Back

Local aromatic interactions define temperature sensitivity of phase separation in an intrinsically disordered protein.

Ohhashi, Y.; Nishinami, S.; Maruyama, Y.; Fukuyama, M.; Shiraki, K.; Chatani, E.; Taguchi, H.

2026-05-08 biochemistry
10.64898/2026.05.07.723405 bioRxiv
Show abstract

Liquid-liquid phase separation (LLPS) of intrinsically disordered proteins is highly sensitive to environmental conditions, yet the molecular basis of sharp temperature responsiveness remains poorly understood. Here, we investigate a sequence-encoded mechanism underlying the pronounced temperature sensitivity of phase separation using Sup35NM, the intrinsically disordered domain of the yeast prion Sup35. We show that a tyrosine-rich local structural region within this domain encodes strong temperature responsiveness of droplet formation. Mutational analyses reveal that tyrosine residues mediate both intramolecular interactions that stabilize local structure and intermolecular interactions required for LLPS. Substitution of the tyrosine residues with alanine disrupts local structure and weakens intermolecular interactions, thereby diminishing temperature sensitivity. In contrast, substitution with phenylalanine promotes rapid droplet gelation, abolishes internal fluidity, suppresses amyloid formation, and confers resistance to temperature-induced dissolution. Based on these findings, we propose a molecular model in which finely tuned, moderately weak aromatic interactions among tyrosines enable reversible local compaction that is sensitive to temperature, thereby generating a sharp phase transition. These results suggest that amino acid sequences encode not only phase separation propensity but also the sensitivity of condensates to environmental perturbations, providing a framework for understanding how intrinsically disordered proteins act as molecular sensors of cellular conditions.

Matching journals

The top 3 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.