INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE MURINE WEAVER DISEASE is caused by the mutation of a single amino acid (G156S) in the putative pore region of the inwardly rectifying K⁺ channel, GIRK2. Previous studies have shown that recombinant GIRK2wv expressed heterologously in Xenopus oocytes and mammalian cells, as well as in cerebellar granule cells from wv mice, formed homomultimeric channels characterized by 1) G protein insensitivity, 2) cation nonselectivity, and 3) sensitivity to QX314, MK-801, and verapamil inhibition.

Indirect evidence from a number of studies supports the hypothesis that the mutant weaver channel might render neurons leaky to Ca²⁺, eventually resulting in cell death. A recent study demonstrated that intracellular Ca²⁺ was elevated in the primary cerebellar neuronal cultures from heterozyogous (wv/+) animals relative to wild type (4). Silverman and colleagues (14) have reported an apparent permeability of GIRK2wv channels to Ca²⁺ over wild-type GIRK channels when expressed in Xenopus oocytes. In the Silverman et al. study, the Ca²⁺ permeability of the channel was inferred from the activation of the endogenous Ca²⁺-activated Cl conductance, which was seen only in GIRK2wv-expressing oocytes (14). In parallel studies of recombinant GIRK2wv channels in oocytes, removal of Ca²⁺ from the incubation medium was shown to significantly enhance oocyte survival (18). Taken together, these observations suggested that GIRK2wv channels are permeable to monovalent as well as divalent cations.

In this study, we directly tested whether homomeric GIRK2wv channels are permeable to Ca²⁺ in both Xenopus oocytes and mammalian cells transiently expressing GIRK2wv channels in culture. Homomeric channels were weakly permeable to Ca²⁺ in the Xenopus oocytes. In contrast, the divalent cation permeability was absent in the GIRK2wv-expressing mammalian cells, suggesting that susceptibility to cell death in GIRK2wv-expressing neurons may simply be due to Ca²⁺ influx through parallel, voltage-dependent channels following prolonged depolarization.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

cDNA clone. GIRK1 was cloned from a RIN cell library and had a predicted amino acid sequence identical to the cardiac clone originally described (8). GIRK2 and GIRK2wv were gifts from Dr. P. Kofuji (California Institute of Technology, CA). The m2 muscarinic receptor was purchased from Clontech (Clontech, CA) in the pGEM3Z vector. All GIRK constructs were subcloned into either the pMXT vector, obtained from Dr. P. Kofuji, for oocyte expression or pEGFPN3 (Clontech, CA) for mammalian expression. The m2 receptor was linearized with Hind III, and cRNA was transcribed using the T7 polymerase mMessage mMachine kit (Ambion, Austin, TX). All GIRK constructs were linearized with Sal I, and cRNA was transcribed using T3 polymerase mMessage mMachine kit (Ambion). The cRNA concentration was determined by ultraviolet light absorption at 260 nm (A₂₆₀) and confirmed by intensity on ethidium bromide stained agarose gels. For mammalian cell expression, GIRK1, GIRK2, and GIRK2wv were fused to enhanced green fluorescent protein (EGFP) at the carboxy terminus in the pEGFPN3 vector (Clontech).

Oocyte electrophysiology. Oocytes were injected with 2 ng of m2 muscarinic receptor cRNA and 5 ng of each GIRK subunit cRNA along with 12.3 ng GIRK5 (KHA1) antisense cRNA. Injected oocytes were maintained in OR-2+ solution containing (in mM) 96 NaCl, 2.5 KCl, 1 CaCl₂, 1 MgCl₂, 5 HEPES, 2.5 sodium pyruvate, and 50 µg/ml gentamicin. Nominally Ca²⁺-free solutions were used to incubate and maintain oocytes expressing GIRK2wv.

Mammalian cell culture and cDNA transfection. COS-7 cells (American Type Culture Collection) were plated in 35-mm dishes and grown in DMEM supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 100 µg/ml streptomycin. Transfection of COS-7 cells was carried out using SuperFect Transfection Reagent (10 µl per dish; QIAGEN, Valencia, CA) mixed with the following amounts of cDNA: 0.5 µg m2R with 1 µg GIRK1-EGFP and GIRK2 or GIRK2wv-EGFP in 100 µl serum-free media (Opti-MEM1; GIBCO, Grand Island, NY). Cells were exposed to the DNA-containing solution for 10 min at room temperature, followed by the addition of 600 µl of serum-containing cell culture media. Cells were then incubated for 2 h, washed once, and incubated at 37°C, 5% CO₂. Electrophysiological recordings from COS-7 cells were made 48 h from transfection initiation on green fluorescent protein positive cells.

Electrophysiological recording from mammalian cells. Whole cell recordings were performed at room temperature 48-72 h posttransfection in an initial bath solution consisting of (in mM) 140 NaCl, 5.4 KCl, 2 CaCl₂, 1 MgCl₂, and 10 HEPES, pH 7.4. After currents in low-K⁺ solutions were recorded, cells were superfused with either a high-K⁺ solution containing (in mM) 140 KCl, 1 MgCl₂, and 10 HEPES, pH 7.4, or a high-Na⁺ solution containing (in mM) 145 NaCl, 1 MgCl₂, and 10 HEPES, pH 7.4. In experiments designed to quantitate the Ca²⁺ permeability, cells were superfused with 5 or 70 mM Ca²⁺ bath solution containing (in mM) 5 CaCl₂ with 140 NMDG-Cl, or 70 CaCl₂ with 35 NMDG-Cl, 1 MgCl₂, 10 HEPES, pH 7.4. Solution osmolarity was kept constant at 270-290 mosM using a vapor pressure osmometer (Wescor, Logan, UT). Patch pipettes were pulled from microhematocrit capillary tubes (Fisherbrand, Fisher, Pittsburgh, PA) to give resistances of 3-6 M. The pipette recording solution contained (in mM) 120 KCl, 2 CaCl₂, 1 MgCl₂, 11 EGTA, 33 KOH, 10 HEPES, 1 NaGTP, 2 MgATP, pH 7.2, or 140 NMDG, 0.2 CaCl₂, 1 MgCl₂, 1 EGTA, 10 HEPES, 1 NaGTP, 2 MgATP, pH 7.2 when Ca²⁺ currents were measured. Whole cell currents were recorded with an EPC-7 (List Electronics, Lambrecht, Germany) patch-clamp amplifier at 2 kHz and low-pass filtered at 1 kHz. Stimulation and data acquisition were controlled by the PULSE software package on a Power Macintosh computer, and data analysis was performed with IGOR software (WaveMetrics, Lake Oswego, OR).

Digital fluorescent imaging. COS-7 cells were grown on 25-mm coverslips to 50% confluence, loaded with 2 µM fura 2-AM (Molecular Probes) in serum-free MEM for 1 h, then equilibrated in Hanks' balanced saline solution (HBSS; GIBCO, Grand Island, NY) for 50 min. Fura 2 loading and equilibration were carried out at 37°C. The coverslip with loaded cells was moved into a Teflon coverslip recording dish (model LU-CSD; Medical Systems, NY) and mounted on the microscope stage. Cells were continuously superfused throughout the experiment with HBSS or switched to a 70 mM Ca²⁺ solution containing (in mM) 70 CaCl₂ with 35 NMDG-Cl, 1 MgCl₂, and 10 HEPES, pH 7.4. Internal Ca²⁺ (Ca_i) was determined using digital fluorescent imaging of cellular fura 2 epiflorescence. Emission was determined at 510 nm following excitation at 340 and 380 nm. Images were obtained every 20 s, and 64 frames were averaged at each excitation wavelength. Background was obtained using an area of the coverslip devoid of cells and subtracted from each excitation wavelength image. After background subtraction, the 340-nm image was divided by the 380-nm image to provide a ratio (R) image. Image analysis was carried out using the ImageMaster Ratio Fluorescence Imaging Software (Photon Technology International, NJ).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Comparison of GIRK1/2 and GIRK2wv channels expressed in Xenopus oocytes and COS-7 cells. The goal of this study was to quantitate the magnitude of the Ca²⁺ permeability of recombinant GIRK2wv channels compared with heteromultimeric GIRK1 + GIRK2 (GIRK1/2) channels. We compared expression of the wild-type and mutant channels in Xenopus oocytes to that in mammalian COS-7 cells.

View larger version (24K): [in this window] [in a new window]   Fig. 1.   Coexpression of GIRK1 with GIRK2 in Xenopus oocytes leads to larger G protein-independent currents than coexpression in mammalian cells. A: 2 microelectrode current recordings from an oocyte injected with cRNA for m2 muscarinic receptors, GIRK1 and GIRK2. A, left: 3 families of superimposed current traces elicited by 1-s voltage steps to potentials between 150 and 50 mV from a holding potential of 80 mV. Dotted lines, zero current level. First set of currents was recorded from an oocyte in a solution in which all the NaCl was replaced with KCl. Second set of currents represents carbachol (Carb)-induced currents obtained by subtracting currents recorded in high-K+ solution from those recorded in a high K+ solution containing in addition 5 µM carbachol. Third set of currents represents residual current remaining after adding 500 µM Ba2+ to high-K+ carbachol-containing solution. All 3 families of current were obtained from the same oocyte. Corresponding current-voltage (I-V) relations are plotted to right of current traces. Current magnitude was determined at the point in the current traces where the currents were maximum. B: whole cell patch voltage-clamp recordings from a single COS-7 cell transiently transfected with the GIRK1-EGFP and GIRK2 cDNA along with m2 receptor cDNA as described in MATERIALS AND METHODS. Cells chosen for electrophysiological recording were identified by their green fluorescence. B, left: 3 families of superimposed current traces elicited by 200-ms voltage steps to potentials between 160 and 50 mV from a holding potential of 80 mV. As in A, the dotted lines indicate zero current level. First set of currents was recorded in a solution in which all the external NaCl was isosmotically replaced with KCl. Second set of currents represents difference currents obtained by subtracting currents in high-K+ solution from those obtained in the identical solution containing in addition 5 µM carbachol. Third set of current traces was obtained when 200 µM Ba2+ was added to high-K+ carbachol-containing solution. Associated I-V relationships are plotted to right of current traces. Note that, compared with the current GIRK1/2 currents recorded in the oocyte in high-K+ solutions, currents recorded from the mammalian cell preparations failed to show a significant expression of G protein-independent (basal) current activation.

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Comparison of GIRK2wv expression in Xenopus oocytes and mammalian cells. A: currents were recorded from oocytes as in Fig. 1A. Three families of current traces shown at left were recorded in solutions containing high K+, high Na+, high Na+ containing 300 µM QX314, and a solution in which all the monovalent cations had been replaced with N-methyl-D-glucamine (NMDG). Associated I-V relations are plotted on the right. B: currents recorded from a single transfected mammalian cell as in Fig. 1B. External solutions are given above each family of current traces. Associated I-V relations are plotted to the right.

Divalent permeability of uninjected oocytes. To quantitate the magnitude of the GIRK2wv-induced Ca²⁺ influx pathway, it was necessary to identify basal divalent permeability through endogenous, voltage-gated Ca²⁺ channels in uninjected oocytes exposed to solutions containing elevated divalent concentrations. Characterization of endogenous voltage-activated Ca²⁺ channels in oocytes has been previously established (2, 9, 12). A comparative summary of the magnitude of both inward and outward current at 150 and 50 mV, in high and low Ca²⁺ solutions is plotted in Fig. 3 as a function of internal 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA) buffering. Charnet and colleagues (1) have previously reported the use of BAPTA injections to buffer the influx of Ca²⁺ in oocytes, thereby preventing activation of the endogenous Ca²⁺-activated anion conductance.

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Summary of inward and outward current amplitudes in uninjected oocytes as a function of increasing external divalent concentrations in the presence and absence of internal BAPTA buffering. All experiments were obtained from uninjected oocytes using voltage protocols described in Fig. 1. Oocytes were injected with high concentrations of BAPTA as described in MATERIALS AND METHODS. A: representative currents and I-V relationship from a single oocyte in solutions containing 5 and 70 mM Ca2+ as the only permeant species. B: currents from an oocyte injected with BAPTA in the presence of the low and high Ca2+- containing solutions. C: comparison of current amplitude for two Ca2+-containing solutions in the presence and absence of internal Ca2+ buffering. Peak current amplitude was determined at 50 and 150 mV. In all experiments, Ca2+ was the only permeant cation in the external solution. Isosmolarity was obtained by addition of sucrose to the low-Ca2+ solution.

Direct measurement of current carried by Ca²⁺ in oocytes and mammalian cells expressing GIRK1/2 and GIRK2wv. To determine the magnitude of divalent current carried by GIRK2wv channels compared with GIRK1/2 channels, experiments on both oocytes and COS-7 cells were performed in solutions in which Ca²⁺ was the only permeant cation in the extracellular solution. Oocytes were injected with 100 mM BAPTA prior to current recording in high divalent solutions to block activation of the contaminating Ca²⁺-activated Cl current. Figure 4 illustrates data obtained from both oocytes and COS-7 cells expressing GIRK1/2. Both basal and carbachol-induced K⁺ currents were recorded to ensure that cells were expressing G protein-activated GIRK channels. Sequential exposure of COS-7 cells to 5 and 70 mM Ca²⁺-containing solutions failed to result in current activation. Current amplitudes at 150 mV for oocytes showed no significant change on switching from low to high external Ca²⁺.

View larger version (31K): [in this window] [in a new window]   Fig. 4.   Elevation in extracellular Ca2+ fails to result in an increase in inward current or change in reversal potential in GIRK1/2-expressing oocytes as well as mammalian cells. A: 2 microelectrode voltage-clamp experiments were carried out on GIRK1/2-expressing oocytes as in Fig. 1A. Left: G protein independent currents and carbachol-induced currents in high-K+ solutions and currents obtained in solutions in which Ca2+ was the only permeant species. Average I-V relationships are given to the right of the current traces for 7 cells in which similar experiments were carried out. Note that there was no significant increase in current in changing from a low to high Ca2+-containing solution as seen in the average I-V relationship. B: identical experiments carried out in transiently transfected COS-7 cells. Average I-V relationships were obtained in a total of 5 cells. Note the absence of a G protein independent current or current increases on exchanging to extracellular solutions containing high Ca2+.

View larger version (19K): [in this window] [in a new window]   Fig. 5.   GIRK2wv expression in oocytes but not mammalian cells gives rise to a significant inward current in 5 mM Ca2+-containing solutions over that seen in GIRK1/2-expressing cells. A: experiments were carried out in divalent solutions using voltage protocols as in Fig. 1. Current traces are from representative oocytes expressing GIRK1/2 and GIRK2wv. Amplitude of the inward current at 150 mV for uninjected, GIRK1/2, and GIRK2wv-expressing oocytes is compared on right. Divalent inward current was significantly greater in the GIRK2wv-expressing oocytes over that observed for the GIRK1/2-expressing oocytes (P < 0.001) as well as for the uninjected oocytes (P < 0.0001). Level of significance between mean currents in 3 populations of oocytes was determined using General Linear Models (SAS, Carey, NC) with a Tukey correction for analysis of unequal cell sizes. B: currents from representative COS-7 cells transiently transfected with GIRK1-EGFP + GIRK2 and GIRK2wv-EGFP fusion protein. There was no significant increase in inward current in 2 mM Ca2+-containing solutions for the GIRK2wv cells over that seen for the GIRK1/2-expressing cells as summarized in the bar graph to the right of the current traces.

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Comparison of representative currents recorded from oocytes and mammalian cells expressing GIRK1/2 and GIRK2wv in 70 mM Ca2+-containing solutions. A: representative families of current traces from 2 different oocytes expressing either GIRK1+2 or GIRK2wv in external solutions containing 70 mM external Ca2+. Summary of maximum inward current at 150 mV is given to right of the current traces. Inward current in the GIRK2wv-expressing oocytes was significantly increased (P < 0.001) over that observed for the GIRK1/2-expressing oocytes. Statistical analysis was performed as in Fig. 5, comparing mean inward current data between uninjected, GIRK1 + GIRK2, and GIRK2wv-expressing oocytes. Results of statistical analysis demonstrated that in the oocytes the difference between the mean current amplitude was significantly different as a function of channel type. There was no statistical difference in mean current between channel types as a function of external Ca2+ concentration. B: families of current traces from 2 representative COS-7 cells expressing GIRK1-EGFP + GIRK2 and GIRK2wv-EGFP. Summary of maximum inward current at 160 mV is given to right of current traces. Note there was no significant difference in current amplitude in the high external Ca2+ solutions between the cells expressing mutant and wild-type channels.

Intracellular calcium measurements. To further investigate whether constitutively active GIRK2wv channels expressed in mammalian cells could give rise to significant changes in levels of intracellular calcium (Ca_i), as has been reported for neurons cultured from weaver mice (4, 21), we carried out digital fluorescent imaging experiments in COS-7 cells transfected with GIRK1/2 or GIRK2wv where GIRK1 and GIRK2wv were tagged with EGFP. Fura 2 was used to detect resting Ca_i in nontransfected COS-7 cells, COS-7 cells cotransfected with GIRK1-EGFP, and GIRK2 or COS-7 cells transfected with GIRK2wv-EGFP. Cells were loaded with fura 2-AM for 1 h before digital fluorescent imaging experiments. The resting 340/380 ratios (R) among the three groups were indistinguishable. These data are summarized in Fig. 7. The dynamic range of the cellular response to changes in Ca_i was determined in experiments in which cells were sequentially exposed to a solution containing 10 µM ionomycin in the presence of 2 mM Ca²⁺, followed by a solution change to one in which the free Ca²⁺ concentration was buffered to zero in the presence of 1 mM EGTA as seen in Fig. 7A. The mean resting 340/380 nm fluorescence intensity ratio (R_340/380) values were 0.6 ± 0.02 (n = 43) for nontransfected cells, 0.55 ± 0.09 (n = 32) for GIRK1/GIRK2-expressing cells, and 0.57 ± 0.03 (n = 34) for the GIRK2wv-expressing cells (Fig. 7B). We were unable to detect a change in the R_340/380 values on changing from low (2 mM) to high (70 mM) external Ca²⁺ in either the GIRK1/2- or GIRK2wv-transfected cells. The average of the change in the R_340/380 in individual cells on increasing extracellular Ca²⁺ from 2 to 70 mM was 0.28 ± 0.02 (n = 5 ) for GIRK1/2 and 0.21 ± 0.03 (n = 9) for GIRK2wv as summarized in Fig. 7C.

View larger version (15K): [in this window] [in a new window]   Fig. 7.   Comparison of resting Ca2+ levels in GIRK1 + GIRK2 vs. GIRK2wv-expressing COS-7 cells. Relative levels of intracellular Ca2+ were determined as the 340/380 nm fluorescence intensity ratio (R340/380) in fura 2 loaded COS-7 cells expressing either GIRK1 + GIRK2 or GIRK2wv using digital imaging techniques. A: resting levels of R340/380 in 13 cells before and after exposure to external solutions containing 10 µM ionomycin. Ionomycin-containing solution was added at arrow. When the R340/380 appeared to reach a maximum value, the external solution was changed to one containing 0 mM Ca2+ and 1 mM EGTA. B: plot of the average resting R340/380 in nontransfected cells, cells expressing GIRK1 + GIRK2, and cells expressing GIRK2wv in an external solution containing 2 mM Ca2+. Average maximal response to ionomycin in the GIRK2wv cells is shown in bar at right. C: comparison of the difference in resting R340/380 for cells in external solutions containing 70 mM Ca2+ minus that obtained in 2 mM Ca2+. Cells were exposed to the Ca2+-containing solutions and the differences in the response in the individual cells averaged. There was no significant difference in the response between the GIRK1 + GIRK2 expressing cells over that obtained for the GIRK2wv-expressing cells.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Expression and activation of GIRK2wv in oocytes has been associated with a large increase in the endogenous Ca²⁺-activated Cl current, indicative of a large divalent influx in Ca²⁺-containing solutions (14). This observation, along with the increased vulnerability of oocytes to cell death in Ca²⁺-containing solutions (18), has suggested that the GIRK2wv channel might allow for a significant Ca²⁺ "leak." These observations prompted our investigation of the magnitude of the GIRK2wv divalent permeability in heterologous expression systems.

In this study, we have compared expression of GIRK1/2 and GIRK2wv channels in both oocytes and mammalian cells. We compared the divalent permeability of the wild-type to the mutant GIRK2wv channels and found that oocytes expressing the mutant channels show 1) an elevation of inward current over those expressing the wild-type GIRK1/2 and 2) a shift in reversal potential on switching from low- to high-Ca²⁺-containing external solutions. These results were obtained in the presence of high concentrations of internal BAPTA to inhibit activation of contaminating Ca²⁺-activated Cl currents. Similar experiments in transfected COS-7 cells failed to demonstrate either a significant increase in Ca²⁺ current through the mutant channels over that observed in GIRK1/2-expressing cells or a shift in reversal potential on increases in extracellular Ca²⁺. We were also unable to observe any significant increase in the resting levels of Ca_i in COS-7 cells expressing GIRK2wv over that observed in nontransfected cells or cells expressing GIRK1/2. Taken together, these data indicate that the divalent cation properties of the expressed GIRK2wv conductance differ between the two heterologous expression systems. Most importantly, however, the data strongly suggest that, if the mammalian cells are a better model for channel expression in mouse neurons than the oocytes, the elevated Ca_i levels observed in weaver neurons (4, 21) may be a result of chronic depolarization and not Ca²⁺ influx through the GIRK2wv channels themselves. Evidence in support of this hypothesis comes from the studies of Liesi and Wright (10) who demonstrated that Ca²⁺ channel function is essential in mediating the weaver gene effect. The rescue effect of weaver granule cell neurons at high but not low concentrations of MK-801 in the studies of Liesi and Wright (10) was consistent with the reported inhibitory effect of verapamil and high concentrations of MK-801 on voltage-gated Ca²⁺ channels; low concentrations of MK-801 (1 µM) had no rescue effect (10).

Determination of the Ca²⁺ permeability of the homomultimeric GIRK2wv channel is more direct in mammalian cells lacking the contaminating endogenous Ca²⁺-activated anion conductance. In our experiments, we found no evidence for Ca²⁺ flux through GIRK2wv channels monitored as either an increase in inward current in high divalent solution or in measurements of changes in resting Ca_i as monitored by ratiometric intracellular fura 2 fluorescence. Consistent with the previous anecdotal observation made by Navarro and co-workers (13), we were unable to observe an increase in inward current in Chinese hamster ovary (data not shown) as well as COS cells transfected with the GIRK2wv gene in 70 mM Ca²⁺ containing external solutions.

It is tempting to generalize that the weaver mutation in the signature sequence of all K⁺ selective channels would produce a similar loss in K⁺ selectivity in the outwardly as well as the inwardly rectifying K⁺ conductances. Interestingly, this same mutation has been found in a member of the six-transmembrane family of K⁺ channels, which points to the contrary. The K⁺ channel, KCNQ4, localizes its expression to cochlear outer hair cells and maps to the DFNA2 locus for a form of nonsyndromic dominant deafness (7). A mutation in this gene in the DNFA2 pedigree exchanges the G for an S (G285S) in the GYG sequence in the pore of that channel, identical to the mutation in GIRK2wv. The G285S mutation in KCNQ4 exerts a strong dominant negative effect on wild-type KCNQ4, and its loss leads to slow cellular degeneration (7), although the precise pathogenesis is unknown. KCNQ4 codes for a six-transmembrane domain K⁺ channel subunit protein that is assumed to form a functional heterotetramer with other members of the KCNQ family. Unlike GIRK2wv, which has an equivalent mutation in the signature sequence, the mutation G285S in KCNQ4 does not appear to form functional homomultimers as does GIRK2wv. Coexpression studies with the mutant KCNQ4 G285S and other members of the KCNQ family carried out to date show that coexpression of the mutant subunit reduces current expression by ~90%. The remaining current is K⁺ selective over Na⁺ or Ca²⁺ (7), unlike the selectivity profile of the mutant GIRK2wv. Thus similar pore mutations in the outward and inwardly rectifying K⁺ channel families would appear to have significantly different functional phenotypes with respect to changes in channel selectivity and ability to form functional homo- and heteromultimers (6, 16). The two transmembrane domain GIRK subunits appear to tolerate changes in pore-forming residues allowing for the formation of hetero- as well as homomultimeric channels. The six-transmembrane domain K⁺ channels appear to require a more rigid scaffolding intolerant of similar changes in pore-forming residues.

In addition to the observed differences in selectivity between GIRK1/GIRK2 heteromultimers and GIRK2wv homomultimers, we observed a consistent difference in the kinetics of current activation for the two channels at the most hyperpolarized potentials when expressed in Xenopus oocytes. Current activation for GIRK2wv expressed in oocytes was instantaneous, whereas the kinetics of activation for GIRK1/GIRK2 were much slower. Similar differences in the time course of current activation on hyperpolarization between recombinant GIRK1/GIRK2 and GIRK2wv channels have been observed by Slesinger et al. (16). Differences in time course of current activation between the wild-type GIRK1/GIRK2 channels and the mutant GIRK2wv channels were not observed when the recombinant channels were expressed in the mammalian cell background. These differences in activation kinetics, which we observed exclusively in the oocyte expression experiments, are consistent with the weak inward rectification of the currents also observed for the mutant GIRK2wv channel in the oocyte system in our studies as well as those of Kofuji et al. (6). Rectification in the inwardly rectifying K⁺ channels is a result of Mg²⁺ and/or polyamine binding to an intracellular site, thereby blocking monovalent cation permeation in the outward direction (3, 11, 17, 19). It may be that the reduced rectification seen for GIRK2wv when expressed in oocytes may be due to weak binding and/or permeation of a class of cytoplasmic polyamines not present in the mammalian cells.

In summary, results for our investigation indicate that the modest Ca²⁺ influx through GIRK2wv homomeric channels expressed in oocytes differs from that observed for channels expressed in mammalian cells and may represent the formation of a functional channel arising from coassembly with an unidentified endogenous subunit of the oocyte. Coassembly with the endogenous Xenopus oocyte subunit GIRK5 is unlikely, in that our experiments were conducted using antisense against GIRK5, which would have prevented its expression. Recombinant GIRK2wv channel expression in mammalian cells was not associated with either an observable Ca²⁺ permeation through the conductance nor an increase in intracellular Ca_i over that observed in nontransfected cells. Our data suggest that the elevation in Ca_i associated with neuronal cell death in murine cells expressing the gene may not be due to divalent permeation through the GIRK2wv homomeric channels but may be due instead to toxicity induced through chronic depolarization, allowing for Ca²⁺ influx through voltage-dependent Ca²⁺-permeable pathways.