We first explored a direct role for CCDC32 in CME by testing if and when CCDC32 is recruited to CCPs. ARPE-HPV cells that stably express mRuby-clathrin light chain a (mRuby-CLCa) and siRNA-resistant, N-terminally GFP tagged full-length CCDC32 (eGFP-CCDC32(FL)) were generated (Figure 1—figure supplement 1) and imaged using Total Internal Reflection Fluorescence Microscopy (TIRFM). In addition to diffuse staining across the inner plasma membrane (PM) surface, suggestive of direct membrane binding, we also observed colocalization of eGFP-CCDC32(FL) and mRuby-CLCa (Figure 1A) at clathrin-coated pits.

Furthermore, cohort-averaged fluorescence intensity traces were obtained using time-lapse imaging and primary (mRuby-CLCa)/subordinate (eGFP-CCDC32) tracking powered by cmeAnalysis (Aguet et al., 2013; Chen et al., 2020; Jaqaman et al., 2008). As a negative control, ARPE-HPV cells that stably express mRuby-CLCa and eGFP showed neither diffuse PM staining nor eGFP recruitment to CCPs (Figure 1—figure supplement 2), whereas more than half of analyzed CCPs were observed to recruit CCDC32. The recruitment curve of CCDC32 follows the assembly of clathrin on CCPs (Figure 1B). Importantly, compared to CCDC32-positive CCPs, those that fail to recruit eGFP-CCDC32 exhibited significantly shorter and exponentially decaying lifetimes, previously shown to be characteristic of abortive CCPs (Figure 1C; Aguet et al., 2013). Together, these data establish that the early recruitment, along with clathrin, of CCDC32 to nascent CCPs regulates their maturation and lifetimes.

To gain further insight into the function of CCDC32 in CME, we next examined the effects of siRNA-mediated knockdown of CCDC32 on CME and CCP dynamics. In agreement with the previous observation (Wainberg et al., 2021), siRNA-mediated knockdown of CCDC32 in ARPE-HPV cells, which resulted in an ~60% depletion of CCDC32 (Figure 2A and B), reduced the cellular uptake efficiency of TfnR (Figure 2C).

A recent study reported that CCDC32 functions as an essential chaperone for AP2 complex assembly and that CRISPR-mediated knockout of CCDC32 led to a severe decrease in AP2 (Wan et al., 2024). Under our conditions of CCDC32 knockdown, we did not detect any decrease in protein levels of the AP2 complex (Figure 2—figure supplement 1A and B), suggesting that the residual CCDC32 was fully capable of fulfilling this catalytic function.

To further define which stages of CME were dependent on CCDC32, we knocked down CCDC32 in ARPE-HPV cells that stably express eGFP-CLCa and employed quantitative live-cell TIRFM to visualize and analyze CCP dynamics (Mettlen and Danuser, 2014). The videos and kymographs of time-lapse imaging showed that CCDC32 depletion resulted in the formation of brighter, static clathrin-coated structures (CCSs) (Figure 2D and E and Figure 2—videos 1 and 2) that dominate the images. These have been seen under other perturbation conditions (Aguet et al., 2013; Chen et al., 2020) and reflect the accumulation of CCPs that are either larger, flatter, or both. However, we also noted a subpopulation of dynamic CCPs (arrows, Figure 2E) that were visually obscured by bright static CCSs, yet represent the majority of total CCSs detected. Indeed, despite the deceptive nature of the images, which we have previously reported (Chen et al., 2020), the percentage of static CCSs (lifetime >150 s), as determined by unbiased quantitative analysis, was only 7.9% (Figure 2—figure supplement 2).

To quantify the dynamic behaviors of CCPs, we performed cmeAnalysis (Aguet et al., 2013; Jaqaman et al., 2008; Loerke et al., 2011) and DASC (disassembly asymmetry score classification; Wang et al., 2020), which together provide a comprehensive and unbiased characterization of CCP intermediates and CME progression when CCDC32 was depleted. DASC, which measures fluctuations of clathrin-assembly/disassembly to accurately distinguish abortive from productive CCPs (Wang et al., 2020), revealed an increased rate of nascent clathrin assembly, reported as the initiation rate of CCSs (Figure 2F), but a reduced rate of bona fide CCP initiation (Figure 2G) and a correspondingly lower percentage of stable, bona fide CCPs (CCP%; Figure 2H), which are both brighter and longer-lived than abortive CCPs (Wang et al., 2020). Together, these data reveal an early defect in the growth and stabilization of nascent clathrin assemblies leading to an increase in the fraction of abortive CCPs. Importantly, the remaining dynamic population of bona fide CCPs exhibited shorter lifetimes upon CCDC32 knockdown (Figure 2I), indicating their more rapid maturation.

Together, these results demonstrate that CCDC32 is an important endocytic accessory protein involved in CCP stabilization and maturation. Strikingly, despite the profound effects of CCDC32 depletion on CCP dynamics, the efficiency of TfnR uptake was only marginally affected (Figure 2C). These paradoxical effects are typical for endocytic accessory proteins and indicative of their functional redundancy and/or the induction of compensatory mechanisms, including in this case, the observed increased rate of CCS assembly (Figure 2F) and the more rapid maturation of bona fide CCPs (Figure 2I; Aguet et al., 2013; Bhave et al., 2020; Wang et al., 2020).

During CCP maturation, successful invagination of the clathrin coat and its underlying membrane has been identified as a key step that determines the fate of CCPs (Aguet et al., 2013; Wang et al., 2020). To determine whether CCDC32 regulates CCP invagination, we first used Epifluorescence (Epi)-TIRF microscopy to measure CCP invagination in live cells (Figure 3A). In this approach, time-lapse Epi and TIRF fluorescence signals were near-simultaneously acquired for ARPE-HPV eGFP-CLCa cells (Figure 3B) and analyzed using primary (TIRF-channel)/subordinate (Epi-channel) tracking powered by cmeAnalysis (Wang et al., 2020). The resulting Epi and TIRF fluorescence intensity traces of CCPs with similar lifetimes were then aligned and averaged to yield intensity cohorts, which were further log-transformed to give average traces of the invagination depth (∆z\begin{document}$\Delta z$\end{document}) of the CCPs' center-of-mass (Figure 3A and C). Here, we chose to present CCPs with lifetimes of 25–35 s because they represent the invagination behavior of the most frequent tracks in control cells (Figure 2I). CCDC32 knockdown strongly inhibited CCP invagination (Figure 3C), suggesting a key role for CCDC32 in regulating this critical early stage in CME.

To independently validate that CCDC32 regulates CCP invagination, we next examined the clathrin-coated structures (CCSs) at higher resolution using Platinum Replica Electron Microscopy (PREM; Svitkina, 2017). Strikingly, CCDC32 knockdown resulted in a substantial increase in the number of flat CCSs (Figure 4A–C, pseudo-colored blue in panel B, see also Figure 4—figure supplement 1A–C) and a corresponding higher occupancy of CCSs on the PM (Figure 4E; Figure 4—figure supplement 1D), consistent with TIRFM and Epi-TIRF microscopy observations. While the flat CCSs we detected in CCDC32 knockdown cells were significantly larger than in control cells (Figure 4D, mean diameter of 147 nm vs. 127 nm, respectively), they are much smaller than typical long-lived flat clathrin lattices (d≥300 nm; Grove et al., 2014). Indeed, the surface area of the flat CCSs that accumulate in CCDC32 KD cells (mean ~1.69 x 10⁴ nm²) remains significantly less than the surface area of an average 100 nm diameter CCV (~3.14 x 10⁴ nm²). Thus, we refer to these structures as ‘flat clathrin assemblies’ because they are neither curved ‘pits’ nor large ‘lattices’. Rather, the flat clathrin assemblies represent early, likely defective, intermediates in CCP formation. Importantly, while significantly decreased in both number and size, dome-shaped (green) and spherical (orange) CCSs were still detected, likely corresponding to intermediates within the larger subpopulation of dynamic, bona fide CCPs able to maintain TfnR uptake (Figure 4A–D; Figure 4—figure supplement 1A–C). The PREM results provide high-resolution structural data supporting a critical role for CCDC32 in regulating CCP invagination.

Next, we explored the mechanism of CCDC32 recruitment to CCPs. The adaptor AP2, which is essential for CCP and CCV formation, is a heterotetramer consisting of α, β2, μ2, and σ2 subunits (Figure 5A). A previous study had reported interactions between overexpressed σ2-mCherry and CCDC32-GFP (Wainberg et al., 2021), but had not demonstrated interactions with the native AP2 complex. A more recent study reported that C-terminally tagged CCDC32 interacts with the α:σ2 hemicomplex, but does not interact with the assembled AP2 heterotetramer (Wan et al., 2024). To explore this apparent discrepancy, we conducted co-immunoprecipitation (co-IP) experiments from cell lysates of ARPE-HPV cells that stably express a fully functional eGFP-tagged α-subunit of AP2 (Mino et al., 2020; AP2-α-eGFP, Figure 5B). Mass spectrometry analysis revealed, as expected, the co-IP of all the subunits of the AP2 complex, as well as well-known AP2 binding proteins (i.e. EPS15, NECAP2, AAK1, and AAGAB; Mettlen et al., 2018). Notably, CCDC32 was also efficiently co-immunoprecipitated with intact AP2 (Figure 5C; Table 1). However, this experiment does not definitively rule out the possibility that CCDC32 only interacts with a small population of immature α:σ2 hemicomplexes that might be present in the lysate. Therefore, we performed co-IP experiments in cell lysates from ARPE-HPV cells stably expressing eGFP-CCDC32(FL) (Figure 5D). Western blotting showed that the α subunit (Figure 5E and F), together with each of the AP2 complex subunits (Figure 5—figure supplement 1), was efficiently co-precipitated with eGFP-CCDC32(FL), but not by GFP (Figure 5D–F). Thus, we conclude that CCDC32 interacts with the mature AP2 complex. The discrepancy with previous findings may be due to the location of the eGFP tag (see Discussion).

Most endocytic accessory proteins interact with AP2 via the appendage domains of the α and β2 subunits (Praefcke et al., 2004; Schmid et al., 2006; Figure 5A, α-AD and β-AD). Therefore, we expressed and purified GST-tagged AP2 α-AD and β-AD (Figure 5G) and used GST pull-down assays to test which might interact with CCDC32. GST-AP2-α-AD, but not GST-AP2-β-AD, was capable of pulling down CCDC32 from ARPE-HPV eGFP-CCDC32(FL) cell lysate (Figure 5H and I). Thus, CCDC32 can interact with AP2 via the α-AD.

We next identified a region on CCDC32 responsible for AP2 binding. Structural predictions of CCDC32 made by AlphaFold 3.0 show it to be a mainly unstructured protein with several isolated α-helices (Figure 6—figure supplement 1A and B), which likely led to its misnomer as a coiled-coil domain containing protein. Indeed, even when modeled as a dimer or trimer, we found no evidence of coiled-coil interactions between the α-helices (data not shown). Moreover, endogenous CCDC32 did not co-immuno-precipitate with eGFP-CCDC32 from ARPE-HPV eGFP-CCDC32(FL) cell lysate (Figure 6—figure supplement 1C), indicating that the protein is a monomer in vivo. Nonetheless, we focused first on a strongly predicted α-helix encoded by aa78-98 and located in the middle of CCDC32 and engineered a CCDC32 construct lacking this α-helix (∆78–98, Figure 6A; Figure 6—figure supplement 1A and B). The corresponding siRNA-resistant CCDC32(∆78–98) mutant was stably expressed in ARPE-HPV cells at comparable expression levels to eGFP-CCDC32(FL) (Figure 1—figure supplement 1A). As previously shown (Figure 5—figure supplement 1), all four subunits of AP2 efficiently co-IP with full-length CCDC32 from cell lysates (Figure 6B and C). Interestingly, the co-IPs of both α and σ2 subunits with CCDC32(∆78–98) were greatly reduced (Figure 6B and C); whereas, unexpectedly, the ability of CCDC32(∆78–98) to interact with β2:µ2 was unaffected. These data are partially consistent with the results of Wan et al., 2024, who reported interactions with the α:σ2 hemicomplex and with µ2, but not with β2. Given the efficiency and selectivity of our co-IP, our data also suggests that the mature AP2 complex exists in a dynamic equilibrium between α:σ2 and β2:µ2 hemicomplexes. Indeed, it has previously been reported that β2:µ2 hemicomplexes can partially support synaptic vesicle recycling in C. elegans bearing null mutations in the α-subunit of AP2 (Gu et al., 2013). These findings demonstrate that aa78-98 are essential for CCDC32 interactions with both mature AP2 complexes and the α:σ2 hemicomplex.

Disrupting AP2 interactions severely impaired the recruitment of CCDC32(∆78–98) to CCPs (Figure 6D and E), indicating that CCDC32 is likely recruited to CCPs through its interactions with AP2. Notably, unlike CCDC32(FL), CCDC32(∆78–98) was unable to rescue either the uptake of the CME cargo, TfnR (Figure 6F) or the stabilization of CCPs (Figure 6G). Together, these results establish that CCDC32, a previously uncharacterized endocytic accessory protein, plays a critical role in regulating CCP stabilization and invagination.

Loss-of-function nonsense mutations in CCDC32 have been reported to result in CFNDS (Abdalla et al., 2022; Harel et al., 2020); however, the disease-causing mechanism remains unknown. The identified frameshift mutations result in premature termination and truncation of the protein at residues 10, 55, or 81 (Harel et al., 2020). Based on our findings, we hypothesize that these C-terminally truncated mutants, all of which lack the α-helix encoded by residues 78–98, will be defective in AP2 binding and unable to function in CME. To test this hypothesis, we generated an siRNA-resistant eGFP-CCDC32(1-54) construct (Figure 7A) based on the clinical report (Harel et al., 2020), and stably expressed this truncated CCDC32 mutant in ARPE-HPV cells (Figure 1—figure supplement 1A).

As predicted, AP2 did not co-IP with CCDC32(1-54) from the cell lysate (Figure 7B and C), confirming that the disease-causing mutation in CCDC32 abolishes its interactions with AP2. Correspondingly, we could not detect CCDC32(1-54) recruitment to CCPs in dual channel TIRFM imaging (Figure 7D and E), consistent with CCP recruitment of CCDC32 being dependent on CCDC32-AP2 interactions. Interestingly, we also noticed that the level of diffuse plasma membrane binding detected by TIRFM decreased to a greater extent than that seen for CCDC32(∆78–98), now corresponding to background, eGFP only levels (see Figure 7—figure supplement 1 for direct comparison). These data suggest that the C-terminus of CCDC32 is required for binding to the inner surface of the plasma membrane. Finally, we observed that expression of eGFP-CCDC32(1-54), after siRNA-mediated knockdown of endogenous CCDC32, was unable to rescue TfnR uptake efficiency (Figure 7F) or CCP stabilization (Figure 7G). These findings show that this loss-of-function nonsense mutation in CCDC32 abolishes its interactions with AP2 and inhibits CME, likely contributing to the development of CFNDS.

The eGFP-CCDC32(FL, human) cDNA in a pEGFP-C1 vector was purchased from Addgene (#110505) and then mutated to be siRNA resistant, which retained amino acid sequence (#98–102) while modifying the nucleotide sequence. Next, aa78-98 was deleted from siRNA-resistant eGFP-CCDC32(FL) to generate eGFP-CCDC32(∆78–98). Finally, eGFP, eGFP-CCDC32(1-54), eGFP-CCDC32(∆78–98), and eGFP-CCDC32(FL) were separately cloned into a pLVx-CMV100 vector (Dean et al., 2016) using NEBuilder HiFi DNA Assembly Master Mix (Catalog #E2621). A full list of primers used for mutagenesis and cloning is available in Table 2. Note that our disease mimic construct CCDC32(1-54) does not contain a 9 aa peptide (VRGSCLRFQ) in the N-terminus and an extra 12 aa in the C-terminus when CFNDS patient mutation was described (p.(Glu64Glyfs∗12)) in Harel et al., 2020.

In addition, mRuby-CLCa in a pLVx-IRES-puro vector was generated in our previous study (Srinivasan et al., 2018). AP2-α-AD (aa701-938, mouse) in a pGEX-2T-1 vector and AP2-β-AD (aa592-937, rat) in a pGEX-4T-1 vector were kind gifts of the late Linton Traub (University of Pittsburgh, PA).

All cell lines used in this study were routinely tested and confirmed negative for mycoplasma contamination. Cell line identities were authenticated by ATCC. None of the cell lines used are listed among commonly misidentified cell lines maintained by the International Cell Line Authentication Committee.

ARPE-19 and ARPE19-HPV16 (herein called ARPE-HPV) cells were obtained from ATCC and cultured in DMEM/F12 (Gibco, Catalog #8122502) with 10% FBS. HEK293T cells were obtained from ATCC and cultured in DMEM (Gibco, Catalog #8122070) with 10% FBS. ARPE-HPV cells that stably express eGFP-CLCa were generated in our previous study (Chen et al., 2020). ARPE-HPV cells that stably express a fully functional AP2-α-eGFP, in which eGFP is inserted into the flexible linker of AP2, at aa649, were generated in our previous study (Mino et al., 2020), and this AP2-α-eGFP construct has been shown to be fully functional.

Lentiviruses encoding mRuby-CLCa were produced in HEK293T packaging cells following standard transfection protocols (Kutner et al., 2009) and harvested for subsequent infections to ARPE-HPV cells to generate ARPE-HPV mRuby-CLCa cells. Lentiviruses encoding eGFP and siRNA-resistant, eGFP-CCDC32(1-54), eGFP-CCDC32(∆78–98), and eGFP-CCDC32(FL) were produced in HEK293T packaging cells and harvested for subsequent infections to ARPE-HPV mRuby-CLCa cells. Infected cells were FACS sorted for homogenous mRuby and eGFP signals after 3 days and passaged for 2 weeks before experiments.

For siRNA-mediated knockdown of CCDC32, ARPE-HPV cells stably expressing eGFP-CLCa were seeded on 6-well plates (250,000 cells/well) and transfected with 2 rounds of siCCDC32 (Silencer Select Pre-designed siRNA ID#: s228444, targets aa98-102 sequence) through 3 days. Cells treated with siControl (Silencer Select Negative Control #1 siRNA, cat#:4390843) were used as negative control. Transfections of siRNA were mediated with Opti-MEM and Lipofectamine RNAi-MAX (Invitrogen) as detailed in Chen et al., 2020.

For CCDC32 rescue experiments, ARPE-HPV cells stably expressing eGFP or siRNA-resistant eGFP-CCDC32(FL) eGFP, eGFP-CCDC32(∆78–98 or eGFP-CCDC32(1-54)) were transfected with siCCDC32 or siControl as described above. Anti-GFP Monoclonal Antibody (Proteintech, # 66002–1-Ig), Anti-Vinculin Polyclonal Antibody (Proteintech, #26520–1-AP), and anti-C15orf57 Polyclonal antibody (Invitrogen, #PA5-98982) were used in Western Blotting to confirm protein expression level and knockdown efficiency.

ARPE-HPV cells with stable expression of AP2-α-eGFP, eGFP, eGFP-CCDC32(1-54), eGFP-CCDC32(∆78–98) or eGFP-CCDC32(FL) were cultured in a 15 cm dish. When confluence reached ~90%, cells were washed 3 x with ice-cold PBS and detached with a cell scratcher. Next, cells were spun down at 500 × g, 4 °C for 3 min and then resuspended in 1 ml ice-cold lysis buffer (50 mM Tris, pH 7.5, 150 mM NaCl, mM EDTA, 0.5% Triton X-100, 1×protease inhibitor, 1 mM PMSF). After 30 min rotation in a cold room and occasional vortexing, the lysed cells were spun at 500 ×g, 4 °C, for 3 min to remove the nuclei. Subsequently, 50 µl anti-GFP magnetic beads (Biolinkedin, #L1016) were added to cell lysate that contained 0.5 mg proteins (determined by BCA assay). The reaction was allowed to proceed by rotation at 4 °C for 2 hr and then precipitated with DynaMag. The sediments were washed with lysis buffer, and then resuspended and heat-denatured in 2 x Laemmli buffer (supplemented with 5% β-Mercaptoethanol). The final samples were run into SDS-PAGE gels before: (1) being transferred to membranes for Western Blotting; or (2) being sent to the proteomics core facility for mass spectrometry analysis. Anti-α-Adaptin 1/2 Antibody (C-8) (Santa Cruz Biotechnology, Catalog #sc-17771) was used in WB to detect AP2 enrichments.

Samples were digested overnight with trypsin (Pierce) following reduction and alkylation with DTT and iodoacetamide (Sigma–Aldrich). Following solid-phase extraction cleanup with an Oasis HLB µelution plate (Waters), the resulting peptides were reconstituted in 10 µl of 2% (v/v) acetonitrile (ACN) and 0.1% trifluoroacetic acid in water. 2 µl of each sample were injected onto an Orbitrap Fusion Lumos mass spectrometer (Thermo Electron) coupled to an Ultimate 3000 RSLC-Nano liquid chromatography systems (Dionex). Samples were injected onto a 75 μm i.d., 75 cm long EasySpray column (Thermo), and eluted with a gradient from 0% to 28% buffer B over 90 min. Buffer A contained 2% (v/v) ACN and 0.1% formic acid in water, and buffer B contained 80% (v/v) ACN, 10% (v/v) trifluoroethanol, and 0.1% formic acid in water. The mass spectrometer operated in positive ion mode with a source voltage of 2.4 kV and an ion transfer tube temperature of 275 °C. MS scans were acquired at 120,000 resolution in the Orbitrap and up to 10 MS/MS spectra were obtained in the Orbitrap for each full spectrum acquired using higher-energy collisional dissociation (HCD) for ions with charges 2–7. Dynamic exclusion was set for 25 s after an ion was selected for fragmentation.

Raw MS data files were analyzed using Proteome Discoverer v2.4 (Thermo), with peptide identification performed using Sequest HT searching against the human reviewed protein database from UniProt. Fragment and precursor tolerances of 10 ppm and 0.6 Da were specified, and three missed cleavages were allowed. Carbamidomethylation of Cys was set as a fixed modification and oxidation of Met was set as a variable modification. The false-discovery rate (FDR) cutoff was 1% for all peptides. To generate the volcano plot, Tubulin Beta 6 was used for normalization.

GST in a pGEX-6P-1 vector, GST-AP2-α-AD in a pGEX-2T-1 vector, and GST-AP2-β-AD in a pGEX-4T-1 vector were transfected and expressed in BL21(DE3) separately, and then affinity purified using GSTrap HP column (Cytiva). The affinity-purified GST fusion proteins were applied to a HiLoad 16/600 Superdex 200 pg column (Cytiva). Target proteins were collected and concentrated using Amicon Ultra-15 10 K Centrifugal filters (Sigma-Aldrich), and stored in 20 mM HEPES, 150 mM NaCl, 1 mM TCEP, pH 7.4.

In GST pull-down assays, purified GST, GST-AP2-α-AD, and GST-AP2-β-AD were pre-bound to anti-GST beads (Biolinkedin, #L-2004). Subsequently, these beads were separately added into 0.5 mg cell lysate of ARPE-HPV eGFP-CCDC32(FL). The reactions were allowed to proceed by rotation at 4 °C for 2 hr, and then the beads were spun down at 1000 × g, 4 °C for 3 min. The sediments were washed with lysis buffer and then heat-denatured in 2 x Laemmli buffer (supplemented with 5% β-Mercaptoethanol) before running into SDS-PAGE gels, which were followed by western blotting analysis.

Internalization of TfnR was quantified by in-cell ELISA following established protocols described (Chen et al., 2020; Conner and Schmid, 2003; Reis et al., 2015; Srinivasan et al., 2018). Briefly, ~15,000 cells were seeded on gelatin-coated 96-well plate overnight. The next day, cells were starved in 37°C-warm PBS⁴⁺ (1×PBS buffer plus 0.2% BSA, 1 mM CaCl₂, 1 mM MgCl₂, and 5 mM D-glucose) for 30 min. After starvation, the cell medium was replaced with ice-cold PBS⁴⁺ containing 5 µg/ml HTR-D65 (anti-TfnR mAb; Schmid and Smythe, 1991). In parallel, some cells were kept at 4 °C for the measurement cell surface TfnR (denoted as S) and blank controls (denoted as B), while some were incubated in a 37 °C water bath for the indicated times for the measurement of internalized TfnR (denoted as I). For B and I, acid wash (0.2 M acetic acid and 0.2 M NaCl, pH 2.3) was applied to remove surface-bound HTR-D65. After washing with cold PBS, all cells were fixed with 4% PFA (Electron Microscopy Sciences) diluted in PBS for 30 min at 37 °C. Subsequently, the cells were permeabilized with 0.1% Triton X-100 and blocked with Q-PBS (PBS, 2% BSA, 0.1% lysine, and 0.01% saponin, pH 7.4) for 2 hr. Surface-bound and internalized HTR-D65 were detected with HRP Goat anti-Mouse IgG (H+L) (Bio-Rad) and o-phenylenediamine dihydrochloride (OPD, Sigma-Aldrich). Well-to-well variation of cell numbers was accounted for by BCA assays.

Cells for imaging were seeded on gelatin-coated glass bottom dishes 35 mm (ibidi, #81218–800) for ~12 hr before data acquisition. Live cell imaging was conducted with a Nikon Eclipse Ti2 inverted microscope that was equipped with: (1) an Apo TIRF/100x1.49 Oil objective; (2) a Prime Back Illuminated sCMOS Camera (Prime BSI, 6.5x6.5 µm pixel size and 95% peak quantum efficiency); (3) a M-TIRF module for epifluorescence (Epi) acquisition; (4) an H-TIRF module for TIRF acquisition, where penetration depth was fixed to 80 nm; (5) an Okolab Cage Incubator for maintaining 37 °C and 5% CO₂.

For time-lapse TIRF imaging, 451 consecutive images were acquired at a frame rate of 1 frame/s for single channel or 0.5 frame/s for dual channels. For time-lapse Epi-TIRF imaging, 451 consecutive Epi and TIRF images were acquired nearly simultaneously at a frame rate of 0.66 frame/s. Perfect Focus System (PFS) was applied during time-lapse imaging.

The acquired data was analyzed using cmeAnalysis (Aguet et al., 2013; Jaqaman et al., 2008) and DASC (Wang et al., 2020). Briefly, cmeAnalysis was used to track the lifetime and fluorescence of clathrin-coated structures, and then DASC was applied to unbiasedly classify bona fide CCPs vs. abortive coats. Next, the classified CCP tracks were used to calculate CCP invagination Δz\begin{document}$\Delta z$\end{document}, as previously described (Wang et al., 2020). Tracks that overlap with others or deviate from the properties of a diffraction-limited particle were excluded from the analysis.

The ‘unroofing’ technique that mechanically removes the upper cell plasma membrane while preserving the ventral plasma membrane with CCSs and other structures for PREM analyses was performed essentially as described previously (Yang et al., 2022). Briefly, ARPE19 cells treated with siCCDC32 or siControl on coverslips were quickly transferred into ice-cold PEM buffer (100  mM PIPES−KOH, pH 6.9, 1  mM MgCl₂ and 1  mM EGTA) containing 2  µM unlabeled phalloidin (Sigma, #P2141) and 10  µM taxol (Sigma-Aldrich, #T7402) and unroofed by a brief (1 s) ultrasonic burst from a 1/8-inch microprobe positioned at ~45° angle ~3 mm above the coverslip and operated by Misonix XL2020 Ultrasonic Processor at 17–20% of output power. After sonication, the coverslips were immediately fixed with 2% glutaraldehyde in 0.1 M Na-cacodylate buffer, pH 7.3 for at least 20 min at room temperature.

Sample processing for PREM was performed as described previously (Svitkina, 2022). In brief, glutaraldehyde-fixed cells were post-fixed by sequential treatment with 0.1% tannic acid and 0.2% uranyl acetate in water, critical-point dried, coated with platinum and carbon, and transferred onto EM grids for observation.

PREM samples were examined using a JEM 1011 transmission electron microscope (JEOL USA, Peabody, MA) operated at 100 kV. Images were acquired by an ORIUS 832.10 W CCD camera (Gatan, Warrendale, PA) and presented in inverted contrast.

Based on the degree of invagination, the shapes of CCSs were classified into: flat CCSs with no obvious invagination; dome-shaped CCSs that had a hemispherical or less invaginated shape with visible edges of the clathrin lattice; and spherical CCSs that had a round shape with the invisible edges of clathrin lattice in 2D projection images. In most cases, the shapes were obvious in 2D PREM images. In uncertain cases, the degree of CCS invagination was determined using images tilted at ±10–20 degrees. The area of CCSs was measured using ImageJ and used for the calculation of the CCS occupancy on the plasma membrane.