Inducible expression of claudin-1-myc but not occludin-VSV-G results in aberrant tight junction strand formation in MDCK cells

Occludin and 18 distinct members of the claudin family are tetra-span transmembrane proteins that are localized in cell-specific tight junctions (TJs). A previous study showed that expression of chick occludin in Madin-Darby canine kidney (MDCK) cells raised transepithelial electrical resistance (TER) and, paradoxically, increased mannitol flux. In the present study, we employed epitope tagged canine occludin expression, under the control of the tetracycline repressible transactivator, to determine the extent to which the unexpected parallel increase in TER and mannitol flux was related to a structural mismatch between avian and canine occludins, which are only 50% identical. To determine whether the paradoxical changes in permeability was specific to occludin, we assessed the effect of over-expressing epitope tagged murine claudin-1. Our data revealed that over-expression of either of the epitope tagged mammalian tight junction proteins increased TER, mannitol and FITC-dextran flux. We observed a 2- and up to 5.6-fold over-expression of occludin-VSV-G and claudin-1-myc, respectively, with no change in ZO-1, endogenous occludin or claudin-1 expression. Confocal microscopy revealed that occludin-VSV-G, claudin-1-myc and ZO-1 co-localized at the TJ. In addition, claudin-1-myc formed aberrant strands along the lateral cell surface without an underlying ZO-1 scaffold. In fracture labeled replicas these strands consisted of claudin-1-myc with little accompanying occludin. These observations suggest that in epithelial cells claudin-1 can assemble into TJ strands without the participation of either ZO-1 or occludin. The proximity of the myc tag to the COOH-terminal YV sequence of claudin-1 appeared to interfere with its interaction with ZO-1, since over-expression of non-tagged claudin-1 increased TER but had a minimal effect on solute flux and no aberrant strands formed. From our data we conclude that differences in structure between avian and mammalian occludin do not account for the observed paradoxical increase in mannitol flux. Levels of ZO-1 remained unchanged despite substantial increases in induced TJ integral protein expression, suggesting that an imbalance between levels of ZO-1 and occludin or claudin-1 leads to altered regulation of pores through which non-charged solute flux occurs. We suggest that ion and solute flux are differentially regulated at the TJ.