Skip main navigation

EHD1 Modulates Cx43 Gap Junction Remodeling Associated With Cardiac Diseases

Originally publishedhttps://doi.org/10.1161/CIRCRESAHA.119.316502Circulation Research. 2020;126:e97–e113

Abstract

Rationale:

Efficient communication between heart cells is vital to ensure the anisotropic propagation of electrical impulses, a function mainly accomplished by gap junctions (GJ) composed of Cx43 (connexin 43). Although the molecular mechanisms remain unclear, altered distribution and function of gap junctions have been associated with acute myocardial infarction and heart failure.

Objective:

A recent proteomic study from our laboratory identified EHD1 (Eps15 [endocytic adaptor epidermal growth factor receptor substrate 15] homology domain-containing protein 1) as a novel interactor of Cx43 in the heart.

Methods and Results:

In the present work, we demonstrate that knockdown of EHD1 impaired the internalization of Cx43, preserving gap junction-intercellular coupling in cardiomyocytes. Interaction of Cx43 with EHD1 was mediated by Eps15 and promoted by phosphorylation and ubiquitination of Cx43. Overexpression of wild-type EHD1 accelerated internalization of Cx43 and exacerbated ischemia-induced lateralization of Cx43 in isolated adult cardiomyocytes. In addition, we show that EHDs associate with Cx43 in human and murine failing hearts.

Conclusions:

Overall, we identified EHDs as novel regulators of endocytic trafficking of Cx43, participating in the pathological remodeling of gap junctions, paving the way to innovative therapeutic strategies aiming at preserving intercellular communication in the heart.

Introduction

Meet the First Author, see p 1322

Intercellular communication between heart cells is vital to sustain the anisotropic propagation of electrical impulses, which is mainly ensured by gap junctions (GJ) localized at the intercalated discs (IDs).1 GJ are formed by 4-pass transmembrane Cx (connexins), of which Cx43 is the most expressed in the heart.2,3 Six connexin subunits oligomerize to form of a hemichannel, which can mediate the exchange of molecules between the intracellular and extracellular milieu. Two hemichannels localized in adjacent cells can dock, forming GJ channels that allow the passage of small metabolites and ions, sustaining electric and metabolic coupling between cells.4 Several cardiomyopathies characterized by conduction system block and arrhythmogenesis have been associated with impaired GJIC (GJ intercellular communication), including acute myocardial infarction and heart failure (HF).2,5

Regulation of GJIC can occur at different levels during the connexin life cycle, and it is typically achieved through alterations on channel abundance, conductance, permeability, and subcellular distribution. Endocytosis of GJ, involving the recognition of Tyr-sorting signals or ubiquitin moieties on Cx43 contributes to the fine-tuned regulation of intercellular communication, under basal and pathophysiological conditions.6,7 Accordingly, studies from our lab demonstrated that Nedd4-mediated ubiquitination of Cx43 triggers the recruitment of the Eps15 (endocytic adaptor epidermal growth factor receptor substrate 15), required for endocytosis of Cx43-containing channels.8

It is well established that increased phosphorylation of Cx43 on residues S368 and S373 contribute to ischemia-associated remodeling of Cx43 channels in cardiomyocytes, which includes lateralization or degradation of Cx43.9–11 Additionally, Cx43-S368 phosphorylation was associated with increased ubiquitination and internalization of Cx43, suggesting that crosstalk between different posttranslational modifications underlies GJ remodeling.7,12–15 Results from our group demonstrated that ubiquitinated Cx43 is degraded in cardiomyocytes during ischemia and ischemia/reperfusion, with the involvement of Eps15 and the autophagy players p62 and LC3 (microtubule-associated protein 1 light-chain 3).10 Nevertheless, little is known about the molecular mechanisms underlying GJ remodeling, before its degradation. Recently, we unveiled the Cx43 interactome in rat hearts, which showed that ischemia positively regulates the interaction of Cx43 with EHD1 (Eps15 homology domain-containing protein 1).16 EHD1 is the best-characterized member of a family of endosomal scaffolding ATPases involved in membrane trafficking, comprising 4 mammalian paralogues—EHD1-4.17 In various cellular systems, EHD1 participates in endocytic recycling of MHC (major histocompatibility complex) proteins, glucose transporters, glutamate, and transferrin receptors, as well as in the internalization of L1/NgCAM (neuron-glia cell adhesion molecule) and IGF-1R (insulin-like growth factor 1 receptor).18–21 Although the interacting partners and effectors remain largely undefined, EHDs have been linked to the regulation of endosome-based trafficking of cardiac ion channels, ultimately affecting calcium (Ca2+) homeostasis and excitability.22 Therefore, in the present work, we hypothesized that the association between Cx43 and EHD1 contributes to the modulation of Cx43 trafficking and lateralization during myocardial ischemia and HF.

Methods

Please see the Major Resources Table in the Data Supplement.

The data that support the findings of this study are available from the corresponding author on reasonable request.

Animal Models

Animal models were performed as previously described, according to the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (85–23, revised 2011) and the Portuguese law (DL 129/92, DL 197/96; P 1131/97).10,16,23 The Faculty of Medicine of the Universidade do Porto (FMUP) is a governmental institution approved to perform animal experiments by the Portuguese Veterinary Authorities (Direção Geral de Alimentação e Veterinária). This protocol was approved by FMUP/Hospital São João ORBEA (Orgão de Bem-Estar e Ética Animal/Animal Welfare and Ethics Body) and Direção Geral de Alimentação e Veterinária. To exclude introducing the variable sex in the aortic banding model and considering that young female rats are more protected from cardiac hypertrophy and remodeling compared with males, we deliberately used male Wistar rats (5–6 weeks of age). Inclusion criteria were based on cardiac function and structure parameters (available in the Data Supplement). None of the animals was excluded from the study.

Biotinylation of Cell-Surface Proteins, Internalization and Recycling Assays

Cell surface biotinylation was performed using sulfo-NHS-SS-biotin (sulfosuccinimidyl-2-[biotinamido]ethyl-1,3-dithiopropionate; No. 21331, Thermo Fisher Scientific) at 4°C.10 Excess biotin was quenched, washed with PBS/Ca2+/Mg2+, and either scraped in radioimmunoprecipitation assay buffer before Western blot analysis or changed to complete cell culture medium before being returned to a 37°C incubator for 30 minutes, to follow internalization of biotin-labeled proteins. Glutathione stripping (50 mmol/L glutathione, 75 mmol/L NaCl, 10 mmol/L EDTA, 0.075 N NaOH, pH 8), to discriminate between INT (internalized Cx43) and plasma membrane-localized Cx43, was performed (2×, 15 minutes) to release biotinylated molecules, followed by extensive washing. To indirectly assess Cx43 recycling, cells were incubated an additional 30 minutes at 37°C, followed by glutathione stripping (REC), to quantify the amount of Cx43 that remained in the cytosol after the recycling pulse. Internalized Cx43 levels were analyzed by Western blot of biotinylated complexes pulled down on neutravidin beads (No. 29200; Pierce).

Antibody Feeding Assays

Cells overexpressing Cx43E2HA were pulsed with antihemagglutinin antibodies for 30 minutes at 37°C. Antibodies remaining at the cell surface were removed by acid stripping (0.5 mol/L NaCl, 0.5% acetic acid pH 3). To label the internalized fraction of Cx43, cells were fixed, permeabilized, and immunostained after this step. To proceed with the recycling assay, cells were washed in PBS and further incubated at 37°C for 15 minutes, in cell culture medium, in the presence of 8 mmol/L phorbol 12-myristate 13-acetate (PMA) and 50 nmol/L Bafilomycin A1. To label recycled Cx43, cells were fixed and immunostained without permeabilization. Images were collected with an Axio Observer.Z1, with a Plan-Apochromat 63x/1.4 Oil DIC M27 objective (Carl Zeiss AG, Jena, Germany) and analyzed with ImageJ.

Statistical Analysis

All data represent at least 3 independent experiments, expressed as individual data points with mean and SD. Shapiro-Wilk was used to test for normality. Correction for multiple comparisons was applied where indicated in the respective figure legends. Representative images reflect the average level of each experiment. Data were analyzed with GraphPad Prism 6, version 6.01 (GraphPad Software, Inc).

Results

Cx43 Colocalizes and Coimmunoprecipitates With EHD Proteins in Cardiomyocytes

Our previous proteomic study provided evidence that interaction of Cx43 with EHD1 and EHD2 vary during heart ischemia and ischemia/reperfusion (Figure IA in the Data Supplement).16 For validation, we performed colocalization and immunoprecipitation experiments using antibodies that recognize endogenous EHD proteins (EHD1-3). Results presented in Figure 1A and 1B show that EHD1-3 partially colocalized and coimmunoprecipitated with Cx43 in isolated adult rat cardiomyocytes.24 Moreover, association between Cx43 and EHD1-3 was enhanced during ischemia in Langendorff-perfused hearts, returning to the basal levels during reperfusion (ischemia/reperfusion; Figure 1C and 1D). Colocalization of EHD1-3 with Cx43 was particularly noticeable at the lateral sarcolemmal membranes and intercellular vesicles in ischemic hearts (Figure 1D, middle insets). However, colocalization between EHD1-3 and phosphorylated-S368 (Cx43-pS368), a modified form of Cx43 restricted to the IDs, remained unchanged during ischemia and ischemia/reperfusion (Figure IB in the Data Supplement).

Figure 1.

Figure 1. Cx43 (connexin 43) colocalizes and coimmunoprecipitates with EHD (Eps15 [endocytic adaptor epidermal growth factor receptor substrate 15] homology domain-containing protein) in cardiomyocytes.A, Adult cardiomyocytes isolated from Wistar rats were immunostained with antibodies against Cx43 (green) and EHD1-3 (red), followed by confocal microscopy imaging. Inset represents magnified colocalization areas. Scale bar 20 μm. B, Cx43 was immunoprecipitated (IP) from isolated cardiomyocytes, after which interaction with EHD1-3 was analyzed by Western blot (WB). C, Rat hearts were maintained in an ex vivo Langendorff apparatus for 20 min (min; control, CT), subjected to no-flow ischemia (ISCH, 20 min), or to 20 min of ischemia followed by 60 min of reperfusion (ischemia/reperfusion [I/R]). Cx43 was IP from heart lysates, after which interaction with EHD1-3 was analyzed by WB. Graph depicts quantification of interaction levels (n=7). D, Colocalization between Cx43 (green) and EHD1-3 (red) in rat hearts was analyzed by confocal microscopy and quantified using ImageJ (n=4). N-cadherin (white) was used as a marker of intercalated discs. Nuclei were stained with DAPI (4′,6-diamidino-2-phenylindole). Insets represent magnified colocalization areas. Scale bars 20 μm. P values were derived by Mann-Whitney test in all plots. n.s indicates not significant.

Additionally, colocalization between Cx43 and EHD1-3 was increased in primary cultures of adult human and neonatal rat cardiomyocytes subjected to in vitro ischemia (Figure IC and ID in the Data Supplement).25,26 Overall, these results suggest that association of Cx43 with EHD proteins in cardiomyocytes mediates Cx43 trafficking during ischemia.

Knockdown of EHD1 Impairs the Internalization of Cx43 in HL-1 Cardiomyocytes

To address the role played by EHD1 in the modulation of ischemia-induced remodeling of Cx43 channels, we performed siRNA (small interfering RNA)-mediated knockdown of EHD1 (siEHD1) in HL-1 cardiomyocytes (Figure IIA in the Data Supplement), after which we evaluated the cell surface levels of Cx43 in basal or ischemic conditions, by biotinylation assays.27,28 Given its higher stability, N-cadherin was used as a control. In basal conditions, the amount of biotinylated Cx43, putatively representing the undocked Cx43 hemichannel population, was not affected by the knockdown of EHD1 (Figure 2A). In control cells (scramble siRNA [small interfering RNA; siCT]), ischemia increased the levels of Cx43 in hemichannels, which was prevented in EHD1-knockdown cells, that is, no further increase in biotinylated Cx43 was observed following ischemia (Figure 2A). These results were confirmed by fluorescence microscopy analysis (Figure IIB in the Data Supplement). Neither ischemia nor EHD1 silencing induced significant alterations in the Cx43 present in the Triton X100-insoluble fraction (representative of Cx43 within GJ plaques). However, in cells silenced for EHD1 subjected to ischemia, we observed decreased levels of Cx43 in the soluble fraction, likely reflecting an impaired Cx43 internalization (Figure IIC in the Data Supplement).

Figure 2.

Figure 2. EHD1 (Eps15 [endocytic adaptor epidermal growth factor receptor substrate 15] homology domain-containing protein 1) mediates internalization of Cx43 (connexin 43).A, siRNA (small interfering RNA)-mediated knockdown of EHD1 was performed in HL-1 cells for 48 h (siEHD1). Controls were performed with transfection of scramble siRNA (siCT). Simulated ischemia (ISCH) was performed for 30 min, where indicated. Levels of plasma membrane-localized Cx43 were analyzed by Western blot (WB) after cell surface biotinylation and neutravidin pull down. Cx43 levels were quantified by densitometry and plotted on a graph (n=5). B, Biotin-based internalization and recycling assays were performed in cells knocked-down for EHD1 (siEHD1). siCT was used as control. The levels of internalized Cx43 (INT) at 30 min (n=5; represented as fold change of Cx43 basal levels) and after the REC (internalized Cx43 after the recycling step; n=4; represented as fold change of Cx43 internalized at 30 min) were analyzed by WB of biotinylated complexes pulled down on neutravidin beads. All experiments were performed in the presence of inhibitors of protein synthesis (5 mmol/L cycloheximide) and lysosomal degradation (50 nmol/L bafilomycin A1). C, A scratch was created in an HL-1 cell monolayer, after which cells were incubated with Lucifer Yellow (LY) for 1 min. After LY washing, dye transfer was assessed after 5 min. Graph depicts quantification of the diffusion distance of LY. Dots represent individual measurements in each analyzed field (n=4). Scale bars 50 μm. B, HEK293 (human embryonic kidney 293) cells with stable overexpression of Cx43 (HEK293Cx43+) were transfected with myc-EHD1WT (myc-tagged wild-type EHD1) for 24 h. ATP depletion was induced by Antimycin A/2DG (Antimycin A and 2-deoxy-D-glucose) treatment for 2 h, after which Cx43 was immunoprecipitated (IP) and interaction with EHD1 evaluated by WB (n=4). E, HEK293 cells were cotransfected with wild-type Cx43 (Cx43WT), Cx43 mutated on residue S373 (Cx43S373A) or S368 (Cx43S368A) and myc-EHD1WT for 24 h. Cells were treated with phorbol 12-myristate 13-acetate (PMA) for 15 min, as indicated, followed by Cx43 IP and WB analysis of interaction with EHD1 (n=5). P values were derived by 2-way ANOVA, followed by Tukey multiple comparison test in A and E and by Mann-Whitney test in B–D. CT indicates control; INT, internalized Cx43; and n.s, not significant.

We proceeded to evaluate the role of EHD1 in Cx43 endocytosis, through internalization assays. Results in Figure 2B show that internalization of Cx43 was impaired by knockdown of EHD1, whereas the recycling was not affected. Furthermore, the half-life of Cx43 was not significantly altered by siEHD1 (Figure IID in the Data Supplement). In accordance, scrape loading/dye transfer showed that cell-cell communication was enhanced in EHD1-knocked down cells (Figure 2C, Figure IIE in the Data Supplement).

PMA-Induced Phosphorylation and Ubiquitination of Cx43 Promote Interaction With EHD1

Given that cardiomyocytes are extremely difficult to manipulate genetically, to unveil the molecular mechanisms and signals driving the interaction between EHD1 and Cx43, we used HEK293 (human embryonic kidney 293) cells stably overexpressing Cx43 (HEK293Cx43+) with ectopic expression of myc-EHD1WT (myc-tagged wild-type EHD1). To mimic the ischemic environment faced by cardiomyocytes, HEK293 cells were incubated with Antimycin A and 2-deoxy-D-glucose to deplete cellular ATP (Figure IIIA and IIIB in the Data Supplement). In agreement with the previous results, interaction between EHD1 and Cx43 was promoted following ATP depletion (Figure 2D).

Given the role played by phosphorylation and ubiquitination on intracellular trafficking and turnover of Cx43, we challenged the cells with PMA, to stimulate phosphorylation of Cx43-S262, S368, and S373.11,14,29,30 Our results show that cells treated with PMA presented increased colocalization between Cx43 and EHD1, as well as increased S368 phosphorylation (Figure 2E, Figure IIIC in the Data Supplement). Additionally, we assessed the impact on the interaction with EHD1 after point mutating the residues S373 and S368 of Cx43, previously described to be phosphorylated during myocardial ischemia. Results depicted in Figure 2E show that both phospho-null mutants were insensitive to PMA treatment, failing to undergo S368 phosphorylation and to further coimmunoprecipitate with EHD1 (Figure 2E). Phosphorylation of S262 was not affected by the mutations (Figure IIID in the Data Supplement), ascribing these effects to the specific phosphorylation of S368.

Grounded on previous studies implicating ischemia-induced ubiquitination of cardiac Cx43 in the regulation of GJ remodeling, we assessed the role of ubiquitin on the interaction between Cx43 and EHD1. Using a chimera in which a ubiquitin moiety was fused to Cx43 (Cx43Ub), we showed that the presence of ubiquitin increased the interaction with EHD1 (Figure IIIE in the Data Supplement).10,11,30

Overexpression of Wild-Type EHD1 Induces Endosomal Accumulation of Cx43

Next, we sought to unveil the role of EHD1 on the subcellular trafficking and localization of Cx43, both in basal conditions and after PMA stimulation, to promote interaction of EHD1 with Cx43. Our results show that localization of Cx43 at EEA1 (early endosome antigen 1)-positive compartments increased in cells overexpressing wild-type GFP (green fluorescent protein)-EHD1 when compared with GFP-overexpressing cells (Figure 3A). However, in cells overexpressing a mutant EHD1, which lacks ATP binding and oligomerization capacity (GFP-EHD1G65R), the levels of Cx43 within early endosomes were partially decreased, in comparison with wild-type EHD1-overexpressing cells, being further reduced after PMA exposure, suggesting that EHD1 is promoting endocytosis of Cx43.31

Figure 3.

Figure 3. Overexpression of wild-type EHD1 (Eps15 [endocytic adaptor epidermal growth factor receptor substrate 15] homology domain-containing protein 1) induces endosomal accumulation of Cx43 (connexin 43). HEK293Cx43+ (human embryonic kidney 293) cells were transfected with GFP (green fluorescent protein)-tagged wild-type EHD1 (GFP-EHD1WT), mutant EHD1 (GFP-EHD1G65R), or GFP alone for 24 h. Cells were treated with phorbol 12-myristate 13-acetate (PMA) for 15 min, where indicated. A, Colocalization between Cx43 (white) and EEA1 (early endosome antigen 1; red) was analyzed by confocal microscopy. Quantification was performed using ImageJ (n=4). Nuclei were stained with DAPI (4′,6-diamidino-2-phenylindole). Insets represent magnified colocalization areas. Scale bars 10 μm. B, Colocalization between Cx43 (white) and Lamp2 (lysosome-associated membrane protein 2; red; n=4). Nuclei were stained with DAPI. Insets represent magnified colocalization areas. Scale bars 10 μm. P values were derived by 2-way ANOVA, followed by Tukey multiple comparison test in all plots. n.s indicates not significant.

Moreover, in comparison with GFP, overexpression of wild-type EHD1 dramatically increased the localization of Cx43 in late endosomes (Lamp2 [lysosome-associated membrane protein 2]-positive), whereas only a modest increase was observed in EHD1G65R-expressing cells (Figure 3B). Subcellular fractionation assays further validated these data, demonstrating that cells expressing EHD1G65R presented lower levels of Cx43 in late endosomes when compared with cells overexpressing EHD1WT (Figure IIIF in the Data Supplement).32 After PMA exposure, the localization of Cx43 in late endosomes was significantly decreased in cells expressing EHD1WT or EHD1G65R, but not GFP alone, implying that EHD1 is either promoting the recycling of Cx43 or a faster lysosomal degradation. Altogether, these results suggest that wild-type EHD1 promotes the internalization of Cx43 and further supports Cx43 to either recycle back to the plasma membrane or to proceed from early to late endosomes.

In addition to its effect in promoting impairment of GJIC and increased Cx43 internalization,7,14 some studies provided evidence that PMA induces a transient increase in the levels of Cx43 at the plasma membrane, before endocytosis of GJ plaques.30 Therefore, we investigated the involvement of EHD1 in this process. Results presented in Figure 4A show that overexpression of the wild-type (myc-EHD1WT) or the mutant EHD1 (myc-EHD1G65R) did not affect the levels of biotinylated Cx43 in basal conditions. Nevertheless, for short periods (5 minutes) of PMA treatment, overexpression of EHD1G65R increased the levels of Cx43 at the cell surface, after which the amount of biotinylated Cx43 started to decrease to basal levels (at 15 minutes). In cells expressing the empty vector or wild-type EHD1, the levels of Cx43 hemichannels were not significantly changed after PMA incubation, which was corroborated by microscopy (Figure IVA in the Data Supplement).

Figure 4.

Figure 4. Overexpression of EHD1G65R promotes Cx43 (connexin 43) recycling.A, HEK293Cx43+ (human embryonic kidney 293) cells were transfected with myc-EHD1WT (myc-tagged wild-type EHD1), myc-EHD1G65R, or the empty vector for 24 h. Cells were treated with phorbol 12-myristate 13-acetate (PMA) for 5 or 15 min, after which cell surface biotinylation was performed. Levels of Cx43 at the plasma membrane were analyzed by Western blot (WB). Values are depicted on the graph (n=4). P values were derived by 2-way ANOVA, followed by Tukey multiple comparison test. B, HEK293Cx43+ cells were cotransfected with hemagglutinin (HA)-tagged Cx43 (Cx43E2HA) and GFP (green fluorescent protein), GFP-EHD1WT or GFP-EHD1G65R for 24 h. Antibody feeding assays were performed, by pulsing cells with anti-HA antibodies for 30 min at 37°C (internalization, t=30’; n=5). Antibodies remaining at the cell surface were removed by acid stripping, following incubation at 37°C, for 15 min in the presence of PMA (recycling, t=30’+15’; n=4). Immunofluorescence analysis was performed to detect internalized Cx43E2HA (white; after permeabilization); the recycled pool was stained without permeabilizing the cells. Insets represent GFP signal of transfected cells. Scale bars 5 μm. P values were derived by Mann-Whitney test. n.s indicates not significant.

Furthermore, live imaging and dynamic tracking of Cx43-positive vesicles in cells overexpressing Cx43-RFP (red fluorescent protein) and GFP-EHD1WT or GFP-EHD1G65R show that Cx43-RFP–positive vesicles were significantly more mobile in cells overexpressing wild-type EHD1 (112.9 nm/s versus 62.9 nm/s in GFP-expressing cells, and 74.2 nm/s in cells overexpressing GFP-EHD1G65R; Figure IVB in the Data Supplement).

Overexpression of EHD1G65R Promotes Recycling of Cx43

To clarify whether the effects mentioned above are due to a role of EHD1 in internalization or recycling of Cx43, we performed antibody feeding assays, using a Cx43 construct with a hemagglutinin-tag on the second extracellular loop (Cx43E2HA). Results on Figure 4B show that, in accordance with previous results with siRNA, internalization of Cx43 was decreased in cells expressing EHD1G65R, when compared with the wild-type EHD1. Surprisingly, the overexpression of the mutant EHD1G65R significantly enhanced the amount of Cx43 being recycled (technical controls included in Figure IVC and IVD in the Data Supplement).

To discard a cell type-specific effect of EHD1 on Cx43 trafficking, these results were confirmed by biochemical analysis in another cell line, Cos-7 cells (Figure VA and VB in the Data Supplement). Additionally, we showed that this altered Cx43 trafficking had an impact on GJIC. Accordingly, the transfer of Lucifer Yellow was impaired after overexpression of wild-type EHD1, whereas the opposite effect was observed in cells overexpressing EHD1G65R (Figure VC in the Data Supplement).

Eps15 Acts as a Bridge Between Cx43 and EHD1

Next, we investigated whether ubiquitin-dependent interaction with Eps15 is involved in EHD1-mediated internalization of Cx43. To address this question, we either silenced Eps15 or overexpressed a mutant version of Eps15ΔUIM (Eps15 that does not contain the ubiquitin-interacting domains) required to recognize ubiquitinated substrates, after which we evaluated the interaction of Cx43 with EHD1. Our results show that, in the absence of Eps15, or after overexpression of Eps15ΔUIM, the amount of myc-EHD1 that coimmunoprecipitated with Cx43 was significantly decreased (Figure 5A and 5B). As expected, Eps15ΔUIM failed to interact with Cx43 (Figure 5B). However, knockdown of EHD1 does not affect the interaction of Cx43 with Eps15 (Figure VIA in the Data Supplement), reinforcing the hypothesis that binding of Eps15 to Cx43 occurs upstream of EHD1 recruitment.

Figure 5.

Figure 5. Eps15 (endocytic adaptor epidermal growth factor receptor substrate 15) acts as a bridge between Cx43 (connexin 43) and EHD1 (Eps15 homology domain-containing protein 1).A, siRNA (small interfering RNA)-mediated knockdown of Eps15 (siEps15) was performed for 48 h in HEK293Cx43+ (human embryonic kidney 293) cells transiently expressing myc-EHD1WT (myc-tagged wild-type EHD1). Controls were performed with transfection of scramble siRNA (siCT). Interaction with EHD1 was analyzed by Western blot (WB), following Cx43 immunoprecipitation (IP; n=4). B, HEK293Cx43+ cells were cotransfected with myc-EHD1WT and wild-type Eps15 (Eps15-WT), Eps15 mutated in the UIM motifs (Eps15-△UIM) or the empty vector. Cx43 was IP, and interaction with EHD1 was analyzed by WB (n=5). C, HEK293Cx43+ cells were cotransfected with myc-EHD1WT and wild-type Eps15 (GFP [green fluorescent protein]-Eps15-WT),Eps15 mutated in the first 2 EH motifs (GFP-Eps15-△95/295) or the empty vector. Cx43 was IP, followed by analysis of interaction with EHD1 (n=4). D, HEK293 cells expressing myc-EHD1WT were cotransfected with wild-type Cx43 (Cx43WT), or Cx43 mutated on Y265 and Y286 (Cx43Y265/286A) for 24 h. Cells were treated with phorbol 12-myristate 13-acetate (PMA) for 15 min, followed by Cx43 IP and WB analysis of interaction with EHD1 (n=7). E, HEK293Cx43+ cells expressing myc-EHD1WT were incubated with 0.45 mol/L sucrose or 10 mmol/L methyl-β-cyclodextrin for 30 min. Cx43 IP was followed by the analysis of interaction with EHD1 by WB (n=7). F, HEK293Cx43+ cells expressing myc-EHD1WT were treated with 80 μmol/L dynasore for 30 min, followed by Cx43 IP (n=4). P values were derived by 2-way ANOVA, followed by Tukey multiple comparison test in D and by Mann-Whitney test in A–C and E and F. n.s indicates not significant.

Next, we evaluated whether the association between Cx43 and EHD1 is maintained in cells overexpressing an Eps15 mutant lacking 2 of its 3 EH-domains (GFP-Eps15Δ95/295), known to mediate protein-protein interactions and the association with phosphorylated phosphoinositides.17 Results in Figure 5C show that in the presence of Eps15Δ95/295, that does not interact with EHD1 (Figure VIB in the Data Supplement), the amount of EHD1 that coimmunoprecipitated with Cx43 was significantly lower, although the interaction between Cx43 and Eps15 was preserved (compare lane 3 and lane 4, Western blot for GFP on the bottom).

Besides ubiquitin-mediated internalization, clathrin-dependent endocytosis of GJ involves recognition of specific sorting motifs on Cx43 by AP (α-adaptin)-2.33 Therefore, we sought to determine whether EHD1 participates in postendocytic trafficking of Cx43 internalized via clathrin-mediated endocytosis. Inhibition of Cx43 internalization, by overexpression of a mutated form of Cx43 in the endocytic Tyr-sorting signals of Cx43 (Cx43Y265/286A), impaired the interaction between EHD1 and Cx43, whereas sucrose, a chemical inhibitor of endocytosis at the level of shallow coated pits formation or receptor clustering did not have an effect (Figure 5D and 5E and Figure VII in the Data Supplement).34,35 Moreover, methyl-β-cyclodextrin and dynasore, which inhibit the progression from shallow clathrin-coated pits to invaginated coated pits, and the dynamin constriction of the clathrin-coated vesicle, respectively, led to increased interaction of EHD1 with Cx43 (Figure 5E and 5F).36,37 Altogether, this suggests that EHD1 preferentially interacts with Cx43 after membrane budding, but before vesicle scission.

Recycling of Cx43 in the Presence of EHD1G65R Involves Retrograde Trafficking Through Trans-Golgi Network

Multiple recycling pathways have been described, including direct recycling from early endosomes to the plasma membrane, through the endocytic recycling compartment, or by retrograde trafficking via the trans-Golgi network (TGN), involving Ras-associated binding (Rab) protein–GTPase, ADP ribosylation factors (ARFs) GTPases, and EHDs.38,39 To address the impact of EHD1 on recycling of Cx43 via the TGN, we evaluated the levels of Cx43 at the plasma membrane after PMA exposure in the presence of brefeldin A (BFA), which not only disrupts the secretory pathway through disassembly of the Golgi complex but can also inhibit Arf1 that acts on the formation of coated vesicles at the TGN.40,41 The treatment with BFA counteracted the PMA-induced increase in the levels of Cx43 at the plasma membrane in cells overexpressing EHD1G65R (Figure VIIIA in the Data Supplement). Furthermore, an increased colocalization of Cx43 with the Golgi COP (coat protein complex)-I was observed in cells overexpressing EHD1G65R, comparing with EHD1WT-expressing cells (Figure 6A, Figure IIIF and VIIIB in the Data Supplement). However, after PMA treatment, the localization of Cx43 within Golgi vesicles was significantly decreased in EHD1G65R-expressing cells (Figure 6A).

Figure 6.

Figure 6. EHD1G65R-mediated recycling of Cx43 (connexin 43) involves retrograde trafficking through the trans-Golgi network.A, HEK293Cx43+ (human embryonic kidney 293) cells transfected with GFP (green fluorescent protein)-EHD1WT (wild-type EHD1) or myc-EHD1G65R for 24 h, were treated or not with phorbol 12-myristate 13-acetate (PMA) for 15 min. Colocalization between Cx43 and COP (coat protein complex)-I was analyzed by confocal microscopy (n=4). Nuclei were stained with DAPI (4′,6-diamidino-2-phenylindole). Insets represent magnified colocalization areas. Scale bars 10 μm. B, Recycling assays were performed in cells overexpressing hemagglutinin (HA)-tagged Cx43 and GFP-EHD1WT or GFP-EHD1G65R, in the presence or absence of 6 μmol/L brefeldin A (BFA). Immunofluorescence analysis was performed to detect recycled Cx43E2HA (white; without permeabilization). Quantification of total fluorescence is expressed as percentage of internalized Cx43 (n=4). Insets represent GFP signal of transfected cells. Scale bars 5 μm. C, Internalization assays were performed in cells overexpressing Cx43E2HA and GFP, GFP-EHD1WT, or GFP-EHD1G65R, by 30 min pulse with anti-HA antibodies at 4°C, followed by incubation at 37°C, for 30 min to follow internalization of Cx43 (t=30 min internalization). Colocalization between internalized Cx43E2HA (cyan) and GCC1 (golgin GRIP and coiled-coil domain-containing 1; magenta) was analyzed by confocal microscopy (n=5). Insets represent magnified colocalization areas. Scale bars 5 μm. P values were derived by 2-way ANOVA, followed by Tukey multiple comparison test in A and B and by Mann-Whitney test in C. n.s indicates not significant.

To confirm the involvement of the TGN in Cx43 recycling, we performed recycling assays in the presence of BFA. Results displayed in Figure 6B show that Cx43 recycling was prevented by BFA in EHD1G65R-expressing cells. Furthermore, colocalization between internalized Cx43 and GCC1 (golgin GRIP and coiled-coil domain-containing 1), one of the Golgins involved in retromer-mediated trafficking, was increased in cells overexpressing EHD1G65R (Figure 6C).

To investigate the impact of EHD1 overexpression on Cx43 turnover, we performed cycloheximide-chase experiments. Our results show that EHD1G65R stabilizes Cx43, whereas the overexpression of wild-type EHD1 does not affect the half-life of Cx43, comparing with control cells (Figure IX in the Data Supplement).

Overexpression of Wild-Type EHD1 Increases Ischemia-Induced Lateralization of Cx43

Finally, to extend the relevance of our data to a pathophysiological context, we performed experiments in isolated adult rat cardiomyocytes subjected to in vitro ischemia, in which GFP, GFP-EHD1WT, or GFP-EHD1G65R were overexpressed through adenoviral infection. Results shown in Figure 7A demonstrate that the ratio between Cx43 localized at the lateral membranes and IDs was higher in response to ischemia in cells overexpressing GFP, and more so in wild-type EHD1-expressing cells, but not in EHD1G65R. N-cadherin staining was used to identify ID structures (Figure XA in the Data Supplement). Importantly, in ischemic cells overexpressing wild-type EHD1, lateralization of Cx43 was accompanied by an increased colocalization with the early endocytic marker EEA1 (Figure 7A), suggesting that EHD1-mediated internalization of Cx43 precedes the remodeling of Cx43 channels.

Figure 7.

Figure 7. Overexpression of EHD1WT (wild-type Eps15 [endocytic adaptor epidermal growth factor receptor substrate 15] homology domain-containing protein 1) increases ischemia (ISCH)-induced lateralization of Cx43 (connexin 43). Isolated adult rat cardiomyocytes were infected with adenovirus encoding GFP (green fluorescent protein), GFP-EHD1WT, or GFP-EHD1G65R for 36 h, before simulation of ISCH. A, Isolated cardiomyocytes were immunostained with antibodies against Cx43 (red) and EEA1 (early endosome antigen 1; white), followed by confocal microscopy imaging. Nuclei were stained with DAPI (4′,6-diamidino-2-phenylindole). Scale bars 20 μm. The amount of lateralized Cx43 (expressed as a ratio to the levels of intercalated disc–localized Cx43) and colocalization with EEA1 were quantified with ImageJ and plotted on graph. Dots represent the values obtained for each individual cardiomyocyte (n=3). Insets represent magnified colocalization areas. B and C, Ca2+ transients were recorded in electrically stimulated (1 Hz) cardiomyocytes via ratiometric measurements obtained from an alternating excitation at 340 and 380 nm of the fluorescent probe FURA-2. Dots represent the values obtained for each individual cardiomyocyte (n=3). P values were derived by 2-way ANOVA, followed by Tukey multiple comparison test in all plots. CT indicates control; and n.s, not significant.

Given its strong correlation with heart contractility, we also evaluated electrically induced Ca2+ transients in this experimental setup. Our data shows that EHD1WT significantly decreased baseline intracellular levels of Ca2+, comparing with GFP-infected cells, an effect that was not present in cells expressing EHD1G65R (Figure 7B, Figure XB in the Data Supplement). Expectedly, ischemia was associated with an increase in the intracellular Ca2+ levels. Moreover, the amplitude of Ca2+ transients was enhanced in cardiomyocytes infected with EHD1G65R both in control conditions and after ischemia without any change in Ca2+ reuptake kinetics, suggesting an improvement in Ca2+ flux by the expression of EHD1G65R (Figure 7C, Figure XC in the Data Supplement).

Cx43 Interacts and Colocalizes With EHD Proteins in Failing Hearts

Myocardial ischemia has been demonstrated to cause both systolic and diastolic dysfunction, being frequently regarded as a major underlying pathogenic factor in both HF with reduced ejection fraction and preserved ejection fraction.42 Given that altered distribution and function of GJ channels have been strongly correlated with HF progression, we proceeded to evaluate whether the association of EHDs and Cx43 is observed in human myocardium biopsies from aortic stenosis patients. Strikingly, the association between Cx43 and EHD1-3 was demonstrated by colocalization of both proteins at the IDs, lateral membranes, and cytoplasm of human cardiomyocytes, as well as by coimmunoprecipitation in total heart lysates (Figure 8A and 8B). Given the restricted access to human controls for comparison, we further evaluated the association of EHDs with Cx43 in an animal model of aortic banding. Our results demonstrate that the interaction between Cx43 and EHD1-3 was significantly increased in hearts subjected to aortic banding (Figure 8C). Although debanding animals presented a trend towards lower Cx43/EHD1-3 interaction levels, when comparing with the banding animals, there was a significant dispersion in these values, likely reflecting the diversity of cardiac reverse remodeling status among this group.23 Strikingly, we found a positive correlation between the levels of Cx43/EHD1-3 interaction and the left ventricle mass, thus reinforcing the role of EHD1 in Cx43-channel remodeling associated with the onset or development of HF (Figure XD in the Data Supplement).

Figure 8.

Figure 8. Cx43 (connexin 43) interacts and colocalizes with EHD (Eps15 [endocytic adaptor epidermal growth factor receptor substrate 15] homology domain-containing protein) 1–3 in failing hearts.A, Human hearts were fixed and stained with antibodies against Cx43 (green) and EHD1-3 (red) and analyzed by confocal fluorescence microscopy. N-cadherin (white) was used as a marker of intercalated discs. Nuclei were stained with DAPI (4′,6-diamidino-2-phenylindole). Insets represent magnified colocalization areas. Scale bar 10 μm. B, Cx43 immunoprecipitation (IP) in human heart biopsies from 3 aortic stenosis patients (No. 1, No. 2, and No. 3). Interaction with endogenous EHD1-3 was analyzed by Western blot. C, Cx43 IP in heart lysates from sham-operated or mice subjected to banding of the ascending aorta (Ba), followed by debanding (Deb). Graph depicts interaction levels (Ba, n=8; Deb, n=6). P values were derived by Mann-Whitney test. n.s indicates not significant.

Discussion

The results gathered in the present study identified EHD proteins as novel players in the pathological remodeling of GJ, by transiently associating with Cx43 during myocardial ischemia. It has been consistently shown that ischemia-induced electrical uncoupling and arrhythmogenesis are closely related to Cx43-channel remodeling, including increased internalization, lateralization and degradation.5 Therefore, the identification of the signals and interacting partners underlying abnormal Cx43 trafficking might open new avenues towards the establishment of therapeutic strategies to prevent ischemia-induced arrhythmias and progression into HF.

Our data show that knockdown of EHD1 impaired the internalization of Cx43-containing channels in HL-1 cardiomyocytes, which was accompanied by an increase of GJIC. However, we did not find an increase in the levels of Triton-insoluble Cx43 after EHD1 silencing. Although these results seem contradictory, previous studies suggested that communicating GJ can be also recovered in the Triton-soluble fraction. In fact, loss of detergent resistance occurs after the removal of Cx43 channels from the tight arrangement within GJ plaques, which can be observed before GJ internalization, while the intercellular coupling is still preserved.43,44

In HL-1 cells subjected to ischemia, we observed an increase in the levels of biotinylated Cx43, likely representing the hemichannel population of Cx43, an effect that was attenuated in cells knocked-down for EHD1. This result is consistent with multiple reports demonstrating that myocardial ischemia-induced Cx43 hemichannel opening, which is implicated in ATP release, as well as intracellular Ca2+ and Na+ overload, ultimately contributing to cell injury.2,45 Nevertheless, the mechanisms underlying ischemia-induced hemichannel opening and accumulation at the cell surface are not clear and might reflect either (1) altered gating that increases single-channel permeability, (2) increased forward trafficking or recycling of Cx43 hemichannels, accompanied by impaired GJ accretion, or (3) increased GJ disassembly, resulting in a higher number of hemichannels present at the plasma membrane. However, our data are insufficient to allow the discrimination between these hypotheses.

Although EHD1 has been mainly associated with functions on endocytic recycling, some studies reported the involvement of EHD1 in the internalization of the LDL-R (low-density lipoprotein receptor), NgCAM, and IGF-1R.18,19 Accordingly, EHD1 was identified as an interacting partner of clathrin and AP-2, forming a complex with SNAP29 (synaptosomal-associated protein 29) to drive clathrin-mediated endocytosis of IGF-1R.18 Nonetheless, the molecular details underlying the roles of EHD1 upon clathrin-dependent and clathrin-independent endocytosis are currently poorly established. The present work, focused on Cx43 trafficking, provides novel mechanistic insights concerning the role of EHD1 on clathrin-mediated internalization. The results of this study demonstrate that interaction of Cx43 with EHD1 was promoted by ischemia, as well as by increased Cx43 phosphorylation and ubiquitination, whereas mutations in the internalization-associated Tyr-sorting signals of Cx43 (Cx43Y265/286A) prevented EHD1 recruitment. Altogether, our data are consistent with a model where interaction with EHD1 occurs downstream of Cx43 internalization, but before clathrin-coated vesicle scission, likely being important to govern postendocytic trafficking of Cx43 channels and impacting GJIC.

Compelling evidence demonstrated that PKC (protein kinase C)-induced phosphorylation of Cx43-S368 precedes ubiquitination and internalization of Cx43 channels.13,46 However, up until now, Cx43 endocytosis driven by the recognition of Tyr-sorting signals or ubiquitin have been tackled as separated events, rather than synergistic mechanisms. Indeed, there are studies demonstrating that Tyr-based motifs on Cx43 enable interaction with clathrin, AP-2, and disabled-2, whereas others reported that Cx43 ubiquitination signals the recruitment of Eps15 to drive GJ endocytosis.6,8,33,47,48 Our data further demonstrate that the interaction of Cx43 with EHD1 requires both the recognition of ubiquitinated Cx43 by Eps15 and functional Tyr-sorting signals.

We show that overexpression of wild-type EHD1 increased the localization of Cx43 within EEA1- and Lamp2-positive endosomes, as well as the internalization rate of the protein. When phosphorylation of Cx43 was stimulated with PMA, in wild-type EHD1-expressing cells, localization of Cx43 within early endosomes was enhanced, whereas colocalization in Lamp2-positive structures was significantly decreased, suggesting that, in these conditions, EHD1 is either (1) trapping Cx43 in early endosomes, (2) stimulating recycling of the channels, or (3) promoting Cx43 degradation. Biotinylation experiments revealed that the levels of plasma membrane-localized Cx43 are significantly higher in cells expressing EHD1G65R in response to short treatment with PMA, compared with EHD1WT, thereby implicating EHD1 in the regulation of Cx43 trafficking and distribution at the plasma membrane. In accordance, antibody feeding assays corroborated the hypothesis of EHD1 impacting Cx43 recycling. Interestingly, in EHD1G65R-expressing cells, we observed higher colocalization levels between internalized Cx43 and GCC1, as well as impaired Cx43 recycling in the presence of BFA, suggesting that recycling of Cx43 involves retrograde trafficking via the TGN, reaching the plasma membrane through the secretory pathway. In agreement, recent studies reported a role for EHD1/3 upon endosome-to-Golgi retrieval of lysosomal hydrolase receptors.49 Overall, our data are consistent with a model where wild-type EHD1 accelerates clathrin-mediated endocytosis of Cx43, followed by rapid dissociation of these protein complexes, contributing to retain Cx43 within early/sorting endosomes.

Lastly, in adult cardiomyocytes subjected to ischemia, overexpression of wild-type EHD1 increased the colocalization of Cx43 with EEA1 and concomitant lateralization of Cx43, suggesting that EHD1-mediated internalization of Cx43 precedes its mislocalization to the lateral cardiomyocyte membranes via a recycling-like mechanism. In cardiomyocytes expressing EHD1G65R, neither of these effects was observed, with Cx43 staining being preserved at the IDs during ischemia. Furthermore, one may argue that the dedifferentiation of our isolated cardiomyocytes might affect the lateralization phenotype observed and, therefore, underestimate our results. Nevertheless, with the currently available tools, it is difficult to maintain adult cardiomyocytes in culture long enough to allow the expression of adenovirus-carrying genes, without losing proper cardiomyocyte architecture. Another drawback of the present strategy is the fact that it does not allow to assess GJIC nor to discriminate between stabilization of Cx43 in the form of hemichannels or GJ, which would have distinct functional outcomes. Given that adult cardiomyocytes are terminally differentiated cells, with specific mechanisms of polarized trafficking that may differ from those described in nonmyocyte or nonpolarized cells, this likely explains some discrepancies between the findings obtained among our cell-based models. In this work, we used different cellular models, in a complementary and comprehensive manner, not only to overcome experimental approach limitations but also to discard a species-specific mechanism of EHD1-dependent trafficking. Despite the overall consistency of the results obtained with the different models used, it may pose some restrictions on the conclusions of our study.

We also measured Ca2+ transients in adult cardiomyocytes, which revealed that wild-type EHD1 significantly decreased the intracellular Ca2+ levels at baseline, whereas EHD1G65R significantly increased the transients’ amplitude, both in control and ischemic cells. The role of EHDs upon the regulation of cardiac membrane proteins is not without precedent. EHD3 was reported to control the trafficking of NCX (Na+/Ca2+ exchanger)-1, L-type and T-type Ca2+ channels, whereas EHD4 regulates abnormal postinfarction trafficking of voltage-gated Na+ channel (Nav)-1.5.50,51 Therefore, by affecting the trafficking of Na+ and Ca2+ channels, EHDs may impact cardiac excitability and function independently of GJIC.

Besides being involved in ischemia-induced lateralization of Cx43, our results demonstrate an association between Cx43 and EHDs in failing human hearts. Interestingly, although the interaction of EHDs with Cx43 was transiently increased during acute myocardial ischemia and returned to the basal levels upon reperfusion, in the animal model of aortic banding, higher levels of interaction between EHD1-3 and Cx43 were maintained even after debanding, suggesting that chronic EHD-associated GJ remodeling in failing hearts is irreversible, ultimately contributing to sustained cardiac damage.23 Similar to what we have shown in the ex vivo model, these correlative data are not enough to determine causation. Nonetheless, the results gathered in the present article, with the use of different cell-based models, consistently suggest in a coherent and integrated manner that EHD1 participates in the remodeling of GJs, both after acute myocardial ischemia and in HF.

Overall, the results presented in this study bring important insights not only to the field of GJ and intercellular communication, by establishing EHD1 as a novel interacting partner involved in the trafficking of Cx43 but also constitute a significant contribution to unveil the molecular mechanisms associated with pathological remodeling of Cx43 in cardiovascular diseases.

Nonstandard Abbreviations and Acronyms

AP-2

α-adaptin

BFA

brefeldin A

COP

coat protein complex

Cx

connexins

EHD1

Eps15 homology domain-containing protein

Eps15

epidermal growth factor receptor substrate 15

Eps15ΔUIM

Eps15 that does not contain the ubiquitin-interacting domains

GFP

green fluorescent protein

GJ

gap junction

GJIC

GJ intercellular communication

HF

heart failure

IDs

intercalated discs

IGF-1R

insulin-like growth factor 1 receptor

LDL-R

low-density lipoprotein receptor

MHC

major histocompatibility complex

NgCAM

L1/neuron-glia cell adhesion molecule

PKC

protein kinase C

PMA

phorbol 12-myristate 13-acetate

RFP

red fluorescent protein

SNAP29

synaptosomal-associated protein 29

TGN

trans-Golgi network

Acknowledgments

We thank Professor Francisca Lago Paz (Servizo Galego de Saude (SERGAS), Instituto de Investigación Sanitaria de Santiago (IDIS), Department of Cellular and Molecular Cardiology, Centro de Investigación Biomédica en Red de Enfermedades Cardiovasculares (CIBERCV), Santiago de Compostela, Spain), whom helped to establish the primary cultures of human adult cardiomyocytes, and Doctor Francisco Caramelo (Faculty of Medicine, University of Coimbra), for the help with the statistical analysis.

Supplemental Materials

Expanded Materials and Methods

Figures I–X.

Major Resources Table

Footnotes

For Sources of Funding and Disclosures, see page e112.

The Data Supplement is available with this article at https://www.ahajournals.org/doi/suppl/10.1161/CIRCRESAHA.119.316502.

Correspondence to: Henrique Girão, PhD, Coimbra Institute for Clinical and Biomedical Research, Faculty of Medicine, University of Coimbra, Health Science Campus, Polo III, Azinhaga de Santa Comba 3000-548, Coimbra, Portugal. Email

References

  • 1. Hulsmans M, Clauss S, Xiao L, Aguirre AD, King KR, Hanley A, Hucker WJ, Wülfers EM, Seemann G, Courties G, et al. Macrophages facilitate electrical conduction in the heart.Cell. 2017; 169:510–522.e20. doi: 10.1016/j.cell.2017.03.050CrossrefMedlineGoogle Scholar
  • 2. Fontes MSC, Van Veen TAB, De Bakker JMT, Van Rijen HVM. Functional consequences of abnormal Cx43 expression in the heart.Biochim Biophys Acta. 2012; 1818:2020–2029. doi: 10.1016/j.bbamem.2011.07.039CrossrefMedlineGoogle Scholar
  • 3. Unudurthi SD, Wolf RM, Hund TJ. Role of sinoatrial node architecture in maintaining a balanced source-sink relationship and synchronous cardiac pacemaking.Front Physiol. 2014; 5:446. doi: 10.3389/fphys.2014.00446CrossrefMedlineGoogle Scholar
  • 4. Aasen T, Johnstone S, Vidal-Brime L, Lynn KS, Koval M. Connexins: synthesis, post-translational modifications, and trafficking in health and disease.Int J Mol Sci. 2018; 19:1296. doi: 10.3390/ijms19051296CrossrefGoogle Scholar
  • 5. Martins-Marques T, Catarino S, Marques C, Pereira P, Girão H. To beat or not to beat: degradation of Cx43 imposes the heart rhythm.Biochem Soc Trans. 2015; 43:476–481. doi: 10.1042/BST20150046CrossrefMedlineGoogle Scholar
  • 6. Catarino S, Ramalho JS, Marques C, Pereira P, Girão H. Ubiquitin-mediated internalization of connexin43 is independent of the canonical endocytic tyrosine-sorting signal.Biochem J. 2011; 437:255–267. doi: 10.1042/BJ20102059CrossrefMedlineGoogle Scholar
  • 7. Leithe E, Kjenseth A, Sirnes S, Stenmark H, Brech A, Rivedal E. Ubiquitylation of the gap junction protein connexin-43 signals its trafficking from early endosomes to lysosomes in a process mediated by Hrs and Tsg101.J Cell Sci. 2009; 122:3883–3893. doi: 10.1242/jcs.053801CrossrefMedlineGoogle Scholar
  • 8. Girão H, Catarino S, Pereira P. Eps15 interacts with ubiquitinated Cx43 and mediates its internalization.Exp Cell Res. 2009; 315:3587–3597. doi: 10.1016/j.yexcr.2009.10.003CrossrefMedlineGoogle Scholar
  • 9. Martins-Marques T, Catarino S, Marques C, Matafome P, Ribeiro-Rodrigues T, Baptista R, Pereira P, Girão H. Heart ischemia results in connexin43 ubiquitination localized at the intercalated discs.Biochimie. 2015; 112:196–201. doi: 10.1016/j.biochi.2015.02.020CrossrefMedlineGoogle Scholar
  • 10. Martins-Marques T, Catarino S, Zuzarte M, Marques C, Matafome P, Pereira P, Girão H. Ischaemia-induced autophagy leads to degradation of gap junction protein connexin43 in cardiomyocytes.Biochem J. 2015; 467:231–245. doi: 10.1042/BJ20141370CrossrefMedlineGoogle Scholar
  • 11. Smyth JW, Zhang SS, Sanchez JM, Lamouille S, Vogan JM, Hesketh GG, Hong T, Tomaselli GF, Shaw RM. A 14-3-3 mode-1 binding motif initiates gap junction internalization during acute cardiac ischemia.Traffic. 2014; 15:684–699. doi: 10.1111/tra.12169CrossrefMedlineGoogle Scholar
  • 12. Fong JT, Nimlamool W, Falk MM. EGF induces efficient Cx43 gap junction endocytosis in mouse embryonic stem cell colonies via phosphorylation of Ser262, Ser279/282, and Ser368.FEBS Lett. 2014; 588:836–844. doi: 10.1016/j.febslet.2014.01.048CrossrefMedlineGoogle Scholar
  • 13. Leithe E, Rivedal E. Epidermal growth factor regulates ubiquitination, internalization and proteasome-dependent degradation of connexin43.J Cell Sci. 2004; 117:1211–1220. doi: 10.1242/jcs.00951CrossrefMedlineGoogle Scholar
  • 14. Leithe E, Rivedal E. Ubiquitination and down-regulation of gap junction protein connexin-43 in response to 12-O-tetradecanoylphorbol 13-acetate treatment.J Biol Chem. 2004; 279:50089–50096. doi: 10.1074/jbc.M402006200CrossrefMedlineGoogle Scholar
  • 15. Ribeiro-Rodrigues TM, Catarino S, Marques C, Ferreira JV, Martins-Marques T, Pereira P, Girão H. AMSH-mediated deubiquitination of Cx43 regulates internalization and degradation of gap junctions.FASEB J. 2014; 28:4629–4641. doi: 10.1096/fj.13-248963CrossrefMedlineGoogle Scholar
  • 16. Martins-Marques T, Anjo SI, Pereira P, Manadas B, Girão H. Interacting network of the Gap Junction (GJ) Protein Connexin43 (Cx43) is modulated by ischemia and reperfusion in the heart.Mol Cell Proteomics. 2015; 14:3040–3055. doi: 10.1074/mcp.M115.052894CrossrefMedlineGoogle Scholar
  • 17. Naslavsky N, Caplan S. EHD proteins: key conductors of endocytic transport.Trends Cell Biol. 2011; 21:122–131. doi: 10.1016/j.tcb.2010.10.003CrossrefMedlineGoogle Scholar
  • 18. Rotem-Yehudar R, Galperin E, Horowitz M. Association of insulin-like growth factor 1 receptor with EHD1 and SNAP29.J Biol Chem. 2001; 276:33054–33060. doi: 10.1074/jbc.M009913200CrossrefMedlineGoogle Scholar
  • 19. Yap CC, Lasiecka ZM, Caplan S, Winckler B. Alterations of EHD1/EHD4 protein levels interfere with L1/NgCAM endocytosis in neurons and disrupt axonal targeting.J Neurosci. 2010; 30:6646–6657. doi: 10.1523/JNEUROSCI.5428-09.2010CrossrefMedlineGoogle Scholar
  • 20. Jović M, Naslavsky N, Rapaport D, Horowitz M, Caplan S. EHD1 regulates beta1 integrin endosomal transport: effects on focal adhesions, cell spreading and migration.J Cell Sci. 2007; 120:802–814. doi: 10.1242/jcs.03383CrossrefMedlineGoogle Scholar
  • 21. Caplan S, Naslavsky N, Hartnell LM, Lodge R, Polishchuk RS, Donaldson JG, Bonifacino JS. A tubular EHD1-containing compartment involved in the recycling of major histocompatibility complex class I molecules to the plasma membrane.EMBO J. 2002; 21:2557–2567. doi: 10.1093/emboj/21.11.2557CrossrefMedlineGoogle Scholar
  • 22. Curran J, Makara MA, Mohler PJ. Endosome-based protein trafficking and Ca(2+) homeostasis in the heart.Front Physiol. 2015; 6:34. doi: 10.3389/fphys.2015.00034CrossrefMedlineGoogle Scholar
  • 23. Miranda-Silva D, Gonçalves-Rodrigues P, Almeida-Coelho J, Hamdani N, Lima T, Conceição G, Sousa-Mendes C Cláudia-Moura, González A, Díez J, et al. Characterization of biventricular alterations in myocardial (reverse) remodelling in aortic banding-induced chronic pressure overload.Sci Rep. 2019; 9:2956. doi: 10.1038/s41598-019-39581-9CrossrefMedlineGoogle Scholar
  • 24. Sequeira V, Najafi A, Mcconnell M, Fowler ED, Bollen IAE, Wüst RCI, dos Remedios C, Helmes M, White E, Stienen GJ, et al. Synergistic role of ADP and Ca2+ in diastolic myocardial stiffness.J Physiol. 2015; 593:3899–3916. doi: 10.1113/JP270354.CrossrefMedlineGoogle Scholar
  • 25. Ribeiro-Rodrigues TM, Laundos TL, Pereira-Carvalho R, Batista-Almeida D, Pereira R, Coelho-Santos V, Silva AP, Fernandes R, Zuzarte M, Enguita FJ, et al. Exosomes secreted by cardiomyocytes subjected to ischaemia promote cardiac angiogenesis.Cardiovasc Res. 2017; 113:1338–1350. doi: 10.1093/cvr/cvx118CrossrefMedlineGoogle Scholar
  • 26. Rodríguez-Penas D, Feijóo-Bandín S, Lear PV, Mosquera-Leal A, García-Rúa V, Otero MF, Rivera M, Gualillo O, González-Juanatey JR, Lago F. Aliskiren affects fatty-acid uptake and lipid-related genes in rodent and human cardiomyocytes.Biochem Pharmacol. 2011; 82:491–504. doi: 10.1016/J.BCP.2011.05.021.CrossrefMedlineGoogle Scholar
  • 27. Claycomb WC, Lanson NA, Stallworth BS, Egeland DB, Delcarpio JB, Bahinski A, Izzo NJ. HL-1 cells: a cardiac muscle cell line that contracts and retains phenotypic characteristics of the adult cardiomyocyte.Proc Natl Acad Sci U S A. 1998; 95:2979–2984. doi: 10.1073/pnas.95.6.2979CrossrefMedlineGoogle Scholar
  • 28. Dias P, Desplantez T, El-Harasis MA, Chowdhury RA, Ullrich ND, Cabestrero de Diego A, Peters NS, Severs NJ, MacLeod KT, Dupont E. Characterisation of connexin expression and electrophysiological properties in stable clones of the HL-1 myocyte cell line.PLoS One. 2014; 9:e90266. doi: 10.1371/journal.pone.0090266CrossrefMedlineGoogle Scholar
  • 29. Cone AC, Cavin G, Ambrosi C, Hakozaki H, Wu-Zhang AX, Kunkel MT, Newton AC, Sosinsky GE. Protein kinase Cδ-mediated phosphorylation of Connexin43 gap junction channels causes movement within gap junctions followed by vesicle internalization and protein degradation.J Biol Chem. 2014; 289:8781–8798. doi: 10.1074/jbc.M113.533265CrossrefMedlineGoogle Scholar
  • 30. Dunn CA, Lampe PD. Injury-triggered Akt phosphorylation of Cx43: a ZO-1-driven molecular switch that regulates gap junction size.J Cell Sci. 2014; 127:455–464. doi: 10.1242/jcs.142497CrossrefMedlineGoogle Scholar
  • 31. Naslavsky N, Rahajeng J, Sharma M, Jovic M, Caplan S. Interactions between EHD proteins and Rab11-FIP2: a role for EHD3 in early endosomal transport.Mol Biol Cell. 2006; 17:163–177. doi: 10.1091/mbc.e05-05-0466CrossrefMedlineGoogle Scholar
  • 32. de Araùjo MEG, Hube LA, Stasyk T. Isolation of endocitic organelles by density gradient centrifugation.Methods in Molecular Biology (Clifton, NJ). 2008; 424:317–331. doi: 10.1007/978-1-60327-064-9_25.CrossrefMedlineGoogle Scholar
  • 33. Fong JT, Kells RM, Falk MM. Two tyrosine-based sorting signals in the Cx43 C-terminus cooperate to mediate gap junction endocytosis.Mol Biol Cell. 2013; 24:2834–2848. doi: 10.1091/mbc.E13-02-0111CrossrefMedlineGoogle Scholar
  • 34. Chen CL, Hou WH, Liu IH, Hsiao G, Huang SS, Huang JS. Inhibitors of clathrin-dependent endocytosis enhance TGFbeta signaling and responses.J Cell Sci. 2009; 122:1863–1871. doi: 10.1242/jcs.038729CrossrefMedlineGoogle Scholar
  • 35. Hansen SH, Sandvig K, van Deurs B. Clathrin and HA2 adaptors: effects of potassium depletion, hypertonic medium, and cytosol acidification.J Cell Biol. 1993; 121:61–72. doi: 10.1083/jcb.121.1.61CrossrefMedlineGoogle Scholar
  • 36. Dutta D, Donaldson JG. Search for inhibitors of endocytosis: intended specificity and unintended consequences.Cell Logist. 2012; 2:203–208. doi: 10.4161/cl.23967CrossrefMedlineGoogle Scholar
  • 37. Rodal SK, Skretting G, Garred O, Vilhardt F, van Deurs B, Sandvig K. Extraction of cholesterol with methyl-beta-cyclodextrin perturbs formation of clathrin-coated endocytic vesicles.Mol Biol Cell. 1999; 10:961–974. doi: 10.1091/mbc.10.4.961CrossrefMedlineGoogle Scholar
  • 38. Grant BD, Donaldson JG. Pathways and mechanisms of endocytic recycling.Nat Rev Mol Cell Biol. 2009; 10:597–608. doi: 10.1038/nrm2755CrossrefMedlineGoogle Scholar
  • 39. Poirier MG, Eroglu S, Marko JF. The bending rigidity of mitotic chromosomes.Mol Biol Cell. 2002; 13:2170–2179. doi: 10.1091/mbc.01-08-0401CrossrefMedlineGoogle Scholar
  • 40. Nebenführ A, Ritzenthaler C, Robinson DG. Brefeldin A: deciphering an enigmatic inhibitor of secretion.Plant Physiol. 2002; 130:1102–1108. doi: 10.1104/pp.011569CrossrefMedlineGoogle Scholar
  • 41. Saint-Pol A, Yélamos B, Amessou M, Mills IG, Dugast M, Tenza D, Schu P, Antony C, McMahon HT, Lamaze C, et al. Clathrin adaptor epsinR is required for retrograde sorting on early endosomal membranes.Dev Cell. 2004; 6:525–538. doi: 10.1016/s1534-5807(04)00100-5CrossrefMedlineGoogle Scholar
  • 42. Hwang SJ, Melenovsky V, Borlaug BA. Implications of coronary artery disease in heart failure with preserved ejection fraction.J Am Coll Cardiol. 2014; 63:2817–2827. doi: 10.1016/j.jacc.2014.03.034CrossrefMedlineGoogle Scholar
  • 43. Mollerup S, Hofgaard JP, Braunstein TH, Kjenseth A, Leithe E, Rivedal E, Holstein-Rathlou NH, Nielsen MS. Norepinephrine inhibits intercellular coupling in rat cardiomyocytes by ubiquitination of connexin43 gap junctions.Cell Commun Adhes. 2011; 18:57–65. doi: 10.3109/15419061.2011.611920CrossrefMedlineGoogle Scholar
  • 44. Sirnes S, Leithe E, Rivedal E. The detergent resistance of Connexin43 is lost upon TPA or EGF treatment and is an early step in gap junction endocytosis.Biochem Biophys Res Commun. 2008; 373:597–601. doi: 10.1016/j.bbrc.2008.06.095CrossrefMedlineGoogle Scholar
  • 45. Su Z, Li F, Sugishita K, Barry WH, Ueda I. Activation of connexin-43 hemichannels can elevate [Ca2+]iand [Na+]iin rabbit ventricular myocytes during metabolic inhibition.J Mol Cell Cardiol. 2002; 33:2145–2155. doi:10.1006/jmcc.2001.1477.Google Scholar
  • 46. Popolo A, Morello S, Sorrentino R, Pinto A. Antiadrenergic effect of adenosine involves connexin 43 turn-over in H9c2 cells.Eur J Pharmacol. 2013; 715:56–61. doi: 10.1016/j.ejphar.2013.06.019CrossrefMedlineGoogle Scholar
  • 47. Gumpert AM, Varco JS, Baker SM, Piehl M, Falk MM. Double-membrane gap junction internalization requires the clathrin-mediated endocytic machinery.FEBS Lett. 2008; 582:2887–2892. doi: 10.1016/j.febslet.2008.07.024CrossrefMedlineGoogle Scholar
  • 48. Kametaka S, Moriyama K, Burgos P V, Eisenberg E, Greene LE, Mattera R, et al. Internalization of large double-membrane intercellular vesicles by a clathrin-dependent endocytic process.Mol Biol Cell. 2007; 18:2991–3001. doi: 10.1091/mbc.E06.CrossrefMedlineGoogle Scholar
  • 49. McKenzie J, Naslavsky N, Sheff D, Altan-Bonnet N, Caplan S. EHD3 regulates early-endosome-to-Golgi transport and preserves Golgi morphology.J Cell Sci. 2009; 122:389–400. doi:10.1242/jcs.037051.CrossrefMedlineGoogle Scholar
  • 50. Gudmundsson H, Hund TJ, Wright PJ, Kline CF, Snyder JS, Qian L, Koval OM, Cunha SR, George M, Rainey MA, et al. EH domain proteins regulate cardiac membrane protein targeting.Circ Res. 2010; 107:84–95. doi: 10.1161/circresaha.110.216713.LinkGoogle Scholar
  • 51. Dun W, Danilo P, Mohler PJ, Boyden PA. Microtubular remodeling and decreased expression of Nav1.5 with enhanced EHD4 in cells from the infarcted heart.Life Sci. 2018; 201:72–80. doi: 10.1016/j.lfs.2018.03.024CrossrefMedlineGoogle Scholar

Novelty and Significance

What Is Known?

  • Several cardiovascular diseases, including ischemia and heart failure, are associated with Cx43 (connexin 43) remodeling and impairment of intercellular communication.

  • The interactome of Cx43 in the heart changes in pathological conditions.

  • Although several proteins have been implicated in Cx43 trafficking, the machinery driving ischemia-induced lateralization of Cx43 is unknown.

What New Information Does This Article Contribute?

  • The interaction between EHD1 (Eps15 [endocytic adaptor epidermal growth factor receptor substrate 15] homology domain-containing protein 1) and Cx43 increases in cardiomyocytes subjected to ischemia.

  • The binding to EHD1 relies on Cx43 phosphorylation and ubiquitination and is bridged by Eps15.

  • EHD1 overexpression enhances ischemia-induced lateralization of Cx43 in cardiomyocytes.

Several cardiomyopathies have been characterized by arrhythmogenesis and may be associated with impaired GJIC (GJ intercellular communication), including acute myocardial infarction and heart failure. Gap junction–mediated intercellular communication through low resistance Cx43 channels localized at the intercalated discs is vital for electrical impulse propagation that underpins synchronized heart contraction. Defects in Cx43 localization with a concomitant impairment of electrical conduction have been associated with irregular heart rhythm, which can ultimately contribute to heart failure. In this study, we show that the interaction between Cx43 and EHD1 increased in cardiomyocytes subjected to ischemia. In the absence of EHD1, Cx43 accumulated at the plasma membrane with a consequent increase of gap junction–mediated intercellular communication. In adult cardiomyocytes, EHD1 was implicated in ischemia-induced removal of Cx43 from the intercalated discs and its further accumulation at the lateral membrane. Overall, this study identifies a novel Cx43 interacting partner that is involved in the intracellular trafficking of Cx43 channels as well as demonstrate a mechanism whereby Cx43 is internalized and lateralized in cardiomyocytes subjected to ischemia.

eLetters(0)

eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.