Clathrin-mediated endocytosis is independent of actin dynamics in many circumstances but requires actin polymerization in others. We show that membrane tension determines the actin dependence of clathrin-coat assembly. As found previously, clathrin assembly supports formation of mature coated pits in the absence of actin polymerization on both dorsal and ventral surfaces of non-polarized mammalian cells, and also on basolateral surfaces of polarized cells. Actin engagement is necessary, however, to complete membrane deformation into a coated pit on apical surfaces of polarized cells and, more generally, on the surface of any cell in which the plasma membrane is under tension from osmotic swelling or mechanical stretching. We use these observations to alter actin dependence experimentally and show that resistance of the membrane to propagation of the clathrin lattice determines the distinction between "actin-dependent" and "actin-independent". We also find that light-chain bound Hip1R mediates actin engagement. These data thus provide a unifying explanation for the role of actin dynamics in coated-pit budding.

We used fluorescence imaging with one nanometer accuracy (FIONA) to analyze organelle movement by conventional kinesin and cytoplasmic dynein in a cell. We located a green fluorescence protein (GFP)-tagged peroxisome in cultured Drosophila S2 cells to within 1.5 nanometers in 1.1 milliseconds, a 400-fold improvement in temporal resolution, sufficient to determine the average step size to be approximately 8 nanometers for both dynein and kinesin. Furthermore, we found that dynein and kinesin do not work against each other in vivo during peroxisome transport. Rather, multiple kinesins or multiple dyneins work together, producing up to 10 times the in vitro speed.

We present deep-learning-enabled super-resolution across different fluorescence microscopy modalities. This data-driven approach does not require numerical modeling of the imaging process or the estimation of a point-spread-function, and is based on training a generative adversarial network (GAN) to transform diffraction-limited input images into super-resolved ones. Using this Reprints and permissions information is available at www.nature.com/reprints.