Dan Tawfik (left) and Nobuhiko Tokuriki.

For much of Nobuhiko 'Nobu' Tokuriki's tenure as a postdoc in Dan Tawfik's lab at the Weizmann Institute of Science in Rehovot, Israel, he was the only Japanese researcher in residence — earning him visits from the Japanese ambassador whenever he was in town. But Tokuriki did not mind his unique status. “I wanted to see a foreign society and culture,” he says. “And I wanted to study protein evolution.” For that, Tawfik's lab was the place to be.

In nature, proteins evolve new functions rapidly and efficiently. Researchers have long tried to reproduce cellular conditions in the test tube to better understand the process of protein evolution, as well as to search for new or improved enzymes with industrial or medical applications. But such methods, typically referred to as 'directed evolution', have had limited success, resulting in processes that are sluggish compared with those in nature. With Tokuriki's help, Tawfik may have found a way to greatly improve the efficiency of directed evolution by boosting the activities of the cell's quality-control officers — the chaperone proteins.

The method has its origin in work that came to Tawfik's attention about seven years ago. At the time, most scientists thought that the degree to which a mutation improved or degraded a protein's activity was what drove its positive or negative selection.

It's clear there are certain mutations you would never see without this method.

But in 2002, work by Brian Shoichet and his colleagues, then at Northwestern University in Chicago, Illinois, seemed to suggest that many mutations never actually see the light of day. Shoichet's group proposed that certain mutations would gravely affect the way that newly made proteins are folded into three-dimensional shapes, affecting their stability. Thus, the proteins would be quickly whisked away as unstable misfolded 'trash'. And even though some such mutations might be 'adaptive', or of improved function, they would never be detected. Later research indicated that this destabilizing effect of mutation is one of the major hurdles to protein evolution.

After Shoichet's work was published, Tawfik recalls, “I read it again and again. But it took time for its implications to penetrate into the field.” Once the concept sank in, however, it was a short jump for Tawfik to the idea that chaperone proteins, which normally rescue misfolded proteins, might be able to rescue some mutant proteins and accelerate directed protein evolution.

As this idea was taking shape, Tokuriki joined Tawfik's lab, keen to work on the project. “Although it is far from perfect, protein evolution is one of the few areas where we are able to get clear biophysical and biochemical clues as to how and why evolution in the test tube happens,” says Tokuriki. “This is why I chose to study this and not organism evolution.”

Over the course of three-and-a-half years, Tokuriki performed directed evolution experiments on selected enzymes from the bacterium Escherichia coli, both in the presence of large amounts of a chaperone protein called GroEL/GroES and in its absence. This chaperone normally isolates unstable, misfolded proteins in the cell and gives them space to try to refold properly so that they can function normally. The work required each of hundreds of mutant proteins to be characterized over several bacterial generations. “Nobu can do the work of a dozen,” Tawfik says.

The gruelling work revealed that, overall, GroEL/GroES rescued about one-third of the adaptive mutant proteins that, without the chaperone's aid, would have been too unstable to be viable. As a result, it allowed twice as many mutations to accumulate in proteins.

Next, the duo tested the ability of GroEL/GroES to speed up the evolution of a new, divergent function in a specific enzyme. In the chaperone's presence, directed evolution produced at least twice as many adapted variations than in its absence, and these were at least ten times more active and specific than those that evolved without the chaperone (see page 668).

Tawfik says that even though the idea was something “we thought would be stupid not to try”, he was still shocked by how well it worked. Using chaperones in directed evolution will almost certainly be embraced by those trying to produce more powerful enzymes for industrial or therapeutic applications, says Tawfik, “because it's clear there are certain adaptive mutations that you would never see without this method”.