Understanding the relationship between an amino acid sequence and its phase separation has important implications for analyzing cellular function, treating disease, and designing novel biomaterials. Several sequence features have been identified as drivers for protein liquid-liquid phase separation (LLPS), leading to the development of a “molecular grammar” for LLPS. In this work, we further probed how sequence modulates phase separation and the material properties of the resulting condensates. Specifically, we used a model intrinsically disordered polypeptide composed of an 8-residue repeat unit and performed systematic sequence manipulations targeting sequence features previously overlooked in the literature. We generated sequences with no charged residues, high net charge, no glycine residues, or devoid of aromatic or arginine residues. We report that all but one of the twelve variants we designed undergo LLPS, albeit to different extents, despite significant differences in composition. These results support the hypothesis that multiple interactions between diverse residue pairs work in tandem to drive phase separation. Molecular simulations paint a picture of underlying molecular details involving various atomic interactions mediated by not just a handful of residue types, but by most residues. We characterized the changes to inter-residue contacts in all the sequence variants, thereby developing a more complete understanding of the contributions of sequence features such as net charge, hydrophobicity, and aromaticity to phase separation. Further, we find that all condensates formed behave like viscous fluids, despite large differences in their viscosities. The results presented in this study significantly advance the current sequence-phase behavior and sequence-material properties relationships to help interpret, model, and design protein assembly.

Heterochromatin protein 1α (HP1α) is a crucial element of chromatin organization. It has been proposed that HP1α functions through liquid-liquid phase separation (LLPS), which allows it to compact chromatin into transcriptionally repressed heterochromatin regions. In vitro, HP1α can undergo phase separation upon phosphorylation of its N-terminus extension (NTE) and/or through interactions with DNA and chromatin. Here, we combine computational and experimental approaches to elucidate the molecular interactions that drive these processes. In phosphorylation-driven LLPS, HP1α can exchange intradimer hinge-NTE interactions with interdimer contacts, which also leads to a structural change from a compacted to an extended HP1α dimer conformation. This process can be enhanced by the presence of positively charged HP1α peptide ligands and disrupted by the addition of negatively charged or neutral peptides. In DNA-driven LLPS, both positively and negatively charged peptide ligands can perturb phase separation. Our findings demonstrate the importance of electrostatic interactions in HP1α LLPS where binding partners can modulate the overall charge of the droplets and screen or enhance hinge region interactions through specific and non-specific effects. Our study illuminates the complex molecular framework that can fine-tune the properties of HP1α and that can contribute to heterochromatin regulation and function.

A variety of membraneless organelles, often termed “biological condensates”, play an important role in the regulation of cellular processes such as gene transcription, translation, and protein quality control. On the basis of experimental and theoretical investigations, liquid–liquid phase separation (LLPS) has been proposed as a possible mechanism for the origin of biological condensates. LLPS requires multivalent macromolecules that template the formation of long-range, intermolecular interaction networks and results in the formation of condensates with defined composition and material properties. Multivalent interactions driving LLPS exhibit a wide range of modes from highly stereospecific to nonspecific and involve both folded and disordered regions. Multidomain proteins serve as suitable macromolecules for promoting phase separation and achieving disparate functions due to their potential for multivalent interactions and regulation. Here, we aim to highlight the influence of the domain architecture and interdomain interactions on the phase separation of multidomain protein condensates. First, the general principles underlying these interactions are illustrated on the basis of examples of multidomain proteins that are predominantly associated with nucleic acid binding and protein quality control and contain both folded and disordered regions. Next, the examples showcase how LLPS properties of folded and disordered regions can be leveraged to engineer multidomain constructs that form condensates with the desired assembly and functional properties. Finally, we highlight the need for improvements in coarse-grained computational models that can provide molecular-level insights into multidomain protein condensates in conjunction with experimental efforts.

TAR DNA-binding protein 43 (TDP-43) is involved in key processes in RNA metabolism and is frequently implicated in many neurodegenerative diseases, including amyotrophic lateral sclerosis and frontotemporal dementia. The prion-like, disordered C-terminal domain (CTD) of TDP-43 is aggregation-prone, can undergo liquid-liquid phase separation (LLPS) in isolation, and is critical for phase separation (PS) of the full-length protein under physiological conditions. While a short conserved helical region (CR, spanning residues 319-341) promotes oligomerization and is essential for LLPS, aromatic residues in the flanking disordered regions (QN-rich, IDR1/2) are also found to play a critical role in PS and aggregation. Compared with other phase-separating proteins, TDP-43 CTD has a notably distinct sequence composition including many aliphatic residues such as methionine and leucine. Aliphatic residues were previously suggested to modulate the apparent viscosity of the resulting phases, but their direct contribution toward CTD phase separation has been relatively ignored. Using multiscale simulations coupled with in vitro saturation concentration (c sat ) measurements, we identified the importance of aromatic residues while also suggesting an essential role for aliphatic methionine residues in promoting single-chain compaction and LLPS. Surprisingly, NMR experiments showed that transient interactions involving phenylalanine and methionine residues in the disordered flanking regions can directly enhance site-specific, CR-mediated intermolecular association. Overall, our work highlights an underappreciated mode of biomolecular recognition, wherein both transient and site-specific hydrophobic interactions act synergistically to drive the oligomerization and phase separation of a disordered, low-complexity domain.

Recent advances in residue-level coarse-grained (CG) computational models have enabled molecular-level insights into biological condensates of intrinsically disordered proteins (IDPs), shedding light on the sequence determinants of their phase separation. The existing CG models that treat protein chains as flexible molecules connected via harmonic bonds cannot populate common secondary-structure elements. Here, we present a CG dihedral angle potential between four neighboring beads centered at Cα atoms to faithfully capture the transient helical structures of IDPs. In order to parameterize and validate our new model, we propose Cα-based helix assignment rules based on dihedral angles that succeed in reproducing the atomistic helicity results of a polyalanine peptide and folded proteins. We then introduce sequence-dependent dihedral angle potential parameters (εd) and use experimentally available helical propensities of naturally occurring 20 amino acids to find their optimal values. The single-chain helical propensities from the CG simulations for commonly studied prion-like IDPs are in excellent agreement with the NMR-based α-helix fraction, demonstrating that the new HPS-SS model can accurately produce structural features of IDPs. Furthermore, this model can be easily implemented for large-scale assembly simulations due to its simplicity.