C-Terminal Truncation of Connexin43 Changes Number, Size, and Localization of Cardiac Gap Junction Plaques
Haplodeficient mice expressing carboxyl-terminally truncated Cx43 (K258stop/KO), instead of the wild-type Cx43 isoform, reach adulthood and reveal no abnormalities in heart morphology. Here, we have analyzed the expression of K258stop protein and the morphology of gap junctions in adult hearts of these mice. Coimmunofluorescence analysis revealed reduced juxtaposition of K258stop with other junctional proteins at the intercalated disc. Immunoprecipitation studies documented changes in the interaction with previously described Cx43 binding proteins. Quantitative transmission electron and confocal microscopy confirmed the localization of K258stop gap junctions to the periphery of the intercalated disc and further revealed an increase in the size of K258stop gap junction plaques and a reduction in their number. Dual whole cell patch clamp analysis confirmed that K258stop gap junctions were functional, with single channel properties similar to those described in exogenous systems. We conclude that the carboxyl-terminal domain of Cx43 (Cx43CT) is involved in regulating the localization, number and size of Cx43 plaques in vivo. Conversely, protein interactions or posttranslational modifications taking place within the Cx43CT are not required for the assembly of functional gap junctions in the intercalated disc.
Intercellular communication between cardiomyocytes via gap junction channels allows fast impulse propagation and thereby the synchronous contraction of the heart.1 Pathological changes in gap junction expression and distribution might provide an arrhythmogenic substrate.2 Gap junction channels are formed between adjacent cells, each one providing a hemichannel, termed connexon. These connexons are formed by oligomerization of 6 individual connexin proteins.3 To date, 20 mouse and 21 human connexins have been identified.4 It is generally accepted that connexins are transmembrane proteins traversing the lipid bilayer four times, with amino- and carboxyl-termini oriented toward the cytoplasm. The main connexin isotype present in the working myocardium is connexin43 (Cx43). This protein is primarily localized at the intercalated disc5,6 where gap junctions juxtapose closely with adherens junctions and desmosomes, forming a functional unit that allows both electrical and mechanical coupling of the myocardium.7 Although cardiac-specific ablation of Cx43 in mice eventually leads to lethal ventricular tachyarrhythmias,8,9 the loss of cardiac gap junctions does not disrupt the organization of adherens junctions and desmosomes at the intercalated disc.10 In contrast, loss of the expression of either desmosomal or adherens junctional proteins alters Cx43 expression and can lead to severe ventricular arrhythmogenesis,11,12 thus suggesting a possible cross-talk between intercalated disc structures.
The localization and number of Cx43 gap junction plaques being expressed is thought to be regulated by the integration of posttranscriptional modifications and protein interactions taking place within the Cx43 carboxyl-terminal domain (Cx43CT). Direct binding to the scaffolding protein ZO-1 occurs through motifs of the Cx43CT.13 This process is implicated in the regulation of plaque size and remodeling of cardiac gap junctions.14 Interaction with tubulin is thought to be mandatory for trafficking and for the incorporation of Cx43-containing vesicles into existing gap junction plaques via the microtubular network.15,16 Other studies have used both colocalization and coimmunoprecipitation methods to demonstrate the indirect interaction of Cx43 with proteins of the adherens junctions such as beta-catenin, N-cadherin, and p120-catenin.17,18 Additionally, phosphorylation has been correlated to gap junction turnover, and all reported phosphorylation sites of Cx43 are located in the carboxyl-terminal domain.19
Recent studies described the generation of mice expressing a carboxyl-terminally truncated Cx43 protein (K258stop) in place of wild-type protein.20 Because of a disruption in epidermal differentiation, more than 97% of all homozygous K258stop animals died within the first 5 days after birth. Decreasing the K258stop gene dosage by crossing with animals heterozygous for a Cx43 knockout allele surprisingly rescued the lethal epidermal phenotype.20 Therefore, these mice harboring 1 K258stop and 1 Cx43 knockout allele (K258stop/KO) provided the first in vivo model to study the influence of Cx43CT on adult heart function. Here, we present data showing that truncation of the carboxyl-terminal domain of Cx43 alters the localization, number, and size of cardiac gap junctions and their association with other components of the intercalated disc.
Materials and Methods
This study was performed in accordance with the Principles for Utilization and Care of Vertebrate Animals Used in Testing, Research and Training as put forth by the U.S. interagency Research Animal Committee.
Animal Resources and Tissue Acquisition
Cx43/KO and K258stop/KO littermates were obtained by crossing Cx43/KO and Cx43/K258stop mice, both backcrossed onto a 94% C57BL/6 background.20 Adult animals (3 to 6 months of age) were brought under deep anesthesia (ketamin (116 mg/kg), acepromazine (11 mg/kg), xylazine (23 mg/kg), and hearts were rapidly excised. For protein assays, ventricles were minced, snap-frozen in liquid nitrogen, and stored at −80°C.
Morphology, Immunohistochemistry, and Protein Analysis
Methods used for histology, transmission electron microscopy, immunohistochemistry, and protein analysis are described in detail in the supplemental materials (available online at http://circres.ahajournals.org).
Dual Patch Clamp Electrophysiology
Methods used to record gap junction channels from cell pairs have been described in detail previously21 and are outlined in the online supplement.
Statistical significance was assessed by the unpaired 2-tailed t test (Origin Version 7.0; Origin Laboratory Corporation). Values are shown as mean±SEM. Statistical significance was set at P<0.05.
Normal Cardiac Morphology in Adult K258stop/KO Animals
As reported recently,20 mice harboring 1 K258stop and 1 Cx43 knockout allele (K258stop/KO) do not develop the epidermal barrier defect that is responsible for postnatal lethality in homozygous K258stop mice (supplemental Figure 1). Adult K258stop/KO mice displayed a slight reduction in body weight compared with heterozygous Cx43 knockout control animals (Cx43/KO) but no difference in either absolute or relative heart weight (supplemental Table I). Likewise, general histological analysis and, in particular, analysis of the morphology of the right ventricular outflow tract revealed no differences between the 2 genotypes (supplemental Figure II).
Reduced K258stop Protein Content in K258stop/KO but Not in Cx43/K258stop Adult Ventricles
We applied conventional immunoblot techniques to assess the relative abundance of K258stop protein in heart lysates. Samples were obtained from adult left ventricular tissue harvested from K258stop/KO animals. Expression levels were compared with those determined from Cx43/KO hearts, using an antibody recognizing an epitope within the amino-terminal domain of Cx43. Our data show a significant decrease in the total amount of K258stop protein when compared with control (Figure 1A and 1B). Interestingly, when samples from Cx43/K258stop animals were tested, the levels of the truncated protein were comparable to those of the wild-type form (Figure 1C). Yet, these animals harbor the same cx43k258stop gene dose as K258stop/KO animals, and no differences in the amount of transcript corresponding to either the cx43 or the cx43k258stop alleles have been detected.20 Proteins of adherens junctions and desmosomes, ZO-1, and Cx45 showed no significant difference in their expressed amounts between genotypes (supplemental Figure III).
K258stop Protein Was Localized Within the Area of the Intercalated Disc
Subcellular localization of cardiac Cx43 was initially assessed using paraffin sections of adult left ventricular tissue (Figure 2A through 2D). Staining with antibodies recognizing an epitope in the N-terminal domain revealed the typical punctuated staining pattern for wild-type Cx43 (Figure 2A). Signals for the truncated Cx43 protein, on the other hand, were detected as enlarged signals, some of them aligned parallel to fiber orientation (Figure 2C). Immunodetection of cadherin suggested that the distribution pattern of adherens junctions was similar for both genotypes (Figure 2B and 2D). Given the low intensity of Cx43-specific fluorescent signals obtained from paraffin sections, additional coimmunolocalization experiments were performed in cryosections of adult left venticular tissue (see also10). Panels E through H show staining for either the gap junction (E and G), or the adherens junction proteins (F and H). Panels I and J show merged images, together with nuclear staining. Fiber orientation was adequately preserved (see also supplemental Figure IIH), and colocalization of Cx43 and N-cadherin was seen less frequently in hearts expressing the truncated protein. High-magnification fluorescent microscopic images confirmed this observation. Figure 3 depicts the localization of Cx43 or K258stop and either alpha catenin, ZO-1 or plakophilin-2. The data show the characteristic interchanging pattern of wild-type Cx43 with other proteins at the intercalated disc (Figure 3, panels A1-A9. See also10). In contrast, K258stop signals appeared to be larger, and connexin-specific signals were less frequently coincident or immediately adjacent to those corresponding to other intercalated disc proteins (Figure 3 panels B1-B9). Overall, these data suggest that truncation of the CT domain affected the organization of gap junction plaques, and their localization within the intercalated disc complex (see also supplemental Figure V). Interestingly, the distribution of intercalated disc proteins in the hearts of animals bearing the Cx43/K258stop genotype was not different from that observed in wild-type animals (supplemental Figure VIA).
Quantification of Gap Junction Number and Size by Laser Scanning Microscopy
Quantitative analysis was performed using images obtained by laser scanning microscopy. Cryosections of left ventricular tissue were co-stained for Cx43 or K258stop (Figure 4B and 4E) as well as N-cadherin (Figure 4C and 4F). Merged images are shown in panels 4A and 4D, for samples obtained from Cx43/KO or K258stop/KO hearts, respectively. Individual intercalated discs in en face orientation were identified by N-cadherin staining, and consecutive optical sections were taken throughout the structure (see supplemental Figures VII and VIII for additional details). Number and diameter of individual junctions, as well as the percent of intercalated disc volume covered by gap junction (Table 1) or adherens junction signals (supplemental Table II) were determined from junctions that covered at least 3, 5 or 7 contiguous pixels.22 Projections of identical sample volumes revealed no changes in the average intercalated disc area recorded from K258stop/KO hearts, as compared with control (113.08 μm2±9.01 and 108.82 μm2±6.39, respectively; P=0.7). Additionally, the number, size, and percentage of intercalated disc area positive for N-cadherin signals were similar regardless of genotype (Figure 4G through 4I; Cx43/KO: gray boxes; K258stop/KO striped boxes). However, the number of Cx43-positive signals was significantly reduced, and the average diameter of individual gap junction signals was increased in K258stop/KO hearts when compared with control (Figure 4G and 4H, respectively; see also Table 1). The variance of the size of gap junction signals recorded from K258stop samples was more than twice that observed from control animals (Table 1 and supplemental Figure IX). The percentage of total intercalated disc volume occupied by Cx43-immunopositive signals (Figure 4I) was significantly decreased in K258stop/KO hearts—to about 60% of control (Figure 4I and Table 1). Finally, the fraction of gap junction signals that localized to the periphery of individual intercalated discs was 71.3±4.7% for K258stop/KO and 55.9±3.6% for Cx43/KO hearts (n=16 intercalated discs; P=0.01).
Gap Junction Plaque Morphology in K258stop-Expressing Hearts: Transmission Electron Microscopy
Transmission electron microscopy (TEM) allowed us to gain insight into the effect that CT truncation had on the morphology of the intercalated disc. Representative electron micrographs of cardiac tissue obtained from Cx43/KO and K258stop/KO animals are depicted in Figure 5A and 5D, respectively, revealing no gross changes in cardiomyocyte ultrastructure between genotypes. The area covered by cardiomyocytes, as well as the overall length of both sarcolemma and intercalated disc profile, displayed no variations between Cx43/KO and K258stop/KO hearts (Table 2). Individual intercalated disc profiles (white box in Figure 5A and 5D) were analyzed at higher magnification (Figure 5B, 5C, 5E, and 5F). The truncated Cx43 protein seemed able to form gap junction plaques between neighboring cardiomyocytes (Figure 5F). Although nonserial TEM images provide insufficient information to quantify the actual size of the plaques,23 a trend consistent with the LSM results was observed. K258stop gap junction profiles were found between end-to-end aligned cardiomyocytes, mainly at the edge of the intercalated disc, where the last plicate segment inserts into the external sarcolemma. Very few K258stop gap junction plaques were found at the lateral ends of plicate segments, or within the interplicate segments, whereas Cx43 gap junction plaques were distributed throughout, interchanging with mechanical junctions (supplemental Figure X). The latter was also observed for Cx43/K258stop adult hearts (supplemental Figure XI). Significant differences were again detected for the number and average size of individual gap junction profiles (see Table 2). An almost 5-fold reduction in the number and a 1.6-fold increase in gap junction profile size were detected in Cx43/KO hearts. Overall, Cx43CT truncation led to important differences in gap junction dimensions.
Immunoprecipitation experiments assessed whether truncation of the CT domain influenced the ability of Cx43 to physically interact with other intercalated disc proteins. Ventricular lysates were incubated with antibodies recognizing epitopes within ZO-1, beta-catenin (supplemental Figure XII), or p120-catenin (data not shown). Whereas full length Cx43 coimmunoprecipitated beta- and p120-catenin, no interaction with K258stop was detectable. However, a small amount of K258stop was immunoprecipitated by ZO-1. This result suggests that the interaction between Cx43 and ZO1 can be maintained, likely through common molecular partners (including Cx45), even in the absence of the PDZ binding domain.
Single Channel Analysis of Functional K258stop Gap Junctions
To assess the functionality of K258stop channels, we performed dual patch clamp analysis in pairs of neonatal cardiac myocytes obtained from K258stop homozygous neonates. Control experiments were conducted in myocytes harvested from Cx43/Cx43 homozygous pups. Cell pairs from both genotypes showed junctional conductance values in control conditions that were higher than our limit of resolution (approx 20 nS)24; similar results were obtained from adult myocytes bearing the heterozygote genotype (data not shown). All cell pairs were completely uncoupled by 1-mmol/L heptanol, implying that electrical coupling was attributable to gap junctions. Single channel activity was studied in neonatal cardiomyocytes by applying a transjunctional voltage (Vj) of 80 mV. The results are shown in Figure 6. For both panels, all-event frequency histograms are shown on the top and examples of single channel traces are shown at the bottom. Data recorded from cells with the wild-type genotype were best described by 3 Gaussian functions (27.3 pS, 78.6 pS, and 106.4 pS; N=7, n=199; panel A; N is number of cell pairs and n represents the number of events). As shown in panel A (bottom), the 78.6-pS peak resulted from transitions between open (O) and residual (r) states, whereas the 106.4-pS peak represented the transition from open (O) to closed (C) states. Transitions between the closed and residual states were occasionally observed, thus yielding a 27.3-pS peak. In contrast, no residual state was observed in events recorded from K258stop channels (see panel B, bottom). Consequently, only one peak (centered at 104.9 pS; N=8, n=182) was detected in the all-events histogram generated with the data from K258stop myocytes. Furthermore, truncation of the CT domain led to a prolongation of the dwell open time, from 107±22 ms (Cx43/Cx43 genotype) to 905±108 ms (K258stop/K258stop). These results were consistent with previous observations obtained from exogenously expressed channels.25 Overall, the data show that truncation of the CT domain affected not only the morphology of the gap junction plaques, but also the function of gap junction channels.
The carboxyl-terminal domain of the Cx43 protein (Cx43CT) and its function in integrating gap junction channel regulation has been a focus of connexin research for the last decade.18,19,26,27 Here, we have analyzed the effect of truncation of the CT domain on the expression and distribution of Cx43 in the adult heart, using a recently generated in vivo mouse model.20 Our experiments show that Cx43CT is necessary for regulating number and size of cardiac gap junction plaques; we further demonstrate that loss of the CT domain alters the spatial organization of Cx43 in the intercalated disc, as well as the function of Cx43 channels.
Morphology of Adult K258stop Hearts
We did not find anatomical alterations in the hearts of adult K258stop/KO mice. In particular, no malformations of the right ventricular outflow tract as those identified in Cx43KO animals were detected.28 The only minor deviation from the control group was a slight reduction in body weight in the K258stop/KO mice. However, this was not accompanied by changes in absolute or relative heart weight. In general, we found no significant alterations in anatomical or basic functional properties of K258stop/KO mice when compared with their wild-type controls.
K258stop Gap Junction Plaques in the Intercalated Disc: Number, Size, and Localization
The present study confirmed previous preliminary observations in homozygous K258stop hearts,20 showing that truncated Cx43 localized at sites of end-to-end apposition of cardiomyocytes. TEM data demonstrated the presence of K258stop gap junctions in the intercalated discs of adult hearts. These results conclusively demonstrate that the CT domain is not necessary for the formation of gap junction plaques. As such, our data depart from the notion that phosphorylation of certain residues in the CT region, or the preservation of the PDZ-binding domain of Cx43, are necessary steps for the formation of functional gap junctions.29,30
We have performed detailed analysis of size and distribution of cardiac gap junctions by 2 complementary methods: TEM and LSM. Whereas TEM allows for a higher level of resolution, the view is limited to one plane. On the other hand, LSM provides a better view of the overall intercalated disc as a contiguous 3-dimensional object; as such, LSM represents a better method to quantify volumetric parameters (see Tables 1 and 2⇑). Although we cannot rule out that some of the large gap junction signals detected by LSM represent aggregates of smaller gap-junctional plaques, results converged with the TEM data, leading us to conclude that truncation of the CT domain causes a reduction in number of gap junction plaques, an increase in mean plaque size, and an increase in variability of plaque sizes. Interestingly, this trend was consistent with in vitro measurements of gap junction plaques formed by GFP-concatenated Cx43, or plaques formed by Cx43 protein in the presence of an inhibitory peptide made up by the Cx43 ZO-1 binding motif.14 Our data suggest that loss of the CT domain disrupts gap junction size regulation in vivo, similar to what has been seen in vitro by interfering with the interaction of the CT domain with ZO-1.14
K258stop gap junctions were found within the intercalated disc but they did not show the typical interchanging pattern with adherens junctions and desmosomes. In fact, these gap junctions showed a tendency to accumulate toward the periphery of the intercalated disc. This result correlated with changes in the formation of intermolecular complexes detectable by coimmunoprecipitation. Interactions between adherens junction proteins and Cx43 are thought to be involved in cotrafficking or coassembly of junctional complexes to the plasma membrane.17,31 The individual contribution of each molecular interaction and the precise program under which the macromolecular complex of the gap junction plaque is formed remain to be determined. Our data show, however, that the presence of the Cx43CT domain is a necessary component in the normal architecture of the final structure.
Our data show that Cx43CT truncation led to an increase in average diameter of gap junctions (see Figure 4H). This result may be related to the fact that CT truncation leads to a prolongation in half-life of Cx43,20 perhaps by interfering with internalization and degradation. Indeed, Cx43CT has recently been shown to interact with the ubiquitin protein ligase Nedd4, and silencing of Nedd4 resulted in accumulation of gap junction plaques.32
Overall, our results are consistent with those obtained consequent to disruption of the Cx43-ZO-1 interaction.14 Yet, it is important to note that we still found K258stop and ZO-1 in the same immuno-precipitate (see supplemental Figure XI). Given that the PDZ-binding domain of Cx43 would be absent from the truncated protein, it seems reasonable to speculate that the observed connexin-ZO1 interaction was not direct, but mediated by other protein(s). Though at first glance surprising, this result is consistent with the fact that K258stop localizes within the intercalated disc structure, an event likely to involve associations with other—yet unknown—molecules. Our results thus support the hypothesis that cytoplasmic domains other than region 258 to 382 of Cx43 participate in intermolecular interactions. In previous studies, we have shown an interaction between the cytoplasmic loop and the carboxyl terminal domain of Cx43.26 Whether disruption of the interaction between these regions is relevant to the changes in gap junction plaque formation hereby reported remains to be determined. However, it is important to keep in mind that Cx45 tends to colocalize with Cx43 in the same junctions. As such, the interaction of Cx45 with ZO-1, and possibly of Cx45 with K258stop, could explain, at least in part, the presence of ZO-1 in the K258stop precipitates.
K258stop Channel Function in Cardiomyocytes: Absence of Transition to Residual State and Prolongation of Mean Open Time
Dual patch clamp analysis revealed that the macroscopic conductance (Gj) between cardiomyocytes obtained from either wild-type or K258stop-expressing hearts was higher than 20 nS (our limit of resolution).24 These results are consistent with those of Revilla et al, showing no difference in the macroscopic junctional conductance recorded from Xenopus oocyte pairs expressing truncated Cx43 channels, when compared with control.33 Moreover, our single channel data were consistent with previous observations indicating that truncation of the CT domain leads to a loss of the residual state, and a prolongation in the open time of Cx43 channels.25,21 At first glance, one could speculate that K258stop gap junctions would be more conductive. Yet, we do not have information regarding the number of functional channels present at the plaques, nor do we know whether truncation of the CT domain affects open probability. As such, an attempt to correlate protein content, dimension of gap junction plaques, and single channel properties with overall electrical coupling (under normal conditions) should be avoided. However, future experiments will address whether these changes could modify the response of the heart to pathological conditions such as ischemia or hypertrophy.
Is There a Link Between Changes in Channel Function and Reduced K258stop Protein Expressed?
The mechanism leading to the significant reduction in K258stop protein levels in K258stop/KO hearts remains unclear. It is interesting to note that there was no reduction in K258stop protein in hearts from animals harboring the heterozygous Cx43/K258stop genotype. This suggests that the cellular processes responsible for dampening total K258stop levels were repressed when full-length Cx43 was present. In that regard, we also observed that full-length Cx43 showed a higher mobility in SDS-PAGE when co-expressed with the truncated form. We speculate that phosphorylation of at least some selective sites in Cx43 might occur in a cooperative manner, requiring the presence of all carboxyl-terminal domains within a connexon. Additional studies will be necessary to address this hypothesis.
Previous studies suggest that when total Cx43 protein content is decreased, preservation of gap junction plaque size is favored over preservation of plaque number. Saffitz et al22 showed a reduction in number but no change in the size of ventricular gap junctions in Cx43/KO mice when compared with control. In our case, we also observed that a reduction in the amount of gap junction protein did not lead to a decrease in size of the gap junctions; in fact, an increase in gap junction size was observed (see Tables 1 and 2⇑). Though the mechanisms responsible for this phenomenon remain to be determined, we speculate that reduced expression of K258stop protein might be a compensatory adaptation, triggered by the loss of regulatory properties in the functional gap junction channel. In this regard, it is worth noting that a reduction in gene dose drastically improved the survival rate of pups expressing only the truncated channel (eg, K258stop/KO versus K258stop/K258stop).20 As such, in heterozygous animals, reduced expression may be an attempt to compensate the inability of the channels to close according to the specific demands of the cellular system at a given point in time.
Studies on the function of the CT domain of Cx43 in exogenous expression systems abound. This is the first characterization of the role of Cx43CT in the adult heart. Our data demonstrate that integrity of the CT domain is not necessary for the formation of gap junction plaques, thus arguing against the notion that Cx43 phosphorylation, or the preservation of the PDZ-binding domain, are mandatory steps in the formation of functional cardiac gap junctions.29,30 Our results also show that the CT domain is not necessary for targeting the Cx43 protein to the intercalated disc. Yet, at the intercalated disc, loss of the CT domain was associated with the formation of fewer gap junction plaques, of larger size, and mainly located to the periphery of the structure. As such, our results provide the first direct evidence that the CT domain, though not a component of the gap junction channel pore, provides plasticity to the gap junction structure and is essential for the integration of the Cx43 molecule as part of a whole macromolecular complex. Whether redistribution of gap junction plaques affects the ability of action potentials to propagate through the myocardium remains to be determined. These and other experiments to assess the effect of CT truncation on the regulation of gap junction channels under pathologic conditions such as ischemia, or hypertrophy, will be the matter of future studies.
We thank Joyce Qi for technical assistance in transmission electron microscopy analysis.
Sources of Funding
This work was supported by grants HL080602, HL39707, and GM GM057691 from the National Institutes of Health (to M.D.). Work in the Bonn laboratory was supported by grants (Wi 270/25-1,2 and 29-1) of the German Research Association (to K.W.). K.M. was recipient of the Kenneth Rosen Fellowship awarded by the Heart Rhythm Society.
Original received January 16, 2007; resubmission received August 28, 2007; revised resubmission received September 26, 2007; accepted September 28, 2007.
Nicholson BJ. Gap junctions - from cell to molecule. J Cell Sci. 2003; 116: 4479–4481.
van Veen AA, van Rijen HV, Opthof T. Cardiac gap junction channels: modulation of expression and channel properties. Cardiovasc Res. 2001; 51: 217–229.
Severs NJ, Coppen SR, Dupont E, Yeh HI, Ko YS, Matsushita T. Gap junction alterations in human cardiac disease. Cardiovasc Res. 2004; 62: 368–377.
Gutstein DE, Morley GE, Tamaddon H, Vaidya D, Schneider MD, Chen J, Chien KR, Stuhlmann H, Fishman GI. Conduction slowing and sudden arrhythmic death in mice with cardiac-restricted inactivation of connexin43. Circ Res. 2001; 88: 333–339.
van Rijen HV, Eckardt D, Degen J, Theis M, Ott T, Willecke K, Jongsma HJ, Opthof T, de Bakker JM. Slow conduction and enhanced anisotropy increase the propensity for ventricular tachyarrhythmias in adult mice with induced deletion of connexin43. Circulation. 2004; 109: 1048–1055.
Gutstein DE, Liu FY, Meyers MB, Choo A, Fishman GI. The organization of adherens junctions and desmosomes at the cardiac intercalated disc is independent of gap junctions. J Cell Sci. 2003; 116: 875–885.
Kostetskii I, Li J, Xiong Y, Zhou R, Ferrari VA, Patel VV, Molkentin JD, Radice GL. Induced deletion of the N-cadherin gene in the heart leads to dissolution of the intercalated disc structure Circ Res. 2005; 96: 346–354.
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.
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.
Giepmans BN. Role of connexin43-interacting proteins at gap junctions. In: Dhein S, ed. Adv Cardiol. 2006; 42: 41–56.
Maass K, Ghanem A, Kim JS, Saathoff M, Urschel S, Kirfel G, Grummer R, Kretz M, Lewalter T, Tiemann K, Winterhager E, Herzog V, Willecke K. Defective epidermal barrier in neonatal mice lacking the C-terminal region of connexin43. Mol Biol Cell. 2004; 15: 4597–4608.
Saffitz JE, Green KG, Kraft WJ, Schechtman KB, Yamada KA. Effects of diminished expression of connexin43 on gap junction number and size in ventricular myocardium. Am J Physiol Heart Circ Physiol. 2000; 278: H1662–H1670.
Hoyt RH, Cohen ML, Saffitz JE. Distribution and three-dimensional structure of intercellular junctions in canine myocardium. Circ Res. 1989; 64: 563–574.
Moreno AP, Chanson M, Elenes S, Anumonwo J, Scerri I, Gu H, Taffet SM, Delmar M. Role of the carboxyl terminal of connexin43 in transjunctional fast voltage gating. Circ Res. 2002; 90: 450–457.
Delmar M, Coombs W, Sorgen P, Duffy HS, Taffet SM. Structural bases for the chemical regulation of Connexin43 channels. Cardiovasc Res. 2004; 62: 268–275.
Reaume AG, de Sousa PA, Kulkarni S, Langille BL, Zhu D, Davies TC, Juneja SC, Kidder GM, Rossant J. Cardiac malformation in neonatal mice lacking connexin43. Science. 1995; 267: 1831–1834.
Musil LS, Goodenough DA. Biochemical analysis of connexin43 intracellular transport, phosphorylation, and assembly into gap junctional plaques. J Cell Biol. 1991; 115: 1357–1374.
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.
Wei CJ, Francis R, Xu X, Lo CW. Connexin43 associated with an N-cadherin-containing multiprotein complex is required for gap junction formation in NIH3T3 cells. J Biol Chem. 2005; 280: 19925–19936.
Leykauf K, Salek M, Bomke J, Frech M, Lehmann WD, Durst M, Alonso A. Ubiquitin protein ligase Nedd4 binds to connexin43 by a phosphorylation-modulated process. J Cell Sci. 2006; 119: 3634–3642.