Intrinsically disordered proteins phase-separate into liquid condensates at concentrations above a critical threshold. The critical temperature depends on the amino acid sequence, but which residues matter? Standard models use sticker-spacer frameworks: some residues (stickers) drive phase separation through attractive interactions, others (spacers) just occupy space. The sequence dependence comes from how stickers are arranged along the chain.
Samanta et al. (arXiv:2603.12534) show that solvent accessibility — how exposed each residue is — reshapes the sequence dependence. A hydrophobic residue buried in a collapsed chain conformation contributes differently to phase separation than the same residue exposed on the chain surface. The contribution depends not just on what the residue is but on where it sits in the 3D ensemble of chain conformations.
The mechanism: disordered proteins fluctuate rapidly between conformations. In each conformation, different residues are exposed to solvent and available for intermolecular contacts. The effective interaction between two chains depends on which residues happen to be exposed when they encounter each other. Residues that are frequently exposed contribute more to phase separation; frequently buried residues contribute less, even if they're chemically identical.
This introduces a conformational averaging effect. The critical temperature isn't determined by the raw sequence composition (how many hydrophobic residues) but by the time-averaged exposure profile (how often each hydrophobic residue faces outward). Two sequences with identical composition but different patterning will have different exposure profiles and therefore different critical temperatures.
For protein design, the implication is that you can tune phase separation by controlling residue accessibility through sequence patterning — not just sticker density but sticker placement relative to chain geometry. The three-dimensional ensemble is a hidden variable in the sequence-to-phase-behavior mapping.