Par complex cluster formation mediated by phase separation

Cell polarity, crucial for various metazoan processes, involves the localized concentration of protein complexes at specific membrane regions. The Par (partitioning defective) complex, conserved from worms to mammals, is a key player in cell polarization, regulating asymmetric cell division (ACD), apical-basal polarity, cell migration, and neuronal polarization. Par complex dysfunction leads to developmental defects and tumorigenesis. The Par complex comprises Par3 (Bazooka in *Drosophila*), Par6, and atypical protein kinase C (aPKC), which interact through multiple domains. Par3 contains three PDZ domains mediating protein-protein interactions, while Par6 interacts with Par3 via its PDZ domains, and aPKC forms a stable subcomplex with Par6. aPKC activation phosphorylates Par3, leading to its dissociation from Par6/aPKC. While the assembly and functions of the Par complex are relatively well-understood, the mechanism by which Par proteins are recruited and highly concentrated at restricted membrane domains remains unclear. In *Drosophila* neuroblasts (NBs), Par proteins are concentrated apically during mitosis, forming a crescent. Previous work in *Drosophila* epithelia and *C. elegans* embryos showed that this enrichment involves Par3 oligomerization via its N-terminal domain (NTD), forming filaments. However, the process by which these filaments form dynamic clusters in vivo is poorly understood. This study investigates the mechanism behind the localized concentration of Par proteins, focusing on the possibility of liquid-liquid phase separation (LLPS).

The literature extensively documents the Par complex's role in establishing cell polarity and its involvement in various cellular processes. Studies have characterized the individual protein interactions within the complex, highlighting the roles of PDZ domains in Par3, PB1 domains in Par6 and aPKC, and the regulatory phosphorylation of Par3 by aPKC. However, the mechanism for the initial concentration and organization of the Par complex at the cell cortex remained elusive. Prior research suggested that Par3 oligomerization through its NTD plays a crucial role in forming higher-order structures, yet the dynamic nature of these structures and their involvement in the formation of the apical crescent was not fully understood. This study builds upon these findings by investigating the role of liquid-liquid phase separation (LLPS) in the formation of the Par complex clusters and their contributions to cell polarization.

The study used a multi-faceted approach combining in vivo imaging of *Drosophila* neuroblasts, in vitro biochemical assays, and heterologous cell-based experiments. **In vivo studies:** Endogenous Par protein localization in dividing *Drosophila* neuroblasts was analyzed using immunofluorescence microscopy at different cell cycle stages. The effect of 1,6-hexanediol, a phase separation inhibitor, was tested to determine the involvement of LLPS. Transgenic flies expressing tagged wild-type and mutant Par proteins were generated to investigate the functional role of phase separation in vivo. **In vitro studies:** Recombinant Par proteins were purified, and their ability to undergo liquid-liquid phase separation (LLPS) was assessed in vitro using both imaging (DIC and fluorescence microscopy) and sedimentation assays. The interaction between Par3 and Par6 was investigated using GST pull-down assays and fluorescence polarization, and the structure of the Par3 PDZ3-Par6β PBM complex was determined by X-ray crystallography. The role of specific domains (NTD, PDZ domains, PB1 domain, PBM) in promoting LLPS was assessed by performing in vitro phase separation assays with mutants lacking these domains. **Cell-based studies:** Mammalian Par3, Par6β, and aPKC were expressed individually or in combination in COS7 cells to investigate their ability to form cytoplasmic puncta. Fluorescence recovery after photobleaching (FRAP) analysis was used to study the dynamics of Par3 and Par6 in the puncta. The effects of various Par3 and Par6 mutations on puncta formation were assessed, and the role of PKC kinase activity was investigated by using both wild-type and mutant forms of aPKC. Quantitative analysis of puncta formation was performed. **Drosophila S2 cell culture:** Drosophila S2 cells were used to assess the expression and interaction of Baz (Drosophila Par3) and Par6 protein variants. **CRISPR-Cas9 mediated genome editing in Drosophila:** This technique was used to generate fly strains with GFP-tagged Baz (WT, ANTD, FUS-Baz) to further elucidate NTD's role in Par clustering at the physiological level. FRAP analysis was performed to study the dynamics of these variants. **Statistical analysis:** Appropriate statistical tests (one-way ANOVA with Tukey's multiple comparison test) were used to analyze the data.

The study's key findings are: 1. **Cell cycle-dependent Par protein clustering:** Endogenous Par proteins (Bazooka, Par6, aPKC) in *Drosophila* neuroblasts form distinct puncta on the apical cortex, with condensation increasing during metaphase and dispersing during anaphase. This clustering is sensitive to 1,6-hexanediol, a phase separation inhibitor. 2. **Spontaneous Par complex condensation:** In COS7 cells, mammalian Par3 and Par6β co-expression leads to the formation of highly concentrated, dynamic cytoplasmic puncta, indicative of LLPS. These puncta exhibit rapid exchange with the surrounding cytoplasm. The open conformation of Par3, achieved by 4N12 deletion or NTD expression, is essential for puncta formation, which is significantly enhanced by Par6β. 3. **In vitro LLPS:** Purified Par3N (NTD+PDZ1-3) undergoes LLPS in vitro, which is significantly enhanced by the addition of Par6β. LLPS is dependent on the concentration of Par3N/Par6β, exhibiting a greater LLPS capacity at lower concentrations. 4. **Specific Par3-Par6 interaction:** Par3 PDZ3 specifically interacts with Par6β PBM with high affinity (~1 µM). This interaction is significantly enhanced by Par3 NTD and Par6β PB1 oligomerization, increasing the valency of interactions and driving LLPS. The crystal structure of Par3 PDZ3 in complex with Par6β PBM peptide elucidates the binding mode. 5. **Multivalency-dependent LLPS:** Mutations in Par3 (deletion of PDZ3, NTD mutation) or Par6β (deletion of PB1 or PBM) impair Par3N/Par6β LLPS in vitro. These mutations have similar effects in living cells, confirming their impact on puncta formation. Par6α, unlike Par6β and Par6γ, lacks a conserved PBM and fails to promote LLPS. 6. **aPKC as an inactive client:** aPKC is recruited to Par3N/Par6β condensates in vitro and in vivo. However, activated aPKC phosphorylates Par3 (at CR3), causing condensate dispersal. The inactive aPKC within condensates is prevented from phosphorylating Par3. 7. **LLPS is essential for Par complex localization and function:** In *Drosophila* neuroblasts, mutations in Baz (Par3) or Par6 that disrupt LLPS impair their apical condensation and lead to defective asymmetric cell division and lineage development. Rescuing experiments in *baz* or *par6* mutants demonstrate the functional relevance of LLPS. 8. **Physiological significance of Baz LLPS:** Endogenous Baz levels in neuroblasts are too low for cytoplasmic LLPS. However, membrane attachment locally concentrates Baz, facilitating LLPS and apical condensation. Knock-in flies expressing GFP-tagged Baz variants (ΔNTD, FUS-Baz) further support this observation. Baz ΔNTD displays reduced apical condensation and smaller brain size compared to the WT. FUS-Baz, which retains LLPS ability to a certain extent, showed normal localization and brain size.

This study demonstrates that liquid-liquid phase separation (LLPS) is a crucial mechanism for the localized condensation of the Par complex during cell polarization. The multivalent interactions between Par3 and Par6, driven by the specific interaction of Par3 PDZ3 and Par6β PBM and amplified by the oligomerization of Par3 NTD and Par6β PB1, induce LLPS, concentrating the Par complex at the apical cortex. aPKC is recruited as an inactive client, but its activation leads to Par3 phosphorylation and condensate dispersal. The results highlight the importance of both LLPS and apical anchoring in regulating the localization of the Par complex. The study carefully addresses the potential artifacts of overexpression, comparing results from overexpression studies in COS7 cells and in vivo studies using both overexpression and knock-in approaches. Discrepancies between overexpression and knock-in studies are attributed to the sensitivity of LLPS to protein concentration, suggesting the importance of considering protein levels when studying phase separation in biological systems. The results imply that LLPS may be a general mechanism for the localized condensation of other polarity complexes, suggesting that this work provides valuable insight into the fundamental mechanisms governing cell polarity.

This study provides strong evidence that liquid-liquid phase separation (LLPS) is a critical mechanism for the localized cortical condensation of the Par complex, a key regulator of cell polarity. The multivalent interactions between Par3 and Par6, specifically involving Par3 PDZ3 and Par6β PBM, are essential for driving LLPS. The findings highlight the dynamic interplay between LLPS, protein interactions, and aPKC activity in regulating Par complex localization and function. Future studies could explore the roles of other regulators, such as Plk1, and investigate the broader applicability of the LLPS mechanism to other cell polarity complexes.

The study primarily focuses on *Drosophila* neuroblasts and mammalian cells, and the generalizability to other cell types and organisms needs to be further investigated. While the study shows a correlation between LLPS and the function of the Par complex, a more direct demonstration of the causal link between the two would be beneficial. The study also acknowledges the challenges of studying LLPS in vivo due to the inherent limitations of microscopy techniques.