Biopunks help solve structure of key viral enzyme

I've been meaning to blog about the program FoldIt for some time.  It's a program out of David Baker's lab (link to University of Washington) where users can manipulate protein structures to improve folding but in a unique twist, the improvement is 'scored', just like a game.  A problem is posed on the website and thousands of players compete to see who can find the best solution. They also have regular competitions to see how well players can predict the structure of a protein from just the amino acid sequence (the Holy Grail of the protein folding world).  I've played on and off for a couple of years and it's pretty fun... but it's a lot like my day job, so if I take the time to play a video game it is usually something like Rock Band.  Still, FoldIt is a pretty easy game to play (the number of rules is limited and the GUI is very intuitive) so the real challenge is in the player's ability to use logic and their skill in 3D visualization. I'll have a more detailed post about it and its less-familiar cousin EteRNA (the RNA equivalent of FoldIt) later. When you are ready to play, click here (it's free, easy, and fun... really!)

This week in the journal Nature Structural and Molecular Biology (I have had several rejection letters from this fine journal) is an article (abstract) describing the use of FoldIt to solve a scientific problem at least ten years old!  Crystallographers have been trying to determine the structure of the monomeric form of MPV protease. MPV stands for Mason Pfizer Virus, a virus that causes an AIDS-like syndrome in monkeys.  Although it is not HIV per se, the protease is a key enzyme in the production of a mature virus and so developing drugs using this structure could be useful in developing an HIV therapy.  This protein was one of those rare cases where good crystals could be grown but interpreting the x-ray data back into the final 3D picture of the protein could not be done. So, researchers looked to crowd-sourcing as a means to solve this problem.  For three weeks, FoldIt players had the chance to optimize a 3D model, based on an NMR structure of the monomer.  Teams competed to see who could come up with the best solution (based on how well the different structural pieces of the protein fit together).  After all that tweaking and optimizing, more than a million different models were created.  Crystallographers used the best of these models as a starting point and one of them successfully generated a solution to the x-ray data using molecular replacement.

Who were the people who provided the key to success?  The top groups were listed as co-authors by their team name (FoldIt Contenders Group and FoldIt Void Crushers Group) and by a few handles listed in the article (spvincent, grabhorn, mimi) so not much is known about these folks, but I'll bet many of them had no knowledge of crystallography or biochemistry and probably had no clue what this enzyme did.  However, the fact that thousands of players worked together on this problem and were ultimately successful highlights the power of crowd-sourcing and that top-flight science can be accomplished through a "game".  I think this paper will also motivate others, both players and scientists, to leverage the power of biopunks!