Extended Data Fig. 1: Isolated GC verification.

a, Enrichment analysis of GC fraction versus starting homogenate (input). Left, western blot detects protein enrichment of GAP43 (GC marker) and depletion of MAP2 (somato-dendritic protein marker) in GC fraction. Middle, native gel electrophoresis shows de-enriched presence of large (28S) and small (18S) ribosomal subunit rRNA in GC fraction. Right, RT–PCR detects mRNA for Actb (ubiquitous) but not Gfap (progenitor and glial marker) from the GC fraction. n = 6 biological replicates for protein and rRNA; n = 3 for mRNA. b, GC protection assay schema: bulk GC fraction isolated after subcellular fractionation is a suspension of GC particles enclosing GC-specific molecules (blue) within a medium that contains dilute soluble cytosolic molecules (red). Treatment with RNase or protease leads to hydrolysis of RNA and protein in the suspension medium not protected within GC particles, leaving only the GC-protected molecules (blue) after treatment. Addition of detergent before treatment results in hydrolysis of both cytosolic as well as GC-encapsulated molecules due to ruptures in the encapsulating GC plasma membrane, providing a control for the efficiency of enzymes. The difference in RNA or protein signal between hydrolysis control and GC-protected samples corresponds to the GC-encapsulated signal. c, d, GC protection assays with non-membrane-permeable degrading enzymes (protease in c and RNase in d) to test GC integrity and GC-specific membrane encapsulation of RNA and proteins in isolated GCs. Treatment with enzyme plus detergent, but neither alone, completely abolishes RNA and protein signal from GC fractions. Signals persisting in treatments with enzyme alone (lanes 3) correspond to RNA and protein encapsulated (protected) by the GC membrane, and correspond to the specific molecular content of isolated GCs. Treatment with protease alone has no effect on the signal from GAP43, confirming GC-specificity. Conversely, the signal from GM130, a Golgi matrix protein known to be excluded from GCs, is abolished with protease treatment alone, indicating there is no non-specific encapsulation in GCs. Reduced presence of both Actb and rRNA in samples treated with enzyme alone is consistent with their ubiquitous presence in both GCs and elsewhere in the homogenate. n = 5 independent biological replicates. e, Bioanalyzer profiles show GC-protected RNA compared to detergent-treated control, with characteristic peaks corresponding to 28S and 18S rRNA, and a spectrum of low intensity signal characteristic of mRNA. Experiments performed in n = 5 biological replicates with consistent results.