In middle school Science class, there was a choice between doing a science fair project and participating in the Science Olympiad. Both were a big deal and took months to work on, but at least there was a choice. The science fair project was for kids who were always curious about why hot air rises or wanted to know whether bread would grow mold more quickly in the refrigerator or in a shoebox in the basement. The quality of the science wasn't usually a matter of a given child's aptitude but rather whether or not his parents had doctoral degrees.

Science Olympiad was a different story, a competition for kids who were a few years away from majoring in physics or math. Tower Building, The Egg Drop, and Trajectory were three of the main events at the Olympiad, which was usually an invitational competition involving several nearby middle schools. Students, mostly male, who participated in Trajectory event worked to build and calibrate a device which would launch a tennis ball a fixed distance. This wasn't a competition of brawn, one where he who created the largest, most powerful catapult was the shoe-in to Regionals. Instead, the group which built the most faithful machine, one that could repeatedly and accurately send the tennis ball fixed yet previously undisclosed distance on the day of the event, would bring home the blue ribbon.

Thus, Trajectory honed the ability of these young engineers to make predictions about how their home-built instrument would function. It was one thing to build a catapult that could send a tennis ball within the range of 25 and 100 yards, but an infinitely more complex problem when the tennis ball had to land spot-on at the 44^1/2 yard mark. What it meant was that the Trajectory participants would spend full weekends with their creations and a basketful of tennis balls, changing small settings in their instruments and recording -- over and over -- how far the tennis ball would travel, hoping that the process would soon become predictable.

These kids would create their own personal Trajectory databases, and augment them they would, for there was usually a significant relationship between how many times a catapult had been tested and how accurately it functioned on game day. It may sound like a flimsy analogy, or come off as corny, but the endless repetition that it took to win Trajectory is not unlike the foundation of knowledge on which most fields of science are based. Repeating an experiment with slightly varied parameters is what fleshes out a tenuous theory or helps create models with which predictions are often made. Data are logged, then crunched, and the resultant tables and graphs can allow complex questions to be answered with a simple reference.

In the field of molecular biology, there is no more critical and burgeoning a topic to which this principle can be applied than to that of protein structure. It is necessary to understand that the word 'protein' here implies far more than a beef steak or what is inside an egg shell -- we're talking about polypeptides, not a food group. In terms of structure, proteins, which are nothing more than chains of amino acids, serve myriad sub-cellular functions. Sometimes they are shaped like small, flexible cages, like hemoglobin, which is used to bind and carry oxygen in the blood stream. Other proteins are shaped like (and essentially function like) short tunnels or passageways, and can allow atoms to selectively enter or exit a cell. These ones are called ion channels. It may not sound like too great of a phenomenon that different ion channels are structurally more alike than any one ion channel is to hemoglobin, and this small wonder has in fact been proven. In fact, much can already be predicted about a protein based on its similarity to well understood proteins. Yet to understand proteins further, scientists are now focusing on the process of their formation, which is as integral to their final form and function as is their amino acid structure.

Protein formation is probably a straightforward process. It is likely to be governed by a number of quantifiable factors which will most certainly be understood well one day. Currently, however, protein formation is understood only well enough to make educated guesses about how the process is carried out in specific situations. Certain basics are well understood: it is known that proteins form as long, single-file chains of amino acids, and that these chains will crumple, twist and smoosh together to form different functional shapes. The twisting and crumpling may sound random, but as it happens it is not at all. In fact, the final shape of the protein relies entirely on whether the twisting and crumpling happens in an appropriate manner. Ever take the end of a shoelace and twist it, over and over? At first it's just a simple coil, but keep twisting and soon that coil doubles over on itself, and later it starts to conglomerate into a tight knot, right? That is perhaps an oversimplified analogy to what is happening as a protein is forming -- the single strand of amino acids is folding, twisting, and doubling over until a specific final shape is achieved.

Yet the unique part of protein folding that differs from the shoelace is that, within a cell, it can form into that same knot shape again and again without fail. If you tried to untwist your shoelace and re-twist it in to the same knot, it is very likely that a different final shape would result. The fact that it happens over and over in cells in exactly the same way, along with an understanding of chemical interactions and the cellular environment, has helped researchers develop theories and begin to understand the principles of the process.

For instance, it is well understood that temperature and acidity both affect the formation of a protein. These two variables can be readily observed with an ordinary egg. Take the egg -- essentially a big mess of protein -- and put it on a hot frying pan, and the heat turns the egg white opaque. On the molecular level, the individual proteins in the egg are being untwisted and, literally, scrambled. As the chain heats up, it becomes possible to take on new shapes that require too much energy to occur at lower temperatures. The regular shape of the protein is then lost permanently as it untwists and re-twists into an entirely new conformation. Acidity has a similar effect, yet for different reasons. Rather than lending more energy to the amino acids, altering the pH of their surroundings gives them different chemical properties -- ones that are initially neutral develop a positive or negative charge, allowing them to interact with other charged amino acids in a new manner. So when you drop an egg into a jar of vinegar, pickling it, what you're really doing is denaturing the protein, or allowing the amino acids to repel and attract one another in a new manner that wouldn’t occur otherwise. Drop that same egg into a glass of neutral water and nothing of the sort will happen, at least in the short run, because the proteins remain in their natural form.

But, these variables -- temperature and acidity -- are but two in an entire array of forces which can ultimately affect the folding of the protein. Within a cell, the temperature and pH are in fact fairly predictable. Things that aren't predictable are other solutes or molecules that can affect the folding of the protein, including salts and ions that are found in the cytoplasm, as well as other proteins, sugars or lipids which are all part of the mix. In various ways, each of these interacts with the protein as it is being formed from a simple chain of amino acids into its final three dimensional structure. Furthermore, there are 20 some-odd amino acids that link up to form proteins, each which is affected by pH, temperature, and its neighbors in the chain slightly differently. While figuring out how one amino acid relates to its immediate neighbors is simple, determining what happens when a whole chain of them interacts is a veritable quagmire.

What this means is that there is a lot going on between a protein's amino acid structure and its final shape and function, and to predict how the atoms each amino acid will interact with their environment collectively can be a hairy mess. If the variables were fewer, these predictions would be manageable, but the sheer numbers make the problem intractable with current technology.

To figure out the structure of an unknown protein, two main strategies are put to use. The first falls back on the tried and true methods of Science Olympiad. In existence is a large, pooled database of known amino acid sequences and the known, characterized protein structures with which they correlate. When a researcher comes across an amino acid sequence and wants a general idea of what the protein does or looks like, all he or she needs to do is enter that sequence into a computer and everything that is similar to it will be retrieved. From there, relevant publications about those proteins can be read and data regarding how specific the known and unknown proteins correlate can be studied. This is called a blast search, and it is an enormous resource for narrowing down and focusing researchers. As the middle-schoolers can, with enough calibrating, match the settings of their catapults to the distances the tennis ball will launch, molecular biologists can use a blast search to get an idea of what a protein is all about.

But while bioinformatics can shed some light on a protein's structure and function, it can only provide clues—it cannot elucidate the functional form of the protein with any specific information. It is here that computer programs, which calculate the various parameters (temperature, acidity, amino acid sequence...) and churn out information about how a polypeptide will fold into its normal shape. At least, computers can do this in theory. If the problem were as simple as figuring out how a chain of 10 amino acids will bunch up when dropped into some neutral solution, a normal desktop computer would have no problem with the calculation. As the polypeptide grows, and parameters change, the time it takes to calculate the molecular interactions that will affect a protein's final shape increases orders of magnitude -- several orders of magnitude. These calculations can't be tackled by the old iBook -- but they're fair game for supercomputers.

Yet in a tour de force to push the field forward, some scientists at Stanford University have come up with another possibility -- they use individual computers from informed owners across the globe to pitch in and work out small bits of the puzzle at a time. The project is called Folding@home. Any computer that is connected to the internet, or presumably any computer on which this article appears, can be used for the project if its owner signs up. In fact, anyone with a new Playstation 3 that doesn't use it constantly can sign up their machine to participate. With the thousands of computers working on small bits of the project all at once, simulations of protein folding can be understood more quickly. It's like my mother used to say: many hands make light work. How right she was.

« Previous Next »