Established methods for assessing epithelial barrier function. Methods for measuring epithelial barrier function can be categorized as measuring global barrier function (top row) or local barrier function (bottom row). Results may be detected after a set waiting time, typically ranging from minutes to hours (left), or in real time (right). (A) Tracer permeability assays track the passage of macromolecules from the apical compartment to the basal compartment after a set period of time. Permeability measurements are averaged across the whole tissue. (B) TER measures the electrical resistance of a monolayer. Measurements can be recorded instantaneously. (C) Sandwich assays immobilize tracers once they cross the barrier, allowing sites of barrier breaches to be detected at the experiment end point. (D) ZnUMBA employs the fluorogenic dye FZ3, which increases in fluorescence at sites of barrier breaches, allowing real-time monitoring of barrier breaches and their repair.

Modification of ZnUMBA for different model systems. (A) X. laevis embryos are injected with mRNAs of interest at early embryo stages (1–4-cell stage). After developing to gastrula stage, FZ3 is injected into the blastocoel, and the embryos are incubated for a minimum of 5 min to allow the injection site injury to heal. Finally, just prior to imaging, embryos are mounted in Zn2+-containing medium and imaged via confocal microscopy. (B) Zebrafish embryos are injected with mRNAs of interest at the 1-cell stage. Embryos are dechorionated, and after developing to 3 hpf (1k-cell stage), embryos are injected with FZ3 and fluorescently labeled dextran into the interstitial fluid. Embryos are incubated for ∼15 min to allow the injection site injury to heal. Finally, embryos are mounted in low-melting-point agarose, and Zn2+-containing medium is added before the start of imaging. (C) Experimental setup for ZnUMBA using MDCK II cultured epithelial cells. The Transwell filter cup is placed upside-down on a clean surface, and 1×105 cells resuspended in 300 µl of DMEM are seeded onto the bottom surface of the filter. The filter is incubated at 37°C in a moist CO2 incubator for 10–14 h. After cells are attached to the surface, the filter cup is inverted and placed into a well of a 12-well plate. The cells are cultured for ∼5 d until the TER increases. For ZnUMBA, 500 µl of HBSS containing 2 mM ZnCl2 is placed on the glass-bottom dish, and the Transwell filter cup with the cell sheet attached is placed onto the Zn2+-containing medium. HBSS containing 10 µM FZ3 and 1 µM CaCl2-EDTA is added into the filter cup (upper compartment), and the fluorescence is observed using an inverted fluorescence microscope. For visualization of the basal compartment, RITC–dextran can be included in the ZnCl2 solution.

ZnUMBA reveals a temporary but weak resealing of the barrier following laser injury and prior to contraction-mediated junction repair. (A–A″) Junction injury was performed by exposing the area indicated by the white dotted circle (A′) to intense 405 nm laser light. The junction that was injured is indicated by white box in A. Following laser injury (A′, time 0), FZ3 fluorescence [shown using the Green Fire Blue lookup table (LUT) applied using FIJI; bottom bar] increases sharply at the site of injury and less intensely along the length of the junction. F-actin (Lifeact–mRFP, shown using the Fire LUT applied using FIJI; top bar) accumulates at the site of the injury, and the barrier breach is repaired. A″ shows side views (x–z) of the images shown in A′. White arrow points out that FZ3 signal at the site of the injury is more apical than the signal along the length of the junction. (B) Quantification of laser injury experiments. The mean pixel intensity of a 1 µm-wide line drawn over the injured junction from vertex to vertex was normalized to a reference junction from the same movie. Graph shows mean normalized intensity (left axis) or mean junction length change (right axis)±s.e.m. (n=25 junctions from nine embryos across three experiments). (C–E) Junction injury was performed as in A. FZ3 is shown in the Green Fire Blue LUT (left bar), as in A, and membrane (mCherry–farnesyl) is shown in the Cyan Hot LUT applied using FIJI (right bar). Intensity values are in arbitrary units. Lifeact–miRFP703 is not shown. Peak FZ3 intensity is marked with an asterisk, evidence of membrane reorganization is indicated by a white arrowhead and endocytic vesicles are indicated by a yellow arrowhead. Laser injury induced a bleb-like membrane protrusion in five of 18 injured junctions (C,D). In three of these cases, the membrane protrusion occurred after FZ3 intensity peaked (C), and in the other cases the FZ3 intensity peak coincided with the expansion of the membrane protrusion (D). In the remaining 13 of 18 junctions, laser injury induced membrane reorganization during the contraction phase of junction repair (E), consistent with bunching or folding of the membrane during contraction. Eight of the eighteen junctions showed evidence of endocytosis (yellow arrowhead), and six of eighteen junctions had multiple FZ3 peaks, consistent with a temporary but weak resealing of the TJ barrier. n=18 junctions from seven embryos across one experiment. (F) A speculative model of how TJ strands form a temporary but weak seal prior to reinforcement via contraction of the strand network. (1) An intact TJ strand network. (2) Laser injury induces breaks in TJ strands, allowing Zn2+ and FZ3 to mix. (3) The strand network is partially re-established by annealing or elongation of existing strands but is susceptible to future breaks because of their dynamic nature. In some cases, damaged areas of the strand network might be removed via endocytosis. (4) Actomyosin-mediated junction contraction establishes robust crosslinking of the TJ network, making it less susceptible to future breaks.

ZnUMBA detects barrier breaches in zebrafish embryos upon pharmacological and genetic perturbations. (A,B) Consecutive images at the indicated times of an imaging plane at the interface of the EVL and deep cells are depicted, with cell membranes (mem) in red, FZ3 in green and Alexa Fluor 647–dextran (Dex) in magenta. Control and EGTA-treated embryos are shown in A, and control and MZpoky mutant embryos are shown in B. First row shows x–y views. Second row shows side views (x–z), with most interstitial fluid accumulation below the EVL cells. Third row displays the ratio of mean intensities of FZ3 fluorescence divided by the fluorescence intensity of Alexa Fluor 647–dextran (FZ3/Dex ratio); the Fire lookup table (LUT) has been applied using FIJI, with a LUT calibration bar shown on the right. Quantification panels (A′,B′) show plots of FZ/Dex ratio for control (black) and EGTA-treated (A′) or MZpoky mutant (B′) embryos (magenta) as a function of time. At the sampled time resolution of 0.5–6 min, significant barrier breaches are detected in both EGTA-treated embryos and MZpoky mutant embryos. Note that the intensity of the FZ3 images for EGTA-treated embryos at 64 min and MZpoky embryos at 36 min was decreased for better display. Control (EGTA), N=4, n=7; EGTA, N=3, n=6. Control (MZpoky), N=3, n=5; MZpoky, N=3, n=5. N, number of independent experiments; n, number of embryos. Scale bars: 10 µm. See also Movie 1.

ZnUMBA detects transient local barrier defects in MDCK II cultured epithelial cells. (A) ZnUMBA of WT cells and Cldn2-KO cells. GFP–nls-labeled WT MDCK II cells and Cldn2-KO MDCK II cells were mixed and co-cultured on the bottom surface of the filter. Images before (left) and after (middle) FZ3 addition are shown. The Fire lookup table (LUT) has been applied using ImageJ. Right panel is the magnified view of the region outlined by a rectangle in the middle panel. Note that the cell–cell junctions between WT cells labeled with nuclear GFP have higher signal compared to those between Cldn2-KO cells. Scale bars: 20 µm. (B,C) ZnUMBA of Ang1-KO cells in the Cldn2-KO background. GFP–nls-labeled Ang1/Cldn2-dKO MDCK II cells and Cldn2-KO MDCK II cells (B), or Ang1/Cldn2-dKO MDCK II cells and mCherry–nls-labeled Cldn2-KO MDCK II cells (C), were mixed and co-cultured. Fire (B) and Gem (C) LUTs from ImageJ have been applied to the FZ3 channel. mCherry signal is shown pseudocolored in blue (C). Boxes indicate regions shown as magnified views in the right-hand panels. Note that tricellular junctions between Ang1/Cldn2-dKO cells (green arrows) have higher signal than bicellular junctions. Scale bars: 20 µm. Images in A–C are representative of four experiments.

ZnUMBA detects naturally occurring leaks at cell–cell boundaries. (A) Brightest-point projections of FZ3 signals using Cldn2-KO cells over four time intervals (min:s). Scale bar: 20 µm. (B) Time-lapse images (min:s) of ZnUMBA using Cldn2-KO cells. RITC–dextran (lower panels, white) was added to the FZ3 solution to visualize the basal compartment of paracellular space. The Fire lookup table from ImageJ has been applied to the FZ3 channel (upper panels). Kymographs of the FZ3 and RITC–dextran (R-dex) signals in the yellow rectangles (a, b and c) are shown (bottom). Note that ZnUMBA signal fluctuates over time, whereas RITC–dextran signal remains unchanged. Scale bar: 20 µm. See also Movie 2. Images in A and B are representative of three experiments.