Interactions between ankyrin-G, Plakophilin-2, and Connexin43 at the cardiac intercalated disc

Abstract

Rationale: The early description of the intercalated disc defined 3 structures, all of them involved in cell-cell communication: desmosomes, gap junctions, and adherens junctions. Current evidence demonstrates that molecules not involved in providing a physical continuum between cells also populate the intercalated disc. Key among them is the voltage-gated sodium channel complex. An important component of this complex is the cytoskeletal adaptor protein Ankyrin-G (AnkG).

Objective: To test the hypothesis that AnkG partners with desmosome and gap junction molecules and exerts a functional effect on intercellular communication in the heart.

Methods and results: We used a combination of microscopy, immunochemistry, patch-clamp, and optical mapping to assess the interactions between AnkG, Plakophilin-2, and Connexin43. Coimmunoprecipitation studies from rat heart lysate demonstrated associations between the 3 molecules. With the use of siRNA technology, we demonstrated that loss of AnkG expression caused significant changes in subcellular distribution and/or abundance of PKP2 and Connexin43 as well as a decrease in intercellular adhesion strength and electric coupling. Regulation of AnkG and of Na(v)1.5 by Plakophilin-2 was also demonstrated. Finally, optical mapping experiments in AnkG-silenced cells demonstrated a shift in the minimal frequency at which rate-dependence activation block was observed.

Conclusions: These experiments support the hypothesis that AnkG is a key functional component of the intercalated disc at the intersection of 3 complexes often considered independent: the voltage-gated sodium channel, gap junctions, and the cardiac desmosome. Possible implications to the pathophysiology of inherited arrhythmias (such as arrhythmogenic right ventricular cardiomyopathy) are discussed.

Figure 1

Immunoblots of PKP2, AnkG, and Cx43 (top to bottom) from samples exposed to protein A/G beads coated with PKP2, rabbit IgG (negative control), Cx43, or AnkG antibodies. Heart lysate (far right lane) used as positive control.

Figure 2

Loss of AnkG expression leads to PKP2 relocalization. Panel A: Western blot of NRVMs untreated (UNT) or treated with an oligonucleotide that prevents, or fails to prevent AnkG expression (AnkG-siRNA and ΦsiRNA). Cumulative data in B. All measurements normalized to those from untreated cells in same blot (n=8; p=NS ΦsiRNAvsAnkG-siRNA). C: Immunolocalization of AnkG (top) and PKP2 (bottom) immunoreactive proteins in cells treated with control oligonucleotide (ΦsiRNA) or AnkG-siRNA. Loss of AnkG expression (right column) led to PKP2 redistribution. Bars, 20μm.

Figure 3

Loss of AnkG expression did not affect abundance/distribution of plakoglobin or N-cadherin in NRVMs. A; left: Western blot for AnkG and plakoglobin from same sample; tubulin used for each blot as loading control. Center and right: immunodetection of plakoglobin (PG) in cells treated with control (ΦsiRNA) or AnkG-silencing constructs (AnkG-siRNA). B: Similar conditions. Detection of N-cadherin-immunoreactive proteins. C-D: Quantification of corresponding Western blots. All measurements normalized to those obtained from untreated cells in same blot. In both cases, n=6; p=NS ΦsiRNAvsAnkG-siRNA. Bars, 20μm.

Figure 4

AnkG abundance/distribution in PKP2-silenced NRVMs. A: Western blot for AnkG and PKP2 with respective tubulin loading controls; cells untreated (UNT), treated with siRNA for PKP2 (PKP2-siRNA) or with non-silencing construct (ΦsiRNA). B: Quantification of AnkG band density, corrected individually by loading control and normalized to UNT (n=5; p<0.05 ΦsiRNA vs PKP2-siRNA). C: Immunolocalization of PKP2 (top) and AnkG (bottom) in NRVMs treated with ΦsiRNA or PKP2-siRNA. Bars=20μm.

Figure 5

Loss of PKP2 expression leads to Na_v1.5 remodeling. A: Western blot for PKP2 and Na_v1.5 with respective tubulin loading controls in NRVMs untreated (UNT), treated with ΦsiRNA, or PKP2-siRNA. B: Quantification of Na_v1.5 band density (n=6; pNS; ΦsiRNA vs PKP2-siRNA). C: Immunolocalization of PKP2 (top) and Na_v1.5 (bottom) in NRVMs treated with ΦsiRNA or PKP2-siRNA. Bars=20μm.

Figure 6

Loss of AnkG expression weakens intercellular adhesive strength. Cells treated with 1.2-2.4 U/mL dispase 4h to release them from attachment to matrix. A: Pictures of monolayers after dispase treatment (time zero) and after 1 and 3 minutes of gentle shaking (field diameter, 35mm). B: Bars indicate number of fragments found at different times in experiment. An intact, lifted monolayer was counted as 1 fragment; UNT n=6, ΦsiRNA n=7, AnkG-siRNA n=6 (#, p<0.05; *, p<0.01).

Figure 7

Role of AnkG expression on abundance/localization of Cx43 and Na_v1.5. A and B: Western blot (left) and immunoreactive signal (center; right) in cells treated with ΦsiRNA (non-silencing control) or AnkG-siRNA. Tubulin used as Western loading control. UNT: Untreated cells. Cells probed for Cx43 or Nav1.5. C: Quantification of Cx43 band density corrected individually by loading controls. Each measurement normalized to that obtained by UNT cells in same blot (p<0.05 AnkG-siRNA vs ΦsiRNA; p<0.01 AnkG-siRNA vs UNT, n=8). D: Quantification of Nav1.5 band density corrected individually by loading controls. Each measurement normalized to UNT cells in same blot. (p=NS, n=5,5,6 for UNT, ΦsiRNA, AnkG-siRNA respectively). E: Dual patch clamp junctional conductance measurements in cell pairs UNT (n=7), treated with ΦsiRNA (n=12) or AnkG-siRNA (n=16). *p<0.01. Bars=20μm.