Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Circ Res. Author manuscript; available in PMC 2010 June 28.
Published in final edited form as:
PMCID: PMC2892879

Structural and Molecular Mechanisms of Gap Junction Remodeling in Epicardial Border Zone Myocytes following Myocardial Infarction


Lateralization of the ventricular gap junction protein connexin (Cx)43 occurs in epicardial border zone myocytes following myocardial infarction (MI) and is arrhythmogenic. Alterations in Cx43 protein partners have been hypothesized to play a role in lateralization although mechanisms by which this occurs are unknown. To examine potential mechanisms we did nuclear magnetic resonance, yeast 2-hybrid, and surface plasmon resonance studies and found that the SH3 domain of the tyrosine kinase c-Src binds to the Cx43 scaffolding protein zonula occludens (ZO)-1 with a higher affinity than does Cx43. This suggests c-Src outcompetes Cx43 for binding to ZO-1, thus acting as a chaperone for ZO-1 and causing unhooking from Cx43. To determine whether c-Src/ZO-1 interactions affect Cx43 lateralization within the epicardial border zone, we performed Western blot, immunoprecipitation, and immunolocalization for active c-Src (p-cSrc) post-MI using a canine model of coronary occlusion. We found that post-MI p-cSrc interacts with ZO-1 as Cx43 begins to decrease its interaction with ZO-1 and undergo initial loss of intercalated disk localization. This indicates that the molecular mechanisms by which Cx43 is lost from the intercalated disk following MI includes an interaction of p-cSrc with ZO-1 and subsequent loss of scaffolding of Cx43 leaving Cx43 free to diffuse in myocyte membranes from areas of high Cx43, as at the intercalated disk, to regions of lower Cx43 content, the lateral myocyte membrane. Therefore shifts in Cx43 protein partners may underlie, in part, arrhythmogenesis in the post-MI heart.

Keywords: arrhythmia, cardiac gap junction connexins, myocardial ischemia, cell–cell coupling

In adult heart gap junctions are formed by connexin (Cx)43 protein localized to the intercalated disk (ID). They are responsible for forming low resistance conduits for electric conduction. Pathological conditions, including myocardial infarction (MI), induce striking changes in the morphology of Cx43 characterized by relocalization from the ID to non disk lateral membranes.1 In the infarct epicardial border zone (EBZ), which is made up of myocytes that survive adjacent to the necrosing area, this lateralization is associated with loss of gap junctional coupling between myocytes leading to conduction block and reentrant arrhythmias.1,2 Whereas lateralization of Cx43 is an almost ubiquitous response to cardiac pathology,1,35 mechanisms of remodeling are unknown. Because this lateralization may be an important factor which alters conduction properties that lead to reentrant arrhythmias, elucidating molecular mechanisms underlying this process is an important undertaking.

Localization of Cx43 at IDs is associated with interaction of Cx43 with the PDZ-containing scaffolding protein zonula occludens (ZO)-1.6,7 Studies performed in cell culture suggest that Cx43/ZO-1 interactions are disrupted under conditions of low intracellular pH.8 An association of Cx43 with activated c-Src (phosphorylated c-Src [p-cSrc]), a plasma membrane-associated tyrosine kinase, has been demonstrated at low intracellular pH and correlates with an inhibition of gap junction intercellular communication.9 Our previous studies have demonstrated that binding of the c-Src SH3 domain has global effects on the Cx43 carboxyl-terminal domain (Cx43CT) structure. Moreover, the presence of SH3 disrupted the association of the Cx43CT/ZO-1 PDZ-2 domain complex.10 This suggests that there is interdependent binding of Cx43 to ZO-1 and Cx43 to p-cSrc, indicating that switches in Cx43 molecular partner interactions facilitates regulation (eg, channel gating, internalization, and remodeling) of channel function.

In the present study, we examined the molecular mechanism of Cx43 lateralization seen in EBZ ventricular myocytes following MI using a canine model of MI in which gap junction lateralization has been shown to occur.1 We found that the interactions of Cx43 to ZO-1 were lost, accompanied by decreased colocalization of these proteins at the ID. Based on our previous work indicating p-cSrc may disrupt the Cx43/ZO-1 interaction10 and our studies here identifying a direct interaction between ZO-1 and p-cSrc, we formulated a model whereby p-cSrc initially interacts with the carboxyl-terminal domain of Cx43 (Cx43CT), causing structural changes and subsequently dissociates ZO-1 via a chaperoning of ZO-1 from Cx43.

Materials and Methods

Canine Model of MI

All animal studies were performed in accordance with Columbia University Institutional Animal Care and Use Committee approval (AAAA6210). MI was produced in 15 mongrel dogs weighing 25 to 30 kg by a 2-stage ligation of the left anterior descending coronary artery near its origin, as described previously.11 EBZ tissue (≈1-mm-thick samples) was dissected from the epicardial surface of the infarcted anterior left ventricle. For details, see the online data supplement, available at

Western Blot and Coimmunoprecipitation

For both Western blot (50 μg) and coimmunoprecipitation (co-IP) studies, normal and EBZ tissues were lysed in complete lysis buffer (50 mmol/L Tris-HCl [pH 7.4], 0.25 mmol/L, Na-deoxycholate, 150 mmol/L NaCl, 2 mmol/L EGTA, 0.1 mmol/L Na3VO4, 10 mmol/L NaF, 1 mmol/L phenyl methyl sulfonyl fluoride, 1% Triton-X 100, one-half tablet of Complete Protease Inhibitor [Roche]), sonicated for 30 seconds, maintained on ice for 30 minutes, and then triturated and spun at 10 000 rpm for 10 minutes. Total protein was assessed using BCA protein assay kit (Pierce) and then used for Western blot and immunoprecipitation studies as published previously.12


Rapidly frozen epicardial samples were sectioned (15 μm) and stained for Cx43 (monoclonal, Chemicon), ZO-1 (polyclonal, Invitrogen), and p-cSrc (polyclonal, Cell Signaling). Images were processed using Leica Software and quantification of Cx43 overlay with p-cSrc and ZO-1 was analyzed using Image J.

Expression and Purification of Recombinant-Tagged Proteins

Unlabeled, 15N-labeled, and 15N13C-labeled Cx43CT wild-type (residues 255 to 382 of rat Cx43), Cx43CT296–382, Cx43CT296–350, Cx43CT256–382, the SH3 domain of c-Src, and the PDZ-2 domain of ZO-1 were expressed and purified as described previously10 and confirmed for purity/degradation by SDS-PAGE.

Yeast Two-Hybrid

The Saccharomyces cerevisiae strain AH109 was maintained on yeast extract, peptone, and dextrose (YPD) agar plates. Cotransformation was performed by the lithium acetate procedure as described in the instructions for the MATCHMAKER 2-hybrid kit (Clontech). For colony growth assays, AH109 cotransformants were streaked on plates lacking Leu and Trp and allowed to grow at 30°C, for 3 days or until colonies were large enough for further assays. An average of 3 colonies were chosen, suspended in water, equilibrated to the same optical density at 600 nm, and restreaked on plates lacking Leu and Trp (+HIS) or plates also lacking his (−HIS). Controls involved cotransformations with GAL4ad-SV40 large T antigen and GAL4bd-pVA3 murine p53.

Nuclear Magnetic Resonance

The full strategy for solving the Cx43CT/c-Src SH3 domain structure is described in the online data supplement and shown in Online Figure I. NMR data were acquired at 7°C using an INOVA 600 Varian spectrometer fitted with a cold probe. The buffer conditions for each protein were 1× PBS at pH 5.8. Experimental data to determine distance constraints, as well as the methods used for structure modeling, refinement, visualization, and evaluation, have been described previously.13 Binding isotherms from 15N-HSQC titration experiments were calculated with ORIGIN 7.0 software (Microcal Software Inc).

Surface Plasmon Resonance

Protein immobilization on a Biacore 3000 sensor chip was performed as described previously.14 Before loading the ZO-1 PDZ-2 domain, the chip was equilibrated with PBS (pH 5.8) containing 0.005% surfactant P-20 (PBS-P). The PDZ-2 domain at various concentrations in the same buffer was injected in duplicate over the immobilized c-Src SH3 domain. After injection of the PDZ-2 domain, PBS-P was reintroduced with a 300-second lag time to start dissociation. The chip was regenerated to baseline by injection of 5 μL of 100 mmol/L HCl.

Translational Diffusion Experiment

Translation diffusion coefficients were measured by a water-sLED experiment, which includes water-selective pulses to suppress the solvent signal, combined with the longitudinal eddy current delay that allows transient perturbations to disappear. A total of 128 scans were collected into a 64K data block using a sweep width of 8510 Hz.

Statistical Analysis

For all biochemical studies, comparison across groups was performed using a standard ANOVA test. Results were considered significant at the P<0.05 level.


SH3 Domain of c-Src Directly Interacts With the PDZ-2 Domain of ZO-1 In Vitro

Scanning the amino acid sequence of the second PDZ domain (PDZ-2) of ZO-1 indicated a putative SH3 binding domain (PXXPXK) upstream from the first β-sheet (Figure 1A). To determine whether the SH3 domain of c-Src bound to this region, we used yeast 2-hybrid (Y2H) techniques with the PDZ-2 domain expressed in AH109 yeast cells as a GAL4 DNA-binding domain (bd) fusion protein and the SH3 domain coexpressed in the same cells as a GAL4 transactivation domain (ad) fusion protein. We found direct binding between the SH3 and PDZ-2 domains (Figure 1B). Using both the surface plasmon resonance (Figure 1C) and NMR 15N-HSQC (Figure 1D) experiments, we confirmed the direct interaction between the SH3 and PDZ-2 domains. Analysis of NMR spectra for PDZ-2 residues affected by interaction with SH3 indicate a significant overlap with PDZ-2 residues affected by interaction with Cx43CT (Figure 1E).

Figure 1
Interaction between the c-Src SH3 and the ZO-1 PDZ-2 domains. A, Sequence of the ZO-1 PDZ-2 domain used in this study. Residues comprising the SH3 binding motif are colored red. B, The S cerevisiae yeast strain AH109 was cotransformed with GAL4ad and ...

Localization of Cx43 With p-cSrc in the EBZ of Myocardial Infarcts

To determine whether p-cSrc was located at the ID during the time of Cx43/ZO-1 unhooking, we examined colocalization of p-cSrc and Cx43 at 1 hour post-CO, when Cx43/ZO-1 interactions were significantly altered (as shown in the next section). There was little p-cSrc staining seen in normal epicardium (Figure 2A) and none where Cx43 was localized (Figure 2B). Overlay images of p-cSrc and Cx43 show no interaction in normal epicardium (Figure 2C and 2D). In contrast, we found p-cSrc in EBZ myocytes at 1 hour post-CO, both intracellularly as well as at the cell membranes particularly in the region of the ID (Figure 2E). Overlay of Cx43 images with the p-cSrc (Figure 2F) showed spots where the 2 proteins colocalized (Figure 2G), seen more clearly in the magnified inset box (Figure 2H). Myocyte orientation is noted by the cartoon at the right. Transverse sections were immunostained for Cx43 and p-cSrc (Figure 2I). Arrows indicate that colocalization occurs within the en face region of the ID. Quantitative analysis of overlay images, including those at 30 minutes and 3 hour post-CO indicates that colocalization of p-cSrc and Cx43 at the ID increased starting at 30 minutes and levels were maintained at 1 and 3 hours post-CO (Figure 2J). Staining for inactive cSrc showed no interaction of the nonactive kinase with Cx43 (Online Figure II).

Figure 2
Immunolocalization of Cx43 and p-cSrc in normal and EBZ myocytes 1 hour post-CO. p-cSrc (green) was not found in normal epicardium (A) but upregulated in hearts 1 hour post-CO (E). Cx43, found at the ID in normal tissue (B), exhibited some lateralization ...

Interactions of ZO-1 With c-Src and Cx43 in the EBZ of Myocardial Infarcts

To determine whether Cx43 protein partners shift in post-MI EBZ, we examined the total levels of nonphosphorylated (np)-cSrc, activated phosphorylated cSrc (p-cSrc), Cx43, and ZO-1 by Western blot and immunoprecipitation. We found that following MI, total np-cSrc levels did not change (Figure 3A); however, activated p-cSrc levels increased significantly (Figure 3B). This, in turn, resulted in an increase of p-cSrc/np-cSrc ratios within the EBZ (Figure 3C). During this same time period, Cx43 levels decreased (Figure 3D), an effect that lasted out to 48 hours; ZO-1 levels were not significantly altered (Figure 2E). Immunoprecipitation studies demonstrated an increased interaction between p-cSrc and Cx43 (Figure 3F) as early as 30 minutes post-CO, which remained elevated at 1 hour post-CO, followed by a decrease to almost normal levels by 3 hours post-CO. In contrast, Cx43/ZO-1 interactions were significantly decreased by 30 minutes post-CO and remained low at 1 hour post-CO (Figure 3G) but recovered by 3 hours post-CO. A concurrent decrease in the Cx43 interaction with p-cSrc was also observed. The most striking change was an increase in the interaction of p-cSrc with ZO-1 at 30 minutes post-CO (Figure 3H).

Figure 3
Determination of protein levels and protein interactions in EBZ following CO. A, Examination of np-cSrc levels indicated no marked change in total levels, whereas p-cSrc increased over time following CO (B). The ratio of p-cSrc to np-cSrc increases over ...

Structure of the Cx43CT/c-Src SH3 Domain Complex

Multiple Cx43CT residues including K264-K287 (the SH3 binding domain15) and A367-S372 (residues affected by the ZO-1 PDZ-2 domain) were affected on binding the c-Src SH3 domain.10 To determine whether these additional Cx43CT modifications were the result of a direct interaction with the SH3 domain or long-range conformational (indirect) changes as a result of interactions at residues K264-K287, we solved the structure by NMR (Figure 4A) (coordinates deposited in the Protein Data Bank; entry 2OM4). SH3 binding caused Cx43CT residues T275-P284 to adopt a left-handed type II helix but had no effect on the 2 preexisting α-helical domains in Cx43CT or the C-terminal residues (G350-I382) (Figure 4B). We identified 24 intermolecular nuclear Overhauser effects corresponding to the interaction of SH3 with the second Cx43CT PXXP motif (A276-S282) and found 2 binding pockets: 1 formed by intercalation of the side chains between Cx43CT residues S279, P280, and S282 and SH3 residues Y92, W118, and Y136; the other formed by intercalation of side chains between Cx43CT residues A276 and P277 and SH3 residues Y90 and Y136 (Figure 4C and 4D). The structure of the SH3 domain in complex with Cx43CT (Figure 4D) consists of 2 three-stranded antiparallel β-sheets, with 2 loops between residues E93-D99 and D112-D117 (Online Table I) similar to previously reported SH3/ligand complex structure (RMSD 1.16±0.12 Å).15

Figure 4
Solution structure of the Cx43CT/c-Src SH3 domain complex. A, Ten lowest-energy structures aligned according to the SH3 domain and Cx43CT residues 275 to 284 (red). B, Representative structure of the complex, with amino acids comprising the α ...

Promiscuity of SH3 Binding

Of the 5 Cx43CT regions affected by SH3 interactions,10 only residues K264-K287 directly interacted with SH3. Although the molar ratios of the Cx43CT and SH3 domains were similar for the structural determination (1:1), our previous study was performed at a different Cx43CT:SH3 molar ratio (1:3).10 Therefore, we used NMR translational diffusion experiments to assess whether multimeric complexes, affecting other Cx43CT regions, were formed when the Cx43CT was in the presence of excess SH3 ligand (Table) and found that at 0.8 mmol/L and pH 5.8, Cx43CT exists mostly as a dimer (consistent with16). The molecular mass of the SH3 domain alone was slightly higher than the monomer. In combination, the molecular mass of the complex was calculated to be 23 kDa, consistent with a 1:1 Cx43CT/SH3 stoichiometry.

NMR Translational Diffusion Experiments

Experiments performed at a 1:3 ratio estimated the complex weight at 35 kDa, indicating a stoichiometry of 1 Cx43CT to 2.5 SH3 domains and suggesting that Cx43CT interacts with more than one SH3 molecule. Using Y2H techniques with a series of Cx43CT truncations (Figure 5A), we confirmed binding of SH3 to Cx43CT wild-type and Cx43CT264–287, both containing the second PXXP motif and also found that the SH3 domain interacted with truncations Cx43CT296–350 and Cx43CT356–382 but not with Cx43CT296–382. Yet a reverse configuration showed this interaction (Figure 5B) consistent with previous observations demonstrating that in the Y2H system steric hindrance or structural modifications can cause a false-negative result in some configurations.17 These 2 novel Cx43 SH3 binding domains were confirmed by NMR 15N-HSQC (Online Figure III).

Figure 5
Characterization of the c-Src SH3 domain interaction with the Cx43CT domain. A and B, The S cerevisiae yeast strain AH109 was cotransformed with GAL4ad and GAL4bd fusion constructs as labeled in the figure.

SH3 Disruption of the Cx43CT/PDZ-2 Interaction

We have assessed the contributions of the various SH3 binding domains on displacing the Cx43CT/PDZ-2 interaction. 15N-labeled Cx43CT296–382 and Cx43CT356–382 (Figure 6A and 6B; in black) were combined with unlabeled PDZ-2 (1:1 molar ratio) at pH 5.8 (Figure 6A and 6B; in red); unlabeled SH3 was titrated into the complex to a final 1:1:3 molar ratio (Figure 6A and 6B; in green), and 15N-HSQC were acquired. The addition of PDZ-2 into Cx43CT296–382 and Cx43CT356–382 decreased signal intensity of the last 22 residues, indicating direct interaction between Cx43CT truncations and the PDZ-2 domain. The addition of SH3 into Cx43CT296–382/PDZ-2 and Cx43CT356–382/PDZ-2 led to reappearance of those Cx43CT resonance peaks (Figure 6A and 6B for close-up view and quantification of the signal intensity for residue I382). These results are identical to those observed when SH3 was added into Cx43CTwild-type/PDZ-2,10 suggesting that binding of SH3 to residues A276-S282 and S330-F337 is not necessary for disruption of the Cx43CT/PDZ-2 complex. Displacement may occur by (1) SH3 has a higher binding affinity and outcompetes PDZ-2 binding to Cx43CT or (2) SH3 binding PDZ-2 lowers the Cx43CT/PDZ-2 affinity.

Figure 6
15N-HSQCs demonstrating the dissociation of the complex between Cx43CT truncations and the ZO-1 PDZ-2 domain by the c-Src SH3 domain. The Cx43CT296–382 (A) and Cx43CT364–382 (B) residues that disappeared because of the interaction with ...

To test these possibilities, we used NMR titration experiments of 15N-Cx43CT with different PDZ-2 concentrations and calculated the binding affinity (Figure 6C). The decreasing signal intensity for a subset of Cx43CT residues affected as a result of the increasing PDZ-2 concentrations (Figure 6C, top; residues I382, R376, and R374) were fit according to the nonlinear least squares method and the titration data afforded a KD of 63±3 μmol/L (mean±SD). The KD for the Cx43CTwild-type, PDZ-2, and Cx43CT296–382 domain interaction with the SH3 domain (331±20, 228±23, and 658±16 μmol/L, respectively) were also obtained by similar titration experiments (Figure 6C). These results demonstrate that SH3 can bind multiple Cx43CT domains with the major site of interaction being the second PXXP motif (P277-P280). We also determined that the binding affinity of SH3 for the PDZ-2 domain was higher than the affinity of SH3 domain for the Cx43CTwild-type. These equilibrium measurements, combined with the overlap of affected residues (Figure 1D and 1E), suggest that elevated levels of p-cSrc would outcompete Cx43 for binding to ZO-1.

Localization of Cx43 and ZO-1 in the EBZ of Myocardial Infarcts

Immunolocalization of Cx43 was performed in normal and EBZ myocytes following 30 minutes, 1 hour, and 3 hours post-CO (Online Figure IV). The most dramatic change was seen 1 hour post-CO. In transverse sections of normal epicardium, ZO-1 extensively covered the intercalated disk region (Figure 7A, a), whereas Cx43 was found primarily around the ends of the myocytes (Figure 7A, b), with overlap of Cx43 with ZO-1 at the edges of the disk (Figure 7A, c). Following 1 hour of CO, transverse sections show that whereas ZO-1 remained at the disk (Figure 7A, d), Cx43 was found in a discontinuous pattern (Figure 7A, e), with little overlap of Cx43 and ZO-1 in the ID region (Figure 7A, f). In longitudinal sections in normal hearts, ID contained both Cx43 and ZO-1 (Figure 7B, a). Multiple disks within the inset show a tripartite staining with the green ZO-1 flanked by 2 separated regions of Cx43 (Figure 7B, b), with overlap of Cx43 and ZO-1 forming a yellow region (Figure 7B, c). In contrast, following 1 hour of CO, Cx43 localized to lateral membranes (Figure 7B, d), with disk regions being comprised of ZO-1 alone (Figure 7B, e) and no overlap after 1 hour of CO (Figure 7B, f). Quantitative analysis of the overlay images at all time points (Figure 7C) indicates that at the ID there was less Cx43/ZO-1 overlap (Figure 7C) following CO. These data clearly show that in early ischemia, Cx43 loses its interaction with ZO-1 and translocates to the lateral membranes of myocytes.

Figure 7
One hour post-CO causes dissociation of Cx43 from ZO-1. A, Transverse sections show that in normal myocardium, ZO-1 is found covering the ID (A, a), whereas Cx43 encircles the disk region (A, b). A, c, Overlay image shows that the colocalization of these ...


Alteration of Cx43 quantity and location accompanies many cardiac pathologies.1,4,5 In surviving EBZ myocytes, there is an increased Cx43 deposition on lateral nondisk sarcolemmal membranes associated with decreases in transverse propagation, leading to reentrant ventricular tachycardia.1,2 Based on the results from our in vitro and in vivo studies, we propose the following molecular mechanism of Cx43 lateralization. First, low intracellular pH that results from a coronary occlusion causes p-cSrc to translocate to the ID, where it binds directly to Cx43 causing the carboxyl terminal domain of Cx43 to wind into a left-handed type II helix. In addition, high p-cSrc levels at the disk facilitates interaction of the SH3 domain of p-cSrc with the PDZ-2 domain of ZO-1, as well as 2 additional lower-affinity Cx43CT domains, causing displacement of the Cx43CT from ZO-1. This enables Cx43 to move in the membrane lipid bilayer from the ID to the lateral membrane, causing decreased conduction velocity and formation of an arrhythmogenic substrate.

The first step in lateralization of Cx43 is the decrease in pH, which has been demonstrated to activate c-Src18 by phosphorylation and to translocate endogenous p-cSrc to the ID. We see in these studies that there is a phosphorylation of Src on Y416. Our data indicate that in cardiac myocytes, p-cSrc is sequestered to the ID, where it binds to Cx43 residues A276-S282 because of a higher affinity at low pH than physiological pH.8 This involves a direct interaction in a manner similar to SH3 class II peptide-binding motifs. We also determined that the SH3/Cx43CT interaction alone was not responsible for displacement of the Cx43CT/PDZ-2 interaction and that instead p-cSrc was binding directly to ZO-1, acting as a competitive ligand for the PDZ-2 domain. This is consistent with our results in EBZ myocytes showing when p-cSrc is at the ID, it also binds directly to ZO-1. Examination of this domain indicates the presence of a consensus SH3-binding domain (PXXPXK) localized at the beginning of the first β-sheet of the PDZ-2 domain in close proximity to the loop between the second and third β-sheets, and our NMR data indicate that this is indeed the region to which the SH3 domain of c-Src is binding with a higher-affinity interaction than that of p-cSrc to Cx43CT. This may be attributable to a lack of the conserved arginine residue in the Cx43CT, which would normally be involved in forming a salt bridge with SH3 residue D99,15 stabilizing one end of the left-handed type II helix. Its loss is not compensated by the multiple Cx43 PXXP motifs, which may explain the weaker Cx43CT/SH3 affinity. This suggests that the p-cSrc has a preference for binding first to ZO-1, outcompeting Cx43, and then to the Cx43CT for phosphorylation and channel closure,19 leading to closed Cx43 channels, which are free to move through the lipid bilayer from the area of high Cx43 concentration at the ID to areas of low Cx43 concentration at the lateral myocyte membranes. Interestingly, the interaction of Cx43 with p-cSrc is a transient interaction, with a loss of the interaction seen by the 3-hour time point. Because Cx43 lateralization exists past 3 hours, there are likely to be multiple mechanisms for Cx43 lateralization; the mechanism proposed here is likely to occur during the acute phase of coronary occlusion, whereas other mechanisms occur at later stages.

Cx43 is a phosphoprotein, with multiple phosphorylation sites for both serine/threonine, as well as for tyrosine phosphorylation.20 The regulation of Cx43 by phosphorylation is a complex subject, with some sites being positive regulators, leading to increased conductance, whereas others negatively regulate or close the channel. Studies have shown that overall Cx43 phosphorylation is decreased in ischemia,21 although subsequent studies have determined that some sites, such as Ser368,22 have increased phosphorylation yet are associated with channel closure. In the studies described here, we have no information about which phosphospecies interacts with ZO-1, although with the interaction occurring primarily in normal tissue, it is most likely the highly phosphorylated P2 form.

Role of the Cx43 Interaction With ZO-1

It does not appear that there is a single role for the Cx43/ZO-1 interaction. Myocyte culture studies indicate that transfection of the N-terminal domain of ZO-1 caused a dominant negative effect inhibiting Cx43 from binding to the full-length ZO-1, resulting in internalization of Cx437 and indicating that ZO-1 was playing a scaffolding and stabilizing role in localization of Cx43. Because cultured myocytes do not have intercalated disks, these studies only indicated that loss of Cx43 interaction with ZO-1 affected scaffolding of Cx43 at the cell membrane. Constitutively active c-Src also inhibited interactions of Cx43 with ZO-1, although the molecular mechanisms by which this occurred were not explored.23 In addition, Gourdie and colleagues have shown that in cultured cells, loss of Cx43/ZO-1 interactions cause rapid increases in gap junctional plaque size, indicating that channel accretion increases in the absence of direct binding of Cx43 to ZO-1.24 This indicates that ZO-1 also acts to maintain gap junctional plaque size. Our present study indicates that under in vivo diseased conditions, there is a loss of the Cx43/ZO-1 interaction and Cx43 moves away from the ID and down the lateral membranes. These lateral connexins do not associate with ZO-1, which may explain the large size of many of the aggregates seen in lateral myocyte membranes. Also, although many cardiac pathologies show lateralization of Cx43, it is clear that the molecular mechanisms for this lateralization are not identical. One study has shown an increase in Cx43/ZO-1 interaction at the intercalated disk25 in heart failure that is possibly attributable to differences in activation of c-Src in heart failure versus MI or other signal transduction pathways activated in heart failure that inhibit unhooking of Cx43 from ZO-1.

In summary, we have used a multidisciplinary approach that combines studies in the canine model of MI with NMR, Y2H, and surface plasmon resonance studies to elucidate the molecular mechanism by which activated tyrosine kinase c-Src is a regulator of junctional complexes in ventricular myocytes of the infarcted heart. We speculate that this early closure of gap junction channels in part regulates early stage arrhythmogenesis in the infarcted heart.

Supplementary Material



Sources of Funding: This work was supported by NIH grants R01GM072631 (to P.L.S.), R01HL083205 (to H.S.D.), and R01HLB066140 (to H.S.D.; A.L.W. co–principal investigator); American Heart Association Grants 0560050Z (to P.L.S.) and SDG 0535084 (to H.S.D.); and Nebraska Health and Human System grant LB506, 2005-27 (to P.L.S.).


Disclosures: None.

Contributor Information

Fabien Kieken, Department of Biochemistry and Molecular Biology, University of Nebraska Medical Center, Omaha.

Nancy Mutsaers, Department of Pharmacology, College of Physicians and Surgeons, Columbia University, New York.

Elena Dolmatova, Department of Cardiology, Beth Israel Deaconess Medical Center, Boston, Mass.

Kelly Virgil, Department of Biochemistry and Molecular Biology, University of Nebraska Medical Center, Omaha.

Andrew L. Wit, Department of Pharmacology, College of Physicians and Surgeons, Columbia University, New York.

Admir Kellezi, Department of Biochemistry and Molecular Biology, University of Nebraska Medical Center, Omaha.

Bethany J. Hirst-Jensen, Department of Biochemistry and Molecular Biology, University of Nebraska Medical Center, Omaha.

Heather S. Duffy, Department of Cardiology, Beth Israel Deaconess Medical Center, Boston, Mass.

Paul L. Sorgen, Department of Biochemistry and Molecular Biology, University of Nebraska Medical Center, Omaha.


1. Peters NS, Coromilas J, Severs NJ, Wit AL. Disturbed connexin43 gap junction distribution correlates with the location of reentrant circuits in the epicardial border zone of healing canine infarcts that cause ventricular tachycardia. Circulation. 1997;95:988–996. [PubMed]
2. Smith JH, Green CR, Peters NS, Rothery S, Severs NJ. Altered patterns of gap junction distribution in ischemic heart disease. An immunohisto-chemical study of human myocardium using laser scanning confocal microscopy. Am J Pathol. 1991;139:801–821. [PubMed]
3. Yao JA, Hussain W, Patel P, Peters NS, Boyden PA, Wit AL. Remodeling of gap junctional channel function in epicardial border zone of healing canine infarcts. Circ Res. 2003;92:437–443. [PubMed]
4. Akar FG, Nass RD, Hahn S, Cingolani E, Shah M, Hesketh GG, DiSilvestre D, Tunin RS, Kass DA, Tomaselli GF. Dynamic changes in conduction velocity and gap junction properties during development of pacing-induced heart failure. Am J Physiol Heart Circ Physiol. 2007;293:H1223–H1230. [PubMed]
5. Kaplan SR, Gard JJ, Carvajal-Huerta L, Ruiz-Cabezas JC, Thiene G, Saffitz JE. Structural and molecular pathology of the heart in Carvajal syndrome. Cardiovasc Pathol. 2004;13:26–32. [PubMed]
6. Giepmans BN, Moolenaar WH. The gap junction protein connexin43 interacts with the second PDZ domain of the zona occludens-1 protein. Curr Biol. 1998;8:931–934. [PubMed]
7. Toyofuku T, Yabuki M, Otsu K, Kuzuya T, Hori M, Tada M. Direct association of the gap junction protein connexin-43 with ZO-1 in cardiac myocytes. J Biol Chem. 1998;273:12725–12731. [PubMed]
8. Duffy HS, Ashton AW, O'Donnell P, Coombs W, Taffet SM, Delmar M, Spray DC. Regulation of connexin43 protein complexes by intracellular acidification. Circ Res. 2004;94:215–222. [PubMed]
9. Solan JL, Lampe PD. Connexin phosphorylation as a regulatory event linked to gap junction channel assembly. Biochim Biophys Acta. 2005;1711:154–163. [PubMed]
10. Sorgen PL, Duffy HS, Sahoo P, Coombs W, Delmar M, Spray DC. Structural changes in the carboxyl terminus of the gap junction protein connexin43 indicates signaling between binding domains for c-Src and zonula occludens-1. J Biol Chem. 2004;279:54695–54701. [PubMed]
11. Ursell PC, Gardner PI, Albala A, Fenoglio JJ, Jr, Wit AL. Structural and electrophysiological changes in the epicardial border zone of canine myocardial infarcts during infarct healing. Circ Res. 1985;56:436–451. [PubMed]
12. Duffy HS, Iacobas I, Hotchkiss K, Hirst-Jensen BJ, Bosco A, Dandachi N, Dermietzel R, Sorgen PL, Spray DC. The gap junction protein connexin32 interacts with the Src homology 3/hook domain of discs large homolog 1. J Biol Chem. 2007;282:9789–9796. [PubMed]
13. Kieken F, Jovic M, Naslavsky N, Caplan S, Sorgen PL. EH domain of EHD1. J Biomol NMR. 2007;39:323–329. [PubMed]
14. Hirst-Jensen BJ, Sahoo P, Kieken F, Delmar M, Sorgen PL. Characterization of the pH-dependent interaction between the gap junction protein connexin43 carboxyl terminus and cytoplasmic loop domains. J Biol Chem. 2007;282:5801–5813. [PubMed]
15. Feng S, Chen JK, Yu H, Simon JA, Schreiber SL. Two binding orientations for peptides to the Src SH3 domain: development of a general model for SH3-ligand interactions. Science. 1994;266:1241–1247. [PubMed]
16. Sorgen PL, Duffy HS, Spray DC, Delmar M. pH-dependent dimerization of the carboxyl terminal domain of Cx43. Biophys J. 2004;87:574–581. [PubMed]
17. MacDonald PN, editor. Methods in Molecular Biology. Vol 177. Two-Hybrid Systems: Methods and Protocols. Totowa, NJ: Humana Press; 2001. [PubMed]
18. Yamaji Y, Tsuganezawa H, Moe OW, Alpern RJ. Intracellular acidosis activates c-Src. Am J Physiol. 1997;272:C886–C893. [PubMed]
19. Zhou L, Kasperek EM, Nicholson BJ. Dissection of the molecular basis of pp60(v-src) induced gating of connexin 43 gap junction channels. J Cell Biol. 1999;144:1033–1045. [PMC free article] [PubMed]
20. Lampe PD, Lau AF. The effects of connexin phosphorylation on gap junctional communication. Int J Biochem Cell Biol. 2004;36:1171–1186. [PMC free article] [PubMed]
21. Beardslee MA, Lerner DL, Tadros PN, Laing JG, Beyer EC, Yamada KA, Kleber AG, Schuessler RB, Saffitz JE. Dephosphorylation and intracellular redistribution of ventricular connexin43 during electrical uncoupling induced by ischemia. Circ Res. 2000;87:656–662. [PubMed]
22. Lampe PD, TenBroek EM, Burt JM, Kurata WE, Johnson RG, Lau AF. Phosphorylation of connexin43 on serine368 by protein kinase C regulates gap junctional communication. J Cell Biol. 2000;149:1503–1512. [PMC free article] [PubMed]
23. Toyofuku T, Akamatsu Y, Zhang H, Kuzuya T, Tada M, Hori M. c-Src regulates the interaction between connexin-43 and ZO-1 in cardiac myocytes. J Biol Chem. 2001;276:1780–1788. [PubMed]
24. Hunter AW, Barker RJ, Zhu C, Gourdie RG. Zonula occludens-1 alters connexin43 gap junction size and organization by influencing channel accretion. Mol Biol Cell. 2005;16:5686–5698. [PMC free article] [PubMed]
25. Bruce AF, Rothery S, Dupont E, Severs NJ. Gap junction remodelling in human heart failure is associated with increased interaction of connexin43 with ZO-1. Cardiovasc Res. 2008;77:757–765. [PubMed]