(A) S1P₂ immunoblot of HBE cells expressing control (left) or S1P₂-specific shRNA (right) with α-tubulin as loading control. (B) Representative images of HBE cells (DNA only) expressing control (left) or S1P₂-specific shRNA (right) grown for 3 weeks. Scale bar, 10 µm. (C) Proliferation assay indicates that S1P₂-knockdown cells proliferate at the same rate as wild type controls cells. (D) Representative DIC micrographs of 5-dpf WT (top) and Mil (S1P₂ mutant) (bottom) zebrafish larvae, where cartoon shows region where fish was imaged. Scale bars, 100 µm where red box indicates region imaged. (E) Quantification of epidermal clumps of 22 zebrafish larvae.

(A) Representative confocal projections of 5-dpf WT (left) and Mil (S1P₂ mutant) (right) zebrafish larvae treated with DMSO or FAK inhibitor PF573228, where high Texas-Red Dextran indicates poor permeability in Mil but is greatly reduced when barrier function is improved with FAK inhibitor treatment. Scale bar = 10 µm and red box indicates region of fish imaged. (B) 5 dpf Mil and wild type zebrafish treated with and without FAK inhibitor. Note that while FAK inhibitor-treated Mil have other developmental defects (heart and circulation), there are no obvious clumps as seen in the untreated fish. Scale bar = 100 µm and red box indicates region of fish imaged. Note FAK inhibitor does not affect WT zebrafish. (C) Quantification of epidermal masses in 5 dpf Mil zebrafish larvae with and without PF573228. (D) Quantification of apoptotic cells within epidermal masses with and without PF573228. For both, error bars are SD and p values are ***<0.0001. (E) Immunoblot showing knockdown of S1P₂ by photo-activatable morpholinos. (F) Stills from a video where PF573228 was added to S1P₂ morphant at 5 dpf, where red arrows show the edge of the epidermis over time, scale bar = 50 µm and red box indicates region of fish imaged. Time is hours:minutes following FAK inhibitor addition. Note: epidermal cells that are sloughed off become embedded in the agarose where fish is mounted.