Novel approaches to study dynamic coupling of the tight junction barrier to the actin cytoskeleton

National Heart, Lung, And Blood Institute

Investigators

Linked publications & trials

Abstract

There is considerable evidence that the paracellular tight junction barrier depends on the underlying perijunctional actin cytoskeleton for its formation and regulation. Unfortunately insight into how the cytoskeleton is organized and dynamically coupled to the barrier is hindered by an incomplete understanding of how perijunctional proteins interact with the cytoskeletal proteins and how these interactions are organized in three dimensions. Further there are currently no experimental models for watching the tight junction contacts of claudin strands in real time to determine what controls their dynamic organization and their coupling to the cytoskeleton. In this project we are developing complementing experimental approaches to address these critical questions. During this report period we developed a model system in Rat-1 fibroblasts to allow observation in super-resolution time-lapse imaging of how claudin strands are coupled to actin filaments through ZO-1 and ZO-2. Because the tight junction strands in epithelial cells are oriented in the Z-axis, even super resolution methods can not resolve individual strands to detect whether they move, break or reanneal and how attachment to scaffolding proteins controls these characteristics. It is important to understand what controls TJ strands dynamics since this is presumably the basis for controlling flux through junction. When fluorescently-tagged claudin is express in claudin-null Rat-1 cells they form a network of anastomosing strand contacts between overlapping cells when viewed in the X-Y plane, reminiscent of genuine TJ networks. When the C-terminal PDZ-binding motif of claudin is removed the patches become much more dynamics with respect to inter-strand distances and the overall area of a patch. This was quantified as statistically significant using Image J. A phosphomimicking (Y to E) mutation in the tail of claudin that we have previously shown decreases binding affinity for ZO-1 resulted in the same increase in dynamics. CRISPR KO of ZO-1 and ZO-2 also results in the same dynamic behavior suggesting that the coupling of claudin through the PDZ domain of ZO proteins to actin stabilizes the stands. This helps explain why the physiologic barrier is so dependent on the state of perijunctional actin. Claudin strands only occasionally overlap with ZO-1 and ZO-1 only occasionally overlaps with actin filaments. Further, FRAP studies of claudin, ZO-1 and actin show that each has a very different recovery time and mobile fraction. Together these observations suggest that the cell-to-cell sealing proteins are not coupled to the cytoskeleton in a static configuration with a fixed stoichiometry but that the links are highly dynamic perhaps to allow cell contacts to move yet retain the paracellular seal (manuscript in preparation). To better define the 3 dimensional relationships among junctional cytoskeletal proteins we are documenting their relative locations around the junction using STED imaging in cultured Caco-2 and MDCK cells. Proteins are located either with antibodies or by fusion to fluorescent protein tags. This has allowed us to spatially define compartments beginning with the transmembrane sealing proteins, then the scaffolding proteins like ZO-1 and farther into the cell to actin and myosin and other cytoskeletal proteins. We have begun to examine changes in the spatial relationships of proteins after exposing cells to cytokine regimens which we have documented disrupt the barrier. We will confirm or further investigate changes in protein proximities using ZO-1as a reference by performing BioID proteomics before and after cytokine exposure. This should allow us to document molecule rearrangements in a model of pathologic barrier disruption. We are also developing a HRP-based fusion protein method (APEX) to be able to identify proteins that are near known TJ proteins and whether these spatial relationships and phosphorylation states change in response to physiologic stimuli that alter the barrier. Intramural collaborative studies include 1) characterization of paracellular permeability defect in the livers of LKB KO mice with Dr. Irwin Arias lab, NICHD (published in FY16), 2) testing AFM methods to quantify the contribution of the apical junctional complex to cortical tension in epithelial cell monolayers with Dr. Richard Chadwicks lab, NIDCD (manuscript in preparation), 3) mutational and computational modeling of claudin strand structure with Dr. Bechara Kachars lab NIDCD (manuscript in preparation). Extramural collaborations include testing a myelin specific claudin 11/BioID transgenic mouse model attempting to identify junction proteins in vivo.