The GFP–COL puncta colabeling for the inner COPII layer do not exceed 450 nm in diameter. (A) Confocal images obtained using STED microscopy of whole RPE-1 cells stably expressing GFP–COL (GFP–COL–RPE; colabeled with an antibody against GFP in the same channel; green) cotransfected with the inner layer COPII-marker mSc-Sec23A and additionally colabeled with an antibody against Sec24C in the same channel (magenta). Time points indicate incubation in presence of asc/biotin (500 µg·ml⁻¹ and 400 µM, respectively) before fixation with PFA. Scale bar, 10 µm; n = 14. (B) Corresponding zoomed in areas of the cells in A, obtained using superresolution microscopy (gSTED). Scale bar, 1 µm. (C) Enlargements of small GFP-positive puncta colocalizing with the COPII marker and large GFP-positive structures extracted from images as displayed in B. Panels show the different channels in gray scale, followed by the overlay image and the overlay image containing the line with a width of 5 (images ii–v, vii, and ix) or 10 pixels (images i, vi, viii, and x) drawn through the object of interest to generate the corresponding line graphs, as shown below. Line graphs display the distance in micrometers on the x axis and the signal intensity as an au on the y axis. The FWHM values corresponding to the estimated maximum object diameters in nanometers or micrometers are displayed above the images for the corresponding line graphs fitted with a Gaussian curve (in green for GFP + GFP–COL and in magenta for mSc-Sec23A + Sec24C). When two curves were used to fit the graphs, the sum of the FWHM was used as an estimate (indicated by asterisk). When Gaussian fitting was not possible, the displayed estimated diameter value was measured via ImageJ (two asterisks). If no peak in the red channel could be identified, values were left blank (indicated by dash).