Online gamers obsessed with puzzles have now topped what scientists had failed to do by themselves, which was to improve the model of an enzyme. This has led to the first crowdsourced redesign of a protein, thanks to the online game Foldit, which was developed to allow players to fiddle at folding proteins at home, to search for the lowest-energy configurations.
Foldit was developed by teams led by Zoran Popovic, director at the Center for Game Science, and biochemist David Baker, of the University of Washington in Seattle. Researchers had previously reported successes with protein folding, but the latest work moves into the realm of protein design, which is a more open-ended problem. The challenge was issued by posing a series of puzzles to Foldit players, and then testing these variations in the lab. Researchers have created an enzyme with an 18-fold higher activity than the original.
Justin Siegel, a post-doctoral researcher in biophysics in Baker’s group, worked for two years trying to make these enzymes better but couldn’t achieve it. Foldit players were able to make a large jump in structural space, and researchers still don’t fully understand how they have accomplished this.
The enzyme in question catalyzes a reaction in synthetic chemistry called Diels-Alder, which are used throughout industry to synthesize everything from drugs to pesticides. Researchers got back nearly 70,000 designs for the remodeling of four amino-acid loops and 100,000 for a design that would stabilize it. It’s out of these that the scientists ultimately solved the riddle.
The scaling involved was unprecedented, as players used a variety of different intuitions to fold the proteins. While there are no immediate applications for the enzyme that catalyzes this particular Diels-Alder reaction, Baker is now looking for more useful targets. Foldit players are working on inhibitors that bind and block the 1918 pandemic influenza virus.