Medicine's newest frontier
By S. A. Unger and Ellen Goldbaum
Illustration by Frank Miller
Structural biology image courtesy of HWI Graphics Department
Last year about this time, the world was abuzz with news that the much-vaunted Human Genome Project was coming to a climax as rival international groups, both private and government-sponsored, raced toward the finish line in what is arguably one of the most important scientific accomplishments of the 20th century.
But what does it mean now that we know the order of the chemical letters (nucleotides) that make up the gene and where each gene lies on a chromosome? What does this sequencing and mapping of genes—as these two processes are referred to, respectively—tell us? More importantly, how can they help scientists develop advanced medical treatments? Much like when a sculptor unveils his prize work of art only to be met with a quizzical “What is it?” scientists are now busy explaining to a highly expectant public why the Human Genome Project is indeed a masterpiece.
In essence, what they are telling us is that although the Human Genome Project was a race, it was, more specifically, a relay race—and only the first leg of it has been run.
The baton—the decoded human genome—is today being handed off to a new set of sprinters who will take it the second leg, bringing scientists just one step closer to what most nonscientists consider the real finish line: the translation of basic science into clinical science, where it has practical applications for patients.
This second leg of the race is being referred to as the postgenomic era because scientists running it are no longer focused on mapping and sequencing genes; instead they have turned their attention to the substances genes produce: proteins.
When a gene is activated-or, as geneticists would say, expressed-the sequence of chemical letters (A, C, G and T) in its DNA serves as a blueprint for the building of a protein. It is these proteins that in some shape or form orchestrate the workings of all the cells in our body, making them extremely pertinent to anyone hoping to find better ways of treating diseases. (Experts estimate that humans have some 30,000 genes that are responsible for churning out thousands of different proteins.) The challenge is that, in most cases, the functions of these genes and their proteins are mysteries.
Toward the goal of unlocking the secrets of proteins, a scientific field called proteomics has sprung into being with the aim of discovering the structures and interactions of all proteins in a given cell. By studying the proteomic landscape of healthy and diseased cells, researchers may better understand the complex ways in which cells communicate at a molecular level (referred to as cell signaling), which in turn can lead to a better understanding of our body's metabolic pathways.
Once these mysteries are solved, pharmaceutical companies can collaborate with basic scientists to develop more sophisticated diagnostic devices, as well as new drug targets. In this new world of pharmacogenetics and pharmacodynamics, scientists will attempt to identify which misguided protein needs targeting and design drugs that bind to it in order to turn it on or off-a form of treatment that, because of its specificity, is predicted to produce few, if any, side effects. This approach is called rational drug design because it poses a contrast to the hit-and-miss experimental approach scientists have generally had to rely on in the past when developing new drugs.
Although scientists do not know the details surrounding how proteins function, they do understand that protein function is often tied to a protein's shape; in other words, the way in which a protein interacts with other molecules is in many cases determined by its three-dimensional architecture.
In this postgenomic era, therefore, much attention and tremendous resources are being marshaled toward the goal of modeling the three-dimensional structure of as many proteins as possible. More specifically, scientists who have expertise in structural biology are seeking to devise ways to better understand, and even predict, how a gene's chemical letters direct amino acids-the components of proteins-to assemble, or fold, in order to create a protein's three-dimensional shape. Based on current knowledge, scientists estimate that there are several thousand different classes of protein folds.
The quest to sequence the amino acids and model the shapes of proteins promises to be as thrilling a race as was the one to decode the human genome.
Over the past few decades, scientists have devised a number of painstaking techniques to determine the shape of a protein. One of these techniques is X-ray crystallography, which involves bombarding single crystals of protein molecules with X-ray beams. The beams are diffracted by the atoms of the protein molecule, thereby generating a diffraction pattern. This pattern is then analyzed by computer to define the protein's molecular structure. A second technique involves use of nuclear magnetic resonance spectroscopy, which uses powerful magnets to determine the chemical shifts of nearly all atoms of the protein. Knowledge of these shifts enables scientists to measure distances between protons, from which protein structure can be calculated.
Like the Human Genome Project, the success of proteomics will depend heavily on the ability of scientists to automate and greatly accelerate these and other protein-modeling techniques-as well as protein-sequencing techniques-in order to process the phenomenal amount of information embedded in all the proteins in a given species, or its proteome. Integral to this automation effort will be the tandem development of new generations of supercomputers and robotics systems.
Underpinning this entire postgenomic quest is the burgeoning field of bioinformatics, in which computer scientists and biologists work side by side toward the goal of harvesting information produced by the genome and proteome projects in order to provide medical researchers with the knowledge they will need to move forward with discoveries that will touch all our lives.
It was with an understandable sense of pride and accomplishment, therefore, that the University at Buffalo learned New York State Governor George Pataki had proposed (in his January 2001"State of the State" address) that Buffalo be designated as the site for a world-class Center of Excellence in Bioinformatics.
In what Bruce Holm, senior associate dean in UB's School of Medicine and Biomedical Sciences, describes as "one of the biggest announcements that has been made at this university in at least a decade," the governor detailed plans for the center. Pataki says this center would be one of three established in the state as part of an ambitious $1 billion high-technology and biotechnology plan aimed at positioning New York State as a worldwide leader in university-based research, job creation and job development.
The other two centers-one in Albany for nanoelectronics, the other in Rochester for photonics and optoelectronics-would, along with the Buffalo Center of Excellence in Bioinformatics, serve to link university researchers directly with private industry as part of what Pataki refers to as "the largest high-tech economic development initiative in our state's history."
Making Buffalo a dedicated research site for bioinformatics is a natural progression of the pioneering work that the proposed center's three partner institutions-University at Buffalo, Roswell Park Cancer Institute (RPCI), and Hauptman-Woodward Medical Research Institute (HWI)-have been doing for years.
Holm says Buffalo is uniquely positioned for a Center of Excellence in Bioinformatics because the scientific areas in which it excels will be critical ingredients in the bioinformatics revolution. Holm also points out that the majority of the DNA used in the Human Genome Project came from volunteers in Western New York, courtesy of RPCI's DNA libraries.
These ingredients, as outlined by Holm and others who will be involved in the center, include the following:
Supercomputing. UB is home to one of the world's leading academic supercomputing centers-the Center for Computational Research (CCR)-which has the large-scale computing and visualization capabilities and the staff expertise necessary for tackling the massive computational problems presented by the data in the human genome. CCR already serves as the computational backbone for research under a $25 million National Institutes of Health (NIH) grant in structural genomics awarded to a consortium of nine institutions, including UB and HWI.
"CCR provides the computing power that is essential for any successful bioinformatics initiative," says Holm. "Without it, we wouldn't even be in the game."
"Proteomics and structural biology require massive supercomputing capabilities," says Jaylan Turkkan, UB vice president for research. "There are tens of thousands of different ways proteins can fold, so only with high-end computing are scientists going to be able to model and predict what those folds are going to look like."
Structural Biology, Genomics and Proteomics. Over the past five years, more than $17 million in key scientific awards has been received by UB's School of Medicine and Biomedical Sciences, RPCI and HWI to study genomics, proteomics, structural biology and neuroimaging as they pertain to disease modeling and drug discovery.
Generation of DNA Microarrays. The DNA microarray facility, jointly operated by UB and RPCI, allows scientists to detect thousands of genes simultaneously and analyze their expression. By creating custom gene "chips," each of which can contain thousands of genes, the facility is a boon to researchers investigating which of the 30,000 human genes are active in a given cell or tissue.
Pharmaceutical Science. Pioneering work in UB's School of Pharmacy and Pharmaceutical Sciences led to the development of the field of pharmacodynamics. Continued work by the same researchers has resulted in new techniques to find markers of pharmacological effect that can be used to optimize new drugs and therapies.
At the same time, HWI and UB are home to the developers of SnB, the molecular structure-determination software based on the algorithm developed by Nobel laureate Herbert Hauptman, UB distinguished professor and president of HWI, and the Shake-and-Bake algorithm developed by George DeTitta, professor and chair of the Department of Structural Biology in the medical school and executive director of HWI; Charles Weeks, senior research scientist at HWI; and Russ Miller, director of CCR and professor of computer science and engineering. SnB is the structure-determination software of choice in more than 500 laboratories across the United States.
Visualization. Also at UB is the New York State Center for Engineering Design and Industrial Innovation (NYSCEDII), which has the virtual-reality capabilities to allow scientists to visualize and interact with three-dimensional molecular structures in large, immersive environments.
The current setup at NYSCEDII allows scientists to interact with biological data through visual data-mining techniques. Its eventual expansion to a fully developed, six-wall (walls, floor and ceiling) "cave"-a capability now available at only one other institution in the nation-will amplify the amount of information that can be displayed by a factor of six, at the very least. It will be possible to "walk through" a compound, evaluate the fit of a series of drug leads to a particular target protein and suggest changes to drug candidates to maximize their potential efficacy.
"We have the right collection of institutions, individuals and facilities to make this [center] happen," says CCR's Miller. "In fact, many of us have been working collaboratively for years."
And, while it is very early, the business community is already starting to respond. A major firm with strong ties to the area and an interest in bioinformatics has stated that it is considering an additional significant investment in resources in Buffalo, now that a Center of Excellence in Bioinformatics has been proposed.
Holm adds that because Buffalo is a very low-overhead place to locate a business, he expects to see more firms become interested in the area as the center gets going. "Heads are turned by this kind of investment," he says.
That's also likely to be the response from federal funding agencies, according to Turkkan. "The federal government wants to see that a state supports a university's activities," she says. "When we can show them that we already have a coherent plan for this center, that we have space identified and major support from Albany, all of these things will make us that much more competitive-especially because they demonstrate to the NIH that recruitment of top faculty will be much easier."
In addition, Turkkan notes, the participating institutions have long histories of collaboration that include formal memoranda and faculty with joint appointments, and she emphasizes that it is the nature of these institutions themselves and the capabilities they possess that also make it possible to investigate bioinformatics in its broadest sense.
In the university, and in labs throughout the world, bioinformatics is driving such collaborations. Buffalo already is making and exploiting those connections, but it is Pataki's announcement that will put the area on the map.
"It's time to recognize that Buffalo has some unique strengths," says Holm. "The money for the new Center of Excellence in Bioinformatics will allow us to take those strengths and turn them into an economic engine for this area."
S.A. Unger is editor of Buffalo Physician.