Identification and characterization of the domain that mediates actin binding and directs basal and TNF‐induced myosin light chain kinase 1 (MLCK1) recruitment to the perijunctional actomyosin ring

Abstract

Two myosin light chain kinase (MLCK) splice variants are expressed within intestinal epithelial cells, MLCK1 and MLCK2. These differ by a single exon that completes an IgCAM domain (IgCAM3) that is only present in MLCK1. We have previously shown that a small molecule targeting IgCAM3 prevents TNF‐induced MLCK1 recruitment to the perijunctional actomyosin ring thereby blocking myosin regulatory light chain (MLC) phosphorylation and tight junction barrier loss. Here, we sought to determine whether IgCAM3 alone was sufficient or if a larger portion of MLCK1 is required to direct recruitment and to define the interactions of these domains with F‐actin. MLCK1‐EGFP or MLCK2‐EGFP were stably expressed in Caco‐2 monolayers. In the absence of exogenous stimuli, MLCK1‐EGFP preferentially localized to the perijunctional actomyosin ring relative to MLCK2‐EGFP, which was most concentrated within basal stress fibers. Truncated EGFP‐tagged constructs consisting only of IgCAM1‐4 of MLCK1 or MLCK2 without the catalytic domain localized similarly to full‐length proteins. Further analyses demonstrated that the minimal MLCK1 trafficking domain consisted of IgCAM2 and IgCAM3 as well as the unstructured sequences preceding each of these domains; IgCAM3‐EGFP was distributed throughout the cytoplasm and nucleus of Caco‐2 cells. Consistent with this intracellular localization, studies using recombinant MLCK constructs showed that MLCK1(IgCAM1‐4) and MLCK1(IgCAM1‐3) bound F‐actin with greater affinity than MLCK1(IgCAM3) or MLCK2(IgCAM1‐4). TNF induced recruitment of MLCK1‐EGFP and MLCK1(IgCAM1‐4)‐EGFP, but not MLCK2‐EGFP or MLCK2(IgCAM1‐4)‐EGFP, to the perijunctional actomyosin ring within Caco‐2 monolayers. Although IgCAM3 was not concentrated at the perijunctional actomyosin ring, it was able to act as a dominant negative inhibitor of TNF‐induced full‐length MLCK1 recruitment. Moreover, IgCAM3 prevented TNF‐induced increases in tight junction permeability. Thus, although IgCAM3 alone is insufficient for perijunctional localization, it can disrupt TNF‐induced MLCK1 recruitment to the perijunctional actomyosin ring. These data confirm our previous results showing that IgCAM3‐mediated interactions are required for TNF‐induced MLCK1 recruitment and increased tight junction permeability but also demonstrate that interactions mediated or stabilized by other MLCK1 domains are also critically involved in this process. The data also suggest that, although direct binding to F‐actin may contribute to MLCK1 localization, it is insufficient for TNF‐induced MLCK1 recruitment and barrier regulation.