Occludin modulates transepithelial migration of neutrophils.

Neutrophils cross epithelial sheets to reach inflamed mucosal surfaces by migrating along the paracellular route. To avoid breakdown of the epithelial barrier, this process requires coordinated opening and closing of tight junctions, the most apical intercellular junctions in epithelia. To determine the function of epithelial tight junction proteins in this process, we analyzed neutrophil migration across monolayers formed by stably transfected epithelial cells expressing wild-type and mutant occludin, a membrane protein of tight junctions with four transmembrane domains and both termini in the cytosol. We found that expression of mutants with a modified N-terminal cytoplasmic domain up-regulated migration, whereas deletion of the C-terminal cytoplasmic domain did not have an effect. The N-terminal cytosolic domain was also found to be important for the linear arrangement of occludin within tight junctions but not for the permeability barrier. Moreover, expression of mutant occludin bearing a mutation in one of the two extracellular domains inhibited neutrophil migration. The effects of transfected occludin mutants on neutrophil migration did not correlate with their effects on selective paracellular permeability and transepithelial electrical resistance. Hence, specific domains and functional properties of occludin modulate transepithelial migration of neutrophils.

Neutrophils cross epithelial sheets to reach inflamed mucosal surfaces by migrating along the paracellular route. To avoid breakdown of the epithelial barrier, this process requires coordinated opening and closing of tight junctions, the most apical intercellular junctions in epithelia. To determine the function of epithelial tight junction proteins in this process, we analyzed neutrophil migration across monolayers formed by stably transfected epithelial cells expressing wild-type and mutant occludin, a membrane protein of tight junctions with four transmembrane domains and both termini in the cytosol. We found that expression of mutants with a modified N-terminal cytoplasmic domain up-regulated migration, whereas deletion of the C-terminal cytoplasmic domain did not have an effect. The N-terminal cytosolic domain was also found to be important for the linear arrangement of occludin within tight junctions but not for the permeability barrier. Moreover, expression of mutant occludin bearing a mutation in one of the two extracellular domains inhibited neutrophil migration. The effects of transfected occludin mutants on neutrophil migration did not correlate with their effects on selective paracellular permeability and transepithelial electrical resistance. Hence, specific domains and functional properties of occludin modulate transepithelial migration of neutrophils.
Mucosal infections result in the accumulation of large amounts of leukocytes at sites of inflammation, with neutrophils being the first phagocytic cell type that arrives. To emigrate from the blood, neutrophils need to become stimulated by chemotactic factors to adhere and cross the endothelium and to migrate across the inflamed epithelium (1)(2)(3). Neutrophils cross simple epithelia by migrating along the normally closed paracellular pathway, requiring a highly coordinated opening and closing of the epithelial intercellular junctions. Breakdown of the epithelial barrier caused by acute inflammations often results in organ failure.
The epithelial junctional complex consists of different types of cell-cell interactions, among which tight junctions (TJ) 1 are the most apical structures (4,5). TJ function as semipermeable diffusion barriers that regulate selective diffusion along the paracellular pathway and act as intramembrane fences that prevent the intermixing of apical and basolateral membrane components in the exoplasmic membrane leaflet (6,7). TJ consist of several transmembrane proteins and a cytoplasmic plaque composed of different proteins that generally function as cytoskeletal linkers and/or signal transducers (8 -10).
Although an elegant in vitro system to analyze transepithelial migration of neutrophils has been known for a long time (11), little is known about how tight junction properties affect this process. It is thought, however, that the permeability of TJ influences the efficiency of migration by regulating diffusion of the chemoattractant (12). Of all the known tight junction proteins, only the junctional adhesion molecule has been shown to modulate transmigration of neutrophils across endothelia (13). It is not known, however, whether the junctional adhesion molecule is also important for migration across epithelia.
We started to analyze the function of epithelial tight junction proteins in the transmigration of neutrophils using the previously established in vitro system in which MDCK cells are cultured on a permeable support, and migration is then measured by adding neutrophils to one and chemoattractant to the other side (11). We combined this experimental system with stably transfected MDCK cells expressing wild-type or mutant occludin, the best characterized transmembrane protein of tight junctions. Occludin possesses four transmembrane domains and two external loops (Fig. 1); the N-and the C-terminal domains are exposed to the cytosol (14). Although occludin is not required for the formation of morphologically normal tight junctions (15), it is involved in the regulation of selective paracellular permeability and in the fence function of TJ (reviewed in Ref. 16). In this study, we show that occludin is indeed important for the regulation of neutrophil migration. Our results indicate that occludin modulates transepithelial migration of neutrophils, but the effects on transmigration do not correlate with the changes in paracellular permeability and transepithelial electrical resistance (TER). Moreover, our results demonstrate for the first time a functional role of the N-terminal cytosolic domain of occludin.

EXPERIMENTAL PROCEDURES
Cell Culture and Cell Lines-MDCK cells expressing occludin, HAoccludin, HAoccludinCT3, and OccL1D were previously generated and characterized; clones exhibiting average phenotypes in terms of TER and paracellular permeability were used (17,34). OccludinCT3 was generated by converting the codon for serine 253 to a stop codon as described for HAoccludinCT3 (17) using polymerase chain reactionbased mutagenesis and previously described cDNA coding for chicken occludin as a template (see Fig. 1 for a schematic representation of the constructed mutants and chimeras). The OccludinCT3 cDNA was then stably expressed in MDCK cells as previously (17). For experiments testing the function of occludinCT3 on tight junction functions, the cells were plated on 12-well Transwell culture inserts with 0.4-m pores (Costar Corp., Cambridge, MA) and, when indicated, were pretreated overnight with 7 mM sodium butyrate. For migration experiments, the cells were cultured on inserts with 3-m pores and were never pretreated with sodium butyrate to avoid secondary effects on the neutrophils. Cells grown on inserts with 0.4-and 3-m pores exhibited comparable TER and paracellular permeability values, but the establishment of monolayers with stable TER values took twice as long on the large pore as on the small pore filters (Ͼ8 days).
Antibodies-An antibody specific for the NH 2 -terminal domain of chicken occludin was raised against a histidine-tagged fusion protein containing the entire NH 2 -terminal domain and was generated with pTricHis2 (Invitrogen Corp., San Diego, CA). The purified fusion protein was emulsified with Specol (Central Veterinary Institute, Lelystad, The Netherlands) and subcutaneously injected into rabbits. The anti-COOH-terminal domain antibody were previously described and recognizes chicken and dog occludin (antibody A) (17). The mouse monoclonal antibody against the hemagglutin (HA) epitope was kindly provided by Drs. P. van der Sluijs (University of Utrecht, The Netherlands) and I. Mellman (Yale University, New Haven, CT) (18). Drs. J. M. Anderson and M. S. Mooseker (Yale University, New Haven, CT) kindly supplied rat monoclonal antibody R40.76 specific for ZO-1 (19).
Immunoblots, Transepithelial Electrical Resistance, and Paracellular Permeability-Total cell extracts were separated on SDS-polyacrylamide gel electrophoresis gradient gels, transferred to nitrocellulose, and probed with primary antibodies, horseradish peroxidase-conjugated secondary antibodies, and ECL (17). Transepithelial electrical resistance and paracellular flux of horseradish peroxidase and [ 3 H]mannitol were measured as described previously (17). To measure paracellular permeability during migration, [ 3 H]mannitol was added together with the neutrophils to the apical chamber. At the end of the migration experiment, apical and basolateral media were collected, and the radioactivity was determined by liquid scintillation counting.
Transepithelial Migration Assay-Human buffy coats were obtained from healthy volunteers from the Hospital of the University of Geneva. Neutrophils were purified by centrifugation through a Ficoll cushion followed by hypotonic lysis of erythrocytes (20). Purified neutrophils were resuspended in Dulbecco's modified Eagle's medium at a concentration of 2 ϫ 10 6 cells/ml.
For the migration assays, MDCK cells were grown for 10 days on 12-well Transwell culture inserts with pores of 3-m diameter (Costar Corp.). This long culture time was necessary for the cells to establish monolayers exhibiting stable values of TER and paracellular flux on the large pore filters. The morphology of all cell lines used in this study was controlled by electron microscopy and thin sectioning of Epon-embedded monolayers grown on large pore filters: no morphological differences between wild-type and transfected cell lines were detected (see below).
Prior to the migration assays, the monolayers were washed once with tissue culture medium and then 10 6 neutrophils in 500 l of the same medium were added to the apical chamber. Migration was induced by adding the chemoattractant N-formyl-Met-Leu-Phe (10 Ϫ7 M) to the lower chamber (11). Control incubations were done by adding normal medium to the apical and chemoattractant to the basolateral chamber. After incubation for 1 h at 37°C, the plates were cooled on ice, and the medium of the lower chamber was centrifuged at 1000 ϫ g for 10 min to pellet the neutrophils that migrated across the epithelial sheet. The neutrophils were washed and resuspended in 100 l of cold phosphatebuffered saline for counting in a Bright-Line hemacytometer (Reichert-Jung).
Immunofluorescence and Microscopy-To stain chicken occludin with the anti-NH 2 -terminal domain antibody, cells were cooled on ice; extracted for 3 min with 0.2% Triton X-100 in 100 mM KCl, 3 mM MgCl 2 , 1 mM CaCl 2 , 200 mM sucrose, and 10 mM Hepes (pH 7.1) at 4°C; and then fixed for 5 min in methanol at Ϫ20°C. To label with the anti-Cterminal domain and the anti-HA antibody, filter-grown monolayers were permeabilized for 2 min on ice with the same extraction buffer as above but were then fixed for 30 min with 95% ethanol on ice (17). The fixed cells were blocked and incubated with antibodies as described (17).
To test the morphology of the monolayers after migration assays, the cells were transferred to ice at the end of the incubation, washed with cold phosphate-buffered saline, fixed for 20 min with 3% paraformaldehyde in phosphate-buffered saline, and permeabilized with 0.3% Triton X-100 in phosphate-buffered saline containing 0.3% bovine serum albumin for 3 min. ZO-1 was labeled with rat monoclonal antibody R40.76 (19) and a fluorescein isothiocyanate-conjugated donkey antirat antibody (Jackson Immunoresearch); F-actin was visualized with tetramethylrhodamine isothiocyanate (TRITC)-conjugated phalloidin (Sigma). For electron microscopy, the cells were fixed, embedded in Epon, and processed for transmission electron microscopy as described (17).
The samples were mounted with the ProLong anti-fade kit (Molecular Probes, Inc., Eugene, OR) and analyzed with a confocal laser scanning microscope (LSM 410 invert; Carl Zeiss, Inc.) equipped with an argon and a helium-neon laser for excitation at 488 and 543 nm and BP510 -525 and LP590 emission filters.

Modification of the N-terminal Cytoplasmic Domain of Occludin Results in Increase
Transmigration-In MDCK cells, stable transfection of chicken occludin and full-length occludin with an N-terminal HA epitope (Fig. 1, HAoccludin) results in increased TER, a measure for the general tightness of the monolayer, and small, if any, increases in selective paracellular permeability (17,21). To test whether occludin is involved in the transmigration of neutrophils, we analyzed the previously described stably transfected MDCK cell lines expressing wildtype chicken occludin or HAoccludin with the migration assay established by Cramer et al. (11).
Wild-type and transfected MDCK cells were cultured on permeable supports with 3-m pores for 10 days to allow the establishment of monolayers with stable TER and paracellular flux values. As in our previous experiments with cells grown on filters with 0.4-m pores (17), monolayers formed by cells expressing chicken occludin or HAoccludin exhibited 2-3-fold increases in TER relative to wild-type MDCK cells, for which were normalized to wild-type MDCK cells because the efficiency of migration varied in different neutrophil preparations. Fig. 2 shows that transfection of chicken occludin did not significantly affect the efficiency of neutrophil transmigration. Because transfected chicken occludin is efficiently targeted to tight junctions (17), this indicates that increased amounts of occludin in tight junctions did not interfere with the migration process. In contrast, when cells expressing HAoccludin were analyzed, a 4-fold increase was found. Modification of the N terminus of occludin thus leads to increased transmigration without affecting TER and paracellular permeability, suggesting that the N-terminal cytosolic domain is involved in the reversible opening and closing of tight junctions during neutrophil migration.
To test whether the increased rates of transmigration affected the integrity of the monolayers, we first compared the TER before and immediately after migration. Table I shows that the TER of monolayers formed by wild-type and transfected MDCK cells was not affected by the neutrophils. We next added [ 3 H]mannitol together with the neutrophils and compared paracellular flux occurring during transmigration with parallel cultures incubated with only [ 3 H]mannitol and chemoattractant, but without neutrophils. Although this resulted in slightly increased ratios, the differences were not significant (p Ͼ 0.1).
To test whether transmigration of neutrophils affected the morphology of the monolayers, we fixed the cells immediately after the migration assay and then labeled the monolayers by indirect immunofluorescence for ZO-1, a submembrane protein of tight junctions that directly interacts with occludin (22)(23)(24), and for F-actin with fluorescent phalloidin. The confocal sections in Fig. 3 show that the distribution of ZO-1 was not different in cells that were incubated with neutrophils compared with cells that were incubated with chemoattractant only. Serial sectioning of the samples did also not reveal any significant differences (not shown). Similarly, effects on the distribution of F-actin could not be detected in either single (Fig. 3) or serial (not shown) sections. Staining of the samples for occludin did also not reveal any differences caused by the neutrophils (not shown).
We next analyzed the morphology of the monolayers and of the junctional region by thin sectioning of Epon-embedded cells and electron microscopy. Because migration was analyzed in the apical to basolateral direction, we were looking for monolayer regions that still contained a trapped neutrophil in the intermembrane space and took images from the junctions connecting the two corresponding cells to ensure that we were looking at junctions between cells that had allowed the passage of neutrophils. Fig. 4 shows examples of junctional regions of wild-type (A without and B with neutrophils) and HAoccludinexpressing cells (C without and D with neutrophils). Fig. 4, E and F, shows examples of neutrophils in the intermembrane space of wild-type (E) and HAoccludin-expressing (F) cells. We could not detect morphological effects caused by neutrophil transmigration on the junctions even when HAoccludin-expressing cells were analyzed that allowed 4 times more neutrophils to migrate than wild-type cells. Thus, modification of the N-terminal cytosolic domain of occludin resulted in increased transmigration of neutrophils without detectably af-

H]mannitol of monolayers formed by wild-type and transfected MDCK cells TER was determined before and after the migration assay and paracellular flux of [ 3 H]mannitol was measured in parallel cultures that
were or were not incubated with neutrophils. For TER, the numbers represent the ratios obtained by dividing the TER values after by those measured before migration of neutrophils; n represents the number of filters analyzed. For paracellular flux, the numbers represent the ratios obtained by dividing values obtained from cultures with neutrophils by those obtained from cultures without neutrophils; n represents the number of independent experiments. fecting the morphology of the junctional regions.
Deletion of the C-terminal Cytoplasmic Domain Does Not Affect Transmigration-To be able to analyze the importance of the C-terminal cytoplasmic domain for transmigration, we constructed a mutant lacking this domain (Fig. 1, occludinCT3) and stably expressed it in MDCK cells. The immunoblot in Fig.  5 shows that a protein with a corresponding molecular weight could be expressed and was recognized by an antibody raised against the N-terminal domain of chicken occludin.
To determine the effects of occludinCT3 on TER and paracellular permeability, we plated wild-type MDCK cells, cells expressing occludinCT3, and control transfectants on filters. Fig. 6A shows that expression of occludinCT3 resulted in monolayers that exhibited twice the transepithelial electrical resistance of those formed by wild-type and control MDCK. When the cells were incubated overnight with sodium butyrate to induce higher expression levels (Fig. 5), only small, if any, further increases in transepithelial electrical resistance could be detected (Fig. 6A). We next measured paracellular flux of [ 3 H]mannitol to determine possible effects on selective paracellular permeability. Clones expressing occludinCT3 exhibited increased paracellular permeability of this low molecular weight tracer (Fig. 6B). This effect was even greater when higher expression levels were induced by preincubation with sodium butyrate. Because the expression of wild-type and mutant occludin does not affect transcellular fluid phase transport (17), this increase must be due to increased paracellular permeability. Because we had observed the same behavior of TER and paracellular permeability in cells expressing HAoc-cludinCT3 (17), this indicates that modification of the N-terminal domain of occludin does not affect the properties of the paracellular diffusion barrier.
We then performed the migration assays with cell lines expressing occludinCT3 and HAoccludinCT3. If the transfected cells were grown on filters with 3-m pores, we found that expression of occludinCT3 and HAoccludinCT3 induced 2-3fold increases in TER and 3-4-fold increases in paracellular flux relative to wild-type cells as shown in Fig. 6 for cultures grown on small pore filters. As in case of cells transfected with cDNAs coding for full-length occludin, neutrophil migration did not cause alterations in TER or paracellular flux (Table I), nor in the morphological appearance of the monolayers (not shown). Fig. 2 shows that expression of occludinCT3 did not affect the efficiency of neutrophil migration, suggesting that deletion of the C-terminal cytosolic domain does not affect transmigration even though it results in increased paracellular flux. Expression of the analogous mutant with the N-terminal modification, HAoccludinCT3, resulted in increased migration. Although the stimulation was a little smaller than the one by HAoccludin, this indicates that deletion of the C-terminal cytosolic domain Proteins of total cell extracts derived from wild-type and stably transfected MDCK cells, which were preincubated without or with sodium butyrate, were separated by SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose. Expression was then tested by incubating the blots with an antibody against the NH 2 -terminal domain of chicken occludin. Bound primary antibodies were then detected with a horseradish peroxidase-conjugated secondary antibody and enhanced chemiluminescence. Note that all analyzed clones expressed the transfected protein without preincubation with sodium butyrate but that the pretreatment resulted in increased expression levels. is not required for the increase in transmigration.
A Dominant Negative Mutant of Occludin Inhibits Transmigration-To test a role of the extracellular loops of occludin in transmigration, we next analyzed a cell line expressing an occludin mutant with a deletion in the first extracellular loop, OccL1D, that causes decreased paracellular permeability. 2 Also, if grown on 3-m-pore filters, expression of this mutant resulted in a reduction of paracellular flux of [ 3 H ]mannitol by about 60%. TER and paracellular flux of cells expressing OccL1D were not affected by the neutrophils (Table I). Fig. 1 shows that expression of OccL1D reduced the number of transmigrating neutrophils to 40%, suggesting that extracellular portions of occludin, also, have a function in neutrophil transmigration. Although this decrease could be due to the decreased paracellular permeability, it could be that the extracellular regions of occludin are more directly involved in the reversible opening of tight junctions. NH 2 -and COOH-terminal Cytoplasmic Domains Are Involved in the Continuous Junctional Arrangement of Occludin-The finding that modification of the N-terminal cytoplasmic domain modified transmigration was surprising, because thus far, all known cytoplasmic interactions and functions of occludin have been mapped to the C-terminal cytosolic domain (for a review, see Ref. 16). Because HAoccludinCT3 is discontinuously distributed along the junction, one could use oc-cludinCT3-expressing cells to test whether the N-terminal cytosolic domain is sufficient to mediate a continuous junctional distribution. To do this, we tested the distribution of oc-cludinCT3 by immunofluorescence using the antibody raised against the N-terminal cytosolic domain of chicken occludin. This antibody does not cross-react with endogenous dog occludin. Fig. 7 shows that occludinCT3 (A1) formed a continuous ring around the cells co-localizing with ZO-1 (A2). In contrast, the distribution of HAoccludinCT3 was also discontinuous when the protein was visualized with this new antibody directed against the NH 2 -terminal cytoplasmic domain (Fig. 7B). Both labelings also revealed intracellular fluorescence. For HAoc-cludinCT3, we have previously shown that most of the intracellular fluorescence is due to the pool of mutant occludin that accumulates in the Golgi complex (Ref. 25; the structure of the Golgi-associated fluorescence was lost because of the harsh permeabilization conditions necessary for efficient labeling of occludin in tight junctions). These results are in agreement with the finding that COOH-terminally truncated occludin tagged at the truncated COOH terminus is continuously distributed along tight junctions in Xenopus embryos (26).
Similarly to transfected occludin, endogenous occludin exhibited a normal continuous junctional pattern in cells expressing occludinCT3 (Fig. 7C), as it does in wild-type MDCK cells (Fig. 7D), and it did not form patches, as in cells expressing HAoccludinCT3 (Fig. 7E). The expression levels of endogenous occludin were not affected by the expression of the two COOHterminally truncated mutants (Ref. 17 and data not shown). Thus, both terminal cytoplasmic domains of transfected occludin have to be inactivated to induce a discontinuous distribution of endogenous and transfected occludin. DISCUSSION Our experiments indicate that occludin is a regulator of the transmigration of neutrophils across epithelial sheets. The Nterminal cytosolic domain of occludin, which is not critical for TER and selective paracellular permeability, is important for this function. Thus, not only the C-terminal but also the Nterminal cytoplasmic domain of occludin is of functional relevance.
Modification of the N-terminal cytosolic domain of transfected occludin results in two phenotypes: an increased efficiency of transmigration and, if the C-terminal domain is also inactivated, a discontinuous distribution of transfected and endogenous occludin. The finding that the N-terminal cytosolic domain is sufficient to mediate a continuous distribution of occludin suggests that this domain also interacts with the submembrane cytoskeleton, similar to the C-terminal domain. Although several components of the submembrane plaque of tight junctions are known, the protein that binds to the Nterminal domain of occludin has not yet been identified.
The C-terminal cytosolic domain of occludin interacts with at least four proteins: the three peripheral membrane proteins ZO-1, ZO-2, and ZO-3 (23,24,27,28) and the membrane protein VAP-33 (29). Because the ZO-1 complex also interacts with the actin-cytoskeleton, this protein may in fact serve as an anchor and therefore be important for the continuous distribution of occludin (24,30). Although occludin is continuously distributed in junctions lacking detectable ZO-1 (31), this might be due to the presence of additional interactions mediated by the N-terminal domain, as we show here in cells expressing occludinCT3.
Interestingly, expression of mutants lacking the C-terminal cytosolic domain did not affect transmigration, suggesting either that this domain and its interactions with the submembrane cytoskeleton are not required for transmigration or, alternatively, that the connections provided by endogenous occludin are sufficient. Because removal of the C-terminal domain results in increased paracellular permeability, this finding also suggests that the amount of chemoattractant that crosses wild-type MDCK cells is sufficient to provide maximal stimulation. In contrast, reduction of paracellular flux by expression of OccL1D (a mutant that inhibits paracellular flux by more than 50%) (34) significantly reduced the number of transmigrating neutrophils. Nevertheless, it cannot yet be excluded that this inhibition is due to a function of this extracellular loop during the transmigration process.
Occludin is important for all functions of tight junctions tested thus far: the formation of a semipermeable paracellular diffusion barrier (17,21,26,32,33), the restriction of intramembrane diffusion between the apical and basolateral cell FIG. 7. Subcellular distribution of occludinCT3. In A1, A2, and B, cells expressing occludinCT3 (A1 and A2) or HAoccludinCT3 (B) were plated on coverslips for 2 days and were then fixed with the Triton X-100/methanol procedure. The cells were subsequently labeled with the anti-NH 2 -terminal domain antibody to detect transfected occludin (A1 and B) and a monoclonal anti-ZO-1 antibody (A2). In C-E, oc-cludinCT3-expressing cells (C), wild-type MDCK cells (D), or HAoc-cludinCT3-expressing cells (E) were cultured on filters for 1 week and were then fixed with the ethanol/acetone procedure. The samples were labeled with the anti-occludin COOH-terminal domain antibody that cross-reacts with dog occludin (antibody A) to visualize the distribution of endogenous occludin. Shown are confocal sections through the junctional area of the monolayers. Note the continuous distribution of transfected and endogenous occludin in cells expressing occludinCT3 (A and C) and the discontinuous distribution in cell expressing HAoc-cludinCT3 (B and E). In mature monolayers grown on filters, the anti NH 2 -terminal domain antibody failed to stain, suggesting that the domain becomes engaged in interactions that block antibody binding. If endogenous occludin was visualized in cells grown on coverslips, the same distribution was observed as in cells grown on filters (not shown). Bars, 10 m.
surface domains (17), and, as we show now, the regulation of neutrophil transmigration. Nevertheless, different structural domains of occludin are important for different functions. The C-terminal cytoplasmic domain modulates paracellular diffusion but does not appear to play a role in neutrophil transmigration. In contrast, the N-terminal cytosolic domain regulates transmigration but not paracellular permeability. On the other hand, modification of both cytoplasmic domains is required for a discontinuous distribution of occludin and disruption of the intramembrane diffusion barrier because only expression of HAoccludinCT3 (17), but not occludinCT3, 2 results in a breakdown of the intramembrane diffusion barrier.
To summarize, this study identifies occludin as a regulator of neutrophil transmigration across epithelia. Specific domains and properties of occludin are involved in the regulation of transmigration. Although the precise role of occludin in the mechanism that allows the reversible opening of tight junctions is not clear, the effects of transfected wild-type and mutant occludin on transmigration do not correlate with their effects on TER and paracellular permeability. Our results also indicate for the first time that the N-terminal cytosolic domain of occludin is of functional relevance.