Expansion of Intrinsically Disordered Proteins Increases the Range of Stability of Liquid–Liquid Phase Separation

Abstract

Proteins containing intrinsically disordered regions (IDRs) are ubiquitous within biomolecular condensates, which are liquid-like compartments within cells formed through liquid–liquid phase separation (LLPS). The sequence of amino acids of a protein encodes its phase behaviour, not only by establishing the patterning and chemical nature (e.g., hydrophobic, polar, charged) of the various binding sites that facilitate multivalent interactions, but also by dictating the protein conformational dynamics. Besides behaving as random coils, IDRs can exhibit a wide-range of structural behaviours, including conformational switching, where they transition between alternate conformational ensembles. Using Molecular Dynamics simulations of a minimal coarse-grained model for IDRs, we show that the role of protein conformation has a non-trivial effect in the liquid–liquid phase behaviour of IDRs. When an IDR transitions to a conformational ensemble enriched in disordered extended states, LLPS is enhanced. In contrast, IDRs that switch to ensembles that preferentially sample more compact and structured states show inhibited LLPS. This occurs because extended and disordered protein conformations facilitate LLPS-stabilising multivalent protein–protein interactions by reducing steric hindrance; thereby, such conformations maximize the molecular connectivity of the condensed liquid network. Extended protein configurations promote phase separation regardless of whether LLPS is driven by homotypic and/or heterotypic protein–protein interactions. This study sheds light on the link between the dynamic conformational plasticity of IDRs and their liquid–liquid phase behaviour.

RNA modulation of transport properties and stability in phase separated condensates

“…This structural behavior allows proteins to maximize their number of intermolecular contacts, and thus, the droplet connectivity as recently shown in Ref. 108 . Phase-separation driven expansion for proteins undergoing homotypic LLPS has been observed for tau-IDP 130 using steady-state fluorescence measurements of pyrene and fluorescein-labeled tau-K18 proteins, a protein associated with Alzheimers disease 57 .…”

“…The specific composition and patterning of the amino acids along the sequence has a huge impact on the protein macroscopic phase behavior 81,97,121 . Moreover, beyond sequence, their conformational ensemble plays a crucial role not only in their ability to phase-separate, but also in the condensate structure 108,121,122,128,129 . 56,80 (including FUS-PLD), hnRNPA1 15 , TDP-43 42,80 and TDP-43-PLD 127 are depicted by intervals to consider concentration uncertainty.…”

“…Moreover, when comparing both TDP-43 variants P (R g ) distributions, we find almost identical protein ensembles, exhibiting the wild-type variant slightly more open conformations. Nonetheless, that small surplus of extended conformations which can enable higher amount of intermolecular contacts 113 is not enough to enhance LLPS as through the α − α helical interactions present in h-TDP-43 127 .…”

“…Simultaneously, a huge effort in developing different levels of coarse-grained (CG) potentials is being devoted, including mean field models [85][86][87][88][89] , lattice-based simulations [90][91][92][93] , minimal models [94][95][96][97][98] , and sequence-dependent models 83,[99][100][101] . By retaining the specific physico-chemical features of proteins, DNA and RNA, while averaging out others for computational efficiency, CG models have been widely used to elucidate key factors behind LLPS and their dependency on the protein length 102,103 , amino acid sequence 83,99,100,104,105 , multivalency 90,[106][107][108][109][110][111] , conformational flexibility 112,113 and multicomponent composition [114][115][116][117][118] . Nonetheless, regarding the role of RNA in LLPS and the molecular driving forces contributing to the formation of RNA-RBP condensates, further work is required 119 .…”

Valency and Binding Affinity Variations Can Regulate the Multilayered Organization of Protein Condensates with Many Components

“…The highest attainable resolution to investigate the process of protein condensation with computer simulations is now approaching that of atomistic models [ 56 , 57 , 58 ]. However, since LLPS is a collective phenomenon that involves thousands of interacting biomolecules [ 41 , 42 ], coarse-grained models with different levels of resolution are still the most effective potentials to decipher the molecular and biophysical forces driving protein demixing and self-assembly [ 34 , 53 , 58 , 59 , 60 , 61 , 62 , 63 , 64 , 65 , 66 , 67 , 68 , 69 , 70 , 71 , 72 , 73 , 74 ]. Indeed, coarse-grained simulation studies have been useful at linking a wide-range of protein characteristics to the modulation of their phase behavior—e.g., valency [ 39 , 65 , 75 , 76 , 77 , 78 ], topological distribution of binding sites [ 53 ], amino acid sequence and patterning [ 79 , 80 , 81 , 82 ], IDR conformation [ 69 , 83 ]—and the emergence of multilayered condensate organizations [ 15 , 33 ].…”