| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
CELLULAR |
1 Department of Pharmacology, SUNY Upstate Medical University, 750 East Adams Street, Syracuse, NY 13210, USA
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
) approximation was decreased by about 50%. The transjunctional voltage (Vj)-dependent gating of homotypic Cx40 H15Q + K16A mutant gap junctions was also significantly reduced. The minimum normalized steady-state junctional conductance (Gmin) increased from 0.17 to 0.72, with an increase in the half-inactivation voltage from 48 to 60 mV. However, the unitary junctional conductance (
j; 160 pS) was only slightly altered, and the relative cation/anion conductance and permeability ratios were unchanged from wild-type Cx40 gap junction channels. The relative K+/Cl– permeability (PK/PCl) ratio increased from six to ten when [KCl] was reduced to 25% of normal. These data suggest that the HKH motif at positions 15–17 is important to the conformational structure of the putative voltage sensor and spermine receptor of Cx40, without causing significant alteration of the electrostatic surface charge potentials that contribute to the ion selectivity of this gap junction channel.
(Received 28 August 2005;
accepted after revision 7 November 2005;
first published online 10 November 2005)
Corresponding author R. D. Veenstra: Department of Pharmacology, SUNY Upstate Medical University, 750 East Adams Street, Syracuse, NY 13210, USA. Email: veenstrr{at}upstate.edu
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The NT domain of the connexins is postulated to control the voltage polarity of Vj gating (Verselis et al. 1994). The connexin NT domain is thought to form an 
–helix that points inward towards the connexin channel pore due to the formation of a helix–loop–helix motif that positions the first half of the NT domain within the Vj field (Purnick et al. 2000a). Polar or charged NT amino acid mutations alter the Vj-gating properties of several connexins, and also affect the corresponding hemichannel and gap junction channel conductances (Verselis et al. 1994; Oh et al. 1999; Musa et al. 2004; Tong et al. 2004). It is generally observed that the Vj-gating polarity is oppositely affected by the valence of the charge substitution on NT amino acids up to position 10 in the ß-group of connexins (Purnick et al. 2000b). This raises the possibility that at least the first half of the connexin NT domain translocates towards the centre of the channel when the appropriate Vj polarity is applied to that side of the gap junction, although the Cx32 R15Q Charcoat-Marie-Tooth disease (CMTX) mutation also slightly alters the Vj gating of homotypic or heterotypic Cx26 and Cx32 gap junctions (Abrams et al. 2001).
Cx40 and Cx43 possess other alternative amino acid loci on their respective NT domains, including positions 15–17 of this initial cytoplasmic connexin domain. We demonstrate that substituting the Q15 and A16 residues found in Cx43 for the naturally occurring H15 and K16 residues of Cx40 dramatically diminished the inhibition of Cx40 gap junction currents (Ij) by intracellular spermine. Significant alterations in the Vj-gating properties of Cx40 were also observed. However, in contrast to previous findings, the H15Q + K16A mutations did not significantly affect the unitary channel conductance of the homomeric homotypic and heterotypic or heteromeric mutant/wild-type (wt) Cx40 gap junction channels. These experiments demonstrate that polar and charged amino acid residues beyond the midpoint of the –type connexin NT domain are vital to the structural channel components required for Vj-dependent gating and spermine block of gap junctions. Furthermore, the Cx40 H15Q + K16A mutations demonstrate a separation of function between the Vj-dependent gating processes and the conductance and ionic selectivity properties of the Cx40 gap junction channel.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Stable murine neuro2a (N2a) neuroblastoma cell clones of rat connexin40 (Cx40) were grown in culture media containing 90% minimum essential medium (MEM), 10% heat-inactivated fetal bovine serum, 0.1 mM non-essential amino acids, 2.5 mML-glutamine, 0.25 g l–1 G418 sulphate, and 10 U µg–1 ml–1 penicillin/streptomycin (Invitrogen, Carslbad, CA, USA). Confluent cell monolayers were passaged weekly and 1–2 x 105 cells were plated in a 35 mm culture dish for electrophysiological examination the next day. Parental N2a cells were grown in MEM culture media minus the G418 sulphate. All T25 culture flasks were kept in a 5% CO2 humidified incubator at 37°C.
Expression of site-directed mutations
The rat Cx40 cDNA was mutated by PCR-based site-directed mutagenesis, and inserted into the pTracer-CMV2 vector (Invitrogen) by directional cloning using two restriction endonucleases. The mutant DNA sequence was confirmed by automated DNA sequence analysis by the Upstate Medical University DNA Core Facility. A plasmid preparation was prepared using the EndoFree plasmid maxi kit according to the manufacturer's instructions and stored at –20°C (Qiagen, Valencia, CA, USA). Transient transfections were performed on 80% confluent parental N2a cells grown in 24-well culture plates on a daily basis. Each well was transfected for 4 h with approximately 1 µg DNA and 3 µl of Lipofectamine 2000 in 100 µl of Opti-MEM, according to manufacturer's instructions (Invitrogen). The cells were then split into two 35 mm culture dishes containing 3 ml of MEM culture medium and incubated overnight. Green-fluorescent-protein (GFP)-positive cell pairs were identified under epifluorescent illumination on the stage of an Olympus IMT-2 inverted phase-contrast microscope at 470 nm excitation and >500 nm emission wavelengths.
Electrophysiological measurements
All dual whole-cell voltage-clamp experiments were performed using two RK-400 patch clamp amplifiers (Molecular Kinetics, Inc., Pullman, WA, USA) at room temperature (20–22°C) on the stage of an Olympus IMT-2 microscope equipped with phase-contrast optics for observation of live cells. The 35 mm cell culture plates were rinsed and bathed with serum-free saline (mM: NaCl 142, KCl 1.3, CsCl 4, TEACl 2, MgSO4 0.8, NaH2PO4 0.9, CaCl2 1.8, dextrose 5.5, Hepes 10, pH 7.4). Patch pipettes were filled with a KCl internal pipette solution (IPS KCl; mM: KCl 140, CsCl 4, TEACl 2, MgCl2 1, CaCl2 3, BAPTA 5, Hepes 25, pH 7.4). IPS potasssium gluconate contained 140 mM potassium gluconate in place of KCl. The final osmolarity of all external and internal solutions was adjusted to 310 mosmol l–1. Raffinose (Sigma, St Louis, MO, USA) was added to the 25, 50 and 75% IPS KCl to maintain the osmolarity. The K+ activities of each IPS were measured using ion-selective electrodes.
Mg2ATP (3 mM) was added fresh daily to the IPS KCl. A stock solution of 0.5 M spermine(HCl)4 (Calbiochem, La Jolla, CA, USA) was diluted daily as required with IPS KCl.
Patch electrodes had a measured bath resistance (Rel) of 3–6 M
and dual whole-cell patch electrode series resistance was corrected for using previously described methods (Veenstra, 2001a). Whole-cell currents were digitally sampled at 1 or 4 kHz using pClamp8.2 software (Axon Instruments) after low-pass filtering at 100 or 500 Hz with a four-pole Bessel filter (LPF-2, Warner Instruments). Statistical curve fitting of experimental data was performed using the least-squares method of iteration in the Clampfit8.2 application (Axon Instruments).
Secondary connexin structure modelling
The full-length rat Cx40, Cx43, connexin26 (Cx26) and connexin32 (Cx32) primary amino acid sequences were analysed using the PREDICTOR application (Combet et al. 2000). This program compares the output of eight different secondary-structure predictor programs (DPM, DSC, GOR4, HNNC, PHD, Predator, SIMPA96, and SOPM) to arrive at a consensus prediction.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Previously it was shown that substitution of the Cx40 position 9 or 13 glutamic acid residues with the corresponding position 9 or 13 lysine residues from Cx43 reduced or eliminated the spermine block of homotypic mutant Cx40 gap junctions (Musa et al. 2004). However, the reciprocal K9E and K13E mutations into Cx43 did not confer spermine sensitivity to the homotypic mutant Cx43 gap junctions. This led us to consider the involvement of other disparate amino acid sequences between Cx40 and Cx43 as structural components of the Cx40 spermine-blocking site. The amino-terminal rat Cx40 and Cx43 primary amino acid sequences and their predicted consensus secondary structures are shown in Fig. 1 along with the same for Cx26 and Cx32. The previously mutated sites are italicized and, for this study, we replaced the Cx40 H15 and K16 residues with the Q15 and A16 residues from Cx43. These amino acid residues are predicted to reside on a cytoplasmic –helix formed by the NT domain of these two –group connexins (Combet et al. 2000).
|
|
|
|
|
Effective electrical distance estimates for wtCx40 and H15Q + K16A gap junctions
Vj-dependent Kd values indicate that the blocking molecule senses a fraction of the voltage field at the site of occlusion. Woodhull (1973) ascribed this phenomenon to a two-barrier binding-site model where the inner well (binding site) lies an electrical distance () from the inside of the channel. The Vj-dependent Kd values from Table 1 were fit with the expression:
|
|
values are provided in Table 2. Spermine has a valence of +4 at physiological pH, so the equivalent values are 0.82 for the wtCx40 and 0.48 for the mutant Cx40 H15Q + K16A gap junctions. The H15Q + K16A mutation reduced the fraction of the Vj field sensed by the spermine molecule at the blocking site to approximately 50% of the assumed maximum (z= 4.0; Musa & Veenstra, 2003). From Table 1, it is also apparent that the H15Q + K16A mutation increased the Vj-dependent minimum Kd by an order of magnitude from 200 µM to 2 mM.
|
|
One of the predictions of the above Woodhull model analysis is that the relative association and dissociation rates should be, respectively, decreased and increased by the H15Q + K16A mutation. The kinetics of spermine association at positive Vj values were quantified by fitting the decay phase of Ij during the positive Vj steps with a first-order exponential function at all Vj and [spermine]. The spermine on-rates (kon) were calculated using the expression:
|
|
decay was the decay constant under both experimental conditions. The corresponding off-rates (koff) were calculated from the expression:
|
|
|
|
rise. At positive Vj values, Popen is the fraction of Ij that remained on at the end of the Vj pulse, and the steady-state Popen= 1 at all negative Vj values. The on-rates were plotted as a function of Vj for the wtCx40 and mutant Cx40 H15Q + K16A gap junctions in Fig. 4A and C. The concentration-dependent on-rates were exponentially distributed with respect to Vj. The first-order equation for the wtCx40 concentration- and Vj-dependent spermine association rates was:
|
|
|
|
value for this mutant gap junction.
|
|
|
|
|
Vj gating of wtCx40 and Cx40 H15Q + K16A gap junctions
The steady state Ij–Vj relationships for the wtCx40 and mutant Cx40 H15Q + K16A gap junctions suggest that the Vj gating properties were altered by these experimental conditions. To examine this aspect, 200 ms mV–1Vj ramps from 0 to ±120 mV were applied to homotypic wtCx40 and Cx40 H15Q + K16A gap junctions (Fig. 5A–D). The ensemble-averaged Ij of five Vj ramps from each experiment were normalized to the ±5 to ±20 mV slope conductance (gj,max) and the data from between six and eight experiments were pooled to construct the steady-state Gj–Vj curves. The Gj–Vj relationships were fit with the Boltzmann equation:
|
|
is the half-inactivation voltage. The Boltzmann distribution for the wCx40 Gj–Vj curve illustrated in Fig. 5 is provided in Table 3. The H15Q + K16A mutation exhibited biphasic gating characteristics, first decreasing and then increasing with increasing Vj of either polarity. This biphasic Gj–Vj relationship was fitted with two Boltzmann functions in series. The solutions to the biphasic Boltzmann curves are provided in Table 3. The V for the two gating phases of the Cx40 H15Q + K16A gap junction were near 40 and 80 mV, and the increase in Gj at higher voltages averaged about 50% of the initial decline in Gj. This complex Vj gating behaviour has not been previously described for homotypic gap junctions.
|
|
Sidedness of spermine block
It was assumed that spermine inhibits Cx40 Ij from the positive Vj side of the junction, consistent with the Vj gating polarity of Cx40 gap junctions and the valence of spermine. As a direct test of this assumption, 2 mM spermine was applied unilaterally to the wtCx40 or the Cx40 H15Q + K16A side of heterotypic Cx40 H15Q + K6A/wtCx40 gap junctions. If the assumption is true, more spermine block will be observed when 2 mM spermine is added to the wtCx40 side of the junction and Vj is positive. When Vj is negative, the first-order kinetics predict that spermine will dissociate from the gap junction. Figure 6A illustrates the findings that the block produced by 2 mM spermine added to the wtCx40 side of the heterotypic gap junction closely resembled the wtCx40 Ij–Vj relationship (compare to Fig. 2B). Conversely, when 2 mM spermine was added to the Cx40 H15Q + K16A side of the heterotypic gap junction, the Ij–Vj relationship closely resembled that of the Cx40 H15Q + K16A gap junction (compare to Fig. 2E). Thus it is concluded that spermine, like Vj-dependent gating, occludes the positive side of the Cx40 gap junction.
|
Coexpression of Cx40 H15Q + K16A gap junctions
The effect of the mutant Cx40 H15Q + K16A protein on the Vj-gating properties of wild-type Cx40 and Cx43 gap junctions was examined by transient coexpression of the Cx40 H15Q + K16A subunit in stable wtCx40 or wtCx43 transfected N2a cell clones. Coexpression of the Cx40 H15Q + K16A protein with its wild-type counterpart eliminated most of the Vj-dependent gating of the heteromeric gap junctions (Fig. 7A and Table 4). The most notable difference between the homomeric Cx40 H15Q + K16A and heteromeric Cx40 H15Q + K16A plus wtCx40 gap junctions was the disappearance of the slight increase in Gj at high Vj values. The coexpressed wtCx40 + Cx40 H15Q + K16A Gj–Vj relationship was fitted with a parallel double Boltzmann function of the form:
|
|
j) properties of wild type and now also the mutant H15Q + K16A Cx40 gap junctions (see Fig. 8).
|
|
|
We also tested the ability of the Cx40 H15Q + K16A protein to alter the Vj-gating and j properties of wtCx43 gap junctions. In contrast to the findings with wtCx40, the steady state Gj–Vj curve for the Cx40 H15Q + K16A and wtCx43 coexpressed gap junctions retained a majority of the Vj-dependent gating properties that could be ascribed to a wtCx43 Boltzmann function (Fig. 7E, Table 4). The Gj–Vj relationship for the wtCx43 + Cx40 H15Q + K16A coexpression experiments was also fitted with the parallel double connexin Boltzmann function:
|
|
was ±69.2 mV, and the average gating charge valence was ±2.95 elementary charges (q) for wtCx43. The thin continuous line in Fig. 7E depicts the solution to the parallel double connexin Boltzmann function where p1 was 0.72 or 0.69 for negative or positive Vj, respectively. Hence, this function estimates that the relative ratio of homotypic homomeric wtCx43 to Cx40 H15Q + K16A gap junctions was 70% to 30% in favour of the wtCx43 phenotype. Again, the correlation coefficients for the two fitted curves were similar (r= 0.97 or 0.96, respectively). Low gj recordings confirmed the presence of 100 and 150 pS gap junction channels (Fig. 7F), indicative of the coexistence of parallel homomeric homotypic wtCx43 and mutant Cx40 H15Q + K16A gap junctions. Channel properties of wtCx40 and Cx40 H15Q + K16A gap junctions
Previous results with the E9K and E13K mutations that altered the spermine blocking ability and Vj gating properties of Cx40 gap junctions also altered their j properties (Musa et al. 2004). However, the coexpression experiments seem to indicate that the H15Q + K16A mutations had no effect on the j properties of Cx40 gap junction channels. The j of homotypic wtCx40 and Cx40 H15Q + K16A j are illustrated in Fig. 8. Representative examples of the unitary channel activity, the all-points amplitude histograms for a single-channel recording, and the composite single-channel current–voltage (ij–Vj) relationships are shown for both the wtCx40 and Cx40 H15Q + K16A homotypic gap junction channels. The slope conductances confirm that j is not reduced in the homotypic Cx40 H15Q + K16A channel and may actually be slightly (6%) elevated. The heterotypic Cx40 H15Q + K16A/wtCx40 channel did not exhibit any noticeable rectification (Fig. 9).
|
j values of the homotypic wtCx40 and Cx40 H15Q + K16A gap junction channels, the reduction in j was similar for both channels when 140 mM KCl was replaced with 140 mM potassium gluconate. Figure 10 illustrates the j values obtained in the presence of 140 mM potassium gluconate. The slope conductance of the wtCx40 gap junction channel was reduced to 146 pS, while the j of the Cx40 H15Q + K16A gap junction channel was similarly reduced to 149 pS. Despite the complete replacement of 140 mM Cl– with gluconate, which is an organic anion with an aqueous mobility (25°C) of only 0.30 compared with 2.03 for Cl–, the j of both channels was reduced by only 4–7%. This is indicative of a relatively cation-selective channel. A simple approximation of the relative cation/anion conductance of a gap junction channel can be obtained by scaling the anionic terms in the Goldman-Hodgkin-Katz (GHK) current equation by a relative permeability factor (Rp) to obtain the same KCl/potassium gluconate j ratio as the equimolar anion substitution experiments (Veenstra et al. 1995).
|
|
|
seal formation (Fig. 11A–C) to compensate for the aqueous Cl– diffusion potential of the 50% KCl patch electrode or set equal to 0 mV (Fig. 11D–F). The pipette offsets were –2.3 ± 1.3 mV and +13.7 ± 0.7 mV for the 140 and 70 mM KCl patch electrodes relative to the bath Ag/AgCl2 reference cell containing 140 mM KCl IPS (n= 7). This interelectrode offset of 16 mV is close to the expected 17 mV for the calculated Cl– electrodiffusion potential. The ij–Vj curve in Fig. 11G indicates that the measured reversal potential (Erev) of the wtCx40 gap junction channel is not due to the nulling of the bath pipette offsets prior to the establishment of the double whole-cell recording configuration. The corresponding K+/Cl– relative permeability ratio (PK/PCl) for an Erev of –14.4 mV, adjusted to the measured K+ activities, according to the GHK voltage equation was 6.2. The previously reported value for PK/PCl was approximately 8.0 (Beblo & Veenstra, 1997).
|
The PK/PCl of the homotypic mutant Cx40 H15Q + K16A gap junction channel was also determined under asymmetric KCl conditions. The channel currents and ij–Vj curve in Fig. 12 clearly indicate that the Erev and PK/PCl values were indistinguishable from the wild type for this mutant Cx40 gap junction channel. Hence, despite the dramatic reductions in the ionic blocking ability of tetrapentylammonium and spermine molecules, the ionic conductance and permeability properties were not altered by the incorporation of the H15Q + K16A mutations.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
–helical turn away from the position 13 locus in the opposite direction from position 9 on the NT domains of these two connexins (Fig. 1).
The tandem H15Q + K16A mutation in Cx40 diminishes the concentration- and Vj-dependent inhibition of Cx40 Ij by spermine (Fig. 2). The Vj sensitivity of the spermine association rates and equilibrium dissociation constants were reduced by 50% and the spermine dissociation rates were essentially instantaneous upon Vj polarity reversal (Figs 3 and 4, Table 2). The Vj-dependent Kd values did not decrease below 2 mM, and there was only partial (
50%) Ij inhibition at the minimum Kd for the mutant Cx40 gap junction compared with 
80% inhibition at the minimum Kd of 200 µM for the wtCx40 gap junction. The Cx40 H15Q + K16A mutation also diminished the magnitude of the homotypic gap junction Vj-dependent gating, although it is not readily apparent if any reduction in gating charge valence resulted from this homotypic charge substitution (Fig. 5, Table 3). The Vj-dependent inhibition of Cx40 Ij by 5 mM tetrapentylammonium (TPeA) ions was further reduced by the H15Q + Q16A mutation than with the E9K and E13K mutations (Fig. 6; Musa et al. 2001, 2004). The heterotypic Cx40 H15Q + K16A/wtCx40 gap junction clearly indicates that there is charge interaction between the mutant and wild-type homotypic hemichannels. The Gmax of this heterotypic gap junction was shifted 25 mV negative by the neutralization of the H15 and K16 amino acids of wtCx40, and the gating charge valence was reduced by approximately 50% at positive Vj relative to the wtCx40 side of the gap junction. This heterotypic gap junction also retained the complex biphasic Vj-dependent gating observed with the Cx40 H15Q + K16A gap junction, although any alteration in the gating charge valence was obscured by the limited 25% reduction in Gj associated with the mutant hemichannel. A biphasic Gj–Vj curve with respect to a single Vj polarity was also observed in homotypic Cx45–Cx45 and heterotypic Cx45–Cx43EGFP gap junctions (Bukauskas et al. 2002). This phenomenon was attributed to a four-state contingent gating model consisting of two open-or-closed gates in series, one on each side of the channel. Our results with the Cx40 H15Q + K16A mutation are similar, but the magnitude of the increase in Gj at high Vj was greater than that observed for Cx45 despite a fourfold reduction in the overall magnitude of Vj gating. This biphasic gating is consistent with a relaxation of Vj-dependent closure as well as a contingent gating mechanism.
Taken together, these results suggest that the spermine block, TPeA block, and intrinsic Vj-dependent gating properties of the Cx40 gap junction share common structural determinants. This concept is further supported by the observation that the gating charge valence for both processes is approximately 3.2 equivalent charges (q) for the wtCx40 gap junction. Given that exogenously applied spermine serves as an inhibitory Vj-dependent ligand, these observations are consistent with the NT domain of Cx40 serving as a receptor for an intrinsic or extrinsic gating particle that inactivates the channel.
Because a homotypic gap junction is bilaterally symmetric, it cannot be definitively determined if the site of channel occlusion by spermine resides on the positive or negative Vj side of the channel. It has been postulated that Cx40 gap junctions close with positive Vj polarity (Valiunas et al. 2000). Since the mutant Cx40 H15Q + K16A gap junction exhibits only about 25% of the normal Cx40 Vj gating and is an order of magnitude less sensitive to spermine, the polarity of Cx40 channel closure could be tested using the heterotypic mutant Cx40 H15Q + K16A/wtCx40 hemichannel combination. When added unilaterally, spermine will enter the gap junction channel when Vj is positive on the same (cis) side. The electrical distance () measurement of 0.8–1.0 for the wtCx40 gap junction indicates that a single spermine molecule senses 80–100% of the Vj field during channel occlusion, not which side of the gap junction is occluded (Fig. 3, Table 2). The inhibition of Ij by unilateral 2 mM spermine resembled the block produced in wtCx40 or mutant Cx40 H15Q + K16A gap junctions depending on which cell contained spermine at positive Vj polarity (Fig. 6). We therefore conclude that Cx40 gap junctions close with positive Vj polarity whether it occurs by the endogenous fast Vj gate or exogenously applied spermine.
The Vj-dependent gating properties of the coexpressed Cx40 H15Q + K16A subunit with the wtCx40 or wtCx43 connexins suggest that this mutant Cx40 subunit readily heteromerizes with wtCx40, but not wtCx43 gap junctions (Fig. 7, Table 4). The Vj-dependent gating properties of the Cx40 H15Q + K16A–wtCx40 gap junctions phenotypically resembled the mutant homotypic gap junction except for the loss of the complex biphasic increase in Gj at higher Vj values. The Cx40 H15Q + K16A–wtCx43 gap junctions more closely resembled the Vj-dependent gating behaviour of wtCx43 gap junctions. This parallel homomeric expression of Cx40 H15Q + K16A and wtCx43 gap junction channels was confirmed by the concomitant observation of distinct 100 and 150 pS channels (Fig. 7F).
The Cx40 H15Q + K16A mutation is distinct from the previous E9K and E13K mutations in that the Cx40 gap junction channel properties are effectively unaltered. The unitary channel conductance (j) of the Cx40 H15Q + K16A gap junction channel was at most 10 pS (6%) higher than the wtCx40 channel (Fig. 8). The heterotypic Cx40H15Q + K16A/wtCx40 gap junction channel possesses the same j profile on the positive side of the junction as the respective wtCx40 and mutant Cx40 H15Q + K16A gap junction channels (Fig. 9). Equimolar substitution of potassium gluconate for KCl also produces similar percentage reductions in j, indicative of similar cation-to-anion conductance ratios of 4.5:1 or 4.3:1 (Fig. 10).
The selective permeability of the wtCx40 gap junction channel was previously postulated to be 8:1 in favour of cations (Beblo & Veenstra, 1997). This hypothesis was reexamined using the same experimental procedures this time with asymmetric KCl gradients across the homotypic wtCx40 gap junction (Veenstra, 2001b). The Erev measured –14.4 mV from the 70/140 mM KCl single channel ij–Vj relationship, which corresponds to a PK/PCl of 6.2 when adjusted to the measured K+ activities of 65 and 146 mM (Fig. 11). This PK/PCl ratio is slightly lower than the estimate of 8.0 previously reported, but the ionic composition of the IPS was different and the PK/PCl ratio was directly measured in these most recent experiments. The mutant Cx40 H15Q + K16A gap junction channel had an identical Erev measurement of –14.4 mV, and hence, the same selective ionic permeability as the wtCx40 channel (Fig. 12). The dependency of the PK/PCl ratio on the ionic strength of the low KCl IPS was essentially identical for both homotypic gap junction channels. The increase in the PK/PCl ratio from 6 to 10 when trans KCl was reduced to 25% of control is indicative of electronegative surface charge effects on the low ionic strength side of the gap junction channel.
A PK/PCl ratio of 10:1 was reported for Cx46 hemichannels, while the endogenous Xenopus Cx38 hemichannels have a PK/PCl ratio of about 4, indicative of modest cation selectivities for other connexin channels (Trexler et al. 1996; Zhang et al. 1998). We are presently re-examining the PK/PCl ratio of the wtCx43 gap junction channel in the same context as the present Cx40 experiments. These observations with the wtCx40 and mutant Cx40 H15Q + K16A gap junction channels confirm our previous findings about this connexin and suggest that the ion permeation pathway is not significantly affected by the H15Q + K16A mutation, despite significant disruption of the Vj-dependent gating and ionic blocking properties. The S26L mutation slightly altered the Cx32 Vj-dependent gating without altering the ionic conductance (in CsCl) or selectivity of this mutant homotypic gap junction channel, but did cause an apparent decrease in the solute accessible pore size (Oh et al. 1997). In contrast, the increased on- and off-rates (Figs 2A and 4C–D) are consistent with an increased spermine accessibility of the Cx40 H15Q + K16A gap junction channel. It remains to be determined if the reduction in open channel block produced by the Cx40 H15Q + K16A mutation has rendered the channel permeable to spermine.
The connexin NT domain was postulated to form a domain-hinge-domain motif, thereby permitting the NT end of the protein to bend inward and forming the ‘voltage sensor’ of the channel that lies within the Vj field (Purnick et al. 2000a). The consensus secondary structure predictions from the Network Protein Sequence Analysis website predicts that the Cx40 and Cx43 NT domains form a –helix through positions 16 while Cx26 and Cx32 do not. In agreement with the Purnick et al. (2000a) model, the NT –helix is predicted to end at the GG or SG hinge at positions 11 and 12 for Cx26 and Cx32 (Fig. 1). It is feasible that this is why charged point mutations in Cx26 and Cx32 through position 10 altered the Vj gating of these two ß-group connexins, although G15R is now also implicated in this process (Purnick et al. 2000b; Abrams et al. 2001).
How the position 15–17 HKH locus contributes to the gating structure of the Cx40 gap junction channel is unknown. The H15 + K16 locus, by virtue of its lack of effect on Cx40 j and ionic selectivity, presumably resides outside the channel pore. This locus also contributes partially to the Vj-dependent gating structure common to the spermine and TPeA blocking sites. The spermine-blocking site also includes the E9 and E13 sites presumed to reside on the NT –helix and inside the channel pore of Cx40. The E9K mutation reduced the Cx40 j by one-third and the Vj-dependent gating was preserved, albeit with a reduction in the gating charge valence of 1 q and the Gmin to zero (Musa et al. 2004). The effects of the E9K mutation are consistent with a loss of electronegative surface charge within the channel and Vj-gating structure. The zero Gmin also indicates that the pore can become completely occluded by the intrinsic Vj gate despite the dramatic reduction in spermine block. The E13K mutation resulted in a flickery Vj-independent, and spermine-insensitive, channel, suggesting that this residue plays a critical role in the channel architecture in addition to the Vj-dependent gating-like processes already implicated in pore occlusion (Musa et al. 2001, 2004). Hence, despite sharing common structural determinants, the intrinsic Vj-dependent gating, spermine, and TPeA blocking mechanisms are not identical. We speculate that the NT domain of the –group connexins forms a -helix that terminates prior to the highly conserved extended ß-strand (HorY)STxxG(KorR) motif that links the connexin NT pore domain with the first transmembrane (M1) domain. Additional experimentation and structural data about the NT domain of the –group connexins are needed to test this structural hypothesis.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Beblo DA & Veenstra RD (1997). Monovalent cation permeation through the connexin40 gap junction channel. Cs, Rb, K, Na, Li, TEA, TMA, TBA, and effects of anions Br, Cl, F, acetate, aspartate, glutamate, and NO3. J Gen Physiol 109, 509–522.
Bukauskas FF, Bukauskiene A, Verselis VK & Bennett MVL (2002). Coupling asymmetry of heterotypic connexin45/connexin43-EGFP gap junctions: Properties of fast and slow gating mechanisms. Proc Natl Acad Sci U S A 99, 7113–7118.
Combet C, Blanchet C, Geourjon C & Deléage G (2000). Macromolecular interactions: tracing the roots. Trends Biochem Sci 25, 147–150.[CrossRef][Medline]
Harris AL (2001). Emerging issues of connexin channels: Biophysics fills the gap. Q Rev Biophys 34, 325–472.[Medline]
Lampe PD & Lau AF (2004). The effects of connexin phosphorylation on gap junction communication. Int J Biochem Cell Biol 36, 1171–1186.[CrossRef][Medline]
Lin X, Gemel J, Beyer EC & Veenstra RD (2005). A dynamic model for ventricular junctional conductance during the cardiac action potential. Am J Physiol Heart Circ Physiol 288, H113–H1123.
Musa H, Fenn E, Crye M, Gemel J, Beyer EC & Veenstra RD (2004). Alternate amino terminal glutamate or lysine residues of rat connexin-40 and -43 affect spermine block and voltage gating. J Physiol 557, 863–878.
Musa H, Gough JD, Lees WJ & Veenstra RD (2001). Ionic blockade of the rat connexin40 gap junction channel by large tetraalkylammonium ions. Biophys J 81, 3253–3274.
Musa H & Veenstra RD (2003). Voltage-dependent blockade of connexin40 gap junctions by spermine. Biophys J 84, 205–219.
Oh S, Ri Y, Bennett MVL, Trexler EB, Verselis VK & Bargiello TA (1997). Changes in permeability caused by connexin 32 mutations underlie X-linked Charcot-Marie-Tooth Disease. Neuron 19, 927–938.[CrossRef][Medline]
Oh S, Rubin JB, Bennett MVL, Verselis VK & Bargiello TA (1999). Molecular determinants of electrical rectification of single channel conductance in gap junctions formed by connexins 26 and 32. J Gen Physiol 114, 339–364.
Purnick PEM, Benjamin DC, Verselis VK, Bargiello TA & Dowd TL (2000a). Structure of the amino terminus of a gap junction protein. Arch Biochem Biophys 381, 181–190.[CrossRef][Medline]
Purnick PE, Oh S, Abrams CK, Verselis VK & Bargiello TA (2000b). Reversal of the gating polarity of gap junctions by negative charge substitutions in the N-terminus of connexin 32. Biophys J 79, 2403–2415.
Sáez JC, Berthoud VM, Brañes MC, Martinez AD & Beyer EC (2003). Plasma membrane channels formed by connexins: their regulation and functions. Physiol Rev 83, 1359–1400.
Tong J-J, Liu X, Dong L & Ebihara L (2004). Exchange of gating properties between rat Cx46 and chicken Cx45.6. Biophys J 87, 2397–2406.
Trexler EB, Bennett MVL, Bargiello TA & Verselis VK (1996). Voltage-gating and permeation in a gap junction hemichannel. Proc Natl Acad Sci U S A 93, 5836–5841.
Valiunas V, Weingart R & Brink PR (2000). Formation of heterotypic gap junction channels by connexins 40 and 43. Circ Res 86, e42–e49.
Veenstra RD (1996). Size and selectivity of gap junction channels formed from different connexins. J Bioenerg Biomembr 28, 327–337.[CrossRef][Medline]
Veenstra RD (2001a). Voltage clamp limitations of dual whole-cell gap junction current and voltage recordings. I. Conductance measurements. Biophys J 80, 2231–2247.
Veenstra RD (2001b). Determining ionic permeabilities of gap junction channels. In Methods in Molecular Biology, Connexin Channels
I2 measured from baseline Vj= 0 mV steps) was reduced in a time-dependent manner only when a positive transjunctional voltage (Vj) pulse (middle pulse) was applied to cell 1. Each current trace represents a ±10 mV increase in Vj. B, the normalized steady-state Ij–Vj curves for representative [spermine]. Each data point represents the mean value for between three and seven experiments. C, the Vj-dependent dose–response curves for inhibition of wild-type Cx40 Ij by spermine calculated from the steady state Ij–Vj relationships. Double whole-cell current recordings (D), normalized steady-state junctional current–voltage relationships (E), and Vj-dependent spermine dose–response curves (F) for homotypic Cx40 H15Q + K16A gap junctions in 140 mM KCl. Each data point in E and F represents the mean value for between three and seven experiments. The Kd values are listed in 
, n= 3) or mutant Cx40 H15Q + K16A (
, n= 4) side of heterotypic wtCx40/Cx40 H15Q + K16A gap junctions demonstrates that occlusion occurs on the positive Vj side of the Cx40 gap junction channel. B, the normalized steady-state Ij–Vj relationships in the presence of unilateral 5 mM tetrapentylammonium for wtCx40 (continuous line, modelled from 
