Adhesion energy controls lipid binding-mediated endocytosis

Abstract

Several bacterial toxins and viruses can deform membranes through multivalent binding to lipids for clathrin-independent endocytosis. However, it remains unclear, how membrane deformation and endocytic internalization are mechanistically linked. Here we show that many lipid-binding virions induce membrane deformation and clathrin-independent endocytosis, suggesting a common mechanism based on multivalent lipid binding by globular particles. We create a synthetic cellular system consisting of a lipid-anchored receptor in the form of GPI-anchored anti-GFP nanobodies and a multivalent globular binder exposing 180 regularly-spaced GFP molecules on its surface. We show that these globular, 40 nm diameter, particles bind to cells expressing the receptor, deform the plasma membrane upon adhesion and become endocytosed in a clathrin-independent manner. We explore the role of the membrane adhesion energy in endocytosis by using receptors with affinities varying over 7 orders of magnitude. Using this system, we find that once a threshold in adhesion energy is overcome to allow for membrane deformation, endocytosis occurs reliably. Multivalent, binding-induced membrane deformation by globular binders is thus sufficient for internalization to occur and we suggest it is the common, purely biophysical mechanism for lipid-binding mediated endocytosis of toxins and pathogens.

Fig. 1. Membrane deformation after polyvalent lipid binding is a common mechanism of viral endocytosis.

A Spinning disc confocal fluorescence microscopy micrographs of polyomavirus-like particles (VLPs) bound to Giant unilamellar vesicles (GUVs) containing receptor gangliosides. 2 µg of each VLP was incubated for 1 h at RT with GUVs containing the indicated gangliosides (98% DOPC, 1% ganglioside, 1% β-BODIPY FL C12-HPC dye) and imaged at the equatorial plane. Scale bar is 2 µm. B Spinning disc confocal fluorescence microscopy micrographs of polyoma VLPs bound to energy-depleted CV1 cells. Cells were starved of cellular energy by 30 min incubation in starvation buffer (PBS +/+ supplemented with 10 mM 2-deoxy-D-glucose and 10 mM NaN3) followed by 1 h incubation with 5 µg of each VLP in starvation buffer and imaged live on a spinning disk confocal microscope. DiI membrane dye was added 10 min prior to imaging at 1 mg/ml final concentration. Scale bars are 5 µm and 1 µm for insets. Arrows mark VLP-filled membrane invaginations. C Quantification of colocalization in confocal fluorescence micrographs between polyomavirus VLPs and lysosomes as marked by Lamp1-GFP in live cells. CV1 cells expressing Lamp1-EGFP were kept at 4 °C for 10 min before incubation with 2 µg of the indicated VLPs for 30 min at 4 °C, followed by further incubation at 37 °C for the indicated times before imaging live on a spinning disk confocal microscope. Means ± s.e.m. from n = 3 independent experiments. D Representative confocal fluorescence micrographs of the Lamp1-EGFP expressing cells containing the indicated VLPs after 6 h incubation at 37 °C. Scale bars are 5 µm and 1 µm for insets. Magenta: VLPs, Cyan: membrane marker or Lamp1-EGFP. Source data are provided as a Source Data file.

Fig. 2. A polyvalent virus-like-particle lipidic-receptor system for endocytosis.

A Schematic representation of the synthetic system. Shown is a genetically encoded nanoparticle (GEM) assembled from 180 copies of the encapsulin protein (dark green) coupled to GFP (light green) scaffold. A GPI-anchored anti-GFP nanobody (purple) inserted into the membrane (beige) serves as receptor. B Fluorescence micrograph of GEMs binding to the cell membrane of CV1 cells. Scale bar is 10 µm. Insets: (upper) magnified region of the GEM-GFP decorated membrane from the overview emphasizing monodisperse binding. A single particle is marked with a box. Scale bar is 2 µm. (lower) Transmission electron micrograph of purified GEM. Scale bar is 15 nm. Experiments have been repeated twice with similar results. C Fluorescence micrograph of GEMs bound to Giant Plasma Membrane Vesicles (GPMVs) of CV-1 cells expressing GPI-anchored nanobody. Cells were incubated with 0.45 nM GEMs for 1 h at RT before imaging at the equatorial plane on a spinning disk confocal microscope. Experiments have been repeated twice with similar results. Scale bar is 2 µm. D Fluorescence micrograph of GEMs bound to energy-depleted CV1 cells expressing GPI-anchored anti-GFP nanobody that were starved of cellular energy by 30 min incubation in starvation buffer (PBS+/+ supplemented with 10 mM 2-deoxy-D-glucose and 10 mM NaN3) followed by 1 h incubation with 2 µg of purified GEMs in starvation buffer and imaged live on a spinning disk confocal microscope. DiI membrane dye was added 10 min prior to imaging at 1 mg/ml final concentration. Experiments have been repeated twice with similar results. Scale bars are 5 µm and 1 µm for the inset. E Fluorescence micrograph of GEMs bound to CV1-cells expressing Clathrin-light-chain-dsRED incubated for 10 min with 2 µg of GEMs before live imaging on a TIRF microscope. Experiments have been repeated three times with similar results. Scale bars are 5 µm and 1 µm for inset. F Fluorescence micrograph of GEMs bound to CV1-cells expressing Caveolin-1-mRFP incubated for 10 min with 2 µg of GEMs before live imaging on a TIRF microscope. Experiments have been repeated three times with similar results. Scale bars are 5 µm and 1 µm for inset. G Correlative confocal fluorescence platinum-replica electron microscopy micrographs of plasma membrane sheets generated after unroofing cells incubated with GEMs. Shown are 4 representative intracellular plasma membrane structures colocalizing with GEMs bound to the outside of cells that are neither positive for clathrin (as shown by antibody-staining) nor caveolae (based on distinct caveolae protein coat). Scale bars are 50 nm. Electron microscopy micrographs are on top, same field of view with correlative GFP fluorescence of the GEMs at the bottom. H Top: Example platinum replica electron microscopy micrographs of a typical clathrin-coated pit, caveola and clathrin/caveolin double-negative invagination. Bottom: Quantification of colocalization of GEM fluorescence with endocytic structures. Means ± S.D. for 6 cells from n = 2 independent experiments. Scale bar is 50 nm. Overview images are provided in Supplementary Fig. S2 and Supplementary Movie 2. Source data are provided as a Source Data file.

Fig. 3. GEMs are endocytosed and are trafficked through the endolysosomal pathway.

A, B Correlative fluorescence light microscopy and transmission electron microscopy of GEMs internalized in CV-1 cells. A Timepoint 1 h after binding. B Timepoint 6 h after binding. Each panel top from left to right: Fluorescence micrograph of GEMs; transmission electron micrograph of same region; correlative images. Each panel bottom left: Transmission electron micrograph of inset above. Each panel bottom right: Volumetric 3D-reconstruction of electron micrographs. GEMs emphasized in green, membrane emphasized in purple. Experiments have been repeated twice with similar results. Scale bars are 500 nm for overview and 100 nm for insets. C Fluorescence micrographs from a time-course experiments of endocytosis showing the distribution of GEMs in CV1 cells expressing anti-GFP nanobody and Lamp1-mRFP. Cells were incubated with 2 µg of GEMs for the indicated time points at 37 °C before live imaging on a spinning disk confocal microscope. Scale bars are 10 µm. D Quantification of colocalization between GEMs and Lamp1-mRFP and between GEMs and Rab7-mRFP from timepoints indicated in C Means ± s.e.m., n = 3 independent experiments. E Quantification of GEM endocytosis upon treatment with genetic (siRNA against clathrin-heavy-chain and expression of dominant negative Dyn2-K44A) or chemical inhibitors (Nystatin/Progesterone, BafilomycinA and Cytochalasin D) as compared to mock treatment of controls (Transferrin endocytosis for siRNA against CHC and overexpression of DynK44A; SV40 endocytosis for Nystatin/Progesterone, BafilomycinA and CytochalasinD). Mean fluorescence intensity ± S.D. was determined from flow cytometry measurements of 6811–27,733 cells from n = 2 independent experiments. Source data are provided as a Source Data file.

Fig. 4. Adhesion energy controls membrane deformation and endocytosis of GEMs.

A Schematic representation of GPI-anchored nanobody constructs with decreasing binding affinity expressed in the outer membrane of cells used in this study. B Fluorescence micrographs of GEMs bound to GPMVs harvested from CV1 cells expressing the panel of GPI-anchored anti-GFP nanobody constructs as indicated and subsequently incubated with 0.45 nM GEM-GFP particles for 1 h at RT before imaging at the equatorial plane on a spinning disk confocal microscope. Experiments have been repeated twice with similar results. Scale bars are 2 µm. C Fluorescence micrographs of GEMs bound to energy-depleted CV-1 cells expressing the panel of GPI-anchored anti-GFP nanobodies. CV1 cells were starved of cellular energy by 30 min incubation at 37 °C in starvation buffer (PBS+/+ supplemented with 10 2-deoxy-D-glucose and 10 mM NaN3) followed by 1 h incubation with 2 nM of GEM-GFP particles in starvation buffer at 37 °C and imaged live on a spinning disk confocal microscope. DiI membrane dye was added 10 min prior to imaging at 1 mg/ml final concentration. Experiments have been repeated twice with similar results. Scale bars are 5 µm and 1 µm for insets. Arrows mark VLP-filled membrane invaginations. D Quantification of GEM-GFP endocytosis as a function of receptor affinity as determined by flow cytometry measurements of the mean cell-associated fluorescence after acid wash. Mean fluorescence intensity ± s.e.m. was determined from flow cytometry measurements of 846–9538 cells/sample from n = 3 independent experiments. E Quantification of GEM-GFP endocytosis as a function of receptor affinity and upon treatment with genetic (siRNA against clathrin-heavy-chain and expression of dominant negative Dyn2-K44A) or chemical inhibitors (Nystatin/Progesterone and BafilomycinA) as compared to mock treatment of controls (Transferrin endocytosis for siRNA against CHC and overexpression of DynK44A; SV40 endocytosis for Nystatin/Progesterone and BafilomycinA). Endocytosis was determined by flow cytometry measurements of the mean cell-associated fluorescence after acid wash. Mean fluorescence intensity ± S.D. was determined from flow cytometry measurements of 1026–29,803 cells from n = 2 independent experiments. Source data are provided as a Source Data file.

Fig. 5. Theoretical modeling of GEM-GFP wrapping and tubulation.

A Schematic representation of the geometric parameters considered in the modeling for the length and tilt angle variations (cartoon) and the corresponding estimative numerical values (table): b = estimated distance from membrane midplane to nanobody C-terminus to which the GPI-anchor is attached; a = estimated vertical extensions of unstructured 12-residue peptide linker connecting the GFP N-terminus to the GEM surface; Ɵ = estimated tilt angle of the complex, i.e of the axis (with length 4.8 nm) connecting the linker attachment sites at the nanobody C-terminus and GFP N-terminus, relative to the membrane normal; l = projected vertical extensions 4.8 nm Cos[Ɵ] of the complex corresponding to tilt angle estimates. These length estimates and variations lead to the mean distance ${\mathrm{l}}_0$ = 2.5 + 1.5 + 4 nm = 8 nm and standard deviation $\mathrm{\sigma }$ = −1 nm. B Table of the percentage of energy-depleted cells and GPMVs containing GEM-filled tubular invaginations from the total amount of cells/GPMVs for the corresponding binding affinities. C Energy gain per central GEM particle in a tubule, compared to individual wrapping of the particle, as a function of the binding affinity. Above the wrapping threshold around Kd = 1000 nM, particle wrapping in tubules is energetically clearly favorable. D Schematic representation of the two membrane wrapping cases: i) GEM wrapping in live cells leading to cellular endocytosis and ii) GEM wrapping resulting in the formation of long tubules in energy-depleted cells, illustrated here for an intermediate conformation with three particles close to the wrapping threshold where the particles are only partially wrapped. Modeled minimum-energy conformations for GEM wrapping are shown in Supplementary Fig. S9. Source data are provided as a Source Data file.