Journal of Cell Science 514 cell functions, including cell proliferation (Ashton et al., 1999;Kitamura et al., 2003), differentiation (Corbit et al., 1999;Pessino et al., 1995), apoptosis (Brodie and Blumberg, 2003;Kajimoto et al., 2004;Zhong et al., 2002) and tumor suppression (Lu et al., 1997;Reddig et al., 1999). Increasing evidence also indicates that PKCδ has a positive role in cell motility (Chen et al., 2007;Gliki et al., 2002;Iwabu et al., 2004;Li et al., 2003) and the metastatic potential of tumor cells (Kiley et al., 1999;Kruger and Reddy, 2003; Alonso-Escolano et al., 2006;Kharait et al., 2006;Villar et al., 2007). However, the role of PKCδ in intercellular junctions remains obscure. In this study, we aim to explore the role of PKCδ in cellcell junctions using Madin-Darby canine kidney (MDCK) cells as a model.
Results
Expression of GFP-PKCδ, but not GFP-PKCα, suppresses the homophilic interactions between the ectodomains of E-cadherins in MDCK cellsTo explore the role of PKC in cell-cell junctions, GFP-PKCδ and GFP-PKCα were stably expressed in MDCK cells (Fig. 1A). Expression of either construct in those cells did not alter their growth (data not shown) or their ability to form cell colonies within which the cells are in contact with each other (Fig. 1B). To examine their effect on adherens junctions and tight junctions, cells stably expressing GFP-PKCδ or GFP-PKCα were co-cultured with parental control MDCK cells and then grown to confluence. The cells were fixed and stained for adherens junctions and tight junctions with anti-E-cadherin and anti-ZO-1, respectively. Neither GFP-PKCδ nor GFP-PKCα affected tight junctions of MDCK cells (Fig. 1C,D). However, the fluorescence intensity of E-cadherin at cell-cell contacts in the cells expressing GFP-PKCδ, but not GFP-PKCα, was much weaker than that in adjacent parental MDCK cells (Fig. 1C,D).As the total and cell surface levels of E-cadherins were not altered by GFP-PKCδ (supplementary material Fig. S1), it is unlikely that the decreased fluorescence intensity of E-cadherin was due to suppression of E-cadherin expression. The rat monoclonal anti-Ecadherin (clone ECCD-2) used in Fig. 1C is known to recognize the extracellular domain of E-cadherin, rendering it possible that GFP-PKCδ causes a conformational change on the ectodomain of E-cadherin, which then prevents E-cadherins from detection by the ECCD-2 antibody. To clarify this, a mouse monoclonal antibody (clone 36) that recognizes the cytoplasmic portion of E-cadherins was used. The fluorescence intensity of E-cadherin with the clone 36 antibody did not show differences between control cells and PKCδ-overexpressed cells ( Fig. 2A), supporting the idea that the conformation of E-cadherins rather than their distribution or expression is affected by PKCδ. In addition, we found that the ability of the ECCD-2 antibody, but not the clone 36 antibody, to detect E-cadherin at cell-cell junctions relied on the presence of Ca 2+ in the culture medium, as demonstrated by Ca 2+ -switch assay (Fig. 2B), indicat...