Protocol for calculating binding free energy of RNA:RNA interactions through molecular dynamics simulations using adaptive biasing force technique

Abstract

The adaptive biasing force (ABF) technique allows sampling to proceed in a flat free energy surface when performing molecular dynamics (MD) simulations. Here, we present a protocol to perform MD simulations using the ABF technique and apply it to calculate the binding free energy of an RNA:RNA interaction. We describe steps for server setup, test running software, and building molecular models. We then detail procedures for running and configuring ABF-MD simulations and analyzing binding free energy and structural change. For complete details on the use and execution of this protocol, please refer to Fujita et al.¹ and Kameda et al.².

Keywords: Biophysics; Biotechnology and bioengineering; Computer sciences.

Graphical abstract

Figure 1

PIC structure models and reaction coordinates in this study (A) mRNA and tRNA around the codon-anticodon base-pairs are shown in black and silver, respectively. Distances d₁, d₂ and d₃ used as collective variables (reaction coordinates) are indicated. To the right, rRNA segments (split) are shown in green, and protein components are shown in other colors. (B) Water molecules are shown in pink, which enclose the system and form a sphere. Adapted from.

Figure 2

Atoms specified for measuring base-pairing distances Atomic models of RNA base pairing are presented, highlighting the mass center (asterisk) of the atoms (boldface) specified as distance information in ABF simulation runs. Dotted line, hydrogen bonding.

Figure 4

Examples of the results from assessing convergence and computing free energy The 25 Å radius PIC model with AUG in the P-site was used for ABF simulation runs at Beocat in Kansas State University. (A) Convergence of free energy profiles with time evolution. The L(τ, τ − Δτ) of P(d₁, d₂, d₃) is shown (Δτ = 100 ns). (B) ΔG_binding computed from this simulation (red circle) was compared to the values reported previously. Error bars show the mean ± S.E.M. of ΔG_binding obtained from each simulation trajectory of previous experiments done for the PIC model with indicated start codons.

Figure 3

Definition of binding free energy Schematic representation of G_bound, G_unbound, and ΔG_binding is presented. Adapted from.

Figure 5

Experimental validation of calculated free energy penalty values Free energy penalty values were calculated for the ribosomal P-site with indicated start codons and were compared to initiation frequencies from the relevant codons in human embryonic kidney 293T cells. Adapted from and presented after modification.