Delayed Rectifier K⁺ Currents,I_K, Are Encoded by Kv2 α-Subunits and Regulate Tonic Firing in Mammalian Sympathetic Neurons

Generation and characterization of dominant negative Kv2 α-subunits

In preliminary studies, a single point mutation was introduced into the pore region of Kv2.1, analogous to the approaches used to generate the dominant negative Kv4.2 α-subunit, Kv4.2W362F (Barry et al., 1998), and the dominant negative Kv1.5 α-subunit, Kv1.5W461F (Li et al., 1999). However, heterologous expression studies revealed that this construct produced outward K⁺currents indistinguishable from those evident on expression of wild-type (rat) Kv2.1. These results are similar to those described byBlaine and Ribera (1998) in studies with Xenopus Kv2.2. Because a functional dominant negative construct was obtained by introducing two mutations in the pore region of XenopusKv2.2 (Blaine and Ribera, 1998), a similar strategy was used here for (rat) Kv2.1 to generate a double pore mutant, Kv2.1W365C/Y380T (see Materials and Methods). The Kv2.1W365C/Y380T construct, Kv2.1DN, was epitope-tagged at the C terminus with the 8 aa FLAG tag to allow direct detection of transgene expression.

To determine the functional properties of this construct, HEK-293 cells were transfected with Kv2.1DN (and EGFP) alone and in combination with Kv2.1, Kv2.2, or Kv1.4 (in a 1:1 ratio; see Materials and Methods), and outward K⁺ currents, evoked in response to 125 msec depolarizing voltage steps to potentials ranging from −40 mV to +50 mV from a holding potential of −70 mV, were recorded (Fig.1). These experiments revealed that mean ± SEM peak outward K⁺ current densities (30 ± 7 pA/pF; n = 9) in HEK-293 cells expressing Kv2.1DN alone (Fig. 1A) are not significantly different from those recorded from wild-type or mock-transfected HEK-293 cells (15 ± 6 pA/pF; n = 11). However, when the Kv2.1DN is coexpressed with wild-type Kv2.1, outward K⁺ currents are attenuated markedly compared with those recorded from cells expressing Kv2.1 (and EGFP) alone (Fig. 1B, Table1). Although current amplitudes are attenuated markedly, the time- and voltage-dependent properties of the residual currents are indistinguishable from those determined in cells expressing Kv2.1 alone. Increasing the relative amount of the Kv2.1DN to wild-type Kv2.1 to 5:1 completely eliminated the Kv2.1-encoded K⁺ currents. Unexpectedly, however, the currents produced on coexpression of Kv2.1DN with Kv2.2 in a 1:1 ratio are indistinguishable from those produced by Kv2.2 alone (Fig.1C, Table 1). Similar results were obtained when the (1:1) ratio of the mutant construct Kv2.1DN to wild-type Kv2.2 was increased to 5:1. Immunohistochemical experiments with the anti-FLAG antibody (see Materials and Methods) revealed that the expression levels of the Kv2.1DN protein were similar in HEK-293 cells expressing Kv2.1 or Kv2.2 (data not shown). Together, these results suggest that, in HEK-293 cells, Kv2.1DN does not assemble with Kv2.2 (see also below). Additional experiments revealed that coexpression of Kv2.1DN does not measurably affect the currents produced on expression of Kv1.4 (Fig.1D, Table 1) or other Kvα-subunits (data not shown), revealing that the Kv2.1DN construct selectively coassembles with Kv2.1.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Kv2.1W365C/Y380T (Kv2.1DN) specifically attenuates Kv2.1-encoded currents in HEK cells. Whole-cell voltage-gated outward K⁺ currents were recorded from HEK-293 cells expressing Kv2.1DN alone (A), Kv2.1 or Kv2.1 plus Kv2.1DN (B), Kv2.2 or Kv2.2 plus Kv2.1DN (C), Kv1.4 or Kv1.4 plus Kv2.1DN (D), and EGFP. Cells were transfected with cDNA constructs (1:1) encoding these subunits, and currents were obtained from EGFP-positive cells as described in Materials and Methods. Schematic of Kv2.1DN is shown, with altered residues. A, In cells expressing Kv2.1DN alone, outward K⁺currents are indistinguishable from those in wild-type HEK-293 cells. B, In HEK-293 cells coexpressing Kv2.1DN and Kv2.1, current densities are reduced markedly compared with those recorded from cells expressing Kv2.1 alone (compare leftand right). C, In contrast, coexpression of Kv2.1DN with Kv2.2 reveals outward currents indistinguishable from those measured in cells expressing Kv2.2 alone. D, Outward K⁺ currents in cells expressing Kv2.1DN and Kv1.4 are indistinguishable from those expressing Kv1.4 alone.

A similar strategy was exploited to generate a double pore mutant of Kv2.2, Kv2.2W373C/Y388T, that also functions as a dominant negative (Kv2.2DN); this construct was myc epitope-tagged at the C terminus (see Materials and Methods). Electrophysiological experiments revealed that mean ± SEM peak outward K⁺ current densities (26 ± 6 pA/pF;n = 10) recorded from HEK-293 cells expressing Kv2.2DN (Fig. 2A) are not significantly different from those recorded from wild-type or mock-transfected HEK-293 cells (15 ± 6 pA/pF; n = 11). However, coexpression of Kv2.2DN with wild-type Kv2.2 (1:1) markedly reduced peak outward K⁺ current densities compared with the currents recorded from cells expressing Kv2.2 (and EGFP) alone (Fig. 2B, Table 1). The time- and voltage-dependent properties of the residual currents are indistinguishable from those determined in HEK-293 cells expressing Kv2.2 alone. Increasing the relative amount of the Kv2.2DN to wild-type Kv2.2 cDNA to 5:1 eliminated the Kv2.2-encoded K⁺ currents. Similar to the findings with Kv2.1DN and Kv2.2 (Fig. 1C), coexpression of Kv2.2DN with wild-type Kv2.1 at a 1:1 (Fig. 2C, Table 1) or 5:1 ratio does not measurably affect Kv2.1-encoded currents. Immunohistochemical experiments with the anti-myc antibody revealed that the expression levels of the Kv2.2DN-myc, however, are similar in Kv2.2- and Kv2.1-expressing HEK-293 cells (data not shown). Therefore, these combined observations are consistent with the hypothesis advanced above that Kv2.1 and Kv2.2 do not coassemble in HEK-293 cells (see also below).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Kv2.2W373C/Y388T (Kv2.2DN) specifically attenuates Kv2.2-encoded currents in HEK cells. Whole-cell voltage-gated outward K⁺ currents were recorded from HEK-293 cells expressing Kv2.2DN (A), Kv2.2 or Kv2.2 plus Kv2.2DN (B), Kv2.1 or Kv2.1 plus Kv2.2DN (C), and EGFP, as described in the legend to Figure 1. A, Schematic of Kv2.2DN is shown, with the altered residues indicated. In cells expressing Kv2.2DN alone, outward K⁺ currents are similar to those in wild-type cells. B, When Kv2.2DN is coexpressed with Kv2.2, however, Kv2.2-encoded outward K⁺ currents are reduced compared with cells expressing Kv2.2 alone (compareleft and right). C, In contrast, Kv2.2DN does not affect Kv2.1-encoded currents.

Expression of Kv2.1DN or Kv2.2DN attenuatesI_K in SCG neurons

To assess the functional consequences of expression of the Kv2.x dominant negative constructs on outward K⁺currents in SCG neurons, cells were transfected with Kv2.1DN-FLAG or Kv2.2DN-myc (and EGFP) using the Biolistics Gene Gun (see Materials and Methods). As reported previously (Malin and Nerbonne, 2000, 2001) EGFP is readily detected in the cell bodies of SCG cells within ∼15 hr of transfection (data not shown). Cells were stained with the anti-FLAG or anti-myc antibody to verify expression of the transgenes, Kv2.1DN and Kv2.2DN, and both constructs were readily detected in all EGFP-positive cells (data not shown). Also similar to previous findings using this methodology (Malin and Nerbonne, 2000, 2001), all EGFP-positive cells in these cultures also expressed the transgene.

Representative whole-cell voltage-gated outward K⁺ currents recorded from three EGFP-positive Kv2.1DN-expressing SCG cells are presented in Figure3 (middle). As in wild-type cells (Fig. 3, left), type I, type II, and type III Kv2.1DN-expressing cells were readily distinguished (Table2). Most (∼70%) cells expressI_Af,I_K, andI_SS, and are therefore classified as type I cells (Malin and Nerbonne, 2000). However, the density ofI_K in Kv2.1DN-expressing type I cells is significantly (p < 0.005) lower thanI_K density in wild-type I cells (Table2). Approximately 20% of the Kv2.1DN-expressing SCG cells are type II (expressing I_Af,I_As,I_K, andI_SS), similar to the percentage (∼25%) of wild-type II cells (Fig. 3, Table 2). In type II, as in type I, cells, Kv2.1DN expression significantly (p < 0.05) reducesI_K density (Table 2). In both type I and type II cells, the effects of Kv2.1DN expression are specific; neither the densities of I_Af,I_As, orI_SS nor the kinetics ofI_Af orI_As decay in these cells are significantly different from those measured in wild-type I and II SCG cells (Table 2). Similar to the findings in wild-type cells, ∼10% of Kv2.1DN-expressing cells are classified as type III SCG neurons, expressing only I_K andI_SS, and, interestingly,I_K density in type III cells is unaffected by Kv2.1DN expression (Table 2). Therefore, expression of Kv2.1DN results in the selective attenuation (by ∼50%) ofI_K in type I and type II SCG cells.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Expression of Kv2.1DN or Kv2.2DN reducesI_K density in SCG neurons. Whole-cell voltage-gated outward K⁺ currents were recorded from isolated SCG neurons in response to 6 sec depolarizing voltage steps to test potentials between −10 and +50 mV from a holding potential of −90 mV. Experiments were conducted as described in Materials and Methods, with 1 μm TTX and 100 μmCdCl₂ in the bath solution to block voltage-gated inward Na⁺ and Ca²⁺ currents, respectively. The records shown to the left,middle, and right were recorded from wild-type, Kv2.1DN-expressing, and Kv2.2DN-expressing cells, respectively. There are distinct and stereotyped differences in the waveforms of the currents in wild-type I, type II, and type III SCG cells (Malin and Nerbonne, 2000). The numbersgiven above the records in each column reflect the percentages of cells studied under each experimental condition that display the type I, type II, or type III outward K⁺ current phenotype. Although the percentages of type I, type II, and type III Kv2.1DN- or Kv2.2DN-expressing cells are not different from those in wild-type SCG cells, expression of either Kv2.1DN (middle) or Kv2.2DN (right) decreases the density of the slowly decaying current, I_K, in type I cells. Expression of Kv2.1DN also decreases I_Kdensity in type II cells (middle).

Similar experiments were conducted on cells expressing Kv2.2DN, and representative outward K⁺ currents recorded from (3) Kv2.2DN- (and EGFP) expressing SCG neurons are shown in Figure 3 (right). As in wild-type and Kv2.1DN-expressing SCG cells, most (∼70%) Kv2.2DN-expressing cells can be classified as type I, expressing I_Af,I_K, andI_SS. In Kv2.2DN-expressing type I cells (Fig. 3), I_K density is significantly (p < 0.005) lower than in wild-type I SCG neurons (Table 2). The densities ofI_Af andI_SS and the kinetics ofI_Af andI_K decay, in contrast, are unaffected by Kv2.2DN expression (Table 2). In ∼20% of the Kv2.2DN-expressing SCG cells, I_Af,I_As,I_K, andI_SS were evident, and these cells were classified as type II (Malin and Nerbonne, 2000). The mean ± SEM densities and inactivation rates ofI_Af,I_As,I_K, andI_SS in Kv2.2DN-expressing type II cells are not significantly different from those recorded in wild-type II cells (Table 2). In contrast to the findings with Kv2.1DN, which reduces I_K density in type I and type II cells, expression of Kv2.2DN reducesI_K density in type I cells without affecting I_K density in type II cells (Table 2). In these experiments, two Kv2.2DN-expressing cells were classified as type III cells (expressingI_K andI_SS). AlthoughI_K densities in these (2) cells are lower than in wild-type III cells (Table 2), the small number of Kv2.2DN-expressing type III cells precluded statistical analysis.

Coexpression of Kv2.1DN and Kv2.2DN eliminates most ofI_K in SCG neurons

The results of the experiments described above suggested that the Kv2.1 and Kv2.2 α-subunits form distinct populations ofI_K channels in SCG cells, and that Kv2.1-encoded I_K channels are expressed in both type I and type II cells, whereas Kv 2.2-encodedI_K channels are expressed only in type I SCG cells (Table 2). To explore this hypothesis further, currents were recorded from cells expressing both Kv2.1DN and Kv2.2DN (and EGFP). In these experiments, the relative amounts of the Kv2.1DN and Kv2.2DN cDNAs were the same, and each was one-half of the amounts of each used in the single transfection experiments (Fig. 3) to avoid any possible complications attributable to gene dosage effects. As is evident in the records shown in Figure 4, the combined expression of Kv2.1DN and Kv2.2DN eliminatesI_K in most cells. In 12 of 16 (75%) of the cells examined, only I_Af andI_SS were detected; these cells were classified as type I cells lacking I_K(Table 2). The remaining four (of 16; 25%) cells are classified as type II cells, expressing I_Af,I_As,I_K, andI_SS, although the mean ± SEMI_K density is reduced markedly in these cells, compared with wild-type II cells (Table 2). However, the mean ± SEM I_K density in type II cells expressing both Kv2.1DN and Kv2.2DN is not significantly different from the I_K density in Kv2.1DN-expressing cells, consistent with the hypothesis that Kv2.1 but not Kv2.2 contributes to I_K in type II cells. Importantly, in both type I and type II cells,I_Af,I_As, andI_SS are unaffected by the combined expression of Kv2.1DN and Kv2.2DN, indicating that Kv2 α-subunits do not contribute to these currents (Table 2). In these experiments, no type III cells (expressing I_K andI_SS) were detected, an observation that likely reflects the small number (16) of cells studied.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Coexpression of Kv2.1DN and Kv2.2DN eliminatesI_K in most SCG neurons. Isolated SCG neurons were transfected with both Kv2.1DN and Kv2.2DN (and EGFP) using the Biolistics Gene Gun (as described in Materials and Methods), and outward K⁺ currents were recorded from EGFP-positive cells as described in the legend to Figure 3. Two distinct current waveforms were evident in these recordings: most (75%) cells were found to express only I_Af andI_SS and therefore are type I cells lackingI_K; the remaining cells (25%) express I_Af,I_As,I_K, andI_SS and are classified as type II cells with reduced I_K density (Table 2). The densities of I_Af,I_As, andI_SS in Kv2.1DN plus Kv2.2DN-expressing type I and type II cells are indistinguishable from those measured in wild-type I and II cells (Table 2).

Expression of Kv2 α-subunits in rat SCG neurons

Together, the simplest interpretation of the results presented above is that Kv2.1- and Kv2.2-encoded K⁺channels underlie I_K in type I cells, whereas only Kv2.1 (and not Kv2.2) channels contribute toI_K in type II cells. These findings suggest that there is an additional component ofI_K in type II SCG neurons that is not encoded by Kv2 α-subunits. In addition, the combined observations in Kv2.1DN-, Kv2.2DN-, and Kv2.1DN plus Kv2.2DN-expressing SCG cells suggest that the Kv2.1 and Kv2.2 α-subunits do not associate to form heteromultimeric K⁺ channels in SCG neurons. Therefore, these results are consistent with the electrophysiological findings in HEK-293 cells suggesting that Kv2.1 and Kv2.2 do not coassemble (Figs. 1, 2). Consequently, subsequent experiments were focused on examining Kv2.1 and Kv2.2 expression in HEK-293 cells and in SCG neurons and on determining whether Kv2.1 and Kv2.2 are associated in situ.

Biochemical experiments were performed on extracts of Kv2.1 and Kv2.2-transfected HEK-293 cells and on fractionated SCG neuronal membranes using specific anti-Kv2.1 and anti-Kv2.2 antibodies (see Materials and Methods). As illustrated in Figure5A, the anti-Kv2.1 antibody detects a single band at ∼114 kDa in HEK-293 cells transfected with Kv2.1 and Kv2.2. Similarly, in blots probed using the anti-Kv2.2 antibody, a single protein band at ∼90 kDa was identified (Fig.5A). Western blot analysis also revealed robust expression of Kv2.1 and Kv2.2 in lysates of rat SCG (Fig. 5A). Importantly, both the anti-Kv2.1 and the anti-Kv2.2 antibodies can be used to immunoprecipitate proteins (against which each of these antibodies was targeted) from homogenates of HEK-293 cells transfected with both the Kv2.1 and Kv2.2 cDNAs (Fig. 5B). In contrast, however, Kv2.2 does not coimmunoprecipitate with the anti-Kv2.1 antibody, and Kv2.1 does not precipitate with the anti Kv2.2 antibody (Fig. 5B). When the immunoprecipitation was performed with the anti-Kv2.1 antibody, the Kv2.2 (nonprecipitating) protein was identified in the supernatant (Fig. 5B, lanes S). In addition, when the anti Kv2.2 antibody was used in the immunoprecipitation, the Kv2.1 (nonprecipitating) protein was readily detected in the supernatants (Fig. 5B, lanes S). Similar results were obtained in experiments performed on lysates of rat SCG. The anti-Kv2.1 antibody precipitates the Kv2.1 protein but not the Kv2.2 protein expressed in SCG neurons (Fig. 5C). Similarly, after immunoprecipitations with the anti-Kv2.2 antibody, the Kv2.2 protein is found in the pellet (Fig. 5C, lane P), whereas the Kv2.1 protein is in the supernatant (Fig.5C, lane S). Together, these data suggest that Kv2.1 and Kv2.2 do not associate either in HEK-293 cells or in SCG neurons but rather preferentially form monomeric (Kv2.1 or Kv2.2) voltage-gated K⁺ channels.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Kv2.1 and Kv2.2 are expressed in rat SCGs but do not appear to associate. A, Lysates of control and transfected HEK-293 cells and of isolated rat SCGs were fractionated in SDS-PAGE gels and immunoblotted (IB) with the monoclonal anti-Kv2.1 (left) and the polyclonal anti-Kv2.2 (right) antibodies. B, C, Lysates prepared from transfected HEK-293 cells (B) and rat SCG neurons (C) were immunoprecipitated (IP) with either the monoclonal anti-Kv2.1 or the polyclonal anti-Kv2.2 antibody, fractionated, and immunoblotted with the same antibodies. Although both the anti-Kv2.1 and anti-Kv2.2 antibodies reliably immunoprecipitate the proteins against which each of these antibodies were generated, the Kv2.1 and Kv2.2 α-subunits do not coimmunoprecipitate from lysates prepared from Kv2.1- and Kv2.2-transfected HEK-293 cells or rat SCGs. After immunoprecipitations with the anti-Kv2.1 antibody, the Kv2.2 protein is evident in the supernatants (lanes S) but not in the pellet (lanes P). Similarly, the Kv2.1 protein is found in the supernatant (S) after immunoprecipitation with the anti-Kv2.2 antibody. Closed arrows indicate Kv2.1; open arrows indicate Kv3.2.

Kv2.1-encoded I_K channels regulate action potential durations in tonic SCG cells

In wild-type rat SCG neurons, three distinct repetitive firing patterns, phasic, adapting, and tonic, are observed in response to prolonged (6 sec) depolarizing current injections (Malin and Nerbonne, 2000). These studies revealed that ∼45% of the cells are phasic, firing one or two action potentials, and then become refractory in response to prolonged (6 sec) current injections. Importantly, phasic cells do not fire additional action potentials in response to larger current injections. In contrast, adapting cells (∼30%) fire trains of action potentials in response to low amplitude current injections. However, when the amplitude of the current injection is increased, the firing rate adapts and these cells become refractory. Adapting cells were further distinguished (from phasic and tonic cells) by increased input resistances and decreased current thresholds for action potential generation. Approximately 25% of wild-type SCG cells fire trains of action potentials in response to depolarizing current injections, and the firing frequency increases with the amplitude of the injected current [i.e., no refractoriness is evident in these (tonic) cells over the range of injected currents examined]. Tonic cells were also distinguished from phasic and adapting cells by briefer action potential durations (Malin and Nerbonne, 2000).

In initial experiments focused on exploring the hypothesis that Kv2.x-encoded I_K channels might play a role in shaping action potential waveforms in SCG neurons, Kv2.x-expressing HEK-293 cells and isolated SCG neurons were held at −48 mV (the mean resting membrane potential of rat SCG neurons), and a voltage-clamp paradigm that simulates action potentials typically recorded in phasic, adapting, and tonic SCG neurons was presented. The action potential clamp paradigms are illustrated in Figure6 (bottom). As illustrated at the top of Figure 6, Kv2.1-encoded K⁺currents in HEK-293 cells are activated similarly using the phasic, tonic, and adapting action potential clamp waveforms. The outward K⁺ currents peak ∼1 msec after the peak of the action potential at current densities of ∼25–40 pA/pF (Table3). This is ∼20% of the current activated in these cells by a step depolarization to +50 mV (289 ± 60 pA/pF; see Table 1). Therefore, substantial currents can be evoked by single neuronal action potential waveforms when Kv2.x subunits are expressed in HEK-293 cells. These findings are consistent with the results observed using standard voltage-step protocols, which reveal activation thresholds between −40 mV and −30 mV for Kv2.x channels in HEK cells (Figs. 1, 2).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Activation of Kv2-encoded K⁺currents during action potentials in SCG neurons. To explore directly the activation of Kv2.x-encoded K⁺ currents during action potential waveforms in SCG neurons, cells were held at the typical resting membrane potential of SCG neurons (∼48 mV; see Table3), and outward K⁺ currents activated by typical phasic, adapting, and tonic action potential waveforms were recorded. The voltage-clamp paradigms are illustrated at thebottom. Representative outward current waveforms in Kv2.1-expressing HEK-293 cells driven by the phasic adapting and tonic action potential waveforms are illustrated at the top. Representative outward K⁺ current waveforms in isolated SCG neurons (activated using the action potential voltage-clamp paradigms shown below the record) are presented at the bottom. Action potential clamp recordings from wild-type (solid line), Kv2.1DN-expressing (short dashed line), and Kv2.2DN-expressing (long dashed line) SCG cells are superimposed for comparison purposes.

To probe directly the activation of Kv2.x channels in SCG cells, similar experiments were performed in SCG cells using the action potential clamp paradigms (Fig. 6). Peak outward K⁺ current densities measured in SCG cells with the action potential clamp paradigm are 50–60 pA/pF (Table 3). Comparison of these values with current densities measured using the conventional step depolarization paradigm (Fig. 3) reveals that peak current densities activated during the action potential clamp are ∼25% of the mean peak current densities measured at +50 mV in response to step depolarizations from −70 mV (Fig. 3, Table 2). In addition, in SCG cells expressing Kv2.1DN or Kv2.2DN, peak outward K⁺ currents activated by phasic, tonic, and adapting action potential clamp waveforms are reduced significantly compared with those in wild-type cells (Table 3). Representative voltage-clamp records are illustrated in the middle of Figure 6. As might be expected from the data in HEK-293 cells, these analyses suggest that the contribution of the Kv2.x-encoded K⁺ currents is greater at the time corresponding to 50% rather than 90% repolarization. Together, these observations suggest that there is substantial outward K⁺ current through Kv2-encoded K⁺ channels in SCG neurons during action potentials, and also that blocking Kv2-encoded K⁺ currents likely would impact action potential durations. Subsequent experiments were aimed at testing this hypothesis directly.

The phasic, adapting, and tonic firing patterns seen in wild-type SCG cells are also evident in recordings from cells expressing Kv2.1DN, and the selective attenuation of the Kv2.1-encodedI_K does not affect the distribution of firing patterns (Fig. 7). The expression of Kv2.1DN also does not affect the mean ± SEM input resistances, resting potentials, action potential amplitudes, or the current thresholds for action potential generation in phasic, adapting, or tonic SCG cells (Table 4). As in wild-type SCG cells, Kv2.1DN-expressing adapting cells are readily distinguished from phasic and tonic cells by increased input resistances and lower current thresholds for action potential generation (Table 4). In tonic cells expressing the Kv2.1DN, however, the mean ± SEM action potential duration at 90% repolarization is significantly (p < 0.02) longer than in wild-type tonic cells (Fig. 7B, Table 4). Interestingly, although action potential durations in Kv2.1DN-expressing tonic cells are prolonged, tonic firing is not eliminated, and the firing frequencies of tonic cells are unchanged. Therefore, brief action potentials, regulated by Kv2.1-encodedI_K, are not the sole determinant of tonic firing.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 7.

Phasic, adapting, and tonic firing patterns in SCG neurons expressing Kv2.1DN. SCG neurons were transfected with Kv2.1DN (and EGFP), and action potentials and repetitive firing patterns were recorded in response to brief or prolonged depolarizing current injections, as described in Materials and Methods. A, Current-clamp recordings from three representative Kv2.1DN-expressing cells are shown. In each cell, single action potentials were elicited by 1.5 msec current injections (left), and repetitive firing patterns were recorded in response to 100 pA (middle) or 200 pA (right) 500 msec current injections. Based on the response(s) to the 500 msec current injections, cells were classified as phasic (top), adapting (middle), or tonic (bottom) (Table 4). Although Kv2.1DN expression does not affect the distribution of firing patterns in SCG neurons, action potentials in tonic cells expressing Kv2.1DN are prolonged (B). Action potential durations in phasic and adapting cells (B), in contrast, are unaffected by Kv2.1DN expression (Table 4). wt, Wild type.

Kv2.2-encoded I_K channels regulate action potential threshold and repolarization

To explore the functional role of Kv2.2-encodedI_K channels, current-clamp recordings were obtained from SCG cells expressing Kv2.2DN (and EGFP). Similar to the findings with Kv2.1DN, all three firing patterns are observed in Kv2.2DN-expressing cells (Fig. 8). Also, similar to the findings with Kv2.1DN, action potential durations at 90% repolarization in tonic cells expressing Kv2.2 are significantly (p < 0.02) longer than in tonic wild-type cells (Table 4). However, in contrast to the findings with Kv2.1DN, expression of Kv2.2DN results in an increase in the percentage of adapting cells and a decrease in the percentage of phasic cells (Fig.8, Table 4). Elimination of Kv2.2-encodedI_K channels also has marked effects on the excitability of SCG neurons. In adapting and tonic cells expressing Kv2.2DN, for example, resting membrane potentials are depolarized significantly (p < 0.001) compared with resting membrane potentials in wild-type adapting and tonic SCG cells (Table4). In addition, the current thresholds for action potential generation are reduced significantly in phasic (p < 0.002) and in tonic (p < 0.02) firing SCG cells expressing Kv2.2DN. However, input resistances and action potential amplitude are unaffected by Kv2.2DN expression (Table 4). Together, these results suggest that, unlike Kv2.1-encodedI_K channels, Kv2.2-encodedI_K channels play a role in setting resting potentials and in regulating action potential generation.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 8.

Expression of Kv2.2DN in SCG neurons increases the number of adapting cells. In SCG neurons transfected with Kv2.2DN (and EGFP), action potentials and repetitive firing patterns were recorded, and cells were classified as phasic (top), adapting (middle), or tonic (bottom), as described in the legend to Figure 7. Similar to the results obtained with Kv2.1DN, expression of Kv2.2DN prolongs action potential durations in tonic cells (B). In addition, Kv2.2DN expression alters the distribution of firing patterns: the percentage of phasic cells is reduced (by ∼40%), and the percentage of adapting cells is increased (Table 4). wt, Wild type.

Elimination of I_K prolongs action potentials and reduces tonic firing

The voltage-clamp experiments detailed above revealed that expression of either Kv2.1DN or Kv2.2DN reducesI_K density in type I SCG cells by ∼50%, whereas only Kv2.1DN expression attenuatesI_K in type II cells. In addition, coexpression of Kv2.1DN and Kv2.2DN eliminatesI_K in all type I cells and reducesI_K density in type II cells. Together with the biochemical data presented (Fig. 5), these results suggest that there are two distinct populations of Kv2.x-encodedI_K channels, and that these channels are differentially expressed: Kv2.1-encodedI_K channels in both type I and type II cells and Kv2.2-encoded I_K channels in type I cells. To determine the effects of removing both populations of Kv2 α-subunit-encoded I_K channels, action potentials and repetitive firing patterns were recorded from SCG neurons expressing both Kv2.1DN and Kv2.2DN. Recordings from representative cells are shown in Figure9. Several of the effects of Kv2.1DN and Kv2.2DN coexpression appear to be the sum of the effects of Kv2.1DN or Kv2.2DN expression alone. As in Kv2.2DN-expressing cells, for example, the mean ± SEM current threshold for action potential generation is decreased in cells expressing both dominant negative constructs. In addition, resting membrane potentials are depolarized in these cells compared with wild-type cells (Table 4). Interestingly, action potential durations in adapting (p < 0.02) and phasic (p < 0.004) Kv2.1DN plus Kv2.2DN-expressing cells are prolonged significantly, whereas expression of either Kv2.1DN or Kv2.2DN alone does not affect phasic and adapting action potential durations (Table 4). The most prominent effect of coexpression of both Kv2 dominant negative constructs, however, is the reduction of tonic cell number from 25% of wild-type SCG neurons to 8% of Kv2.1DN plus Kv2.2DN-expressing neurons. These observations are consistent with a role forI_K in action potential repolarization and in the regulation of tonic firing.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 9.

Coexpression of Kv2.1DN and Kv2.2DN markedly reduces tonic firing. Isolated SCG neurons were transfected with Kv2.1DN, Kv2.2DN, and EGFP. Action potentials and repetitive firing patterns were recorded, and cells were classified as phasic (top), adapting (middle), or tonic (bottom), as described in the legend to Figure 7. In contrast to the results obtained with expression of Kv2.1DN or Kv2.1DN alone (Figs. 7, 8), expression of both dominant negative constructs markedly decreases the percentage of tonic firing cells (Table4).

Expression of wild-type Kv2.1 α-subunits increases tonic firing

The results of the experiments described here suggest that Kv2-encoded I_K channels play a role in determining tonic firing. To explore this hypothesis further, action potentials and repetitive firing patterns were recorded from SCG cells transfected with wild-type Kv2.1 and EGFP (Fig.10). These experiments revealed that the membrane properties of Kv2.1-transfected and wild-type phasic, adapting, and tonic SCG cells are indistinguishable (see on-line Table, available at www.jneurosci.org). However, expression of Kv2.1 markedly reduces the number of phasic (∼25%) and adapting (∼25%) cells and increases the number of tonic cells (50%). In addition, mean ± SEM action potential durations at 50 and 90% repolarization in Kv2.1-expressing tonic cells are significantly (p < 0.02) briefer than in phasic and adapting cells.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 10.

Expression of Kv2.1 in SCG neurons increases tonic firing. Action potentials and repetitive firing patterns, recorded as described in the legend to Figure 7, were obtained from isolated SCG neurons 24 hr after transfection with wild-type Kv2.1 and EGFP. As in wild-type cells, the phasic (top), adapting (middle), and tonic (bottom) firing patterns were seen in recordings from cells transfected with Kv2.1. However, the percentage of tonic cells is higher and the percentages of adapting and phasic cells are lower in Kv2.1-expressing cells than in either wild-type, Kv2.1DN-expressing (Fig. 7), or Kv2.2DN-expressing (Fig. 8) SCG cells (Table 4).

Two distinct components of I_K encoded by Kv2.1 and Kv2.2 in SCG neurons

The studies described here were designed to probe the role of Kv2 α-subunits in the generation of voltage-gated K⁺ currents in SCG neurons. Biochemical experiments revealed that both Kv2.1 and Kv2.2 are expressed in SCG neurons, and that these α-subunits do not appear to associate in these cells or in HEK-293 cells (Fig. 5). To explore the functional roles of Kv2 α-subunits, pore mutants of Kv2.1 (Kv2.1DN) and Kv 2.2 (Kv2.2DN), designed to function as dominant negatives, were constructed. Heterologous coexpression studies in HEK-293 cells revealed that the dominant negative effects of Kv2.1DN and Kv2.2DN are specific for Kv2.1 and Kv2.2 channels, respectively (Figs. 2, 3, Table1). The simplest interpretation of these observations is that Kv2.1 and Kv2.2 α-subunits do not coassemble, at least in HEK-293 cells (Table1). When expressed in SCG cells, Kv2.1DN selectively attenuatesI_K in type I (expressingI_Af,I_K,I_SS) and type II (expressingI_Af,I_As,I_K,I_SS) cells;I_Af,I_As, andI_SS are unaffected by Kv2.1DN expression (Table 2). Expression of Kv2.2DN, in contrast, reducesI_K only in type I cells (Table 2).

The experiments here also revealed that the effects of Kv2.1DN and Kv2.2DN are additive. Coexpression of Kv2.1DN and Kv2.2DN, for example, eliminates I_K in type I SCG cells (Fig. 4, Table 2). Importantly, in these experiments, the total amount of cDNA used in the transfections was the same as with Kv2.1DN or Kv2.2DN alone. The additivity of the effects of Kv2.1DN and Kv2.2DN, therefore, does not simply reflect a dose-dependent attenuation of the current. Rather, the experimental observations suggest that Kv2.1DN and Kv2.2DN affect different populations ofI_K channels, consistent with the hypothesis that Kv2.1 and Kv2.2 do not coassemble in type I SCG neurons. The results presented here also reveal that Kv2-encoded K⁺ channels underlieI_K in all type I SCG neurons. In type II cells, in contrast, I_K is not eliminated with Kv2.1DN and Kv2.2DN coexpression. This might reflect the slow turnover rate of some Kv2.x-encoded K⁺ channels in type II SCG neurons. Alternatively, it is also possible that the expression levels of the dominant negative constructs were too low in these experiments to effectively eliminate I_K in type II cells. Both of these hypotheses seem unlikely, however, in light of the results obtained in type I cells and the similarities in the reductions in I_K in Kv2.1DN- and Kv2.1DN plus Kv2.2DN-expressing type II cells. It seems more likely that only a subset of I_K channels in type II cells are encoded by Kv2 (Kv2.1) α-subunits, and that the residualI_K channels reflect the contribution of another Kvα-subunit subfamily. Nevertheless, additional experiments aimed at determining the molecular identity of the individual I_K channels in type II cells will be necessary to test this hypothesis directly.

Kv2.1- and Kv2.2-encoded I_K channels are differentially expressed in type I and type II SCG cells: Kv2.1-encoded I_K is expressed in both type I and type II cells, whereas Kv2.2-encodedI_K channels are only evident in type I cells (Table 2). Although the numbers and the percentages of type III cells are small, the experiments completed to date suggest thatI_K density is reduced in type III cells expressing Kv2.2DN but not in cells expressing Kv2.1DN (Table 2). No type III cells were seen in recordings from cells coexpressing Kv2.1DN and Kv2.2DN, an observation that likely reflects the small number (n = 16) of cells studied in these experiments.

I_K density regulates action potential repolarization and tonic firing

Expression of either Kv2.1DN or Kv2.2DN selectively increases action potential durations in tonic cells, suggesting thatI_K channels contribute to action potential repolarization in these cells (Table 4). Although expression studies in Xenopus oocytes suggest relatively depolarized (∼0 mV) thresholds for activation (Fink et al., 1996), Kv2.x-encoded K⁺ channels in mammalian cell lines activate at more depolarized (approximately −30 to −40 mV) potentials (Figs. 1, 2) (Murakoshi et al., 1997). In addition, action potential clamp experiments reveal substantial activation (∼20% maximal) of Kv2.x-encoded currents in HEK-293 cells during a single (SCG) action potential (Fig. 6). Furthermore, expression of Kv2.1DN or Kv2.2DN markedly reduces the currents elicited during action potentials in SCG cells (Fig. 6). These results suggest an important role for Kv2.x-encoded K⁺ channels in action potential repolarization in SCG neurons and suggest that these channels, like other K⁺ channels, are differentially modified by expression environment (Peterson and Nerbonne, 1999). Interestingly, a similar conclusion was reached independently by Du et al. (2002) in studies focused on the functioning of Kv2.1-encoded currents in hippocampal neurons.

Although I_K shapes individual action potential waveforms in tonic cells, reductions inI_K density do not affect the percentage of cells that fire tonically (Table 4), revealing that brief action potentials are not the sole determinant of tonic firing. However, coexpression of Kv2.1DN and Kv2.2DN and the resulting reductions (in type II cells) or elimination (in type I cells) ofI_K markedly reduce the percentage of tonic firing cells (Fig. 9). Although action potential repolarization in phasic and adapting cells is not affected by removal of either Kv2.1- or Kv2.2-encoded I_K, elimination of both Kv2.1- and Kv2.2-encodedI_K significantly (p < 0.001) increases action potential durations in these cells (Table 4). Together, therefore, these observations suggest that I_K channels contribute to action potential repolarization in phasic, adapting, and tonic SCG cells, and that total I_Kdensity is a critical determinant of the tonic firing pattern. Consistent with this hypothesis, expression of Kv2.1 increasesI_K density and the percentage of tonic cells (see Fig. 10 and on-line Table, available atwww.jneurosci.org).

Kv2.2-encoded I_K regulates neuronal excitability

Although both Kv2.1- and Kv2.2-encodedI_K channels play roles in action potential repolarization in SCG neurons, only Kv2.2-encodedI_K channels appear to contribute to the regulation of (resting) membrane excitability and to affect action potential firing. Expression of Kv2.2DN, for example, depolarizes all SCG cells and reduces the current thresholds for action potential generation in phasic and tonic cells (Table 4). Thus, unlike Kv2.1-encoded channels, Kv2.2-encodedI_K channels contribute to setting the resting membrane potentials of SCG neurons and function to regulate membrane excitability. Interestingly, the properties of Kv2.1- and Kv2.2-encoded K⁺ channels are very similar in heterologous expression systems (Figs. 1, 2), suggesting that there are cell-type-specific regulatory pathways that differentially modulate Kv2.1- and Kv2.2-encoded K⁺ channels. The functional distinctions between the Kv2.1- and Kv2.2-encodedI_K channels revealed in the experiments reported here suggest that differential regulation of these two I_K components, either by post-translational modification (Summers and Gelband, 1998; Colbert and Pan, 1999; Gelband et al., 1999; Zhu et al., 1999) or by coassembly with regulatory subunits (Chiara et al., 1999; Kerschensteiner and Stocker, 1999), would have dramatically different effects on SCG cell membrane excitability. In addition, because Kv2.1- and Kv2.2-encodedI_K channels are differentially expressed in type I, type II, and type III SCG cells, differential modulation of Kv2.x-encoded I_Kchannels would be expected to regulate neuronal excitability in a cell-type-specific manner. Additional experiments aimed at testing this hypothesis directly will clearly be of interest.

Previous studies have shown that the current thresholds for action potential generation are lower in adapting than in phasic or tonic cells, which appears to reflect, at least in part, the reduced density of the fast transient current I_Af in adapting cells (Malin and Nerbonne, 2000, 2001). The observations presented here that expression of Kv2.2DN reduces action potential thresholds and increases adapting cell number (Table 4) suggest a similar role of Kv2.2-encoded I_Kchannels. Furthermore, these results suggest that the adapting phenotype is correlated with low current thresholds for action potential generation, and that these low threshold values in adapting cells arise from reduced density of Kv2.2-encodedI_K (present study), as well asI_Af (Malin and Nerbonne, 2000, 2001), in these neurons.