Molecular Dynamics Simulations of RNA Stem-Loop Folding Using an Atomistic Force Field and a Generalized Born Implicit Solvent

Figures

• 

  1. Secondary structures of 26 RNA stem-loops simulated in this study and their corresponding Protein Data Bank IDs. (A) Eighteen class I stem-loops without bulges or internal loops. (B) Eight class II stem-loops with bulges or internal loops. The G-C/A-U Watson–Crick and the G-U wobble base pairs are represented by black and red lines, respectively. Wobble base pairs A7-C21 in 2LDL and C6-A18 in 2KF0 were observed by NMR spectroscopy and are denoted by plus signs (+). A G8-A27 mismatch base pair in 1N8X was observed by NMR spectroscopy and is denoted by an asterisk (*).

• 

  2. Time evolution of the fraction of native base pairs Q for 18 class I stem-loops without bulges or internal loops (Figure A). Three independent simulations (red, green, and blue lines) were performed for each model, starting from the extended conformations.

• 

  3. Time evolution of the fraction of native base pairs Q for eight class II stem-loops with bulges or internal loops (Figure B). Three independent simulations (red, green, and blue lines) were performed for each model, starting from the extended conformations.

• 

  4. Comparison of class I stem-loop structures between the experiment (blue) and cluster centroid (red) exhibiting the lowest non-hydrogen root mean square deviation (RMSD) values for the entire molecule. Below the RNA name of each structure, the cluster number, RMSD value, and fraction of native base pairs (Q) are shown and separated by slashes. The k-means method with k = 20 was used to cluster the RNA structures sampled from the three independent molecular dynamics simulations starting from extended conformations.

• 

  5. Comparison of the class II stem-loop structures with bulges or internal loops between the experiment (blue) and cluster centroid (red) exhibiting the lowest non-hydrogen root mean square deviation (RMSD) for the entire molecule. For models 1ESY, 2JWV, and 1R7W, which did not correctly fold, the centroid structure of the first cluster is shown. Below the RNA name of each structure, the cluster number, RMSD value, and fraction of native base pairs (Q) are shown and separated by slashes. The k-means method with k = 20 was used to cluster the RNA structures sampled from the three independent molecular dynamics simulations starting from extended conformations.

• 

  6. Folding trajectory of the 1SZY class I stem-loop from the first molecular dynamics simulation using the DESRES-RNA force field and the GB-neck2 model. (A) (Upper) Non-hydrogen root mean square deviation (RMSD) for the stem region. The dashed line indicates an RMSD of 2 Å. (Lower) Base pairs formed during simulation, where the native and non-native base pairs are represented by red and blue, respectively. (B) Secondary structure of 1SZY. The stem region is indicated in red. (C) Snapshots of the trajectory at various simulation time points. RNA backbones are represented by ribbons. The nucleotide ring structures are colored according to nucleotide type: adenine, red; cytosine, yellow; guanine, green; and uracil, cyan.

• 

  7. Folding trajectory of the 1ANR class II stem-loop from the third molecular dynamics simulation using the DESRES-RNA force field and the GB-neck2 model. (A) (Upper) Non-hydrogen root mean square deviation (RMSD) for the two stem regions, where stem 1 (red lines) and stem 2 (orange lines) represent the double helical regions directly connecting the hairpin loop and those near the 5′ and 3′ termini of the RNA, respectively. The dashed line indicates an RMSD of 2 Å. (Lower) Base pairs formed during simulations, where the native and non-native base pairs are represented in red and blue, respectively. (B) Secondary structure of 1ANR. Stems 1 and 2 regions are indicated in red and orange, respectively. (C) Snapshots of the trajectory at various simulation time points. RNA backbones are represented by ribbons. The nucleotide ring structures are colored according to nucleotide type: adenine, red; cytosine, yellow; guanine, green; and uracil, cyan.

• 

  8. Comparison of the loop and bulge structures in four representative RNA stem-loops from NMR and molecular dynamics (MD) simulations with explicit and implicit solvent models. Loop regions of 1F85, 1FHK, 2KOC, and 1R2P, as well as the bulge region of 1R2P, are shown for three structural sources: NMR-derived structures (left column), simulations using the DESRES-RNA force field with the TIP4P-D explicit solvent model (middle column), and simulations using the same force field with the GB-neck2 implicit solvent model (right column). Dashed black lines indicate hydrogen bonds. In the NMR structures, key hydrogen bonds stabilizing the loop conformations are highlighted with red dashed circles. For the TIP4P-D explicit solvent MD simulations, snapshots at 1 μs are shown with non-hydrogen atom root mean square deviation (RMSD) values for the loop and bulge regions. For the GB-neck2 implicit solvent MD simulations, the centroid structures with the lowest RMSD values for the entire molecule are shown with their cluster numbers and RMSD values for the loop and bulge regions. In the structures from the explicit and implicit solvent MD simulations, hydrogen bonds consistent with those…

• 

  9. FRET-averaged end-to-end distances ⟨R FRET⟩ (Å) of rU40 single-stranded RNA obtained from molecular dynamics (MD) simulations using the DESRES-RNA and AMBER-OL3 force fields with the GB-neck2 implicit and explicit solvent models at various ionic concentrations. Standard errors for the simulation data obtained using the GB-neck2 model are too small to be visible in the plot. Values for explicit solvent simulations were obtained from the plot in ref , where the TIP4P-D and TIP3P water models were used with the DESRES-RNA and AMBER-OL3 force fields, respectively, for the MD simulations of rU40 at 300 K. The experimental values were obtained from the plot reported in ref . In both the experiment and the explicit solvent MD simulations, NaCl was used as the background electrolyte.