringing phone this morning at 2 a.m. “I checked to see if they talked in a Swedish accent just to be sure.”