The Impact of Pathogenic and Artificial Mutations on Claudin-5 Selectivity from Molecular Dynamics Simulations

Tight junctions (TJs) are multi-protein complexes at the interface between adjacent endothelial or epithelial cells. In the blood-brain barrier (BBB), they are responsible for sealing the paracellular spaces and their backbone is formed by Claudin-5 (Cldn5) proteins. Despite the important role in preserving brain homeostasis, little is known on how Cldn5 oligomers assemble. Different structural models have been suggested, where Cldn5 protomers from opposite cells associate to generate paracellular pores that do not allow the passage of ions or small molecules. Recently, the first Cldn5 pathogenic mutation, G60R, was identified and shown to induce anion selectivity in the BBB TJs. This offers an excellent opportunity to further assess the structural models. In this work, we performed umbrella sampling molecular dynamics simulations to study the permeation of single Na ⁺ , Cl ⁻ and H ₂ O through two distinct G60R Cldn5 paracellular models. Only one of them, called Pore I, reproduces the functional modification observed in the experiments, displaying a free energy (FE) minimum for Cl ⁻ and a barrier for Na ⁺ at the central constriction, consistent with the formation of an anionic channel. To further test the validity of the model, we performed the same calculations for the Q57D and the Q63D mutants, which affect two side-chains in the constriction site. In particular, Q57 is conserved among various Cldns, with few exceptions such as the two cation permeable homologs Cldn15 and Cldn10b. In both cases, we obtain that the FE profiles are modified with respect to the wild-type system, facilitating the passage of cations. Our calculations are the first in-silico description of the effect of a Cldn5 pathogenic mutation, and provide a further assessment of the Pore I model for Cldn5-based TJ architectures, yielding new atom-detailed insight on the selective permeability of the paracellular spaces in BBB.