BPS2026 – Fluorescent protein dimer-based biosensor for quantitative hydrophobicity mapping of biomolecular condensates in live cells

Biomolecular condensates (BMCs), formed via liquid-liquid phase separation (LLPS), represent dense molecular assemblies governed by a balance of weak, multivalent interactions. Electrostatic attraction, π-π stacking, and hydrophobic forces collectively tune the phase boundary, yet the relative contribution of hydrophobicity to condensate microenvironments remains poorly quantified. A major limitation has been the absence of tools capable of probing hydrophobic interactions. We present the first genetically encoded hydrophobicity sensor. Our sensor enables high-resolution spatial mapping of hydrophobicity in living cells dynamically. The design leverages the moderate dimerization affinity between two fluorescent proteins, which undergo reversible association-dissociation depending on solvent hydrophobicity. The sensor is designed to partition into condensates and dynamically reports on their internal microenvironments . Reduced FRET efficiency signifies higher hydrophobicity. First, we directly probed sensor responsiveness to hydrophobicity in glycerol solutions, subjecting the sensor to a hydrophobic fluidic environment. As expected, we observed decrease (37%) in FRET efficiency in 75% glycerol solution compared to PBS. Application to poly-L-lysine/ATP condensates revealed preferential probe enrichment at condensate surface and a 16% efficiency decrease in the relatively dense interior, possibly due to altered solvent polarity at the condensate-buffer interface. In live-cell experiments with MDA-MB-231 breast cancer cells, the untagged sensor exhibited negligible phototoxicity and uniform cytoplasmic distribution, which is desirable feature for whole-cell hydrophobicity mapping. In summary, we demonstrated our sensor is suitable for quantifying hydrophobicity dynamically in living cells.