Index Introduction Materials and Methods Chemicals Cell culture Construction of and transfection with mutant N-cadherin and E-cadherin Immunoprecipitation Cell adhesion and compaction assay Wound healing assay siRNA Other procedures Results MIA2 cells supported strong cell adhesion activity of N-cadherin and epithelial cell-like localization of junctional proteins Binding of p120-catenin to N-cadherin was not required for strong cell adhesion activity or adherens junction formation Binding of p120-catenin to N-cadherin was required for N-cadherin-mediated linear localization of ZO-1 and occludin p120-catenin was involved in the formation of stable junctional structures and directional cell movement in a wound healing assay p120-catenin was involved in the specific localization of junctional proteins and orientation of the centrosome under dynamic conditions in the wound healing assay p120-catenin was essential for E-cadherin-mediated stable localization of ZO-1 and occludin and the cell polarity formation Discussion Acknowledgments References

Introduction

Cadherins are a large group of integral membrane proteins that mediate Ca²⁺-dependent intercellular adhesion. Their extracellular domains are characterized by multiple repeats of the cadherin-specific motif. Many types of cadherins have been found in various tissues of different multicellular organisms (Takeichi, 1995; Angst et al., 2001). Among them, E-cadherin and its homologous proteins were initially identified; and these cadherins were shown to constitute a subgroup of the cadherin family named “classical cadherins.”

Classical cadherins have been extensively investigated. It is well known that these cadherins have profound effects on cell and tissue structure and function (Derycke and Bracke, 2004). Indeed, E-cadherin, the most studied protein as a model of classical cadherins, is required for the formation of cell junctions such as the adherens junction, tight junction, desmosome, and gap junction (Takeichi, 1995; Angst et al., 2001). N-cadherin, another classical cadherin, also forms adherens junction in some tissues. N-cadherin is also involved in synapse formation, axon bundling, and axon guidance (Matsunaga et al., 1988; Volk and Geiger, 1984; Inuzuka et al., 1991; Redies and Takeichi, 1993; Huntley and Benson, 1999; Erdmann et al., 2003; Fu et al., 2006).

The molecular mechanism by which classical cadherins play their various roles is poorly understood. It is thought, however, that the function of classical cadherins in various processes is mainly attributable to their highly specific and strong intercellular adhesion activity; for their extracellular domains mediate the highly specific homophilic interaction, and their well-conserved cytoplasmic domains support strong cell adhesion activity. These cytoplasmic domains are known to associate with several proteins named catenins (Ozawa et al., 1989). Among them, β-catenin directly binds to both the C-terminal side of the cytoplasmic domain of classical cadherins and α-catenin, which associates with the actin cytoskeleton. Thus, classical cadherins and the cytoskeleton establish a stable junctional structure in which β-catenin plays a pivotal role (Drees et al., 2005; Yamada et al., 2005). It is also known that β-catenin functions as a signaling molecule (Bienz, 2005). Plakoglobin, a close homolog of β-catenin, also binds to classical cadherins, which binding can support strong cell adhesion activity without β-catenin (Fukunaga et al., 2005), but the physiological role in the function of classical cadherins has not been established yet (Zhurinsky et al., 2000).

Another catenin, p120-catenin is an interesting protein that binds directly but weakly to the well-conserved juxtamembrane domain of classical cadherins (Ohkubo and Ozawa, 1999). Knockout experiments on p120-catenin have revealed that the protein is also indispensable in higher vertebrates (Perez-Moreno et al., 2006). However, the biological role of p120-catenin seems to be very complex, and its role in the function of classical cadherins is still enigmatic. Hence, the role of p120-catenin has recently attracted the interest of many investigators; and various studies have been carried out (McCrea and Park, 2007). Some researchers reported that p120-catenin is necessary for strong cell adhesion activity (Ohkubo and Ozawa, 1999; Thoreson et al., 2000), but others denied such a function (Murase et al., 2000). During further studies, another role of p120-catenin was recently uncovered: E-cadherin is endocytosed and actively recycled at least under certain conditions (Le et al., 1999), and the dileucine motif (DM) in the juxtamembrane region is essential for the endocytosis of E-cadherin into a cytoplasmic compartment (Miyashita and Ozawa, 2007a, 2007b). Interestingly, occupation of the p120-catenin- and β-catenin-binding sites apparently inhibits the endocytosis of E-cadherin. The molecular mechanism of the function of these proteins remains to be determined. Recently, however, Ishiyama et al. (2010) clarified the 3-dimensional structure of the p120-catenin-E-cadherin cytoplasmic domain complex and discussed the mechanism of E-cadherin endocytosis. In this context, N-cadherin has an interesting feature in that it is not endocytosed well in cultured cells even without the p120-catenin- (Chen et al., 2003) or β-catenin-binding site (Ozawa, 2003), although N-cadherin was reported to show activity-regulated endocytosis in nervous tissues (Tai et al., 2007; Yasuda et al., 2007).

Meanwhile, Davis et al. (2003) reported a possible role of p120-catenin in cell polarity formation in the mouse salivary gland by using a conditional knockout approach. Unfortunately, the results were inconclusive, because the knockout also reduced significantly the amount of E-cadherin on the cell surface, which is required for the formation and maintenance of cell polarity of epithelial cells, by causing endocytosis of the cadherin. Recently, Smalley-Freed et al. (2010) reported knockdown experiments on p120-catenin in cell culture, but the experiments also had the same problem. Meng et al. (2008), on the other hand, suggested that p120-catenin-binding proteins participate in cell polarity formation. It is well known that formation of cell polarity of epithelial cells requires proper localization of junctional proteins such as classical cadherins, occludin, and ZO-1: Occludin forms tight junction strands (Furuse et al., 1996) and ZO-1 constitutes the undercoat of tight junctions (Itoh et al., 1993). However, the role of p120-catenin in the process remains virtually undetermined.

N-cadherin was reported to have a unique feature of strongly affecting cell motility (Derycke and Bracke, 2004). Moreover, Thoreson et al., (2000) reported that N-cadherin promotes cell invasion of tumor cells, and Yanagisawa and Anastasiadis (2006) proposed that p120-catenin is involved in the process. N-cadherin was postulated to modulate Rho GTPase through p120-catenin (Anastasiadis et al., 2000) and to regulate cell motility. On the other hand, p120-catenin was reported to exert a strong effect on cell motility through regulating the cytoskeleton even in the absence of classical cadherins (Ichii and Takeichi, 2007; Boguslavsky et al., 2007). Apparently, p120-catenin seems to have various effects on cell motility as well as other processes.

For elucidating the functions of classical cadherins, various properties of classical cadherins have been studied by using E-cadherin and a fibroblast cell line, the L cell, as a host cell. However, we may have missed some important features of classical cadherins by these studies, since E-cadherin is mainly expressed in epithelial cells, not in fibroblasts. Hence, in the present study we used the epithelial cell line MIA PaCa-2 (MIA2), a pancreatic adenocarcinoma, which does not express classical cadherins, but can support the strong cell adhesion activity of classical cadherins when transfected with them (Fukumoto et al., 2008; Ozaki et al., 2010). In addition, we chose N-cadherin as a model classical cadherin and studied the properties of its cytoplasmic domain by mutagenesis in this study, since N-cadherin is not endocytosed much even in the absence of p120-catenin- or β-catenin-binding site (Chen et al., 2003; Ozawa, 2003), thus providing a simpler experimental system for the study of function of the cytoplasmic domain than that of E-cadherin. We mainly investigated the role of p120-catenin, especially its role in the formation of junctional structures in epithelial cells under static and dynamic conditions. Herein we provide evidence that p120-catenin is essential for classical cadherin-mediated formation of proper junctional structures needed for the establishment of cell polarity in epithelial cells, and discuss the possible physiological role of this catenin.

Materials and Methods

Chemicals

Antibodies against E-cadherin, β-catenin, p120-catenin, and plakoglobin were purchased from BD Transduction Laboratories (Lexington, KY, USA); and those against claudin-1, N-cadherin, and ZO-1 were from Zymed (South San Francisco, CA, USA). Antibodies against α-catenin, α-tubulin, γ-tubulin, and occludin were obtained from Sigma-Aldrich (St. Louis, MO, USA). Antibody against HA tag was purchased from MBL (Nagoya, Japan). Anti-mouse and anti-rabbit antibodies conjugated with Alexa Fluor-488 or Alexa Fluor-568 and Alexa-568 conjugated phalloidin were obtained from Molecular Probe (Eugene, OR, USA). Anti-mouse IgG and anti-rabbit IgG conjugated Dynabeads, Stealth RNAi, Lipofectamine RNAiMAX, and pEF1/Myc-His were the products of Invitrogen (Carlsbad, CA, USA). Alkaline phosphatase-conjugated anti-mouse and anti-rabbit antibodies were from Promega (Madison, WI, USA). DAPI was obtained from Wako Chemical Co (Osaka, Japan); and Matrigel, from BD Biosciences (NJ, USA).

Cell culture

MIA2 cells were cultured in Dulbecco’s modified Eagle’s medium-F12 (1:1) containing 10% fetal bovine serum in 5% CO₂ atmosphere at 37°C. For immunostaining, the cells were cultured on cover glasses or in collagen gel. For time-lapse microscopy, cells were cultured on glass-bottomed culture dishes.

Construction of and transfection with mutant N-cadherin and E-cadherin

Constructs of human N-cadherin and E-cadherin mutants were made by PCR, which was carried out by using human N-cadherin or E-cadherin cDNA as a template and the appropriate primers. The primers used were as follows: Ncad-s, GGAAGGGAAGGGGGTGG; Ncad-middle-s, GCAACCGTGTCTGTTAC; Ncad-as, ACTAGTGTCATCACCTCCACCAT; Ncad/CP97-as, ACTAGTAGCCGCTTTAAGGCCCTC; Ncad/CP0-as, ACTATGCCGGCGTTTCATCCATAC; Ncad/Δp120-s, TATTTTAAAACAGCAGCCTGACACTG; Ncad/Δp120-as, CAGGCTGCTGTTTTAAAATATTATCTCTTACAT; Ncad/p120A-s, CTGGGCCGCTGCTCCTCCACCTTCTTCATC; Ncad/p120A-as, GGAGCAGCGGCCCAGGACTATGACTTGAGC; Ecad-s, GAATTCCGGAACTGCAAAGCACCT; Ecad-middle-s, CTACACTGCCCAGGAGC; Ecad-as, GCGGCCGCTCAGTCATCACCTCCAC; Ecad/CP81-as, GCGGCCGCTCAAGCCGCTTTCAGATTTTC; Ecad/p120A-s, GGAGCAGCGGCCCAGGACTTTGACTTGAGC; Ecad/p120A-as, CTGGGCCGCTGCTCCGCCTCCTTCTTCATCATA; Ecad/ΔDM-s, GGAGAGCGCCCCCAGAGGATGACA; Ecad/ΔDM-as, TCTGGGGGCGCTCTCCTCCGAAGAAA. The constructs prepared in this study were as follow: Ncad/p120A, N-cadherin with the EED sequence (amino acids 621–623) in its p120-catenin-binding region substituted by AAA; Ncad/CP97, N-cadherin lacking the amino acid sequence (685–747) in the C-terminal half region of the cytoplasmic domain; Ncad/CP0, N-cadherin lacking the entire cytoplasmic domain (amino acid 588–747). HA-tag was attached to the C-termini of these N-cadherin constructs. E-cadherin mutants corresponding to the above N-cadherin mutants were described previously (Ozaki et al., 2010). Amino acid sequences of cadherins are numbered from the N-terminus of the mature molecules in this study. The prepared constructs are schematically shown in Fig. 1. The resultant cDNAs were subcloned into the expression vector pEF1 by using appropriate restriction sites. Then, MIA2 cells were transfected with the construct DNAs, and stable clones were obtained by G418 screening.

View Details

Fig. 1. Constructs of N- and E-cadherin mutants prepared. Constructs of N- and E-cadherin mutants are schematically drawn. β-cat, β-catenin-binding region; p120, p120-catenin-binding region.

Immunoprecipitation

Cultured cells were solubilized for 20 min in 1% NP-40 in RIPA buffer containing protease inhibitors (Roche Diagnostics, Indianapolis, IN, USA) and were centrifuged at 25,000×g for 20 min at 4°C. The resultant solution was reacted with an appropriate antibody conjugated with anti-rabbit IgG- or anti-mouse IgG-Dynabeads for 2 hr with constant agitation. Then, the immunocomplex was collected and washed 3 times with RIPA buffer. Finally, the resultant Dynabeads were mixed with SDS-PAGE sample buffer and subjected to SDS-PAGE.

Cell adhesion and compaction assay

Cell aggregation assays were carried out as described previously (Hirano et al., 1999). During the assays, some of the cell aggregates were taken and put on Matrigel-coated cover glasses and incubated in a CO₂ incubator for up to 12 hr to examine their compaction activity. The resultant cells were fixed with paraformaldehyde or cold methanol and examined with a Nikon Eclipse TE2000-U fluorescence microscope equipped with a differential interference contrast apparatus (Nikon, Tokyo, Japan).

Wound healing assay

Wound healing assays were carried out according to the method described by Nakao et al. (2008), and the cell motility was analyzed by time-lapse microscopy with the Nikon Eclipse TE2000-U microscope. The speed of directional cell movement was determined by measuring the distance between the moving-cell front and the original position of these cells (indicated by the scratch line) divided by the length of time of the assay. The speed of cell protrusion extension was measured by the length of elongation divided by the time required. The location of centrosomes in the cells was described by the angle between the line drawn along the edge of the scratch on the cell sheet and the line that passed through the centrosome and the center of the nucleus. The speed of general cell movement of a given single cell was measured as follows: the movement of the center of the nucleus was traced by time-lapse microscopy, and the total distance the cell moved was divided by the time required for the movement.

siRNA

For p120-catenin knockdown, MIA2 cells were transfected with p120-catenin-targeted Stealth siRNAs. The transfection was performed according to the manufacturer’s protocol. Briefly, 10 pmol of RNAi duplex was mixed with 1 μl of Lipofectamine RNAiMAX in 100 μl of Opti-MEM medium without serum (Gibco) and the resultant solution was added in a well of a 24 well plate containing 25,000 cells in 500 μl of culture medium. The following siRNAs were used: sip120-1, 5'-GGCUAGAGGAUGACCAGCGUAGUAU-3'; sip120-2, 5'-GCAGCUCCCAAUGUUGCCAACAAUA-3'. Negative control Stealth siRNAs were obtained form Invitrogen. After 5 day of incubation, the expression of p120-catenin was analyzed by immunostaining.

Other procedures

Immunofluorescence staining, electron microscopy, SDS-PAGE, and immunoblot analysis were carried out as described previously (Ozaki et al., 2010). Statistical analyses were performed by using Student’s t-test. A P-value of less than 0.05 was considered significant.

Results

MIA2 cells supported strong cell adhesion activity of N-cadherin and epithelial cell-like localization of junctional proteins

MIA2 cells showed a round cell shape and did not express appreciable amount of classical cadherins (Fig. 2A). Indeed, MIA2 cells did not exhibit significant activity of Ca²⁺-dependent cell aggregation or cell compaction (Fig. 2B and Fig. 3). When wild type N-cadherin (Ncad(WT)) was expressed in these cells, however, the transfectants displayed an epithelial morphology (Fig. 2B) and strong activities of Ca²⁺-dependent cell aggregation and cell compaction (Fig. 2B and Fig. 3).

View Details

Fig. 2. Expression of Ncad(WT) in MIA2 cells. (A) Immunoblot analysis of expression of Ncad(WT) and catenins in MIA2 cells. MIA2 cells did not express appreciable amounts of classical cadherins, but catenins, ZO-1, and occludin (occ) were expressed. After expression of Ncad(WT) by transfection, the expression levels of catenins except for p120-catenin (p120) significantly increased. IB, immunoblot analysis. (B) Cell shape, cell compaction activity, and N-cadherin staining of MIA2 Ncad(WT) transfectants. After expression of Ncad(WT), the transfectants showed an epithelial morphology in contrast to the round shape of the parental MIA2 cells. The insets in the panels of cell morphology indicate cell compaction activity. Ncad(WT) was diffusely localized at the intercellular contact sites (arrows), but it also showed linear localization in some areas (arrowhead). (C) Staining for β-catenin, plakoglobin, p120-catenin, ZO-1, and occludin in MIA2 Ncad(WT) transfectants. Immunofluorescence staining showed that β-catenin (β-cat), plakoglobin (PG), and p120-catenin (p120) were diffusely localized throughout the lateral membrane (arrows). ZO-1 and occludin (occ) showed linear localization there (arrowheads), but claudin-1 (cla-1) was not expressed. (D) Staining of ZO-1 and occludin in MIA2 cells. Immunostaining showed that ZO-1 and occludin (occ) were sporadically stained (arrowheads).

View Details

Fig. 3. Cell aggregation activity of MIA2 transfectants expressing mutant cadherins. Cell aggregation activity was measured by the conventional method. The results are described by No-Nt/No, where No and Nt represent the initial cell-aggregate number and the final aggregate number after incubation, respectively. *=p<0.05.

Immunoblot analysis indicated that the expression levels of catenins except for p120-catenin significantly increased after Ncad(WT) expression (Fig. 2A). The bands of p120-catenin became broader and showed slightly slower mobility than those of parental MIA2 cells in the immunoblot analysis, as Fukumoto et al. (2008) had earlier described.

Immunofluorescence staining revealed diffuse localization of Ncad(WT) and catenins at cell-cell contact sites in the transfectants (Fig. 2B and 2C). Moreover, Ncad(WT) showed a linear localization in some area. Meanwhile, tight junction proteins ZO-1 and occludin showed a linear localization in Ncad(WT) transfectant (Fig. 2C), whereas they sporadically formed short clusters in MIA2 cells (Fig. 2D). Close examination of the double staining of Ncad(WT) and ZO-1 or occludin suggested that ZO-1 and occludin were possibly located at the subapical region of the lateral membrane.

Binding of p120-catenin to N-cadherin was not required for strong cell adhesion activity or adherens junction formation

To clarify the role of p120-catenin in classical cadherin function, we replaced the amino acid sequence EED in the p120-binding region of N-cadherin (amino acid 621–623) with the AAA sequence (Ncad/p120A) and expressed this construct in MIA2 cells (Fig. 4A). The resultant transfectant showed slightly irregular epithelial morphology and strong cell adhesion and cell compaction activities (Fig. 3 and Fig. 4C), as had been reported for the corresponding mutant of E-cadherin (Ozaki et al., 2010). The N-cadherin mutant lacking the p120-catenin-binding site (Ncad/Δp120) gave similar results (unpublished observation). The transfectant with its β-catenin-binding region deleted (Ncad/CP97) and that with its entire cytoplasmic domain deleted (Ncad/CP0) showed weak but significant activity of cell aggregation (Fig. 3), but no cell compaction activity (Fig. 4C). These transfectants showed a round cell shape similar to that of parental MIA2 cells. Immunoprecipitation experiments confirmed that Ncad/p120A associated with β-catenin but not with p120-catenin. On the other hand, Ncad/CP97 associated with p120-catenin, showing a broad band in immunoblot analysis, but not with β-catenin; whereas Ncad/CP0 did not associate with p120-catenin or β-catenin (Fig. 4B).

View Details

Fig. 4. Properties of N-cadherin with mutated p120-catenin-binding site. (A) Immunoblot analysis of expression of N-cadherin and p120-catenin in N-cadherin mutant transfectants of MIA2 cells. IB, immunoblot analysis. (B) Immunoprecipitation of N-cadherin in the mutant transfectants. Ncad/p120A was co-precipitated with β-catenin (β-cat) but not with p120-catenin (p120), whereas Ncad/CP97 was co-precipitated with p120-catenin but not with β-catenin. IP, immunoprecipitation. (C) Cell shape and cell compaction activity of MIA2 transfectants of Ncad/p120A, Ncad/CP97, and Ncad/CP0. The Ncad/p120A transfectant showed irregular epithelial morphology and strong cell compaction activity (insets in the panels of cell morphology). In contrast, Ncad/CP97 and Ncad/CP0 transfectants had a fibroblastic shape and did not exhibit strong cell compaction activity. (D) Electron microscopy of MIA2 transfectants of Ncad/p120A, Ncad/CP97, Ncad/CP0, and Ncad(WT). MIA2 transfectants expressing various N-cadherin mutants and parental MIA2 cells were examined by electron microscopy for the formation of cell junctions. The Ncad(WT) and Ncad/p120A transfectants formed adherens junctions or adherens junction-like structures (asterisks), whereas the Ncad/CP97, Ncad/CP0 or parental MIA2 transfectant did not.

Electron microscopy revealed that the Ncad/p120A transfectant formed adherens junctions or adherens junction-like structures as the Ncad(WT) transfectant did, but the location was not restricted to the subapical region. Ncad/CP97 or Ncad/CP0 transfectant did not form these junctions (Fig. 4D).

Binding of p120-catenin to N-cadherin was required for N-cadherin-mediated linear localization of ZO-1 and occludin

In contrast to the case of E-cadherin (Miyashita and Ozawa, 2007b), the N-cadherin with the mutated p120-catenin-binding site (Ncad/p120A) mostly remained in the lateral membrane, as described previously (Fig. 5A), even though it still retained the DM in the juxtamembrane region (Davis et al., 2003). The scattered punctate localization of Ncad/p120A at cell-cell contact sites was similar to but broader than that of Ncad(WT) (Fig. 5A), suggesting that p120-catenin is involved in the formation of stable adherens junction or apico-basal polarity of epithelial cells. Then, we examined the effect of p120-catenin on the localization of ZO-1 and occludin (Fig. 5A). In MIA2 Ncad(WT) transfectant, ZO-1 and occludin were linearly concentrated in the lateral membrane, possibly at the subapical region as described above (Fig. 2C). In contrast, in the Ncad/p120A transfectant, ZO-1 and occludin sporadically formed short clusters as the case of MIA2 cells (Fig. 5A), indicating that p120-catenin was involved in the N-cadherin-mediated linear localization of these tight junction proteins.

View Details

Fig. 5. Localization of ZO-1 and occludin in transfectants expressing various N-cadherin mutants. (A) Localization of N-cadherin mutants, p120-catenin, ZO-1, and occludin. In contrast to the case of Ncad(WT) transfectant, ZO-1 or occludin (occ) only showed short sporadic accumulation in the lateral membrane in Ncad/p120A, Ncad/CP97, and Ncad/CP0 transfectants (arrowheads); whereas N-cadherin mutants (Ncad) showed weak accumulation at the cell-cell contact sites (arrows). p120-catenin (p120) was localized at intercellular contact sites of Ncad/CP97 transfectant (arrows) but not those of other transfectants. (B) Knockdown of p120-catenin by siRNA. MIA2 cells expressing Ncad(WT) were transfected with p120-catenin siRNA. The resultant cells revealed fibroblast-like cell shape. The localization of ZO-1 as well as occludin (occ) became weak at some but not all intercellular contact sites (asterisks), whereas N-cadherin (HA) localization was not affected much.

Ncad/CP97 was co-localized with p120-catenin at intercellular contact sites in the transfectant, but ZO-1 and occludin sporadically formed short clusters (Fig. 5A). In the Ncad/CP0 transfectant, the localization of ZO-1 and occludin was essentially the same as that in the Ncad/CP97 transfectant and MIA2 cells.

Next, knockdown experiments of p120-catenin were carried out by an siRNA method (Meng et al., 2008) to examine the role of p120-catenin in the localization process by another approach. In the resultant transfectant expressing N-cadherin, ZO-1 as well as p120-catenin showed weak or no localization at some but not all intercellular contact sites, whereas N-cadherin localization was not affected much (Fig. 5B), suggesting that p120-catenin is essential for N-cadherin-mediated linear localization of ZO-1 and occludin. The transfectant showed fibroblast-like cell shape which was different from that of Ncad/p120A. The knockdown might affect the cell shape, since p120-catenin plays multiple roles.

p120-catenin was involved in the formation of stable junctional structures and directional cell movement in a wound healing assay

Since the above experiments suggested that p120-catenin was essential for the formation of stable junctional structures, the effect of p120-catenin binding to N-cadherin on the formation of junctional structures was further examined under dynamic conditions of cell movement in a wound healing assay and was analyzed by time-lapse microscopy (Fig. 6 and Supplementary Movie No. 1–3). Moreover, the effect on cell movement was also examined in the wound healing assay. Ncad(WT) transfectant actively moved toward the scratched open area after a short time lag of about 30 min, roughly in the direction perpendicular to the scratched edge of the cell sheet (Fig. 6A and Supplementary Movie No. 1). In contrast, the Ncad/p120A transfectant showed frequent changes in cell shape, irregular cell movement, and cell escape from the cell sheet, and significantly slower directional cell movement than did the Ncad(WT) transfectant (Fig. 6A and 6B). These results clearly indicate that Ncad/p120A transfectant formed apparently unstable cell junctions. Meanwhile, the MIA2 transfectants of Ncad/CP97 and Ncad/CP0 as well as parental MIA2 cells did not show appreciable directional cell movement, although these cells actively moved in various directions (Fig. 6A and Supplementary Movie No. 3). Interestingly, the speed of the general cell movement of these cells was comparable to that of the directional cell movement of Ncad(WT) transfectant (data not shown), suggesting that p120-catenin was important for the directionality of cell movement.

View Details

Fig. 6. Directional cell movement of N-cadherin mutants in the wound healing assay. (A) Cell movement of transfectant cells expressing different N-cadherin mutants in the wound healing assay. MIA2 N-cadherin mutant transfectants were subjected to the wound healing assay and examined for cell movement by time-lapse microscopy. The Ncad(WT) transfectant formed relatively long protrusions and showed directional cell movement, whereas the Ncad/p120A one formed short protrusions (arrow and arrowhead) and showed compromised directional cell movement. The MIA2 Ncad/CP97 and Ncad/CP0 transfectants or parental MIA2 cells did not show appreciable directional cell movement. (B) Directional cell movement of various N-cadherin mutant transfectants in the wound healing assay. The Ncad(WT) transfectant showed significantly faster directional cell movement than the Ncad/p120A one. (C) Maximum length of cell protrusion of various transfectants. The MIA2 Ncad(WT) transfectant formed the longest cell protrusion among the various transfectants. (D) Extension speed of cell protrusion of various transfectants. The transfectants expressing the various N-cadherin mutants showed similar extension speeds of their cell protrusions. *=p<0.05.

The Ncad(WT) transfectant formed relatively long filopodia-like cell protrusions (Fig. 6A and 6C). Meanwhile, MIA2 cells transfected with Ncad/p120A actively extended filopodia-like cell protrusions, but these protrusions often collapsed compared with those of the Ncad(WT) transfectant, resulting in the formation of shorter protrusions. Moreover, the direction of the extension varied frequently (Fig. 6A and Supplementary Movie No. 2). Thus, the Ncad/p120A transfectant seemed to show slower directional cell movement. The Ncad/CP97 and Ncad/CP0 transfectants as well as the parental MIA2 cells formed short and round cell protrusions (Fig. 6A and 6C).

The above transfectants and the parental MIA2 cells showed similar speeds of protrusion extension (Fig. 6D), suggesting again that regulation of cell protrusions was quite important for the directional cell movement.

p120-catenin was involved in the specific localization of junctional proteins and orientation of the centrosome under dynamic conditions in the wound healing assay

Next, we examined the localization of N-cadherin and ZO-1 in the moving cells at the scratched edge in the wound healing assay. In the Ncad(WT) transfectant, N-cadherin and ZO-1 showed clear localization at intercellular contact sites, whereas in the Ncad/p120A transfectant, the mutant N-cadherin showed diffuse localization, and ZO-1 did not show clear clustering (Fig. 7A). We also examined the cytoskeleton, since the cell shape change and cell movement of the transfectants of various N-cadherin mutants in the wound healing assay suggested that the cytoskeleton was compromised by the disruption of the p120-catenin-binding site (Fig. 6). The results indicated that the microtubules in the cell protrusions of the Ncad/p120A transfectant were sparse compared with those in the Ncad(WT) transfectant, whereas the actin filaments in the former were only slightly disorganized (Fig. 7A).

View Details

Fig. 7. Staining for N-cadherin, ZO-1, F-actin, and α-tubulin in the cells at the scratched edge in the wound healing assay. (A) Localization of N-cadherin, ZO-1, α-tubulin, and F-actin. Ncad(WT) in the cells moving into the scratched area was clearly localized at intercellular contact sites, whereas Ncad/p120A in the transfectant showed a diffuse localization at a comparable site (Ncad, arrow). Microtubules (α-tubulin) in the protrusions of the Ncad/p120A transfectant were decreased compared with those in the protrusions of the Ncad(WT) transfectant, whereas actin filaments (F-actin) in the protrusions of the former were only slightly disorganized (arrowheads). (B) Staining for γ-tubulin. The location of γ-tubulin in the Ncad(WT)-transfected MIA2 cells in the cell sheet was random (control), but the location in the cells at the scratched edge was mostly between the nuclei and the leading edge of the cells after the incubation (3 hr, arrows). In contrast, the orientation of γ-tubulin did not change much in the MIA2 Ncad/p120A and Ncad/CP97 transfectants (arrowheads).

Then, we characterized the localization of the centrosomes (Fig. 7B), since Dupin et al. (2009) recently reported that classical cadherins determined the location of centrosomes in various types of cells. The position of γ-tubulin, a marker of centrosomes, in the cells located in the inner region of the cell sheets did not show any special orientation (Fig. 7B, control). However, γ-tubulin in Ncad(WT) transfectants at the edge of the scratched cell sheet became localized mostly between the nuclei and the leading edge of the cells after several hours of incubation. The angle between the edge line of the cell sheet and the line that passed through the γ-tubulin site and center of the nucleus was 78°±9. In contrast, γ-tubulin in Ncad/p120A transfectants at a comparable scratched edge showed less clear orientation (41°±25). When the Ncad/CP97 transfectant was examined, no specific orientation of the centrosomes was obtained. These results suggest that p120-catenin was essential for the formation of proper junctional structures needed for the specific localization of centrosomes in the cells at the edge of the scratched cell sheet, thereby for the formation of cell polarity and directional cell movement.

p120-catenin was essential for E-cadherin-mediated stable localization of ZO-1 and occludin and the cell polarity formation

The above experiments indicated that p120-catenin was involved in the formation of stable cell junctions, hence the formation of cell polarity, in the Ncad(WT) transfectant. Then, we examined whether the effect of p120-catenin would also be observed in the wild-type E-cadherin (Ecad(WT)) transfectant. In this experiment, Ecad/p120A/ΔDM transfectant was used, since Ecad/p120A containing DM was mostly endocytosed in the transfectant (Miyashita and Ozawa, 2007b). The results showed that the effect of p120-catenin was not specific to N-cadherin: ZO-1 and occludin were linearly localized at a possible subapical region of the lateral membrane in the Ecad(WT) transfectant, whereas these proteins were not well localized in the lateral membrane in the Ecad/p120A/ΔDM transfectant (Fig. 8A).

View Details

Fig. 8. E-cadherin-mediated localization of junctional proteins and directional cell movement. (A) Localization of junctional proteins in Ecad(WT) transfectants. In the Ecad(WT) transfectant, E-cadherin (Ecad), p120-catenin (p120), ZO-1, and occludin (occ) were localized at cell-cell contact sites (arrows and arrowheads). In contrast, ZO-1 or occludin showed sporadic localization at the cell-cell contact sites in the Ecad/p120A/ΔDM transfectant, whereas the mutant E-cadherin showed diffuse localization (arrows). (B) Directional cell movement of E-cadherin transfectants in the wound healing assay. Ecad(WT) transfectant showed directional cell movement, although the speed was much slower than that of the Ncad(WT) one, whereas Ecad/p120A/ΔDM-transfected cells showed compromised activity. The protrusions of Ecad/p120A/ΔDM were unstable (arrow and arrowhead). (C) Staining for γ-tubulin. The location of γ-tubulin in the Ecad(WT)-transfected MIA2 cells at the scratched edge was mostly between the nuclei and the leading edge of the cells after the incubation (arrows). In contrast, the location of γ-tubulin in the Ecad/p120A/ΔDM-transfected MIA2 cells did not show specific orientation (arrowheads).

The Ecad(WT) transfectant also showed directional cell movement in the wound healing assay, although the speed was significantly slower than that of Ncad(WT) transfectant; and the former displayed relatively short cell protrusions (Fig. 8B). In contrast, Ecad/p120A/ΔDM transfectants did not show good directional cell movement. Moreover, centrosomes in Ecad(WT) transfectant at the scratched sites showed the specific orientation at the leading edge as observed in the Ncad(WT) transfectant, whereas those in the Ecad/p120A/ΔDM transfectant did not (Fig. 8C). These results suggest that p120-catenin was also required for E-cadherin-mediated formation of proper cell junctions and cell polarity.

Discussion

This study was carried out based on the assumption that unknown features of classical cadherins may be uncovered by the examination of N-cadherin properties in epithelial cells, since most of the properties of classical cadherins have been clarified by using the combination of E-cadherin and the L cell fibroblast cell line. The present results support this assumption at least partially and also revealed some novel properties.

Among the various catenins, the p120-catenin has a unique feature: its binding to classical cadherins is relatively weak. Therefore, p120-catenin seems to play a regulatory role rather than a role of physical linkage. However, some previous results suggested the involvement of p120-catenin in strong cell adhesion. This conclusion needs to be examined further, since those previous studies did not take into account the endocytosis of E-cadherin, which significantly affected the amount of E-cadherin on cell surface and thereby the degree of cell adhesion activity. Indeed, the present results would indicate that the binding of p120-catenin to N-cadherin was not directly required for strong cell adhesion activity or the formation of adherens junction (Fig. 3 and Fig. 4), and that p120-catenin seems to indirectly affect the cell adhesion activity through endocytosis.

Previous studies using a conditional knockout approach for p120-catenin in vivo and knockdown experiments for p120-catenin in cell culture (Davis et al., 2003; Smalley-Freed et al., 2010) suggested a possible role of p120-catenin in the establishment of cell polarity. Unfortunately, those results were inconclusive because the treatments also reduced the level of E-cadherin on the cell surface due to endocytosis. Hence, we investigated the role of p120-catenin by using new experimental setups in the present study. Our results showed that p120-catenin was essential for the N-cadherin-mediated stable accumulation and linear localization of ZO-1 and occludin in the lateral membrane as well as for the compact localization of N-cadherin in the lateral membrane, thereby the establishment of cell polarity. Disruption of the p120-catenin binding site of N-cadherin resulted in the loss of linear localization of ZO-1 and occludin in the lateral membrane as well as in broader localization of N-cadherin in the lateral membrane. Moreover, similar results were obtained when p120-catenin knockdown experiments were carried out. The molecular mechanism of the linear localization of ZO-1 and occludin in the lateral membrane of the Ncad(WT) transfectants is unclear at present. However, the localization required strong cell adhesion activity of classical cadherins as already reported (Itoh et al., 1993), although ZO-1 and occludin can cluster sporadically without classical cadherins, and p120-catenin apparently stabilized the cell junctions in the process. Weak cell adhesion activity of Ncad/PC97 did not support the linear localization of ZO-1 and occludin, although the mutant N-cadherin retained the p120-catenin binding activity. One explanation may be that strong and stable cell-cell adhesion by classical cadherins stabilizes the intercellular contact sites to form stable localization of ZO-1 and occludin. Indeed, the time-lapse microscopy suggested that Ncad/p120A transfectant formed unstable cell-cell adhesions (Supplementary Movie No. 2).

In contrast to our results, however, Smalley-Freed et al. (2010) reported that knockout or knockdown of p120-catenin resulted in the down regulation of E-cadherin and breakage of adherens junction as predicted; but unexpectedly ZO-1 and occludin were still localized at cell-cell contact sites. We do not have a good explanation as to why the tight junction proteins were localized without adherens junction in the experiments, in contrast to the general notion that E-cadherin is required for the localization (Itoh et al., 1993).

It is known that p120-catenin interacts with microtubules, and Meng et al. (2008) recently reported a study on p120-binding proteins. Interestingly, their results were slightly different from our present results. They showed that the knockdown of p120-binding proteins that also bind to microtubules affected the localization of E-cadherin, but not the localization of ZO-1. In contrast, our results indicated that p120-catenin was essential for the clustering and linear localization of the tight junction proteins. The activity of p120-catenin that we found and the activity of p120-catenin-binding proteins that Meng et al. (2008) identified may participate in different processes, since p120-catenin appears to have various activities (McCrea and Park, 2007).

Phosphorylation, which is well known to regulate various processes, has been suggested to be involved in the function of p120-catenin, and p120-catenin is indeed phosphorylated in classical cadherin transfectants (Alema and Salvatore, 2006). However, no convincing evidence for this hypothesis is available.

The present study indicated that the expression of N-cadherin in epithelial MIA2 cells resulted in increased directional cell movement in the wound healing assay, which finding appears to be consistent with the notion that expression of N-cadherin increases the cell movement of various types of cells (Derycke and Bracke, 2004; Yanagisawa and Anastasiadis, 2006). Closer examination of the present results, however, revealed that N-cadherin expression increased the directional cell movement, but did not generally stimulate cell movement of different cells in the present experiments. Indeed, the speed of random cell movement of MIA2 cells (13±6.4 μm/hr) was similar to that of the directional cell movement of the Ncad(WT) transfectants (Fig. 6A). It may be necessary to examine more precisely the effect of N-cadherin on cell movement. In fact, the previous studies examined the cell movement without differentiating these two types of cell movement.

Moreover, the present results obtained by using our new experimental setups indicated that p120-catenin was required for the N-cadherin-mediated directional cell movement, but not for general cell movement (Yanagisawa and Anastasiadis, 2006). Since Dupin et al. (2009) recently showed that classical cadherins determine the polarity of cell orientation by regulating the position of the centrosomes, we hypothesized that p120-catenin participated in the process. Indeed, further study showed that p120-catenin was involved in the specification of the centrosome location and regulation of the directional cell movement (Fig. 7), suggesting that p120-catenin was essential for the N-cadherin-mediated formation of cell polarity in the process, and thereby for directional cell movement.

In light of the above, it would be interesting to determine whether p120-catenin is directly involved in the classical cadherin-mediated formation of other types of cell polarity, such as apico-basal polarity, through specifying the centrosome location. Unfortunately, we cannot directly answer this question using MIA2 cells, since these cells are very flat with a thin cell height and only form partially-developed apico-basal polarity. This hypothesis, however, seems possible, for ZO-1 and occludin showed tight junction-like linear localization at specific sites in MIA2 transfectants of N-cadherin and E-cadherin, although it was not complete. N- and E-cadherins showed linear localization in some areas, as do epithelial cells. Importantly, disruption of p120-catenin-binding site abolished the localization and knockdown of p120-catenin by an siRNA approach resulted in the decrease in the localization of ZO-1 and occludin at intercellular sites. Thus, p120-catenin seems to be essential for the classical cadherin-mediated formation of apico-basal polarity as well. Obviously, the present results are far from complete, but should provide a basis for further research on the role of p120-catenin in the function of classical cadherins in epithelial cells, which function is indispensable for various processes.

Acknowledgments

This study was supported in part by grants-in-aid from the Ministry of Education, Culture, and Technology (Advanced Program of High-profile Research grant to Department of Bioscience, KAKENHI 18659055 to S.O.); the Ministry of Health, Labour, and Welfare (H21-G-209 to S.T.S.); and from Kwansei Gakuin University (to S.T.S.). We wsih to thank to Drs. S. Hirano (Kochi University) and K. Owaribe (Nagoya University) for their kind technical assistance and helpful discussions.