Unique Conductance, Gating, and Selective Permeability Properties of Gap Junction Channels Formed by Connexin40
Abstract Connexin40 is selectively expressed in specialized cardiac conduction (nodal and His-Purkinje) tissues and the atrium, yet the channel properties formed by this gap junction protein have not been investigated. The conductance, gating, and selective permeability of rat connexin40 (Cx40) gap junction channels between pairs of Cx40-transfected mouse neuroblastoma (N2A) cells in culture were studied by using dual whole-cell voltage-clamp techniques. The macroscopic steady state junctional conductance gating was dependent on transjunctional voltage with a Boltzmann half-inactivation voltage of ±50 mV, a residual voltage-insensitive normalized junctional conductance of 35% of maximum, and a gating charge valence of 3. In the presence of 120 mmol/L potassium glutamate, the slope conductance of single rat Cx40 gap junction channels measured 158±2 pS (n=4). Lower conductance states equal to 21% to 48% of the main open-state conductance were also occasionally observed in two of the four cell pairs. Multichannel open probabilities were found to be heterogeneous. Ion substitution and dye transfer experiments were performed to determine the relative chloride/potassium conductance and dye permeability of anionic fluorescein derivatives in rat Cx40 channels. The rat Cx40 channel had a maximum conductance of 180±18 pS (n=3) in 120 mmol/L KCl and a detectable chloride permeability of 0.29 relative to potassium, indicating some selectivity for cations over anions. Cx40 gap junctions were permeable to 2′,7′-dichlorofluorescein (diCl-F) and also to the more polar 6-carboxyfluorescein dye; however, diCl-F dye transfer was not observed to increase with increasing junctional conductance.
- channel conductance ratio
- intercellular communication
Gap junction channels provide a low-resistance pathway for ion flow between cells by forming an aqueous pore permeable to hydrophilic molecules up to 1 kD in mass or 10 to 14 Å in diameter.1 2 3 Multiple protein subunits called connexins form the oligomeric channel assemblies. Connexins are a family of at least 16 structurally related proteins that contain highly conserved sequences corresponding to four putative transmembrane and two extracellular regions and also contain unique, predicted cytoplasmic loop and carboxyl terminal domains.4 The diversity of connexins expressed by various tissues suggests that the cell-specific distribution of physiologically distinct gap junctions may contribute to the tissue-specific regulation of intercellular communication. Gap junction channels formed by specific cloned connexins have unique biophysical properties and gating characteristics that correlate with the properties of the channels in cells from which those sequences were derived.5 6
The number, size, and spatial distribution of gap junctions are important determinants of passive electrophysiological properties and contribute to the velocity and anisotropy of conduction in different regions of normal and diseased myocardium.7 8 It is likely that differential patterns of expression of multiple connexins (including Cx40, Cx43, and Cx45) within the heart also contribute to differing regional conductive properties.9 For example, the fivefold increase in longitudinal conduction velocity in Purkinje fibers relative to the ventricle likely results from the relatively linear tissue arrangement of myocytes and their connections, higher sodium conductance, and tissue-specific patterns of connexin expression. Immunofluorescence and quantitative in situ hybridization studies have demonstrated that although both ventricular myocytes and Purkinje fibers contain Cx43, expression of Cx40 is increased threefold to fivefold in the Purkinje fibers relative to the ventricle.10 Various studies have demonstrated that although Cx40 is relatively scarce in the ventricle, it is found in the atrium, sinus and atrioventricular nodes, His-Purkinje system, and endothelium and in noncardiac organs, including lung, kidney, uterus, and ovary.7 9 10 11 12 13 14 15 16 Thus, a determination of the properties of Cx40 channels may have implications for understanding intercellular communication in multiple systems.
Most cells express more than one connexin, making it difficult to determine unique connexin channel properties. To study the function of a single connexin type, connexins have been exogenously expressed or transfected into communication-deficient cell lines. As in our previous investigations of connexin-specific gap junctions,5 17 18 19 we expressed Cx40 channels in mouse neuroblastoma (N2A) cells. These studies have indicated that gap junctions formed from different connexins have distinct conductance states and Vj sensitivities and, most recently, unique ionic and dye permeabilities. In the present investigation, we have combined double whole-cell patch recording20 21 22 with equimolar ion substitution and dye transfer experiments18 19 to determine the conductance, gating, and permeability properties of Cx40 channels. We demonstrate at both the macroscopic and single-channel levels that Cx40 gap junctions close with increasing Vj but with different sensitivities from other cardiac gap junctions. Interchanging between 120 mmol/L KG and KCl internal pipette solutions allowed us to investigate the ion selectivity of Cx40 channels. Cx40 was found to exhibit a 4:1 selectivity for cations over anions. Dye transfer experiments revealed that Cx40 gap junctions are only slightly permeable to diCl-F and 6-CF dyes.
Materials and Methods
Cell Culture and Transfection
Mouse N2A cells were cultured as described previously.5 17 Genomic DNA containing the coding sequence of rat Cx40 was cloned into the EcoRI site of the eukaryotic expression vector pSFFV-neo. N2A cells were transfected with 20 μg of linearized plasmid. Colonies were selected in 0.5 mg/mL G418 (GIBCO/BRL), and Cx40 expression was verified by blotting of total RNA prepared from selected clones.5 17
Electrophysiological Recordings and Solutions
Connexin-transfected N2A cell cultures were plated at low density (2′105 cells per 35-mm dish) for 12 to 24 hours, washed with HEPES-buffered saline (mmol/L: NaCl 142, KCl 1.3, MgSO4 0.8, NaH2PO4 0.9, CaCl2 1.8, dextrose 5.5, and HEPES 10, pH 7.2), and examined on the stage of an inverted phase-contrast light microscope (Olympus IMT-2). Cx40-induced coupling in the transfected N2A cells was studied by dual whole-cell recording procedures as described previously.22 Patch electrodes had resistances of 2 to 5 MΩ when filled with one of five (IPS 1, IPS KG, IPS KCl, KG/6-CF, and KG/diCl-F) internal pipette solutions. The compositions of the internal pipette solutions are listed in Table 1⇓. All experiments were performed at room temperature (20°C to 22°C). The osmolarity of all bath and internal pipette solutions was measured by the freezing-point method (model 3W2 osmometer, Advanced Instruments Inc).
To generate net macroscopic Ijs, Vj pulses were elicited by stepping the holding potential of the prejunctional cell (V1) from a common holding potential (V1=V2=0 mV, where V2 is the holding potential of the postjunctional cell) to a new value (V′1) for a minimum of 6 seconds. Each 6-second Vj pulse was separated by a 5-second recovery interval. When single-channel activity was observed, Vj pulse durations were increased to 120 seconds. Vj equals V′1−V2 when the input resistances of cells 1 and 2 are 100-fold higher (eg, 2 GΩ) than the resistances of the respective whole-cell patch electrodes (eg, 20 MΩ). Ij signals appear as simultaneous events of equal amplitude and opposite polarity, and the true Ij equals −ΔI2(1+Rel2/Rin2), where ΔI2 is the difference in holding current of cell 2, Rel2 is the patch-electrode series resistance, and Rin2 is the cellular input resistance. When Rin2 and the junctional resistance (Rj) are >2 GΩ and are 100-fold higher than the resistance of the recording electrode (Rel2/Rin2, <0.01), −Ij is ≈ΔI2 (<2% error), and the ΔI2 signal is used to measure the junctional channel current amplitudes. Junctional conductance (gj) or Rj was determined from the expression gj=1/Rj=Ij/Vj, provided that the above conditions are met. All current and voltage recordings were stored on VCR tape using a four-channel digitizing unit (DR-484 Neuro-corder, Neuro Data Instruments Corp) and a VCR tape recorder (JVC HR-D600U) for off-line analysis.
The instantaneous macroscopic Ij was determined from the average of the first 2 milliseconds of recorded ΔI2, and the macroscopic steady state Ij was determined from the average of the last second of the 6-second pulse. Because instantaneous Ij remains linear with Vj, gss was normalized to ginst of each pulse and plotted as a function of Vj. If the linear fit of the instantaneous Ij-Vj curve for each experiment did not intersect 0 mV, indicative of an offset potential, the offset (0.89±0.77 mV, mean±SEM) was subtracted from Vj before calculating the conductance. Gj (gss/ginst) was fit assuming a two-state Boltzmann distribution of the following form: where V0 is the Vj at which the voltage-sensitive conductance is half maximal, and A is a constant expressing the strength of the interaction of the open channel with Vj.23 24 The constant A is equivalent to nq/kT, where n is the number of electron charges (q) that act as the gating mechanism by sensing changes in Vj, and k and T represent, respectively, Boltzmann’s constant and absolute temperature. Gmax is the maximum normalized conductance, Gmin is the normalized Vj-insensitive (residual) conductance, and G∞ is the experimentally derived steady state Gj value.
Channel Current Analysis
For clarity, junctional channel currents were displayed as paired whole-cell currents, and all-points current amplitude histograms were compiled from the last 110 seconds of the 2-minute ΔI2 tracing for each experiment. The first 10 seconds of each 2-minute Vj pulse was omitted to exclude nonstationary data. When recording single gap junction channel currents from a cell pair with gj of <0.5 nS, all of the above conditions are met, and Ij is ≈−ΔI2. All analog signals were low pass–filtered (eight-pole Bessel, LPF-30, WPI, Inc) at 100 Hz or 1 kHz and digitized at 2 kHz and 10 kHz, respectively, by using a DT2801A A/D board (Data Translation, Inc) installed in an IBM PC/AT clone (Everex 386SX/20). The dead time of the recording instrumentation was 1.8 milliseconds.
The gaussian distributions present in the all-points current amplitude histograms were fitted with a pdf18 25 assuming i independent channels, where i is one less than the total number of peaks observed. Each channel was assigned a Pi value.26 In certain instances, the amplitude histograms cannot be accurately described by using expressions for independent channels because of the existence of intermediate conductance states or cooperative gating behavior.25 Alternative expressions assuming the existence of long-lived channel substates have been derived elsewhere.27
The respective increases in channel conductance were determined by using the GHK current equation with the permeability terms set equal to D for each of the monovalent ions. Since D is directly proportional to the aqueous mobility of each ion at a given temperature, the predicted conductance ratios of the internal pipette solutions reflect changes in the osmolarity and aqueous mobilities of the constituent ions. Dion values are as follows (10−5 cm2/s): K+ 1.96, Na+ 1.33, Cs+ 2.06, TEA+ 0.87, Cl− 2.03, and glutamate− (estimated) 0.70. An effective anion-to-cation permeability ratio was determined by scaling an anionic permeability coefficient in the GHK equation until the theoretical results matched the experimental IPS KCl–to–IPS KG unitary channel conductance ratio.28 Reducing the glutamate permeability (eg, to zero) increases the predicted IPS KCl–to–IPS KG conductance ratio to 1.72 from 1.38 and effectively reduces the anion-to-cation selectivity ratio by an additional ≈50% from the experimentally reported values.
Dye Transfer Assays
6-CF (376 D, Molecular Probes) was dissolved in 120 mmol/L potassium citrate (pH 9.0) to yield a final stock concentration of 20 mmol/L and titrated to pH 7.0. Stock 6-CF solution was stored in the dark at −20°C. For dual whole-cell recording experiments using 6-CF, an aliquot of the 20 mmol/L stock solution was diluted to a final concentration of 2 mmol/L with IPS KG. Only one patch electrode of the pair was filled with 2 mmol/L KG/6-CF, and gj measurements were obtained by following normal procedures. After a 10-minute Ij recording period, the presence or absence of dye transfer was observed under epifluorescent illumination. Phase-contrast and fluorescent micrograph images were recorded by using an automatic exposure Olympus 35-mm camera body attached to the microscope’s camera port. The same procedures were followed for diCl-F (401 D, Eastman Kodak Co).
Induction of Electrical Coupling by Cx40 Expression
The communication-deficient N2A cell line was stably transfected with rat Cx40 genomic DNA, and clones testing positive for Cx40 expression were examined for functional coupling by using the double whole-cell recording technique. Electrical communication was evident in 73% of the Cx40 pairs examined while using IPS 1, IPS KG, and IPS KCl (Table 1⇑, n=49). When IPS 1 was used, gj averaged 2.26±1.57 nS (mean±SD, n=28, Fig 1⇓). Cell pairs using IPS KG or IPS KCl that did not result in single-channel recordings (gj, >0.5 nS) were not subjected to further analysis.
Voltage-Dependent Gating of gj
The Vj dependence of rat Cx40 gj was investigated by stepping V1 from 0 to ±100 mV in 10-mV increments while holding V2 constant at 0 mV. Representative Ijs are presented in Fig 2A⇓. Instantaneous Ij increased linearly with Vj in either direction. Ij underwent a time-dependent decay when Vj exceeded ±50 mV. The instantaneous and steady state Ij values are plotted as functions of Vj in Fig 2B⇓, clearly illustrating the voltage dependence of Ij. The instantaneous Ij-Vj relation for the data shown in Fig 2B⇓ approximates a straight line with a slope of 3.5 nS over the entire voltage range. The steady state Ij-Vj relation has a linear slope of 3.9 nS over the range of ±30 mV but levels off above ±70 mV.
If the linear fit of the instantaneous Ij-Vj curve for each experiment did not intersect 0 mV, the resulting offset potential (0.89±0.77 mV, mean±SEM, n=10) was subtracted from Vj before calculating ginst and gss. gss was normalized to the ginst value of each pulse and plotted as a function of Vj. The results are illustrated in Fig 3⇓. Each point represents the normalized conductance, Gj, at each Vj. The solid line represents the best fit of the data assuming a two-state Boltzmann distribution. For this graph, Gmax=1.0 and G∞ are the experimentally derived steady state Gj values. For the curve shown in Fig 3⇓, the Boltzmann parameters (Table 2⇓) were Gmin=0.33, n=3.2, and V0=−54 mV for negative Vj values and Gmin=0.28, n=2.8, and V0=+47 mV for positive Vj values.
Single-Channel Conductance and Ionic Selectivity
Since the gj of Cx40-transfected N2A cell pairs is the product of the number of open channels and their γj values, we continued our study with an investigation of the γj and gating properties of the Cx40 channel. Fig 4A⇓ illustrates one example of whole-cell currents containing single Cx40 gap junction channel openings and closings in the presence of symmetrical IPS KG solutions at a Vj of −20 mV. Four seconds of a full 2-minute continuous recording are shown. The amplitude histogram in Fig 4B⇓, compiled from the −ΔI2 tracing from the same experiment, has four peaks corresponding to current levels L1 through L4 in Fig 4A⇓. The dots represent the detected number of occurrences (digitized points) at each current level. The current-amplitude histogram is fitted (solid line) with a pdf assuming n=3 independent channels, with a single-current amplitude of −3.25 pA and a baseline (closed-channel) variance of 0.31 pA. The open-channel variance was 0.31 pA. However, to accurately define the amplitude of each current peak, two different open probability values of .75 (for P1) and .25 (for P2) were assigned to one and two channels, respectively. The single-channel Ij-Vj relation is shown in Fig 4C⇓. A slope conductance of 157 pS (r>.99), obtained by linear regression fit of the Ij-Vj relation, was determined in the presence of IPS KG (Table 1⇑). Similar Ij-Vj relations were obtained from three other Cx40 cell pairs, and the individual slope conductances were 156, 157, and 162 pS. The slope conductances for all IPS KG experiments averaged 158±2 pS (mean±SD, n=4).
To examine the anion permeability of Cx40 channels, channel current amplitudes were determined in separate experiments by using a modified internal pipette solution (IPS KCl, Table 1⇑), where an equimolar concentration of KCl (120 mmol/L) replaced KG. Channel currents were again recorded at different Vjs. Fig 5A⇓ is a 3-second segment of the paired channel currents obtained during a 2-minute Vj pulse to −30 mV. Numerous transitions between five different conductance states were observed. The current amplitude distribution for the last 110 seconds of the 2-minute recording is shown in Fig 5B⇓. Four independent open channels were assumed for the pdf fit (solid line). The channels had a current amplitude of 4.05 pA and a baseline variance of 0.19 pA. The open-channel variance was set to 0.19 pA. Three different open probabilities were assigned: .11 (P1), .49 (P2), and .89 (P3 and P4). The single-channel Ij-Vj relation is plotted in Fig 5C⇓. The slope conductance in the presence of IPS KCl was 172 pS (r>.99). Single-channel Ij-Vj relations were obtained from two additional Cx40 cell pairs, and the individual slope conductances were 167 and 200 pS. The slope conductances for all IPS KG experiments averaged 180±18 pS (mean±SD, n=3).
The GHK current equation predicts a 36% increase in Ij and gj between IPS KG and IPS KCl if the permeabilities of all monovalent ions are assumed to be directly proportional to the respective aqueous mobilities of each ion (see “Materials and Methods”).28 The IPS KCl–to–IPS KG single-channel junctional slope conductance (γj) ratio (180/158 pS) was 1.14. This 14% increase in maximum conductance is 61% less than the expected 36% increase in γj (ie, to 215 pS in IPS KCl), assuming that the local ion concentrations near the mouth of the Cx40 channels are equivalent to their respective values in the internal pipette solution. The experimental Cx40 IPS KCl–to–IPS KG conductance ratio can be approximated by lowering the relative anion-to-cation permeability within the GHK current equation to 0.29.
Multiple Conductance States
In two of four cell pairs using IPS KG and one of three cell pairs using IPS KCl, the Cx40 channel was observed to transit from either the maximum open (main) state or the ground (closed, nonconducting) state to various nonzero conducting states (data not shown). The occurrence of multiple conductance states was identified by the presence of an additional peak located between the closed state and the fully open-state peaks in the corresponding current amplitude histogram, which could not be defined by assuming multiple independent channel types.29 The additional conductance states were observed infrequently, not at every applied Vj for each cell pair, and ranged from 21% to 48% of the main state conductance. Additional recordings are required for statistical analysis by following previously described procedures.18
Channel Open Probability
Channel open probabilities at each Vj were obtained by determining the channel amplitude histogram pdf for each 2-minute Vj pulse. The single-channel gj and open probabilities were determined for cell pairs with five or less open channels observed. The open probabilities at each Vj for the four cell pairs using IPS KG and the three cell pairs using IPS KCl are listed in Table 3⇓. Table 3⇓ also includes the number of events or transitions between discrete current levels at each Vj. At some Vjs, distinct open-channel peaks in the all-points current amplitude histogram were readily observed for amplitude and event count determinations, but independent channel pdf values could not be obtained (because of either substate activity or cooperativity among multiple channels); therefore, channel open probabilities could not be estimated. The maximum number of open channels was not observed at each Vj. At most Vjs, the open-channel probabilities for each channel in a multichannel recording were not equal.
To examine the dye permeability of Cx40 and correlate it with electrical coupling, dual whole-cell recording experiments were performed on Cx40 N2A cell clones with 2 mmol/L 6-CF or 2 mmol/L diCl-F added to IPS KG in one of the patch pipettes. These anionic dyes were selected because of their molecular weights (diCl-F, Mr=401; 6-CF, Mr=376) and hydrophilic character (>99% ionized at pH 7.0). diCl-F, an anion at physiological pH, is reported to have a nexal membrane permeability that is 4 to 20 times greater than that for 6-CF, which has a valence of −2 at physiological pH.30 Both diCl-F and 6-CF have little binding affinity (<7%) for cytoplasmic proteins.30 31 The unhydrated radius of the fluorescein derivatives are approximately equal (9.5 Å), but the hydration radius of 6-CF would be expected to be greater than that of diCl-F because of its greater polarity.30
Using diCl-F, dual whole-cell recording experiments were performed on 16 Cx40-transfected N2A cell pairs with conductances ranging from 0.4 to 8.3 nS. As illustrated in Fig 6A⇓, of the 16 experiments, only one pair with a gj of 1.2 nS (6%) exhibited diCl-F dye transfer. In one additional experiment, dye coupling developed rapidly, and the second cell of the pair fluoresced with an intensity equal to the dye-injected cell. The gj of this cell pair was 5.25 nS, and Vj- and time-dependent decay of Ij was not evident. These properties are consistent with the low incidence (15% to 20% among pairs)18 of cytoplasmic bridges between cells, and this pair was excluded in the final analysis of Cx40-mediated dye transfer.
The same procedures were repeated with 2 mmol/L 6-CF added to IPS KG of cell 1 in each of 10 experiments. gj ranged from 0.4 to 12.6 nS, and all three (30%) of the pairs with conductances of >7 nS exhibited dye transfer as shown in Fig 6B⇑. The other seven pairs were negative for dye transfer (gj, <7 nS).
The results of the present study demonstrate that rat Cx40 forms a functional gap junction channel with Vj-dependent gating, channel conductance, and selective ionic and dye permeability properties distinct from other reported cardiovascular connexins. Typically, the functional expression of each newly identified connexin is accompanied by a Boltzmann curve describing the Vj dependence of steady state gj as originally described for Vj-dependent gap junctions between amphibian blastomeres.23 24 This two-state Boltzmann model provides a fingerprint for the equilibrium properties of the different Vj-dependent connexins. When mouse Cx40 was expressed in oocytes, a V0 of 35 mV, Gmin of 0.19, and slope factor of 0.32 (gating charge valence, 8) were observed.12 Our average values for rat Cx40 (V0, 50 mV; Gmin, 0.30; and slope factor, 0.12 [gating charge valence, 3]) are indicative of less Vj-sensitive gating than the murine isoform of Cx40. These variations may be attributed to differences in the two expression systems, as reported for Cx32 (Moreno et al, 1994). Mouse and rat Cx40 amino acid sequences are 92% identical, and their disparate Vj sensitivities may reflect primary sequence variations. The greatest similarity Cx40 shares with any of the other known connexins is ≈70% identity with chick Cx42.11 12 The rat Cx40 Boltzmann parameters compare well with those reported for chick Cx42 (V0, 41 mV; slope factor, 0.108; and Gmin, 0.38).5 The Vj dependence of rat Cx40 also differs substantially from that of chick Cx45 (V0, 39 mV; slope factor, 0.055; and Gmin, 0.09) and chick Cx43 (V0, ±77 mV; slope factor, 0.064; and Gmin, 0.53).5
If γj and the number of channels in the membrane remain constant, then Vj-dependent changes in conductance must be manifested as changes in channel gating properties. To date, gap junction channel gating has been modeled by using a simple two-state (all-or-none) gating scheme23 24 32 33 despite recent evidence to the contrary (ie, the existence of multiple or voltage-sensitive subconductance states).15 19 27 Under the two-state gating scheme, rapid reductions in gj are likely to result from a reduction in the number of open channels. The Vj-dependent behavior of several gap junction channels has been attributed to Vj-dependent modulation of channel open probability.34 35 The heterogeneity of multichannel open probabilities precludes us from making any direct comparisons between macroscopic (Fig 3⇑) and unitary (Table 3⇑) probability distributions. The factors producing these multiple open probabilities (possibly the gradual loss of Vj sensitivity as channel number increases within a gap junction plaque34 36 ) remain to be determined.
Cx40 single-channel conductance (158 pS) is unique relative to other cardiovascular connexins; it is less than that for Cx37 (300 pS, the largest γj reported for all connexins under similar ionic conditions)18 and greater than that for Cx45 (26 pS, the smallest γj reported for all known connexins).19 Cx40 channels could thus be considered midrange. Chick and rat Cx43 exhibit multiple conductance states, with the most prevalent γj being 45 to 60 pS, depending on ionic conditions.5 34 Hence, rat Cx40 has a greater γj than the predominant Cx43 state by a factor of 3. Chick Cx42 exhibits six γj states in multiples of 40 pS, with the 160 pS γj state being the most prevalent at 22°C.32 Although Cx40 and Cx42 possess common 160-pS γj states, there are functional differences in the number of observed γj states, and the permeability properties of Cx42 remain to be determined. The maximum γj value (158 pS) for the Cx40 channel is in close agreement with the recently published γj value (153 pS) for mouse Cx40 channels expressed in transfected HeLa cells.37 In these same experiments, a second γj of 121 pS for the Cx40 channel was also reported. It is not known whether these 121-pS transitions were observed in single or multichannel recordings; therefore, it remains to be determined whether this intermediate γj is a substate of the 153-pS channel. From our single-channel recordings, distinct intermediate γj transitions ranging from 21% to 48% of the main state conductance were infrequently observed; we are preliminarily attributing this to the presence of subconductance states of the main Cx40 channel. Unfortunately, the data are not sufficient to definitively demonstrate that these intermediate γj transitions are substates of the main 157-pS channel, as has been demonstrated for other channels.18 19 35 These differing observations for Cx40 from two different species may also reflect functional differences due to primary amino acid sequence variations.
The KG/KCl ion substitution experiments (Figs 4⇓ and 5⇑) provide evidence for a reduced anionic permeability of a midrange γj gap junction channel. Cx40 γj is approximately three times greater than the most prevalent Cx43 γj state, yet the anion permeability is about five times less.38 Furthermore, Cx40 anion permeability is only 50% greater than Cx45 anion permeability, even though γj is fivefold greater.19 For a diffusion-limited ionic channel conductance, as ascribed to the simple pore model, the monovalent ion selectivity sequence should be determined by the aqueous mobilities. Because of the relative aqueous mobilities of all pertinent permeant ions (K+, 1; Cs+, 1.05; TEA+, 0.44; glutamate−, 0.36; and Cl−, 1.04), a 36% increase in maximum γj of the Cx40 channel was expected with the substitution of IPS KCl for IPS KG. One possible explanation for the observed 14% increase in γj is reduced relative anionic permeabilities. A relative anion-to-cation permeability of 0.29, which reduces the relative glutamate−/K+ and Cl−/K+ permeabilities to 0.10 and 0.3, respectively, could account for the lower conductance. The precise mechanism for this modest selectivity of the Cx40 channel cannot be determined from our results. Selectivity can result from specific binding affinities of the pore for various ions or can be the result of electrostatic attractive and repulsive forces acting on the electrolyte solution to alter local ionic concentrations near the mouth of the pore.39 A fixed ring of acidic amino acids (eg, nicotinic acetylcholine receptor subunits40 ) in the Cx40 sequence associated with the pore, which reciprocally increases the local cation concentration and decreases the local anion concentration by a factor of 1.85 while maintaining ionic strength constant near the mouth of the pore, could account for the 14% increase in channel conductance. For a homotypic connexin channel (all subunits are identical) in symmetrical solutions, no Nernst potentials would develop in response to the electrostatic field.
Mouse Cx40 was also observed by Traub et al37 to have a low permeability to Lucifer yellow relative to Cx43, an observation that they attributed to differences in the number of open and equally permeable Cx43 and Cx40 channels at a common gj value. However, their dye transfer and gj measurements of Cx40- and Cx43-transfected HeLa cells were performed in separate experiments. Our combined electrical and dye-coupling experiments using 6-CF or diCl-F have the added advantage of directly correlating dye transfer with gj in each cell pair. The incidence of dye transfer was low for both dyes: 6% for diCl-F and 30% for 6-CF. Transfer of diCl-F was not directly correlated with the magnitude of gj. A low incidence of dye transfer was previously found with human Cx37, where the poor dye permeability was attributed to the presence of long-lived subconductance states having only 20% of the maximum γj.18 We propose that the relatively low dye permeability of Cx40 channels to all three anionic dye molecules (diCl-F, 6-CF, and Lucifer yellow CH) is primarily due to the 4:1 cation-to-anion selectivity of the 160-pS channel. Although dye transfer studies have often been used to assay for the presence of functional gap junctions, the present study illustrates the necessity of considering dye permeability in the final interpretation of dye transfer experiments. The presence or absence of dye transfer may not indicate the full extent of functional gap junction coupling, depending on the pore size and selectivity, conductance state, and the number of open channels. The low incidence of 6-CF and diCl-F dye transfer in Cx40-transfected N2A cell pairs may be due to the presence of a negatively charged electrostatic field associated with the channel, since both dyes are negatively charged at physiological pH. This mechanism for relative cation selectivity could be similar to that of the nicotinic acetylcholine receptor, where three rings of negatively charged residues directly associated with the pore influence the channel conductance, cationic selectivity, and degree of Mg2+ block.40
So what are the physiological consequences of these Vj-dependent gating, cationic selectivity, and gj properties of the Cx40 channel in the heart? Cx40 displays a distinct pattern of expression in cardiac tissues relative to Cx43 and Cx45. Cx40 is present in gap junction plaques between myocytes in the cardiac atrioventricular node, atrioventricular bundle, and Purkinje fibers.10 16 41 It has been suggested that the coexpression of Cx40 and Cx43 in the atrioventricular nodal tissues and Purkinje fibers may have an important functional role in action potential propagation.10 15 16 Propagation of action potentials in the direction parallel to the long myocyte axis is much more rapid in Purkinje fibers than in ventricular muscle by a factor of 5 to 10.10 A greater density of higher γj Cx40 gap junctions in intercalated disks of Purkinje fiber myocytes compared with ventricular muscle gap junctions may partially account for the longer resting space constant and faster action potential propagation in Purkinje fibers. However, Cx40 is also abundant in the slow-conducting nodal cardiac tissues, which suggests that the presence of Cx40 channels alone does not account for increased conduction velocity of a particular tissue. Cell excitability and the arrangement of cells and their interconnections must also be considered; Purkinje fibers also have more sodium current and a more linear geometry, which contribute to increased action potential propagation.
If Cx40 and Cx43 open probabilities for low Vj are comparable, then a higher γj for Cx40 implies that fewer open channels are required to produce effective electrical coupling between the vascular smooth muscle cells or cardiac myocytes expressing Cx40. Although this may imply a higher safety factor for propagation, a lower number of open channels may conversely render the tissue more susceptible to conduction block if the Cx40 channels are gated closed. Hence, the differential regulatory and permeability properties of connexin-specific channels will also be relevant to the overall tissue physiology. The regulatory properties of Cx40 channels in response to second messengers and intracellular ions (H+ and Ca2+) remain to be examined. We previously proposed that Vj-dependent gating may be an efficient protective mechanism that inhibits cell-cell communication between cells with different resting potentials (eg, cells damaged by ischemia or depolarized cells developing spontaneous electrical activity).31 Within seconds, the developing transjunctional potential between depolarized cells and fully polarized cells would be expected to induce closure of a significant proportion of gap junction channels composed of Cx40 and to inhibit the spread of ischemic injury (in the form of depolarizing current or chemical factors) to neighboring cells. The reduced anionic permeability of Cx40 channels implies that many second messengers (cAMP, ATP, and inositol trisphosphate), which are all anionic at physiological pH, will have reduced intercellular permeabilities in tissues expressing Cx40 relative to those expressing predominantly Cx43. The major role of Cx40 channels may be to facilitate the intercellular passage of ions.
In conclusion, the present study demonstrates that Cx40 has conductance, permeability, and gating properties that are distinct from the other two mammalian cardiac connexins, Cx43 and Cx45. Cx40 channels have a fivefold higher cationic selectivity, lower anionic dye permeability, greater Vj sensitivity, and a threefold higher γj than do Cx43 channels. The functional expression of Cx40 in N2A cells provides a useful system for examining the differential gating, regulatory, and selective permeability properties of connexin-specific gap junction channels. Future investigations in connexin-transfected and native cells will further delineate the unique physiological properties imparted on a specific tissue by the expression of distinct connexins.
Selected Abbreviations and Acronyms
|ΔI2||=||difference in holding current of cell 2|
|Cx40, Cx43, and Cx45||=||connexin40, connexin43, and connexin45, respectively|
|G∞||=||experimentally derived steady state Gj value|
|ginst||=||instantaneous junctional conductance|
|Gmax||=||maximum normalized conductance|
|Gmin||=||normalized Vj-insensitive conductance|
|gss||=||steady state junctional conductance|
|Gss||=||normalized steady state junctional conductance|
|IPS 1||=||internal pipette solution 1|
|IPS KCl||=||KCl internal pipette solution|
|IPS KG||=||KG internal pipette solution|
|=||probability density function|
|Pi||=||open probability of channel, where i=1 to 5|
|Rel2||=||patch-electrode series resistance|
|Rin2||=||cellular input resistance|
|V0||=||Vj at which voltage-sensitive conductance is half maximal|
|V1||=||holding potential of the prejunctional cell|
|V2||=||holding potential of the postjunctional cell|
This study was supported by grants HL-42220, HL-45466, and HL-31299 of the National Institutes of Health and a Monsanto-Searle/Washington University Biomedical Research Agreement. Dr Veenstra and Dr Beyer are Established Investigators of the American Heart Association. This work was conducted with the technical assistance of M. Chilton. We wish to thank Dr J. Jalife for a partial loan of equipment and Drs Peter Brink and S.V. Ramanan for helpful discussions and the channel analysis software.
- Received November 11, 1994.
- Accepted June 22, 1995.
- © 1995 American Heart Association, Inc.
Imanaga I, Kameyama M, Irisawa H. Cell-to-cell diffusion of fluorescent dyes in paired ventricular cells. Am J Physiol. 1987;252:H223-H232.
Beyer EC, Veenstra RD. Molecular biology and electrophysiology of cardiac gap junction proteins. In: Peracchia C, ed. Handbook of Molecular and Cellular Physiology of Membrane Channels. New York, NY: Academic Press Inc; 1994:379-401.
Veenstra RD, Wang H-Z, Westphale EM, Beyer EC. Multiple connexins confer distinct regulatory and conductance properties of gap junctions in developing heart. Circ Res. 1992;71:1277-1283.
Saffitz JE, Kanter HL, Green KG, Tolley TK, Beyer EC. Tissue-specific determinants of anisotropic conduction velocity in canine atrial and ventricular myocardium. Circ Res. 1994;74:1065-1070.
Spach MS. Changes in topology of gap junctions as an adaptive structural response of the myocardium. Circulation. 1994;90:1103-1106.
Kanter HL, Laing JG, Beau SL, Beyer EC, Saffitz JE. Distinct patterns of connexin expression in canine Purkinje fibers and ventricular muscle. Circ Res. 1993;72:1124-1131.
Hennemann H, Suchyna T, Lichtenberg-Fraté H, Jungbluth S, Dahl E, Schwarz J, Nicholson BJ, Willecke K. Molecular cloning and functional reexpression of mouse connexin40, a second gap junction gene preferentially expressed in lung. J Cell Biol. 1992;117:1299-1310.
Haefliger JA, Bruzzone R, Jenkins NA, Gilbert DJ, Copeland NG, Paul DL. Four novel members of the connexin family of gap junction proteins: molecular cloning, expression, and chromosome mapping. J Biol Chem. 1992;267:2057-2064.
Bastide B, Neyses L, Ganten D, Paul M, Willecke K, Traub O. Gap junction protein connexin40 is preferentially expressed in vascular endothelium and conductive bundles of rat myocardium and is increased under hypertensive conditions. Circ Res. 1993;73:1138-1149.
Gourdie RG, Severs NJ, Green CR, Rothery S, Germroth P, Thompson RP. The spatial distribution and relative abundance of gap-junctional connexin40 and connexin43 correlate to functional properties of components of the cardiac atrioventricular conduction system. J Cell Sci. 1993;105:985-991.
Reed KE, Westphale EM, Larson DM, Wang H-Z, Veenstra RD, Beyer EC. Molecular cloning and functional expression of human connexin37, an endothelial cell gap junction protein. J Clin Invest. 1993;91:997-1004.
Veenstra RD, Wang H-Z, Beyer EC, Brink PR. Selective dye and ionic permeability of gap junction channels formed by connexin45. Circ Res. 1994;75:483-490.
Veenstra RD, DeHaan RL. Measurement of single channel currents from cardiac gap junctions. Science. 1986;233:972-974.
Veenstra RD, Brink PR. Patch clamp analysis of gap junctional currents. In: Stevenson B, Paul DL, Gallin W, eds. Cell-Cell Interactions: A Practical Approach. Oxford, UK: IRL Press; 1992:167-201.
Harris AL, Spray DC, Bennett MVL. Kinetic properties of a voltage-dependent junctional conductance. J Gen Physiol. 1981;77:95-117.
Spray DC, Harris AL, Bennett MVL. Equilibrium properties of a voltage-dependent junctional conductance. J Gen Physiol. 1981;77:77-93.
Colquhoun C, Hawkes AG. Principles of stochastic interpretation of ion-channel mechanisms. In: Sakmann B, Neher E, eds. Single Channel Recording. New York, NY: Plenum Publishing Corp; 1983:135-175.
Hille B. Ionic Channels of Excitable Membranes. Sunderland, Mass: Sinauer Associates Inc; 1992;268:345-357.
Veenstra RD. Voltage-dependent gating of gap junction channels in embryonic chick ventricular cell pairs. Am J Physiol. 1990;258:C662-C672.
Rook MB, Jongsma HJ, van Ginneken AC. Properties of single gap junctional channels between isolated neonatal rat heart cells. Am J Physiol. 1988;255:H770-H782.
Chanson M, Chandross KJ, Rook MB, Kessler JA, Spray DC. Gating characteristics of a steeply voltage-dependent gap junction channel in rat Schwann cells. J Gen Physiol. 1993;102:925-946.
Veenstra RD, Wang H-Z, Beblo DA, Chilton MG, Harris AL, Beyer EC, Brink PR. Selectivity of connexin-specific gap junctions does not correlate with channel conductance. Circ Res. In press.
Gourdie RG, Green CR, Severs NJ, Anderson RH, Thompson RP. Evidence for a distinct gap-junctional phenotype in ventricular conduction tissues of the developing and mature avian heart. Circ Res. 1993;72:278-289.