Goldstein group ::
Modern proteins are the result of an evolutionary process. In order to understand these biological macromolecules, we need to investigate the evolutionary pressures that have determined their form and function. Conversely, because proteins encode this evolutionary heritage, studies of the properties of proteins can provide insight into the process of natural selection. Finally, following the evolutionary path of specific proteins can provide important information about their biochemical characteristics. Combining insights from physical chemistry with concepts borrowed from condensed matter physics, artificial intelligence, complexity theory, and mathematical biology, we are developing computational and theoretical methods to explore these aspects.
Proteins are generally active only in their correctly folded state, and must find this state sufficiently rapidly to avoid aggregation or proteolysis. In this way, evolutionarily-selected proteins differ from random sequences of amino acids. We seek to investigate the thermodynamic requirements necessary for rapid folding, and the consequences of these requirements for the folding process and the other properties of biological proteins. Using simplified models of the protein free-energy landscape, we try to understand the interplay between various energetic interactions, and to explain the experimentally-observed folding dynamics. By modeling how the need-to-fold affects the fitness of proteins, we can understand how sequences evolve, why proteins have the structures that they do, and how these structures change to meet new needs.
We are also using a physical chemistry perspective to directly model the evolutionary process, through the construction of mutation matrices that represent the rate of substitution of one residue for another in the protein. These matrices allow us to reconstruct the evolutionary process, and let us query nature directly about the properties that are conserved during evolution, and thus are likely to be important for structural or functional reasons. We can also use these evolutionary models to help predict secondary and tertiary structure.