Figure 4. Sequence drivers govern PLCD thermoresponsive condensation.

Molecular dynamics simulations are performed using the Mpipi-T model to simulate single chain and multiple chain protein systems. (a) For single chain systems, the Flory scaling exponent (v) is calculated from the simulated protein’s radius of gyration to determine the transition from coil (expanded) to globule (compact) state. The dashed gray line indicates v = 0.5, which is the coil-to-globule transition. (b) For multiple chain systems, direct coexistence simulations are performed to obtain the density profile along the long axis of the simulation box at various temperatures, from which the dense-phase density is extracted. The results from the simulations show that PLCDs sufficient for condensation display (c) lower coil-to-globule transition temperatures in single-chain simulations and (d) higher dense phase densities under heat stress in multi-chain simulations. Bars represent the mean across proteins, and individual proteins are shown as points. These behaviors reflect underlying sequence heuristics: (e) full-length proteins do not differ significantly in residue composition based on PLCD sufficiency, but (f) sufficient PLCDs are enriched in aliphatic and aromatic residues and depleted in polar residues compared to insufficient PLCDs. Bars represent the mean residue fraction across sequences, and error bars represent the standard deviation.