Molecular Dynamics Investigations of the α-Helix to β-Barrel Conformational Transformation in the RfaH Transcription Factor

Abstract

The C-terminal domain (CTD) of the transcription antiterminator RfaH folds to an α-helix bundle when it interacts with its N-terminal domain (NTD) but it undergoes an all-α to all-β conformational transformation when it does not interact with the NTD. The RfaH-CTD in the all-α topology is involved in regulating transcription whereas in the all-β topology it is involved in stimulating translation by recruiting a ribosome to an mRNA. Because the conformational transformation in RfaH-CTD gives it a different function, it is labeled as a transformer protein, a class that may eventually include many other functional proteins. The structure and function of RfaH is of interest for its own sake, as well as for the value it may serve as a model system for investigating structural transformations in general. We used replica exchange molecular dynamics simulations with implicit solvent to investigate the α-helix to β-structure transformation of RfaH-CTD, followed by structural relaxation with detailed all atom simulations for the best replica. The importance of interfacial interactions between the two domains of RfaH is highlighted by the compromised structural integrity of the helical form of the CTD in the absence NTD. Calculations of free-energy landscape and transfer entropy elucidate the details of the RfaH-CTD transformation process.

Structural conversion of the transformer protein RfaH: new insights derived from protein structure prediction and molecular dynamics simulations

“…Indeed, helix 1 rapidly unfolds whereas a significant portion of helix 2 remains structured in the simulations. This finding is in line with previous MD studies carried out on this system by using rather different approaches (replica exchange in implicit solvent, targeted and high-temperature MD in implicit solvent) (Gc, Bhandari, Gerstman, & Chapagain, 2014;Li et al, 2014). In both of these reports, which were published while the present work was in progress, a higher stability of helix 2 was observed.…”

Differential local stability governs the metamorphic fold-switch of bacterial virulence factor RfaH

“…To calculate the free-energy between both structures, implicit solvent MD simulations were performed on Amber16 along with CUDA as previously reported (28,29). Although water molecules are not being directly computed, the per-residue root mean square fluctuations (rmsf) of RfaH in both structures is comparable to explicit solvent models previously reported (30)(31)(32). Furthermore, this method requires an initial step of deep energy minimization of the system, suggesting that the phase change the solvent would undergo during this process would result in more artifactual dynamics than those arising from a steady solvation potential.…”

Multi-Funnel Landscape of the Fold-Switching Protein RfaH-CTD

“…However, probing such fold switching and mapping their energy landscape by experiments or in silico is a challenge. [10][11][12][13][14] Computationally, the problem is that the exploration of the ensemble of possible structures and the conversion between these structures happens on timescales that, on general-purpose computers, are not accessible in all-atom molecular dynamics simulations with explicit solvent. Enhanced sampling techniques such as Replica Exchange Molecular Dynamics (REMD) [15][16][17][18][19][20] promise to overcome this problem by realizing a random walk in temperature which allows the system to escape out of traps and cross barriers by explorations to higher temperatures.…”

“…Only when continuing the simulations from the best configurations by including solvent molecules explicitly was the correct β-barrel fold found. 11 We have recently proposed to overcome some of the limitations that hold back REMD by a Replica-Exchange-with-Tunneling (RET) approach. 21 We have shown that RET in conjunction with a Hamilton-Replica-Exchange 22,23 of systems where the "physical" system is coupled to varying degrees with biasing Go-models, allows efficient simulation of proteins and protein assemblies that can take more than one state.…”