- Protein-induced membrane deformations -
 |
|
| |
|
| Membrane traffic requires
membrane deformation to generate vesicles and tubules. Strong evidence
suggests that assembly of curvature-active proteins can drive such
membrane shape changes. Well-documented pathways often involve
protein scaffolds, in particular coats (clathrin or COP). However,
specialized protein coats are not always be required for cargo
internalization. Indeed, membrane curvature can in principle
be influenced by any protein binding asymmetrically on a membrane,
which aggregation can trigger large membrane morphological changes (see
our
model for caveolae
formation). We have recently reported that the binding of Shiga toxin, a bacterial toxin protein, to its cellular glycolipid receptor (Gb3) induces membrane deformation without needing contributions from other cell machinery. These results strongly suggest that cargo alone can sometimes trigger its own internalisation by influencing membrane curvature. In the case of Shiga toxin or viral matrix proteins, tubules and buds appear to result from the cargo-driven formation of protein-lipid nanodomains, showing that collective protein behaviour is crucial in the process. Membrane tubulation induced by the aggregation of membrane protein can be reconstituted in vitro using Giant Unilamellar Vesicles (GUV, see below)). Using similar experimental techniques, the same behaviour has been observed for virus-like-particle (VLP), SV40 viral capsids devoid of genetic material, which specifically bind to another type of glycolipid membrane receptor (GM1). Tubules are observed upon adsoprtion for the spherical VLPs, but also upon adsorption of individual protein pentamers that form the VLPs. This means that although the spherical geometry of the viral capsid helps tubule formation, it is not required for membrane tubulation, as the interaction of individual capsid constituent with membrane receptors is sufficient to induce tubulation. Interestingly, several physical and biochemical properties (low membrane tension and saturation of the receptor hydrophobic tails) are required for the tubulation of requirement for tubulation of individual pentamer units, but not of etire spherical capsids. We believe that these differences stem from the Thermodynamics of protein aggregation and tubule formation. indeed, while capsids posses an intrisic curvature, which is directly imparted to the membrane upon protein-receptor binding, the individual capsids don't, and it is the level of saturation of the receptor tail that impart a spontaneous curvature to the membrane. Furthermore, while high membrane tension should always prevent membrane tubulation, the tension is to be compared to the gain of line energy upon aggregation of individual proteins, whie it is to be compared to the much higher receptor adhesion energy of in the case of capsid, so high membrane tension is in practice unable to prevent capsid-induced membrane tubulation.
The tubules observed in GUV seem to bear similar properties than the membrane tubules that can be seen in cells deprived of energy. If energy is present, long, toxin-covered tubules are almost never observed in-vivo, suggesting that energy is used to severe the tubule (e.g by the activitiy of Dynamin), rather than to actually trigger membrane deformation. A parallel can be made with the model we suggested for the regulation of coated vesicle secretion, in which protein aggregation and membrane deformation is a spontaneous process, while energy is used to control the process by preventing membrane deformation (in the case of COP protein, this is done the GTP-dependent recycling of coat proteins in the cytosol). A combination of in vitro experiments on Giant unilamellar vesicles and theoretical modelling based on statistical physics is ideally suited to tackle these collective effects. Physical parameters such as membrane tension or lateral membrane heterogeneity are expected to influence protein assemblies on cell membranes. These parameters, which are probably regulated in cells (see our model for tension regulation of cell membranes), can be controlled in biomimetic membrane systems. A complete picture of large-scale protein-induced membrane deformation needs to address at least three levels of physical description. At the molecular scale (1-10 nm), the properties of the lipids and the proteins and their interaction are critical to understand how proteins can influence membrane composition and shape [28-31]. At the scale of a protein aggregate (0.1-1 µm), molecular interactions can be translated into physical couplings in terms of stresses or deformations. The optimal membrane shape can then be derived from the mechanical energy of deformation. At the scale of many aggregates (more than 1 µm – the whole cell), thermodynamics can be used to compute the kinetics of protein aggregation and the evolution of membrane morphology. In the case of shiga toxin, we have put forward a model that involves the asymetric compaction of lipids beneath an adsorbed toxin, which could result from the high affinity of the toxin to its membrane glycolipid receptor Gb3. Other molecular mechanisms can be invoked to explain the affinity of certain membrane proteins to region of high membrane curvature, but protein organisation at large lengthscales can not critically depend on molecular details, and obey general physical principles, which are reviewed in our COCB paper . |
