Abstract
The molecular nature of tight junction architecture and permeability is a long-standing mystery. Here, by comprehensive biochemical, biophysical, genetic, and electron microscopic analyses of claudin-16 and -19 interactions--two claudins that play key polygenic roles in fatal human renal disease, FHHNC--we found that 1) claudin-16 and -19 form a stable dimer through cis association of transmembrane domains 3 and 4; 2) mutations disrupting the claudin-16 and -19 cis interaction increase tight junction ultrastructural complexity but reduce tight junction permeability; and 3) no claudin hemichannel or heterotypic channel made of claudin-16 and -19 trans interaction can exist. These principles can be used to artificially alter tight junction permeabilities in various epithelia by manipulating selective claudin interactions. Our study also emphasizes the use of a novel recording approach based on scanning ion conductance microscopy to resolve tight junction permeabilities with submicrometer precision.