UBQLN2 450−624 oligomerizes and undergoes temperature-responsive liquid−liquid phase transitions following a closed-loop temperature−concentration phase diagram. We recently showed that disease-linked mutations to UBQLN2 450−624 impart highly varying effects to its phase behavior, ranging from little change to significant decrease of saturation concentration and formation of gels and aggregates. However, how single mutations lead to these properties is unknown. Here, we use UBQLN2 450−624 as a model system to study the sequence determinants of phase separation. We hypothesized that UBQLN2 450−624 regions previously identified to promote its oligomerization are the "stickers" that drive interchain interactions and phase separation. We systematically investigated how phase behavior is affected by all 19 possible single amino acid substitutions at three sticker and two "spacer" (sequences separating stickers) positions. Overall, substitutions to stickers, but not spacers, substantially altered the shape of the phase diagram. Within the sticker regions, increasing hydrophobicity decreased saturation concentrations at low temperatures and enhanced oligomerization propensity and viscoelasticity of the dense phase. Conversely, substitutions to acidic residues at all positions greatly increased saturation concentrations. Our data demonstrate that single amino acid substitutions follow a molecular code to tune phase transition behavior of biopolymers.

Liquid‐liquid phase separation (LLPS) is hypothesized to be the dominant mechanism that underlies the formation of membraneless organelles, including stress granules formed under cellular stress containing sequestered RNA and proteins. Our lab has recently shown that Ubiquilin‐2 (UBQLN2), a protein involved in protein quality control through proteasomal degradation and autophagy pathways, is recruited to stress granules in vivo and undergoes LLPS in vitro under physiological conditions. Using NMR spectroscopy, we identified both intrinsically‐disordered and folded regions that are involved in UBQLN2 self‐association that also mediate UBQLN2 LLPS. One of these regions includes the proline‐rich (Pxx) segment, where known disease‐linked mutations have been shown to cause 1–2% of familial ALS or ALS/dementia cases. Using size‐exclusion chromatography, NMR spectroscopy, and microscopy, we demonstrated that Pxx mutations modulate UBQLN2 self‐association and phase separation in vitro. Pxx mutations at P497, P506 and P525 affected droplet morphology and dynamics to significantly different extents. Our data suggested that increased hydrophobicity of the amino acid promoted UBQLN2 LLPS and lowered the phase transition temperature at which UBQLN2 LLPS is first observed. To systematically investigate the molecular basis for how amino acid composition affects UBQLN2 phase separation, we generated all 19 possible amino acid replacements at residues P497, P506, P525, and V538. Our data indicate that the phase behavior of UBQLN2 is significantly modulated by mutations at positions 497 and 506, but weakly modulated by mutations at positions 525 and 538. These results suggest that positions 497 and 506 are ‘stickers’ whereas positions 525 and 538 are ‘spacers’ using the language of associative polymers. Consistent with our data above, hydrophobic amino acid substitutions promoted UBQLN2 LLPS. Our experiments suggest that UBQLN2 may be used as a model system to understand the sequence determinants of phase separation. Hydrophobic mutations modulate UBQLN2 oligomerization and LLPS, and potentially alter material properties of UBQLN2‐containing biomolecular condensates in the cell, promoting disease states.Support or Funding InformationC.A.C. graciously acknowledges funding from the ALS Association via grants 17‐IIP‐369 and 18‐IIP‐400, as well as from the National Science Foundation (CAREER award 1750462).This abstract is from the Experimental Biology 2019 Meeting. There is no full text article associated with this abstract published in The FASEB Journal.