Unfolding Proteins to Learn Nature’s Origami and Process Statistical Fingerprints of Folding Landscapes

Chemistry: Sheila Jaswal (Amherst College)
Mathematics: Amy Wagaman (Amherst College)

Information on protein stability and folding kinetics is critical to understanding the normal biological function of a protein, as well as the misfolding and aggregation properties of a growing number of proteins found to be involved in neurodegenerative and other diseases of conformation. We are conducting a large-scale analysis of more than one hundred proteins, and investigating new experimental methods (Hydrogen exchange mass spectrometry) to reveal insights relating to protein folding landscapes.

For the analysis side, we are investigating relationships between energetic quantities related to protein folding thermodynamics and kinetics (beyond the known formulaic relationships) and protein structure and function.  We are also investigating certain “outlier” proteins in terms of kinetics/thermodynamics in depth to see if we can ascertain from structural properties why they are outliers (do they have a different protein “fingerprint” than other proteins?). Several possible spin-off problems exist where studying homologues and protein families may be of interest.  A variety of multivariate statistical tools are necessary for the analysis including regression and clustering methods. There is also some possible application of dimension reduction methods, as we are still dealing with a large variable selection problem, due to the size of our database.

Traditional protein folding approaches destabilize the native state.  However, for many proteins, including amyloid precursor proteins and chaperone substrates, significant destabilization of the native state leads to aggregation.  For such proteins, Hydrogen Exchange Mass Spectrometry (HXMS) offers an equilibrium approach to explore their folding landscape at equilibrium. We have developed a numerical simulations approach to model simple HXMS behavior for proteins. By systematically varying conditions of the simulation in analogy to the experimental conditions of temperature and pH, we will probe the relationship between the experimental observables of HXMS and the underlying folding landscape.   This will allow us to optimize methods of analysis to extract folding information from experimental HXMS profiles of proteins. To validate our simulations and analysis, we will first perform HXMS on simple proteins whose landscapes have already been determined through traditional methods.  After validation, we will apply our HXMS approach to proteins not accessible to tradition folding approaches.