HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 30, 27501-27509, July 26, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/30/27501    most recent M201177200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map 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 Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Roh, M. H. Articles by Margolis, B. Search for Related Content PubMed PubMed Citation Articles by Roh, M. H. Articles by Margolis, B. Social Bookmarking                      What's this?

The Carboxyl Terminus of Zona Occludens-3 Binds and Recruits a Mammalian Homologue of Discs Lost to Tight Junctions^*

Michael H. Roh^§, Chia-Jen Liu^¶, Stephanie Laurinec, and Ben Margolis^¶**

From the  Howard Hughes Medical Institute, Departments of  Biological Chemistry and ^¶ Internal Medicine, University of Michigan Medical School, Ann Arbor, Michigan 48109

Received for publication, February 5, 2002, and in revised form, April 29, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Mammalian homologues of the Drosophila polarity proteins Stardust, Discs Lost, and Crumbs have been identified as Pals1, Pals1-associated tight junction protein (PATJ), and human Crumbs homologue 1 (CRB1), respectively. We have previously demonstrated that PATJ, Pals1, and CRB1 can form a tripartite tight junction complex in epithelial cells and that PATJ recruits Pals1 to tight junctions. Here, we observed that the Pals1/PATJ interaction was not crucial for the ultimate targeting of PATJ itself to tight junctions. This prompted us to examine if any of the 10 post-synaptic density-95/Discs Large/zona occludens-1 (PDZ) domains of PATJ could bind to the carboxyl termini of known tight junction constituents. We found that the 6^th and 8^th PDZ domains of PATJ can interact with the carboxyl termini of zona occludens-3 (ZO-3) and claudin 1, respectively. PATJ missing the 6^th PDZ domain was found to mislocalize away from cell contacts. Surprisingly, deleting the 8^th PDZ domain had little effect on PATJ localization. Finally, reciprocal co-immunoprecipitation experiments revealed that full-length ZO-3 can associate with PATJ. Hence, the PATJ/ZO-3 interaction is likely important for recruiting PATJ and its associated proteins to tight junctions.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Polarized epithelia primarily serve to separate and regulate the movement of ions and metabolites between physiologically distinct compartments (1, 2). Substances are transported across epithelia mainly via two mechanisms: the paracellular route, which allow molecules to cross through spaces existing between epithelial cells, and the transcellular route, which involve vectorial transport of molecules through epithelial cells. The latter pathway relies on the demarcation of the apical and basolateral plasma membranes that are composed of distinct proteins and lipids. Tight junctions (TJs)¹ are essential for both pathways in that they serve as selective paracellular barriers ("gates") and "fences" preventing the intermixing of apical and basolateral membrane constituents (3, 4).

Upon visualization by freeze-fracture electron microscopy, TJs simply appear as a network of continuous, anastomosing fibrils at the apical aspect of cell-cell contacts (5). However, the number of proteins known to compose TJs has grown considerably making it clear that TJs have a complex molecular architecture (6-8). The first integral membrane protein found to compose TJs was occludin (9) and it was initially thought that this protein was responsible for generating TJ fibrils. However, the claudin family of proteins was shown to be directly responsible for forming TJ strands (10). Junctional adhesion molecule (JAM) is another integral membrane protein that has been shown to co-localize with occludin and claudins to TJs (11, 12).

Although only three distinct classes of integral membrane proteins are known to exist at TJs, many peripheral membrane TJ proteins have been identified (3, 8, 13). Some of these interact directly with the cytoplasmic carboxyl-terminal tails of occludin, claudins, and/or JAM. For instance, three members of the membrane-associated guanylate kinase (MAGUK) family, ZO-1, ZO-2, and ZO-3, are known to bind the last 150 residues of occludin (14-16). They also interact via their first PSD-95/Discs Large/ZO-1 (PDZ) domain with the extreme carboxyl termini of claudins, most of which end in YV (17). These MAGUK proteins also interact with the cortical actin cytoskeleton allowing for the tethering of the TJ fibrils to F-actin (7). Interestingly, although the extreme carboxyl terminus of ZO-1 does not conform to any known PDZ binding consensus sequence, ZO-2 and ZO-3 terminate in TEL and TDL, respectively. These carboxyl termini fit the class I PDZ ligand consensus sequence, (S/T)X(V/I/L)-COOH, where X represents any amino acid (18). To date, there are no reports describing the PDZ domain proteins that interact with the carboxyl termini of ZO-2 and ZO-3.

Recently, two TJ complexes involved in the establishment and/or maintenance of epithelial cell polarity have been described (19). First, the Par-6·ASIP·aPKC complex was shown to be important for the formation of TJs (20-23). This complex is conserved in Drosophila (DPar-6/Bazooka/DaPKC) and Caenorhabditis elegans (Par-6/Par-3/aPKC). The other polarity complex consists of the mammalian homologues of Drosophila Stardust, Discs Lost, and Crumbs: Pals1, Pals1-associated tight junction protein (PATJ), and human Crumbs homologue 1 (CRB1), respectively. Fly embryos missing Discs Lost, Stardust, or Crumbs fail to form polarized epithelia during development and fail to progress to adulthood (24, 25). We recently reported that PATJ, Pals1, and CRB1 can form a ternary complex at TJs (26).

PATJ was isolated based on its ability to bind and recruit the MAGUK protein, Pals1, to TJs. PATJ consists of an amino-terminal MAGUK recruitment (MRE) domain followed by 10 PDZ domains. The MRE domain of PATJ interacts specifically with Pals1 and is directly responsible for targeting Pals1 to TJs (26). Nonetheless, it was unclear as to the exact domain(s) that contributes to the proper targeting of PATJ itself to TJs. PATJ is a paralogue of multiple PDZ protein 1 (MUPP1), a protein containing a single MRE domain and 13 PDZ domains. Like PATJ, MUPP1 was reported to localize to TJs based on its association with the carboxyl termini of JAM and claudins (27). Here, we investigate the PDZ domain-based protein interactions between PATJ and known TJ constituents. Furthermore, we extend this analysis to identifying the domains that are responsible for targeting PATJ to TJs.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

DNA Constructs-- The cloning of full-length PATJ into the pcDNA3.1()/HA and pEYFP-N1 vectors has been described previously (26). The PATJ open reading frame was amplified by PCR and inserted into the pRK5-Myc vector. The cDNA encoding for PATJ missing its last two or five PDZ domains was also generated by PCR and cloned into both pcDNA3.1()/HA and pEYFP-N1. The first 125 residues (encompassing the MRE domain) were deleted from the pEYFP-N1-PATJ construct using a previously described method (28). Internal PDZ domain deletions were generated in the pRK5-Myc-PATJ and pEYFP-N1-PATJ constructs in a similar fashion. Specifically, we deleted residues 1065-1156 (PATJPDZ6), 1239-1315 (PATJPDZ7), and 1432-1522 (PATJPDZ8). The DNA encoding the regions of PATJ encompassing the 6^th through 7^th and the 8^th through 10^th PDZ domains (PDZ6-PDZ7 and PDZ8-PDZ10, respectively) were generated by PCR and inserted into pRK5-Myc. The EYFP-PATJ mutant constructs were used to study the targeting of PATJ mutants in MDCK cells. The full-length and mutant HA-PATJ and Myc-PATJ constructs were used to transiently express the various PATJ proteins in HEK293 cells for use in in vitro binding studies.

The pGSTag and pET15b vectors were used to express GST- and His₆-tagged fusion proteins in BL21 bacterial cells, respectively. Constructs encoding for GST fused to the last 20 amino acids of human paracellin-1/claudin-16 (PCLN-1), ZO-2, ZO-3, JAM, and mouse claudin 1 were made using a previously described method (29). DNA encoding the first 181 amino acids of Pals1 was generated by PCR and cloned into the pGSTag vector. The DNA encoding residues 1-238 of PATJ was cloned into both pET15b and pGSTag.

Finally, the carboxyl-terminal vesicular stomatitis viral G protein (VSVG)-tagged ZO-3 cDNA was subcloned from the pBK-CMV-ZO3 plasmid (a gift from Erika Wittchen and Bruce Stevenson) into pRK5-Myc using BamHI and XbaI sites. To remove the VSVG epitope tag, the pBK-CMV-ZO3 plasmid was digested with XhoI and XbaI. This removes ~100 bp at the 3' end of the ZO3-VSVG plasmid insert. Two complementary oligos, 5'-TCGAGATGTGGAGTCCTCCGATGA GGACGGCTATGACTGGGGACCAGCAACAGACCTGTAAT-3' and 5'-CTAGATTACAGGTCTGTTGCTGGTCCCCAGTCATAGCCGTCCTCATCGGAGGACTCCACATC-3' were annealed and subsequently ligated into the digested pBK-CMV-ZO3 plasmid. The non-tagged ZO-3 cDNA in the resulting construct was also subcloned into pRK5-Myc using BamHI and XbaI restriction sites. The sequences of the inserts in all plasmid constructs were verified by automated sequencing at the University of Michigan DNA Sequencing Core.

Antibodies-- The affinity purification of rabbit polyclonal antibodies directed against Pals1 has been described previously (26). These antibodies were used at 1:500 dilution for immunoblotting experiments. For generation of anti-PATJ antisera, rabbits were injected with a GST-PATJ-(1-238) fusion protein. Anti-PATJ antibodies were affinity purified using an Affi-Gel 10 column (Bio-Rad) previously coupled to His₆-tagged PATJ-(1-238).

The anti-Myc mouse monoclonal antibody (9E10) was used at 1:1,000 dilution for immunoblotting experiments. The mouse anti-hemagglutinin (HA) monoclonal antibody (clone 12CA5), obtained from Roche Molecular Biochemicals, was used for immunoblotting. For immunoprecipitation and immunoblot detection of EYFP fusion proteins, rabbit polyclonal and mouse monoclonal anti-EYFP antibodies (CLONTECH) were used, respectively. Mouse monoclonal anti-ZO1 antibodies used for immunostaining were obtained from Zymed Laboratories Inc. Mouse monoclonal anti-ZO-3 antibodies (Chemicon) were used in immunostaining and immunoprecipitation experiments. Finally, secondary antibodies conjugated to Alexa Fluor 488, 594, and 647 were purchased from Molecular Probes, Inc. (Eugene, OR).

Cell Culture and Transfection-- MDCK and HEK293 cells were grown and propagated in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 100 units/ml penicillin, 100 µg/ml streptomycin, 2 mM L-glutamine, and 10% fetal bovine serum. FuGENE 6 was used according to the manufacturer's instructions to transfect MDCK and HEK293 cells. 48 h after transfection, HEK293 cells were harvested in Triton lysis buffer (50 mM HEPES, pH 7.5, 10% glycerol, 150 mM NaCl, 1% Triton X-100, 1.5 mM MgCl₂, and 1 mM EGTA) supplemented with 1 mM phenylmethylsulfonyl fluoride, 10 µg/ml aprotinin, and 10 µg/ml leupeptin. MDCK cells were split into and grown in selection media containing 600 µg/ml G418.

MDCK clones expressing the various EYFP-PATJ fusion proteins were isolated and propagated in medium containing 300 µg/ml G418. For immunostaining experiments, wild type MDCK cells and stable MDCK clones were seeded at high confluency onto Transwell membrane filters (0.4-µm pore size; Corning Costar Corp.) and grown for 3-5 days.

Immunostaining and Confocal Microscopy-- Immunostaining was performed as described previously (26, 29). Briefly, MDCK cells grown on filters were fixed in 4% formaldehyde/PBS for 15-20 min, permeabilized with 0.1% Triton X-100/PBS, and blocked with 2% goat serum/PBS. Filters were then excised from their plastic casing and incubated with primary antibodies in 2% goat serum/PBS (rabbit anti-PATJ at 1:100, mouse anti-ZO-1 at 1:1000, and mouse anti-ZO-3 at 1:200) overnight at 30 °C in a humidified chamber. After extensive washes with 2% goat serum/PBS, filters were incubated with fluorochrome-conjugated secondary antibody (at 1:1,000 dilution in 2% goat serum/PBS) for 1-2 h at 30 °C. Subsequently, filters were washed several times with 2% goat serum/PBS and mounted on glass slides with ProLong antifade reagent (Molecular Probes, Inc.). Immunofluorescence microscopy was performed at the University of Michigan Morphology and Image Analysis Laboratory with a Zeiss LSM510 Axiovert 100M inverted confocal microscope (Carl Zeiss, Inc., Thornwood, NY).

In Vitro Binding Experiments, Gel Electrophoresis, and Immunoblotting-- GST pull-down and immunoprecipitation experiments were performed as described previously (30). Briefly 5 µg of GST fusion proteins or 5 µg of antibody (immobilized on glutathione-agarose or Protein A-Sepharose beads, respectively) were incubated overnight at 4 °C with 293 or MDCK cell lysates. Precipitated proteins were washed 3 times with cold PBS supplemented with 10% glycerol, eluted with sample buffer, and resolved on 4-12% gradient bis-Tris gels using the NuPAGE electrophoresis system (Invitrogen). Proteins were electrophoretically transferred from the gels onto nitrocellulose membranes. Membranes were blocked in 5% bovine serum albumin/TBS for 30 min and then incubated with primary antibody in 5% bovine serum albumin/TBS for 2 h at room temperature. After extensive washing with 0.1% Triton X-100/TBS, membranes were soaked in TBS supplemented with 5% skim milk and secondary antibody conjugated to horseradish peroxidase. Incubation with secondary antibody was performed for 1 h at room temperature or 4 h at 4 °C. Membranes were washed with 0.1% Triton X-100/TBS and bands were visualized using ECL reagent (Amersham Biosciences).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Deletion of the MRE Domain of PATJ Does Not Completely Abolish Targeting of PATJ to TJs-- Recently, we isolated and identified a novel multiple PDZ domain protein, PATJ, as a binding partner for Pals1 in MDCK cells (26). PATJ contains 10 PDZ domains in tandem (Fig. 1A) and is a paralogue of MUPP1 that contains 13 PDZ domains. PATJ is the orthologue of Discs Lost, a Drosophila polarity protein containing 4 PDZ domains (25). PATJ and MUPP1 contain single amino-terminal MRE domains that are capable of binding to the L27N domain of Pals1. Discs Lost also contains an MRE domain that binds to the single L27 domain of the Pals1 orthologue, Stardust (26). MUPP1 was previously shown to localize to TJs in MDCK cells (27). Likewise, we also demonstrated by confocal microscopy that an EYFP-PATJ fusion protein expressed in MDCK cells targets to TJs (26).

View larger version (38K): [in this window] [in a new window]   Fig. 1.   Endogenous PATJ localizes to TJs in polarized MDCK epithelial cells. A, displayed is a schematic diagram of PATJ consisting of an amino-terminal MAGUK recruitment (MRE) domain followed by 10 PDZ domains. The bracket denotes the region of PATJ used to immunize rabbits for the production of anti-PATJ antisera. B, MDCK cells were immunostained with affinity purified rabbit anti-PATJ and mouse monoclonal anti-ZO-1 antibodies as described under "Experimental Procedures." The square and rectangular panels represent the X-Y dimension of the Z-series and the X-Z sections (Z-sections), respectively.

We first sought to verify the subcellular localization of endogenous PATJ in polarized MDCK cells. Rabbit polyclonal antibodies were raised against GST fused to the first 238 amino acids of PATJ. Previously, we showed that this antibody could recognize endogenous PATJ in MDCK cell lysates (26). Here, we performed immunostaining experiments using this antibody on fixed, permeabilized MDCK cells. Confocal microscopy revealed that endogenous PATJ localizes to TJs (Fig. 1B).

The MRE domain of PATJ is responsible for targeting Pals1 to the TJ. We next wondered if the association between PATJ and Pals1 was important for localizing PATJ itself to TJs. To address this issue, we expressed an EYFP-PATJ mutant protein missing the MRE domain (EYFP-PATJMRE) in MDCK cells. This protein still localized to TJs albeit with decreased efficiency (Fig. 2A). As predicted, we could show in MDCK cells that full-length EYFP-PATJ could co-immunoprecipitate with endogenous Pals1 (Fig. 2B), which was previously shown to appear as a ~70-kDa doublet on immunoblots (26). In contrast, EYFP-PATJMRE failed to bind Pals1. In Fig. 2B, there was a slight upward shift in mobilities of the proteins in the immunoprecipitated ("IP") lanes relative to those in the "Input" lanes. However, we have consistently observed these molecular weight shifts when immunoprecipitating other proteins and resolving them using the NuPAGE electrophoresis system. Thus, we believe that the mobility shifts of the immunoprecipitated proteins in Fig. 2B represent artifacts attributable to the NuPAGE bis-Tris gels. Given the above results, we deduced that the PATJ/Pals1 interaction most likely contributes to the stability of PATJ at TJs but is not crucial for ultimately targeting PATJ to TJs.

View larger version (48K): [in this window] [in a new window]   Fig. 2.   Deleting the MRE domain of PATJ does not completely abolish the targeting of PATJ to TJs. A, PATJ missing its MRE domain was fused to enhanced yellow fluorescent protein (EYFP) and expressed in MDCK cells. EYFP-PATJMRE was directly visualized, whereas the TJ marker, ZO-1, was visualized by indirect immunofluorescence. B, EYFP-PATJ and EYFP-PATJMRE were immunoprecipitated (IP) from MDCK stable cell lines expressing these proteins. Immunoblotting with affinity purified anti-Pals1 antibody reveals that the endogenous Pals1 doublet co-immunoprecipitates with EYFP-PATJ but not EYFP-PATJMRE. Input denotes straight lysate.

The Carboxyl Terminus of ZO-3 Binds to PATJ in Vitro-- Deleting the MRE domain of PATJ did not completely affect the localization of PATJ to TJs. Thus, we hypothesized that one or more of the PDZ domains were involved in targeting PATJ to TJs. We examined the extreme carboxyl termini of known TJ constituents to search for TJ proteins that could bind class I or class II PDZ domains. Those with PDZ binding consensus sequences were shown in Fig. 3A. The carboxyl termini of ZO-2, ZO-3, and PCLN-1 (claudin 16) are predicted to bind class I PDZ domains, nonetheless, the PDZ proteins that bind to them remain elusive. On the other hand, the carboxyl terminus of JAM interacts with class II PDZ domains present in proteins such as CASK/mLin-2, ZO-1, AF-6, and ASIP/Par-3 (12, 31-33).

View larger version (31K): [in this window] [in a new window]   Fig. 3.   The carboxyl terminus of ZO-3 interacts with PATJ. A, the last 20 amino acids of TJ proteins predicted to bind class I or class II PDZ domains via their carboxyl termini are shown. The last 20 residues of human ZO-2, ZO-3, paracellin-1/claudin-16 (PCLN-1), and JAM were fused to GST and used in the GST pull-down experiment shown in B. B, GST fusion proteins were incubated with lysates derived from 293 cells expressing Myc-PATJ. Bound Myc-PATJ was detected by immunoblotting with anti-PATJ antibodies. The first 181 residues of Pals1 were also fused to GST and used in the experiment. GST-Pals1-(1-181) bears the L27N domain of Pals1 which was previously shown to bind to the amino terminus of PATJ (26).

To determine whether any of these four proteins could associate with PATJ via their carboxyl termini, we fused the last 20 residues of each protein onto GST and the resulting GST fusion proteins were used in in vitro binding experiments. GST fused to the carboxyl terminus of ZO-3 (GST-ZO-3) interacted with PATJ (Fig. 3B). However, the carboxyl termini of ZO-2, PCLN-1, and JAM clearly failed to pull down PATJ. We could also show that PATJ binds to a GST-Pals1-(1-181) fusion protein which contains the L27N domain of Pals1 (positive control).

Next, we sought to narrow down the region of PATJ that could bind to the ZO-3 tail. We expressed truncated versions of PATJ in 293 cells and investigated whether any of these mutants could bind GST-ZO-3. Deleting the last two PDZ domains of PATJ does not abolish binding to the ZO-3 tail. However, PATJ missing its last five PDZ domains fails to interact with the carboxyl terminus of ZO-3 (Fig. 4A). We then wondered if these two truncated PATJ mutants localized to sites of cell contacts. EYFP-PATJPDZ (9-10), but not EYFP-PATJPDZ (6-10), was observed to localize to cell adhesions (Fig. 4B). This suggests that PATJ might be recruited to the TJ through its association with ZO-3 and that the region encompassing the 6^th through 8^th PDZ domain of PATJ may mediate this targeting.

View larger version (38K): [in this window] [in a new window]   Fig. 4.   The region encompassing the 6th through 8th PDZ domains of PATJ was responsible for binding ZO-3 and localizing PATJ to TJs. A, GST- and GST-ZO-3 pull-down experiments were performed utilizing 293 cell lysates containing HA-PATJ and HA-PATJ missing the last two or five PDZ domains: HA-PATJPDZ-(9-10) and HA-PATJPDZ-(6-10), respectively. Precipitated proteins were visualized by blotting with anti-HA antibody. B, EYFP-PATJPDZ-(9-10) and EYFP-PATJPDZ-(6-10) were expressed in MDCK cells and directly visualized by fluorescence microscopy.

The Extreme Carboxyl Termini of ZO-3 and Claudin 1 Interact with PATJ in a PDZ-dependent Fashion-- The above data alone do not provide unequivocal evidence for the notion that the ZO-3/PATJ interaction involves PDZ domains. To better understand the nature of the ZO-3/PATJ association, we sought to determine whether the last three residues (TDL) of ZO-3 were necessary for ZO-3 to bind PATJ. Whereas GST-ZO-3 could interact with PATJ, GST-ZO-3 missing its last three amino acids (GST-ZO-3TDL) could not (Fig. 5A). Based on Fig. 4A, we suspected that the 6^th, 7^th, or 8^th PDZ domain of PATJ represented the binding site for the extreme carboxyl terminus of ZO-3. Therefore, we individually deleted each of these three PDZ domains and repeated the in vitro binding assay with GST-ZO-3 and GST-ZO-3TDL beads. As expected, GST-ZO-3TDL beads failed to pull down any of the PATJ deletion mutants. GST-ZO-3 interacted with both Myc-PATJPDZ7 and Myc-PATJPDZ8 but not with Myc-PATJPDZ6 (Fig. 5B).

View larger version (40K): [in this window] [in a new window]   Fig. 5.   The extreme carboxyl termini of ZO-3 and claudin 1 interact with the 6th and 8th PDZ domains of PATJ, respectively. A, the indicated GST fusion proteins were incubated with lysates prepared from 293 cells expressing full-length Myc epitope-tagged PATJ. GST-ZO-3 and GST-CLDN-1 refer to GST fusion proteins consisting of the last 20 residues of ZO-3 and claudin 1, respectively. GST-CLDN-1YV and GST-ZO-3TDL lack the last two and three residues present in GST-ZO-3 and GST-CLDN-1, respectively. B, GST-ZO-3 or GST-ZO-3TDL were incubated with the indicated PDZ domain deletion PATJ mutants. Bound proteins were resolved by SDS-PAGE. In A and B, the expressed PATJ proteins were visualized by immunoblotting with rabbit polyclonal anti-PATJ. C, GST-CLDN-1 pull-down assays were performed using lysates containing Myc-PATJ or the indicated Myc-PATJ PDZ deletion mutants. D, the ability of the indicated GST fusion proteins to bind two fragments of PATJ (PDZ6-PDZ7 and PDZ8-PDZ10) were examined. In C and D, the bound PATJ proteins were resolved by SDS-PAGE and visualized by anti-Myc immunoblot.

While this study was in progress, it was shown that the carboxyl-terminal YV sequence of claudin 1 (and presumably many other members of the claudin family) was crucial for the ability of this protein to bind the 10^th PDZ domain of MUPP1 (27). This domain of MUPP1 is highly similar in sequence to the 8^th PDZ domain of PATJ. In fact, these respective PDZ domains represent the region of the highest amino acid sequence identity between MUPP1 and PATJ (26). As the claudin 1 carboxyl terminus was shown to interact with MUPP1, we could also demonstrate that a GST-claudin 1 fusion protein (GST-CLDN1) binds PATJ (Fig. 5A). As expected, the GST-CLDN1 missing the last two amino acids (GST-CLDN1YV) failed to interact with PATJ. To further characterize the interaction between PATJ and claudin 1, we performed GST-CLDN1 pull downs with Myc-PATJPDZ6, Myc-PATJPDZ7, and Myc-PATJPDZ8. The GST-CLDN1 fusion protein bound both Myc-PATJPDZ6 and Myc-PATJPDZ7 but failed to interact with Myc-PATJPDZ8 (Fig. 5C).

We also investigated the binding between the carboxyl termini of ZO-3 and claudin 1 and two short fragments of PATJ: PDZ6-PDZ7 and PDZ8-PDZ10. We again observed that the tails of ZO-3 and claudin 1 associate with PATJ at distinct sites as GST-ZO-3 could only bind to a PDZ6-PDZ7 containing fragment and GST-CLDN1 could only pull down a fragment of PATJ containing PDZ8-PDZ10 (Fig. 5D). As expected, both GST-ZO-3TDL and GST-CLDN1YV did not significantly interact with either of the two PATJ fragments. Therefore, we conclude that the extreme carboxyl termini of ZO-3 and claudin 1 interact with the 6^th and 8^th PDZ domains of PATJ, respectively.

The 6^th PDZ Domain Targets PATJ to TJs-- In Fig. 4B, we show that the region of PATJ encompassing the 6^th through 8^th PDZ domain was important for the recruitment of PATJ to cell contacts. Upon closer examination of the data in Fig. 5, we noticed that the interaction between GST-CLDN1 and PATJ was weaker than that between GST-ZO-3 and PATJ. However, the relative affinities observed in these in vitro pull-down assays may not accurately reflect the true affinities between these proteins in vivo. Thus, we wanted to determine the degree to which the 6^th, 7^th, and 8^th PDZ domains each contributed to the targeting of PATJ. We individually deleted these three PDZ domains from EYFP-PATJ and expressed the resulting deletion mutants in MDCK cells. EYFP-PATJ missing the 6^th PDZ domain was predominantly mislocalized from cell contacts (Fig. 6). This was observed in several clones expressing this protein (results not shown). However, EYFP-PATJ missing the 7^th PDZ domain still co-localized with ZO-1 to TJs. Surprisingly, deleting the 8^th PDZ domain, which binds to the carboxyl terminus of claudin 1, had little effect on the targeting of the EYFP-PATJPDZ8 protein to TJs (Fig. 6). These results suggest that the interaction between claudin 1 and PATJ may be dispensable in the targeting of PATJ to TJs, whereas that between ZO-3 and PATJ may be crucial for properly localizing PATJ to TJs.

View larger version (56K): [in this window] [in a new window]   Fig. 6.   Targeting of PDZ deletion PATJ mutants in MDCK cells. EYFP-PATJ missing the 6th, 7th, or 8th PDZ domains were expressed in MDCK cells. The ability of these PATJ mutants to localize to TJs were studied by immunostaining these cells with anti-ZO-1 antibody (red). EYFP-PATJ mutant proteins were directly visualized by fluorescence microscopy (green).

ZO-3 Co-immunoprecipitates with PATJ in Vitro and Co-localizes with PATJ to TJs in Vivo-- To determine whether full-length ZO-3 could form a complex with full-length PATJ, we performed reciprocal co-immunoprecipitation experiments on lysates derived from 293 cells expressing Myc-ZO-3 and Myc-PATJ. Anti-ZO-3 monoclonal antibodies and anti-PATJ antisera were used to immunoprecipitate Myc-ZO-3 and Myc-PATJ, respectively. On the other hand, the anti-Myc monoclonal antibody was used to simultaneously detect both proteins. We observed that PATJ indeed co-precipitates with ZO-3 (Fig. 7A). As a negative control, the identical co-immunoprecipitations were carried out using 293 lysates containing Myc-PATJ and Myc-ZO-3 containing a carboxyl-terminal VSVG epitope tag (denoted as Myc-ZO-3vsvg). The VSVG tag masks the extreme carboxyl-terminal PDZ binding TDL sequence of ZO-3. Hence, Myc-ZO-3vsvg was not expected to bind the PDZ domains. As predicted, the Myc-ZO-3vsvg protein failed to co-immunoprecipitate with Myc-PATJ (Fig. 7A). Finally, we were able to demonstrate that endogenous PATJ and ZO-3 co-localize to TJs in an MDCK epithelial monolayer (Fig. 7B). Collectively, our results indicate that ZO-3 binds to the 6^th PDZ domain of PATJ via its extreme carboxyl terminus. This interaction is likely important for recruiting PATJ and its binding partners such as Pals1 to TJs (Fig. 8).

View larger version (37K): [in this window] [in a new window]   Fig. 7.   ZO-3 and PATJ form a complex and co-localize to TJs in MDCK cells. A, full-length Myc-PATJ was co-expressed with either Myc-ZO3 or Myc-ZO3 tagged with VSVG at the carboxyl terminus (Myc-ZO3vsvg). The latter ZO-3 protein no longer terminates in the TDL sequence and was not predicted to bind a PDZ domain. Immunoprecipitations with anti-ZO-3 monoclonal antibody, preimmune serum (Control IP), or PATJ antisera were performed on the 293 lysates containing the indicated proteins. Bound proteins were resolved by SDS-PAGE. PATJ and the ZO-3 proteins were detected simultaneously by immunoblotting with anti-Myc antibody. The arrowhead and arrow indicate the full-length Myc-PATJ and the ZO-3 proteins, respectively. The asterisk corresponds to a background band observed in lysates derived from 293 cells expressing Myc-PATJ. This band was not recognized by anti-PATJ antibodies (see Anti-PATJ IP lanes and also Fig. 5A). B, MDCK cells were immunostained simultaneously with rabbit anti-PATJ and mouse anti-ZO-3 antibodies. Endogenous PATJ and ZO-3 are shown in green and red, respectively.

View larger version (26K): [in this window] [in a new window]   Fig. 8.   Model depicting the interaction of the PATJ·Pals1·Crumbs complex with TJ constituents in epithelial cells. The 6th PDZ domain of PATJ interacts with the extreme carboxyl-terminal TDL sequence of ZO-3. The 8th PDZ domain of PATJ binds to the carboxyl termini of members of the claudin family which end in the YV sequence. The 1st PDZ domain of ZO-3, as well as ZO-1 and ZO-2 (not shown), also interacts with the carboxyl-terminal YV sequence of claudins (26). The L27N domain of the MAGUK protein, Pals1, binds to the amino-terminal MRE domain of PATJ. In addition, the single PDZ domain of Pals1 binds to the extreme carboxyl-terminal ERLI sequence of the integral membrane protein, Crumbs. It was previously shown that PATJ, Pals1, and Crumbs co-localize to TJs (26). Green crescents represent PDZ domains and are numbered in order from the amino to carboxyl terminus. L27, Lin-2/Lin-7 domain; SH3, Src homology 3 domain; GUK, guanylate kinase-like domain.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We previously showed that a mammalian homologue of Drosophila Discs Lost, PATJ, can exist in a complex with mammalian homologues of Stardust (Pals1) and Crumbs (CRB1) at TJs (26). PATJ contains an amino-terminal MRE domain that mediates the interaction between PATJ and Pals1. Whereas this interaction was crucial for properly targeting Pals1 to TJs, we observed that this was not the case for PATJ (Fig. 2A). In addition, PATJ contains 10 tandemly arranged PDZ domains whose ligands were unknown. We hypothesized that one or more of these PDZ domains conferred the ability of PATJ to localize to TJs. Here, we have identified ZO-3 and claudin 1 as ligands for the 6^th and 8^th PDZ domains of PATJ, respectively. Because many claudin family members terminate in the YV carboxyl-terminal sequence, it seems likely that other claudins also bind to the 8^th PDZ domain of PATJ in epithelial cells.

MUPP1, a paralogue of PATJ, was also shown to localize to TJs in MDCK cells (27). MUPP1 contains 13 PDZ domains in tandem and JAM and claudin 1 were identified as binding partners for the 9^th and 10^th PDZ domains of MUPP1, respectively. The expression of JAM in non-polarized L cell fibroblasts, which lack endogenous claudins and other cell adhesion molecules, was sufficient for recruiting exogenously expressed MUPP1 to the plasma membrane. This suggests that in polarized epithelial cells, the JAM/MUPP1 interaction may be sufficient for and that the claudin/MUPP1 interaction may not be crucial for recruiting MUPP1 to TJs. Here, we find that in contrast to MUPP1, PATJ does not bind to the extreme carboxyl terminus of JAM. Similar to MUPP1, we observe that PATJ can indeed bind the tail of claudin 1. Furthermore, the 8^th PDZ domain of PATJ, which was highly similar in sequence to the 10^th PDZ domain of MUPP1, was shown to mediate this interaction. Nonetheless, we found that claudin 1/PATJ association does not contribute to the targeting of PATJ to TJs as PATJ that was missing the 8^th PDZ domain still localizes properly to TJs. Deleting the 6^th PDZ domain of PATJ, the binding site for the ZO-3 carboxyl terminus, abolishes the targeting of PATJ to TJs. This suggests that the ZO-3/PATJ interaction could be responsible for recruiting PATJ to TJs. The converse, however, was not true considering that the amino-terminal half of ZO-3, which lacks the Src homology 3 domain, guanylate kinase-like domain, and the extreme carboxyl terminus, localizes to TJs (15). It should also be noted that the 6^th PDZ domain of PATJ is a class I PDZ domain (34). Our observation that the extreme carboxyl-terminal TDL sequence of ZO-3 binds this PDZ domain was consistent with that notion.

The PATJ·Pals1·Crumbs and Discs Lost·Stardust·Crumbs complexes exist just superior to adherens junctions in mammalian and Drosophila epithelia, respectively (19, 26). The localization of the latter complex to the subapical complex (SAC) was shown to be crucial for establishment and maintenance of polarity in ectodermally derived epithelia (25, 35-38). In fly embryos lacking Stardust expression, Discs Lost fails to localize to the SAC in epithelial cells established during cellularization (35). In these embryos, Crumbs also fails to localize properly (36). Therefore, Stardust seems to be required for the proper recruitment of Crumbs and Discs Lost to the SAC. In contrast, we find that while the PATJ/Pals1 interaction was necessary for targeting Pals1 to TJs, it does not seem to be crucial for localizing PATJ itself to TJs. In agreement with the Drosophila studies, we have also observed that Crumbs targeting was dependent on its interaction with Pals1.² Therefore, in mammalian epithelia, it appears that the individual components of the PATJ·Pals1·Crumbs complex could be recruited to TJs in sequential fashion. Specifically, ZO-3 may function to recruit PATJ to TJs and in turn PATJ targets Pals1 and Crumbs to TJs.

Interestingly, the Drosophila SAC has been compared with mammalian TJs because several homologues of known TJ proteins such as Par-6, aPKC, and ASIP (DPar-6, DaPKC, and Bazooka, respectively) were known to localize to the SAC (19). Because these three proteins were shown to play a crucial role in the biogenesis of TJs, the SAC likely represents the region of the plasma membrane where TJs will form. It should still be noted that no claudin, occludin, ZO-2, or ZO-3 homologues have been shown to exist within the SAC. The ZO-1 homologue, Polychaetoid (Pyd), has been recently reported to exist as two isoforms, one of which localizes to adherens junctions and the other that displays a broader plasma membrane localization (39). Nonetheless, the precise relationship between Pyd and the SAC, if there is any, remains unclear. Neither the carboxyl termini of ZO-1 nor Pyd are predicted to bind PDZ domains. Collectively, it is tempting to speculate that the presence of additional PDZ domains in PATJ, relative to Discs Lost, correlates with its localization to TJs in mammalian epithelia.

Previously, it was reported that the Par-6·ASIP·aPKC polarity complex was recruited to TJs possibly through the association of the ASIP PDZ domain and the extreme carboxyl terminus of JAM. Here, we have provided insight as to how the PATJ·Pals1·Crumbs polarity complex is localized to TJs as two of the 10 PDZ domains of PATJ were shown to interact with ZO-3 and claudin 1. It will be interesting to identify the proteins that bind to the other eight PATJ PDZ domains. Surely, this will shed more light on how the PATJ·Pals1·Crumbs complex may function and add to the growing list of molecules that contribute to the complex molecular architecture of TJs in mammalian epithelial cells.

	ACKNOWLEDGEMENTS

We are grateful to Erika Wittchen and Bruce Stevenson (University of Alberta) for generously providing the pBK-CMV-ZO3 plasmid. We also thank the University of Michigan Morphology and Image Analysis Core for allowing us to use their confocal microscope.

	FOOTNOTES

^* This work was supported in part by NIDDK, National Institutes of Health Grant DK58208.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ Supported by the Medical Scientist Training Program Grant T32 GM07863 and Predoctoral Genetics Training Program Grant T32 GM07544-24 to the University of Michigan during the course of these studies.

^** Investigator of the Howard Hughes Medical Institute. To whom corresponding should be addressed: Howard Hughes Medical Inst., University of Michigan Medical Center, 4570 MSRB II, Box 0650, 1150 W. Medical Center Dr., Ann Arbor, MI 48109-0650. Tel.: 734-764-3567; Fax: 734-763-9323; E-mail: bmargoli@umich.edu.

Published, JBC Papers in Press, May 20, 2002, DOI 10.1074/jbc.M201177200

² M. Roh and B. Margolis, unpublished data.

	ABBREVIATIONS

The abbreviations used are: TJ, tight junction; CRB, Crumbs; PATJ, Pals1-associated tight junction; Pals1, protein associated with Lin Seven; PDZ, postsynaptic density 95/discs large/zona occludens 1; ZO-3, zona occludens 3; JAM, junctional adhesion molecule; MAGUK, membrane-associated guanylate kinase; MRE, MAGUK recruitment; MUPP1, multiple PDZ protein 1; PCLN, paracellin; SAC, subapical complex; MUPP1, multiple PDZ protein 1; MDCK, Madin-Darby canine kidney; GST, glutathione S-transferase; HA, hemagglutinin; bis-Tris, 2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol; PBS, phosphate-buffered saline; TBS, Tris-buffered saline; VSVG, vesicular stomatitis viral G protein.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				M. S. Balda and K. Matter Tight junctions at a glance J. Cell Sci., November 15, 2008; 121(22): 3677 - 3682. [Full Text] [PDF]

				S. Angelow, R. Ahlstrom, and A. S. L. Yu Biology of claudins Am J Physiol Renal Physiol, October 1, 2008; 295(4): F867 - F876. [Abstract] [Full Text] [PDF]

				K. Duning, E.-M. Schurek, M. Schluter, M. Bayer, H.-C. Reinhardt, A. Schwab, L. Schaefer, T. Benzing, B. Schermer, M. A. Saleem, et al. KIBRA Modulates Directional Migration of Podocytes J. Am. Soc. Nephrol., October 1, 2008; 19(10): 1891 - 1903. [Abstract] [Full Text] [PDF]

				C. J. Mee, J. Grove, H. J. Harris, K. Hu, P. Balfe, and J. A. McKeating Effect of Cell Polarization on Hepatitis C Virus Entry J. Virol., January 1, 2008; 82(1): 461 - 470. [Abstract] [Full Text] [PDF]

				V. M. Stucke, E. Timmerman, J. Vandekerckhove, K. Gevaert, and A. Hall The MAGUK Protein MPP7 Binds to the Polarity Protein hDlg1 and Facilitates Epithelial Tight Junction Formation Mol. Biol. Cell, May 1, 2007; 18(5): 1744 - 1755. [Abstract] [Full Text] [PDF]

				C. H. Storrs and S. J. Silverstein PATJ, a Tight Junction-Associated PDZ Protein, Is a Novel Degradation Target of High-Risk Human Papillomavirus E6 and the Alternatively Spliced Isoform 18 E6 J. Virol., April 15, 2007; 81(8): 4080 - 4090. [Abstract] [Full Text] [PDF]

				D. Massey-Harroche, M.-H. Delgrossi, L. Lane-Guermonprez, J.-P. Arsanto, J.-P. Borg, M. Billaud, and A. Le Bivic Evidence for a molecular link between the tuberous sclerosis complex and the Crumbs complex Hum. Mol. Genet., March 1, 2007; 16(5): 529 - 536. [Abstract] [Full Text] [PDF]

				G. Mandicourt, S. Iden, K. Ebnet, M. Aurrand-Lions, and B. A. Imhof JAM-C Regulates Tight Junctions and Integrin-mediated Cell Adhesion and Migration J. Biol. Chem., January 19, 2007; 282(3): 1830 - 1837. [Abstract] [Full Text] [PDF]

				A. G.S.H. van Rossum, W. M. Aartsen, J. Meuleman, J. Klooster, A. Malysheva, I. Versteeg, J.-P. Arsanto, A. Le Bivic, and J. Wijnholds Pals1/Mpp5 is required for correct localization of Crb1 at the subapical region in polarized Muller glia cells Hum. Mol. Genet., September 15, 2006; 15(18): 2659 - 2672. [Abstract] [Full Text] [PDF]

				M. Tanaka, R. Kamata, and R. Sakai EphA2 Phosphorylates the Cytoplasmic Tail of Claudin-4 and Mediates Paracellular Permeability J. Biol. Chem., December 23, 2005; 280(51): 42375 - 42382. [Abstract] [Full Text] [PDF]

				I. J. Latorre, M. H. Roh, K. K. Frese, R. S. Weiss, B. Margolis, and R. T. Javier Viral oncoprotein-induced mislocalization of select PDZ proteins disrupts tight junctions and causes polarity defects in epithelial cells J. Cell Sci., September 15, 2005; 118(18): 4283 - 4293. [Abstract] [Full Text] [PDF]

				D. Michel, J.-P. Arsanto, D. Massey-Harroche, C. Beclin, J. Wijnholds, and A. Le Bivic PATJ connects and stabilizes apical and lateral components of tight junctions in human intestinal cells J. Cell Sci., September 1, 2005; 118(17): 4049 - 4057. [Abstract] [Full Text] [PDF]

				K. Shin, S. Straight, and B. Margolis PATJ regulates tight junction formation and polarity in mammalian epithelial cells J. Cell Biol., February 28, 2005; 168(5): 705 - 711. [Abstract] [Full Text] [PDF]

				R. Vogelmann and W. J. Nelson Fractionation of the Epithelial Apical Junctional Complex: Reassessment of Protein Distributions in Different Substructures Mol. Biol. Cell, February 1, 2005; 16(2): 701 - 716. [Abstract] [Full Text] [PDF]

				M. Cereijido, R. G. Contreras, and L. Shoshani Cell Adhesion, Polarity, and Epithelia in the Dawn of Metazoans Physiol Rev, October 1, 2004; 84(4): 1229 - 1262. [Abstract] [Full Text] [PDF]

				G. Bazzoni and E. Dejana Endothelial Cell-to-Cell Junctions: Molecular Organization and Role in Vascular Homeostasis Physiol Rev, July 1, 2004; 84(3): 869 - 901. [Abstract] [Full Text] [PDF]

				E. E. Schneeberger and R. D. Lynch The tight junction: a multifunctional complex Am J Physiol Cell Physiol, June 1, 2004; 286(6): C1213 - C1228. [Abstract] [Full Text] [PDF]

				S. W. Straight, K. Shin, V. C. Fogg, S. Fan, C.-J. Liu, M. Roh, and B. Margolis Loss of PALS1 Expression Leads to Tight Junction and Polarity Defects Mol. Biol. Cell, April 1, 2004; 15(4): 1981 - 1990. [Abstract] [Full Text] [PDF]

				K. Ebnet, A. Suzuki, S. Ohno, and D. Vestweber Junctional adhesion molecules (JAMs): more molecules with dual functions? J. Cell Sci., January 1, 2004; 117(1): 19 - 29. [Abstract] [Full Text] [PDF]

				M. H. Roh and B. Margolis Composition and function of PDZ protein complexes during cell polarization Am J Physiol Renal Physiol, September 1, 2003; 285(3): F377 - F387. [Abstract] [Full Text] [PDF]

				T. Ben-Yosef, I. A. Belyantseva, T. L. Saunders, E. D. Hughes, K. Kawamoto, C. M. Van Itallie, L. A. Beyer, K. Halsey, D. J. Gardner, E. R. Wilcox, et al. Claudin 14 knockout mice, a model for autosomal recessive deafness DFNB29, are deaf due to cochlear hair cell degeneration Hum. Mol. Genet., August 15, 2003; 12(16): 2049 - 2061. [Abstract] [Full Text] [PDF]

				M. H. Roh, S. Fan, C.-J. Liu, and B. Margolis The Crumbs3-Pals1 complex participates in the establishment of polarity in mammalian epithelial cells J. Cell Sci., July 15, 2003; 116(14): 2895 - 2906. [Abstract] [Full Text] [PDF]

				E. Knust and O. Bossinger Composition and Formation of Intercellular Junctions in Epithelial Cells Science, December 6, 2002; 298(5600): 1955 - 1959. [Abstract] [Full Text] [PDF]

				S. Poliak, S. Matlis, C. Ullmer, S. S. Scherer, and E. Peles Distinct claudins and associated PDZ proteins form different autotypic tight junctions in myelinating Schwann cells J. Cell Biol., October 28, 2002; 159(2): 361 - 372. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 277/30/27501    most recent M201177200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map 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 Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Roh, M. H. Articles by Margolis, B. Search for Related Content PubMed PubMed Citation Articles by Roh, M. H. Articles by Margolis, B. Social Bookmarking                      What's this?