Nanodissected elastically loaded clathrin lattices relax to increased curvature

Abstract

Clathrin-mediated endocytosis (CME) is the major endocytosis pathway for the specific internalization of large compounds, growth factors, and receptors. Formation of internalized vesicles from the flat plasma membrane is accompanied by maturation of cytoplasmic clathrin coats. How clathrin coats mature and the mechanistic role of clathrin coats are still largely unknown. Maturation models proposed clathrin coats to mature at constant radius or constant area, driven by molecular actions or elastic energy. Here, combining high-speed atomic force microscopy (HS-AFM) imaging, HS-AFM nanodissection, and elasticity theory, we show that clathrin lattices deviating from the intrinsic curvature of clathrin form elastically loaded assemblies. Upon nanodissection of the clathrin network, the stored elastic energy in these lattices drives lattice relaxation to accommodate an ideal area-curvature ratio toward the formation of closed clathrin-coated vesicles. Our work supports that the release of elastic energy stored in curvature-frustrated clathrin lattices could play a major role in CME.

Fig. 1. HS-AFM 3D morphology analysis of CCPs on plasma membranes of freshly unroofed PTK2 cells.

(A) CCPs of various sizes and morphologies observed on plasma membranes of freshly unroofed PTK2 cells (arrows indicate examples of pentagonal insertions). (B) 3D representation of an individual CCP. (C) Spherical sector (yellow spherical cap) fit of an individual CCP. The fit determines the CCP radius of curvature and surface area (above 0-height level defined by the background height of the surrounding membrane). (D) Partial skeleton of clathrin lattice (green lines). The lattice skeleton describes the conformation of each clathrin triskelion in 3D space. (E) Distribution analysis of n = 128 CCPs visualized by HS-AFM according to their surface area versus radius of curvature. The observed distribution of the CCP area and radius of curvature does not follow the predictions of the constant area (horizontal red line) and constant curvature (vertical red line) models obtained from the mean CCP surface area and radius of curvature, respectively (see fig. S1).

Fig. 2. Clathrin inter-arm angles and triskelia conformations in CCPs in 3D space and their correlation with the CCP global curvature.

Distributions of inter-arm angles (gray bins) of (A) pentagons, (B) hexagons, and (C) heptagons. The fitted normal distributions of pentagons [red line in (A)], hexagons [green line in (B)], and heptagons [blue line in (C)] peak at mean values approximately equal to the internal angles of the corresponding regular polygons. Fitting statistics are shown in the graphs. (D) Distribution of the sum of the three inter-arm angles in triskelia (α + β + γ). Clathrin triskelia conformations were mainly convex (α + β + γ < 360°, red trace) or flat (α + β + γ = 360°), with a residual number of concave (α + β + γ > 360°, blue trace) conformations. The observed convex triskelia do not display the inter-arm angle sums associated with assemblies of regular polygons [insets: (9)]. (E) Correlation of CCP radius of curvature and inter-arm angle sum in triskelia. CCPs with smaller radius of curvature tend to have more convex clathrin triskelia.

Fig. 3. Predicted CCP energy landscape and comparison to experimental data.

(A to E) Heatmaps showing the CCP energy landscape as a function of CCP area and radius of curvature: (A) membrane bending energy in Eq. 2, (B) membrane tension energy in Eq. 3, (C) clathrin coat bending energy in Eq. 4, and (D) clathrin polymerization energy in Eq. 6; (E) combined CCP energy landscape in Eq. 7. The bold curves and black circles in (A) to (F) indicate the surface area–radius of curvature ratio of spheres, following A = 4πR²; CCPs can only populate the energy landscapes to the right sides of these curves (plot regions on the left sides of A = 4πR² are shaded dark gray). The thin white curves in (A) to (E) denote equal-energy contours. For (A), equal-energy contours correspond to fixed ratios h/R, as indicated schematically. Arrows in (A) to (D) indicate the directions in which the respective contributions to the CCP energy landscape drive CCP maturation. The red bold circle in (E) indicates the lowest-energy CCP state with A ≤ 4πR², which lies on the A = 4πR² curve. (F) Probability density plot of CCP observations: The highest probability region approximately corresponds to the low-energy CCP states predicted by the combined CCP energy landscape. (G) Clathrin bending energy per clathrin arm-arm interaction as a function of the CCP radius of curvature (red line). Clathrin cages, probability distribution of clathrin cages by Morris et al. (9); CCPs, probability distribution of CCP radius of curvature reported here.

Fig. 4. HS-AFM nanodissection: Low-curvature clathrin lattices are loaded with elastic energy.

(A) Details of the clathrin lattice nanodissection experiments. A clathrin lattice on a plasma membrane before (left), after the first (middle), and after the second (right) nanodissection, with the corresponding targets for nanodissection (dashed squares) and single line scans over individual clathrin arms at gradually increased applied force to dissect individual clathrin-clathrin arm-arm interactions (insets). (B) Section analysis along the white dashed lines “1” and “2” in (A): The lattice is initially rather flat along line 1. Following the first and second nanodissections, the section analyses show the lattice separating into two separate CCPs with increased curvature (top profiles 1). In contrast, the nanodissected region is pulled down and flattened (bottom profiles 2). (C) Clathrin nanodissections lead to release of clathrin lattice constraints, allowing low-curvature CCPs to converge to energetically more favorable high-curvature CCP states (CCP energy landscape reproduced from Fig. 3E). As in Fig. 3E, the red bold circle indicates the lowest-energy CCP state with A ≤ 4πR², which lies on the A = 4πR² curve.