QUICK SEARCH:   [advanced] Author: Keyword(s):
Home     Help     Feedback     Subscriptions     Archive     Search     Table of Contents

First published online 8 January 2003
doi: 10.1242/jcs.00300

This Article Summary Full Text Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Hopkins, A. M. Articles by Nusrat, A. Search for Related Content PubMed PubMed Citation Articles by Hopkins, A. M. Articles by Nusrat, A.

Constitutive activation of Rho proteins by CNF-1 influences tight junction structure and epithelial barrier function

¹ Epithelial Pathobiology Research Unit, Department of Pathology and Laboratory Medicine, Emory University, Whitehead Biomedical Research Building, Atlanta, GA 30322, USA
² Max Planck Institute of Molecular Cell Biology and Genetics, Pfotenhauerstrasse 108, 01307 Dresden, Germany
³ INSERM Unite 452, IFR 50, Faculté de Medecine, 28 Avenue de Valombrose, F-06107, Nice, France

View larger version (43K): [in a new window]   Fig. 1. CNF-1 activates RhoA in T84 intestinal epithelial cells. Confluent T84 intestinal epithelial monolayers were incubated with either CNF-1 (1 nM) or vehicle for 24 hours, and RhoA, Rac1 and Cdc42 activation status was examined by pull-down assays involving binding to the GTPase-binding domains of rhotekin or PAK-1 conjugated to agarose beads. Total Rho, Rac and Cdc42 protein levels were similar in lysates from control and CNF-treated monolayers. However, in pull-down activation assays, more Rho, Rac and Cdc42 were recovered from CNF-treated cells relative to controls, indicating increased activation of these GTPases.

View larger version (14K): [in a new window]   Fig. 2. CNF-1 disrupts TJ gate function in a polarized intestinal epithelial monolayer. T84 monolayers were incubated basolaterally with increasing concentrations of CNF-1 (0.005-5 nM; n=4-12 monolayers) for 24 hours. CNF-1 evoked a concentration-dependent reduction in TER that was maximal at 1 nM (A). In parallel, monolayers were treated either apically (light gray, center) or basolaterally (dark gray) with 1 nM CNF-1 or with vehicle alone (black) for 24 hours (B). Permeation of apically loaded fluorosceinated dextran (FD-3; MW 3000) into the basal compartment over 2 hours was used as an index of passive paracellular transport. Flux of FD-3 across control monolayers and those treated apically with CNF-1 were virtually identical irrespective of CNF-1 incubation time. However, basolateral treatment with CNF-1 augmented FD-3 flux in a time-dependent fashion, which was statistically significant at t=24 and 48 hours (n10 monolayers per condition).

View larger version (82K): [in a new window]   Fig. 3. CNF-1 alters morphological localization of candidate TJ proteins. T84 monolayers were exposed to CNF-1 or vehicle alone for 24 hours, and TJ proteins occludin, ZO-1 and JAM were localized by immunofluorescence and confocal microscopy. In control monolayers (a), occludin localized sharply in the apical region of the lateral membrane, visualized en face as a ring pattern. Monolayers treated apically for 24 hours with CNF-1 (b) were identical. Monolayers exposed basolaterally to CNF-1 for 24 hours (c) displayed dramatic redistribution of occludin away from the lateral TJ membrane (arrow). ZO-1 staining in control monolayers (d) resembled that of occludin, outlining TJs just below the apical plane. Apical exposure to CNF-1 for 24 hours did not affect ZO-1 immunolocalization (e); however, discontinuities in ZO-1 ring structures were evident in monolayers treated basolaterally with CNF-1 for the same time period (f). JAM in control monolayers (g) was enriched in the TJ plane, and unchanged by apical treatment with CNF-1 toxin for 24 hours (h). However, basolateral exposure to CNF-1 for the same time period (i) somewhat disrupted JAM distribution, with diffusion away from the TJ membrane into the cytoplasm. Levels of occludin (j), ZO-1 (k) or JAM (1) proteins were minimally different in lysates prepared from control (lane 1) versus CNF-treated (lane 2) epithelial monolayers. Bar10 µm.

View larger version (44K): [in a new window]   Fig. 4. CNF-1 stimulates occludin internalization at an ultrastructural level. Confluent control and CNF-1-treated T84 monolayers were processed for cryo-immunoelectron microscopy (A) by labeling occludin with 15 nm gold particles (pseudo-colored in red) and caveolin-1 with 10 nm gold particles (black). In control cells (a), occludin localized mainly to the apical-most part of the basolateral plasma membrane (arrow). In CNF-1 treated cells, there was a reduction in occludin at the plasma membrane of the TJ domain (b, #). Some labeling was now found to co-localize with caveolin-1 in intracellular caveolae and endosomal-like compartments (c). Bar, 100 nm. Following quantitation of occludin (black bars) and caveolin-1 (gray bars) gold particles at the TJ, statistically significant reductions in both proteins were observed in CNF-treated monolayers relative to controls (B).

View larger version (77K): [in a new window]   Fig. 5. CNF-induced internalization of occludin is via caveolae and early-/recycling endosomes but not late endosomes. T84 epithelial monolayers were incubated with vehicle or CNF-1 (1 nM) for 24 hours, and the internalization of occludin monitored in relation to markers of endosomal/caveolar pathways by immunofluorescence/confocal microscopy. Occludin in the en face plane of control cells (a1, c1, e1, g1, i1) was characteristically distributed in discrete TJ rings at cell-cell borders. Caveolin-1 staining in control cells (a2) consisted of a large pool of apical staining, together with a sub-pool of caveolin-1 localizing in a ring pattern at TJs (arrowhead). This was observed to co-localize with occludin (a3, #). Occludin internalization in CNF-treated cells (b1, arrows) was seen to correspond closely with a pool of internalized caveolin-1 in the same cells (b2, arrowheads) and in a composite image (B3, #). The transferrin receptor in control cells (c2) localized mainly to the apical surface and underneath the membrane, showing no overlap (c3) with the staining pattern for occludin. Following CNF-1 incubation, internalized occludin (d1, arrow) and internalized transferrin receptor (d2, arrowhead) did not appear to co-localize with each other (d3, #). Another early endosomal marker, EEA-1, localized to sub-membranous structures with some additional staining at the TJ membrane (E2, arrowhead). This, like caveolin-1 staining in control cells, overlapped with occludin (e3, #). Occludin internalization upon CNF-1 incubation (F1, arrows) overlapped with EEA-1 distribution in the same cells (F2, arrowheads). This overlap in the merged image (f3, #) suggests internalization of occludin in EEA-1-positive early endosomes. Distribution of the recycling endosomal marker Rab11 was primarily submembranous in control cells (g2), showing no co-localization with occludin (g3). However, occludin (h1, arrow) and Rab11 (h2, arrowhead) internalization in CNF-treated monolayers were observed to overlap significantly (h3, #). The late endosomal marker LAMP-1 localized to the membrane at the level of intercellular junctions in control cells (i2), in a pattern that resembled but did not overlap with (i3, #) that of occludin. Occludin internalization after CNF-1 treatment (j1) did not co-localize with LAMP-1-positive late endosomes in the same cells (j3, #). Results are representative of 6 experiments. Bar10 µm.

View larger version (100K): [in a new window]   Fig. 5. CNF-induced internalization of occludin is via caveolae and early-/recycling endosomes but not late endosomes. T84 epithelial monolayers were incubated with vehicle or CNF-1 (1 nM) for 24 hours, and the internalization of occludin monitored in relation to markers of endosomal/caveolar pathways by immunofluorescence/confocal microscopy. Occludin in the en face plane of control cells (a1, c1, e1, g1, i1) was characteristically distributed in discrete TJ rings at cell-cell borders. Caveolin-1 staining in control cells (a2) consisted of a large pool of apical staining, together with a sub-pool of caveolin-1 localizing in a ring pattern at TJs (arrowhead). This was observed to co-localize with occludin (a3, #). Occludin internalization in CNF-treated cells (b1, arrows) was seen to correspond closely with a pool of internalized caveolin-1 in the same cells (b2, arrowheads) and in a composite image (B3, #). The transferrin receptor in control cells (c2) localized mainly to the apical surface and underneath the membrane, showing no overlap (c3) with the staining pattern for occludin. Following CNF-1 incubation, internalized occludin (d1, arrow) and internalized transferrin receptor (d2, arrowhead) did not appear to co-localize with each other (d3, #). Another early endosomal marker, EEA-1, localized to sub-membranous structures with some additional staining at the TJ membrane (E2, arrowhead). This, like caveolin-1 staining in control cells, overlapped with occludin (e3, #). Occludin internalization upon CNF-1 incubation (F1, arrows) overlapped with EEA-1 distribution in the same cells (F2, arrowheads). This overlap in the merged image (f3, #) suggests internalization of occludin in EEA-1-positive early endosomes. Distribution of the recycling endosomal marker Rab11 was primarily submembranous in control cells (g2), showing no co-localization with occludin (g3). However, occludin (h1, arrow) and Rab11 (h2, arrowhead) internalization in CNF-treated monolayers were observed to overlap significantly (h3, #). The late endosomal marker LAMP-1 localized to the membrane at the level of intercellular junctions in control cells (i2), in a pattern that resembled but did not overlap with (i3, #) that of occludin. Occludin internalization after CNF-1 treatment (j1) did not co-localize with LAMP-1-positive late endosomes in the same cells (j3, #). Results are representative of 6 experiments. Bar10 µm.

View larger version (68K): [in a new window]   Fig. 6. CNF-1 does not dramatically influence AJ protein localization in the lateral membrane. E-cadherin and ß-catenin were immunolocalized in confluent T84 monolayers exposed basolaterally to CNF-1 (1 nM, 48 hours) or vehicle. In control monolayers, both E-cadherin (a) and ß-catenin (c) were visualized en face in characteristic ring formations in the AJ. Basolateral exposure to CNF-1 did not visibly alter the ring pattern of E-cadherin (b), although minor thickening of ß-catenin staining intensity was observed at cell-cell borders (d). Total protein levels of E-cadherin (e) and ß-catenin (f) were unaffected by CNF-1 treatment (lanes 1-3 control, 4-6 CNF-1). Bar10 µm.

View larger version (72K): [in a new window]   Fig. 7. CNF-1 disrupts morphological localization of a TJ-associated pool of phosphorylated MLC. In control T84 monolayers double-labeled with antibodies to ZO-1 (a; green) and p-MLC (b; red), extensive co-localization (c; yellow) was observed in the TJ plane. In monolayers treated apically for 24 hours with CNF toxin, ZO-1 (d) and p-MLC (e) were also seen to co-localize (f). However, monolayers treated basolaterally with CNF-1 for the same time period displayed not only profound disruption of ZO-1 staining (g) but also p-MLC (h) in the TJ plane. Fragments of each protein from the disassembled ring structures co-localized in a punctate pattern (i). Bar10 µm.

View larger version (62K): [in a new window]   Fig. 8. CNF-1 induces polarized restructuring of F-actin and actin-binding proteins in epithelial monolayers. Rhodamine phalloidin was used to highlight F-actin organization in the apical and basal poles of T84 epithelial cells following polarized incubation with CNF-1 or vehicle control. While control monolayers (a) and those incubated apically with CNF-1 (b) revealed identical organization of F-actin in en face confocal micrographs, basolateral exposure to CNF-1 (c) induced dramatic polarized F-actin restructuring. Slight increases in staining intensity were observed in the perijunctional F-actin ring, but the normal pattern of punctate microvillous staining was virtually abolished in the apical pole of these monolayers (arrow) relative to controls. These changes were accompanied by altered immunolocalization of the brush border actin-binding protein villin in CNF-treated monolayers (f) relative to control cells (d) and those treated apically with toxin (e). F-actin distribution at the basal epithelial pole was similar in both control (g) and apically-treated monolayers (h), and consisted of diffuse meshworks of short stress fibers. By contrast, basolateral exposure to CNF-1 (i) stimulated the aggregation of stress fibers into thicker, `cabled' bundles. Parallel changes were detected in immunolocalization of the actin-binding protein paxillin at the basal epithelial pole. Control monolayers (j) displayed `plaque-like' paxillin immunoreactivity, similar to that in monolayers exposed apically to toxin (k). Basolateral treatment with CNF-1 (l) was associated with a slight increase in plaque number and size. CNF-induced changes in F-actin, villin and paxillin (m, n, o, respectively) were not accounted for by changes in the total levels of these proteins as assessed by western blot analysis of lysates from control (lanes 1-3) and CNF-treated monolayers (lanes 4-6) for the times indicated. Slight increases in the molecular mass of paxillin upon CNF incubation reflected tyrosine phosphorylation of the protein, as shown by the increased level of phospho-specific paxillin detected in lysates from CNF-treated cells relative to controls (p). Bar10 µm.

View larger version (64K): [in a new window]   Fig. 9. CNF-1 impairs the assembly of junctions in calcium switch assays. (A) T84 monolayers exposed to EGTA (2 mM) in calcium- and magnesium-free media were allowed to recover in complete culture media (containing calcium and serum) in the presence of CNF-1 (1 nM) or vehicle. TER was monitored over time and used to assess the recovery of cell-cell contact. Monolayers exhibited high resistances prior to EGTA treatment (#), which dropped to 10 .cm2 following EGTA incubation (not shown). During the recovery period, control monolayers recovered high TER values of >700 .cm2 after approximately 6 hours. By contrast, monolayers exposed to CNF-1 during recovery failed to reach TER values >170 .cm2 (n=4) by 24 hours, indicating impaired junctional reassembly. (B) Occludin and E-cadherin staining was examined in monolayers immediately after EGTA treatment, and following recovery for 9 hours in complete media containing either CNF-1 (1 nM) or vehicle. Following EGTA treatment, both proteins were seen to move away from the lateral membrane into an intracellular compartment (a,b). During recovery in complete media with vehicle, both occludin (c) and E-cadherin (d) redistributed correctly into characteristic ring structures at the lateral membrane. By contrast, monolayers exposed to CNF-1 appeared to assemble only E-cadherin but not occludin into ring structures at the same time period (e,f).