Many proteins function only when associated with other large molecules. Interested in the architecture of such macromolecular assemblies, Stephen Harrison set out in 1964 to determine the detailed structure of a virus, since viral coats are assembled from protein subunits. In 1978, he became the first person to achieve that goal. Since then, he has determined the structures of nine additional viruses or clusters of viral proteins. And he has studied protein assemblies that regulate genes, help cells divide, or form lattices around vesicles that transport cargo through cells.
Harrison was first exposed to science by his parents, who were biologists. He became intrigued with biological structures when, as an undergraduate in physics and chemistry at Harvard, he heard lectures by James Watson and Francis Crick, the co-discoverers of the structure of DNA. But he became frustrated when he started postgraduate studies in biophysics at Harvard in 1963. "My initial laboratory experience convinced me that I needed to be able to picture the molecules with which I was trying to work," he recalls. "Just as you need an architect's drawing to understand what keeps a building from collapsing or to know how to build it in the first place, so you need a molecular structure to understand how a virus particle assembles." A year in Aaron Klug's laboratory in Cambridge, U.K., convinced him that x-ray crystallography would be the right approach. X-rays make diffraction images of crystals on photographic film. Using those images, a computer determines how the crystal's atoms are arranged.
Returning to Harvard in 1965, Harrison took root in the laboratory of Don Caspar, who had earlier produced the first x-ray diffraction images of single crystals of tomato bushy stunt virus (TBSV). At that time, plant viruses were about the only readily available examples of macromolecular assemblies.
He spent the next 3 years improving x-ray crystallography by devising better ways to beam x-rays at crystals and writing smarter computer programs to decode the results. "But the computers of the time simply were not powerful enough to do what I needed," Harrison says. "Fortunately, by the time other problems had been solved, available computer power had grown."
After Harrison received his Ph.D. in 1968, he spent 3 years in Caspar's lab obtaining an electron density map of TBSV from diffraction images of mercury derivatives of the viral protein. The level of resolution, 25 angstroms (Ǻ, one 10-billionth of a meter), was not sufficient to reveal the positions of the electrons around individual atoms, which are separated by distances on the order of 1 Ǻ.
Harrison joined the Harvard faculty in 1971. Thanks to improved x-ray sources, computer programs written by a French colleague, and 3 weeks of nonstop work, Harrison obtained a 5.5 Ǻ map of TBSV in January 1974. At that resolution, it was possible to make some deductions about the architecture of the viral coat. Each of the 180 identical protein subunits of this bumpy ball had two domains hinged together. When the hinge rotated, the subunit's conformation changed. In one conformation, there was a stretch of electron density between the two domains, which initially was thought to be RNA.
By December 1977, a very detailed map—2.9 Ǻ resolution—was in hand. For the first time, it was possible to see the position of the protein subunit's polypeptide chain. Moreover, it was clear that the structure thought to be RNA was actually protein—an arm protruding from one of the domains. The arm was folded in one conformation and unfolded in the other. Arm-like interactions have since emerged as important organizational features of many protein assemblies. Thus, Harrison had produced the first high-resolution structure of any virus. The seminal paper was published in Nature in 1978.
Subsequent projects were chosen because they related to major problems, such as viral assembly or public health. "In the case of HIV, my late colleague, Don Wiley [an HHMI investigator at the time of his death] and I decided early in the epidemic that it would have been irresponsible for two leading structural virology groups not to attempt to make some contribution to so pressing a global health problem," Harrison says. His laboratory is currently trying to understand how the envelope protein of HIV changes shape when it binds to the receptor that admits it to cells. Understanding the shape changes might lead to ways to prevent entry, a crucial step in HIV infection.
Viral entry into cells has become a dominant theme in Harrison's lab over the past decade. "The virus particle is not a passive package, but a delivery agent that actively participates in the penetration process," Harrison says. He found that the viral coat subunits of nonenveloped viruses undergo dramatic changes in shape that help them cross the cell's outer membrane. Viruses that have a membranous envelope around their protein coat produce fusion proteins that help the viral membrane fuse with the host cell's membrane. "The demonstration that there is a single kind of fusion mechanism [for enveloped viruses] has been a particularly satisfying outcome of our work," Harrison says, noting that proteins involved in viral entry are potential targets for antiviral drugs.
Harrison's 25-year study of the protein clathrin (performed in collaboration with Tom Kirchhausen at Harvard Medical School) relates to viral entry because clathrin creates a supportive net around vesicles that carry viruses and other particles into cells. By 2004, the two investigators had a detailed atomic model of clathrin, which proved to be a three-armed protein that interacts with other clathrin molecules to form lattices. Because only a few connections clip the structure together, shells can adopt different curvatures and therefore carry cargoes of varying sizes and shapes. "Perhaps even more important," Harrison says, "local interactions allow a localized mechanism for taking the structure apart, which must occur promptly after the vesicle buds."
Another Harvard colleague, Mark Ptashne, inspired Harrison to study protein assemblies that bind DNA to regulate genes. Their collaboration led to the first structure of such a regulatory protein/DNA complex. Harrison has gone on to analyze the concerted participation of multiple DNA-binding proteins in gene regulation, seeking to understand the complex control circuits embodied by these proteins.
From an architectural point of view, the kinetochore, the structure that anchors the centromere of a chromosome to the spindle of a dividing cell, is also a superstructure that recognizes DNA and recruits many additional components, Harrison points out. "As I believe that the future of structural biology involves making realistic molecular movies of subcellular processes, then starting to work on one of the key bits of machinery in cell division was too tempting to resist," he says. Biochemistry, electron microscopy, and preliminary forays into x-ray crystallography are uncovering the functions of each component, he adds.
During the next 5 years, Harrison hopes to determine how nonenveloped viruses, such as rotavirus (a major cause of gastrointestinal disease in children), cross the cell membrane. Just as structural studies of viral fusion proteins have led to a unified view of enveloped viral entry, so Harrison believes that there will be one or a few mechanistic themes in nonenveloped viral entry. He also hopes to learn how an enzyme uncoats clathrin-coated vesicles and how more than 50 different kinds of proteins assemble together into a kinetochore. In the long term, he hopes to create experimentally based molecular movies by combining data from x-ray crystallography, nuclear magnetic resonance studies, electron microscopy, live-cell imaging, and single-molecule biophysics. And he would like to develop a structure-based approach to vaccines. Such vaccines would faithfully mimic the 3-D structures of parts of viral proteins, improving the immune system's response to particular pathogens. Convinced that this could happen within a decade, Harrison is participating in the NIH-funded Center for HIV/AIDS Vaccine Immunology. "I want to help educate others in the consortium about how to think about structures, as well as to make our own contributions if possible," he says.