Arrhythmogenic cardiomyopathy related DSG2 mutations affect desmosomal cadherin binding kinetics

Abstract

Cadherins are calcium dependent adhesion proteins that establish the intercellular mechanical contact by bridging the gap to adjacent cells. Desmoglein-2 (Dsg2) is a specific cadherin of the cell-cell contact in cardiac desmosomes. Mutations in the DSG2-gene are regarded to cause arrhythmogenic (right ventricular) cardiomyopathy (ARVC) which is a rare but severe heart muscle disease. The molecular pathomechanisms of the vast majority of DSG2 mutations, however, are unknown. Here, we investigated the homophilic binding of wildtype Dsg2 and two mutations which are associated with ARVC. Using single molecule force spectroscopy and applying Jarzynski's equality we determined the kinetics and thermodynamics of Dsg2 homophilic binding. Notably, the free energy landscape of Dsg2 dimerization exposes a high activation barrier which is in line with the proposed strand-swapping binding motif. Although the binding motif is not directly affected by the mutations the binding kinetics differ significantly from the wildtype. Furthermore, we applied a dispase based cell dissociation assay using HT1080 cell lines over expressing Dsg2 wildtype and mutants, respectively. Our molecular and cellular results consistently demonstrate that Dsg2 mutations can heavily affect homophilic Dsg2 interactions. Furthermore, the full thermodynamic and kinetic description of Dsg2 dimerization provides a consistent model of the so far discussed homophilic cadherin binding.

Figure 1

(A) Crystal structure of Dsg2 homodimer shown in ribbon representation [, pdbID: 5ERD]. Coordinated Ca²⁺ are shown as blue spheres. (B) Schematic of cantilever tip and sample substrate functionalized with Dsg2 for AFM single molecule force spectroscopy. Both, cantilever and substrate are modified with covalently immobilized Dsg2 fragments. (C) SMFS working principle and typical force-distance cycle exposing a specific dissociation event. The striated area under the curve represents the dissipated work to break a Dsg2 homodimer. (D) Principle of dispase assay for measuring cell-cell adhesion. Cells grown to confluency (a) are lifted up by dispase treatment and exposed to mechanical stress by shaking of cell cultures. Resulting islets of the monolayer are counted and recombinant cell cultures of mutant and wildtype are compared (b).

Figure 2

(A) Force loading rate graph showing the most probable dissociation force $F^{⁎}$ plotted versus the logarithmic loading rate $\ln (r)$. Colors represent Dsg2 wildtype in green, Dsg2- p.D105E (p.D154E) in red and Dsg2-p.V343I (p.V392I) in yellow. The error bars of $F^{⁎}$ are estimated by the standard error of the mean (SEM) and were used for a weighted fit of the Evans and Ritchie model (Eq. 2). Inset: The uncertainty of the estimated parameters $k_{-}^0$ and $x^{‡}$ result from the covariance matrix of the fitted model and the uncertainty of the cantilever spring constant. (B) Proposed free energy landscape of Dsg2 strand exchange binding exposes a small energy difference between bound and unbound state $\mathrm{\Delta }{G}_0$ and a high activation barrier $\mathrm{\Delta }{G}^{‡}$. Dsg2 monomers dimerize when their intermolecular distance go below the reaction length $x^{‡}$.

Figure 3

(A) Series of control experiments for different Ca²⁺ concentrations and sample substrates. The dissociation probability distributions are scaled for the absolute binding frequency. Experiments with Dsg2 modified AFM cantilever and sample substrate in 2 mM calcium-buffer (green) and EDTA-buffer (orange) expose a clear dependency of the binding frequency on the calcium concentration. The background of unspecific adhesion was tested in an experiment conducted on substrates lacking Dsg2 in calcium (red) and EDTA-buffer (gray). (B) Evolution of the Jarzynski equality. The exponentially averaged work $W_n$ plotted versus the number of dissociation events $n$. For wildtype Dsg2 and mutations thereof (green, red and yellow) measured in a 2 mM Ca²⁺ buffer-solution the Jarzynski equality converges fast yielding a robust estimate of the free energy difference $\mathrm{\Delta }{G}_0$. The evolution of $\mathrm{\Delta }{G}_0$ for wildtype Dsg2 (blue) in calcium-free EDTA-buffer solution does not converge for $n<2000$.

Figure 4

Confluent monolayers of transfected cells were subjected to dispase assay. Decreased number of fragments indicates relatively enhanced cell-cell adhesion. Box and whiskers plots represent the number of fragments in six independent experiments for Dsg2 wildtype, p.D105E (p.D154E; $p<0.001$ versus wildtype) and p.V343I (p.V392I). Box and whiskers plot: boxes extend from the 25th–75th, whiskers from the 5th–95th percentiles, the medians are shown as lines; outliers are given as dots. Means are indicated as crosses.