a The dynamin inhibitor MiTMAB (20 μM) prevents EGFP-claudin-2 (cyan) removal from tight junctions following induction (+Dox) of mCherry-claudin-4 (magenta) expression. Maximum projection images representative of 5 independent experiments. Scale: 20 µm, 5 µm, 2 µm. b Dynamin inhibition (dark symbols) reduces numbers of EGFP-claudin-2 vesicles (−Dox, cyan) in the absence of claudin-4 and blocks EGFP-claudin-2 endocytosis triggered by mCherry-claudin-4 expression (+Dox, magenta symbols). n = 8. Each data point represents 4–5 cells, representative of 3 independent experiments. 1-way ANOVA. ***P < 0.0001. c MiTMAB (dark symbols) increases EGFP-claudin-2 expression at the tight junction in the absence of claudin-4 (−Dox, cyan symbols) and prevents EGFP-claudin-2 endocytosis after mCherry-claudin-4 expression (+Dox, magenta symbols). n = 9. Each data point represents 4–5 cells, representative of 5 independent experiments. 1-way ANOVA. ***P < 0.0001. Scale: 5 µm. d Endocytic inhibitors (dark symbols) MiTMAB, chlorpromazine (CPZ, 30 μM), or dynamin inhibitory peptide (DIP, 25 μM) do not affect TER in the absence of claudin-4 (cyan symbols) but fail to prevent TER increases induced by mCherry-claudin-4 expression (+Dox, magenta symbols). n = 3, representative of 3 independent experiments. Two-tail unpaired t-test. ***P < 0.0001. e MitMAB (dark symbols) does not affect EGFP-claudin-2 anchoring in the absence of claudin-4 (+Dox, magenta symbols) but is insufficient to prevent EGFP-claudin-2 mobilization after mCherry-claudin-4 expression (+Dox, magenta symbols). n = 4, representative of 3 independent experiments. 1-way ANOVA. **P = 0.0063, ***P < 0.001. Data are presented as mean ± SD and included in the Source Data file.