Patients of the future may one day be grateful that Patricia Babbitt abhors hot weather. Otherwise this UCSF biopharmaceutical scientist might have chosen a career toiling away at archaeological sites instead of digging through databases seeking the secrets - not of the Pyramids - but of proteins. It's a trade off Babbitt accepts cheerily. "In a way I'm still the archaeologist I wanted to be. I search through evidence about the evolutionary past and make predictions." But unlike her peers in archaeology, Babbitt and her colleagues in the School of Pharmacy track how protein structure determines protein function and then use this information to explain, extrapolate, experiment with and, sometimes engineer entirely new versions of these life-giving treasures.

It has long been known, of course, that proteins - assembled at the instructions of genes - play an essential role in constructing diverse life forms (think: skin and bone) and then maintaining and managing the life form's internal environment (think: immune system). Humans harbor approximately 30,000 genes that may encode 70,000 to 100,000 proteins, some small, some enormous, and all distinctively shaped. For proteins, which consist of interlinked and folded chains of amino acids, shape determines function. And no matter their shape, each protein has a binding site, a place where other proteins or substances can attach, triggering a reaction that influences other reactions. The resulting cascade constitutes a task, such as building muscle tissue or regulating genes, or contributes to a problem, such as cancer.

It is this "binding" relationship that interests Babbitt. "What are the underlying design principles that determine how structures deliver function?" she asks. To find out,she needs to make comparisons. And in that task she has at her disposal a growing army of newly discovered proteins. Unfortunately, most of these "new" proteins are of unknown function and unknown gene parentage, although researchers like Babbitt can surmise certain attributes by comparing them to others now "housed and evaluated" in enormous - and expanding - databases devoted to protein sequences as well as to three-dimensional protein

structures. What this and related research has demonstrated is evolution's conservative attitude. It seems that if a protein scaffold works well in one case, it may be reused. As Babbitt explains, "There are perhaps about a thousand scaffolds or templates from which hundreds of thousands of proteins [throughout nature] have been created."

Within these scaffolds are recognizable patterns, or similar protein sequences called forth by the parent genes. Patterns can be detected by computers and grouped accordingly, one of the many sorting and relationship tasks encompassed by the term "bioinformatics." It is these groupings that constitute a protein's family tree. Unfortunately, when you know one, you don't know them all. True, there is a syllogism among protein superfamilies, e.g., if known protein A interacts with known protein B, or regulates gene X, and unknown protein C resembles protein A, unknown protein C might interact with B or regulate X in a similar way. But finding commonalities at great distances - like comparing your facial characteristics with those of a first millennium ancestor - is difficult.

Tougher still is pinpointing details that enable Babbitt and her colleagues to separate the mere look-alikes from the close relatives. Babbitt's way around the problem has been to narrow the field by concentrating on enzymes - those cut-and-paste proteins that speed up chemical reactions - and their particular scaffolds. Her goal, as always, is to find how the structure of related enzymes actually translates into function. "Related enzymes often will have a partial reaction in common. We then go back into their structure and look for similarities and dissimilarities around their active sites."

This "look" is literal, an on-the-screen investigation of computer-generated enzymatic structure - as well as experimental, a "let's see if this look-alike enzyme acts the same way" approach. The computational and experimental techniques support each other, since what is inferred in cyberspace can be tested in real space, providing a feedback loop that refines the computer analysis.

"We've found that as long as we use the same or very similar structures of two related enzymes to do the same thing [react the same way chemically], they can be interchangeable," says Babbitt. This is the root of the superfamily tree. Move farther "up the trunk," however, and the function starts to change. It is at these critical change points where Babbitt and her colleagues tag and study the structural elements, trying to discern the precise mechanism that stamps the protein as different. The evidence gathered so far demonstrates that even very close relatives - with only slight structural differences - can act very differently. "We're using this information to create a new set of definitions for enzyme chemistry and developing a new database in collaboration with the UCSF Research Resource in Biocomputing, Visualization and Informatics."

To the uninitiated, Babbitt's work may seem more of an esoteric insight into protein evolution than a significant factor in maintaining human health. But all breakthroughs have a context and a history. And in spreading the wealth of her data to other researchers, Babbitt is speeding the time when inference of protein function can help to find a protein's drug target - the first step in any successful new therapy. Moreover, by literally shuffling around segments of protein structure (based upon established inferences), Babbitt and her colleagues in the private sector are developing entirely new proteins with important social benefits - such as biodegrading toxic pollutants. At the same time, researchers seeking new antibiotics are now armed with the knowledge that small structural changes can produce significantly different chemical reactions - knowledge that may be used to create new drugs against antibiotic-resistant strains of bacteria.

For Babbitt, a former UCSF graduate student who fashioned a career at her alma mater, moving to the new campus at Mission Bay is a welcome change, particularly now that the actual relocation is over. "I like the idea of being closer to my colleagues." The synergy, she believes, will help to advance her work in what has suddenly become a very hot field. "What we do may not seem that practical, but it is essential to practical applications." And no matter how hot the work becomes, Babbitt will never have to worry about a sunburn.