Dual clathrin and integrin signaling systems regulate growth factor receptor activation

The crosstalk between growth factor and adhesion receptors is key for cell growth and migration. In pathological settings, these receptors are drivers of cancer. Yet, how growth and adhesion signals are spatially organized and integrated is poorly understood. Here we use quantitative fluorescence and electron microscopy to reveal a mechanism where flat clathrin lattices partition and activate growth factor signals via a coordinated response that involves crosstalk between epidermal growth factor receptor (EGFR) and the adhesion receptor β5-integrin. We show that ligand-activated EGFR, Grb2, Src, and β5-integrin are captured by clathrin coated-structures at the plasma membrane. Clathrin structures dramatically grow in response to EGF into large flat plaques and provide a signaling platform that link EGFR and β5-integrin through Src-mediated phosphorylation. Disrupting this EGFR/Src/β5-integrin axis prevents both clathrin plaque growth and dampens receptor signaling. Our study reveals a reciprocal regulation between clathrin lattices and two different receptor systems to coordinate and enhance signaling. These findings have broad implications for the regulation of growth factor signaling, adhesion, and endocytosis. The integration and organization of growth factor signaling, adhesion, and endocytosis is poorly understood. Here the authors use light and electron microscopy to shed light on the role of flat clathrin lattices and cell adhesion in growth factor signaling.

C ellular communication begins with a cascade of molecular interactions initiated by plasma membrane (PM) receptors of which the receptor tyrosine kinases (RTKs) are one of the major superfamilies. RTKs are ubiquitous integral membrane proteins in eukaryotes that perform numerous actions including the regulation of proliferation, differentiation, survival, and migration 1 . Epidermal growth factor receptor (EGFR) is one of the most widely studied members of the RTK superfamily and regulates epithelial tissue development and homeostasis. In lung, breast, and head and neck cancers, EGFR is a primary driver of tumorigenesis and a major target for therapy 2 . EGFR spans the plasma membrane and contains a ligand-interacting domain facing the extracellular space and a tyrosine kinase domain in the cytoplasm 3 . The binding of EGF to EGFR triggers receptor dimerization and cross-phosphorylation of tyrosine residues within its cytosolic domain 4 . This provides docking sites for the recruitment of scaffold proteins including Grb2, and the activation of downstream tyrosine kinases such as ERK and Src 5,6 . The EGFR pathway can be activated either directly by its cognate ligands or through transactivation by other signaling proteins including integrins 7,8 .
Integrins are adhesion molecules that are responsible for cell-cell and cell-matrix interactions, and relay mechanical signals bidirectionally between the extracellular space and the cytoplasm 9 . In this way, integrins sense the local environment and control tissue rigidity, cell growth, and movement 10 . Dysregulation of integrin signaling contributes to cancer and metastasis 11 . Integrins are heterodimers of α and β subunits, each containing a large multidomain extracellular region (>700 residues) for ligand binding, a single transmembrane helix, and a short cytoplasmic tail (13-70 residues) 12 . β-integrin cytoplasmic tails lack enzymatic activity. Instead, they harbor distinct regulatory sequences including two NPxY motifs (Asn-Pro-x-Tyr) that can be phosphorylated 13 . NPxY motifs have a high affinity for phosphotyrosine-binding domain proteins 14 . This allows them to bind to partners including clathrin-mediated endocytosis (CME) accessory proteins 15,16 . While these proteins are known to interact generally, how they are dynamically and spatially organized at the PM to control signaling and crosstalk is unclear.
EGFR and integrins can be regulated through CME-the main endocytic pathway used by eukaryotic cells to internalize receptors into the cytoplasm 17 . During CME, clathrin and adapters assemble to bend the membrane into Ω-shaped pits 18,19 . Scission of clathrin-coated pits (CCPs) from the PM yields closed spherical vesicles with an average diameter of~100 nm 20 . Besides CCPs, cells exhibit a subset of clathrin coats known as flat clathrin lattices (FCLs) or plaques 21 . In contrast to CCPs, FCLs are longlived on the PM and display a variety of two-dimensional shapes 22,23 . FCLs are abundant in myocytes and can bind cortical actin during muscle formation and function 24 . FCLs are also important for the adhesion between osteoclasts and bone 25,26 . During cell communication, different types of receptors including EGFR are clustered at FCLs 27,28 . Additionally, FCLs are enriched with β5-integrin [29][30][31] . Yet, it is unknown how flat clathrin lattices are regulated and their general functions remain unclear. Furthermore, the signaling pathways responsible for their initiation, growth, maintenance, and disassembly are unknown.
Here, we identified an EGFR/β5-integrin/Src signaling axis that regulates flat clathrin lattice expansion during growth factor stimulation. Using a combination of quantitative fluorescence and electron microscopy, we showed that EGF triggers large ultrastructural changes to the membrane of human squamous carcinoma (HSC3) cells. These changes include the generation and dramatic expansion of large FCLs and required EGFR activation by EGF as well as Src kinase and β5-integrin. Agonist stimulation leads to persistent recruitment of EGFR, Grb2, and β5-integrin into clathrin structures, and a corresponding loss of Src kinase.
We provide evidence of β5-integrin phosphorylation mediated by Src that regulates this signaling domain. These data reveal a mutual regulation of FCLs and two different receptor systems: EGFR and β5-integrins. Thus, an EGFR/β5-integrin/Src axis contributes to the expansion of FCLs, which in turn act as dynamic signaling platforms to partition and enhance growth factor signaling at the adherent plasma membrane of human cells.

Results
EGF triggers changes in the ultrastructure of FCLs. First, we tracked the ultrastructure of clathrin at the plasma membrane during growth factor stimulation. To accomplish this, we used genome-edited HSC3 cells endogenously expressing EGFR tagged with GFP, an established model to study EGFR endocytosis and human EGFR-dependent head and neck carcinoma [32][33][34] . We plated HSC3 cells on collagen-coated glass, grew them for 2 days, starved, and stimulated them with vehicle (Ctrl) or EGF for 2, 5, 15, 30, and 60 min. Then, we mechanically unroofed and fixed cells to directly visualize clathrin at the cytoplasmic face of the PM using platinum replica transmission electron microscopy (PREM) 35 . In these images, we measured the nanoscale structure and distribution of clathrin-coated structures (CCSs) across the ventral PM of control and EGF stimulated cells. Remarkably, CCSs were 4.6-fold more abundant in cells stimulated with EGF for 15 min (Fig. 1a-d and Sup. Fig. 1). This increment in CCSs abundance was also observed in cells grown on fibronectinand laminin-coated glass (Sup. Fig. 1).
PREM allowed us to segment distinct CCSs based on their curvature into three subclasses: (1) flat clathrin lattices (FCLs), i.e., where no curvature is evident (shown in green); (2) dome, i.e., curved but the clathrin edge is visible (blue); and (3) sphere, i.e., highly curved and the edge is not observable (magenta) (Fig. 1a-d and Sup. Fig. 1) 36 . Representative segmentation masks illustrate the densities of diverse CCSs densities at different time points of EGF stimulation (Fig. 1e). We compared the differences in densities by quantifying the membrane occupation of all CCSs (Sup. Fig. 1) and individual subclasses (Fig. 1f-h). All CCSs in control cells occupied 0.84 ± 0.2% membrane area and markedly increased 2.5-, 4.6-and 5-fold after EGF stimulation for 5, 15, and 30 min, respectively. The area of all CCSs decreased to nearbaseline levels after 60 min (0.67 ± 0.33%) (Sup. Fig. 1). Notably, membrane area occupation of FCLs followed a similar time course to all CCSs classes, with a peak at 15 and 30 min post-EGF (3.2 ± 0.58%, 6.5-fold; 3.3 ± 1.92%, 6.7-fold), reaching nearbaseline levels at 60 min (0.5 ± 0.22%) (Fig. 1f). In contrast, the mean membrane occupation of domed and spherical clathrin structures remained below 1% with no significant differences across the times tested (Fig. 1g, h).  1 EGF modifies the ultrastructure of clathrin at the plasma membrane. a Montaged PREM image of an unroofed control HSC3-EGFR-GFP cell and the mask created after segmentation of the full membrane outlined (yellow). Flat, dome, and sphere clathrin-coated structures (CCSs) are shown in green, blue, and magenta, respectively. b High-magnification image of the cropped PREM in (a); the different segmented CCSs are color-coded as in (a), with grayscale in magnified insets. c Montaged PREM image of an unroofed HSC3-EGFR-GFP cell treated with 50 ng/mL EGF for 15 min and the mask created after the segmentation. d High-magnification of a representative region of the PREM in (c), the magnification insets are shown at the same scale and are outlined with dashed squares in each image. e Representative clathrin masks of the EGF stimulation time course for 0, 2, 5, 15, 30, and 60 min. PREM images corresponding to the masks and cropped images are shown in Supplementary Fig. 1. f-h Morphometric analysis of the percentage of plasma membrane (PM) area occupation for each CCS subtype. I-shaped box plots show median extended from 25th to 75th percentiles, and minimum and maximum data point whiskers with a coefficient value of 1.5. i-k Morphometric analysis of the size of flat, dome, and sphere CCSs during the EGF time course for the three clathrin subtypes. Dot plots show every structure segmented, the bar is the median. FCLs were substantially larger after EGF treatment (Fig. 1a-d). The 95% confidence interval range of the size of hundreds of FCLs increased starting at 5 min EGF treatment (34023.3-40528.6 nm 2 ), with a peak after 15 min (34197.1-40416.1 nm 2 ) and 30 min (41783.9-49003.3 nm 2 ) and approached control levels after 60 min (25758.9-32889.3 nm 2 ) (Fig. 1i). The average size of domed (21,993.9 ± 4294.8 nm 2 ) and spherical clathrin structures (11,075.2 ± 1714.7 nm 2 ) were similar to their respective controls across the time course (Fig. 1j, k). Thus, EGF increased both the size and total area specifically occupied by FCLs at the ventral PM.
Does clathrin behave differently on the top of the cell? To answer this question we modified the PREM protocol to image the structure of the inner face of the dorsal, non-adherent, PM. All CCSs occupation was~30% less in control dorsal PMs  Fig. 2). Of note, EGF did not trigger an increase in the total amount of CCSs at the top of the cell, but a specific two-fold increase in the proportion of FCLs was measured. Interestingly, we did not observe an expansion in the size of individual FCLs in response to EGF at the dorsal surface (Sup. Fig. 2). Together, our nanoscale analysis revealed a direct connection between the relative abundance of FCLs at the dorsal surface and a dramatic expansion in both the size and abundance of FCLs at the ventral adherent surface of the plasma membrane in response to EGF stimulation.
EGFR, Src, and β5-integrin are required for FCL formation. EGF triggers phosphorylation cascades that activate distinct cellular effectors. We considered that signaling can regulate FCL expansion during growth factor stimulation. To identify the possible members of the EGFR signaling pathway involved in this process, we performed both pharmacological and genetic screens. First, we targeted EGFR using gefitinib (Gefi), a drug that blocks receptor activity by binding to the ATP-binding pocket in the tyrosine kinase domain 37 . PREM of HSC3 cells preincubated with gefitinib followed by EGF treatment showed a percentage of FCLs membrane area occupation (Gefi + EGF = 0.66 ± 0.13%) similar to control cells (Ctrl = 0.48 ± 0.11%) (Fig. 2a-c). Gefitinib decreased the effect of EGF on FCL formation by 4.8-fold (Fig. 2f). We observed a similar effect with EGFR knockdown cells (EGFR siRNA + EGF) (Fig. 2d, f). Representative segmentation masks illustrate the structure and populations of CCSs in the different conditions analyzed (Fig. 2e). Both approaches that interfered with EGFR showed a consistent inhibitory effect on the expansion of FCLs (Fig. 2g).
We then aimed to evaluate the role of the tyrosine kinase Src for two reasons. First, Src is a master effector immediately downstream of EGFR. Second, Src has been reported to directly phosphorylate proteins key to the endocytic machinery [38][39][40] . We used PP2, a specific Src family kinase inhibitor 41 , and then challenged the cells with EGF. We observed a 4.1-fold decrease in FCLs abundance after drug treatment compared to EGF alone, and similar as compared to control cells (PP2 + EGF = 0.77 ± 0.24%) (Fig. 3a-f). Src knockdown (Src siRNA + EGF) showed similar effects on preventing the expansion of FCLs area and size after EGF stimulation (Fig. 3f, g).
Next, we targeted the β5-integrin, a known component of FCLs 16,29 . We treated cells with cilengitide acid (CTA), a molecule that specifically prevents β5-integrin interaction with the extracellular matrix 42 or a siRNA to knock down β5-integrin followed by EGF stimulation (Fig. 4a-e). We detected a 2.2-fold inhibition of FCL formation by CTA (CTA + EGF = 1.48 ± 0.73%), and a similar effect with β5-integrin siRNA (β5 siRNA + EGF = 0.77 ± 0.24%) (Fig. 4f). Both approaches targeting β5-integrin showed a similar inhibitory effect on the growth of FCLs after EGF treatment (Fig. 4g). Neither the drugs nor the different siRNAs, changed the percentage of membrane area or the size of flat, domed, and spherical clathrin structures, in the presence or absence of EGF (Sup. Figs. 4,5). β5-integrin interacts with actin through adapter proteins, and the actin cytoskeleton has been implicated in the regulation of FCLs 28,43,44 . Therefore, we next used Cytochalasin D (CytoD), an acting disrupting drug, to test if perturbing filamentous actin affects clathrin structure. Platinum replica EM showed that CytoD had an inhibitory effect on the relative growth of FCLs in cells treated with EGF (Sup. Fig. 6). FCLs are also known to control the activation of downstream signals such as the mitogen-activated protein kinase ERK1/2 29,45 . Yet, we did not detect major differences in the ultrastructure of clathrin in cells treated with a specific ERK1/2 inhibitor after EGF stimulation (Sup. Fig. 7). Overall, these data indicated that activated EGFR, the downstream tyrosine kinase Src, actin, and β5-integrin are all required for EGF-induced expansion of FCLs.
EGFR, Src, and β5-integrin are differentially located in CCSs. We envisioned that EGFR, Src, and β5-integrin are spatially located in close proximity to FCLs to regulate their formation during EGF signaling. Using total internal reflection fluorescence microscopy (TIRFM) along with a high-throughput two-color correlation analysis 46 , we mapped the location of EGFR, Src, and β5-integrin at thousands of individual clathrin-coated structures in an unbiased manner (see Methods). Correlation analysis between single clathrin sites and the protein of interest measured as the degree of correlation (correlation coefficient) is a quantitative local measure of spatial colocalization across the entire population of structures [46][47][48] .
First, we assessed the correlation between EGFR and clathrin. To do this, we used the genome-edited HSC3 endogenously expressing EGFR-GFP and transfected them with mScarlet tagged clathrin light chain A (mScarlet-CLCa). Figure 5a shows clathrin as diffraction-limited puncta, whereas EGFR appears as a more diffuse fluorescent signal across the ventral PM in control cells (C = 0.15 ± 0.06). EGF stimulation caused the appearance of pronounced EGFR clusters and a 2.4-fold increase in its correlation with clathrin (C = 0.36 ± 0.15) (see white in overlay). We observed a similar increase in wild-type (WT) HSC3 cells cotransfected with EGFR-GFP and mScarlet-CLCa stimulated with EGF (Sup. Fig. 8). We then imaged control cells expressing Src-GFP that showed mottled fluorescence (Fig. 5a). The Src signal correlated with clathrin (C = 0.41 ± 0.1), but decreased 2.2-fold after EGF treatment for 15 min (C = 0.18 ± 0.1) (Fig. 5b). We further confirmed the loss of Src from CCSs by showing that the correlation between Src and clathrin continuously decreases at intervals between 0 and 15 min of stimulation with EGF (Sup. Fig. 9).
Interestingly, the β5-integrin signal appeared as discreet puncta that markedly correlated with flat, domed, and spherical clathrin as shown by TIRF correlative platinum replica electron microscopy images (Sup. Fig. 10). Correlation, however, remains high in both control (C = 0.79 ± 0.11) and EGF-treated cells (C = 0.76 ± 0.07) (Fig. 5a, b). Thus, EGFR, Src, and β5-integrin differentially associate with clathrin. To further confirm this, we measured the EGFR and Src signal using β5-integrin as a reference. Similarly, EGF stimulation increased the β5-integrin correlation with EGFR but decreased correlation with Src (Sup. Fig. 11). Of note, we tested other members of the β-integrin subfamily including β1, β3, and β6 as well as αV, the most frequent partner of β5-integrin. However, the correlation values were 2-to 4-fold smaller when compared to the β5-integrin correlation with clathrin (Sup. Fig. 12). We also found that the correlation between clathrin and F-tractin, an actin filament marker, decreases after 15 min of stimulation with EGF. However, we observed no substantial change in colocalization between this probe and clathrin in cells pretreated with CytoD, neither in control cells nor EGF stimulated cells (Sup. Fig. 6). Altogether, these data support the idea that EGF leads to persistent recruitment of EGFR and β5-integrin into CCSs with a concurrent loss of Src.
EGFR and β5-integrin are connected through phosphorylation. Next, we assessed the biochemical connection between EGFR, Src, and β5-integrin. EGF binding elicits EGFR dimerization leading to Src tyrosine kinase activation 49 . β5-integrin cytoplasmic domain has three tyrosine residues (Y766, Y774, and Y796) that are highly conserved among their orthologues and paralogues (Sup. Fig. 13). In silico analysis suggested Src, among other kinases, has a substantial likelihood of catalyzing phosphorylation at all tyrosines of the β5-integrin cytoplasmic domain (Sup. Fig. 13). We, therefore, hypothesized that these proteins form a signaling loop that can regulate clathrin remodeling through phosphorylation. To support this hypothesis, we used a luciferase-coupled system to measure the phosphorylation of synthetic peptides corresponding to the β5-integrin cytoplasmic domain (Fig. 6a). We detected a 13.6-fold increase in the phosphorylation of the WT peptide in the presence of Src, as compared to the negative control containing only the WT peptide ( Fig. 6b). By contrast, this effect was reduced when we tested the non-phosphorylatable peptide 3Y-F. As a positive control, we incubated a peptide corresponding to amino acids 6-20 of p34 cdc2 , a well-characterized Src substrate 50 , in the presence of purified and active Src, and we detected phosphorylation. While β5-integrin has been reported to be phosphorylated by PAK4 at S759 and S762 51 , we did not detect phosphorylation when the WT β5-integrin peptide was incubated with purified PAK4 (Fig. 6b). These data indicate that the intracellular domain of β5integrin is a Src substrate in vitro.
FCLs regulate adhesion and sustained signaling at the PM. Next, we used Interference Reflection Microscopy (IRM) to test if CCSs are cell adhesion sites. In IRM, the adhesion of a cell to the surface is measured as a change in intensity due to destructive interference of reflected light between the cell/substrate interface 52 . In inverted images, the plasma membrane of unroofed cells transfected with mSca-CLCa appeared bright against a dark background indicating adhesion (Sup. Fig. 14). In these images, bright spots are an indicator of how tight the membrane adheres to the substrate. In parallel, we used TIRF to visualize CCSs and then automatically measured the correlation between the clathrin signal and the IRM signal across thousands of individual clathrin spots. IRM images showed a distinctive IRM signal at sites of clathrin, indicating increased amounts of adhesion relative to the surrounding membrane. We detected a subtle but measurable increase in clathrin/IRM correlation (adhesion) between control and EGF-treated cells (Sup. Fig. 14). Our IRM data directly show that single clathrin sites are nanoscale adhesion sites.
We hypothesized that FCLs mediate EGFR membraneassociated signal transduction by regulating the distribution of active EGFR and downstream interactors in both space and time. Using TIRFM, we next visualized the presence of active EGFR at the PM by immunolabeling endogenous phosphor-Y1068 (P-EGFR), a well-established marker of EGFR activity (Fig. 7a). In stimulated cells, the correlation of P-EGFR with clathrin increased 3.5-fold (C = 0.74 ± 0.08) with respect to the control (C = 0.24 ± 0.06). Conversely, disruption of FCLs formation by β5-integrin knockdown (β5 siRNA + EGF), decreased the P-EGFR correlation with clathrin by~25% (C = 0.56 ± 0.1) (Fig. 7b). A similar trend was observed when we measured the fluorescence intensity of P-EGFR at the PM (Fig. 7c). Furthermore, we measured the fluorescence intensity of the total EGFR-GFP signal (T-EGFR) at the PM as an indicator of receptor internalization. We observed that EGF led to a 34.6% decrease of the T-EGFR, and β5-integrin siRNA further decreased receptor levels at the PM (68.2%) (Fig. 7d). We also examined the location of the downstream master scaffold Grb2 in FCLs (Fig. 7e). While EGF caused an increase in both Grb2 correlations with clathrin and recruitment of the adapter to the PM, β5-integrin knockdown blocked these effects (Fig. 7f, g). These observations were reproduced when interfering with β5-integrin using a pharmacological approach (Sup. Fig. 15). Together these data highlight the function of FCLs as sustained signaling hubs for growth factor receptors. These results suggest that FCLs capture and organize growth factor signals at the ventral PM by delaying the endocytosis of a population of active EGFR along with key partner proteins such as Grb2. These molecular events are

Discussion
Crosstalk between signaling systems allows for biological processes to be integrated, responsive, and adaptable 53 . There is an emerging hypothesis that FCLs can act as signaling zones or adhesion sites at the PM, filling unique roles outside of endocytosis. Here, we show that activated EGFR, Src, and β5-integrin are coupled to dramatic growth and maintenance of FCLs in human cells. These flat clathrin sites in turn partition and enhance growth factor signals at the PM (Fig. 8). Thus, two receptor systems (growth factor and integrins) are connected, clustered, and controlled at the nanoscale by endocytic proteins. We propose that a reciprocal feedback loop operates where FCLs facilitate local crosstalk between EGFR, β5-integrin, and other signaling proteins to create dynamic signaling hubs at the PM. First, we observed clathrin using platinum replica electron microscopy to distinguish flat from curved clathrin structures. Surprisingly, EGFR activation resulted in a dramatic increase in FCLs size and density. Domed and spherical clathrin were mostly unchanged. Blocking the receptor abolished these effects. Of note, these structural changes follow a time course similar to the activation dynamics of downstream kinases 54 , further supporting the direct connection between clathrin remodeling and signal transduction. Although we observed changes in the number of caveolae across the membrane, this represents an exciting opportunity for future research. Our pharmacological and genetic perturbations linked the tyrosine kinase Src and the adhesion receptor β5integrin to FCL growth and EGFR activation. Historically, growth factor receptors and integrins have been biochemically connected to Src in several ways 7 . Here, we show a direct spatial connection. We also found that FCLs are preloaded with a subpopulation of Src. Src was partially released from the complex over time in response to EGF. In contrast, β5-integrin is enriched in FCLs 16,29,30 , and we found that this correlation with clathrin persists in response to EGF. Thus, EGFR, Src, and β5-integrin are dynamically coupled through FCLs to regulate EGF signaling.
We showed that this new pathway is controlled by phosphorylation. Specifically, in silico analysis and biochemical assays indicated that the β5-integrin intracellular domain is a Src substrate. Src activation kinetics are fast and occur within 5 min 55 . Thus, it is possible that at an early stage of clathrin growth, Src phosphorylates targets and is then released. While our results point toward an early phosphorylation event in β5-integrinmediated by Src, it is also possible that Src continually cycles on and off clathrin during receptor activation and plays a more extended role in the process. Likewise, Src is a wide-ranging kinase and might phosphorylate additional substrates located on other PM structures such as caveolae and focal adhesions 56,57 . Additionally, Src is known to phosphorylate clathrin heavy chain 39 , whether this directly regulates clathrin assembly and dynamics is an exciting opportunity for future experimentation.
Second, we found that deletion of the β5-integrin cytoplasmic domain and non-phosphorylatable mutations (3Y-F) block integrin association with clathrin. These effects were rescued by a phosphomimic mutant (3Y-E), suggesting that tyrosine phosphorylation of β5-integrin is a molecular switch in this process. Interestingly, the β5-integrin cytoplasmic domain interacts with endocytic proteins including Eps15, ARH, and Numb 16 . Integrin cytoplasmic tails can also induce profound differences in the behavior of integrins 58 . Thus, we propose that a phosphorylation switch in β5-integrin is the regulator for the orchestrated recruitment of the endocytic machinery to sites where FCLs form and grow. In this regard, the growth factor response is directly linked to cellular adhesion proteins through activation of endocytic proteins and controlled by phosphorylation.
A recent hypothesis proposes that long-lived FCLs arise from adhesive forces generated from integrins that physically prevent clathrin from curving, a process called frustrated endocytosis 29,22,45 . Interestingly, we observed the FCL formation and the enrichment of active EGFR and Grb2 peaks after 15 min of stimulation with EGF ( Figs. 1 and 7). At the same time, we detected a decrease in the overall level of EGFR at the PM. This decrease suggests that EGF triggers the internalization of a subpopulation of EGFR into endosomal compartments known to control the specificity and robustness of the cytoplasmic signaling response 59,60 . Signaling is thought to continue in these endosomal carriers in some cases. In parallel, we propose that some phosphorylated and active receptors remain at the PM in clathrin. By preventing endocytosis of a subpopulation of EGFR, this clathrin/ adhesion/receptor complex could prolong EGF signaling at the membrane. These domains might also act as diffusion traps for additional EGFR and other growth factor receptors whose diffusion decreases after agonist stimulation [61][62][63][64] . Using fluorescence microscopy, EGFR and other structurally diverse receptors have been reported to form long-lived complexes 22,[65][66][67][68] . In contrast, stimulation of the LPA1 receptor has been shown to trigger the depolymerization of FCLs through a process that depends on actin cytoskeleton polymerization 28 . Thus, different systems might activate or deactivate these structures to regulate their activity. How endocytosis, adhesion, receptor diffusion, and actin cytoskeletal dynamics cooperate across the entire population of active receptors to control signaling will be an important future area of study. Furthermore, how the formation of these domains impacts the overall cell and tissue signaling behaviors are important areas for future study.
Interestingly, we find measurable differences in the growth and size of flat clathrin lattices between the ventral and dorsal plasma membrane of the cell. We propose that these differences are a direct result of the interaction between the β5-Integrin and the extracellular matrix. These interactions would not occur along the dorsal surface, further supporting a direct mechanical role of adhesion to the maintenance and expansion of these growth factor signaling domains. This bimodal behavior adds a layer of complexity to the EGFR signaling system in cells. Here, the top and bottom of the cell could be acting differently in response to EGF and the cumulative effects could influence signaling in complex ways. Future work on the integration of different regional cellular signals into global signaling outcomes are key to understanding how cells and tissues respond to EGF.
Is this mechanism unique to EGFR? Many receptors including seven-transmembrane receptors and B cell receptors trigger clathrin nucleation at the PM upon biding their ligands 22,[65][66][67][68] . For the β 2 -adrenergic receptor, the increase in clathrin occurs with a delay in clathrin-coated vesicle maturation and no differences in the overall rate of vesicle scission 66 . Our data reveal a similar Fig. 4 Flat clathrin lattice formation requires β5-integrin. a Representative PREMs of control HSC3-EGFR-GFP cells (Ctrl), b treated either with 50 ng/mL EGF alone (EGF) for 15 min, or in presence of c 10 μM cilengitide acid (CTA + EGF), and d β5-integrin siRNA (β5 siRNA + EGF). β5-integrin siRNA validation is shown in Supplementary Fig. 3. The magnification insets are shown at the same scale and are outlined with dashed squares in each image. Flat, dome, and sphere clathrin-coated structures (CCSs) are shown in green, blue, and magenta, respectively, with native grayscale in magnified insets. e Representative masks of segmented cells treated as in (a-d). PREM images corresponding to the masks and cropped images are shown in Supplementary  Fig. 4 increase in clathrin nucleation during EGF stimulation, but the major changes to clathrin occur specifically and exclusively with dramatic growth in FCLs. It is possible that other receptor cargos also stabilize flat clathrin coats to act as nanoscale receptor signaling domains. Thus, FCLs could be generalized signaling hubs at the PM. Future work is needed to test this hypothesis and to map the nanoscale location of the proteins identified here and other possible interactors within the clathrin lattice.
Growing evidence suggests that signaling systems are locally organized by organelles and cytoskeletal structures. We propose that FCLs are a unique plasma membrane scaffold that dynamically organizes receptors and signaling molecules in space and time through multivalent interactions at the nanoscale. We suggest that crosstalk between EGFR and β5-integrin through Src phosphorylation occurs in FCLs and simultaneously regulates their expansion and maintenance. Finally, because of the importance of EGFR/Src/β5-integrin in the physiology of cells and tissues, and the connection between dysregulation of these systems and cancer, FCLs likely play a broader role in coordinating the responses to chemical and mechanical stimuli. For example, how the growth and regulation of these clathrin signaling domains impacts cell motility, metastasis, growth, division, gene expression, or differentiation are important areas for future investigations. Likewise, how these clathrin adhesion signaling domains might behave in complex tissues are unknown. Future work is needed to provide insight into how these signaling platforms dynamically organize, coordinate, and regulate this essential biology at the nanoscale.
Plasma membrane preparation and fixation. After EGF pulse-chase stimulations, cells were rinsed briefly with stabilization buffer (30 mM HEPES, 70 mM KCl, 5 mM MgCl 2 , 3 mM EGTA, pH 7.4). Ventral cell membranes were obtained with the application of unroofing buffer containing 2% paraformaldehyde in stabilization buffer using a 10 mL syringe with a 22 gauge, 1.5 needle. The syringe was held vertically within 1 cm of the coverslip during the mechanical unroofing. To prepare dorsal plasma membranes, we inverted cell-containing coverslips onto PDL-coated coverslips flanked with two 3 mm × 3000 mm strips of parafilm and covered with 400 μL of stabilization buffer. A pencil eraser was applied to the top center of the coverslips with light pressure for 10 s before the coverslips were lifted away. Then, dorsal membranes were found around the pencil eraser application area of the bottom coverslip. After either of the mechanical manipulations, coverslips were washed once and moved to a fresh stabilization buffer containing 2% paraformaldehyde for 20 min. They were rinsed 4× with PBS followed by electron or fluorescent microscopy preparation.
Platinum replica electron microscopy (PREM). Coverslips were transferred from glutaraldehyde into 0.1% w/v tannic acid for 20 min. Then, coverslips were rinsed 4× with water and placed in 0.1% w/v uranyl acetate for 20 min. The coverslips were dehydrated, critical point dried, and coated with platinum and carbon as previously described 35 . The replicas were separated from glass coverslips using hydrofluoric acid and mounted on glow-discharged Formvar/carbon-coated 75mesh copper TEM grids (Ted Pella 01802-F). Transmission Electron Microscopy imaging was performed as previously described 69 at 15,000× magnification (1.2 nm per pixel) using a JEOL1400 (JEOL) and SerialEM 3.8 software for montaging. Montages were stitched together using IMOD 4.0.26 69 . Images are presented in inverted contrast. Each montage was manually segmented in ImageJ 1.53n 70 by outlining the edge of the membrane, flat clathrin structures (no visible curvature), domed clathrin structures (curved but can still see the edge of the lattice), and sphere clathrin structures (curved beyond a hemisphere such that the edge of the lattice is no longer visible). The percentage of occupied membrane area was defined as the sum of areas from clathrin of the specified subtype divided by the total area of the visible membrane.
Immunofluorescence. Unroofed cells were incubated in PBS containing 3% w/v bovine serum albumin (Fisher Bioreagents, BP9703) and 0.1% v/v Triton X-100 Total internal reflection microscopy (TIRFM). Cells were imaged on an inverted fluorescent microscope (IX-81, Olympus), equipped with a 100x, 1.45 NA objective (Olympus). Combined green (488 nm) and red (561 nm) lasers (Melles Griot) were controlled with an acousto-optic tunable filter (Andor) and passed through an LF405/488/561/635 dichroic mirror. Emitted light was filtered using a 565 DCXR dichroic mirror on the image splitter (Photometrics), passed through 525Q/50 and 605Q/55 filters, and projected onto the chip of an electron-multiplying chargecoupled device (EMCCD) camera. Images were acquired using the Andor IQ2 software. Cells were excited with alternate green and red excitation light, and images in each channel were acquired at 500-ms exposure at 5 Hz. We used Matlab software previously reported 46 to automatically identify clathrin spots in one channel and extract small, square regions centered at the brightest pixel of each object. Matched regions from the same cellular location in the corresponding image were extracted. An equal number of randomly positioned regions were also extracted to test for nonspecific colocalization. The mean correlation between thousands of clathrin spots and their corresponding image pairs across several cells from independent experiments was calculated using Pearson's correlation coefficient. The corrected total cell fluorescence (CTCF) at the PM ( Fig. 7 and Sup. Fig. 15) was assessed with ImageJ software by measuring the integrated density of manually segmented membranes using the following formula: CTCF = Integrated Density -(Area of the segmented cell * Mean fluorescence of three background readings)/area of the segmented cell.
Correlative light and electron microscopy (CLEM). For TIRF/CLEM, we followed the protocol previously reported 48 . In brief, HSC3 WT cells were plated on collagen-coated coverslips, transfected with β5-integrin-GFP, and processed 1 day later. Cells were starved, stimulated, and unroofed as described above. Membranes were unroofed and fixed for 20 min in 2% PFA. After fixation, membranes were stained for Clathrin Heavy Chain and phalloidin-Alexa Fluor 647 as mentioned in the immunofluorescence section. TIRF microscopy was used to image a 15 × 15 image montage of a large field of unroofed cells. After fluorescence imaging, the coverslips were marked with a diamond objective marker and fixed in 2% glutaraldehyde in PBS overnight at 4°C. PREM of the marked region was performed as reported previously 35 . The replica was lifted and placed onto copper TEM grids and imaged with ×20 phase-contrast objective to find the same regions that were originally imaged in fluorescence. Each cell of interest was located on the grid prior to TEM imaging. The fluorescence images were fitted to the EM images using an affine spatial transformation with nearest-neighbor interpolation to map the Gaussian centers of clathrin structures visible in both TIRF and EM images. The correlation was performed using MATLAB code already reported 35,36 . TIRF/EM overlay were created to observe the colocalization of clathrin-coated structures in EM with fluorescence image of β5-integrin-GFP in TIRF.
Interference reflection microscopy (IRM)/GFP TIRF. IRM/TIRF was implemented on a Nikon inverted fluorescence microscope (Ti2 Eclipse, Nikon) with an In vitro phosphorylation assay using purified Src or PAK4 and peptides corresponding to the β5integrin carboxyl domain (742-799) WT and mutants in (a). Significance was tested by a one-way ANOVA test, *P = 1.51 × 10 −4 , **P = 0.002, ***P = 1.09 × 10 −5 , ns P = 0.0205. N = 3 biologically independent experiments with consistent results. c Representative TIRF images of HSC3 WT cells cotransfected with mScarlet-CLCa and β5-integrin-GFP WT or containing the different mutations shown in (a), either before (Ctrl) or after 50 ng/mL EGF stimulation for 15 min. Scale bars are 10 μm; insets are 7.3 μm × 7.3 μm. d Automated two-color correlation analysis of (a). Dot box plots show median extended from 25th to 75th percentiles, mean (square), and minimum and maximum data point whiskers with a coefficient value of 1.5. Significance was tested by a two-tailed t-test, ns P β5-WT = 0.0811, ***P β5-ΔC = 4.03 × 10 −7 , ***P β5-3Y-F = 8.9 × 10 −5 , ns P β5-3Y-E = 0.9149, ns P β5-2S Epi-illumination for IRM was performed with a mercury lamp illuminator (Intensilight C-HGFI, Nikon) filtered with a 515 long-pass filter. A dichroic below the objective (Omega 490-600DBDR) was used to reflect a 488 nm TIRF illumination laser and the IRM epi-illumination sequentially into the objective. Images were acquired using the Nikon (NIS) Elements version 5.2 software. Unroofed HSC3 transfected with GFP-CLCa were illuminated for IRM and GFP TIRF separately and fluorescence or reflected light was collected through two different emission filters (525/50 nm for GFP, 600/50 nm for IRM) onto an Andor iXon Ultra 897 EMCCD. For IRM, the aperture diaphragm of the epi-illumination was adjusted to optimize contrast. Large montages were generated in IRM and the GFP channel for each sample. IRM image processing was performed as follows. Each image in the montage was averaged to form a field background image which was subtracted from each image. The stack of images was then baseline shifted by subtracting the minimum pixel value of the entire stack. Finally, the stack was normalized to the mean pixel value such that its max intensity was three times the mean. Single images of cells from this montage were contrast inverted and used with their corresponding GFP images for further analysis. In each whole-cell GFP image (GFP-CLCa), all the local maxima above a user-defined threshold and >2 μm away from the edge of the cell boundary were identified. Each local maximum was centered in a 4 μm × 4 μm cropped region. The mean correlation between thousands of clathrin spots and their corresponding IRM image pairs across several cells from independent experiments was calculated as described above.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
The protein sequences were consulted in the Uniprot database  Fig. 8 Model of flat clathrin lattices expansion during growth factor response. a Small flat clathrin lattices are in proximity to Src and are enriched with β5-integrin. b EGF triggers the dimerization, clustering, and cross-phosphorylation of EGFR at growing FCLs. This in parallel allows the biding of the downstream scaffold Grb2 and locally activates Src, which in turn phosphorylates β5-integrin cytoplasmic domain. The maintenance of the EGFR/Src/β5integrin axis promotes flat clathrin lattice growth. A key implication of this model is that two different receptor systems are spatially organized at the nanoscale within flat clathrin lattices. EGF epidermal growth factor, EGFR epidermal growth factor receptor, FCLs flat clathrin lattices.
Correspondence and requests for materials should be addressed to Justin W. Taraska.
Peer review information Nature Communications thanks Stephen Royle, Stéphane Vassilopoulos and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.
Reprints and permission information is available at http://www.nature.com/reprints Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. This is a U.S. Government work and not under copyright protection in the US; foreign copyright protection may apply 2022