Contribution of energy values to the analysis of global searching molecular dynamics simulations of transmembrane helical bundles

Jaume Torres, John A.G. Briggs, Isaiah T. Arkin*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

22 Scopus citations

Abstract

Molecular interactions between transmembrane α-helices can be explored using global searching molecular dynamics simulations (GSMDS), a method that produces a group of probable low energy structures. We have shown previously that the correct model in various homooligomers is always located at the bottom of one of various possible energy basins. Unfortunately, the correct model is not necessarily the one with the lowest energy according to the computational protocol, which has resulted in overlooking of this parameter in favor of experimental data. In an attempt to use energetic considerations in the aforementioned analysis, we used global searching molecular dynamics simulations on three homooligomers of different sizes, the structures of which are known. As expected, our results show that even when the conformational space searched includes the correct structure, taking together simulations using both left and right handednesses, the correct model does not necessarily have the lowest energy. However, for the models derived from the simulation that uses the correct handedness, the lowest energy model is always at, or very close to, the correct orientation. We hypothesize that this should also be true when simulations are performed using homologous sequences, and consequently lowest energy models with the right handedness should produce a cluster around a certain orientation. In contrast, using the wrong handedness the lowest energy structures for each sequence should appear at many different orientations. The rationale behind this is that, although more than one energy basin may exist, basins that do not contain the correct model will shift or disappear because they will be destabilized by at least one conservative (i.e. silent) mutation, whereas the basin containing the correct model will remain. This not only allows one to point to the possible handedness of the bundle, but can be used to overcome ambiguities arising from the use of homologous sequences in the analysis of global searching molecular dynamics simulations. In addition, because clustering of lowest energy models arising from homologous sequences only happens when the estimation of the helix tilt is correct, it may provide a validation for the helix tilt estimate.

Original languageEnglish
Pages (from-to)3063-3071
Number of pages9
JournalBiophysical Journal
Volume82
Issue number6
DOIs
StatePublished - 2002

Bibliographical note

Funding Information:
This work was supported by grants from the Wellcome Trust and the Biotechnology and Biological Sciences Research Council to I.T.A.

Fingerprint

Dive into the research topics of 'Contribution of energy values to the analysis of global searching molecular dynamics simulations of transmembrane helical bundles'. Together they form a unique fingerprint.

Cite this