Rab13 mediates the continuous endocytic recycling of occludin to the cell surface.

During epithelial morphogenesis, adherens junctions (AJs) and tight junctions (TJs) undergo dynamic reorganization, whereas epithelial polarity is transiently lost and reestablished. Although ARF6-mediated endocytic recycling of E-cadherin has been characterized and implicated in the rapid remodeling of AJs, the molecular basis for the dynamic rearrangement of TJs remains elusive. Occludin and claudins are integral membrane proteins comprising TJ strands and are thought to be responsible for establishing and maintaining epithelial polarity. Here we investigated the intracellular transport of occludin and claudins to and from the cell surface. Using cell surface biotinylation and immunofluorescence, we found that a pool of occludin was continuously endocytosed and recycled back to the cell surface in both fibroblastic baby hamster kidney cells and epithelial MTD-1A cells. Biochemical endocytosis and recycling assays revealed that a Rab13 dominant active mutant (Rab13 Q67L) inhibited the postendocytic recycling of occludin, but not that of transferrin receptor and polymeric immunoglobulin receptor in MTD-1A cells. Double immunolabelings showed that a fraction of endocytosed occludin was colocalized with Rab13 in MTD-1A cells. These results suggest that Rab13 specifically mediates the continuous endocytic recycling of occludin to the cell surface in both fibroblastic and epithelial cells.

Polarized epithelial cells create apical and basolateral plasma membrane (PM) 1 domains that face the lumen and basement membrane, respectively. These two domains have distinct functional properties and display unique lipid and protein compositions. Tight junctions (TJs) act as a fence preventing the lateral diffusion of proteins and lipids between the two domains and as a gate regulating solute flux through the paracellular space (1). Proteins constituting TJs include the transmembrane proteins mediating cell-cell adhesion and the cytosolic plaque proteins linking TJs to the cytoskeleton and participating in intracellular signaling (2). Occludin, claudins, and junctional adhesion molecules constitute the transmembrane proteins in TJs, whereas a number of scaffolding proteins and signaling molecules such as zonula occludens-1 and Par3-Par6-atypical protein kinase C complex have been identified as the cytosolic plaque proteins in TJs (3)(4)(5)(6)(7)(8). However, the molecular mechanisms governing the de novo formation of TJs are poorly understood.
The maintenance of epithelial polarity depends on the continuous sorting of membrane proteins and lipids into distinct cell surface domains. This sorting occurs both in the trans-Golgi network during biosynthetic vesicular transport and, after endocytosis, in endosomes (9,10). These polarized vesicular transport pathways must be strictly regulated to ensure the controlled loss and recovery of epithelial polarity during dynamic morphogenetic events. The Rab family small G proteins consist of more than 60 family members in mammalian cells and play a crucial role in determining the specificity of vesicular transport pathways. Each Rab family member that is localized to a distinct compartment in the exocytic or endocytic pathway functions as a molecular switch, cycling between the GTP-and GDP-bound conformations, at the compartment where they reside (11)(12)(13). Two Rab family members, Rab3B and Rab13, localize to TJs, and it has been proposed that they function as putative regulators of polarized vesicular transport to and/or from TJs in polarized epithelial cells. Indeed, recent evidence suggests that Rab13 regulates the assembly of functional TJs and that the role of Rab13 in polarized vesicular transport from the trans-Golgi network to the cell surface is distinct from that of Rab3B (14 -17).
Adherens junctions (AJs) and TJs are points of adhesion between epithelial cells that ensure the maintenance of appropriate epithelial cell polarity. Cadherins are essential adhesion molecules within AJs supporting not only stable cell-cell contacts but also dynamic morphogenetic events such as epithelial-mesenchymal transitions (EMT) and mesenchymal-epithelial transitions (MET) (18). During EMT, E-cadherin, the prototypical epithelial cadherin, is down-regulated by transcriptional silencing and/or protein degradation through the ubiquitin-proteasome pathway (19). The endocytosis and recycling of E-cadherin have recently emerged as alternative mechanisms allowing cells to undergo rapid changes in morphology in response to extracellular stimuli (20 -22). The small G protein ARF6 is implicated in the regulation of the endocytic recycling of E-cadherin (23). Among the transmembrane proteins in TJs, occludin and claudins, a family composed of more than 20 different members, are thought to be responsible for the formation of TJs and epithelial polarization (4,5). Like cadherins, it has been shown that both occludin and claudins are transcriptionally silenced during EMT, and occludin is subject to post-transcriptional protein degradation by the ubiquitin-proteasome pathway (24,25). To begin to understand the molecular mechanisms underlying the dynamic change in epithelial polarity during EMT and/or MET, we examined the intracellular transport of occludin and claudins to and from the cell surface in both fibroblastic BHK and epithelial MTD-1A cells. When occludin was expressed in BHK cells, it was continuously endocytosed and recycled back to the cell surface. Endogenous occludin was also subjected to continuous endocytic recycling in MTD-1A cells. Our results indicate that Rab13, one of the cytosolic plaque proteins in TJs, directs the continuous endocytic recycling of occludin in both fibroblastic and epithelial cells.
Recombinant Adenovirus Infection-Recombinant adenoviruses expressing EGFP and EGFP-Rab13 Q67L (Ad-EGFP and Ad-EGFP-Rab13 Q67L) were constructed using the Transpose-Ad Adenoviral Vector System (Qbiogene, Carlsbad, CA) according to the manufacturer's instructions. Briefly, EGFP and EGFP-Rab13 Q67L cDNAs were cloned into a pCR259 transfer vector. Recombinant adenoviral plasmid was generated by Tn7-mediated transposition in Escherichia coli. The resulting plasmid was linearized by PacI digestion and transfected into QBI-HEK293 cells using an MBS Mammalian Transfection kit (Stratagene). After a 24-h transfection, cells were split into a 96-well plate and incubated at 37°C for 10 -14 days. Screening of recombinant adenovirus was done by PCR and immunoblot analysis. Recombinant ad-enovirus was amplified in QBI-HEK293 cells, and its titer was determined by multiplicity of infection test. MTD-1A cells were infected with Ad-EGFP (mock) or Ad-EGFP-Rab13 Q67L at a multiplicity of infection of 100. After a 36-h culture, cells were subjected either to endocytosis or recycling assay.
Endocytosis Assay-Endocytosis assay was performed as described previously (21). Briefly, cell surface proteins were biotinylated with 0.5 mg/ml Sulfo-NHS-SS-Biotin (Pierce) in PBS containing 0.9 mM CaCl 2 and 0.33 mM MgCl 2 (PBS/CM) at 4°C for 30 min, quenched with 50 mM NH 4 Cl in PBS/CM at 4°C for 15 min, and incubated at 37°C or 18°C for the indicated periods of time to allow endocytosis. The remaining biotin on the cell surface was stripped with 50 mM MESNA in 100 mM Tris/HCl (pH 8.6) containing 100 mM NaCl and 2.5 mM CaCl 2 at 4°C for 30 min and quenched with 5 mg/ml iodoacetamide in PBS/CM at 4°C for 15 min. After lysis with 50 mM Tris/HCl (pH 8.0) containing 1.25% Triton X-100, 0.25% SDS, 150 mM NaCl, 5 mM EDTA and 10 g/ml APMSF, an aliquot was taken to determine the total amount of cargo proteins expressed in the cells. Biotinylated cargo proteins were then isolated with UltraLink Immobilized NeutrAvidin Plus beads (Pierce). The samples were prepared for immunoblot analysis. The values for biotinylated cargo proteins protected from MESNA treatment were normalized to total cargo proteins expressed in the cells.
Immunoblot-Samples were separated on SDS-PAGE, and proteins were transferred to a polyvinylidene difluoride membrane. Membrane blocking and antibody dilutions were done in Block Ace (Dainippon Pharmaceutical, Osaka, Japan). Blots were developed by chemiluminescence using horseradish peroxidase-coupled secondary antibody (Jackson ImmunoResearch Laboratories, West Grove, PA) and an ECL-Plus kit (Amersham Biosciences). Quantitation was performed after scanning of the autoradiograph film of nonsaturating signals by using a NIH Image 1.62 program.
Recycling Assay-Recycling assay was performed as described previously (21). Briefly, cell surface proteins were biotinylated with 0.5 mg/ml Sulfo-NHS-SS-Biotin (Pierce) in PBS/CM at 4°C for 30 min, quenched with 50 mM NH 4 Cl in PBS/CM at 4°C for 15 min, and incubated at 37°C for the indicated periods of time to allow endocytosis. The remaining biotin on the cell surface was stripped with 50 mM MESNA in 100 mM Tris/HCl (pH 8.6) containing 100 mM NaCl and 2.5 mM CaCl 2 at 4°C for 30 min, and quenched with 5 mg/ml iodoacetamide in PBS/CM at 4°C for 15 min. Cells were again incubated at 37°C for the indicated periods of time to allow recycling of endocytosed cargo proteins back to the cell surface. Then newly appeared cell surface biotin was again stripped with 50 mM MESNA in 100 mM Tris/HCl (pH 8.6) containing 100 mM NaCl and 2.5 mM CaCl 2 at 4°C for 30 min, and quenched with 5 mg/ml iodoacetamide in PBS/CM at 4°C for 15 min. After lysis with 50 mM Tris/HCl (pH 8.0) containing 1.25% Triton X-100, 0.25% SDS, 150 mM NaCl, 5 mM EDTA, and 10 g/ml APMSF, an aliquot was taken to determine the total amount of cargo proteins expressed in the cells. Biotinylated cargo proteins were then isolated with UltraLink Immobilized NeutrAvidin Plus beads (Pierce). The samples were prepared for immunoblot analysis. The values for biotinylated cargo proteins protected from MESNA treatment were normalized to total cargo proteins expressed in the cells.
Bafilomycin A1 (BAF) Treatment-BHK cells were preincubated in 1 M BAF at 4°C for 30 min and then incubated in 1 M BAF at 37°C for the indicated periods of time to allow recycling of endocytosed cargo proteins back to the cell surface. The recycling assay was performed as described above.
Immunofluorescence Microscopy-MTD-1A cells grown on glass coverslips were fixed with either 1% formaldehyde in PBS for 15 min at room temperature (for immunostaining of occludin, occludin-HA, HA-GFP, HA-TfR, occludin-GFP, occludin-Rab4, occludin-Rab11, and occludin-TfR), Ϫ20°C methanol for 5 min on ice (for immunostaining of claudin-1), or 10% trichloroacetic acid in PBS for 15 min on ice (for immunostaining of HA-claudin-1). After permeabilization with 0.2% Triton X-100 in PBS for 15 min and blocking with 5% goat serum in PBS for 60 min at room temperature, cells were incubated with primary antibodies for 60 min and with Alexa 488 or 594-conjugated secondary antibodies (Molecular Probes) for 60 min at room temperature. Fluo-rescent images of the cells were acquired using a Radiance 2000 confocal laser scanning microscope (Bio-Rad).
Microinjection-MTD-1A cells were seeded onto glass coverslips and grown for 24 h. Either pCI-neo-HA-Rab3B T36N, pCI-neo-HA-Rab3B Q81L, pCI-neo-HA-Rab13 T22N, or pCI-neo-HA-Rab13 Q67L expression plasmid (0.05 mg/ml) was microinjected into the nuclei of the cells. Microinjections were performed using the Inject Man NI2 Micromanipulator and FemtoJet Microinjector Systems (Eppendorf, Hamburg, Germany) mounted on an inverted microscope TE-2000-U (Nikon, Tokyo, Japan). Nuclear injections were operated using a Z (depth) limit option, a 0.1-s injection time, and an injection pressure of 70 hectopascals. Cells were grown for an additional 18 h to allow the expression of proteins before fixation.

Occludin Is Endocytosed Continuously in Fibroblastic BHK
Cells-To begin to understand the molecular basis for the dynamic rearrangement of TJs, we first examined the transport of TJ membrane proteins to and from the cell surface in fibroblastic BHK cells (17). Although the transport of claudin-1, the original member of claudins, to the cell surface was measured easily, the detection of the cell surface transport of occludin was more difficult. If occludin was endocytosed more rapidly than transported, detection of occludin would be difficult. To test this possibility, we examined the endocytic transport of occludin using a well established biochemical assay based on cell surface biotinylation (21,26). As a control, a well characterized endocytosed and recycled protein, TfR, was examined. TfR expressed in BHK cells was not recovered on avidin beads without biotinylation (Fig. 1A). Endocytosed TfR increased in a time-dependent manner up to 15 min and stayed at a constant level until 120 min at 37°C (Fig. 1A). When the endocytosis assay was performed at 18°C, a temperature that causes the accumulation of endocytosed proteins in early/sorting endosomes (28), endocytosed TfR accumulated progressively and did not reach the maximum level within 120 min (Fig. 1A). When occludin was expressed in BHK cells and subjected to cell surface biotinylation, occludin was efficiently biotinylated and isolated on avidin beads. Importantly, no occludin was detected when biotin was omitted (Fig. 1B). If the endocytosis assay was performed at 37°C, endocytosed occludin increased linearly up to 15 min and showed a steady level until 120 min (Fig. 1B). In contrast, endocytosed occludin at 18°C accumulated progressively up to 120 min (Fig. 1B). These results indicated that occludin as well as TfR was indeed endocytosed in BHK cells.
Because claudin-1 was not endocytosed significantly when expressed in BHK cells, occludin seemed to be selected for endocytosis (17). To characterize further the endocytosis of occludin in BHK cells, we performed the endocytosis assay in K ϩ -free media, a technique that has been shown to inhibit clathrin-dependent endocytosis of low density lipoprotein receptor and other receptors (21,27,29,30). K ϩ depletion blocked the endocytosis of occludin to an extent comparable to that of TfR in BHK cells (Fig. 2, A and D), suggesting a clathrin-dependent endocytosis of occludin. Because the TJ-associated Rab family members, Rab3B and Rab13, are good candidates for regulators of occludin transport, we next assessed their function in the endocytosis of occludin. For this purpose, we generated dominant active mutants (Rab3B Q81L and Rab13 Q67L) that are defective in GTP hydrolysis as well as dominant negative mutants (Rab3B T36N and Rab13 T22N) that have a lower affinity for GTP than GDP. When these mutants were cotransfected with occludin into BHK cells, they were expressed at comparable levels and did not affect the expression level of occludin (Fig. 2B). Occludin was endocytosed in Rab3B T36N-, Rab3B Q81L-, Rab13 T22N-, or Rab13 Q67L-transfected cells as efficiently as in empty vector (mock)-transfected cells (Fig. 2, C and D). These observations indicate that occludin is endocytosed from the cell surface in a Rab3B/Rab13-independent manner in BHK cells.
Endocytosed Occludin Is Recycled Continuously Back to the Cell Surface in BHK Cells-The saturable and progressive accumulation of endocytosed occludin at 37 and 18°C, respectively, indicates recycling of endocytosed occludin back to the cell surface in BHK cells. To examine the recycling of endocytosed occludin, we performed a biochemical recycling assay, in which a decrease of biotinylated cargo molecules represents their recycling back to the cell surface. When endocytosed TfR was allowed to be recycled back to the cell surface, the amount of biotinylated TfR decreased in a time-dependent manner as expected (Fig. 3). Like TfR, endocytosed and biotinylated occludin was diminished up to 5 min (Fig. 3). This clearly demonstrates that endocytosed occludin is indeed recycled back to the cell surface in BHK cells.
Rab13 Q67L Mutant Inhibits Recycling of Endocytosed Occludin, but Not TfR, Back to the Cell Surface in BHK Cells-To investigate further the recycling of occludin in BHK cells, the recycling assay was performed in the presence of BAF, a compound that inhibits the recycling of endocytosed proteins back to the cell surface by interfering with endosomal acidification (31). Consistent with a previous report, the recycling of TfR in the presence of BAF was inhibited compared with that in control media (21) (Fig. 4A). BAF restrained the recycling of occludin to an extent comparable to that of TfR (Fig. 4, A and  C), revealing that endocytosed occludin is indeed recycled back to the cell surface in BHK cells.
Next, we examined the effect of Rab3B and Rab13 mutants on the recycling of occludin. For this purpose, the recycling of occludin was analyzed in BHK cells cotransfected with occludin and mock, Rab3B T36N, Rab3B Q81L, Rab13 T22N, or Rab13 Q67L. Occludin was recycled in Rab3B T36N-and Rab3B Q81L-transfected cells as efficiently as in mock-transfected cells, indicating that occludin is recycled in a Rab3B-independent manner (Fig. 4, B and C). In contrast, the recycling of occludin was impaired in Rab13 Q67L-transfected cells but not in Rab13 T22N-transfected cells, compared with mock-transfected cells (Fig. 4, B and C). These results demonstrate that occludin is recycled back to the cell surface via a Rab13-dependent pathway in BHK cells.
The question naturally arises as to whether the Rab13-dependent recycling pathway is specific for occludin. To address this question, we examined the endocytosis and recycling of TfR in BHK cells cotransfected with TfR and Rab13 Q67L. TfR was endocytosed in Rab13 Q67L-transfected cells as efficiently as in mock-transfected cells. In contrast to occludin, the recycling of TfR was not affected by the presence of the Rab13 Q67L mutant (Fig. 5, A and B). These results demonstrate that Rab13 directs a specific endocytic recycling pathway for occludin, but not for TfR, in BHK cells.
Intracellular Accumulation of Occludin in Epithelial MTD-1A Cells-The above results show that occludin is endocytosed from and recycled back to the cell surface when expressed in fibroblastic BHK cells. Next we investigated the transport pathways of claudins and occludin in epithelial MTD-1A cells. Both claudin-1 and occludin were localized primarily at cell-cell contacts of the PM in MTD-1A cells (Fig. 6). If claudin-1 and occludin are continuously endocytosed from and recycled back to the cell surface, the inhibition of their recycling should result in their intracellular accumulation. To test this possibility, we incubated MTD-1A cells for 4 h at 18°C. The intracellular punctate staining of occludin, but not claudin-1, was clearly evident at 18°C whereas the staining of cell-cell contacts was intact (Fig. 6). Because 18°C incubation FIG. 2. K ؉ -dependent endocytosis of occludin in BHK cells. A, BHK cells expressing TfR or occludin were subjected to K ϩ depletion and endocytosis assay as in Fig. 1A. B, BHK cells coexpressing occludin and mock, Rab3B T36N, Rab3B Q81L, Rab13 T22N, or Rab13 Q67L were subjected to immunoblot to determine the expression level of occludin and Rab. C, BHK cells coexpressing occludin and mock, Rab3B T36N, Rab3B Q81L, Rab13 T22N, or Rab13 Q67L were subjected to endocytosis assay as in A. D, the effects of K ϩ depletion, Rab3B mutants, and Rab13 mutants on endocytosis of occludin were quantitated. Endocytosed proteins were expressed as the percentage of total biotinylated proteins. The data shown in D are the means Ϯ S.E. of three independent experiments.

FIG. 3. Recycling of occludin in BHK cells. BHK cells expressing
TfR or occludin were cell surface biotinylated and incubated at 37°C for 5 min to allow endocytosis of biotinylated TfR or occludin on the cell surface. After stripping the remaining biotin from the cell surface, cells were again incubated at 37°C for the indicated periods of time to allow recycling of biotinylated TfR or occludin back to the cell surface. After the second stripping of cell surface biotin, biotinylated TfR or occludin were isolated with avidin beads, detected by immunoblot, and quantitated. Recycled proteins were expressed as the percentage of endocytosed proteins. The data shown are the means Ϯ S.E. of three independent experiments. also blocks the early biosynthetic transport of newly synthesized proteins and causes their intracellular accumulation, we pretreated MTD-1A cells with cycloheximide (CHX) to inhibit protein synthesis (32). The intracellular staining of occludin was present after CHX treatment (Fig. 6), indicating that a pool of occludin was endocytosed from the cell surface. To ascertain the recycling of occludin, we next treated MTD-1A cells with BAF. As observed with 18°C incubation, BAF treatment also caused an intracellular accumulation of occludin (Fig. 6). These results suggest that a pool of endogenous occludin, but not claudin-1, is continuously endocytosed from and recycled back to the cell surface in MTD-1A cells.
Endogenous Occludin Is Continuously Endocytosed and Recycled Back to the Cell Surface in MTD-1A Cells-To clarify FIG. 4. Rab13-dependent recycling of occludin in BHK cells. A, BHK cells expressing TfR or occludin were subjected to BAF treatment and recycling assay as in Fig. 3. After the endocytosis reaction, BAF treatment was performed by preincubating on ice for 30 min and incubating at 37°C for 5 min in the presence of 1 M BAF to allow the recycling reaction. B, BHK cells coexpressing occludin and mock, Rab3B T36N, Rab3B Q81L, Rab13 T22N, or Rab13 Q67L were subjected to recycling assay as in A. C, the effects of BAF treatment, Rab3B mutants, and Rab13 mutants on recycling of occludin were quantitated. Recycled proteins were expressed as the percentage of endocytosed proteins. The data shown in C are the means Ϯ S.E. of three independent experiments. further the morphological observations, we performed a biochemical endocytosis and recycling assay. When MTD-1A cells were subjected to cell surface biotinylation, endogenous occludin was efficiently biotinylated and isolated on avidin beads (Fig. 7A). If biotin was omitted, no occludin was detected (Fig.  7A). When the endocytosis assay was performed at 37°C, endocytosed occludin was increased rapidly for 15 min and accumulated gradually until 120 min (Fig. 7, A and B). Endogenous occludin was endocytosed progressively at 18°C, and the extent of endocytosis did not reach the maximum level within 120 min (Fig. 7, A and B). In contrast to occludin, the endocytosis of claudin-1 was not detected during the 120-min incubation at 18°C (Fig. 7, A and B). When endocytosed occludin was allowed to be recycled back to the cell surface, the biotinylated occludin decreased in a time-dependent manner (Fig. 7C). Taken together, these results show that endogenous occludin is continuously endocytosed from and recycled back to the cell surface even in MTD-1A cells as well as BHK cells. Interestingly, the kinetics of endocytosis and recycling of occludin in MTD-1A cells seems to be slower than that in BHK cells.

Rab13 Q67L Mutant Specifically Inhibits Continuous Endocytic Recycling of Occludin to the Cell Surface in MTD-1A
Cells-Because multiple pathways for the endocytic recycling of E-cadherin were reported in different cellular contexts (20 -22), we examined the role of Rab3B and Rab13 in the endocytic recycling of occludin in MTD-1A cells. For this purpose, we microinjected Rab3B T36N, Rab3B Q81L, Rab13 T22N, or Rab13 Q67L cDNA and determined the intracellular localization of endogenous occludin in MTD-1A cells. Microinjection of Rab3B T36N or Rab3B Q81L cDNA did not change the PM localization of occludin, suggesting that the endocytic recycling of occludin was independent of Rab3B (Fig. 8A). In contrast to Rab3B, Rab13 Q67L, but not Rab13 T22N, caused an intracellular accumulation of occludin (Fig. 8B). To characterize further the effect of the Rab13 Q67L mutant, we next examined the intracellular localization of endogenous claudin-1 in MTD-1A cells microinjected with Rab13 Q67L cDNA. In contrast to the continuously recycling occludin, no significant change in the localization of claudin-1 was detected in Rab13 After stripping the remaining biotin from the cell surface, cells were again incubated at 37°C for the indicated periods of time to allow recycling of biotinylated occludin back to the cell surface. After the second stripping of cell surface biotin, biotinylated occludin was isolated with avidin beads, detected by immunoblot, and quantitated. Endocytosed proteins and recycled proteins were expressed as the percentage of total biotinylated proteins and endocytosed proteins, respectively. The data shown in B and C are the means Ϯ S.E. of three independent experiments.

Rab13-dependent Recycling of Occludin
Q67L-microinjected MTD-1A cells (Fig. 8B). These results suggest that Rab13 is involved in the endocytic recycling of occludin in epithelial MTD-1A cells as well as fibroblastic BHK cells.
To gain an insight into where and when the Rab13 Q67L mutant worked on the endocytic recycling of occludin, we first compared the intracellular localization of Rab13 with other Rab family members, TfR and claudin-1 in MTD-1A cells. When HA-Rab13 was transiently transfected into MTD-1A cells, it localized to the PM and the cytosolic vesicular structures as reported previously (15). Then we performed the double immunolabelings of Rab13 with Rab3B, Rab4, Rab11, TfR, and claudin-1 in MTD-1A cells. The Rab13-positive cytosolic vesicular structures near the proximity of PM segregated from those positive for Rab3B, Rab4, Rab11, and TfR (Fig. 9A). The colocalization of Rab13 with claudin-1 was detected at the PM, but not in the cytosolic vesicular structures (Fig. 9A). These results suggest that Rab13 defines the specific membrane compartments distinct from those defined by Rab3B and known recycling endosomal markers (Rab4, Rab11, and TfR).
Next we examined whether occludin was colocalized with Rab13, Rab3B, Rab4, Rab11, and TfR in MTD-1A cells. Although the occludin-positive vesicular structures were hardly detected in control MTD-1A cells, occludin was accumulated in the cytosolic vesicular structures in MTD-1A cells transiently expressing the Rab13 Q67L mutant as shown in Fig. 8 and Fig.  9B. Double immunolabelings of occludin with Rab13, Rab3B, Rab4, Rab11, and TfR in MTD-1A cells transiently expressing the Rab13 Q67L mutant revealed that a fraction of Rab13 was colocalized with occludin in the PM and the cytosolic vesicular structures near the proximity of PM (Fig. 9B). However, the colocalization of occludin with Rab3B, Rab4, Rab11, and TfR was hardly detected in MTD-1A cells (Fig. 9B).
The above results show that the endocytic recycling of occludin is specificity controlled by Rab13 among other Rab family members. Then we investigated whether Rab13 specifically regulated the endocytic recycling of occludin among other recycling transmembrane proteins. For this purpose, we chose TfR and pIgR. A number of previous studies have established that both TfR, which recycles to basolateral PM, and pIgR, which mediates both apical-to-apical recycling and basolateralto-apical transcytosis, pass through a series of endosomal compartments and are sorted from each other. The compartments responsible for this sorting are variously termed the apical recycling endosome, common endosome, or subapical compartment in epithelial cells (9,33,34). Because we are able to detect the endogenous TfR and pIgR of MTD-1A cells by immunoblot (Fig. 10A), we analyzed the endocytosis and recycling of TfR and pIgR using the same biochemical assay as used for occludin (see Fig. 7). When MTD-1A cells were infected with Ad-EGFP or Ad-EGFP-Rab13 Q67L, they were expressed at comparable levels and did not affect the expression level of TfR, pIgR, and occludin (Fig. 10A). Although the endocytosis of occludin was not affected by the Rab13 Q67L mutant, its recycling was inhibited in Rab13 Q67L-expressed cells compared with GFPexpressed cells (Fig. 10, B and C). In contrast, the expression of the Rab13 Q67L mutant had no effect on the endocytosis and recycling of TfR and pIgR (Fig. 10, B and C). Taken together, Rab13 specifically directs the recycling of endocytosed occludin back to the cell surface both in BHK cells and MTD-1A cells. DISCUSSION Epithelial morphogenesis encompasses a variety of processes such as cell migration, wound healing, and tubulogenesis, which play a central role in animal development and tissue regeneration (35). In the tubulogenesis of the mammary gland, lung, and kidney, TJs are subjected to dynamic remodeling without complete loss of cell-cell contacts (9,10). Epithelial cells need to undergo EMT and/or MET to complete these complex and diverse events. However, the molecular mechanism by which epithelial cells rapidly remodel their polarity and TJs during EMT and/or MET, remains elusive. To begin to address this issue, we investigated the intracellular transport of the transmembrane proteins in TJs to and from the cell surface in both fibroblastic BHK and epithelial MTD-1A cells.
We have initially found that the cell surface transport of occludin was difficult to analyze in BHK cells compared with claudin-1 (17). This prompted us to investigate the intracellular transport of TJ transmembrane proteins to and from the cell surface in both fibroblastic BHK and epithelial MTD-1A cells. Our results revealed that occludin was subjected to continuous endocytic recycling in both BHK and MTD-1A cells, explaining our initial findings. Furthermore, we identified Rab13, one of the cytosolic plaque proteins in TJs, as a crucial regulator for the recycling of endocytosed occludin back to the cell surface (15).
The intracellular accumulations of claudins and occludin were reported previously under several experimental conditions including Ca 2ϩ depletion, wound healing, proinflammatory cytokine treatment, and Ras-transformed epithelial cells, in which intercellular adhesive bindings of claudins and occludin on the cell surface were disrupted (36 -39). In contrast, we identified a specific intracellular accumulation of occludin in MTD-1A cells with intact cell-cell contacts. In Ca 2ϩ -depleted T84 cells, Ivanov et al. (39) reported that occludin was endocytosed by the clathrin-dependent endocytosis and was not localized to the classical recycling endosome marked by TfR but accumulated in the storage compartment marked by syn- taxin-4. We here demonstrated that occludin was endocytosed in a clathrin-dependent manner and was recycled via the Rab13dependent pathway in contacting MTD-1A cells. It will be important to determine the endocytic recycling pathway of occludin in each cellular context.
The function of PM proteins can be regulated by altering their number expressed at the cell surface (40). One aspect of this type of regulation involves the synthesis of new proteins and/or degradation of existing proteins. Traweger et al. (24) reported that occludin was subjected to ubiquitination by the E3 ubiquitin-protein ligase itch and degradation by the proteasome. We also noticed that occludin expressed in BHK cells was subjected to post-translational degradation with a half-life of ϳ6 h. 2 Continuous recycling is another means to exert this type of control. Continuously recycled proteins have a pool of molecules available in intracellular compartments for their rapid insertion into the PM. Their proportion at the cell surface and intracellular compartments will depend upon the relative rates of endocytosis and recycling. An intracellular pool of occludin is minimal in contacting epithelial cells under physiological conditions. However, a variety of experimental conditions including 18°C incubation, BAF treatment, and Rab13 Q67L mutant expression can shift occludin from a PM pool to an intracellular pool. Because polarized recycling through endocytic compartments has been shown to be a very efficient and rapid mechanism to remobilize a number of PM proteins to the specific regions of the PM (41), the continuous recycling of occludin likely plays a role in the dynamic and rapid remodeling of epithelial polarity and TJs.
Occludin, claudins, and junctional adhesion molecules are three distinct transmembrane proteins localized to TJs. Occludin, possessing four transmembrane domains, was identified as the first constituent of TJ strands (4). Later, claudins, with four transmembrane domains, were identified as another component of TJ strands (5), and junctional adhesion molecules, bearing a single transmembrane domain, was shown to not constitute TJ strands but associate laterally with them (7,42). Although claudins are now thought to be an indispensable structural component of TJ strands, well developed TJ strands and complex histological abnormalities were observed in occlu-2 S. Morimoto, N. Nishimura, and T. Sasaki, unpublished results. A, MTD-1A cells infected with Ad-EGFP (mock) or Ad-EGFP-Rab13 Q67L were subjected to immunoblot to determine the expression level of TfR, pIgR, occludin, EGFP, and EGFP-Rab13 Q67L. B, MTD-1A cells infected with mock or Ad-GFP-Rab13 Q67L were subjected to endocytosis and recycling assays for TfR, pIgR, and occludin as in Fig. 7. C, the effects of Rab13 Q67L on endocytic recycling of TfR, pIgR, and occludin were quantitated. Endocytosed proteins and recycled proteins were expressed as the percentage of total biotinylated proteins and endocytosed proteins, respectively. The data shown in C are the means Ϯ S.E. of three independent experiments. dinϪ/Ϫ mice (43,44). Although a variety of experiments have suggested a role for occludin in controlling the formation of TJs (1), it is possible that occludin works as a more complex modulator rather than an essential structural constituent of TJs.
To our knowledge, this work provides the first direct evidence that Rab13, a cytosolic plaque protein in TJs, regulates the vesicular transport of occludin. Our results demonstrate that Rab13 Q67L, but not Rab13 T22N, inhibits the recycling of endocytosed occludin back to the cell surface. Although Marzesco et al. (16) recently showed little change in the distribution of occludin in MDCK cells stably expressing Rab13 Q67L, we observed an intracellular accumulation of occludin in MTD-1A cells transiently expressing Rab13 Q67L. Although the exact reason for these variable observations is currently unknown, the difference in cell lines may influence the pathways for endocytic recycling of occludin. The MTD-1A cell line used in this study derives from a mouse mammary tumor, in contrast to the MDCK cells used in the other study. Alternatively, the difference between stably Rab13 Q67L-expressing MDCK cells and transiently Rab13 Q67L-expressing MTD-1A cells may reflect the amount of Rab13 Q67L proteins expressed within a cell.
According to the current model in which a single Rab protein acts in conjunction with multiple effectors to determine vesicular transport specificity and organelle identity (11)(12)(13), Rab13 would interact with effector(s) that regulate the recycling of endocytosed occludin back to the PM. Although the precise site and step in which Rab13 Q67L blocks the recycling of occludin remain to be determined, the accumulation of occludin-containing vesicles and the colocalization of Rab13 with occludin near the proximity of PM suggest that the tethering/docking/fusion of occludin-containing vesicles with the PM would be under the control of Rab13 and its effector(s). Further studies are clearly needed to address these crucial issues.
It is now becoming clear that TJs are not an absolute and static barrier but a very dynamic cellular structure (1). The dynamic turnover of TJs and epithelial polarity are essential for epithelial morphogenesis, in which cells often undergo EMT and/or MET (35). In the present study, we revealed that a TJ transmembrane protein, occludin, is recycled continuously even in epithelial cells with intact cell-cell contacts. We also showed that Rab13, a TJ cytosolic plaque protein, regulated the endocytic recycling of occludin in both fibroblastic and epithelial cells. Similarly, distinct apical and basolateral transport pathways from the trans-Golgi network to the PM are identified and operated in both fibroblastic and epithelial cells (45)(46)(47). Taken together, it is likely that Rab13, expressed in both fibroblastic and epithelial cells (15), would regulate the transport of occludin during EMT and/or MET.