Conformational Ensembles from Experimental Data and Computer Simulations

Conformational Ensembles from Experimental Data

and Computer Simulations

Poster Abstracts

6-POS

Board 6

Hierarchical, Structural Basis for Motions Encoding HDXMS Data

Dominik Budday

, Sigrid Leyendecker

, Henry Van den Bedem

University of Erlangen-Nuremberg, Erlangen, Germany,

Stanford University, Menlo Park, CA,

USA.

Hydrogen-Deuterium Exchange Mass Spectroscopy (HDXMS) can provide important

experimental insights into functional dynamics based on neutron exchange between protein and

solvent. Differences in Deuterium exchange by the protein between wild-type and mutants, or

across different members of a protein family can relate structural dynamics with function.

However, the long time-scales of HDXMS experiments often make data interpretation

challenging. The availability of computational methods capable of resolving these spatio-

temporal scales could result in broader adoption of HDXMS.

Here, we adapt Kino-Geometric Sampling (KGS) to provide a structural basis for motions with

hierarchically increasing hydrogen bond violations. Our geometric approach encodes hydrogen

bonds as holonomic constraints, imposing collective motions on the dihedral degrees of freedom

to maintain cycle-closure. A singular value decomposition of the constraint Jacobian ranks

independent, orthonormal motion modes by constraint violation, reminiscent of normal modes

that describe functionally relevant motions at low eigenfrequencies. The method is based on the

hypothesis that violations of the hydrogen bond network geometrically encoded in KGS is

related to functional, dynamic exchange in the molecule. Our kinematic, time-independent

analysis is very fast and applicable to proteins and RNA, making it suitable to study motions

across spatio-temporal scales in a matter of seconds.

Predictions from KGS hierarchical motions on exchanging hydrogen bonds show qualitative

agreement with HDXMS measurements in the α-subunit of protein Gs. We further compare our

predictions with the Start2Fold database which provides a collection of sparse data on folding

cooperativity and stability of proteins measured by HDXMS and related experimental methods.

Initial results are promising and strengthen our hypothesis, displaying a fast, yet detailed

computational tool to interpret and predict hydrogen-deuterium exchange in macromolecules.