Claudins form ion-selective channels in the paracellular pathway. Focus on "Claudin extracellular domains determine paracellular charge selectively and resistance but not tight junction fibril architecture".

THE ABILITY OF EPITHELIA to form a diffusion barrier between cellular compartments of very different fluid and solute composition is dependent not only on asymmetrically distributed transcellular transport mechanisms (transcellular pathway) but also on structures that regulate the diffusion of ions and small, noncharged solutes through the paracellular pathway. Although those involved in regulating solute and water transport across the transcellular route are well understood, the mechanisms governing the paracellular pathway have only recently begun to be examined. At the apical end of the paracellular space, adjacent cell membranes are in close apposition, a site that was termed by early anatomists as the “terminal bar” (2) and was considered to be an impermeable barrier in the paracellular space. It was not until the elegant ultrastructural studies of Farquhar and Palade (8) that the terminal bar was shown, in fact, to consist of a junctional complex composed of an apical tight junction (zonula occludens), an intermediate junction (zonula adherens), and a desmosome (macula adherens). Analysis of freeze-fracture replicas indicated that tight junctions form a circumferential network of anastamosing strands of varying complexity located in the plane of the plasma membrane (20). Furthermore, it was suggested that the number of parallel tight junction strands might correlate with the level of measured transepithelial electrical resistance (TER) (5). When the number of strands was plotted against TER, however, it was found that the relationship between the increase in TER with each additional strand was not a linear, but an exponential, one. This led to the suggestion that tight junction strands contain pores that flicker between an open and closed conformation (4). Electrophysiological measurements had indicated that whereas the plasma membrane contains pores of 0.4 nm radius (18), pores of 3–4 nm radius are present in the paracellular space (13). In other words, the tight junction barrier in the paracellular space is considerably more permeable to water and small solutes than the plasma membrane. Furthermore, tight junctions appeared to be capable of discriminating between ions of similar charge and, in general, to be predominantly cation selective (15). This led to the prediction that tight junctions must contain aqueous pores lined by proteins, the amino acid composition of which would determine the charge selectivity of the tight junction pores. Initial studies designed to identify integral tight junction proteins resulted in the discovery of two important tight junction-associated cytoplasmic proteins, ZO-1 (22) and cingulin (3). It was not until 1993, however, that the molecular composition of the tight junction strands began to be clarified in a series of groundbreaking studies by the Tsukita group. The first of these was occludin (11), a 58-kDa tetraspan phosphoprotein, which, when overexpressed in MDCK cells, resulted in an elevation of TER and a paradoxical increase in the flux of small, water-soluble solutes (1, 14). However, when occludin-null embryonic stem cells were found to differentiate into tight junction-expressing cells (16), it became clear that other protein(s) must contribute to the formation of tight junction strands and that the precise function of occludin in the tight junction remained to be established. This was further supported by the observation that occludin-null mice survived to adulthood, although they developed a complex phenotype (17). In their quest to identify other integral tight junction proteins, the Tsukita group re-examined the junction fraction from their original chicken liver preparations. The search yielded two novel, 23-kDa integral tight junction proteins, claudins-1 and -2 (9). They, like occludin, are tetraspan proteins; however, the claudins share no sequence homology with occludin. In addition, in contrast to occludin, when claudin-1 cDNA was transfected into fibroblasts, they formed a network of tight junction strands in the plasma membrane, indicating that claudin is necessary and sufficient to form tight junction strands (12). To date, a 20-member family of claudins has been recognized (23) and these, either singly or in combination with several different claudins, are expressed in a celland tissue-specific pattern. Although clearly indicating that claudin(s) are Address for reprint requests and other correspondence: E. E. Schneeberger, Molecular Pathology Unit, Rm. 7147, 149 13th St., Charlestown, MA 02129 (E-mail: [email protected]). Am J Physiol Cell Physiol 284: C1331–C1333, 2003; 10.1152/ajpcell.00037.2003.