Cell cycle-dependent palmitoylation of protocadherin 7 by ZDHHC5 promotes successful cytokinesis

Cell division is central to life, driving many vital cellular events such as proliferation, propagation, development and regeneration (Rieder and Khodjakov, 2003). As the cell progresses into mitosis, its morphology and surface undergo dramatic reshaping. Adherent cells transiently round up during mitosis to fulfill the geometric demand of cell division for accurate chromosome segregation (Lancaster et al., 2013; Stewart et al., 2011b). As chromosomes segregate during anaphase, the contractile ring machinery assembles to physically divide cells into two, and the daughter cells spread back to regain their interphase morphology during cytokinesis (Ramkumar and Baum, 2016). Morphological changes during mitosis and cytokinesis are primarily driven by the reorganization of the actomyosin cytoskeleton network and adhesive systems (Cramer and Mitchison, 1997; Eggert et al., 2006; Rosenblatt, 2008).

At the onset of mitosis, ECT2 activates RhoA at the plasma membrane, which activates its downstream effectors, formins and Rho-associated protein kinase (ROCK) (Ramkumar and Baum, 2016). A member of formins, mDia1 localizes to the cell cortex during mitosis and promotes cortical actin nucleation and polymerization (Bovellan et al., 2014). ROCK is a serine/threonine kinase that activates myosin through the phosphorylation of its two main substrates. Phosphorylation of the myosin-binding subunit of myosin phosphatase (MYPT1) by ROCK inactivates myosin phosphatase (Kimura et al., 1996). ROCK can also directly phosphorylate the myosin regulatory light chain at Ser19 (Amano et al., 1996), which results in the activation of myosin ATPase that stimulates actin crosslinking and actomyosin contractility (Narumiya et al., 2009).

As a cell rounds up, myosin II progressively accumulates at the cell cortex and the amount of myosin at the cortex is correlated with the rounding pressure (Ramanathan et al., 2015). RNAi-based depletion of cortical myosin II substantially impairs the rounding pressure of mitotic cells (Toyoda et al., 2017). Thus, myosin II plays a fundamental role in cortical reorganization during mitosis by promoting cortical tension (Taubenberger et al., 2020). During cytokinesis, communication between the midzone and actin cortex through Rho signaling drives the assembly of the actomyosin contractile ring machinery between segregating chromosomes (Eggert et al., 2006; Wadsworth, 2021). During cytokinesis, the actin cytoskeleton and the ezrin/radixin/moesin (ERM) family, a linker between the plasma membrane and the underlying cortical actin cytoskeleton (Tsukita and Yonemura, 1997), have been shown to play roles in the translocation of plasma membrane proteins to the cleavage furrow through their interaction with the cytoplasmic domains of plasma membrane proteins (Yoshigaki, 1997; Yonemura et al., 1993; Uretmen Kagiali et al., 2020).

Protocadherins (PCDHs) are the largest subgroup of cell surface proteins in the cadherin superfamily (Morishita and Yagi, 2007; Nollet et al., 2000). Although identified as adhesion molecules, the adhesive roles of PCDHs are context dependent. PCDHs can form homophilic and heterophilic interactions that regulate cell–cell adhesion and downstream signaling events in embryonic and adult tissues (Bradley et al., 1998; Chen et al., 2007; Kahr et al., 2013; Kim et al., 1998; Kuroda et al., 2002; Tai et al., 2010). In our previous study (Ozlu et al., 2015), we investigated how cell surface proteins change during cell division and compared the cell surface proteome of interphase and mitotic cells. Our proteomic analysis identified PCDH7 as one of the proteins that is enriched at the mitotic cell surface and retraction fibers. Knockdown of PCDH7 using siRNAs caused a decrease in the mitotic rounding pressure albeit not as strong as myosin II (Ozlu et al., 2015).

Here, we aimed to unravel the underlying mechanism of the cell cycle-dependent localization of PCDH7 and its role in cell division. PCDH7 localizes to the cell surface at the onset of mitosis and concentrates at the cleavage furrow during cytokinesis. A palmitoyltransferase, ZDHHC5, was identified as a proximal interactor of PCDH7. ZDHHC5 interacts with PCDH7, and it targets PCDH7 to the mitotic cortex and cleavage furrow. Cell cycle-dependent localization of PCDH7 depends on palmitoylation and requires ZDHHC5 catalytic activity. PCDH7 depletion reduced active RhoA and myosin II levels at the cleavage furrow and increased the multinucleation rate. We propose that spatiotemporal regulation of PCDH7 through palmitoylation by ZDHHC5 promotes successful cytokinesis in mammalian cells.

Our previous study revealed that PCDH7 gets enriched at the cell surface as the cell progresses into mitosis (Ozlu et al., 2015). To probe the spatiotemporal regulation of PCDH7, we analyzed its subcellular localization throughout the cell cycle. For this, we used the PCDH7-GFP-BAC cell line, in which GFP-tagged PCDH7 is expressed under its own promoter using bacterial artificial chromosome (BAC) transgenomics in HeLa Kyoto cells (Poser et al., 2008). Immunostaining of the PCDH7-GFP-BAC cells showed that PCDH7 localized to the cell–cell contacts during interphase (Fig. 1A, top), was enriched at the cell surface and retraction fibers at the onset of mitosis (Fig. 1A, middle), and concentrated at the contractile ring and to a narrow zone within the furrow (Normand and King, 2010) (Fig. 1A, bottom). We observed a similar cleavage furrow (0 min) and equatorial accumulation (3–12 min) of PCDH7 in HeLa S3 cells expressing PCDH7::GFP (Fig. 1B; Movie 1). We observed that the localization of PCDH7 at the cleavage furrow was slightly asymmetric (Fig. 1B, 0 min; Fig. S1A). The cell cycle-dependent localization of PCDH7 prompted us to test whether cytokinesis-specific localization is due to cleavage furrow formation but not due to freshly forming cell–cell contact regions of two daughter cells. To this end, we utilized the monopolar cytokinesis approach (Hu et al., 2008; Karayel et al., 2018; Ozlü et al., 2010), in which cells are arrested in monopolar mitosis by using the kinesin-5 inhibitor S-trityl-L-cysteine (STC), followed by induction of monopolar cytokinesis with the CDK inhibitor purvalanol A. The chromosomes stay around one pole, whereas the plasma membrane forms a bud-like extension where the cleavage furrow is greatly expanded without a cell–cell contact region (Fig. S1B). The cleavage furrow proteins, myosin II, anillin, mDia and RhoA localize to the bud-like extension and it biochemically mimics the cleavage furrow of bipolar cytokinetic cells (Hu et al., 2008; Karayel et al., 2018; Ozlü et al., 2010). PCDH7 was enriched at the plasma membrane of monopolar mitotic cells (Fig. 1C, top), and in monopolar cytokinetic cells, it was concentrated at the budding site, which corresponds to the cleavage furrow-like cortex (Fig. 1C, bottom). These results suggest that the cleavage furrow localization of PCDH7 is independent of its cell–cell contact localization.

To examine the mitosis-dependent translocation of PCDH7, we performed fluorescence recovery after photobleaching (FRAP) experiments. The mobility of PCDH7::GFP in interphase and mitotic cells was compared using the PCDH7-GFP-BAC cell line. The GFP signal within the region of interest (ROI) at the plasma membrane of both mitotic (Fig. S1C, top) and interphase (Fig. S1C, bottom) cells was photobleached using a focused laser beam. The fluorescence recovery within the ROI was analyzed (Fig. 1D) using a double normalization algorithm (Phair et al., 2004) and fixed-sized ROI areas (Kappel and Eils, 2004). After photobleaching, a significant difference was not observed between the recovery levels of PCDH7::GFP in cell–cell contacts during interphase (average recovery 46.6%) and at the plasma membrane during mitosis (average recovery 45%), as shown in Fig. 1E (left). However, PCDH7::GFP recovered significantly faster at the mitotic cell surface with an average FRAP half time of around 62 s, compared to its recovery at the cell–cell contacts in interphase, where the average FRAP half time was around 109 s (Fig. 1E, right). We conclude that PCDH7 is much more dynamic and shows a high turnover rate at the mitotic cell surface in comparison to the cell–cell contact regions.

Next, we asked about the function of PCDH7 in cell division. Our previous investigation by siRNA-based depletion of PCDH7 revealed that PCDH7 is required for the development of mitotic rounding pressure at the onset of mitosis (Ozlu et al., 2015). To obtain more rigorous data, we used CRISPR-Cas9 genome editing (Ran et al., 2013) to knockout PCDH7 in HeLa S3 cells. PCDH7 was targeted using three individual single guide RNAs (sgRNAs). Control cells were treated with a non-targeting (NT) guide RNA (sgNT) in parallel. The knockout (KO) of PCDH7 in isolated colonies was validated by western blotting analysis (Fig. S1D). One colony (PCDH7sg1) was selected to proceed with phenotypic characterization and sgNT-treated cells were used as a control. PCDH7 expression levels in PCDH7 KO, rescue (PCDH7 KO+PCDH7::GFP) and control (PCDH7 KO+eGFP) cells were also validated by western blotting analysis (Fig. S1E).

Then, we examined the extent of cell division failure in PCDH7 KO cells. When we analyzed the multinucleation percentage in fixed cells, we observed a moderate but statistically significant increase in the multinucleation rate in PCDH7 KO cells (Fig. 2A, middle; Fig. 2B). The expression of PCDH7::GFP in PCDH7 KO (Fig. 2A, right) decreased the multinucleation rate statistically close to the levels seen in the control and rescued the phenotype (Fig. 2B, blue).

The weak but statistically significant multinucleation rate could be due to possible adaptations in PCDH7 KO HeLa S3 cells, which can also adapt to proliferate in suspension (Thilly, 1976). Therefore, we employed siRNA- and endonuclease-prepared siRNA (esiRNA)-based depletion of PCDH7 in adherent HeLa Kyoto and HEK293T cells. HeLa Kyoto cells were treated either with control siRNA or PCDH7 siRNA mix, and PCDH7 depletion at the protein level was confirmed by western blotting (Fig. S2A).

The multinucleation rate increased about twofold in PCDH7 siRNA-treated cells (7.7%) in comparison to control siRNA-treated cells (4.3%) (Fig. 2C). Similarly, in HEK293T cells, which have relatively high PCDH7 expression (Human Protein Atlas, accessed 2022, https://www.proteinatlas.org/ENSG00000169851-PCDH7/cell+line; Karlsson et al., 2021), PCDH7 knockdown (Fig. S2B) significantly increased the multinucleation rate, proportional to the extent of protein depletion (Fig. S2C). HeLa Kyoto cells were treated either with control siRNA or PCDH7 siRNA mix, and PCDH7 depletion at the protein level was confirmed by western blotting (Fig. S2A).

To examine the multinucleation phenotype of PCDH7, we performed live-cell analysis using LifeAct::RFP-expressing cells. In contrast to control cells (Fig. 2D, top; Movie 2), in PCDH7 KO cells, the cleavage furrow ingression started; however, cells were not able to complete cytokinesis and merged back to form multinucleated cells (Fig. 2D, bottom; Movie 3). We noticed that the onset of cleavage furrow formation (27 min in control versus 36 min in PCDH7 KO) and completion of furrow ingression (33 min in control versus 42 min in PCDH7 KO) were prolonged in PCDH7 KO cells (Fig. 2D). Quantifying these in siRNA-treated HeLa Kyoto cells revealed a significant delay in both mitosis and cleavage furrow ingression durations (Fig. 2E–G).

To address whether the interaction partners of PCDH7 are involved in its cell cycle-dependent translocation, we employed a proteomic approach using the proximity-dependent biotinylation (BioID) method (Roux et al., 2012). In this method, the protein of interest is fused to a promiscuous biotin ligase BirA*, which biotinylates proteins in proximity to the bait, and the biotinylated proteins are then identified using mass spectrometry. Initial immunofluorescence analysis revealed that the PCDH7::BirA* recombinant protein exhibited the expected localization pattern in both mitotic and interphase cells, and specifically biotinylated the vicinity (Fig. S3A). To biochemically assess the biotinylation efficiency and specificity, we performed streptavidin affinity purification. Western blotting analysis revealed that PCDH7::BirA* exhibited efficient and distinct biotinylation patterns (Fig. S3B), and streptavidin affinity purification successfully pulled down PCDH7 (Fig. S3C).

To discover the interphase and mitosis-specific interaction partners of PCDH7, the biotinylated proteins that were isolated from PCDH7::BirA*-expressing mitotic and interphase cells using streptavidin beads were analyzed using liquid chromatography and tandem mass spectrometry (LC-MS/MS). Non-transfected but biotin-supplemented cells were used as a control. The significant interactors were then identified by calculating the spectral count-based fold change between control and PCDH7::BirA*-expressing cells from four biological replicates with a false discovery rate (FDR) of 0.05 (Choi et al., 2015). We identified 78 proteins for mitosis (Table S1) and 129 for interphase cells (Table S2), 47 of which were common in both groups (Table S3). Those proteins were then analyzed using the STRING database (Szklarczyk et al., 2019) and clustered according to Gene Ontology (GO) and KEGG pathway enrichments (Raudvere et al., 2019) using StringApp (Cline et al., 2007; Doncheva et al., 2019). Significant clusters included actomyosin network-related proteins, cell adhesion proteins, cadherin-binding proteins, vesicular proteins and ERM family proteins (Fig. 3A).

Among many interesting candidates, one striking interaction partner of PCDH7 was a palmitoyltransferase (ZDHHC5), which adds the palmityl group to its target proteins. Palmitoylation is a reversible post-translational modification that is known to affect the hydrophobicity, membrane domain interactions and conformation of transmembrane proteins (Blaskovic et al., 2013). To address whether PCDH7 cooperates with ZDHHC5 during cell division, we first visualized the subcellular localization of PCDH7 and ZDHH5 in PCDH7-GFP-BAC cells by using anti-GFP and anti-ZDHH5 antibodies, respectively. During interphase, both PCDH7 and ZDHHC5 localized to the cell–cell contact regions (Fig. 3B, top). During mitosis, both proteins decorated the cell surface and retraction fibers (Fig. 3B, middle) and, as cells proceeded to cytokinesis, they both accumulated at the cleavage furrow (Fig. 3B, bottom). To further verify their colocalization, we applied the proximity ligation assay (PLA) (Söderberg et al., 2006). For this, we used fixed PCDH7-GFP-BAC cells and anti-GFP and anti-ZDHHC5 antibodies to visualize target protein interaction during cytokinesis (Fig. 3C). As expected, we observed a significantly higher signal in cells treated with both primary antibodies than in the control cells that were treated with only one primary antibody (Fig. 3D). The PLA signals were accumulated around the equatorial zone between segregating chromosomes during cytokinesis, which is in line with the immunofluorescence assay (Fig. 3C, bottom).

Given their colocalization, we tested whether PCDH7 physically interacts with ZDHHC5 during mitosis by performing co-immunoprecipitation using the GFP-trap approach in mitotic PCDH7::GFP-expressing cells. We observed that PCDH7 interacts with ZDHHC5 in mitotic cells (Fig. 3E).

The interaction between PCDH7 and ZDHHC5 prompted us to test whether PCDH7 is palmitoylated. To this end, we performed a palmitoylation assay by using acyl-biotin exchange (ABE) chemistry (Wan et al., 2007). PCDH7::GFP-expressing HeLa S3 cells were used to enrich the PCDH7 protein in the samples. Briefly, after performing a high-speed-centrifugation-based membrane enrichment, unmodified cysteine thiol groups were blocked using N-ethylmaleimide (NEM). Half of the sample was treated with hydroxylamine (HA), which cleaves the palmitate groups from cysteines that were then biotinylated using a thiol-reactive biotin molecule. As a negative control, the other half was not treated with HA (HA−). Subsequently, thiol-biotinylated proteins were purified by streptavidin beads (Fig. 4A). Palmitoylated proteins were expected to be present in the hydroxylamine-treated (HA+) eluates. Palmitoylation of PCDH7 was tested by western blotting using an anti-PCDH7 antibody. In parallel, antibodies against EGFR and calnexin were used as positive controls that are known to be palmitoylated. Along with EGFR and calnexin, we observed PCDH7 in the palmitoylated fraction. As expected, PCDH7 was not detectable in the HA− fraction (Fig. 4B).

To further examine the palmitoylation of PCDH7, we used a standard palmitoylation inhibitor, 2-bromopalmitate (2BP), which is a non-metabolizable palmitate analog that inhibits the incorporation of palmitate into proteins (Webb et al., 2000). We first tested the effect of the palmitoylation inhibitor on the hydrophobicity of PCDH7 using Triton X-114 extraction, which separates proteins into detergent and aqueous phases according to their hydrophobic properties (Bordier, 1981). The inhibition of palmitoylation decreased the proportion of PCDH7 in the detergent phase, suggesting a reduction in the hydrophobicity of PCDH7 (Fig. S4A).

To examine the role of palmitoylation in mitosis-dependent PCDH7 cell surface localization, we tested the association of PCDH7 to the cell surface after palmitoylation inhibitor (2BP) treatment. For this, we purified cell surface proteins and compared PCDH7 levels in control and 2BP-treated cells. Briefly, HeLa S3 cells were treated with 2BP or DMSO (control) overnight, while being synchronized to interphase or mitosis using thymidine and STC, respectively. To enrich for cell surface-exposed proteins, intact cells were labeled with a non-permeable sulfo-NHS-SS-biotin reagent, followed by affinity purification of biotinylated proteins using streptavidin (Özkan Küçük et al., 2018). Although 2BP treatment did not affect biotinylated PCDH7 levels in interphase cells (Fig. 4C, left), it notably decreased the amount of biotinylated PCDH7 in mitotic cells [Fig. 4C, right, elute (E)] and increased the amount of unbiotinylated cytoplasmic PCDH7 [Fig. 4C, right, unbound (U)]. These results suggest that palmitoylation of PCDH7 is required for its mitotic cell surface localization. On the other hand, the amounts of EGFR in the input (I) and elute (E) lanes were comparable in both DMSO- and 2BP-treated cells, suggesting that during mitosis, this effect is specific to PCDH7.

To further examine the impact of the palmitoylation inhibitor on the subcellular localization of PCDH7 in mitotic and interphase cells, we performed immunostaining and live imaging of a PCDH7-GFP-BAC cell line treated with 2BP. In line with our previous findings (Fig. 4C), the inhibition of palmitoylation did not show a noticeable effect on the localization of PCDH7 at the cell–cell contacts during interphase (Fig. S4B). However, it significantly affected the surface localization of PCDH7 in mitotic cells (Fig. 5A) and decreased the plasma membrane enrichment (Fig. S4C) of PCDH7 (Fig. 5B). Live imaging of PCDH7 also supported these results (Fig. S4D; Movies 4 and 5).

To further investigate the effect of the palmitoylation inhibitor on the dynamic behavior of PCDH7 at the mitotic cell surface, we again took the FRAP approach and measured the extent of PCDH7::GFP recovery at the plasma membrane after photobleaching in the absence or presence of 2BP. Inhibition of palmitoylation did not have a significant effect on the FRAP recovery of PCDH7 at the cell–cell contacts during interphase (Fig. 5C, red and orange). On the other hand, 2BP treatment dramatically increased the FRAP recovery time of PCDH7 at the mitotic plasma membrane (Fig. 5C, dark blue and blue). These results suggest that inhibition of palmitoylation has a detectable effect on the dynamics of PCDH7 at the mitotic cortex but not at the cell–cell cortex.

Finally, we analyzed the effect of 2BP on the localization of PCDH7 during cytokinesis (Fig. 5D). Inhibition of palmitoylation explicitly perturbed the cleavage furrow and equatorial localization of PCDH7 during cytokinesis (Fig. 5D; Fig. 5E, left) and led to a decrease in the levels of PCDH7 at the cleavage furrow (Fig. 5E, right). We conclude that palmitoylation is dispensable for stable localization of PCDH7 at cell–cell contacts during interphase, but essential for its cortical and cleavage furrow localization during mitosis and cytokinesis, respectively.

Building upon the results of palmitoylation-dependent localization of PCDH7 in mitosis and cytokinesis, we next sought to examine the role of ZDHHC5 in this process. To this end, siRNA- (Fig. 6A) or shRNA- (Fig. S5A) mediated knockdowns of ZDHHC5 and control (non-targeting shRNA and siRNAs) HeLa PCDH7-GFP-BAC cells were used. In line with our previous observation, in control siRNA-treated cells, ZDHHC5 and PCDH7 largely colocalized at the mitotic cell surface and retraction fibers (Fig. 6B, top). Strikingly, ZDHCC5 depletion significantly decreased the PCDH7 signal at the mitotic cell surface (Fig. 6B, bottom; Fig. 6C). shRNA-mediated knockdown of ZDHHC5 (Fig. S5A) also gave similar results (Fig. S5B). The cell surface localization of PCDH7 during mitosis was significantly decreased upon shZDHHC5 treatment (Fig. S5B, right).

Next, we tested whether restoring ZDHHC5 expression could rescue impaired plasma membrane localization of PCDH7 during mitosis. Expression of murine ZDHHC5 in siZDHHC5 cells significantly restored the plasma membrane localization of PCDH7 in mitotic cells. However, the catalytically inactive ZDHHC5 mutant C134S was not able to rescue this phenotype. PCDH7 levels at the plasma membrane were similar to those following siZDHHC5 treatment (Fig. 6D,E). These results indicate that the palmitoylation activity of ZDHHC5 translocates PCDH7 to the plasma membrane at the onset of mitosis.

Next, we examined the role of ZDHHC5 in targeting PCDH7 to the cleavage furrow. In ZDHHC5-depleted cells, cleavage furrow localization of PCDH7 was perturbed (Fig. 7A,B). PCDH7 was no longer enriched at the cleavage furrow or the equatorial zone during late cytokinesis in ZDHHC5 knockdown cells (Fig. 7A,B).

As ZDHHC5 localizes to the cleavage furrow together with PCDH7, we next examined whether ZDHHC5 depletion affected cytokinesis. Intriguingly, we observed a significantly increased multinucleation rate (8.5%) in shZDHHC5 cells in comparison to that in cells treated with control shRNA (4.6%) (Fig. S5C; Fig. 7C). To conclude, ZDHHC5 localizes and targets PCDH7 to the cleavage furrow and equatorial zone during cytokinesis.

To gain more insight into the cytokinesis defect of PCDH7-depleted cells, we monitored active RhoA and myosin levels at the cleavage furrow. Control and PCDH7 KO cells were transfected with an eGFP–RhoA Biosensor (Piekny and Glotzer, 2008) that binds only to the active form of RhoA, enabling tracking of RhoA activity. Live-cell imaging was performed to compare RhoA activity during cytokinesis in control (Fig. 8A, top; Movie 6) and PCDH7 KO (Fig. 8A, bottom; Movie 7) cells. Active RhoA levels at the cleavage furrow were significantly decreased in PCDH7 KO cells compared with those in the control cells (Fig. 8B; Fig. S5). Although active RhoA levels were impaired, the cells were able to initiate furrow ingression in PCDH7 KO cells. However, a significant portion of those cells (11.7%) failed to complete furrow ingression and they merged back (Fig. S6B,C).

Next, we examined active myosin levels in PCDH7-depleted cells by using an anti-phospho-myosin II (S19) antibody (Matsumura et al., 1998) (Fig. 8C). The phosphorylated myosin was significantly enriched at the cleavage furrow in control cells (Fig. 8D, gray). In PCDH7 KO cells, phospho-myosin intensity was significantly reduced at the cleavage furrow (Fig. 8D, red). The expression of ectopic PCDH7 in PCDH7 KO cells significantly replenished the phospho-myosin levels (Fig. 8D, blue).

In agreement with ZDHHC5 dependency of PCDH7, similar to PCDH7, depletion of ZDHHC5 caused a significant decrease in phospho-myosin (S19) levels at the cleavage furrow (Fig. 8E,F). Taken together, our data suggest that the ZDHHC5–PCDH7 axis has an impact on RhoA and myosin activity during cytokinesis, thus contributing to the fidelity of cell division.

This study aimed to understand the function and the underlying mechanism of cell cycle-dependent localization of PCDH7. Our analysis revealed an unprecedented role of palmitoylation in the translocation of PCDH7 to the mitotic cell surface and cleavage furrow, which promotes myosin phosphorylation at the cleavage furrow. Palmitoylation is a reversible post-translational modification that plays a key role in controlling protein targeting by increasing the hydrophobicity of a protein (Linder and Deschenes, 2007). Our data support a mechanism by which palmitoylation of PCDH7 stabilizes its cell surface and cleavage furrow localization. Indeed, its cell surface and cleavage furrow localization during mitosis and cytokinesis, respectively, was dependent on palmitoylation, whereas the localization to cell–cell contacts during interphase was not affected by palmitoylation inhibition. Our FRAP analysis revealed that the turnover rate and dynamics of molecules at the cell–cell contact region and cell surface are different: they are faster at the cell surface during mitosis. It is possible that trans interactions are more stable and palmitoylation is dispensable or less critical for trans-clustered molecules at the cell–cell contact regions. Our analysis revealed that PCDH7 is palmitoylated; however, our attempts at finding palmitoylation sites by mutating potential cysteine residues that affect protein localization failed. It is challenging to determine palmitoylation sites because of the lack of a consensus motif. Besides cysteine, palmitoylation can also occur on serine and lysine residues. In addition, a cumulative effect of palmitoylation on multiple cysteines and even on serine or lysine (Brownlee and Heald, 2019) residues might be involved; thus, mutating individual residues might not be sufficient to mimic the unpalmitoylated form. Similarly, PCDH1, which shares 46% homology with PCDH7, was also found to be palmitoylated and localized to the membrane in a palmitoylation-dependent manner (Kahr et al., 2013). However, mutating one palmitoylated cysteine residue was not sufficient to mimic treatment with the palmitoylation inhibitor and exhibit the membrane localization phenotype (Kahr et al., 2013). More work is required to understand the molecular details in the palmitoylation of non-clustered protocadherins.

Our BioID-based proximity interactome identified the palmitoyl transferase ZDHHC5 as a significant interactor of PCDH7. In contrast to the majority of palmitoyl acyl transferases that localize to the Golgi or endoplasmic reticulum, ZDHHC5 localizes to the plasma membrane (Ohno et al., 2006). ZDHHC5 has been implicated in multiple cellular processes such as endocytosis, cell adhesion, Na⁺ pump activity and pathogen–host interactions, partly by regulating the localization of related proteins (Plain et al., 2020; Pradhan et al., 2021; Woodley and Collins, 2019, 2021). In this study, for the first time, we analyzed the function of ZDHHC5 in the context of cell division. At metaphase, ZDHHC5 localizes to the mitotic cell surface and retraction fibers, and it concentrates at the cleavage furrow during cytokinesis. The translocation of PCDH7 to the mitotic cortex and cleavage furrow occurs in a ZDHHC5 palmitoylation activity-dependent manner.

Loss of ZDHHC5 caused cytokinesis defects and a statistically significant increase in the rate of multinucleated cells. Imaging of protein fatty acylation in cells undergoing cell division has revealed that at metaphase, S-palmitoylation is enriched at the cell surface and around the spindle, and as cells progress into cytokinesis, palmitoylated proteins are concentrated at the cleavage furrow (Hannoush and Arenas-Ramirez, 2009). Based on those findings, it is entirely possible that palmitoyl acyl transferases, and in part ZDHHC5, might have a wider role in organizing the mitotic cell cortex and targeting cell division-related proteins to the plasma membrane and cleavage furrow. PCDH7 might be the very first example of many cell cycle-dependent palmitoylated proteins. Although many studies have indicated the role of protein palmitoylation in protein trafficking (Linder and Deschenes, 2007) and different intracellular signaling pathways (Resh, 2006), the importance of palmitoylation in the context of cell division has been emerging only recently. Depalmitoylation activity is required for the unequal partitioning of Notch and Wnt signaling during asymmetric cell division (Stypulkowski et al., 2018). Palmitoylation-dependent membrane association of importin α has been shown to affect the mitotic spindle and nuclear scaling during Xenopus embryogenesis (Brownlee and Heald, 2019). Our findings attribute new functional aspects to the role of palmitoylation in cell division. Future studies will expand upon the palmitoylation-dependent regulatory mechanisms during cell division by investigating palmitoyltransferases and their target molecules that function in mitosis and cytokinesis.

What is the function of PCDH7 during cell division? Our previous study showed that PCDH7 is required for the development of full mitotic rounding pressure (Ozlu et al., 2015). Rounding pressure is dependent on myosin II activity (Stewart et al., 2011a,b). In the present study, we have shown that the knockout of PCDH7 caused a statistically significant increase in cytokinesis failure in HeLa S3 cells and a delay in mitosis. One possible explanation for the increased duration of mitosis in PCDH7-depleted cells is that rounding pressure that is regulated by PCDH7 is needed for the accurate timing of mitosis.

The multinucleation phenotype caused by PCDH7 depletion was moderate but significant. In our previous study, although we observed a reduced rounding pressure in PCDH7 siRNA cells, the cells were still able to round up. The weakness of the observed phenotypes might be due to the functional redundancy between PCDH1 and PCDH7 and the fact that multiple parallel pathways act together for mitotic rounding and cytokinesis (Cramer et al., 1994; Ramkumar and Baum, 2016). We observed that RNAi-based depletion of PCDH7 in HeLa Kyoto and HEK293T cells induces a higher multinucleation rate than CRISPR-based knockout of PCDH7 in HeLa S3 cells. The stronger phenotype observed in RNAi experiments might be because PCDH7 KO cells might have developed compensatory changes in the absence of PCDH7 and such adaptations were not developed in acute depletions. Besides, HeLa Kyoto cells and HEK293T cells are more adherent than HeLa S3 cells and this suggests that adherent cells are affected more by the depletion of PCDH7.

In addition, we observed that PCDH7 accumulates asymmetrically at the cleavage furrow. PCDH7 might have a role in apical-basal polarity and its major impact in cytokinesis might only be evident in the epithelium where cells are tightly packed and polarized, and not when cells are grown separately in culture. True epithelial cells in culture and three-dimensional cultures should be used in future studies to circumvent this issue.

The actomyosin network is important for mitotic rounding and cleavage furrow formation, and both processes appear to be affected in the absence of PCDH7. A previous genome-wide RNAi screen also reported binucleation and cell migration defects in PCDH7-depleted cells (Neumann et al., 2010). Indeed, we observed that active RhoA and myosin levels at the cleavage furrow were significantly reduced in the absence of PCDH7 when compared to the control cells. Previously it was reported that PCDH7 increases phospho-myosin light chain levels to enhance anchorage-independent cell growth (Wang et al., 2020). Similarly, in a parallel study, we observed that PCDH7 overexpression enhances phospho-myosin levels during cell migration in Retinal Pigment Epithelial (RPE) cells and PCDH7 interacts with PP1cβ, the catalytic subunit, and MYPT1, the myosin targeting subunit of myosin phosphatase (Qureshi et al., 2022 preprint). Previous studies also reported that PCDH7 behaves as a signaling molecule more than an adhesion molecule and inhibits the phosphatase activities of PP1aa and PP2A (Wang et al., 2020; Zhou et al., 2017). Recently, it has been reported that PCDH7 functions in osteoclast differentiation by regulating RhoA and Rac1. Loss of PCDH7 expression impairs RhoA activation in pre-osteoclasts (Kim et al., 2021). In agreement with those studies, we also observed that depletion of PCDH7 impairs active RhoA and myosin levels at the cleavage furrow, albeit RhoA and myosin activity were sufficient for promoting actomyosin ring constriction and furrow ingression. Instead, we observed defects in the completion of furrow ingression and higher rates of merged cells upon PCDH7 depletion. Further studies are necessary to elucidate the molecular details of PCDH7-dependent regulation of the actomyosin cortex and the association of PCDH7 expression with RhoA and myosin activity during cytokinesis.

In summary, our study suggests a new pathway for cell cycle-dependent protein localization during cell division. We suggest that ZDHHC5-dependent targeting of PCDH7 to the cell cortex and cleavage furrow contributes to cortical remodeling. The molecular details of palmitoylation-dependent cortical regulation during cell division will unravel the fundamental redundant pathways existing in human cells.

HeLa S3 (American Type Culture Collection, CCL-2.2, female), HeLa Kyoto (a gift from Ulrike Eggert, Randall Centre for Cell and Molecular Biophysics, School of Basic and Medical Biosciences, King's College London, London, UK), and HEK293T [a gift from Dr Tamer Önder, Koç University Research Center for Translational Medicine (KUTTAM), Istanbul, Türkiye., Koç University School of Medicine, Istanbul, Türkiye] cells were grown in Dulbecco's modified Eagle medium (DMEM) (Sigma-Aldrich, D6429) supplemented with 1% penicillin-streptomycin (P/S) (Capricorn Scientific, PS-B) and 10% fetal bovine serum (FBS) (Gibco, 10270106). The PCDH7-GFP-BAC (Poser et al., 2008) transgenic cell line was a kind gift from Dr Ina Poser (Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany) and was grown in DMEM supplemented with 1% P/S, 10% FBS and 400 μg/ml G418 (Santa Cruz Biotechnology, sc-29065A). HeLa S3 cells stably expressing PCDH7::GFP were created by lentiviral delivery of the PCDH7::GFP construct, acquired by cloning PCDH7 into the pLenti CMV GFP Puro (pLenti) (Addgene, 17448) mammalian expression vector, and were grown in DMEM supplemented with 1% P/S, 10% FBS, and 2 µg/ml puromycin. To inhibit palmitoylation, cells were incubated with 100 µM 2BP (Sigma-Aldrich, 21604) overnight (Webb et al., 2000).

Cells synchronization to interphase, mitosis (Ozlü et al., 2010) and cytokinesis (Hu et al., 2008; Karayel et al., 2018) was performed as previously described. Cells were synchronized to interphase by a double thymidine block using 2 mM thymidine (Santa Cruz Biotechnology, sc-296542). Subsequently, to induce monopolar mitosis, cells were treated with 10 µM S-trityl-L-cysteine (STC) (Sigma-Aldrich, 164739). For synchronization of monopolar cytokinesis, cells were treated with 10 µM STC and then 100 µM purvalanol A (Tocris Bioscience, 1580). For synchronization of bipolar cytokinesis, cells were incubated with 10 ng/ml of nocodazole (Calbiochem, 487928) for 5 h for mitosis and released from nocodazole for 1 h for cytokinesis.

For immunostaining, cells were plated on coverslips (12 mm), fixed with 3% paraformaldehyde, blocked, and incubated with primary and secondary antibodies in 2% bovine serum albumin (BSA) in PBS containing 0.1% Triton X-100. The following antibodies and reagents were used: anti-β-tubulin (Cell Signaling Technology, CS2128S, 1:500), anti-α-tubulin (Cell Signaling Technology, 3873S, 1:1000), anti-tubulin (Abcam, ab6160, 1:500), anti-GFP (non-commercial antibody, 1:5000; Invitrogen, A11120, 1:1000), anti-HA (Abcam, ab16918, 1:200), anti-phospho-myosin (Cell Signaling Technology, CS3675, 1:500), anti-ZDHHC5 (Atlas Antibodies, HPA014670, 1:500), Alexa Fluor 488 (Invitrogen, A21202, A21206, 35552, 1:1000)- and Alexa Fluor 555 (Invitrogen, A11077, 1:1000) conjugated secondary antibodies, Streptavidin–Alexa Fluor 488 (Invitrogen, S32354, 1:1000), Phalloidin iFluor 555 (Abcam, ab176756, 1:1000) and DAPI (Sigma-Aldrich, D8417).

For live imaging, ibiTreat μ-Slide 8 Well (ibidi, 80826) or μ-Dish 35 mm (ibidi, 81156) plates were used. Confocal microscopy was performed with a Leica DMi8/SP8 TCS-DLS (LAS X Software) laser scanning confocal microscope using 40× Plan Apo 1.3 NA and 63× Plan Apo 1.4 NA oil-immersion objectives. Live-imaging experiments were performed using a Leica DMi8 widefield fluorescence microscope (LAS X Software) or Leica DMi8/SP8 TCS-DLS confocal microscope equipped with 37°C and 5% CO₂ chambers. 63× Plan Apo 1.4 NA oil-immersion or 20× PL FLUOTAR L 0.40 NA objectives were used. Single images or z-stacks were acquired every 3 min and a single focal plane was used in the figures unless specified in the figure legends.

Apart from FRAP data, all images were analyzed in Fiji. Graphs and statistical data were generated in GraphPad Prism, except for multinucleation data. Statistical tests for multinucleation analysis were performed in R: to test the significance of changes in the proportion of ‘normal’ cells across conditions, we reformulated the 2×2 contingency tables as data tables suitable for logistic regression analysis. We applied the Bonferroni test. Specifically, we used the glm function in R with family=‘binomial’ option to test the changes in the log odds ratio of ‘normal’ cells given the condition. We used the replicate of the origin of each cell as a covariate term for the logistic regression model. Finally, we used the ‘p.adjust’ function with the method=‘bonferroni’ option to adjust P-values from coefficients of multiple logistic regression models. Statistical details of each experiment including the statistical test used, the exact value of n and definition of error bars are given in the figure legend for each figure.

The microscopy setup included a frequency-doubled femtosecond-pulsed Ti:Sa solid-state tunable laser source (Chameleon Ultra II, Coherent) equipped with a second harmonic generator. The laser output was tuned to 488 nm and the beam was directed through mirrors and a Keplerian telescope to the inverted microscope (Eclipse TE2000-U; Nikon) equipped with a dichroic mirror (Chroma, Q495LP) and 60× oil-immersion objective (Nikon Apo TIRF, NA 1.49). A 300 mm focal length lens was placed right before the microscope to focus the laser at the back focal plane of the microscope objective to obtain widefield illumination. The microscope was equipped with two different cameras for brightfield and fluorescence image acquisitions. Brightfield images were captured by a CCD camera (Thorlabs, DCU223M). Fluorescence images were captured by an EMCCD camera (Hamamatsu ImagEM C9100-13) placed after an emission filter with a pass band of 530±30 nm. Photobleaching was performed by removing the neutral density optical filter and the lens right before the microscope, and simultaneously focusing the laser light at the desired area for 1 s. FRAP image acquisition was done every 5 s using an automated shutter and a minimum of 130 frames were captured following the photobleaching of each sample.

Cells were transfected using polyethylenimine (PEI) (Polysciences, 23966), Lipofectamine 2000 (Invitrogen, 11668019) or Lipofectamine RNAiMAX (Thermo Fisher Scientific, 13778075) protocols. For the PEI protocol, transfection mixtures were prepared in Opti-MEM reduced serum medium (Invitrogen, 31985047) using PEI with 60 μg DNA/150 mm dishes in a 3:1 ratio. After 30 min of incubation at room temperature, the transfection mixture was added to the cells. For the Lipofectamine 2000 protocol, the manufacturer's instructions were followed.

Plasmids encoding murine ZDHHC5 and its catalytically inactive mutant ZDHHC5 (C134S) were kindly provided by Prof. William Fuller (Institute of Cardiovascular and Medical Sciences, College of Medical, Veterinary and Life Sciences, University of Glasgow, UK). The pEGFP-RhoA Biosensor plasmid was acquired through Addgene (#68026).

For viral transduction, lentiviruses were packaged in HEK293T cells using PEI transfection of the target sequence-containing vector (pLenti), packaging vector (psPAX2, Addgene #12260) and envelope vector (pCMV-VSV-G, Addgene #8454). Viral particles were collected after 48 and 72 h of transfection and used to infect HeLa S3 in the presence of 2 μg/ml protamine sulfate (Sigma-Aldrich, P4505) as a coadjutant. Transduced cells were then selected with the appropriate selective antibiotic.

The PCDH7 sequence was amplified from the PCDH7::eGFPN1 construct (Ozlu et al., 2015) using specific primers (5′-GTCAGCTAGCACCATGCTGAGGATGCGGACC-3′ and 5′-GCTAGAATTCGCCCTCCCTGGGATATTTAAATATATTTG-3′) and cloned into the BioID vector [pcDNA3.1 MCS-BirA* (R118G)-HA, Addgene #36047].

Proximity-dependent biotinylation was performed as previously described (Roux et al., 2012). Briefly, cells were transfected with the BioID vector and incubated with 50 μM biotin (Invitrogen, B20656) during interphase and mitosis synchronization. Cell pellets were lysed in lysis buffer [50 mM Tris-HCl, pH 7.4; 500 mM NaCl; 0.4% SDS; 5 mM EDTA; 2% Triton X-100; 1 mM dithiothreitol (DTT; Sigma-Aldrich, 43815); and protease inhibitor] and incubated with streptavidin beads (Pierce, 53117) overnight at 4°C in a tube rotator. Whole-cell lysates and unbound fractions were kept at −20°C for further analysis. Beads were washed with wash buffer 1 (2% SDS in dH₂O), wash buffer 2 (2% deoxycholate; 1% Triton X-100; 50 mM NaCl; 50 mM HEPES, pH 7.5; and 1 mM EDTA), wash buffer 3 (0.5% NP-40; 0.5% deoxycholate; 1% Triton X-100; 500 mM NaCl; 1 mM EDTA; and 10 mM Tris, pH 8.1) and wash buffer 4 (50 mM Tris, pH 7.4; 50 mM NaCl) sequentially. For western blotting analysis, bound proteins were eluted from the streptavidin beads with 50 µl of Laemmli-DTT sample buffer containing 500 nM D-biotin (Invitrogen, B1595) at 98°C by shaking at 1000 rpm in a ThermoMixer (Eppendorf). For mass spectrometry analysis, on-bead tryptic digestion was performed. Briefly, beads were washed with urea buffer [8 M urea (Sigma-Aldrich, A2383); 0.1 M Tris-HCl, pH 8.5]. Then, bead-bounded proteins were reduced with 100 mM DTT and alkylated using 100 mM iodoacetamide (AppliChem, A1666). After alkylation, beads were washed with 50 mM ammonium bicarbonate (AppliChem, A3583) and incubated with trypsin (Thermo Fisher Scientific, 25247) at 37°C overnight (14–16 h) in the ThermoMixer. The resulting digested peptides were collected and desalted using C18 STAGE tips (Rappsilber et al., 2007). The experiment was performed in four biological replicates with a minimum of two technical replicates for each condition.

Peptides were analyzed by online C18 nanoflow reversed-phase nLC (NanoLC-II, Thermo Fisher Scientific) or C18 nanoflow reversed-phase HPLC (Dionex Ultimate 3000, 3500 RSLC nano, Thermo Fisher Scientific) connected with an orbitrap mass spectrometer (Q Exactive Orbitrap, Thermo Fisher Scientific). Samples were separated in an in-house packed 100 μm (internal diameter)×23 cm C18 column (Reprosil-Gold C18, 5 μm, 200 Å, Dr. Maisch) using 80-min linear gradients from 5–25%, 25–40% and 40–95% acetonitrile in 0.1% formic acid with 300 nl/min flow in 100 min total run time. The scan sequence began with an MS1 spectrum [Orbitrap analysis; resolution 70,000; mass range 400–1500 m/z; automatic gain control (AGC) target 1e6; maximum injection time 32 ms]. Up to 15 of the most intense ions per cycle were fragmented and analyzed in the orbitrap with data-dependent acquisition. MS2 analysis consisted of collision-induced dissociation (higher-energy collisional dissociation) (resolution 17,500; AGC 1e6; normalized collision energy 26; maximum injection time 85 ms). The isolation window for MS/MS was 2.0 m/z.

Raw files were processed with the Proteome Discoverer 2.3 (Thermo Fisher Scientific) software. Carbamidomethylation of cysteine residues was used as a fixed modification, and acetylation (of protein N-termini) and oxidation of methionine residues were used as variable modifications. A maximum of two missed cleavages was allowed for the tryptic peptides. The precursor mass tolerance was set to 10 parts per million and fragment mass tolerance was set to 0.02 Da. Both peptide and protein FDRs were set to 0.01. The other parameters were used with default settings. The database search was performed against the human UniProt database (release 2015) containing 21,039 entries using the SEQUEST HT search engine integrated into the Proteome Discoverer environment.

The spectral counts of proteins were used to calculate fold-change ratios and FDR values for identified proteins using the qspec-param program of qprot_v1.3.5 (Choi et al., 2015). Proteins were filtered with a 0.05 cut-off for FDR values. Significant protein hits were loaded into the STRING database v11.0 (Szklarczyk et al., 2019) using the Cytoscape StringApp (Doncheva et al., 2019) with 0.7 confidence. MCODE clustering of the network was performed by Cytoscape (v3.7.2) and its plugin clusterMaker (Cline et al., 2007). GO and KEGG enrichment analyses of the network were performed via g:Profiler (Raudvere et al., 2019).

The Duolink PLA Kit (Sigma-Aldrich, 92101) was used based on the manufacturer's instructions. Briefly, HeLa S3 PCDH7-GFP-BAC cells were seeded onto glass coverslips and synchronized to cytokinesis with bipolar synchronization. The cells were fixed using 3.2% paraformaldehyde and permeabilized with 0.1% Triton X-100 in TBS. After blocking with Duolink blocking solution for 30 min at 37°C, the cells were incubated overnight at 4°C with corresponding pairs of primary antibodies: anti-GFP (Invitrogen, A11120, 1:1000) for PCDH7 and anti-ZDHHC5 (Sigma-Aldrich, HPA014670, 1:2000). Samples treated with only one corresponding primary antibody were used as controls. The cells were washed and incubated with PLA probes for 1 h at 37°C. After another wash, the cells were treated with the ligase for 30 min at 37°C. The washing step was repeated and the cells were incubated with the polymerase for 100 min at 37°C. After the final washes, the slides were mounted with a coverslip using the Duolink Mounting Medium with DAPI and incubated for 15 min before sealing.

HeLa S3 cells expressing GFP alone or PCDH7::GFP were arrested in mitosis and used for pull-down assays. The pellets were dissolved in PBS with 1% Triton X-100, EDTA-free protease inhibitor (Thermo Fisher Scientific/Pierce, 88666) and PhosSTOP (Roche, 4906845001), and then homogenized and centrifuged at 17,000 g for 10 min at 4°C. Protein concentrations were determined by BCA protein assay (Pierce, 23225) and equal protein amounts from each sample were loaded into pre-conditioned GFP-Trap A beads (gta-20; ChromoTek). Aliquots of the input sample were saved for further analysis. After incubation for 3 h at 4°C, the unbound samples were collected, the GFP-Trap A beads were washed and proteins were eluted in 2× Laemmli buffer with 100 mM DTT by boiling for 10 min at 95°C.

Samples were separated by molecular mass using 10% SDS-PAGE gels and transferred to a nitrocellulose membrane. The membrane was blocked with 4% w/v nonfat dry milk in TBS containing 0.1% Tween-20 and probed with the primary antibody diluted in 2% BSA in TBS containing 0.1% Tween-20. The signal was visualized using ECL (Pierce, 32106; Bio-Rad, 1705061) detection of the HRP-conjugated secondary antibodies (Cell Signaling Technology, 7074S and 70765; 1:2000). The following primary antibodies were used for western blotting: anti-PCDH7 (Abcam, ab139274, 1:400), anti-EGFR (Santa Cruz Biotechnology, SC-03, 1:100), anti-tubulin (Cell Signaling Technology, 3873, 1:2000), anti-actin (Abcam, ab6276, 1:5000), anti-phospho-histone H3 (Upstate, 06-570, 1:500) and anti-biotin (custom antibody, 1:10,000), GFP (custom antibody, 1:5000), anti-ZDHHC5 (Atlas Antibodies, HPA014670, 1:2000) and anti-calnexin (Abcam, ab22595, 1:1000). Source data for all western blots are available in Fig. S7.

For detection of protein palmitoylation, the ABE procedure was performed as previously described (Wan et al., 2007). Briefly, cell pellets were lysed in ice-cold lysis buffer [150 mM NaCl, 50 mM Tris-HCl, 5 mM EDTA, pH 7.4 with 10 mM NEM (Pierce, 23030), 1× protease inhibitor cocktail (PI) and 2× phenylmethylsulfonyl fluoride (PMSF; Cayman, 14333)]. After homogenization, membrane proteins were enriched by using high-speed centrifugation (Optima MAX-XP Ultracentrifuge, TLA-120.2 rotor) at 200,000 g for 30 min at 4°C. The membrane-enriched pellet was dissolved in lysis buffer with 10 mM NEM, 1× PI, 1× PMSF and 1.7% Triton X-100 and incubated at 4°C for 2 h. To remove particulates, the sample was centrifuged at 250 g for 5 min at 4°C. Chloroform-methanol (CM) precipitation was applied to precipitate the proteins. The pellet was air-dried for 2–3 min and 4% SDS buffer (4SB) (4% SDS, 50 mM Tris, 5 mM EDTA, pH 7.4) with 10 mM NEM was added into the sample and incubated for 20 min at 37°C to dissolve the protein pellet completely. After NEM incubation overnight at 4°C (1 mM NEM, 1× PI, 1 mM PMSF and 0.2% Triton X-100), three sequential CM precipitations were applied to remove NEM from the sample. After the final CM precipitation, the pellet was dissolved in 4SB, and the sample was equally divided into two fractions as HA− and HA+, and incubated for 1 h at room temperature with HA− [50 mM Tris, 1 mM HPDP–biotin (Thermo Fisher Scientific, 21341), 0.2% Triton X-100, 1 mM PMSF, 1× PI, pH 7.4] and HA+ (0.7 M hydroxylamine, 1 mM HPDP–biotin, 0.2% Triton X-100, 1 mM PMSF, 1× PI, pH 7.4) buffers, respectively. Then, proteins were precipitated by CM precipitation and the dissolved pellets were incubated in low HPDP–biotin buffer (150 mM NaCl, 50 mM Tris, 5 mM EDTA, 0.2 mM HPDP–biotin, 0.2% Triton X-100, 1 mM PMSF, 1× PI, pH 7.4) for 1 h at room temperature. Three sequential CM precipitations were performed to remove unreacted biotin. The protein pellets were dissolved in 4SB and then SDS was diluted to 0.1% by the addition of 0.2% Triton X-100, 1× PI and 1 mM PMSF, and samples were incubated at room temperature for 30 min. Then, samples were loaded on pre-conditioned Streptavidin Plus UltraLink Resin (Pierce, 53117) and incubated for 90 min at room temperature. Unbound fractions from both samples were saved, beads were washed three times with lysis buffer containing 0.1% SDS and 0.2% Triton X-100, and bound proteins were eluted in 2× Laemmli buffer with 1% β-mercaptoethanol by boiling for 10 min at 95°C.

Hydrophobic proteins were extracted from the hydrophilic ones using the Triton X-114 (TX-114) extraction protocol as described previously (Bordier, 1981; Taguchi et al., 2013). Briefly, precondensation of TX-114 (Biomatik, A4026) was performed by repeated cycles of clarifying at 4°C and incubation at 37°C to separate the detergent phase. Cells were lysed using lysis buffer (2% TX-114 in PBS), and the lysate was cleared by centrifugation at 16,100 g for 3 min at 4°C. The lysate was then centrifuged at 22,000 g for 10 min at room temperature for phase separation. The aqueous phase was removed, and the detergent phase was washed with wash buffer (0.1% TX-114 in PBS), clarified on ice, and incubated at 37°C for phase separation. Centrifugation and washing steps were repeated two more times and the detergent and aqueous phases were collected for further analysis.

Plasma membrane proteins were labeled with sulfo-NHS-SS-biotin for membrane enrichment as previously described (Özkan Küçük et al., 2018). Briefly, cells were incubated with 5 mM S-NHS-SS-biotin (Pierce, 21331) for 30 min at 4°C with gentle shaking, the reaction was quenched with glycine, and cells were snap frozen. Cells were lysed in a buffer (10 mM Tris-HCl pH 7.6, 0.5% SDS, 2% NP40, 150 mM NaCl, 1 mM EDTA and 10 mM iodoacetamide) supplemented with protease inhibitors (Pierce, 88666) and the lysates were incubated with pre-conditioned Streptavidin Plus UltraLink Resin overnight at 4°C. Unbound samples were collected and the beads were washed with lysis buffer three times. Biotinylated surface proteins were eluted by boiling at 70°C for 20 min in SDS sample buffer including 100 mM DTT with agitation. All fractions including whole-cell lysates (input) and unbound and plasma membrane-enriched (elute) fractions were analyzed by western blotting.

ZDHHC5 expression was knocked down using shRNA and siRNA approaches. shZDHHC5 was gifted by Dr G. Ekin Atilla-Gökcümen (Department of Chemistry, University at Buffalo, The State University of New York, Buffalo, USA) (Pradhan et al., 2021) and pLKO.1 was gifted by Dr Elif Nur Fırat Karalar (Department of Molecular Biology and Genetics, Koc University, Istanbul, Türkiye). The lentiviral particles of shZDHHC5 (target sequence 5′-CCCAGTTACTAACTACGGAAA-3′ in pLKO.1 vector) and empty pLKO.1 (Addgene, #8453) were packaged in HEK293T cells and used to transduce PCDH7-GFP-BAC cells. Virus-incorporated stable cells that stably expressed shRNAs were obtained after puromycin selection (2 µg/ml).

siGENOME siRNA pools that target ZDHHC5 (siZDHHC5, Dharmacon, D-026577-01-0020, target sequence: 5′-GGACUAAGCCUGUAUGUGU-3’) and non-targeting siRNA (Dharmacon, D-001210-01-05) were used for ZDHHC5 knockdown. PCDH7 expression was knocked down using different siRNA and esiRNA pools: FlexiTube GeneSolution for PCDH7 (QIAGEN, GS5099), FlexiTube siRNA PCDH7 6 (QIAGEN, SI03080469), FlexiTube siRNA PCDH7 5 (QIAGEN, SI03040261) and esiPCDH7 (Sigma-Aldrich, EHU077631). AllStars Negative Control siRNA (QIAGEN, 1027280) was used as control. All siRNAs were transfected with Lipofectamine RNAiMAX transfection reagent (Thermo Fisher Scientific) according to the manufacturer's instructions. Briefly, cells were seeded and transfected with 15 pmol of siRNA two times, at 24 h and 48 h. After 72 h after seeding, cells were either pelleted for western blotting analysis or fixed for immunofluorescence analysis.

sgRNAs that target PCDH7 were designed using the online ‘CRISPR design tool’ (http://crispr.mit.edu). The following oligonucleotide sequences were used as top/bottom pairs: sg1, 5′-CACCGCGACGTCCGCATCGGCAACG-3′ and 5′-AAACCGTTGCCGATGCGGACGTCG-3′; sg5, 5′-CACCGCATCGTGACCGGATCGGGTG-3′ and 5′-AAACCACCCGATCCGGTCACGATG-3′; and sg6, 5′-CACCGCGGGCTTCTCTTTGGCGCGC-3′ and 5′-AAACGCGCGCCAAAGAGAAGCCCGC-3′. All sgRNAs were cloned into the lentiCRISPR plasmid (Shalem et al., 2014) (pXPR_001; Addgene, #49535) as described in Ran et al. (2013).

PCDH7 knockout cell lines were generated by lipofectamine transfection of the CRISPR plasmid to HeLa S3 cells followed by antibiotic selection. Single colonies were isolated with serial dilution of the pool population and PCDH7 knockout clones were selected after verifying the absence of PCDH7 protein expression with western blotting. CRISPR rescue cell lines were generated by viral transduction of the pLenti (Campeau et al., 2009) PCDH7::eGFP plasmid to PCDH7 knockout cells.

We gratefully acknowledge Büşra Aytül Akarlar and the Proteomics Facility of Koç University (KUPAM) for the technical assistance in mass spectrometry analyses. We thank Dr Bilal Ersan Kerman for critical discussions throughout the project. We also thank Dr Alexandr Jonas for critical discussions and interpretations of FRAP data. We thank Dr Timothy Mitchison for the anti-biotin antibody and critical reading of the manuscript. We thank Dr G. Ekin Atilla-Gökcümen for sharing ZDHHC5 shRNAs, Dr Elif Nur Fırat Karalar for sharing the pLKO.1 plasmid, Dr Tamer Önder for sharing HEK293T cells, Dr Ulrike Eggert for sharing HeLa Kyoto cells, and Prof. William Fuller for sharing murine ZDHHC5 plasmids. We thank Artür Manukyan from Max Delbrück Center for his help with statistical analysis. We also thank Dr Nazan Saner and Dr. Aydanur Şentürk for the critical reading of the manuscript. We gratefully acknowledge the permission to use the facilities of the Cellular and Molecular Imaging Core of Koç University Research Center for Translational Medicine funded by the Ministry of Development, Republic of Türkiye.

The peer review history is available online at https://journals.biologists.com/jcs/lookup/doi/10.1242/jcs.260266.reviewer-comments.pdf