Abstract

Intracellular protons and calcium ions are two major chemical factors that regulate connexin43 (Cx43) gap junction communication and the synergism or antagonism between pH and Ca2+ has been questioned for decades. To assess the ability of Ca2+ ions to modulate Cx43 junctional conductance (gj) in the absence of pH-sensitivity, patch clamp experiments were performed on Neuroblastoma-2a (N2a) cells or neonatal mouse ventricular myocytes (NMVMs) expressing either full-length Cx43 or the Cx43-M257 (Cx43K258stop) mutant protein, a carboxyl-terminus (CT) truncated version of Cx43 lacking pH-sensitivity. The addition of 1 μM ionomycin to normal calcium saline reduced Cx43 or Cx43-M257 gj to zero within 15 min of perfusion. This response was prevented by Ca2+-free saline or addition of 100 nM calmodulin (CaM) inhibitory peptide to the internal pipette solution. Internal addition of a connexin50 cytoplasmic loop calmodulin-binding domain (CaMBD) mimetic peptide (200 nM) prevented the Ca2+/ionomycin-induced decrease in Cx43 gj, while 100 μM Gap19 peptide had minimal effect. The investigation of the transjunctional voltage (Vj) gating properties of NMVM Cx43-M257 gap junctions confirmed the loss of the fast inactivation of Cx43-M257 gj, but also noted the abolishment of the previously reported facilitated recovery of gj from inactivating potentials. We conclude that the distal CT domain of Cx43 contributes to the Vj-dependent fast inactivation and facilitated recovery of Cx43 gap junctions, but the Ca2+/CaM-dependent gating mechanism remains intact in its absence. Sequence-specific connexin CaMBD mimetic peptides act by binding Ca2+/CaM non-specifically and the Cx43 mimetic Gap19 peptide has negligible effect on this chemical gating mechanism.

Introduction

Formed by a 20 member protein family, gap junctions are composed of a hexamer of connexin (Cx) subunits in each cell membrane that dock extracellularly to form an intercellular aqueous pore that facilitates the passage of ions, second messengers, fluorescent dyes, small nucleic acids, etc., virtually any solute under 1 kDa in molecular mass and 14 Å in width [1,2]. Gap junctions are thought to be regulated by two general gating mechanisms, a fast transjunctional voltage (Vj) gating mechanism and a slow chemical gating mechanism [3]. The calcium-dependent uncoupling of junctional membranes was first demonstrated in Chironomus salivary gland cells and the ‘healing-over’ of lesioned sheep heart tissue in the 1960s [4,5]. Thus, the initial hypothesis that the permeability of intercellular junctions was dependent on the local cytosolic calcium activity was firmly established, at least until Turin and Warner demonstrated that respiratory acidosis induced by 100% CO2 could similarly uncouple Xenopus embryo blastomeres [68]. Since calcium-dependent junctional uncoupling typically depended on massive Ca2+ influx across injured cell membranes, intracellular pH was then thought to physiologically modulate junctional coupling via respiratory gas exchange. Hence, a controversial dilemma has existed ever since about the ability of intracellular protons (H+) and/or Ca2+ ions to physiologically modulate the conductance of gap junction membranes with different studies in different tissues touting the independent virtues of cytosolic Ca2+ or H+ as the physiologically relevant biochemical factor controlling the junctional membrane conductance [9,10]. Complicating this matter further, intercellular H+ and Ca2+ concentrations are interdependent in numerous tissues including cardiac muscle, supporting the notion that they act synergistically to regulate gap junction coupling [1113].

A molecular basis for the pH-dependent gating of connexin43 (Cx43), the connexin with the most widespread expression pattern including the myocardium, was provided when truncation of last 125 amino acids of the cytoplasmic carboxyl terminus (CT) abolished the pH sensitivity of Cx43 gap junction conductance (gj) [14]. This pH gating mechanism for Cx43 gap junctions was further defined when it was demonstrated that the distal CT domain (pH gating particle) binds to region 119–144 of the Cx43 cytoplasmic loop (CL) in a pH-dependent manner [15]. The calcium gating hypothesis has evolved to include the action of calmodulin (CaM) based on the inhibition of gap junction uncoupling by CaM inhibitors and evidence that CaM binds to Cx32 and lens gap junctions [1618]. Since the identification of CaM-binding domains (CaMBDs) on the cytoplasmic amino- and carboxyl termini of Cx32, additional connexin CaMBDs were identified on the CL domain of Cx43, Cx50, Cx46 (sheep Cx44), and Cx45 [1922]). Previous studies in this laboratory indicated that Ca2+/CaM causes a gated closure of Cx43 and Cx50 gap junctions as evidenced by the reduced open probability of gap junction channels during perfusion of coupled cell pairs with the calcium ionophore ionomycin [21,23].

To test the ability of Ca2+ to modulate the conductance of Cx43 gap junctions independent of pH, we employed the pH-insensitive CT-truncated version of Cx43 (Cx43-M257) and the Cx43+/K258stop mouse, which heterologously expresses the Cx43-M257 (K258stop) protein [24], and examined the ability of Ca2+ and CaM to uncouple Cx43 gap junctions in an endogenous mouse neuro2a (N2a) cell expression system and homozygous Cx43-M257 neonatal mouse ventricular myocytes (NMVMs). We found that Cx43-M257 and NMVM Cx43K258stop/K258stop gap junction conductances (gj) were inhibited in a calcium- and calmodulin-dependent manner using previously published procedures performed on full-length wild-type (WT) Cx43 gap junctions [23]. We further tested the ability of connexin CaMBD mimetic peptides to inhibit the Cx–Ca2+/CaM gating mechanism in a connexin- and sequence-specific manner and found that the higher affinity Cx50-3 CaMBD mimetic peptide could block the Ca2+/CaM-induced uncoupling of Cx43 gap junctions Conversely, the Gap19 peptide, which targets a CL sequence immediately adjacent to the known Cx43-3 CaMBD site [25], did not inhibit the Ca2+/CaM-induced uncoupling process. Lastly, we examined the Vj-gating properties of NMVM Cx43K258stop/K258stop gap junctions and found that the fast kinetic component of the Vj-dependent gating mechanism was abolished as previously reported in Xenopus oocytes and N2a cells [26,27]. Additionally, the increased slope gj during the recovery from Vj-dependent inactivation previously observed only in primary NMVM gap junctions was also eliminated by the CT truncation of Cx43. We conclude that the Ca2+/CaM-dependent gating mechanism of Cx43 gap junctions remains intact after deletion of the distal CT pH-sensitive domain, that connexin-specific CaMBD peptides function non-specifically to bind CaM, that the Gap19 peptide does not influence this gating mechanism, and the Cx43 distal CT domain is somehow involved in the ‘facilitated’ recovery of gj after Vj-dependent inactivation.

Materials and methods

Cell cultures

Murine Neuro2a neuroblastoma (N2a) cells, grown to 70–90% confluency, were cultured in 12-well culture dishes containing MEM media supplemented with 10% fetal bovine serum and transiently transfected with 1 μg of plasmid cDNA containing full-length (WT) Cx43 or Cx43-M257 cDNA sequences using Lipofectamine2000 and OptiMEM according to the manufacturer's directions (ThermoFisher Scientific). Full-length rat Cx43 pTracer-CMV2 and Cx43-M257 pIRES2-EGFP plasmids were purified using the EndoFree plasmid minikit (Qiagen). N2a cells were lightly trypsinized after 4 h and plated in 35 mm culture dishes overnight for patch clamp electrophysiology studies the next day.

All cardiomyocyte experiments were performed on enzymatically dissociated neonatal C57BL/6 murine ventricular myocytes cultured for 48–72 h according to the published procedures [28]. The newborn mice were killed under isoflurane anesthesia in accordance with procedures approved by the Institutional Animal Care and Use Committee (IACUC) conforming to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). Newborn litters from heterozygous Cx43-M257 (Cx43K258stop) mouse matings were collected and the ventricle and tail of each newborn pup were numbered for separate enzymatic dissociation and genotyping. Genomic DNA was obtained using the DNeasy Blood and Tissue kit (Qiagen) and subjected to polymerase chain reaction (PCR) analysis using Taq polymerase (ThermoFisher Scientific) and M257 forward (5′-CAA AAC ACC CCC CAA GGA ACC TAG) and reverse (5′ GCA TCC TCT TCA AGT CTG TCT TCG) primers as originally described [24].

Electrophysiology

Gap junctional current (Ij) recordings were acquired using previously established dual whole-cell patch clamp procedures with a pair of BioLogic RK-400 patch clamp amplifiers (Bio-Logic Science Instruments, www.bio-logic.net) [29]. Whole-cell input resistances (Rinput) and capacitances (Cinput) were measured using a −140 to +60 mV ramp (20 ms/mV) and a +5 mV, 10 ms voltage step applied to both cells of an N2a or NMVM cell pair from a holding potential of −40 mV. Whole-cell patch electrode (access) resistances (Rel1 and Rel2), beginning with a 4–5 MΩ resistance in the bath, were calculated using the equation Rel = τcap/Cinput, where τcap is the single-cell capacitive current decay time constant [29,30]. The macroscopic junctional conductance (gj) was calculated using the equation: gj = −ΔI2/[(V1 + ΔV) − V2], where I1 and I2 are whole-cell currents, V1 and V2 are command (holding) voltages, (ΔV) is a voltage step applied to one cell while maintaining the holding potential (Vh) of the partner cell, ΔI2 is the change in whole-cell current of the partner cell, and (V1 + ΔV) − V2 is the transjunctional potential, Vj. For all gj experiments, Vh = 0 mV for both cells, ΔV was a 5 s, −20 mV pulse applied four times/min, and gj was normalized by dividing the time-dependent gj by the initial gj measurement, or Gj = gj(time)/gj(initial). For all ionomycin perfusion experiments, gj(initial) was calculated as the gj averaged over the first 2  min preceding perfusion (8 data points). To more accurately calculate gj, the whole-cell patch electrode resistances, Rel1 and Rel2, were subtracted from the gj = −ΔI2/Vj calculation where the actual gj = apparent gj – (1/Rel1 + 1/Rel2), i.e. the observed rj = actual rj + Rel1 + Rel2). The bath saline contained (in mM): 142 NaCl, 1.3 KCl, 4 CsCl, 2 tetraethylammonium chloride, 0.8 MgSO4, 0.9 NaH2PO4, 1.8 CaCl2, 5.5 dextrose, and 10 HEPES (titrated to pH 7.4 with 1 N NaOH). An amount of 1.8 mM CaCl2 was omitted from the Ca2+-free saline. Ionomycin (1 μM, I3909, Sigma–Aldrich) was added to the perfusion saline fresh daily. The internal pipette solution (IPS) contained (in mM): 140 KCl, 1.0 MgCl2, 0.3 KBAPTA, 25 HEPES (titrated to pH 7.40 with 1 N KOH).

The peptide inhibition studies of Ca2+/CaM-dependent uncoupling of Cx43 gap junctions were performed by adding the inhibitory peptides to the IPS of both cells on the day of use at concentrations dependent upon their measured Kd. CaM inhibition was achieved by adding 100 nM CaM inhibitory peptide corresponding to the CaMBD of CaM kinase II (Enzo Life Sciences, #BML-P200-0500) with a reported Kd of 52 nM. The Cx50 CaMBD peptide, corresponding to the Cx50 CL sequence 141-SSKGTKKFRLEGTLLRTYVCHIIFKT-166 and its scrambled control (FKLYKCISFGGTEITTRSHVLTKKRL) were published previously [21]. The CaMBD peptide was added to the IPS at a concentration of 200 nM. The Gap19 peptide (128-KQIEIKKFK-136, Sigma–Aldrich, #SML1426), derived from the Cx43 TAT-L2 peptide (TAT-119-DGANVDMHLKQIEIKKFKYGIEEHGK-144, was added to the IPS at a concentration of 100 μM [25,31].

Immunocytochemistry

For immunocytochemical studies, NMVMs were cultured on poly-l-lysine-coated coverslips for 48–72 h and immunolabeled for Cx43 protein using previously published procedures [28]. Coverslips were rinsed in phosphate-buffered saline (PBS, pH 7.0), fixed with 4% paraformaldehyde in PBS, rinsed, permeabilized with 1% Triton X100 in PBS, and blocked with 2% goat serum in 1% Triton X100 PBS, all for 15 min at room temperature. Anti-Cx43 rabbit polyclonal amino-terminal (Abgent, #AP1541b) and mouse monoclonal carboxyl-terminal (ThermoFisher #35-5000) antibodies were diluted 1:200 in 2% goat serum, 1% triton X-100 PBS and added to the 12 mm diameter coverslip overnight at 4°C. Each well was rinsed 3–5 times with PBS the next day and incubated with goat anti-rabbit Alexa Fluor 488 (ThermoFisher #11008) and anti-mouse Alexa Fluor 546 (ThermoFisher #11003) secondary antibodies, diluted 1:1500 in 10% goat serum PBS, for 30 min at room temperature. After rinsing, the cells were labeled with DAPI for 10 min, rinsed with PBS and finally with pure water, blotted dry, and mounted on a glass slide using ProLong antifade mounting reagent and cured overnight in the dark.

Sealed glass coverslips were viewed with an Olympus IX-70 microscope using a Sutter Instruments LS 175 W Xenon arc lamp epifluorescence illumination system and Lambda 10-2 filter wheel controller with 484/15 or 555/25 nm band-pass excitation filters and FITC or FITC/Cy3/Cy5 dichroic mirror/emitter filter sets (Chroma Technology Corp., #41001 or #62005). DNA staining was observed with a 365/10 nm excitation filter and DAPI 460/50 nm dichroic mirror/emitter filter set (Chroma, 31000v2-dapi-hoechst-amca). Fluorescent micrographs were acquired with an Andor iXon 885 ECCD camera using Imaging Workbench 6.0 software (INDEC Systems, Santa Clara, CA, U.S.A.). Exported TIF files were background subtracted (≈10%) and green/red/blue color processed using ImageJ software. Magnification was 600× using an Olympus PlanApo 1.40/0.17 aperture 60× oil immersion and 10× C-mount objectives.

Results

To examine the Ca2+/CaM gating of Cx43-M257 gap junctions, we employed the same ionomycin perfusion assay from our previous study of WT Cx43 [23]. When coupled WT or Cx43-M257 N2a cell pairs were superfused with normal saline containing 1 μM ionomycin and 1.8 mM CaCl2, Ij recordings in response to a −20 mV Vj pulse declined steadily from initial levels to zero during 8–12 min of perfusion at a rate of 1 ml/min (Figure 1A,D). The average initial gj was 26.4 ± 6.2 nS for the WT Cx43 and 29.1 ± 7.7 nS for the Cx43-M257 cell pairs (mean ± SEM, n = 7, 6). In all experiments, the normalized Gj = gj(time)/gj(initial) declined by ≥95% within 10 min of 1.8 mM CaCl2 + 1 μM ionomycin saline perfusion (Figure 1B,E). The average normalized Gj declined to 0 nS within 8 min from the onset of perfusion for both WT Cx43 and Cx43-M257 gap junctions (Figure 1C,F, P-value >0.05, one-way ANOVA).

Calcium/ionomycin-induced uncoupling of Cx43 gap junctions expressed in N2a cells.

Figure 1.
Calcium/ionomycin-induced uncoupling of Cx43 gap junctions expressed in N2a cells.

(A and D) Whole-cell current traces from the partner cell 2 of an N2a cell pair during a −20 mV, 5 s transjunctional (Vj) voltage pulse (inset, upper left) applied to cell 1 before (control) and during perfusion with 1 μM ionomycin, 1.8 mM CaCl2 saline. Current traces from representative experiments illustrate the decline in junctional current (Ij = −ΔI2) during ionomycin perfusion of an N2a cell pair expressing either full-length WT Cx43 (A) or carboxyl-tail truncated Cx43 (Cx43-M257, D) gap junctions, or the prevention of the calcium-induced decline in Ij by calmodulin (CaM) inhibition (C). (B) Time course of the decline in normalized Gj for all seven WT N2a-Cx43 cell pairs. (C) The average (mean ± SEM, n = 7) decline in junctional conductance (Gj) during CaCl2/ionomycin perfusion of N2a-Cx43 cell pairs illustrating 100% Gj inhibition within 15 min of perfusion. (E) Time course of the decline in normalized Gj for all six N2a-Cx43-M257 cell pairs. (F) Average Gj of N2a-Cx43-M257 cell pairs (n = 6) during perfusion with 1.8 mM CaCl2 saline + ionomycin (filled square) illustrating complete uncoupling in the absence of the Cx43 CT domain.

Figure 1.
Calcium/ionomycin-induced uncoupling of Cx43 gap junctions expressed in N2a cells.

(A and D) Whole-cell current traces from the partner cell 2 of an N2a cell pair during a −20 mV, 5 s transjunctional (Vj) voltage pulse (inset, upper left) applied to cell 1 before (control) and during perfusion with 1 μM ionomycin, 1.8 mM CaCl2 saline. Current traces from representative experiments illustrate the decline in junctional current (Ij = −ΔI2) during ionomycin perfusion of an N2a cell pair expressing either full-length WT Cx43 (A) or carboxyl-tail truncated Cx43 (Cx43-M257, D) gap junctions, or the prevention of the calcium-induced decline in Ij by calmodulin (CaM) inhibition (C). (B) Time course of the decline in normalized Gj for all seven WT N2a-Cx43 cell pairs. (C) The average (mean ± SEM, n = 7) decline in junctional conductance (Gj) during CaCl2/ionomycin perfusion of N2a-Cx43 cell pairs illustrating 100% Gj inhibition within 15 min of perfusion. (E) Time course of the decline in normalized Gj for all six N2a-Cx43-M257 cell pairs. (F) Average Gj of N2a-Cx43-M257 cell pairs (n = 6) during perfusion with 1.8 mM CaCl2 saline + ionomycin (filled square) illustrating complete uncoupling in the absence of the Cx43 CT domain.

Negative control experiments were performed on paired Cx43-M257 N2a cells by omitting 1.8 mM CaCl2 from the 1 μM ionomycin saline (Figure 2A). The omission of the 1.8 mM CaCl2 from the perfusate prevented the rundown of Cx43-M257 Gj, thus demonstrating the Ca2+-dependence of the ionomycin-induced uncoupling (Figure 2A). The average initial gj was 18.6 ± 6.9 nS for the zero Ca2+ experiments (mean ± SEM, n = 6). To test for the involvement of CaM in the ionomycin/Ca2+-induced uncoupling of Cx43-M257 gap junctions, the CaM kinase II calmodulin-binding domain inhibitory peptide (CaMKII 290–309) was added to both whole-cell patch pipettes at a concentration of 100 nM (≈2 × Kd). As observed previously with WT Cx43 [23], 100 nM CaM inhibitory (CaMi) peptide completely antagonized the decrease in Cx43-M257 Ij and Gj during 1.8 mM CaCl2 + 1 μM ionomycin saline perfusion (Figure 2B,C). The average initial gj was 30.3 ± 6.9 nS for the CaM inhibitory peptide Cx43-M257 experiments (mean ± SEM, n = 6). These findings are consistent with previous findings with WT Cx43 of a calcium/calmodulin-dependent gating mechanism even in the absence of pH-sensitivity [19,23].

Calcium/CaM-dependent uncoupling of Cx43-M257 gap junctions expressed in N2a cells.

Figure 2.
Calcium/CaM-dependent uncoupling of Cx43-M257 gap junctions expressed in N2a cells.

(A) Omission of 1.8 mM CaCl2 from the 1 μM ionomycin saline prevented the decline in the average Gj of N2a-Cx43-M257 cell pairs (open circle). The Ca2+-dependent decline in Cx43-M257 Gj from Figure 1F is shown in gray (filled circle) for comparison. (B) Whole-cell current traces from the partner cell 2 of an N2a Cx43-M257 in response to −20 mV, 5 s Vj pulses before (1 min, control) and during perfusion with 1 μM ionomycin, 1.8 mM CaCl2 saline illustrating the lack of decline in Ij in the presence of 100 nM CaM inhibitory (CaMi) peptide. (C) The average Gj of six Cx43-M257 N2a cell pairs did not decline below initial values during 12–13 min superfusion with 1.8 mM CaCl2 + 1 μM ionomycin saline when 100 nM CaMi peptide was added to both whole-cell patch pipettes. Again, the control decay of Cx43-M257 Gj in the presence of Ca2+/ionomycin is shown in gray.

Figure 2.
Calcium/CaM-dependent uncoupling of Cx43-M257 gap junctions expressed in N2a cells.

(A) Omission of 1.8 mM CaCl2 from the 1 μM ionomycin saline prevented the decline in the average Gj of N2a-Cx43-M257 cell pairs (open circle). The Ca2+-dependent decline in Cx43-M257 Gj from Figure 1F is shown in gray (filled circle) for comparison. (B) Whole-cell current traces from the partner cell 2 of an N2a Cx43-M257 in response to −20 mV, 5 s Vj pulses before (1 min, control) and during perfusion with 1 μM ionomycin, 1.8 mM CaCl2 saline illustrating the lack of decline in Ij in the presence of 100 nM CaM inhibitory (CaMi) peptide. (C) The average Gj of six Cx43-M257 N2a cell pairs did not decline below initial values during 12–13 min superfusion with 1.8 mM CaCl2 + 1 μM ionomycin saline when 100 nM CaMi peptide was added to both whole-cell patch pipettes. Again, the control decay of Cx43-M257 Gj in the presence of Ca2+/ionomycin is shown in gray.

The above Cx43-M257 experiments were performed using the exogenous N2a cell expression system and a Cx43K258stop mouse was later developed that expresses the CT-truncated (K258stop = M257) form of Cx43 in a heterozygous manner [24]. To confirm the Ca2+/CaM-dependent uncoupling of pH-insensitive Cx43-M257 gap junctions in an endogenous expression system, mouse ventricular cardiomyocytes were cultured from the hearts of homozygous newborn littermates from heterozygous Cx43+/K258stop mice matings. The genotype of each newborn pup was determined by PCR analysis of tail DNA samples and confirmed by immunocytochemical labeling of cultured cardiomyocytes using amino-terminal and carboxyl-terminal anti-Cx43 antibodies (Figure 2A,B). The Cx43-NT antibody recognizing both forms of Cx43 was secondarily labeled with Alexa Fluor488 (green) and the Cx43-CT antibody, which would recognize only the full-length WT Cx43, was labeled with Alexa Fluor546 (red). Homozygous Cx43-M257 cardiomyocytes were devoid of Cx43-CT immunolabeling (Figure 3A), whereas the homozygous WT Cx43 (and heterozygous Cx43+/K258stop) cardiomyocytes were immunolabeled by both anti-Cx43 antibodies (Figure 3B). The perfusion of homozygous Cx43-M257 and WT Cx43 paired cardiomyocytes with 1.8 mM CaCl2 + 1 μM ionomycin saline resulted in complete uncoupling within 8 min of perfusion, similar to the results obtained in N2a cells. The average initial gj was 54.2 ± 6.2 and 39.3 ± 4.7 nS for the Cx43-M257 and WT-Cx43 cardiomyocyte pairs (n = 7, 7). Connexin mimetic peptides corresponding to the CaMBD of Cx43 and Cx50 were used previously to demonstrate that the connexin-specific sequence was capable of preventing the uncoupling of their respective gap junctions during Ca2+-ionomycin perfusion [21,23]. However, since these Cx mimetic peptides correspond to a full-length CaMBD capable of binding CaM only with different affinities [19,21], we tested the ability of the higher affinity Cx50-3 peptide (Kd ≈ 5 nM) to antagonize the Ca2+/CaM-dependent uncoupling of WT Cx43 N2a cell pairs (Figure 4A). We found that 200 nM Cx50-3 peptide was sufficient to completely prevent the decline in Gj during Ca2+/ionomycin perfusion of Cx43 gap junctions. The 200 nM of the scrambled control Cx50-3 peptide significantly delayed the uncoupling by ≈4 min (P-value <0.01, one-way ANOVA), but, failed to prevent uncoupling during 10–12 min of Ca2+-ionomycin perfusion. These results suggest that any Cx mimetic peptide capable of binding CaM antagonizes the Ca2+/CaM-dependent gating mechanism in a connexin non-specific manner simply by binding CaM. The average initial gj was 14.9 ± 5.7 and 34.4 ± 8.1 nS for the Cx50-3 and scrambled control peptide experiments (n = 5, 3).

Calcium/CaM-dependent uncoupling of Cx43-M257 ventricular cardiomyocytes.

Figure 3.
Calcium/CaM-dependent uncoupling of Cx43-M257 ventricular cardiomyocytes.

(A) Immunofluorescent labeling of Cx43-M257 cardiomyocyte gap junctions with anti-Cx43 amino-terminal (NT) and Alexa Fluor488 antibodies (green) and anti-Cx43 CT and Alexa Fluor555 antibodies (red) illustrating the lack of the Cx43 CT domain in cardiomyocytes cultured from homozygous Cx43-M257 neonatal mice. (B) Immunocytochemical labeling of WT Cx43 gap junctions illustrating the presence of the NT and CT domains in the littermate control cardiomyocytes. (C) Time course of the decline in normalized Gj for all seven homozygous Cx43-M257 NMVM pairs. (D) Time course of the decline in normalized Gj for all seven littermate homozygous WT Cx43 NMVM pairs. (E) The average Gj of both WT and Cx43-M257 cardiomyocyte gap junctions was completely inhibited by calcium + ionomycin saline perfusion, confirming the results obtained by exogenous expression of Cx43 constructs in N2a cells.

Figure 3.
Calcium/CaM-dependent uncoupling of Cx43-M257 ventricular cardiomyocytes.

(A) Immunofluorescent labeling of Cx43-M257 cardiomyocyte gap junctions with anti-Cx43 amino-terminal (NT) and Alexa Fluor488 antibodies (green) and anti-Cx43 CT and Alexa Fluor555 antibodies (red) illustrating the lack of the Cx43 CT domain in cardiomyocytes cultured from homozygous Cx43-M257 neonatal mice. (B) Immunocytochemical labeling of WT Cx43 gap junctions illustrating the presence of the NT and CT domains in the littermate control cardiomyocytes. (C) Time course of the decline in normalized Gj for all seven homozygous Cx43-M257 NMVM pairs. (D) Time course of the decline in normalized Gj for all seven littermate homozygous WT Cx43 NMVM pairs. (E) The average Gj of both WT and Cx43-M257 cardiomyocyte gap junctions was completely inhibited by calcium + ionomycin saline perfusion, confirming the results obtained by exogenous expression of Cx43 constructs in N2a cells.

Effects of connexin mimetic peptides on calcium/CaM-dependent uncoupling of Cx43 gap junctions.

Figure 4.
Effects of connexin mimetic peptides on calcium/CaM-dependent uncoupling of Cx43 gap junctions.

(A) Connexin mimetic peptide of a confirmed Cx50 cytoplasmic loop (CL) calmodulin-binding domain (CaMBDs) previously shown to inhibit Ca2+/CaM-dependent uncoupling was tested on Cx43 WT gap junctions [21]. Complete inhibition of Cx43 gap junction uncoupling was achieved with 200 nM Cx50-3 mimetic peptide added to both patch pipettes. The scrambled control Cx50 CaMBD peptide delayed, but did not prevent the Ca2+/CaM-dependent decline in Cx43 Gj. (B) A distinct Cx43 CL mimetic peptide reported to inhibit Cx43 hemichannel activity, but not Cx43 gap junction coupling, was examined for possible effects on Cx43 Ca2+/CaM-induced uncoupling [25]. The 100 μM Gap19 peptide added to both whole-cell patch pipettes delayed but did not prevent the decline in Cx43 Gj induced by 1.8 mM CaCl2/ionomycin saline perfusion.

Figure 4.
Effects of connexin mimetic peptides on calcium/CaM-dependent uncoupling of Cx43 gap junctions.

(A) Connexin mimetic peptide of a confirmed Cx50 cytoplasmic loop (CL) calmodulin-binding domain (CaMBDs) previously shown to inhibit Ca2+/CaM-dependent uncoupling was tested on Cx43 WT gap junctions [21]. Complete inhibition of Cx43 gap junction uncoupling was achieved with 200 nM Cx50-3 mimetic peptide added to both patch pipettes. The scrambled control Cx50 CaMBD peptide delayed, but did not prevent the Ca2+/CaM-dependent decline in Cx43 Gj. (B) A distinct Cx43 CL mimetic peptide reported to inhibit Cx43 hemichannel activity, but not Cx43 gap junction coupling, was examined for possible effects on Cx43 Ca2+/CaM-induced uncoupling [25]. The 100 μM Gap19 peptide added to both whole-cell patch pipettes delayed but did not prevent the decline in Cx43 Gj induced by 1.8 mM CaCl2/ionomycin saline perfusion.

The Cx43-3 CaMBD (136–158) is adjacent to the CL ‘L2’ (119–144) region and the L2 peptide is known to bind to the Cx43 CT domain in a pH-dependent manner and inhibit Cx43 hemichannel function [15,31]. Recently, a Gap19 peptide targeting the central portion of the L2 region, 128-KQIEIKKFK-136, was developed and shown to inhibit Cx43 hemichannel function without inhibiting gj [25]. Given the close proximity of the Gap19 region to the known Cx43 CaMBD, we wanted to test for possible effects of the Gap19 peptide on the Ca2+/CaM-dependent uncoupling of Cx43 gap junctions. In seven Ca2+-ionomycin perfusion experiments, inclusion of 100 μM Gap19 peptide in both patch pipettes again delayed the uncoupling by ≈3 min (P-value <0.01, one-way ANOVA), but did not prevent the uncoupling of WT Cx43 N2a cells pairs (Figure 4B). The average initial gj was 31.8 ± 5.9 nS (n = 7). These results suggest that Gap19 does not prevent the Ca2+/CaM-dependent gating of Cx43 gap junction.

Truncation of the Cx43 CT was also reported to alter the Vj-gating of Cx43 gap junctions by eliminating the ‘fast’ Vj-gating mechanism to the 50 pS subconductance state of Cx43 gap junction channels [26,27]. We had previously observed a hysteresis in the steady-state GjVj curve only in NMVMs termed ‘facilitation’, wherein a slow (200 ms/mV) Vj ramp from ±120 mV back to 0 mV Vj resulted in a higher linear slope conductance at low Vj values than measured initially with increasing ±Vj values from 0 mV [28]. The previous Vj-gating studies of Cx43-M257 gap junctions were performed in Xenopus oocytes and N2a cells. Thus, we examined the steady-state and kinetic Vj-gating properties of myocardial homozygous Cx43-M257 gap junctions using cultured NMVMs. The application of the slow ±120 mV Vj ramp revealed the loss of facilitation in Cx43-M257 ventricular cardiomyocyte gap junctions during the return recovery phase of the GjVj curve (Figure 5A,B). The average initial gj of these Vj-gating experiments was 7.2 ± 1.4 nS since Vj-dependent inactivation can be detected only in low gj cell pairs. The steady-state Cx43-M257 NMVM GjVj curve parameters for the Boltzmann equation fitted curves are listed in Table 1. The steady-state GjVj parameters for WT Cx43 NMVM gap junctions from the original study are listed in italics in Table 1 [28].

Vj-gating properties of Cx43-M257 ventricular gap junctions.

Figure 5.
Vj-gating properties of Cx43-M257 ventricular gap junctions.

(A) Average GjVj relationship from six Cx43-M257 ventricular cardiomyocyte cell pairs obtained during a continuous 200 ms/mV Vj ramp from 0 to ±120 mV (inactivation). The Gj data (—) were fitted with the Boltzmann equation () listed in Table 1 and the fitted curve for the WT Cx43 NMVM data from Lin et al. [32] is illustrated by the gray line () for comparison. (B) Average GjVj curve for the same six Cx43-M257 NMVM cell pairs during the continuous 200 ms/mV ramp from ±120 to 0 mV (recovery). Absent from these traces is the increased linear slope conductance (‘facilitation’) observed during the recovery phase observed in WT Cx43 ventricular cardiomyocyte gap junctions illustrated by the light gray line () [32]. (C) The kinetics of Vj-dependent inactivation of WT and Cx43-M257 ventricular cardiomyocytes were determined from five to seven cell pairs in response to Vj pulses from −60 to −140 mV in 10 mV increments according to the methods of Lin et al. [32]. Exponential fits of the slow inactivation rates (Kon,slow) from WT mice revealed an e-fold change in rate for every 19.8 ± 0.7 mV compared with 22.8 ± 1.7 mV for Cx43-M257 myocytes. The fast inactivation component was absent from Cx43-M257 myocytes, consistent with previous findings [26,27]. (D) Gj inactivation during a 1 Hz ventricular action potential waveform from WT () and Cx43-M257 () cardiomyocytes. Owing to the slower inactivation kinetics and reduced Vj-sensitivity of the Cx43-M257 gap junctions, Vj-dependent inactivation of Gj during the action potential was reduced to 10% compared with 50% for WT Cx43 cardiomyocyte gap junctions. BCL = Basic Cycle Length.

Figure 5.
Vj-gating properties of Cx43-M257 ventricular gap junctions.

(A) Average GjVj relationship from six Cx43-M257 ventricular cardiomyocyte cell pairs obtained during a continuous 200 ms/mV Vj ramp from 0 to ±120 mV (inactivation). The Gj data (—) were fitted with the Boltzmann equation () listed in Table 1 and the fitted curve for the WT Cx43 NMVM data from Lin et al. [32] is illustrated by the gray line () for comparison. (B) Average GjVj curve for the same six Cx43-M257 NMVM cell pairs during the continuous 200 ms/mV ramp from ±120 to 0 mV (recovery). Absent from these traces is the increased linear slope conductance (‘facilitation’) observed during the recovery phase observed in WT Cx43 ventricular cardiomyocyte gap junctions illustrated by the light gray line () [32]. (C) The kinetics of Vj-dependent inactivation of WT and Cx43-M257 ventricular cardiomyocytes were determined from five to seven cell pairs in response to Vj pulses from −60 to −140 mV in 10 mV increments according to the methods of Lin et al. [32]. Exponential fits of the slow inactivation rates (Kon,slow) from WT mice revealed an e-fold change in rate for every 19.8 ± 0.7 mV compared with 22.8 ± 1.7 mV for Cx43-M257 myocytes. The fast inactivation component was absent from Cx43-M257 myocytes, consistent with previous findings [26,27]. (D) Gj inactivation during a 1 Hz ventricular action potential waveform from WT () and Cx43-M257 () cardiomyocytes. Owing to the slower inactivation kinetics and reduced Vj-sensitivity of the Cx43-M257 gap junctions, Vj-dependent inactivation of Gj during the action potential was reduced to 10% compared with 50% for WT Cx43 cardiomyocyte gap junctions. BCL = Basic Cycle Length.

Table 1
Boltzmann Equation Parameters for the Cx43-M257 Gj − Vj Curves
 Gmax Gmin V½ (mV) Valence (zR N 
Inactivation 
 −Vj 1.06 ± 0.006 0.32 ± 0.015 −74.3 ± 0.8 1.22 ± 0.05 0.95 
 1.00 ± 0.001 0.31 ± 0.001 −57.06 ± 0.04 −3.48 ± 0.02   4 
 +Vj 1.02 ± 0.002 0.22 ± 0.003 +66.7 ± 0.2 2.15 ± 0.04 0.97  
 1.00 ± 0.001 0.37 ± 0.001 +55.87 ± 0.09 +3.06 ± 0.03   
Recovery 
 −Vj 0.86 ± 0.007 0.40 ± 0.016 −71.1 ± 1.4 1.13 ± 0.08 0.94 
 1.60 ± 0.002 0.30 ± 0.001 −43.36 ± 0.06 −2.35 ± 0.01  4 
 +Vj 0.92 ± 0.005 0.44 ± 0.007 +65.0 ± 0.6 1.26 ± 0.05 0.96  
 2.46 ± 0.005 0.43 ± 0.002 +44.78 ± 0.09 +2.09 ± 0.01   
 Gmax Gmin V½ (mV) Valence (zR N 
Inactivation 
 −Vj 1.06 ± 0.006 0.32 ± 0.015 −74.3 ± 0.8 1.22 ± 0.05 0.95 
 1.00 ± 0.001 0.31 ± 0.001 −57.06 ± 0.04 −3.48 ± 0.02   4 
 +Vj 1.02 ± 0.002 0.22 ± 0.003 +66.7 ± 0.2 2.15 ± 0.04 0.97  
 1.00 ± 0.001 0.37 ± 0.001 +55.87 ± 0.09 +3.06 ± 0.03   
Recovery 
 −Vj 0.86 ± 0.007 0.40 ± 0.016 −71.1 ± 1.4 1.13 ± 0.08 0.94 
 1.60 ± 0.002 0.30 ± 0.001 −43.36 ± 0.06 −2.35 ± 0.01  4 
 +Vj 0.92 ± 0.005 0.44 ± 0.007 +65.0 ± 0.6 1.26 ± 0.05 0.96  
 2.46 ± 0.005 0.43 ± 0.002 +44.78 ± 0.09 +2.09 ± 0.01   

.

The Vj-dependent decay kinetics in NMVM cell pairs were examined using −Vj pulses and the ventricular action potential as previously described [28,32]. The decay kinetics during Vj pulses from −60 to −140 mV confirmed the loss of the fast inactivation time constant and the presence of a distinct slow inactivation time constant relative to WT Cx43 NMVM gap junctions (Figure 5C). The elimination of the Cx43 fast inactivation time constant also eliminated nearly all of the Vj-dependent inactivation during the 1 Hz simulated ventricular action potential (Figure 5D). The decay rates (Kon) were calculated using the expression Kon (ms−1) = (1 – Popen)/τinactivation = A0•exp(Vj/Vk) + C, where A0 is the rate constant amplitude in ms−1, C is the y-intercept (ms−1), and Vk is the voltage constant for the decay rate [28,32]. Exponential fits of the WT slow decay rates and the M257 decay rates yielded rates of 0.0000469•exp(Vj/19.8)  − 0.000086 for WT Cx43 NMVM gap junctions and 0.0000504•exp(Vj/22.8)  − 0.000058 for Cx43-M257 NMVM gap junctions. These data reveal that deletion of the CT terminus of Cx43 not only eliminated the fast decay of Cx43 gap junctions, but also reduced the Vj-sensitivity of the remaining slow decay component of Cx43 gap junctions by 15%. Together, these data from Cx43-M257 NMVMs confirm the elimination of the fast decay component of Cx43 gap junctions, quantify the reduction in the kinetics and Vj-sensitivity of the remaining slow decay component, and demonstrates the removal of the facilitated recovery of gj from inactivating potentials of Cx43 gap junctions.

Discussion

The primary purpose of this study was to determine if the recently described Ca2+/CaM-dependent gating mechanism exists in the absence of Cx43 pH-sensitivity abolished by truncation of the Cx43 CT domain at position 257 [14,19,23]. The perfusion of Cx43-M257 N2a or NMVM cell pairs with 1 μM ionomycin saline containing normal 1.8 mM CaCl2 routinely induced 100% inhibition of Cx43 gj within 8–12 min from the onset of 1 ml/min perfusion (Figures 1A–F and 3C–E). Consistent with previous findings using full-length Cx43 expressed in N2a cells or WT NMVMs [23], omission of the 1.8 mM CaCl2 from the 1 μM ionomycin bath saline or addition of 100 nM CaMKII 290–309 CaM inhibitory peptide to both whole-cell patch pipettes prevented the rundown of Cx43-M257 gj during 13 min of perfusion (Figure 2A–C). Taken together, these data support the conclusion that the Ca2+- and CaM-dependent chemical gating mechanism of Cx43 gap junctions does not require the presence of the Cx43 distal CT domain essential to the pH-sensitive gating of Cx43 gj. The K258stop truncation of the rat and mouse Cx43 CT domain at residue S257 (A257 in human Cx43) also deletes 12 of 15 identified in vitro CaM kinase II phosphorylation sites, with the exception of S244, S255, and S257 in mouse and rat Cx43 only [33]. Furthermore, the Cx43-M257 truncation mutant protein eliminates a newly identified putative Cx43 CT CaMBD located between residues K264 and T290 [34]. Thus, the involvement of CaMKII phosphorylation and binding of CaM to the Cx43 CT domain seems unlikely, but cannot be completely ruled out. Of course, under natural conditions, pH and Ca2+/CaM gating are both likely to occur during ischaemia since both chemical gating domains will be present in the native, full-length Cx43.

The existence of Cx CaMBDs has been demonstrated by in vitro CaM-binding assays with sequence-specific Cx mimetic peptides comprising the entire 15–26 amino acid CaMBDs of Cx32, Cx43, sheep Cx44 (Cx46), Cx45, and Cx50 [1822]. In past studies, these Cx CaMBD mimetic peptides were used as inhibitory peptides to validate the functional role of the Cx-specific sequence in the Ca2+/CaM gating of the parent connexin [21,23]. However, like the CaMKII 290–309 inhibitory peptide which corresponds to the high-affinity CaMBD of CaMKII, the inhibitory function of these peptides is likely based on the binding of CaM, thus acting as a ‘CaM sponge’ when added in excess to an intracellular pipette solution. To test this hypothesis, we applied increasing concentrations of the Cx50 CaMBD peptide to WT Cx43 gap junctions and found that 200 nM of the Cx50-3 peptide was sufficient to prevent the Ca2+/CaM-dependent inhibition of Cx43 gj (Figure 4A). This observation indicates that inhibition of the Ca2+/CaM gating process by these Cx-sequence-specific CaMBD mimetic peptides does not distinctly prove the functionality of the corresponding domain, but it also does not necessarily disprove the functional relevance of these identified Cx CaMBD domains. Additional experiments and novel approaches, like functional mutagenesis of known Cx CaMBDs or non-CaM-binding sequence-specific peptides, will be required to test the functionality of known Cx CaMBDs in a selective manner.

Another Cx43 mimetic peptide, the Gap19 nonapeptide, targeting CL residues 128–136 in the middle of the L2 pH receptor domain (residues 119–144) was shown to inhibit Cx43 hemichannel activity with an intracellular IC50 of 6.5 μM without affecting gj at substantially higher concentrations (400 μM), presumably by interfering with the CL–CT interaction [15,25,31]. Since the Cx43 CaMBD peptide corresponds to CL residues 136–158, it is possible that Gap19 might interfere in the Ca2+/CaM gating mechanism [19]. To test this hypothesis, we performed the Ca2+-ionomycin perfusion experiments on N2a-Cx43 cell pairs with 100 μM Gap19 added to both whole-cell patch pipettes (Figure 4B). A 100% inhibition of Cx43 Gj was still achieved within 10–12 min of perfusion though a 2–3 min delay in gj inhibition was evident. Our results confirm that Gap19 does not inhibit Cx43 gj nor prevent the Ca2+/CaM-dependent uncoupling mechanism.

The Vj-gating and γj properties of Cx43-M257 gap junctions were previously studied in exogenous Xenopus oocyte and N2a cell expression systems [26,27]. Both studies reported a loss of the fast kinetic component of the Vj-gating mechanism and a lower Gmin for the steady-state GjVj curve. Additionally, gap junction channel recordings from Cx43-M257 N2a cell pairs revealed a loss of the low γj subconductance state, an increased open time for the remaining ≥100 pS main conductance state with primarily slow transitions between the open and closed states of the channel [27]. Previously, we had observed a hysteresis in the steady-state GjVj curves obtained during the application of slow 24 s, 0 to ±120 mV Vj ramps found only in primary NMVM cell pairs, not N2a-Cx43 cell pairs [28]. The observed increase in the linear slope conductance at low Vj potentials, corresponding to the Gmax of the steady-state GjVj curves, occurred during the recovery (from inactivation) phase of the ±120 to 0 mV Vj ramp in NMVMs and was called ‘facilitation’. This facilitated recovery of Gj was not affected by non-specific serine/threonine protein kinase inhibitors (e.g. 12 μM H7, data not shown) nor 100 nM rotigaptide, though the inactivation phase was reduced by these treatments [32].

Thus, we examined the effect of the Cx43 CT truncation on the Vj-gating properties of homozygous Cx43-M257 NMVMs. Our results confirmed the loss of the fast inactivation component of Cx43 gap junctions seen in exogenous expression systems, but also the abolition of the facilitated recovery of Gj during decreasing Vj values after achieving steady-state inactivation to Gmin (Figure 5A–C). We did not, however, observe a reduction in Gmin compared with control WT NMVM gap junctions (Table 1) [28]. The reduced rate and Vj-sensitivity of the remaining slow kinetic decay component of Cx43-M257 gap junctions implies that the Cx43 CT domain contributes some of the charge for the slow Vj-gating mechanism of Cx43 gap junctions [26]. Furthermore, the abolition of the facilitated recovery of Gj from inactivating potentials was only previously attained by non-selective histone deacetylase inhibition (pan-HDACI) with 100 nM trichostatin A or 1 μM vorinostat, implying that protein acetylation directly (e.g. Cx43 CT domain) or indirectly (e.g. tubulin?) affects the Vj-gating properties of Cx43 gap junctions [35].

In summary, we conclude that the calcium/calmodulin-dependent chemical gating mechanism of Cx43 gap junctions does not require the distal CT domain of Cx43. The Gap19 peptide targeting a portion of the Cx43 CL domain also does not prevent the Ca2+/CaM gating mechanism of Cx43, but the full-length sequence-specific Cx-CaMBD mimetic peptides function as non-specific CaM-binding domains that do not necessarily infer the distinct modulatory function of these domains in their respective connexin-specific gap junctions. Lastly, we confirmed in Cx43-M257 cardiomyocyte gap junctions that deletion of the Cx43 CT domain eliminates the fast inactivation component of Cx43 gap junctions, but also eliminates the facilitated recovery of gj from inactivation.

Abbreviations

     
  • CaM

    calmodulin

  •  
  • CaMBD

    calmodulin-binding domain

  •  
  • CaMi

    calmodulin inhibition

  •  
  • CL

    cytoplasmic loop

  •  
  • CT

    carboxyl terminus

  •  
  • Cx

    connexin

  •  
  • Cx43

    connexin43

  •  
  • Cx45

    connexin45

  •  
  • Cx46

    connexin46

  •  
  • Cx50

    connexin50

  •  
  • gj

    junctional conductance

  •  
  • Gj

    normalized junctional conductance

  •  
  • γj

    single gap junction channel conductance

  •  
  • Ij

    junctional current

  •  
  • IPS

    internal pipette solution

  •  
  • Kon

    decay (inactivation) rate

  •  
  • N2a

    neuro2a

  •  
  • NMVM

    neonatal mouse ventricular myocytes

  •  
  • nS

    nanoSiemen

  •  
  • NT

    amino terminus

  •  
  • PBS

    phosphate-buffered saline

  •  
  • pS

    picoSiemen

  •  
  • Vj

    transjunctional voltage

  •  
  • WT

    wild-type

Author Contribution

S.W. performed most of the experiments, data analysis, some of the primary NMVM culturing and Cx43+/K258stop mouse genotyping, and assisted with manuscript preparation. C.C. performed some of the experiments on N2a-Cx43-M257 cell pairs. X.L. performed the Vj-gating and kinetics experiments on Cx43-M257 NMVM cell pairs. R.D.V. performed some experiments on N2a-Cx43, NMVM homozygous Cx43-M257 and WT Cx43 cell pairs, some primary NMVM culturing and Cx43+/K258stop mouse genotyping, immunocytochemical labeling, and final manuscript preparation.

Funding

This work was supported by grants from the NIH HL042220, AHA 17GRNT33710031, Hendricks Fund, and Joseph C Georg Fund from the CNY Community Foundation to R.D.V.

Acknowledgements

We thank Dr Steven Taffet for the gift of the Cx43-M257 pIRES2-EGFP plasmid. We thank Dr Karen Maass for providing us with the Cx43+/K258stop mice and the methodology for genotyping these mice.

Competing Interests

The Authors declare that there are no competing interests associated with the manuscript.

References

References
1
Harris
,
A.L.
(
2007
)
Connexin channel permeability to cytoplasmic molecules
.
Prog. Biophys. Molec. Biol.
94
,
120
143
2
Valiunas
,
V.
,
Cohen
,
I.S.
and
Brink
,
P.R.
(
2018
)
Defining the factors that affect solute permeation of gap junction channels
.
Biochim. Biophys. Acta
1860
,
96
101
3
Bukauskas
,
F.F.
and
Verselis
,
V.K.
(
2004
)
Gap junction channel gating
.
Biochim. Biophys. Acta
1662
,
42
60
4
Loewenstein
,
W.R.
,
Nakas
,
M.
and
Socolar
,
S.J.
(
1967
)
Junctional membrane uncoupling. Permeability transformations at a cell membrane junction
.
J. Gen. Physiol.
50
,
1865
1891
5
Délèze
,
J.
(
1970
)
The recovery of resting potential and input resistance in sheep heart injured by knife or laser
.
J. Physiol.
208
,
547
562
6
Rose
,
B.
,
Loewenstein
,
W.R.
and
R
,
W.
(
1975
)
Permeability of cell junction depends on local cytoplasmic calcium activity
.
Nature
254
,
250
252
7
Turin
,
L.
and
Warner
,
A.
(
1977
)
Carbon dioxide reversibly abolishes ionic communication between cells of early amphibian embryo
.
Nature
270
,
56
57
8
Turin
,
L.
and
Warner
,
A.E.
(
1980
)
Intracellular pH in early Xenopus embryos: its effect on current flow between blastomeres
.
J. Physiol.
300
,
489
504
9
Spray
,
D.C.
,
Stern
,
J.H.
,
Harris
,
A.L.
and
Bennett
,
M.V.
(
1982
)
Gap junctional conductance: comparison of sensitivities to H and Ca ions
.
Proc. Natl Acad. Sci. U.S.A.
79
,
441
445
10
Peracchia
,
C.
(
1990
)
Increase in gap junction resistance with acidification in crayfish septate axons is closely related to changes in intracellular calcium but not hydrogen ion concentration
.
J. Membr. Biol.
113
,
75
92
11
Vaughan-Jones
,
R.D.
,
Lederer
,
W.J.
and
Eisner
,
D.A.
(
1983
)
Ca2+ ions can affect intracellular pH in mammalian cardiac muscle
.
Nature
301
,
522
524
12
Burt
,
J.M.
(
1987
)
Block of intercellular communication: interaction of intracellular H+ and Ca2+
.
Am. J. Physiol. Cell Physiol.
253
,
C607
C612
13
White
,
R.L.
,
Doeller
,
J.E.
,
Verselis
,
V.K.
and
Wittenberg
,
B.A.
(
1990
)
Gap junctional conductance between pairs of ventricular myocytes is modulated synergistically by H+ and Ca++
.
J. Gen. Physiol.
95
,
1061
1075
14
Morley
,
G.E.
,
Taffet
,
S.M.
and
Delmar
,
M.
(
1996
)
Intramolecular interactions mediate pH regulation of connexin43 channels
.
Biophys. J.
70
,
1294
1302
15
Duffy
,
H.S.
,
Sorgen
,
P.L.
,
Girvin
,
M.E.
,
O'Donnell
,
P.
,
Coombs
,
W.
,
Taffet
,
S.M.
et al.  (
2002
)
pH-dependent intramolecular binding and structure involving Cx43 cytoplasmic domains
.
J. Biol. Chem.
277
,
36706
36714
16
Welsh
,
M.J.
,
Aster
,
J.C.
,
Ireland
,
M.
,
Alcala
,
J.
and
Maisel
,
H.
(
1982
)
Calmodulin binds to chick lens gap junction protein in a calcium-independent manner
.
Science
216
,
642
644
17
Peracchia
,
C.
,
Wang
,
X.
,
Li
,
L.
and
Peracchia
,
L.L.
(
1996
)
Inhibition of calmodulin expression prevents low-pH-induced gap junction uncoupling in Xenopus oocytes
.
Pflügers Arch.
431
,
379
387
18
Török
,
K.
,
Stauffer
,
K.
and
Evans
,
W.H.
(
1997
)
Connexin 32 of gap junctions contains two cytoplasmic calmodulin-binding domains
.
Biochem. J.
326
,
479
483
19
Zhou
,
Y.
,
Yang
,
W.
,
Lurtz
,
M.M.
,
Ye
,
Y.
,
Huang
,
Y.
,
Lee
,
H.-W.
et al.  (
2007
)
Identification of the calmodulin binding domain of connexin 43
.
J. Biol. Chem.
282
,
35005
35017
20
Zhou
,
Y.
,
Yang
,
W.
,
Lurtz
,
M.M.
,
Chen
,
Y.
,
Jiang
,
J.
,
Huang
,
Y.
et al.  (
2009
)
Calmodulin mediates the Ca2+-dependent regulation of Cx44 gap junctions
.
Biophys. J.
96
,
2832
2848
21
Chen
,
Y.
,
Zhou
,
Y.
,
Lin
,
X.
,
Wong
,
H.-C.
,
Xu
,
Q.
,
Jiang
,
J.
et al.  (
2011
)
Molecular interaction and functional regulation of connexin50 gap junctions by calmodulin
.
Biochem. J.
435
,
711
722
22
Zou
,
J.
,
Salarian
,
M.
,
Chen
,
Y.
,
Zhuo
,
Y.
,
Brown
,
N.E.
,
Hepler
,
J.R.
et al.  (
2017
)
Direct visualization of interaction between calmodulin and connexin45
.
Biochem. J.
474
,
4035
4051
23
Xu
,
Q.
,
Kopp
,
R.F.
,
Chen
,
Y.
,
Yang
,
J.J.
,
Roe
,
M.W.
and
Veenstra
,
R.D.
(
2012
)
Gating of connexin 43 gap junctions by a cytoplasmic loop calmodulin binding domain
.
Am. J. Physiol. Cell. Physiol.
302
,
C1548
C1556
24
Maass
,
K.
,
Ghanem
,
A.
,
Kim
,
J.-S.
,
Saathoff
,
M.
,
Urschel
,
S.
,
Kirfel
,
G.
et al.  (
2004
)
Defective epidermal barrier in neonatal mice lacking the C-terminal region of connexin43
.
Mol. Biol. Cell
15
,
4597
4608
25
Wang
,
N.
,
De Vuyst
,
E.
,
Ponsaerts
,
R.
,
Boengler
,
K.
,
Palacios-Prado
,
N.
,
Wauman
,
J.
et al.  (
2013
)
Selective inhibition of Cx43 hemichannels by Gap19 and its impact on myocardial ischemia/reperfusion injury
.
Basic Res. Cardiol.
108
,
309
26
Revilla
,
A.
,
Castro
,
C.
and
Barrio
,
L.C.
(
1999
)
Molecular dissection of transjunctional voltage dependence in the connexin-32 and connexin-43 junctions
.
Biophys. J.
77
,
1374
1383
27
Moreno
,
A.P.
,
Chanson
,
M.
,
Anumonwo
,
J.
,
Scerri
,
I.
,
Gu
,
H.
,
Taffet
,
S.M.
et al.  (
2002
)
Role of the carboxyl terminal of connexin43 in transjunctional fast voltage gating
.
Circ. Res.
90
,
450
457
28
Lin
,
X.
,
Gemel
,
J.
,
Beyer
,
E.C.
and
Veenstra
,
R.D.
(
2005
)
Dynamic model for ventricular junctional conductance during the cardiac action potential
.
Am. J. Physiol. Heart Circ. Physiol.
288
,
H1113
H1123
29
Veenstra
,
R.D.
(
2016
)
Establishment of the dual whole cell recording patch clamp configuration for the measurement of gap junction conductance
.
Methods Mol. Biol.
1437
,
213
231
30
Marty
,
A.
and
Neher
,
E
. (
1983
) Whole-cell recording. In
Single Channel Recording
(
Sakmann
,
B.
and
Neher
,
E.
eds), pp.
107
122
Plenum Press
,
New York
31
Ponsaerts
,
R.
,
De Vuyst
,
E.
,
Retamal
,
M.
,
D'hondt
,
C.
,
Vermeire
,
D.
,
Wang
,
N.
et al.  (
2010
)
Intramolecular loop/tail interactions are essential for connexin 43-hemichannel activity
.
FASEB J.
24
,
4378
4395
32
Lin
,
X.
,
Zemlin
,
C.
,
Hennan
,
J.K.
,
Petersen
,
J.S.
and
Veenstra
,
R.D.
(
2008
)
Enhancement of ventricular gap junction coupling by rotigaptide
.
Cardiovasc. Res.
79
,
416
426
33
Huang
,
R.Y.
,
Laing
,
J.G.
,
Kanter
,
E.M.
,
Berthoud
,
V.M.
,
Bao
,
M.
,
Rohrs
,
H.W.
et al.  (
2011
)
Identification of CaMKII phosphorylation sites in Connexin43 by high-resolution mass spectrometry
.
J. Proteome Res.
10
,
1098
1109
34
Sorgen
,
P.L.
,
Trease
,
A.J.
,
Spagnol
,
G.
,
Delmar
,
M.
and
Nielsen
,
M.S.
(
2018
)
Protein–protein interactions with connexin 43: regulation and function
.
Int. J. Mol. Sci.
19
,
1428
35
Xu
,
Q.
,
Lin
,
X.
,
Andrews
,
L.
,
Patel
,
D.
,
Lampe
,
P.D.
and
Veenstra
,
R.D.
(
2013
)
Histone deacetylase inhibition reduces cardiac connexin43 expression and gap junction communication
.
Front. Pharmacol.
4
,
44