Giant Vesicle Platform for Synthetic Endoplasmic Reticulum Networks

Abstract

The endoplasmic reticulum (ER) is a membrane-bound organelle found in all eukaryotic cells. This organelle has a fascinating morphology, which involves a network of membrane nanotubes and three-way junctions connecting three nanotubes. Previous reconstitution studies were based on proteo-liposomes with a size of 100 to 200 nanometers, using electron microscopy as the main method to image the resulting membrane morphologies. These studies revealed the importance of membrane GTPases, that is, membrane proteins that hydrolyse GTP. Here, we introduce an alternative platform for the formation of synthetic ER networks starting from giant unilamellar vesicles with a typical size of 10 mirometers and containing a single membrane GTPase such as atlastin from fruit flies or Sey1 from yeast. This giant vesicle approach makes the networks accessible to optical microscopy, by which one can elucidate both their structure and their dynamics. We first demonstrate that the membranes' propensity to form synthetic ER networks depends strongly on the lipid composition. More precisely, we show that this propensity varies non-monotonically with the mole fraction of cholesterol and that the frequency of three-way junctions has a maximum for an optimal mole fraction between 10 and 15 mol%. This optimal value is comparable to the mole fraction of cholesterol as estimated for ER membranes in vivo. We also demonstrate the formation of synthetic ER networks when the membranes contain only the membrane GTPase Sey1 but no other membrane protein, which implies that Sey1 is able on its own to generate a high membrane curvature. Based on the reconstitution approach introduced here, the synthetic ER networks can now be further investigated using the versatile toolbox that has been developed for giant vesicles.