The 293 cells were provided by Dr Tatsuhiko Furukawa (Department of Molecular Oncology, Kagoshima University, Kagoshima, Japan). The cells were cultured in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal calf serum. DLD1, a human colorectal adenocarcinoma cell line, was provided by Dr Shintaro T. Suzuki (Kwansei Gakuin University, Kobe, Japan). The PKE cells, provided by Dr Takami Matsuyama (Department of Immunology, Kagoshima University), were cells primarily cultured from the kidneys of Clawn miniature swine [a swine strain established by Japanese scientists (21)] and passaged 10–15 generations. Madin-Darby canine kidney (MDCK) cells were provided by Dr Yasushi Daikuhara (Kagoshima University Dental School, Kagoshima, Japan). The expression vector encoding hemagglutinin (HA)-tagged full-length E-cadherin was prepared as previously described (22). The vector contains neo gene, which confers G418-resistance. As a control, an empty vector without E-cadherin gene was used yielding nH-2 and nH-6 clones. Transfection of 293 cells with the HA-tagged E-cadherin vector resulted in EH-5 and EH-13 clones. The cells (5×10⁵) were transfected with the expression vector (10 µg) using the calcium-phosphate transfection method as previously described (15), and stably transfected cells were selected in G418-containing medium. Isolated G418-resistant clones were tested for the expression of the transfected construct by immunofluorescence microscopy and immunoblot analysis.

Mouse monoclonal antibodies (mAbs) against E-cadherin (catalogue no. 610182), β-catenin (catalogue no. 610153), fibronectin (catalogue no. 610077) and plakoglobin (γ-catenin; catalogue no. 610253) were obtained from BD Transduction Laboratories (Lexington, KY, USA). Pan-cadherin mAb (catalogue no. C1821-100UL) was obtained from Sigma-Aldrich (St. Louis, MO, USA). Rabbit polyclonal anti-occludin (catalogue no. 71-1500), anti-ZO-1 (catalogue no. 61-7300) and mouse monoclonal anti-vimentin (catalogue no. 18-0052) antibodies were purchased from Zymed Laboratories (South San Francisco, CA, USA). A mouse mAb specific for cytokeratins 5/8 (catalogue no. MAB3228) was acquired from Merk Millipore Ltd. (Tokyo, Japan). A mouse mAb against desmoglein 2 was obtained from Progen Biotechnik GmbH (Heidelberg, Germany). A rat mAb (3F10; catalogue no. 11867423001) directed against HA was purchased from Roche Applied Science (Mannheim, Germany). All secondary antibodies (fluorescein-, rhodamine- and peroxidase-conjugated) were obtained from Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA, USA).

Immunoblot analysis was performed essentially as previously described (22). Briefly, the cells were lysed by boiling in SDS sample buffer for 5 min. Proteins (30–50 µg) were separated by SDS-PAGE and then transferred onto nitrocellulose membranes (Schleicher & Schuell, Keene, NH, USA). The membranes were incubated with the appropriate primary antibodies diluted at 1:1,000, followed by incubation in horseradish peroxidase-conjugated secondary anti-mouse (catalogue no. 315-036-045) or anti-rabbit (catalogue no. 111-036-045) antibodies (Jackson ImmunoResearch Laboratories, Inc.) diluted at 1:1,000. Proteins were visualized using the enhanced chemiluminescence (ECL) system (Amersham Pharmacia Biotech, Piscataway, NJ, USA).

Immunofluorescence staining of the cells was performed as previously described (23) with some modifications. The cells were cultured on coverslips for 48 h prior to fixation. The cells were then fixed with 3% paraformaldehyde in phosphate-buffered saline (PBS) for 20 min at room temperature. Following 3 washes in PBS containing 50 mM NH₄Cl, the cells were permeabilized using 0.1% Triton X-100 in PBS for 5 min. After washing in PBS, the cells were soaked in blocking solution (PBS containing 5% fetal calf serum) for 15 min, and then incubated for 30 min with primary antibodies diluted in blocking solution. After washing 3 times in PBS, the cells were incubated with rhodamine- or fluorescein-conjugated secondary antibodies. The cells were analyzed as previously described (24) using an Olympus microscope.

E-cadherin is a major cell adhesion molecule in epithelial cells and an epithelial cell marker (25). Although the immunofluorescence staining of MDCK cells, a typical epithelial cell line, with E-cadherin antibodies revealed strong membrane staining, the PKE and the 293 cells exhibited no membrane staining (Fig. 1A, upper panels). In a control experiment, we noted clear membrane staining for β-catenin, a cytoplasmic subunit of the transmembrane cell adhesion cadherin complex, in these 3 cell lines (Fig. 1A, lower panels).

The immunoblot detection of E-cadherin in these cells revealed that E-cadherin expression in the PKE and the 293 cells was markedly decreased compared with its expression level in the MDCK cells (Fig. 1B).

The immunoblot detection of N-cadherin in the PKE and the 293 cells revealed high expression levels of N-cadherin (Fig. 2A). The MDCK cells, on the other hand, did not express N-cadherin (Fig. 2A), which was in agreement with previous findings (26). Immunoblot analysis using pan-cadherin antibodies raised against a synthetic peptide corresponding to the C-terminal amino acids of chicken N-cadherin, revealed that the PKE and the 293 cells expressed a protein with the same electrophoretic mobility as N-cadherin (Fig. 2A). Although the MDCK cells do not express N-cadherin, they express a protein that migrates slowly relative to N-cadherin. This protein has been shown to be K-cadherin (23,27). The 293 cells also expressed a protein with a slightly slower electrophoretic mobility (Fig. 2A). Although this protein has not been identified definitively, we conjecture that this protein is the precursor form of N-cadherin, containing the prosequence.

The immunofluorescence staining of the cells with pan-cadherin antibodies revealed clear membrane staining (Fig. 2B). Thus, the proteins recognized by the pan-cadherin antibodies were present on the membrane. No heterogeneity of expression within the cell lines was noted. These results are consistent with a previous observation that N-cadherin is expressed endogenously at cell-cell contact sites in 293 cells (28).

The MDCK cells exhibited strong and marked cytoplasmic and filamentous staining with a pan-keratin antibody mix containing mAb specific for keratin 5 and mAb specific for keratin 8 (Fig. 3A, upper panels). This staining pattern was also observed in the PKE and the 293 cells (Fig. 3A, upper panels). No heterogeneity of expression within the cell lines was noted as expected for typical kidney-derived cell lines and the 293 cells.

Although the decreased expression of E-cadherin and the unexpected expression of N-cadherin argue against the epithelial cell origin of 293 cells, the expression of keratin 5/8 supports the notion that they are epithelial cells. To examine the epithelial nature of 293 cells, we determined the expression levels of ZO-1 and occludin in these cells. These proteins are components of tight junctions in epithelial and endothelial cells (11). Immunoblot analysis of the cells revealed that the PKE and 293 cells, as well as the MDCK cells, expressed ZO-1 and occludin (Fig. 3B).

Immunofluorescence staining of these cells with ZO-1 antibodies revealed strong membrane staining (Fig. 3A, lower panels). Thus, transport of the tight junction component, ZO-1, from its site of biosynthesis occurred. No heterogeneity of expression within the cell lines was noted. Immunostaining for occludin revealed similar membrane staining patterns (data not shown).

The immunoblot detection of vimentin in the MDCK and the PKE cells revealed high expression levels of vimentin (Fig. 4A). The 293 cells, on the other hand, expressed vimentin at a reduced level (Fig. 4A). Immunoblot analysis of fibronectin revealed that the PKE cells expressed a large amount of fibronectin (Fig. 4A). Although the MDCK cells expressed a small amount of fibronectin, the 293 cells did not express detectable amounts of fibronectin.

Consistent with the results of the immunoblot analysis, immunofluorescence staining of the cells with vimentin and fibronectin antibodies revealed that the PKE cells expressed these proteins (Fig. 4B). The 293 cells expressed vimentin at a low level.

The immunoblot detection of desmoglein 2 in the 293 cells revealed high expression levels (Fig. 5A). Immunofluorescence staining of the 293 cells, however, revealed no membrane staining, although the staining of the DLD1 cells revealed clear membrane staining (Fig. 5B). Thus, desmoglein 2 remained localized in intracellular compartments. No heterogeneity of expression within the cell lines was noted.

These observations suggested that low E-cadherin expression levels were responsible for the failure of desmosome assembly. In an attempt to restore the localization of desmosomes in the membrane through increased E-cadherin expression, the 293 cells were transfected with the expression vector for HA-tagged E-cadherin and stable transfectants were isolated following selection with G418 (Fig. 5C). In these cells, E-cadherin was detected mainly on the surface membrane as revealed by staining with anti-HA antibodies (Fig. 5D). Immunofluorescence staining of these E-cadherin-expressing cells with desmoglein 2 antibodies revealed no membrane staining (Fig. 5E). Thus, E-cadherin expression in 293 cells is insufficient for the cell surface localization of desmoglein 2.