Dissecting the propensity of RIM1 subdomains to form phase condensates

Emerging evidence indicates that liquid-liquid phase separation (LLPS) may orchestrate synaptic vesicle and active zone (AZ) protein organization in presynaptic terminals via biomolecular condensates. A protein's LLPS propensity is determined by intrinsic factors such as structural disorder, amino acid composition, multivalent interactions, and regulation by post-translational modifications, highlighting the need for a systematic dissection of condensate-forming regions in a cellular context. Here, we systematically examined the condensation properties of the AZ protein RIM1 using an established cell-based overexpression assay and found that, while full-length RIM1 exhibits the highest condensation propensity in our assay, even fragments as short as ~ 250 amino acids form condensate-like assemblies in HEK cells. For all but four fragments, their propensity to be enriched in droplets over the intracellular solution quantitatively correlates with their length. We identified a minimal region that closely reproduces key full-length RIM1 droplet properties with regard to abundance per cell and volume, comprising the zinc finger domain, the intrinsically disordered region 1 (IDR1), the proline-rich motif 1 (PRM1), the PDZ- and C2A-domains, and two non-condensing regions, the sequences containing either PRM2 and IDR2 or PRM1 till the end of the C2A-domain. In contrast to in vitro data, RIM-BP2 promotes phase condensation of full-length RIM1 only at lower RIM1 expression levels and differentially impacts the phase condensation capacity of different RIM1 truncation mutants. Taken together, our results define the phase condensation behavior of different RIM1 sequence elements and show that enrichment of the protein in the phase condensates increases exponentially with the length of the fragments.