Axel Brunger credits his academic success to the many disciplines he has studied and the excellent high school courses in Germany that influenced his career. “It has been tremendously important to me to have an education in applied mathematics and physics and then to later learn biology and biochemistry,” he says. “There are so many opportunities to apply tools and methods developed in one discipline to another.”
Brunger develops tools for interpreting x-ray crystallography diffraction data. X-ray crystallography explores the structures of large molecules such as proteins by crystallizing the molecules and bombarding the crystals with x-rays. The images produced by the diffracted rays are then translated into a molecule’s 3D structure. Before Brunger devised a computer program called X-PLOR, crystallographers often spent months poring over x-ray results or gave up altogether because the diffraction data were inadequate for the task. But X-PLOR enabled scientists to refine structures in just days.
More recently, Brunger incorporated X-PLOR into a software suite called CNS, which includes most of the steps for mining reduced x-ray diffraction data. The article describing CNS was widely cited, and it still generates about 1,000 citations per year. Moreover, Brunger developed a major computational tool called the free R value, which indicates whether molecular models are likely correct. His expertise in computational chemistry also led him to explore the structures of proteins that enable one nerve cell to communicate with another. “That is one of the most fundamental processes in neurobiology,” he says.
X-PLOR broke new ground with its application of simulated annealing, which Brunger explains by evoking a model of a mountainous landscape with many valleys. To find the lowest point on the model, you roll marbles down the mountain and see where they settle. If you shake the model, some of the marbles bounce out of their valleys into even lower valleys. By shaking the model many times, you eventually find the lowest (or nearly lowest) point on the model. The many iterations of simulated annealing are analogous to shaking the model many times, and finding the lowest point on the model is analogous to finding an accurate structure from x-ray crystallography diffraction data. Although simulated annealing had been used to solve other problems, such as how to plan the best route for a traveling salesman who must visit many cities, no one had applied it to x-ray crystallography before. “[X-PLOR] allowed people to solve certain problems that before were considered intractable or required new diffraction data,” Brunger explains.
To extend X-PLOR into CNS, Brunger and his colleagues developed an intuitive computer language that allows researchers—in crystallography and other fields—to try out many methods for interpreting x-ray diffraction data. “It was one of the first easy-to-use systems that noncrystallographers could use to solve structures,” Brunger says. “Therefore, it allowed people to focus more on biology, which is ultimately what you want to get out of it.”
A better understanding of neurotransmission is Brunger's goal for his biological studies. This process is at the heart of nervous system communication, because it transfers chemicals called neurotransmitters from one neuron to another. In the mid-1990s, Brunger began to study the structures of SNARE proteins, which are found in cell membranes. The SNARE complex involved in neurotransmission consists of syntaxin-1, synaptobrevin, and SNAP-25. Synaptobrevin is found on the bubbles of membrane (synaptic vesicles) that carry neurotransmitter to the inner face of a neuron’s outer membrane. That face contains syntaxin and SNAP-25. As the two membranes make contact, their respective SNAREs zip them together into one membrane, ejecting neurotransmitter from the neuron.
Brunger showed that those SNAREs are corkscrew-shaped proteins that can assemble into quartets of one syntaxin-1, one synaptobrevin, and two SNAP-25 helices. The proteins lie in parallel, with their heads pointing in the same direction. “That lends support to the notion that assembly of the complex will pull membranes together and set the stage for [membrane] fusion,” Brunger explains. “So the structure suggested a molecular mechanism.” This mechanism is reminiscent of that used by certain viral fusion proteins, such as the one that enables the flu virus to enter cells.
Brunger has also studied the interactions between SNARE proteins and clostridial neurotoxins (CNTs), which cause botulism and tetanus. After a CNT enters a neuron, it emits an enzyme that severs SNAREs, destroying an essential participant in neurotransmission. X-ray crystallography of one of these protein-cutting enzymes attached to the SNAP-25 SNARE showed a string-like molecule wrapped around the protease. This revealed that SNAREs can exist as long strings as well as in helical form in neuronal cytoplasm. Brunger suggests that SNAREs are strings before membranes fuse and that contact with their counterpart SNAREs makes them fold into helices, driving the membranes together.
After his move from Yale to Stanford University in 2000, Brunger began to study the synaptic release machinery at the single-molecule and -vesicle levels. He was introduced to this technology by Dr. Steven Chu who was then at Stanford University. Over the years, Brunger expanded his efforts in the single-molecule/particle optical microscopy field; he also began developing novel computational tools for structure determination based on single-molecule localization and FRET data to model biomolecular complexes. Brunger believes that a combination of optical approaches using reconstituted systems, optical imaging of biological specimens, high-resolution structural techniques, and computer simulation will eventually allow elucidation of the molecular mechanism of Ca2+-triggered synaptic vesicle fusion.