CI2Daggett Research Group | Research | Protein Folding

University of Washington - College of Engineering - School of Medicine - Department of Bioengineering

Selected Projects

Chymotrypsin Inhibitor 2
Cytochrome b5
Cytochrome b562
Engrailed Homeodomain
p13 suc1
Protein A
WW domain
Model Peptides

Protein Folding

WW domain How a protein becomes a folded, biologically active structure remains elusive despite years of intense research activity. To examine this folding reaction we need to characterize all states, including the native, transition, intermediate(s), denatured, as well as the mechanism of conversion between these states. The heterogeneous, dynamic, and often transient nature of the nonnative states does not lend themselves to easy experimental observation. As a result, we use molecular dynamics simulations to complement, extend, and, in some cases, predict data from protein folding and unfolding experiments to more fully map the folding/unfolding reaction coordinate.

Protein folding is hypothesized to be a microscopically reversible process, meaning that for a system at equilibrium, the number of molecules unfolding via a given pathway is equal to the number of molecules refolding along the same pathway in the reverse direction. Or, in other words, folding and unfolding follow the same path, just in opposite directions. This implies we can learn about protein folding by studying the unfolding pathway and considering the order of events in reverse.

EnHD Folding/Unfolding We have directly observed microscopic reversibility in protein simulations of two proteins at temperatures where their folding and unfolding rates are equal. In the case of CI2, we saw the protein move from the native state (N), to a nearly native state (N′), through the transition state (TS) to the denatured state (D), then in the reverse order back to N′. Major events in unfolding, including losing native packing and hydrogen bonds, helix breaking, and side-chain interactions shifting, occurred in the reverse order in the subsequent refolding. A follow-up study in the engrailed homeodomain showed passage through the same states (N → N′ → TS → D → TS → N′ → N) twice. It additionally illustrated the similarity of structures along the folding and refolding pathways in RMSD space.

In order to gain a better understanding of the important states along the unfolding pathway and the transitions between these states, we have performed both native and unfolding molecular dynamics simulations on over 800 structurally diverse proteins as part of our Dynameomics project. We have identified and characterized the TS ensembles of 183 structurally diverse proteins, representing the largest, most comprehensive study of the TS to date. There was no significant difference in the overall protein properties when the native structures were separated into four different fold classes (all proteins, all-α, all-β, and mixed α/β and α+β). Instead, the global, average global properties of the TS ensemble converged with low standard deviations across all proteins, suggesting there are generic rules for the structure and properties of the TS.

TPR Protein Another area of interest is the folding/unfolding pathways of the elongated tetratricopeptide repeat (TPR) proteins, made of consecutive copies of homologous structural units called "repeats". The repeats, consisting of a helix-turn-helix motif, stack together with only a few inter-repeat contacts to neighboring repeats, lacking interactions between residues that are distant in the sequence. These properties allow repeat proteins such as TPRs to add or delete individual repeats without destroying the overall protein structure, but varying their surfaces for protein-protein interactions. We use molecular dynamics simulations to characterize unfolding intermediates, both within a single repeat and in the larger TPR protein structure.

Relevant Publications