Fig. 4: The cytoplasmic tail of the SARS-CoV-2 S protein has a suboptimal ER retrieval motif.

a Schematic of the cytoplasmic tail of S protein showing three different mutations made in the COPI-binding site. b Beads coated with the indicated recombinant S tail GST-fusions were incubated with clarified lysate from 293T cells and the eluates immunoblotted for β-COP (COPI), Sec23A (COPII) and moesin. Representative blots from three independent experiments, input 1/50 of that applied to beads. c Quantification of the blots in (b), data from the three independent experiments (arbitrary units), showing mean values and SEMs. d Flow cytometry analysis of U2OS cells expressing the indicated forms of S. The histograms represent the ratio of extracellular S (AF488 signal) to that of intracellular S (AF647 signal). Histograms normalised to the mode value and represent ~10,000 events. Chi-squared tests show the differences between the median ratios of the wild-type (0.87) and those of K1269A (1.1), H1271K (0.11), and T1273A (0.12) to all be statistically significant (P = 0.01, 99% confidence). Representative of four independent experiments. e Micrographs of U2OS cells transiently expressing N-terminally HA-tagged wild-type S or the indicated mutants. Cell surface S was initially stained using AF647-labelled anti-HA under non-permeabilising conditions. Cells were then permeabilised and stained with AF488-labelled anti-HA to detect the internal S, and for the ER marker calreticulin. Both H1271K and T1273A accumulate in the ER, with the latter’s additional perinuclear location presumably reflecting subtle differences in relative ER-to-Golgi, and Golgi-to-ER transport kinetics due to differential effects on COPI and COPII binding. Scale bars 10 μm, with the images for the individual channels taken at the same magnification as the merged image that has the scale bar. The experiment was repeated twice.