The coxsackievirus and adenovirus receptor interacts with the multi-PDZ domain protein-1 (MUPP-1) within the tight junction.

The coxsackievirus and adenovirus receptor (CAR) is a component of the epithelial cell tight junction. In a yeast two-hybrid screen we identified the multi-PDZ domain protein MUPP1 as an interaction partner for the CAR cytoplasmic domain. CAR and MUPP1 were found to colocalize at the tight junction, to coprecipitate from epithelial cells, and to interact in vitro. The interaction was found to specifically involve the PDZ-binding motif within the CAR C terminus and MUPP1 PDZ domain 13. In transfected cells, CAR recruited MUPP1 to cell-cell contacts. The inhibition of CAR expression with small interfering RNA inhibited MUPP1 localization to the tight junction. The results indicated that CAR interacts with MUPP1 and is involved in MUPP1 recruitment to the tight junction.

The coxsackievirus and adenovirus receptor (CAR) 1 was first identified as a cellular protein involved in attachment and infection by group B coxsackieviruses and adenoviruses (1)(2)(3)(4). CAR also functions as a transmembrane component of epithelial tight junctions (TJs) (5). The TJ is the apical-most component of the junctional complex of epithelia and endothelia and is critical in maintaining and restricting the paracellular flow of ions and solutes. Though CAR localizes to the TJ, the nature of its interactions with other TJ-associated proteins remains unclear.
Several PDZ domain-containing proteins have been localized to the cytoplasmic region of the TJ. These include the structurally related membrane-associated guanylate kinase proteins zonula occludens (ZO)-1, ZO-2, and ZO-3, the membrane-associated guanylate kinase inverted proteins, the partitioning proteins, and the multi-PDZ domain protein 1 (MUPP1). These proteins likely function to link TJ-associated transmembrane proteins to intracellular signaling molecules.
The CAR C terminus resembles hydrophobic C-terminal peptide motifs known to interact with PDZ protein domains, such as those contained within membrane-associated guanylate kinase proteins and other scaffolding proteins localized to the TJ region. CAR associates with ZO-1, as demonstrated by coprecipitation from polarized epithelia and by the relocalization of ZO-1 that occurs in CAR-transfected Chinese hamster ovary (CHO) cells (5). CAR may also interact with a component of the adherens junction, ␤-catenin, as has been reported in A549 cells (6). In addition, yeast-two-hybrid studies have shown that CAR interacts with Ligand of Numb-X (LNX) (7), a PDZ protein believed to regulate Notch signaling in the central nervous system; however, LNX is not known to associate with TJs (8,9).
To define the associations mediated by the cytoplasmic tail of CAR, we performed a yeast two-hybrid screen and identified MUPP1 as a CAR-interacting protein. The localization of MUPP1 to the area of the TJ was disrupted when CAR was absent, although the distribution of other TJ-associated components such as ZO-1 remained unchanged. Taken together, these data indicate that CAR interacts directly with MUPP1 and that this interaction is involved in restricting MUPP1 to the cytoplasmic region of the TJ.

MATERIALS AND METHODS
Cell Culture-CHO cells stably transfected with cDNA constructs encoding human CAR (CHO-CAR cells), CAR lacking its cytoplasmic tail (CHO-CAR ⌬tail ), CAR with deletion of its hydrophobic C-terminal motif (CHO-CAR AQS ), or with vector alone (CHO-pcDNA) were cultured in nucleoside-free ␣-minimum essential medium containing 10% dialyzed fetal calf serum (5).
Antibodies-CAR-specific antiserum was generated in rabbits and affinity-purified as described (12). Rabbit polyclonal antibodies against MUPP1 were kindly provided by Shoichiro Tsukita and Ronald T. Javier (10,11). Mouse anti-ZO-1 and rabbit anti-proteasome subunit antibodies were purchased from Zymed Laboratories, Inc.. Antibodies to HA, GAPDH, early endosome antigen-1, Rab7, lysosomal integral membrane proteins-II (LIMP II), and lysosome-associated membrane protein 1 (LAMP I) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-CAR monoclonal antibody RmcB and sheep anti-MUPP1 were purchased from Upstate (Lake Placid, NY). Control MOPC21 was purchased from Sigma. Oregon Green-conjugated anti-HA antibody was purchased from Clontech. Fluorochrome-conjugated secondary antibodies for immunofluorescence were purchased from Jackson ImmunoResearch (West Grove, PA). Horseradish peroxidase-conjugated antibodies were obtained from Santa Cruz Biotechnology.
Yeast Two-hybrid Screening-DNA encoding the mCAR1 cytoplasmic domain (CCHRK 3 DGSIV) was inserted into plasmid pGBKT7 (Clontech), in-frame with the yeast GAL4 DNA binding domain. The resulting plasmid (CAR-BD) was used as bait in a yeast two-hybrid screen of an 11-day mouse embryo library constructed in pGADT7 using protocols provided with the Clontech Matchmaker Two-Hybrid System 3. Saccharomyces cerevisiae strain AH109 was transformed with both library and bait plasmids, and ϳ2.5 ϫ 10 6 dual transformants were screened for growth on medium deficient in histidine and adenine and for production of ␤-galactosidase. Library plasmids were recovered from putative positive clones and further tested to eliminate any that induced GAL4 activation independently of any bait plasmid, in the presence of binding domain alone, or in the presence of binding domain fused to a nonspecific control protein (human lamin C). Of 11 cDNAs that were obtained, one encoded PDZ domain 13 of MUPP1; the other 10 encoded a PDZ domain derived from another protein.
Immunofluorescence Microscopy-CHO cells plated on glass chamber slides or siRNA-transfected Caco-2 cells on collagen-coated chamber slides were washed in phosphate-buffered saline and fixed in 1% paraformaldehyde in phosphate-buffered saline for 15 min at room temperature, washed, and permeabilized with 1% Triton X-100 for 10 min at room temperature. Cells were incubated with the indicated primary antibodies for 1 h at room temperature, washed, incubated with fluorochrome-conjugated secondary antibodies for 30 min at room temperature, washed, and mounted with Vectashield (Vector Laboratories, Burlingame, CA) containing 4Ј,6-diamidino-2-phenylindole (DAPI). All cells were postfixed with 4% paraformaldehyde. Images were captured with an Olympus fluorescence microscope (Melville, NY).
Caco-2 cells grown on transwell inserts were permeabilized with methanol for 5 min at Ϫ20°C, washed briefly in phosphate-buffered saline, and blocked with 5% bovine serum albumin for 30 min at 37°C. Cells were incubated with monoclonal anti-CAR RmcB and polyclonal anti-MUPP1 for 1 h at room temperature. Cells were washed and incubated with anti-mouse fluorescein isothiocyanate and anti-rabbit Texas Red in 10% goat serum for 30 min at room temperature, washed, and postfixed with 4% paraformaldehyde. Images were captured with a confocal laser-scanning microscope (Leica, Exton, PA).
Reverse Transcription-PCR-Total RNA was isolated with RNA STAT 60 (Tel-Test, Friendswood, TX) and treated with DNase (Ambion). For complementary DNA synthesis, 1 g of total RNA was used in a 20-l reaction containing 1 mM dNTPs, 2.5 mM oligo(dT), 0.1 volume of 10ϫ buffer (supplied by manufacturer), and 2500 units/ml murine leukemia virus reverse transcriptase (Invitrogen). The reverse transcription reaction was carried out at 1 cycle in a thermal cycler at 42°C for 50 min, followed by a 15-min incubation at 70°C. PCR was carried out with TaqDNA polymerase (Roche) for 25 cycles with primers to CAR (sense primer 5Ј-GCT TAG TCC CGA AGA CCA-3Ј and antisense 5Ј-GTG GCA CAT CTT CCC TGA-3Ј) and those to GAPDH (sense primer 5Ј-ACC ACC AAC TGC TTA GCA-3Ј and antisense 5Ј-CCC TGT TGC TGT AGC CAA-3Ј) added simultaneously to the PCR. PCR products were separated on a 1.5% agarose gel containing ethidium bromide Immunoprecipitations and Western Blotting-Lysates of CHO cells transiently transfected with HA-MUPP1 were prepared with 0.1% Triton X-100 extraction buffer containing phenylmethanesulfonyl fluoride. Cells were collected with a rubber policeman in lysis buffer and sonicated briefly. Insoluble material was cleared by centrifugation for 5min at 4°C. For coimmunoprecipitation of MUPP1 and CAR followed by immunoblotting for CAR, lysates were incubated with an anti-HA monoclonal antibody, anti-CAR monoclonal antibody RmcB, or control monoclonal antibody MOPC 21 for 1 h at 4°C. For coimmunoprecipitation of MUPP1 and CAR followed by immunoblotting for MUPP1, lysates were incubated with anti-MUPP1 polyclonal antibody, RmcB, or control MOPC 21 overnight at 4°C. Sepharose G beads were added for an additional 1 h. After centrifugation, the beads were washed in cell lysis buffer and then heated at 95°C for 10 min in Laemmli sample buffer. Following a brief centrifugation, the supernatant was run on a 4 -15% Tris-HCl gel (Bio-Rad) and transferred to a polyvinylidene difluoride membrane. The membrane was blocked overnight in Trisbuffered saline containing Twenn-20 (TBS-T) with 5% milk and probed with rabbit anti-CAR or blocked in 10% donkey serum and probed with anti-sheep MUPP1 followed by horseradish peroxidase-conjugated antibody to rabbit or sheep immunoglobulin (Jackson ImmunoResearch), developed with ECL reagents (Amersham Biosciences), and exposed to film.
Streptag-SUMO-hCAR Fusion Protein-A bacterial expression vector encoding the full-length human CAR cytoplasmic domain was constructed and expressed in a modified pET expression vector (provided by Steven Weeks, Drexel University, Philadelphia, PA) encoding a linear fusion consisting of an N-terminal streptactin-tag (IBA-GmbH) for affinity purification followed by SUMO (Invitrogen) linked to the cytoplasmic domain of CAR. The cytoplasmic domain of CAR was amplified by PCR and cloned into the modified pET vector between a BSAI site created at the 3Ј-end of SUMO, and an XhoI site using the following primers, sense primer, 5Ј-TTTTTTGGTCTCACATGTGCTGTCGTAAA-AAGCGC-3Ј and antisense primer, 5Ј-TTTTTTCTCGAGCTATACTAT-AGACCCATCCTTGCT-3Ј.
The pET-CAR expression vector was introduced into competent Escherichia coli Rosetta (DE3) cells (Novagen). Overnight cultures grown in LB were diluted 1:5 into a final culture volume of 25 ml. At an A 600 of 0.6 the culture was exposed to 1 mM isopropyl-1-thio-␤-D-galactopyranoside (Fisher), and bacteria were collected after 4 h. The bacterial cell pellet was resuspended in phosphate-buffered saline containing 1 mM phenylmethanesulfonyl fluoride, 0.5 M NaCl, and 1% Triton X-100. The cells were lysed by sonication, and the cell debris was removed by centrifugation at 10,000 rpm for 10 min. The supernatant was incubated with 2 ml of a 50% slurry of Streptactin-MacroPrep beads (IBA-GmbH) for 2 h. at 4°C. The beads were washed four times with cold phosphate-buffered saline, and the protein was eluted from the beads with 2.5 mM desthiobiotin. The desthiobiotin was removed by dialyzing the protein into 20 mM Hepes, pH 6.7.
In Vitro Binding Assays-GST fusion proteins expressed in E. coli were purified using glutathione-Sepharose 4B beads (Amersham Biosciences). SUMO or SUMO-hCAR (1 g) was added to a pull-down buffer (1% Triton X-100, 100 mM NaCl, 0.2 mM phenylmethanesulfonyl fluoride, 5% fetal bovine serum) and incubated with GST, GST-MUPP1 PDZ10 , or GST-MUPP1 PDZ12-13 bound to glutathione-Sepharose 4B beads for 1 h at 4°C. Beads were sedimented by centrifugation, washed in pull-down buffer, and heated at 95°C for 10 min in Laemmli sample buffer. Following a brief centrifugation, the supernatant was run on a 4 -15% Tris-HCl gel and transferred to a polyvinylidene difluoride membrane. The membranes were blocked overnight in 5% bovine serum albumin at 4°C and incubated with anti-Streptavidin monoclonal antibody. Alternatively, samples were run on 4 -15% gels and stained with Coomassie Blue to control for loading of GST and GST fusion proteins.

Identification of MUPP1 as a CAR Binding Partner-To
identify the proteins that interact with the C-terminal tail of murine CAR1, we performed yeast two-hybrid screening. The mCAR cytoplasmic domain was fused to the yeast GAL4 DNA binding domain, and used as bait to screen an 11-day mouse embryo library. A cDNA encoding the thirteenth PDZ domain of the multi-PDZ domain protein MUPP1 was isolated; the two-hybrid interaction with the CAR C terminus appeared to be specific (Table I).
Direct in Vitro Interaction between CAR and MUPP1-The yeast two-hybrid screen indicated that the CAR cytoplasmic pGADT7-T Ag ϩ domain interacts directly with PDZ domain 13 of MUPP1. We tested whether a GST fusion protein containing MUPP1 PDZ domain 13 could precipitate the CAR cytoplasmic domain prepared as a streptactin-tagged SUMO fusion protein (SUMO-hCAR). MUPP1 GST fusion protein was immobilized on glutathione-Sepharose beads and incubated with SUMO-hCAR in buffer containing detergent and bovine serum albumin. Beads were washed, and precipitated proteins were blotted for streptactin. A GST protein containing MUPP1 PDZ domains 12 and 13 (MUPP PDZ 12-13 ) precipitated SUMO-hCAR but not streptacin-tagged SUMO (Fig. 1). GST alone, or a control fusion protein containing MUPP1 PDZ domain 10, did not precipitate SUMO-hCAR. These results confirm that the CAR C terminus interacts directly with MUPP1. CAR Associates with MUPP1 and Recruits MUPP1 to Intercellular Junctions-In a human polarized epithelial cell line, Caco-2, both CAR and MUPP1were localized to the apical region of the lateral cell membrane as determined by confocal immunofluorescence microscopy ( Fig. 2A). ZO-1 colocalized with both CAR and MUPP1 (not shown), consistent with previous reports that all three proteins are components of the epithelial tight junction. When CAR was precipitated from Caco-2 cell lysates with monoclonal antibody RmcB, MUPP1 was detected in the precipitates by immunoblotting with a polyclonal anti-MUPP1 antibody (Fig. 2B). These results demonstrated that CAR and MUPP1 are physically associated at the tight junction of human epithelial cells.
CHO cells do not naturally form tight junctions and do not express many tight junction components. When HA-tagged human MUPP1 was introduced into CHO cells engineered to express human CAR (CHO-hCAR), CAR and MUPP1 could be coprecipitated with either anti-CAR or by anti-HA antibodies (Fig. 3), and HA staining was seen exclusively at cell borders (Fig. 4A). In sparsely plated cultures, staining for both CAR and HA-MUPP1 was accentuated at sites of cell-cell contact (Fig. 4B). In contrast, HA-tagged MUPP1 introduced into control CHO cells (CHO-pcDNA) was expressed in a diffuse cytoplasmic pattern (Fig. 4A) with no staining at cell junctions. These results indicated that transfected MUPP1 and CAR associate in non-polarized CHO cells and suggested that as-sociation with CAR recruits MUPP1 to sites of intercellular contact.
Association with MUPP1 Depends on the CAR C Terminus-HA-MUPP1 introduced into CHO cells expressing truncated human CAR, with deletion of the entire cytoplasmic domain (CAR ⌬tail ) or deletion of the C-terminal DGSIV motif (CAR AQS ), was expressed throughout the cytoplasm, with no localization to cell-cell contacts (Figs. 4A). Unlike full-length CAR, CAR ⌬tail and CAR AQS did not coprecipitate with MUPP1, and MUPP1 did not coprecipitate with truncated CAR (Fig. 3). Thus, the association between CAR and MUPP1, and recruitment of MUPP1 to sites of cell contact, depends on the presence of the putative PDZ-binding motif at the CAR C terminus. Given the results of the yeast two-hybrid screen, and the in vitro pulldown experiment shown above, it is likely that CAR association with MUPP1 involves an interaction between the C-terminal motif and MUPP1 PDZ domain 13.
Localization of MUPP1 in the Absence of CAR-The experiments in CHO cells indicated that CAR recruits MUPP1 to sites of contact between transfected CHO cells. To determine whether endogenous CAR is involved in MUPP1 recruitment to the tight junction of polarized epithelial cells, we used RNA interference technology to silence CAR expression. siRNAs specific to CAR were transiently transfected into Caco-2 cells. RT-PCR was performed at 48-h post-transfection to determine the extent of CAR mRNA knockdown by siRNA transfection. Caco-2 cells transfected with a control siRNA with no sequence similarity to any human gene sequence showed a high level of CAR expression, indicating that transfection of cells with a nonspecific siRNA had no effect on CAR mRNA expression (Fig. 5A, right). However, transfection of cells with a siRNA specific for CAR significantly reduced endogenous CAR mRNA levels (Fig. 5A, left). To determine whether the decrease of mRNA levels resulted in a reduction in protein expression, control and CAR siRNA-transfected lysates were subjected to Western blot analysis. CAR protein expression was significantly reduced in CAR siRNA-transfected cells compared with controls (Fig. 5B). Immunoblots were stripped and reprobed for GAPDH to control for equal protein loading. The residual levels of CAR mRNA and protein most likely To establish whether the loss of CAR expression by siRNA transfection would affect the localization or expression of TJassociated proteins, particularly MUPP1, we performed immunostaining and Western blot analysis. Lysates from CAR or control siRNA-transfected Caco-2 were collected and probed with antibodies to several TJ proteins including ZO-1 and MUPP1. Down-regulation of CAR had no effect on the expression of either ZO-1 or MUPP1 (Fig. 5B). Loss of CAR expression also did not affect the distribution of ZO-1 as assessed by immunostaining (Fig. 6A). Several other TJ and adherens junction proteins were unaffected by CAR silencing, including ZO-2, junctional adhesion molecule-1, claudin-1, and ␤-catenin (data not shown).
In cells transfected with a nonspecific siRNA, MUPP1 distribution remained intact, as characterized by a continuous ring circumscribing each cell (Fig. 6B). In contrast, MUPP1 was dramatically relocalized to punctate bodies within the cytoplasm in cells transfected with a siRNA directed against CAR (Fig. 6B). As determined by serial confocal cross-sections, these areas of punctate MUPP1 staining resided in the apicalmost domain of the epithelium (within 1.75 m) with little localization below the level of the apical TJ (data not shown).
To determine whether these bodies resulted from the relocalization or recycling of MUPP1 through endocytic vesicles or from its being targeted for degradation, we attempted to stain MUPP1 intracellular bodies with antibodies specific for endosomal markers and for components of the ubiquitin degradation pathway. MUPP1-positive bodies did not express the early endosome marker early endosome antigen 1 (13), the late endosome marker Rab7 (14), the lysosomal markers LIMP II (Fig.  7), or the late endosome/lysosomal marker LAMP I (data not shown) (15). In addition, MUPP1-positive bodies did not colo- calize with internalized fluorochrome-labeled dextran (data not shown), further indicating that MUPP1 was not relocalized to endocytic compartments. MUPP1-positive intracellular bodies did not stain with antibodies specific for ubiquitin or the 26 S proteasome (data not shown). These data indicated that CAR is involved in the localization of MUPP1 to the TJ and that in the absence of CAR, MUPP1 may be retargeted to specialized vesicles within the cytoplasm. DISCUSSION We found that the CAR cytoplasmic domain interacts with the multi-PDZ domain protein MUPP1. The interaction specifically involves the PDZ-binding motif within the CAR C terminus and MUPP1 PDZ domain 13. In transfected cells, localization of CAR to cell-cell contact appears to recruit MUPP1. In polarized epithelial cells, endogenous CAR and MUPP colocalize and are physically associated within the tight junction. Inhibition of endogenous CAR expression in polarized epithelia by siRNA transfection leads to relocalization of MUPP1.
MUPP1 localizes to the TJ and interacts with junctional adhesion molecule and with members of the claudin family of transmembrane proteins (11,16). The loss of CAR within the TJ by siRNA transfection led to pronounced relocalization of MUPP1 from the apical TJ complex to punctate bodies within the apical domain. We were unable to identify or define the nature of the MUPP1 bodies, and it remains unclear whether they result from recycling of MUPP1 from the TJ or the whether targeting of MUPP1 to the TJ is disrupted in the absence of CAR. MUPP1 is known to interact with the junctional adhesion molecule and claudin-1, and it is not known whether loss of these components would have a similar effect. The evidence in CHO cells indicates that CAR is capable of recruiting MUPP1 to sites of cell contact and therefore may indicate that loss of CAR leads to improper targeting of MUPP1 to the TJ. Other junctional proteins including junctional adhesion molecule, claudin-1, and ZO-1 have been shown to reside in unique storage organelles following a calcium switch (17). Therefore, it is possible that the absence of CAR leads to the destabilization of the junction and relocalization of MUPP1 to similar organelles. The loss of membrane-associated MUPP1 has been postulated to correlate with the transforming ability of two viral oncoproteins, adenovirus E4-ORF1 and high risk Papillomavirus type 18 E6 (10). Interestingly, in the presence of E4-ORF1, MUPP1 has been noted to relocate to intracellular bodies within the cytoplasm similar in appearance to those induced by silencing of endogenous CAR.
In addition to TJ proteins, MUPP1 interacts with several important mediators of cellular signaling, which may indicate its role as an adaptor for intracellular signaling molecules. By yeast two-hybrid screening, MUPP1 interacts with serotonin 5-hydroxytryptamine type 2C receptors (18,19), kinase negative c-Kit (20), the tandem plextrin homology domain protein (21), and the membrane-spanning NG2 proteoglycan (22).
In several human tumors, CAR expression is down-regulated during progression to malignancy (23)(24)(25), and the expression of transfected CAR in CAR-deficient tumor cells leads to alterations in cell cycle regulation and decreased proliferation (23). It is interesting to note that MUPP1 is targeted by viral transforming proteins and that other junctional PDZ proteins regulate cell proliferation (26,27). It remains to be determined whether the effects of CAR on cell proliferation are mediated through its interaction with MUPP1 or other junctional proteins.
It has been reported that intercellular junctions are disrupted when adenovirus fiber is applied to the basal surface of a polarized epithelial monolayer (6). The mechanism by which this occurs has not been determined, and it will be interesting to learn whether CAR interactions with TJ proteins are affected during the process of virus entry into cells.