Conformational Ensembles from Experimental Data and Computer Simulations

Conformational Ensembles from Experimental Data

and Computer Simulations

Poster Abstracts

59-POS

Board 19

Constructing Ensembles Using Site-Specific Vibrational Spectroscopy Probe Groups

Casey H. Londergan

, Rosalind J. Xu, Kristen L. Kelly, Shannon R. Dalton, Alice R. Vienneau,

Daniel M. Konstantinovsky.

Haverford College, Haverford, PA, USA.

Vibrational spectroscopy has an inherent advantage over other experimental techniques for

ensemble determination due to its very fast intrinsic time scale (10s of fs to a few ps), which

means that protein conformational changes are in slow exchange in vibrational spectra and the

full conformational distribution is thus present in some form in the spectrum. A recently

popularized approach uses functional groups with unique vibrational frequencies (i.e. the CN

stretch of nitriles) as reporters of the environment around specific sites in proteins. The infrared

or Raman lineshapes of these probe groups contain the local structural distribution, and the fast

intrinsic time scale also means that there is a direct match between the decay times of the

vibrational correlation functions and the time steps in all-atom molecular dynamics simulations.

We have placed the SCN vibrational probe group into several proteins, including model peptides,

alpha synuclein, calmodulin, and fuzzy viral complexes. The SCN group in particular is

surprisingly non-perturbative in most cases and this lack of perturbation suggests that it could be

placed in many systems, including directly along protein-protein and protein-membrane binding

interfaces. Our extensive experimental data provides new and previously unreported information

about the range of environments around specific sites in these proteins, especially about the

dynamic structures of bound protein-protein and protein-membrane complexes. We have also

performed initial molecular dynamics simulations intended to provide an interpretive guide to the

data, and it appears that there is at least semi-quantitative agreement between simulated probe

solvent exposure distributions and the CN frequencies and lineshapes of the SCN group. While

there are current challenges associated with making a direct connection between vibrational

probe data and simulations, this general two-pronged strategy is a promising new methodology

for determining and representing protein ensembles.