Investigating the Cytoskeleton of DRGs Using Cryo-Electron Microscopy and Deep Learning.

We describe a method for determining the ultrastructural organization of axons and varicosities of cultured dorsal root ganglion (DRG) neurons using cryogenic electron microscopy (cryo-EM). Cryo-EM reveals the dimensions, proximity, and overall organization of biological specimens in a near-native state, avoiding artifacts of fixation and heavy metal staining employed in classic thin section ultramicroscopy. Cryo-EM excels with thin specimens, and the axons of cryo-preserved cultured DRG neurons are the ideal thickness for high-resolution cryo-electron tomography. DRG neurons are a particularly interesting neuronal preparation because they can be isolated from animals of any age, providing a unique resource to examine age-related changes in axonal morphology. We provide a detailed, step-by-step protocol from DRG isolation and culturing, through cryo-preservation and data acquisition and analysis. We also provide a description for data processing in batch and how implementing deep-learning strategies can facilitate taking tilt-series data through the process of semi-automated tomographic segmentation required for quantitative descriptions of ultrastructural features. We use segmentations focused on the cytoskeletal elements of axons and varicosities of young and old cultured DRG neurons to highlight the approach.