Location and concentration of aromatic-rich segments dictates the percolating inter-molecular network and viscoelastic properties of ageing condensates

Abstract

Maturation of functional liquid-like biomolecular condensates into solid-like aggregates has been linked to the onset of several neurodegenerative disorders. Low-complexity aromatic-rich kinked segments (LARKS) contained in numerous RNA-binding proteins can promote the aggregation process by forming inter-protein beta-sheet fibrils that accumulate over time - ultimately driving the liquid-to-solid transition of the condensate. Here, we combine atomistic molecular dynamics simulations with sequence-dependent coarse-grained models of various resolutions to investigate the role of LARKS abundance and position within the amino acid sequence in the maturation of biomolecular condensates. We find that the location of the LARKS motifs along the protein sequence crucially modulates the rate of cross-beta sheet transitions and the associated loss of liquid-like behaviour over time. Our simulations show that shifting the location of the LARKS in fused in sarcoma (FUS) protein towards its center slows down condensate aggregation. More strikingly, our simulations further predict that adding RNA to FUS with re-located LARKS fully inhibits the accumulation of beta-sheet fibrils, maintaining functional liquid-like behaviour without ageing.

Time-dependent material properties of ageing biomolecular condensates from different viscoelasticity measurements in molecular dynamics simulations

“…Therefore, every 100 simulation timesteps, our algorithm evaluates whether the conditions around each fully disordered LARKS are favourable to undergo an ‘effective’ disorder-to-order β -sheet transition. In our model, the structural transition is recapitulated by enhancing the interaction strength of four LARKS-LARKS peptides based on the results of our atomistic potential-of-mean force simulations 73,75,76 . In the atomistic simulations, we estimate the binding free energy difference between disordered interacting LARKS peptides, and interacting LARKS peptides that are forming interprotein cross β -sheets.…”

“…In this section, we investigate the progressive rigidification of phase-separated condensates due to the gradual accumulation of inter-protein structural transitions over time 13,114,128 . It has been recently shown, both experimentally and computationally, that the interaction landscape of proteins can be significantly transformed by structural transitions 50,73,75,76,114 . The low complexity domains (LCD) of many naturally occurring phase-separating proteins-including FUS 49 , TDP-43 50,129 , or hnRNPA1 13,128 among many others 114,175 contain short regions termed Low-complexity Aromatic-Rich Kinked Segments (LARKS), which are prone to forming inter-protein β-sheets in environments of high protein concentration 13,116,176 .…”

“…In the atomistic simulations, we estimate the binding free energy difference between disordered interacting LARKS peptides, and interacting LARKS peptides that are forming interprotein cross β-sheets. We estimate these changes for the three FUS LARKS within the LCD ( 37 SYSGYS 42 , 54 SYSSYGQS 61 , and 77 STGGYG 82 49,73 ), the A-LCD-hnRNPA1 LARKS ( 58 GYNGFG 63 75,114 ), and that of TDP-43 76 ( 58 NFGAFS 63 ; also located in the protein LCD 50 ). Therefore, by employing the HPS-Cation-π model coupled to our dynamical ageing algorithm, we can effectively investigate the viscoelastic behaviour of the FUS-LCD, A-LCD-hnRNPA1, and TDP-43-LCD condensates prior and post-maturation.…”

“…We find an exceptional agreement between G ′ and G ′′ as a function of frequency using both OS and SSRMI techniques for FUS-LCD, A-LCD-hnRNPA1, and TDP-43-LCD condensates. Then, we activate the ageing dynamical algorithm 73,75,76 and perform 0.4 µs simulations under bulk condensate conditions to allow inter-protein structural transitions to accumulate over time. In Fig.…”

“…In that respect, computer simulations are a powerful tool to uncover the molecular mechanisms that explain the changes in viscosity within biomolecular condensates over time 31,51,63,[73][74][75][76] . From atomistic force fields [77][78][79][80][81][82] to coarse-grained (CG) models [83][84][85][86][87][88][89][90][91][92] , including lattice-based simulations [93][94][95] and mean-field theory [96][97][98] , computer simulations have significantly contributed to elucidating factors behind condensate phaseseparation such as protein and RNA length [99][100][101] , amino acid patterning 91,[102][103][104][105] , multivalency 35,[106][107][108][109] , conformational flexibility 89,110 or multicomponent composition 90,111,…”