Extracting Organelle Membrane Topology from Super-Resolution Microscopy Data

Abstract

Constant remodeling of the endoplasmic reticulum (ER) is essential for maintaining its complex structure and function, but difficult to study due to the ER's convoluted topology: the ER exhibits a broad continuum of morphologies ranging from flat, sheet-like domains to highly curved parking-garage structures, tubules, and nanoscale holes. It can transition between these morphologies on a timescale of seconds to minutes. While many proteins that are key players in ER organization have been identified, how the ER maintains its shape is poorly understood. It is particularly unclear how curvature-associated proteins know where and when to migrate along the ER. Due to the structural heterogeneity in the ER, we believe that remodeling processes must be studied in the context of local membrane morphology. To enable this, we have developed super-resolution microscopy techniques that provide a detailed view of ER morphology and dynamics. Single-molecule localization microscopy (SMLM), in particular, provides 3D data at 20-nm resolution. Rather than directly imaging membrane surfaces, however, SMLM gives us the positions of proteins that localize to the ER membrane. Here, we present our tool to create and visualize biophysically-constrained representations of a membrane passing through an SMLM point cloud and its evolution in time. These algorithms are implemented in the PYthon Microscopy Environment (PYME), an open-source nanoscopy imaging and analysis suite (https://www.biorxiv.org/content/10.1101/2020.09.29.315671v1). We are able to quantify the curvature of the ER membrane, and examine the correlation of curvature-associated protein distributions and local nanoscale curvature, which we can use to understand possible mechanisms behind ER remodeling.