Suzuki, H., Tani, K., Tamura, A., Tsukita, S. (2015) Model for the Architecture of Claudin-Based Paracellular Ion Channels through Tight Junctions. J Mol Biol. 2015 Jan 30; 427(2):291-297
Epithelial cells form a border between the inside and outside of the body. They cover both the external surface of the body (skin) and the lining of internal cavities such as the lungs or the gastrointestinal tract. These cells are sealed together by tight junctions, which are just beneath the external surface.
The tight junctions form a cell to cell adhesion network that simultaneously blocks the passage of some substances across the epithelia and allows the highly regulated movement of specific ions and solutes across the barrier and between cells.
The primary components of tight junctions are claudins and recently there have been great advances in explaining their structure and function.
The first crystal structure of a single claudin monomer, solved in 2014, is shown below.
Image adapted from (Suzuki, H. et al., (2014)). Crystal structure of mouse claudin-15 (PDB 4P79)
The claudin protomer (monomer) has two distinct sections: a tight bundle of four α helices, which are embedded in the membrane of an epithelial cell, and five β strands, which form a curved sheet on the extracellular side. These extracellular β sheets contain several negatively charged residues which have previously been shown to be necessary for both the formation of a tight barrier between cells and its selectivity for positive ions (cations). They are also thought to form the paracellular channels that traverse the epithelium.
Two of these functions have been demonstrated by elegant experiments that substituted the acidic residues on the β sheet with positively charged ones, resulting in paracellular channels that were selective for anions.
The monomer structure alone, however, could not provide explanations for either the structure of tight junction strands, which form the mesh that creates cell adhesion, or the transcellular channels that allow interaction between cells.
A recent study by the same team that solved the first claudin structure, (Suzuki, H. et al., (2015)), uses cysteine crosslinking experiments to suggest an arrangement of the claudin monomers which would produce paracellular channels which run across the epithelium, transcellular channels which are perpendicular to the paracellular channels, and linear tight junction strands.
The proposed model uses a basic building block of two claudin monomers lying alongside each other but antiparallel so that the outer edges of the β sheets connect via hydrogen bonds to form half of a β barrel structure. These dimers then associate to form long polymers of the double row of claudins (see figure).
Image adapted from (Suzuki, H. et al., (2015)). The first image shows the association of claudin monomers and their polymerization. The second image shows the same structure rotated through 90˚ so that we are looking down the half β barrel .
This model was tested by substituting Cys residues for key Asn positions at the proposed interface between the outer β strands. The mutant claudins were found to form dimers that reverted to ther usual monomeric structure in reducing conditions. This illustrates that the molecules normally associate face to face, along the proposed β strands, so that once the substitutions were made disulphide bridges could be formed.
More corroboration was obtained using electron microscopy (EM), which is studied in the second-year TSMB course. A form of EM known as freeze fracture electron microscopy, which is often used to examine proteins embedded in lipid membranes, was used to provide images of tight junction strands, which appear as a network of linear strands whose width is consistent with the model.
In the epithelial sheet, a claudin double row would be anchored in the cell membrane with its half β barrel projecting out. By lining itself up against a similarly projecting double row in the adjacent cell we can see how tight junction strands with paracellular channels could be formed. The diameter of this channel would be less than 10 Å, enough to allow the passage of hydrated ions but to restrict their flow.
Image adapted from (Suzuki, H. et al., (2015)). The image shows the joining of a row of half β barrels from adjacent cells, forming a row of paracellular channels.
Epithelial cells maintain the separation of several internal spaces within the body and each of the interfaces has its own requirements for substances that are allowed to cross the border. This means that the paracellular channels need different ion specificity depending on which boundary they maintain. This could be provided by the charges and orientation of the residues in the channel lining and may also be influenced by flexible loops that link the β strands. It is also possible that these loops, which are not well conserved among the 27 members of the claudin family in the mouse (or human) proteome, could function as recognition regions so that only claudins of the same type can associate.
Possibly the most remarkable aspect of this model, however, can be seen by rotating it through 90˚.
Image adapted from (Suzuki, H. et al., (2015)). In this image the paracellular channels are seen in magenta whilst the α helices are aligned perpendicular to the page.
Further gaps can now be seen; these could provide transcellular channels allowing the flow of ions and solutes between epithelial cells.
In short this model appears to provide solutions to all of the functions required by claudins in tight junctions between epithelial cells and it will be very interesting to see if it proves to be correct as further crystal structures are solved.
Do look at the original paper if you have the chance as it includes an excellent short animation that shows the association of single claudin monomers into tight junction strands and then rotates the structure to show the different channels. (Follow the web link to the Supplementary Data.)