PAPIN

A neural plakophilin-relatedarmadillo repeat protein (NPRAP)/δ-catenin interacts with one of Alzheimer disease-related gene products, presenilin 1. We have previously reported the interaction of NPRAP/δ-catenin with synaptic scaffolding molecule, which is involved in the assembly of synaptic components. NPRAP/δ-catenin also interacts with E-cadherin and β-catenin and is implicated in the organization of cell-cell junctions. p0071, a ubiquitous isoform of NPRAP/δ-catenin, is localized at desmosomes in HeLa and A431 cells and at adherens junctions in Madin-Darby bovine kidney cells. We have identified here a novel protein interacting with NPRAP/δ-catenin and p0071 and named this protein plakophilin-related armadillo repeat protein-interacting PSD-95/Dlg-A/ZO-1 (PDZ) protein (PAPIN). PAPIN has six PDZ domains and binds to NPRAP/δ-catenin and p0071 via the second PDZ domain. PAPIN and p0071 are ubiquitously expressed in various tissues and are localized at cell-cell junctions in normal rat kidney cells and bronchial epithelial cells. PAPIN may be a scaffolding protein connecting components of epithelial junctions with p0071.

The armadillo repeat is a repeated motif of about 40 amino acids originally identified in the Drosophila segment polarity gene, armadillo (reviewed in Ref. 1). The list of proteins containing this repeat includes ␤-catenin, plakoglobin, adenomatous polyposis coli gene product, a regulatory protein for small G protein named smg GDP dissociation stimulator, and smg GDP-dissociation stimulator-associated protein (reviewed in Refs. 1 and 2). Among them, p120 ctn and its related proteins form a family. The p120 ctn family is composed of p120 ctn , B6P/ plakophilin 1, plakophilin 2, armadillo repeat gene deleted in velo-cardiofacial syndrome, p0071, and neural plakophilin-related armadillo repeat protein (NPRAP) 1 /␦-catenin (3)(4)(5)(6)(7)(8)(9)(10). p120 ctn has been identified as a major substrate of tyrosine kinase phosphorylation enriched at adherens junctions and contains 10 armadillo repeats (3,(11)(12)(13). p120 ctn directly interacts with E-cadherin. B6P/plakophilin 1 is a protein localized at desmosomes and interacts with keratin (4,5). Plakophilin 2 is also localized at desmosomes (6). armadillo repeat gene deleted in velo-cardiofacial syndrome has been isolated as a gene involved in a human disorder that affects the development of tissues related to pharyngeal arches and pouches and encodes a protein containing 10 armadillo repeats (10). p0071 also has 10 armadillo repeats and is colocalized with desmoplakin in HeLa and A431 cells, but localized at adherens junctions in Madin-Darby bovine kidney cells (8). NPRAP/␦-catenin was discovered independently by two groups as a protein homologous to B6P/plakophilin 1 and as a protein interacting with presenilin 1, respectively (9,10). NPRAP/␦-catenin has a structure similar to that of p0071 and is considered to be a neural isoform of p0071. When transfected in Madin-Darby canine kidney cells, NPRAP/␦-catenin is localized at adherens junctions and interacts with E-cadherin and ␤-catenin (14). We have reported that NPRAP/␦-catenin interacts with a multiple PSD-95/Dlg-A/ZO-1 (PDZ) domain-containing protein named synaptic scaffolding molecule, which interacts with various synaptic components (15). These observations suggest that NPRAP/␦-catenin and p0071 play roles in the organization of cell-cell junctions.
We have here identified a novel multiple PDZ domain-containing protein in the yeast two-hybrid screening with NPRAP/ ␦-catenin as a bait. This protein is expressed ubiquitously and colocalized with p0071 in epithelial cells, suggesting that this protein is a p0071-binding protein. We have named this protein plakophilin-related armadillo repeat protein-interacting PDZ protein (PAPIN).

MATERIALS AND METHODS
Yeast Two-hybrid Screening and cDNA Screening-Yeast two-hybrid screening was performed using a rat brain cDNA library as described previously (16). cDNA library screenings were performed by plaque hybridization using rat brain cDNA libraries (Stratagene) (16).
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF16941.
Antibodies-The rabbit anti-PAPIN antibodies were raised against the products of pGex5X-3 PAPIN-2, -3, and a synthetic peptide, TYN-GNSGNSSEPGETPTLELC. The rabbit anti-p0071 antibody was raised against the product of pGex5X-3 p0071-1. Mouse anti-Myc-tag monoclonal and anti-E-cadherin monoclonal antibodies were obtained from American Type Culture Collection and from Transduction Laboratories, respectively. Rhodamine-conjugated and fluorescein isothiocyanateconjugated second antibodies for dual labeling were purchased from Chemicon.
Pull-down Experiment with Glutathione S-Transferase (GST) Fusion Protein of PAPIN Using Urea/Detergent Extracts of Rat Brain Synaptosomes-Crude synaptosomes were prepared from rat brains as described previously (17). Crude synaptosomes from two rat brains were homogenized in 4 ml of 20 mM Hepes/NaOH, pH 7.4, containing 1% (w/v) Triton X-100 and 6 M urea and then centrifuged at 100,000 ϫ g again for 15 min. The supernatant was dialyzed against 2 liters of 20 mM Hepes/NaOH, pH 8.0, containing 100 mM NaCl overnight and centrifuged at 100,000 ϫ g for 30 min. 2 ml of the supernatant was incubated with 200 pmol of the products of pGex5X-3-PAPIN-9 or GST fixed on glutathione-Sepharose 4B beads. After the beads were washed with 20 mM Hepes/NaOH, pH 7.4, containing 100 mM NaCl and 0.5% (w/v) Triton X-100 four times, the proteins attached to the beads were detected with the immunoblotting.
In Vitro Binding Experiment Using GST Fusion Proteins and COS Cell Extracts-COS cells were cultured in Dulbecco's modified Eagle medium with 10% fetal bovine serum under 10% CO 2 at 37°C and transfected with various Myc-tagged constructs using the DEAE-dextran method (18). COS cells of two 10-cm dishes were homogenized in 1.6 ml of 20 mM Hepes/NaOH, pH 7.4, containing 6 M urea and 1% (w/v) Triton X-100 and centrifuged at 100,000 ϫ g for 15 min. The supernatant was dialyzed against 20 mM Hepes/NaOH, pH 8.0, containing 100 mM NaCl and centrifuged at 100,000 ϫ g for 15 min again. The supernatant was used as an urea/detergent extract. Each 0.5-ml aliquot of the COS cell extracts was incubated with 200 pmol of various GST fusion proteins fixed on 20 l of glutathione-Sepharose 4B beads. After the beads were washed with 20 mM Hepes/NaOH, pH 8.0, containing 100 mM NaCl and 0.5% (w/v) Triton X-100, the proteins on the beads were detected with the immunoblotting using the anti-Myc antibody.
Immunoprecipitation from COS Cells-COS cells were transfected with pClneo PAPIN with either pClneo p0071-1 or -3. The urea/detergent extracts of COS cells were prepared as described above. Each 0.5-ml aliquot was incubated with the anti-p0071 antiserum or the preimmune serum fixed on protein G-Sepharose 4 Fast Flow beads. After the beads were washed, the proteins in the immunoprecipitates were immunoblotted with the anti-PAPIN antibody.
Immunocytostaining and Immunohistochemistry-NRK cells were fixed with methanol at Ϫ20°C for 15 min, incubated in 0.2% (w/v) Triton X-100 in phosphate-buffered saline (PBS), and blocked with 1% (w/v) bovine serum albumin in PBS at room temperature for 30 min. For immunohistochemistry, adult rats were perfused with 4% (w/v) paraformaldehyde in PBS. The fixed samples were treated with 0.2% (w/v) Triton X-100 in PBS and washed with PBS. After being blocked with 10% (v/v) goat serum, the cells were incubated with various first antibodies and visualized with the rhodamine-or fluorescent isothiocyanate-conjugated second antibodies. The images were obtained by a confocal microscopy (Olympus FV300-BX).
Miscellaneous Procedures-The subcellular fractionation of rat brain, SDS-polyacrylamide gel electrophoresis, and determination of protein concentrations were performed as described (16 -18). Western and Northern blottings were performed using ECL reagents (Amersham Pharmacia Biotech) and multiple tissue Northern blots (CLON-TECH), respectively.

RESULTS
To identify a NPRAP/␦-catenin-interacting protein, we performed a yeast two-hybrid screening using various bait constructs of NPRAP/␦-catenin. From 1 ϫ 10 6 transformants, only one positive clone, pPrey 5001, was obtained. The sequence analysis of pPrey 5001 revealed that it contained three PDZ domains. To obtain a full-length clone, we screened a random-primed rat brain cDNA library using the insert from pPrey 5001 as a probe. No single clone contained both of a putative initiation methionine and a termination codon, but from the analysis of several overlapping clones, we tentatively identified a protein composed of 2766 amino acids (Fig. 1A). We have named this protein PAPIN. PAPIN has four PDZ domains in the N-terminal region and two PDZ domains in the C-terminal region. The middle region between the fourth and fifth PDZ domains did not show any significant homology to known proteins (Fig. 1B, a). pPrey 5001 contained the amino acids 231-811 covering the second to fourth PDZ domains (Fig. 1A, underlined). A clone, p5001-52, had an insert of 52 amino acids between the second and third PDZ domains (Fig. 1B, b). Another clone, p5001-61, also had an insert and a termination codon after the fourth PDZ domain (Fig. 1B, c).
pPrey 5001 interacted only with pBTM116 ␦-catenin-1 and not with pBTM116 ␦-catenin-3 or -4, which did not contain the C-terminal region of NPRAP/␦-catenin (data not shown) (9, 10). The PDZ domains of PAPIN were speculated to bind the Cterminal PDZ-binding motif of NPRAP/␦-catenin. Actually, the GST fusion protein containing the first and second PDZ domains of PAPIN interacted with the native NPRAP/␦-catenin from rat crude synaptosomes ( Fig. 2A). Furthermore, we confirmed that PAPIN was enriched in the synaptic plasma membrane and postsynaptic density fractions as well as NPRAP/␦catenin in the rat brain subcellular fractionations (Fig. 2B).
Next, we performed Northern blot analysis using various rat tissues. The messages with 12 kilobase pair were detected in heart, brain, spleen, lung, liver, skeletal muscle, kidney, and testis (Fig. 3A). The tissue distribution of the messages of PAPIN was similar to that of p0071 (Fig. 3B), but not that of NPRAP/␦-catenin, which was specifically expressed in brain as previously reported (data not shown and Ref. 10). In the hybridization with the probe of ␤-actin, all the lanes gave similar signals (data not shown).
p0071 has a PDZ-binding motif similar to that of NPRAP/␦catenin, although the motif is not the C terminus of p0071 (Fig. 4A), suggesting that p0071 may also interact with PAPIN. To confirm this possibility, we tested whether PAPIN interacted with the PDZ-binding motif of p0071 in COS cells. For this purpose, we prepared various constructs of p0071 (Fig.  4B). p0071-1 contained the full-length of p0071 (Fig. 4B, a). p0071-2 lacked the C terminus but contained the PDZ-binding motif (Fig. 4B, b). p0071-3 lacked the PDZ-binding motif (Fig. 4B, c). First, we overexpressed PAPIN with either p0071-1 or -3 in COS cells and immunoprecipitated p0071. PAPIN was coimmunoprecipitated with p0071-1 but not with p0071-3 (Fig. 4C). We also examined the interaction of GST-PAPIN and various Myc-tagged constructs of p0071. The products of pClneo Myc p0071-1 and -2 interacted with the GST fusion protein containing the first and second PDZ domains of PAPIN, whereas that of pClneo Myc p0071-3 did not (Fig. 4D). These results indicate that PAPIN interacts with the PDZbinding motif of p0071.
To determine which PDZ domain of PAPIN is involved in the interaction with p0071, we prepared various Myc-tagged constructs of PAPIN (Fig. 5A). The products of pClneo Myc PAPIN-2 with the first and second PDZ domains and pClneo Myc PAPIN-9 with the second and third PDZ domains bound to the GST fusion protein containing the C terminus of p0071 (Fig. 5B, b and e). The products of pClneo Myc PAPIN-1 with the first PDZ domain and pClneo Myc PAPIN-11 with the third and fourth PDZ domains did not bind (Fig. 5B, a and f). These data suggest that the second PDZ domain of PAPIN is the p0071-binding domain.
Finally, we compared the localizations of p0071 and PAPIN in vivo. The anti-PAPIN antibody recognized a protein with a molecular mass of about 300 kDa in NRK cells (Fig. 6A). In the immunocytostaining of NRK cells, both PAPIN and p0071 were localized at cell-cell contacts (Fig. 6B). Because the message of PAPIN was detected remarkably in the lung, we examined the localization of PAPIN in rat lung. PAPIN and p0071 were detected similarly in bronchial epithelial cells (Fig. 7, A and B).
Both proteins were localized at cell-cell contacts, although the signals were also detected at the apical side of bronchial epithelial cells.

DISCUSSION
In this paper, we have identified a novel multiple PDZ domain-containing protein, PAPIN, through the yeast two-hybrid screening using NPRAP/␦-catenin as a bait. PAPIN interacts with NPRAP/␦-catenin and its ubiquitous isoform, p0071, in vitro and in transfected cells. We could not directly show the interaction of endogenous PAPIN and p0071, because the endogenous p0071 was highly resistant to the detergent extrac-tion and even the detergent/urea extraction. However, because PAPIN and p0071 are localized similarly at cell-cell junctions of NRK cells, both proteins are likely to interact with each other in vivo. The interaction of PAPIN with NPRAP/␦-catenin or p0071 is mediated by the second PDZ domain of PAPIN and the PDZ-binding motif of NPRAP/␦-catenin or p0071. The PDZ domain is a protein-interacting module and was initially reported to bind the C terminus of various proteins (19). However, the PDZ domain of Drosophila INAD protein binds to the PDZ-binding motif of TRP, which is localized inside the sequence (20). The second PDZ domain of PAPIN similarly binds to the PDZ-binding motif of p0071, which is localized inside the sequence.
There are now three reports for the interactions between PDZ domain-containing proteins and armadillo repeat-containing proteins. Adenomatous polyposis coli gene product interacts with PSD-95/SAP90 and SAP97/human Discs-large tumor suppressor gene (36). NPRAP/␦-catenin interacts with synaptic scaffolding molecule (15), and NPRAP/␦-catenin and p0071 bind to PAPIN. Because both the PDZ domain-containing protein family and the armadillo repeat-containing protein family are localized at cell-cell junctions, these interactions may be important in the architecture of cell-cell junctions.
Presenilin 1 is reported to interact with the armadillo repeats of ␤-catenin, p0071, and NPRAP/␦-catenin (37-39). Recent studies have revealed the implication of presenilin 1 in both Notch and Wnt/Wingless pathways (40 -44). Presenilin 1 is localized mainly at the endoplasmic reticulum and Golgi complex (45)(46)(47), but a recent study has revealed that presenilin-1 is recruited to the cell-cell contact through the complex formation with E-cadherin and ␤ catenin (48). We have not examined whether p0071 or NPRAP/␦-catenin interacts simultaneously with PAPIN and presenilin 1. But ligands of PAPIN may form a complex with p0071/NPRAP/␦-catenin and presenilin 1 and may play roles in Notch or Wnt/Wingless pathways.