Proteins have an astonishing structural repertoire which partially accounts for their ability to perform a diverse array of biological functions. There are over 900 basic types of protein structure, or "folds", but few current models for how this amazing structural diversity originated and developed. One possibility is that some folds or topologies evolved from others through mutationally induced structural rearrangements. We are studying one example of evolutionary structural change which has occurred in the Cro family of DNA-binding proteins. We have shown that the fold of lambda Cro, which is a mixture of alpha-helix and beta-sheet, evolved from an all-alpha helical ancestor through an alpha-helix to beta-sheet conformational switching process (Newlove et al, 2004). We have recently identified pairs of Cro proteins with over 40% sequence identity that have different folds, suggesting that this transformation took place relatively recently, and that it may be possible to identify specific mutations which played a role in structural conversion.

We are using bacteriophage Cro proteins, which bind to specific DNA sequences and regulate the lifestyle of certain viruses, as a microcosm of functional evolution in DNA-binding proteins. Different members of the Cro family have evolved to bind a remarkable variety of sequences. What are the mutational mechanisms behind evolutionary changes in DNA-binding specificity? Are they simple or complex? We have used bioinformatics to uncover a partial "evolutionary code" (Hall et al, 2005) in which simple changes in amino acid residues involved in direct nucleotide base contacts determine much of the evolution of specificity in Cro. We are interested in the questions of how robust and broadly applicable this code is, and how its ability to encode specificity depends upon other factors such as evolutionary changes in the protein structure.

Some human diseases such as Alzheimer's are related to protein misfolding events that lead to aggregation and fibril formation in vivo. One major question of interest is what features of an amino acid sequence promote such processes or prevent them from occurring, and whether protein sequences evolve to avoid aggregation-prone sequences. For example, there is considerable evidence that short highly hydrophobic sequence fragments promote misfolding. A recent database survey from our laboratory (Patki et al, 2006) showed that evolution seems to avoid such sequences even in situations where they would be expected to play an important role in stabilizing protein structures.

We are studying a brown spider venom toxin called sphingomyelinase D, which is solely responsible for dermonecrotic lesions in humans bitten by these spiders. This TIM barrel enzyme evolved from a ubiquitous family of housekeeping enzymes known as the glycerophosphoryl diester phosphodiesterases (GDPDs). Interestingly, a similar enzyme is found in pathogenic corynebacteria and we, in collaboration with Greta Binford at Lewis & Clark College, recently showed that horizontal gene transfer between spiders and bacteria accounts for the presence of this enzyme in these very different organisms (Cordes & Binford, 2006). We are currently interested in the mechanism by which the housekeeping ancestors of these proteins acquired toxic function.