Multidimensional Replica Exchange Molecular Dynamics Yields a Converged Ensemble of an RNA Tetranucleotide

Abstract

A necessary step to properly assess and validate the performance of force fields for biomolecules is to exhaustively sample the accessible conformational space, which is challenging for large RNA structures. Given questions regarding the reliability of modeling RNA structure and dynamics with current methods, we have begun to use RNA tetranucleotides to evaluate force fields. These systems, though small, display considerable conformational variability and complete sampling with standard simulation methods remains challenging. Here we compare and discuss the performance of known variations of replica exchange molecular dynamics (REMD) methods, specifically temperature REMD (T-REMD), Hamiltonian REMD (H-REMD), and multidimensional REMD (M-REMD) methods, which have been implemented in Amber’s accelerated GPU code. Using two independent simulations, we show that M-REMD not only makes very efficient use of emerging large-scale GPU clusters, like Blue Waters at the University of Illinois, but also is critically important in generating the converged ensemble more efficiently than either T-REMD or H-REMD. With 57.6 μs aggregate sampling of a conformational ensemble with M-REMD methods, the populations can be compared to NMR data to evaluate force field reliability and further understand how putative changes to the force field may alter populations to be in more consistent agreement with experiment.

Structural fidelity and NMR relaxation analysis in a prototype RNA hairpin

“…The results presented here indicate that the experimentally observed NMR structure and fluctuations are closely reproduced by the simulations using the most current force field, suggesting that experimental ensemble is well-represented by the free energy basin sampled in the simulations. This need not imply that we have found the global minimum for these force fields, and more global studies of tetraloop structures (Banas et al 2010;Cheatham and Case 2013;Bergonzo et al 2014), including some with a UUCG tetraloop similar to that studied here (TE Cheatham and N Henriksen, pers. comm.…”

Highly sampled tetranucleotide and tetraloop motifs enable evaluation of common RNA force fields

“…Common assessments of convergence instead focus on reporting the minimized differences between at least two independent simulations. Simulations can start from different initial velocities or solvent/ion distributions and be considered independent, or, as reported in this study and others, from completely different starting structures, commonly from folded and unfolded or different ensembles of the folded structure (Chen and García 2013;Kührová et al 2013;Bergonzo et al 2014). Additional measures of convergence examined in this study include the use of different methods of biasing the M-REMD simulations, resulting in equivalent sampling between two runs.…”

“…The GACC tetranucleotide was built, minimized, and equilibrated as previously described using each of the force fields that appear in Table 1 (Henriksen et al 2013;Bergonzo et al 2014). Structures built natively in CHARMM36 were converted to AMBER topology files using CHAMBER (Crowley et al 2009).…”

“…These files are provided in the Supplemental Information. CCCC tetranucleotide structures were built using the sequence command in tleap, following the same methods described previously (namely, neutralizing charge with Na + during the addition of 2500 solvent molecules) (Henriksen et al 2013;Bergonzo et al 2014).…”

“…A step in the right direction has been the application of enhanced sampling methods which have been used to identify problems and validate updates to force field parameter sets (Mlýnský et al 2010;Chen and García 2013;Nguyen et al 2014). Widely applied among enhanced sampling methodologies are replica exchange molecular dynamics (REMD) simulations, which have become a common way to evaluate conformational ensembles of RNA (Sorin et al 2002;Vaiana and Sanbonmatsu 2009;Chen and García 2013;Henriksen et al 2013;Kührová et al 2013;Bergonzo et al 2014;Roe et al 2014). In REMD, several independent simulations are run at different temperatures or Hamiltonians (referred to as T-REMD and H-REMD, respectively), and exchanges are attempted between them at specific intervals (Hansmann 1997;Sugita and Okamoto 1999).…”