Figures and data
Simulation slab box containing MUT-16 FFR chains, each represented by a distinct color.
The protein chains are solvated in water (cyan) with ions, Na+ (red) and Cl− (green). Representative specific interactions observed in the simulation are highlighted, including cation-π interactions between the guanidinium group of Arg and the phenol group of Tyr, π-π stacking between phenol groups of Tyr residues, salt bridges between the guanidinium group of Arg and the side-chain carboxylate group of Glu, and hydrogen bonds between the hydroxyl group of Tyr and the carboxylate group of Glu.
Sequence and compositional similarity of human IDRs to MUT-16 FFR and FUS LCD.
A. Box-and-whisker plots showing the distribution of EBAmin similarity scores for human IDRs, evaluated with respect to the MUT-16 FFR (blue) and the FUS LCD (orange). The median is indicated by the horizontal line and the mean by a green triangle; outliers are shown as translucent pink stars. B. Box-and-whisker plots showing the distribution of RMSE values quantifying compositional similarity of human IDRs relative to the MUT-16 FFR (blue) and the FUS LCD (orange). C. Box-and-whisker plots showing residue abundance distributions across the same set of human IDRs, with residue abundances for the MUT-16 FFR (blue) and the FUS LCD (orange) overlaid.
MUT-16 FFR is a low-temperature condensation (UCST) protein.
(A) Schematic representation of the experimental workflow for the inPhase method. In the first step, serial dilutions of MUT-16 fragments are prepared, and phase separation is induced by cleaving the MBP tag using 3C protease. In the second step, the reaction mixture is immediately encapsulated into oil droplets. In the third step, the emulsion droplets are loaded onto a temperature-controlled stage between two Parafilm strips to spatially compartmentalize each condition (e.g., different protein concentrations). In the final step, the droplets are imaged, and the volume fraction (Vcon/Vtot) is determined from the condensate and droplet volumes. (B) Representative single-plane images of emulsion droplets containing the MUT-16 773–944 aa (FFR) fragment at different concentrations (10 µM, 15 µM, 20 µM) and temperatures (20◦C, 30◦C, 40◦C). Left panels show bright-field images, and right panels show fluorescence images. Scale bar: 100 µm. White boxes in the 20 µM panel indicate regions shown in the zoomed-in images below (scale bar: 20 µm). (C) Quantification of temperature effects on condensates. The volume fraction was normalized to the value at 20◦C for each protein concentration. Each experiment comprises data points from multiple emulsion droplets. Bars represent mean ± SD (n = 47 for 10 µM, n = 138 for 15 µM, and n = 116 for 20 µM)
Residue-resolved contact frequencies and persistence of side-chain interactions in the MUT-16 FFR condensate.
A, B. Residue resolved contact maps derived from atomistic simulations of the MUT-16 FFR condensate, shown in both unnormalized and abundance-normalized heatmaps. The maps report contact frequencies for side chain-side chain (SC:SC) interactions between all amino acid pairs, where unnormalized frequencies reflect the raw occurrence of contacts, and abundance-normalized frequencies account for sequence compo-sition effects to highlight intrinsically preferred interactions. C. Box-and-whisker plots of lifetimes for SC:SC contacts between amino-acid pairs. The central line indicates the me-dian, the green triangle denotes the mean, and high-persistence outliers beyond the upper tail are shown as translucent pink stars.
Representative persistence time distributions and interaction dynamics of side-chain contacts.
A. Probability density distributions of lifetimes for selected amino-acid pairs. B. Distance time series for a representative Arg–Asp salt bridge; shaded regions indicate bound states defined by a 4 Å donor–acceptor cutoff. C. Distance time series for a representative Tyr–Phe π–π interaction; bound states are defined by a 5 Å center-of-mass cutoff between aromatic rings. D. Distance time series for a representative Thr–Asp hydrogen bond; bound states are defined by a 3.5 Å donor–acceptor cutoff. Representative snapshots corresponding to selected time points are shown above each time series.
Distinct noncovalent interaction modes and their persistence in the MUT-16 FFR condensate.
A,B. Hydrogen-bond interactions in the MUT-16 FFR condensate: (A) heatmap of contact frequencies and (B) mean lifetimes. C,D. Salt-bridge interactions: (C) fraction of residue-residue contacts forming salt bridges, indicating the proportion of all salt bridges in the system contributed by each residue pair; and (D) box-and-whisker distri-butions of salt-bridge lifetimes. E,F. Cation–π interactions: (E) fraction of contacts and (F) persistence-time distributions. G,H. π–π stacking interactions: (G) fraction of residue-residue contacts, indicating the contribution of each residue pair and (H) persistence-time distributions.
Residue-resolved ion interactions within the phase-separated MUT-16 FFR con-densate.
(A,B) Average number of Na+ ions interacting with amino-acid side chains (A) and backbones (B), highlighting preferential association of sodium with negatively charged and polar residues. (C,D) Average number of Cl− ions interacting with amino-acid side chains (C) and backbones (D), showing substantially weaker interactions relative to Na+. Ion–residue interactions were defined using distance cutoffs of 4 Å for Na+ and 5 Å for Cl−. Values represent averages over 10 independent replicas, with error bars indicating the stan-dard error of the mean.
Ion-mediated interactions within the MUT-16 FFR condensate.
(A) Box-and-whisker plots showing the lifetimes of ion–residue side-chain interactions. Distributions for Na+ interactions are shown in the upper panel (blue), whereas distributions for Cl− interac-tions are shown in the lower panel (red). (B) Frequency of ion-bridged interactions between similarly charged residue pairs, normalized by amino-acid pair abundance. (C) Represen-tative snapshots illustrating Cl−- and Na+-mediated bridging interactions between Arg–Arg and Glu–Asp side chains, respectively.
Interactions of water molecules within the MUT-16 FFR condensate.
A. Density profiles of protein, water, and acidic, basic, and polar residues, averaged over ten independent simulation replicas. Condensate interface indicated by the shaded areas. B. Water–amino acid contacts in the condensate interior and at the condensate–solvent interface. C. Total number of water-bridged contacts among basic, acidic, and polar residue pairs. D. Water-bridged contact frequencies normalized by residue-pair abundance.
Cutoff parameters used for interaction and persistence analyses.
Residue abundance profiles of the MUT-16 FFR (blue) and the FUS LCD (or-ange).
For comparison, residue abundances of representative human IDRs identified as most similar to the MUT-16 FFR based on embedding-based alignment (EBA) similarity (purple) and root-mean-square error (RMSE) (green) are overlaid as scatter points.
Relationship between contact frequency, persistence, and residue abundance in the MUT-16 FFR condensate.
A. Heat map showing the total number of binding events observed for each residue-residue pair that were included in the statistical analysis of contact distributions and mean lifetimes. The color scale represents the number of events on a logarithmic scale. B. Heat map of the mean lifetimes of side-chain-mediated interactions, highlighting the relative stability and typical lifetimes of residue-residue contacts within the condensate. C. Correlation between unnormalised contact frequency and mean persistence time, illustrating how frequently formed contacts relate to their temporal stability. D. Correlation between abundance-normalised contact frequency and mean persistence time, isolating intrinsic interaction stability from effects arising due to residue abundance.
Heat map showing the median lifetimes of side-chain-mediated interactions, em-phasizing the typical (central) persistence of residue–residue contacts within the condensate.
By focusing on the median rather than the mean, the map captures the representative inter-action timescales while minimizing the influence of rare, long-lived events, thereby providing a robust measure of relative interaction stability across residue pairs.
A. Cumulative lifetime distribution for Arg 875-Glu 869 and Arg 904 and Asp 917, illustrating the overall distribution of contact persistence. B. Median lifetimes of individual Arg-Glu residue pairs in the MUT-16 FFR sequence, highlighting the typical interaction timescales. C. Mean lifetimes of specific Arg-Glu residue pairs, reflecting the influence of longer-lived interactions. D. Number of binding events observed for Arg-Glu pairs in the MUT-16 FFR sequence, indicating interaction frequency. E. Median lifetimes of Arg-Asp residue pairs, providing a robust measure of typical contact stability. F. Mean lifetimes of specific Arg-Asp residue pairs, capturing contributions from long-lived contacts. G. Number of binding events for Arg-Asp pairs in the MUT-16 FFR sequence, reflecting the occurrence of these interactions across the trajectory.
Prevalence of distinct noncovalent interaction types in the MUT-16 FFR conden-sate.
A. Average number of hydrogen bonds formed between backbone-backbone (BB:BB), backbone-side chain (BB:SC), and side chain-side chain (SC:SC) contacts in the MUT-16 FFR condensate. Values are averaged over all simulation frames and across 10 independent replica trajectories, with error bars representing the standard error of the mean. B. Heat map of unnormalised contact frequencies for residue pairs capable of forming hydrogen bonds, reflecting the overall prevalence of hydrogen-bonded interactions. C. Representative snap-shot of a hydrogen bond formed between Asn and Gln residues, illustrating a typical polar side-chain interaction observed in the condensate. D. Bar graph showing the unnormalised fraction of contacts for residue pairs that form salt bridges, highlighting electrostatically driven interactions. E. Bar graph showing the unnormalised fraction of contacts for residue pairs that form cation–π interactions. F. Bar graph showing the unnormalised fraction of contacts for residue pairs that form π–π stacking interactions, emphasising the contribution of aromatic interactions to condensate organisation.
Representative lifetime of a π-π stacking interaction between Tyr and Phe residues, showing transitions between bound (pink) and unbound (black) states.
Binding is identified using two criteria: a distance-only cutoff and a combined distance-and-angular cutoff, demonstrating how the inclusion of orientational constraints refines the detection of stacked configurations.
Density profiles of Na+ and Cl− ions compared with acidic and basic residues along the condensate slab normal.
Residue-resolved ion association within the MUT-16 FFR condensate.
A. Average number of Na+ ions interacting with all atoms of each residue (backbone + side chain). B. Average number of Cl− ions interacting with all atoms of each residue (backbone + side chain).
Representative ion-mediated interaction motifs observed in the MUT-16 FFR condensate.
(A) Na+-mediated bridging interaction between the side chain and backbone of an Asp residue. (B) Na+-mediated bidentate interaction involving both the side chain and backbone of a Thr residue. (C) Interaction between Lys residues and Na+ facilitated by a bridging Cl− ion. (D) Time series depicting representative Asp–Na+–Asp and Glu–Na+–Glu bridging interactions. Distances are measured between the negatively charged oxygen atoms of Asp and Glu side chains and the Na+ ion.
Visual representation and time series of multiple Na+-mediated bridging event between two Asp residues, involving both side-chain carboxylate oxygen atoms and backbone oxygen atoms.
Shaded areas indicate the interface between the dense and dilute phases.
