Non-enveloped viruses make use of multivalent lipid binding to mediate their uptake into cells. It is established that the tumor-inducing Simian Virus 40, a member of the polyomavirus family, can induce dramatic membrane curvature in the plasma membrane of cells by binding multivalently to 360 copies of its glycolipid receptor GM1. This ultimately leads to its internalization through clathrin-independent endocytosis and productive infection of the cells. It remains unclear whether this is a generic biophysical mechanism employed by non-enveloped lipid binding viruses to mediate their uptake in a clathrin-independent manner. Here, I found that several members of the polyomavirus family deform membranes in vitro and in cells, strengthening the hypothesis that multivalent lipid binding might be a common biophysical mechanism for membrane deformation and internalization. To further test this hypothesis, I designed a synthetic cellular system for the investigation of endocytosis mediated by globular particles multivalently binding to lipidic receptors in the plasma membrane. This system is composed of a recombinantly expressed globular virus-like-particle and corresponding lipid-anchored receptors. I made use of the encapsulin protein from the archaeon Pyroccoccus furiosus to which GFP was genetically linked and then self-assembles from 180 subunits into a 37 nm diameter capsid bearing regularly-spaced GFP molecules on the surface, here onward called GEM. As receptors for this globular virus-like-particle, I attached a GPI-anchor to 7 different anti-GFP nanobodies with individual binding affinities increasing from the µM to the pM range. This enabled me to range the adhesion energy of synthetic pentavalent lipid binding particles to membranes over 7 orders of magnitude and study the biophysics required for efficient membrane deformation and subsequent internalization to occur. For this, I reconstituted this receptor-ligand system in vitro and in cells and found that particles deform membranes and become internalized in a clathrin-independent manner, provided that the adhesion energy is high enough. Based on experimental work on cells and liposomal membranes and theoretical considerations, a physical model was derived to explain membrane wrapping by GEMs into long, tubular invaginations as a function of binding affinity. My work shows that polyvalent lipid binding alone is sufficient for membrane deformation and clathrin-independent internalization and provides mechanistic insight into the biophysical basis of multivalent lipid-binding mediated membrane deformation by non-enveloped, tumor-causing virions. Secondly, non-enveloped viruses are known to traffic through the lysosomes towards their final destination in the cell, the endoplasmic reticulum, as they require the acidic environment of the endosomal compartments for disassembly and release of their genetic material. On the other hand, bacterial toxins such as the β-subunit of the Cholera toxin traffic to the Golgi apparatus instead, even though they share a similar pentavalent organization of binding sites with the core structural protein of polyomaviruses and bind to the same lipidic moiety in the plasma membrane of cells. The mechanisms conferring the intracellular trafficking specificity to these two pathogens is not yet understood. Here I investigate whether the nanoscale configurations and multivalency of lipid binding sites on these pathogens dictates their intracellular routes. First, I found that several lipid-binding viruses are transported through the endo-lysosomal pathway after endocytosis while completely bypassing the Golgi apparatus, suggesting the existence of a global mechanism conferring specificity to their intracellular sorting. To further understand this, I used the synthetic cellular system described above to test the trafficking of a globular viral mimic with similar architecture of binding sites as the polyomaviruses. I found that the synthetic particles traffic through the endo-lysosomal system in a similar fashion to the virions. Secondly, I studied the minimum requirements conferring the Golgi transport specificity to bacterial toxins. I investigated the intracellular transport of toxin mimics with similar number and flat configuration of lipid binding sites. I found that none of the toxin mimics used in this study could successfully reproduce the Golgi trafficking of bacterial toxins, suggesting that multivalent lipid binding arranged on a flat configuration is not sufficient to provide such intracellular trafficking specificity. My work opens questions as to what other factors are at play in the binding and entry of such pathogens that provide the necessary cues for intracellular transport.
