Nucleoprotein Phase-Separation Affinities Revealed via Atomistic Simulations of Short Peptide and RNA Fragments

Abstract

Liquid-liquid phase separation of proteins and nucleic acids into condensate phases is a versatile mechanism for ensuring the compartmentalization of cellular biochemistry. RNA molecules play critical roles in these condensates, particularly in transcriptional regulation and stress responses, exhibiting a wide range of thermodynamic and dynamic behaviors. However, deciphering the molecular grammar that governs the stability and dynamics of protein-RNA condensates remains challenging due to the multicomponent and heterogeneous nature of condensates. In this study, we employ atomistic simulations of 20 distinct mixtures containing minimal RNA and peptide fragments which allows us to dissect the phase-separating affinities of all 20 amino acids in the presence of RNA. Our findings elucidate chemically specific interactions, hydration profiles, and ionic effects that synergistically promote or suppress protein-RNA phase separation. We map a ternary phase diagram of interactions, identifying four distinct groups of residues that promote, maintain, suppress, and disrupt protein-RNA clusters.

Figure 1:

Condensation propensity of central X residues in GXG peptides with Uracil bases of RNA. (A) Representative snapshots of simulations for four groups based on their condensation propensity: Group P+: strong promoter (ARG), Group P: promoter (PHE), Group S: suppressor (PRO), and Group S+: strong suppressor (MET). Cluster size is quantified by (B) the fraction of GXG peptides and (C) the fraction of PolyU fragments inside the cluster. (D) Cluster density vs. cluster size for all 20 residues. (E) SASA per atom in the cluster for each system.

Figure 2:

Atomistic profiles of interaction patterns of [U]₃ and GXG peptide fragments. (A) 3D plot showing average contacts fraction of RNA-peptide, RNA-RNA, and peptide-peptide pairs computed for twenty mixtures containing unique central residues. (B) Number of hydrogen bonds between RNA and GXG for all 20 residues. (C) Number of pi-stacking interactions between RNA and GXG for all 20 residues. (D) Number of hydrophobic interactions between RNA and GXG for all 20 residues. Interaction values in B-D are shown as interactions per fragment.

Figure 3:

Specific interactions of RNA and peptide. Radial distribution function (RDF) of residue side chains with different parts of RNA (phosphate, sugar, and base) for various types of residues: (A) ARG showing a stronger preferential interaction with phosphate via electrostatic interactions, (B) ASN showing a balanced interaction with the base, sugar, and phosphate of RNA, and (C) TYR showing a stronger preferential interaction with the base via pi-stacking. The peptides are depicted in yellow, while [U]₃ fragments are shown in blue, with their specific interactions highlighted.

Figure 4:

Water interactions with RNA and peptide clusters for all twenty residues. (A) The average number of water molecules per RNA (shaded region) and peptide fragment within the cluster. (B) The number of single water bridges (WB1) connecting RNA and GXG. (C) Number of double water bridges (WB2) connecting RNA and GXG. The schematic representation of (D) WB1 and (E) WB2 illustrates single and double water bridge formations between RNA and peptide fragments.