Morphological changes of plasma membrane and protein assembly during clathrin-mediated endocytosis

Abstract

Clathrin-mediated endocytosis (CME) proceeds through a series of morphological changes of the plasma membrane induced by a number of protein components. Although the spatiotemporal assembly of these proteins has been elucidated by fluorescence-based techniques, the protein-induced morphological changes of the plasma membrane have not been fully clarified in living cells. Here, we visualize membrane morphology together with protein localizations during CME by utilizing high-speed atomic force microscopy (HS-AFM) combined with a confocal laser scanning unit. The plasma membrane starts to invaginate approximately 30 s after clathrin starts to assemble, and the aperture diameter increases as clathrin accumulates. Actin rapidly accumulates around the pit and induces a small membrane swelling, which, within 30 s, rapidly covers the pit irreversibly. Inhibition of actin turnover abolishes the swelling and induces a reversible open-close motion of the pit, indicating that actin dynamics are necessary for efficient and irreversible pit closure at the end of CME.

Fig 1. Aligning the confocal image and the AFM image.

(A) Schematic illustration of the sample stage. A cross-shaped movable XY-stage (orange) is mounted on the base plate (light green) of the inverted optical microscope (IX83) via a stage guide (gray) equipped at each of the 4 ends of the cross. A 3-point support plate (purple) for mounting the AFM scanner unit is fixed on the base plate with a configuration that does not hinder the sliding of the XY-stage along the x-axis and y-axis. These setups allow the sample stage to move independently of the AFM unit and the optical axis. (B) Side view of the HS-AFM unit mounted on the stage illustrated in panel A. (C) Overlaying a confocal image and an AFM image. COS-7 cells expressing EGFP-CLCa were fixed with 5% paraformaldehyde and subjected to AFM (left) and CLSM (middle) imaging. The x-y position of the probe tip was determined as described in S1 Fig. Two images were overlaid (right) based on the x-y center position. Scale bar: 1 μm. Autofluorescence of the probe was much weaker than clathrin spot and could not be detected during the fast scanning. (D) AFM images of CCP on the cytoplasmic surface of the plasma membrane. COS-7 cells were “unroofed” by mild sonication as described in Materials and methods and then fixed with glutaraldehyde. Scale bar: 0.1 μm. AFM, atomic force microscopy; CCP, clathrin-coated pit; CLSM, confocal laser scanning microscopy; COS-7, CV-1 in origin with SV40 gene line 7; EGFP, enhanced green fluorescent protein; EGFP-CLCa, EGFP-fused clathrin light chain a; HS-AFM, high-speed AFM.

Fig 2. Hybrid AFM imaging reveals morphological changes of CCPs in living cells.

In culture medium, live COS-7 cells expressing EGFP-CLCa were subjected to time-lapse hybrid AFM imaging. The position of the probe tip was aligned as described in S1 Fig. (A) The fluorescence image (top) and AFM image (bottom) of a representative CCP. Image size: 0.6 × 0.6 μm². (B) The fluorescence intensity (green) and diameter of the CCPs in the AFM images (dotted black) in panel A are plotted against time. Based on the changes of the clathrin signal, the whole process is divided into a growing phase, a stable phase, and a closing phase based on the procedure described in S3 Fig. (C) Variations of CCPs observed by hybrid time-lapse imaging. Individual CCPs have large variations in the duration of the growing and stable phases. Underlying data may be found in S1 Data. AFM, atomic force microscopy; a.u., arbitrary unit; CCP, clathrin-coated pit; COS-7, CV-1 in origin with SV40 gene line 7; EGFP, enhanced green fluorescent protein; EGFP-CLCa, EGFP-fused clathrin light chain a; FL, fluorescence.

Fig 3. Morphological changes of the plasma membrane and protein assembly during CME.

(A, B) Time-lapse hybrid imaging of COS-7 cells expressing EGFP-CLCa and mCherry-epsin. Fluorescence images for epsin and clathrin, as well as AFM images are shown every 10 s (panel A). Image size: 1.2 × 1.2 μm². The signal intensities of the fluorescence spots for clathrin (green) and epsin (red), and the diameter of the membrane invagination in the AFM image (dotted black) were plotted against time (panel B). The FL intensity is plotted relative to its maximum value. (C, D) Time-lapse hybrid imaging of COS-7 cells expressing EGFP-CLCa and Dyn2-mCherry. Fluorescence images for dynamin and clathrin, as well as AFM images are shown every 10 s (panel C). The signal intensities of the fluorescence spots for clathrin (green) and dynamin (red), and the diameter of the membrane invagination in the AFM image (dotted black), were plotted against time (panel D). The FL intensity is plotted relative to its maximum value. (E) A summary of protein assembly at CCPs. Based on the AFM images, the time when the plasma membrane started to invaginate (left half) and when the pit had completely closed (right half) were defined as t = 0. Arrowheads indicate the time when fluorescence signal was maximum. The results of hybrid imaging for clathrin, epsin, dynamin, and actin are summarized along this time scale. The timings when the fluorescence signal appeared on the CCP (left half) and disappeared (right half) are plotted. n = 35 for clathrin, n = 8 for epsin, n = 13 for dynamin, and n = 14 for actin. Error bars represent SD. Underlying data may be found in S1 Data. AFM, atomic force microscopy; CCP, clathrin-coated pit; CME, clathrin-mediated endocytosis; COS-7, CV-1 in origin with SV40 gene line 7; EGFP, enhanced green fluorescent protein; EGFP-CLCa, EGFP-fused clathrin light chain a; FL, fluorescence; n.d., not determined.

Fig 4. Hybrid AFM imaging distinguishes CCPs from caveolae in live cells.

COS-7 cells expressing both EGFP-CLCa and mCherry-caveolin1 in culture medium were subjected to hybrid time-lapse AFM imaging. (A) The x-y alignment of the obtained AFM image (left), EGFP image (middle), and mCherry image (right) was performed as described in S1 Fig. Small membrane invaginations that colocalized with an EGFP signal or with an mCherry signal are indicated by white and black arrowheads, respectively. Scale bars: 1 μm. (B) The diameters of the membrane invaginations that colocalized with a clathrin signal (top panel) or with a caveolin signal (bottom panel) were measured by section analysis and are summarized as a histogram. (C, D) Lateral movements of CCPs and caveolae were analyzed. The trajectories of the pit center (green for clathrin, red for caveolae) were superimposed on the AFM image (panel C), and the MSD was plotted against time (panel D). Top panel: CCP; bottom panel: caveolae. A lateral drift of the specimen was estimated by measuring the average displacement of all pits between the initial and the last frame of the analysis. Because the average displacement of randomly diffusing spots is zero (|d| = 0), the average displacement of all endocytic pits observed in a series of AFM images corresponds to the drift of specimen during the observation. Underlying data may be found in S1 Data. AFM, atomic force microscopy; CCP, clathrin-coated pit; COS-7, CV-1 in origin with SV40 gene line 7; EGFP, enhanced green fluorescent protein; EGFP-CLCa, EGFP-fused clathrin light chain a; MSD, mean square displacement.

Fig 5. Variations in the closing motions of CCPs.

Membrane morphologies of CCPs when they close were analyzed from the same data set shown in Fig 3A, 3B and 3C. Three different motions identified in the CCPs: capping (panel A), two-step (panel B), and re-opening (panel C). AFM images and fluorescence images of EGFP-CLCa are shown in the panels at left. Fluorescence intensity (green) and the diameter of the CCP (dotted black) are also plotted in the time course (right panels). Image size: 0.6 × 0.6 μm². Membrane swelling in the capping motion, and the small aperture in the two-step motion, are indicated by black and white arrowheads, respectively. (D) The frequency of each closing motion. In total, 113 CCPs were analyzed. The frequency of 3 closing motions (capping, re-open, and two-step) were counted and represented as a percentage. The pits without any of these motions were not counted here. Three motions were not mutually exclusive, and sometimes multiple motions were seen in the identical pits. (E) Relationship between capping motion and re-opening motion. Among the CCPs that showed re-opening motion, the ratios of the CCPs with or without a capping motion are plotted. Error bars represent SD. (F) Distribution of the CCP lifetime for 3 different closing motions. Underlying data may be found in S1 Data. AFM, atomic force microscopy; a.u., arbitrary unit; CCP, clathrin-coated pit; EGFP, enhanced green fluorescent protein; EGFP-CLCa, EGFP-fused clathrin light chain a; FL, fluorescence.

Fig 6. Effects of actin-related inhibitors on CCP lifetime.

(A) COS-7 cells on the microscope stage were treated with cytochalasin B, jasplakinolide, or CK666 and were subjected to time-lapse hybrid imaging. The total lifetime and durations of large and small apertures in the two-step motion were measured and summarized as histograms. (B) Time-lapse AFM and fluorescence images obtained from COS-7 cells expressing EGFP-CLCa after treatment with cytoB (left panels). Image size: 0.6 × 0.6 μm². The fluorescence intensity (green) and the diameter of the CCPs in the AFM image (dotted black) are plotted against time (right panels). The process is divided into growing (red), stable (orange), and closing (blue) phases as described in Fig 2. Underlying data may be found in S1 Data. AFM, atomic force microscopy; a.u., arbitrary unit; CCP, clathrin-coated pit; COS-7, CV-1 in origin with SV40 gene line 7; cytoB, cytochalasin B; EGFP, enhanced green fluorescent protein; EGFP-CLCa, EGFP-fused clathrin light chain a; FL, fluorescence; jasp, jasplakinolide.

Fig 7. Role of actin in the closing step of CME.

(A) The frequencies of capping, re-opening, and two-step motions were counted as described in Fig 5D in the absence or presence of cytochalasin B, CK666, or jasplakinolide. (B) The frequency of CCP formation was quantified in the absence or presence of the inhibitors and summarized. Error bars represent SD. (C) Time-lapse AFM images obtained in a living COS-7 cell before and after treatment with cytoB, CK666, or jasp. Image size: 0.5 × 0.5 μm². Capping and the small aperture are indicated by black and white arrowheads, respectively. (D, E) A summary of actin assembly at a CCP from time-lapse hybrid imaging of COS-7 cells expressing EGFP-Lifeact and mCherry-CLCa. Fluorescence images for Lifeact and clathrin, as well as AFM images, are shown every 10 s (left panels). Image size: 1.2 × 1.2 μm². Two representative CCPs are shown here. The CCP closes with (panel D) and without (panel E) capping. The signal intensities of the fluorescence spots (for Lifeact [green] and clathrin [red]) relative to their maximum values and the diameter of the membrane invagination in the AFM image (black) are plotted against time. Underlying data may be found in S1 Data. AFM, atomic force microscopy; CCP, clathrin-coated pit; CME, clathrin-mediated endocytosis; COS-7, CV-1 in origin with SV40 gene line 7; cytoB, cytochalasin B; EGFP, enhanced green fluorescent protein; EGFP-CLCa, EGFP-fused clathrin light chain a; FL, fluorescence; jasp, jasplakinolide.

Fig 8. Effects of dynamin knockdown on CME.

(A) Knockdown efficiency of dynamin 2 in COS-7 cells was examined by western blotting. Total cell lysates of COS-7 cells transfected with siRNA targeting the dynamin 2 gene (si-DNM2) or control siRNA (Luciferase, si-ctrl) were prepared 24 and 48 h after the transfection and were subjected to western blotting using anti-dynamin 2 and β-actin antibodies. The density of the band corresponding to dynamin was quantified using β-actin as a loading control. The knockdown efficiency was 88% after 24 h and 92% after 48 h. (B) The frequency of CCP formation was counted and compared in si-ctrl and si-DNM2 cells. Error bars represent SD. (C) The frequency of capping, re-opening and two-step motions were analyzed in dynamin- and control-depleted cells as described in Fig 5D. (D) Time-lapse AFM images obtained in a living COS-7 cell transfected with si-ctrl or si-DNM2. Image size: 0.5 × 0.5 μm². Capping and small apertures are indicated by black and white arrowheads, respectively. (E) The total lifetime and the durations of large and small apertures, measured in control- and dynamin-depleted cells, are summarized. Underlying data may be found in S1 Data. AFM, atomic force microscopy; CCP, clathrin-coated pit; CME, clathrin-mediated endocytosis; COS-7, CV-1 in origin with SV40 gene line 7; siRNA, small interfering RNA.