148 / 161
Conformational Ensembles from Experimental Data
and Computer Simulations
Poster Abstracts
144
107-POS
Board 27
Combining Small Angle X-ray Scattering Experiments with Accelerated Molecular
Dynamics Simulations to Determine the Conformational Ensemble of Tri-ubiquitin Chains
Jeff Wereszczynski
.
Illinois Institute of Technology, Chicago, IL, USA.
Small angle X-ray scattering (SAXS) has become an increasingly popular structural technique
for characterizing the ensemble of solution states of flexible biomolecules. However, data
resulting from SAXS is typically noisy and low-dimensional and may therefore be difficult to
interpret without additional structural knowledge. In principle, this information can be provided
by molecular dynamics (MD) simulation, but conventional MD trajectories rarely sample
sufficient phase space to probe the range of structures that contribute to the observed
experimental data. Accelerated MD (aMD) can overcome these sampling inadequacies by
introducing a bias to the underlying energy landscape that lowers the height of energy barriers
and encourages conformational transitions, albeit at the cost of distorting the Boltzmann
distribution of states. Here, we present a method for combining the results of aMD simulations
with experimental SAXS data to accurately model the relative populations of representative
solution states. Scattering states are first identified from aMD trajectories, and their populations
are then re-weighted against empirical data through a Bayesian Monte Carlo approach. Special
care is taken to avoid ensemble over-fitting by iteratively considering increasing subsets of
scattering states along with the associated Akaike Information Criterion, and by reducing
experimental data to the Shannon sampling limit. We apply this technique to several ubiquitin
trimers and find that aMD trajectories typically outperform conventional MD simulations in both
goodness-of-fit and model convergence speed. Furthermore, we observe that different ubiquitin
linkages yield distinct ensembles, which points to their unique roles in biological signaling.
These methods are being implemented in the “SASSIE” webserver, which aims to provide an
easy-to-use modeling interface for interpreting data from scattering experiments.