BPS2026 – Uncovering the domains of pericentrin that might enable its phase-separation propensity

Cells can create membraneless organelles through liquid-liquid phase separation (LLPS), a process which relies on multivalent associative interactions between proteins. Our previous simulation studies demonstrate that coiled-coil (CC) proteins can readily LLPS because of multivalent interactions between CC domains. These results suggest that CC domains might also play a role in promoting LLPS in natural proteins, but more experimental evidence is needed to understand how CC domains affect LLPS propensity. Recent studies suggest that some centrosomal proteins, such as human pericentrin (PCNT), undergo LLPS to help form a mature centrosome. It is unknown, however, what protein motifs in PCNT enable its supposed LLPS and assembly at the centrosome. Previous work has identified that a pericentrin fragment, PCNT(854-1960), is sufficient for LLPS in cells and recruitment of centrosomal-binding proteins; importantly, this fragment is predicted to contain many CC domains. We hypothesize that CC domains facilitate the assembly and apparent LLPS behavior of PCNT(854-1960). We present improved computational tools, including sequence-based predictors of CC interactions and molecular simulation models, to help identify possible interacting motifs in PCNT(854-1960) that enable its LLPS. We additionally present the results of deletion mutations studied via a yeast microscopy platform to determine regions in PCNT(854-1960) that might contribute to its LLPS propensity. Preliminary results suggest an important role for C-terminal CC domains in the assembly and localization of PCNT(854-1960) droplets. This work may have implications for the LLPS of proteins at the centrosome, as CC domains are common in proteins associated with the centrosome and LLPS propensity is predicted to be a shared feature of the broader centrosomal family.