Charged peptides enriched in aromatic residues decelerate condensate ageing driven by cross-β-sheet formation

Abstract

Biomolecular condensates play wide-ranging roles in cellular compartmentalization and biological processes. However, their transition from a functional liquid-like phase into a solid-like state-usually termed as condensate ageing-represents a hallmark associated with the onset of multiple neurodegenerative diseases. In this study, we design a computational pipeline to explore potential candidates, in the form of small peptides, to regulate ageing kinetics in biomolecular condensates. By combining equilibrium and non-equilibrium simulations of a sequence-dependent residue-resolution force field, we investigate the impact of peptide insertion-with different composition, patterning, and net charge-in the condensate phase diagram and ageing kinetics of archetypal proteins driving condensate ageing: TDP-43 and FUS. We reveal that small peptides composed of a specific balance of aromatic and charged residues can substantially decelerate ageing over an order of magnitude. The mechanism is controlled through condensate density reduction induced by peptide self-repulsive electrostatic interactions that specifically target protein regions prone to form cross-β-sheet fibrils. Our work proposes an efficient computational framework to rapidly scan the impact of small molecule insertion in condensate ageing and develop novel pathways for controlling phase transitions relevant to disease prevention.

Fig. 1. Condensate ageing process.

a Schematic representation of the accumulation of inter-protein β-sheets within a bulk condensate simulation. The zoomed images depict the nature of the initially intrinsically disordered protein regions (left) and the cross-β-sheet structures formed over time (right). b TDP-43 LCD and FUS LCD sequence representation, highlighting the charged and aromatic residues, as well as the low-complexity aromatic-rich kinked segments (LARKS) as indicated in the legend. c Storage (G') and shear loss (G") moduli as a function of frequency (ω) for non-aged (top) and aged (bottom) FUS LCD condensates. d Time-evolution of cross-β-sheet concentration for FUS LCD and TDP-43 LCD condensates. The continuous lines depict simulation data while the dashed lines represent fits to the data according to a secondary nucleation dominated mechanism (see Section II C for details on these fits). e Simulation of a FUS LCD condensate in bulk conditions containing small peptides that hinder the nucleation of cross-β-sheet structures.

Fig. 2. Stability bounds of biomolecular condensates upon peptide insertion.

a Normalized temperature (where T_{c,FUS LCD} = 322 K) vs. density phase diagram for FUS LCD in presence and absence of two peptides modulating its phase behaviour. Increasing concentrations of R₂₄ augment the condensate density and critical solution temperature, while (SYSYKRKK)₃ induces the opposite behaviour. Filled points represent direct measurements under coexistence conditions, while the empty point depicts the critical temperature for LLPS. The continuous lines are shown as a guide for the eye. At the bottom we represent typical snapshots of the NPT simulations employed to compute the phase diagram in presence of small peptides, where at temperatures below the critical one, the condensed phase remains stable at P = 0 bar, while above T_c, the system forms a low density phase. b Normalized critical temperature of FUS LCD mixtures with different peptide sequences (as indicated in the legend) as a function of the peptide concentration. c Normalized critical temperature (where T_{c,TDP-43 LCD} = 308 K) of TDP-43 LCD mixtures with different peptides as a function of the peptide concentration. The green and red shaded areas in (b, c) represent an increase and decrease in T_c, respectively. The error bars are obtained as the interval between the highest temperature at which the condensate is stable and the lowest one at which it is not, under P = 0 conditions.

Fig. 3. Ageing kinetics is controlled by the presence of small peptides.

a Relative cross-β-sheet concentration—determined from the amount of inter-protein β-sheets formed—as a function of time for FUS LCD condensates in absence vs. presence of different peptide sequences as indicated in the legend. Filled points represent measured β-sheet concentration from non-equilibrium simulations, while dashed lines depict fits to a secondary nucleation dominated kinetic model performed with Amylofit. We present only 3 representative trajectories that illustrate the peptide-mediated ageing deceleration. b Half-time (t_1/2) for the different FUS LCD condensate mixtures studied. t_1/2 is defined as the point where the relative β-sheet concentration value is halfway between the initial value and final plateau value. In the box plot representations, the average value is represented with a cross symbol, the boxes represent the 25−75 quartile range (with the horizontal bar being the median), and the whiskers indicate the highest and lowest values. c Relative cross-β-sheet concentration as a function of time for TDP-43 LCD condensates in absence vs. presence of different peptide sequences as indicated in the legend. Filled points and dashed curves represent the same quantities as in (a). d Half-time for the different TDP-43 LCD mixtures studied. The average value is represented with a cross symbol, the boxes represent the 25−75 quartile range (with the horizontal bar being the median), and the whiskers indicate the highest and lowest values.

Fig. 4. Intermolecular interactions determining the ageing kinetic modulation.

Intermolecular frequency contact maps for FUS LCD homotypic contacts in pure FUS LCD condensates (a) and FUS LCD/(SYGSYEGS)₃ mixtures (b). The pairwise residue-residue contact frequency is expressed in percentage in all panels. The homotypic LARKS-LARKS interactions are highlighted with a dashed box. c Intermolecular frequency contact maps of the different inserted peptides with the FUS LCD sequence in the condensate. The contacts are calculated between all residues of the inserted peptide and the FUS LCD sequence. Intermolecular frequency contact maps for TDP-43 LCD homotypic contacts in pure TDP-43 LCD condensates (d) and TDP-43 LCD/(SYGSYEGS)₃ mixtures (e). The homotypic LARKS-LARKS interactions are highlighted with a dashed box. f Intermolecular frequency contact maps of the different inserted peptides with TDP-43 LCD. In panels (c) and (f), the LARKS of each protein sequence are highlighted with dashed squares.