This is cool:
As many of you already know, a protein causing AIDS in rhesus monkeys that hadn’t been solved for 15 years was resolved by Foldit players and confirmed
by x-ray crystallography. That paper was selected for an Advance Online Publication (AOP) today in Nature Structural & Molecular Biology; link to
Viruses instruct the cells that they infect to make the building blocks needed to make more viruses. Retroviruses, such as HIV and its close cousin, Mason-Pfizer monkey virus (M-PMV), give these instructions in such a way that a series of different pieces are stuck together and must be cut apart before they are used to build a virus. The virus needs a metaphorical pair of scissors, a protein called a protease, to do the cutting. Otherwise, it is stuck with the unusable bundles of virus parts and cannot replicate itself.
Obviously, knowing how this pair of scissors works would be very helpful in designing drugs to slow down a retroviral infection. For the past fifteen years, however, we haven’t even known what the M-PMV retroviral protease looks like, let alone how it works. After years of research, the final breakthrough came from online gamers playing the puzzle game Foldit.
A protein is a string of molecues (amino acids). Each amino acid is characterized by its side chain, which extends from the string. Each side chain has a certain size and chemical property, and can be hydrophilic (water-loving, like salt) or hydrophobic (water-fearing, like oil). The string spontaneously folds itself into a certain shape, which turns it into the ultimate micromachine. It can transport things (like hemoglobin, which carries oxygen in the blood), repair stuff, act as scaffolding, make chemistry happen, and all the other gazillion little things needed to keep life going.
In Biochemistry, we study how these micromachines work, but in order to know how they work, we first need to know how they are shaped, which means we need to know how they fold. Theoretically, all the information necessary for the protein to fold into the proper shape is right there in the order in which the amino acids are strung together. Given the proper temperature, pH, and salt concentration, proteins with the same amino acid sequence will generally fold into the same shape every time, whether it is in a beaker or in a cell. Problem is, we can’t predict how a protein will fold from its amino acid sequence. This is known as the Protein Folding Problem.
So, here’s the puzzle: take a string of amino acids, and fold it as compact as it will go, with all the water-fearing side chains on the inside and all the water-loving side chains on the outside
[Esoteric, possibly unnecessary scientific stuff that will either fascinate you or make your eyes roll back in your head in incomprehension and boredom. Skip it if you want, but I think it is way cool and it’s my blog :P]
In spite of the protein folding problem, scientists have several tools to point them in the right direction. Nuclear magnetic resonance provide parameters on where certain atoms within the molecule must be relative to each other. Proteins can be compared to other proteins with similar amino-acid sequences. Computer programs can predict the likelihood of certain shapes within a certain amino acid sequence. Scientists can even make an educated guess about the shape of a protein, plug the guess into a computer, and calculate how stable that shape would be. Scientists can also divide the protein into segments, determine the structure of each segment, then put the segments back together. Finally, x-ray crystallography provides experimental data to confirm and refine correct guesses.
[/End esoteric science stuff]
All of these techniques eliminate most of the nearly infinite number of conformations a protein can take (1 with 17 zeros behind it is a conservative estimate of how many shapes even a small protein can make). However, these techniques often do not provide enough information to build a complete picture of how a protein is folded.
Enter Foldit. Players are given puzzles, which consist of computer models of partially folded proteins (based on similar proteins, NMR data, or segments within the protein in which the structure is solved) and are told to adjust the structure based on the parameters above. The most stable structure (according to computer calculation) wins, and also becomes a candidate for comparison with NMR and x-ray crystallography results to confirm if it is correct. Players can even collaborate and improve on each others structures.
The Foldit project accomplishes three things. First, it provides a new tool for solving protein structures. This project is further evidence that the best thinking will be done by humans in collaboration with computers.
Second, it provides data to tease out rules of protein folding. These rules may be accessible to the intuition of a human puzzle-solver, but not to current computers. By watching for patterns in how these puzzles are solved, scientists can improve computer programs used to predict protein structure.
Third, it expands the community of science. Science always has been and always will be a community activity. Stereotypes of the maverick/ loner/ crazy genius aside, us scientists need other scientists. We depend upon each other for verification of our findings and build on each other’s work. Almost never can a discovery be attributed to a single person, especially in the modern age. Crowd sourcing science invites the public to take part in the discovery. You, yes, you, can do science without jumping through the academic hoops. That is really, really cool.
PS: If Astronomy, rather than Biochemistry, is your cup of tea, checkout this other science crowdsourcing project, Galaxy Zoo.