SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References
  7. Appendix

Nav1.2 and Nav1.6 are two voltage-gated sodium channel isoforms found in adult CNS neurons. These isoforms differ in their electrophysiological properties, even though the major regions that are known to be involved in channel activation and inactivation are conserved between them. To determine if the terminal domains of these channels contributed to their activation and fast inactivation differences, we constructed chimeras between the two isoforms and characterized their electrophysiological properties. Exchanging the N-terminal 205 amino acids of Nav1.6 and the corresponding 202 amino acids of Nav1.2 completely swapped the V_1/2 of steady-state activation between the Nav1.2 and Nav1.6 channels in an isoform-specific manner. Exchanging the C-terminal 436 amino acids of Nav1.6 and the corresponding region of Nav1.2 altered the voltage dependence and kinetics of steady-state inactivation, but the changes did not reflect a direct transfer of inactivation properties between the two isoforms. Finally, the N- and C-terminal domains from Nav1.6 demonstrated functional cooperation. These results suggest that the terminal sequences of the sodium channel are important for isoform-specific differences between the channels.

Voltage-gated sodium channels are transmembrane proteins responsible for controlling the influx of sodium currents that depolarize the membrane of neuronal cells. Nine different mammalian voltage-gated sodium channels have been identified (Goldin et al. 2000). These channels share a basic structure but differ in their electrophysiological properties, toxin sensitivity, and tissue expression (Goldin, 2001). A voltage-gated sodium channel comprises a pore-forming α subunit and one or more modulatory β subunits in some tissues (Hartshorne & Catterall, 1984). The α subunit alone is sufficient to produce a functional channel (Noda et al. 1986), but the properties of the channel can be modulated by the β subunits (Isom et al. 1992, 1995). In mammals, there are nine α subunit isoforms (Nav1.1 through Nav1.9) and four β subunit isoforms (β1 through β4) (Goldin, 2001; Yu et al. 2003).

Many structural features that contribute to the function of the sodium channel α subunit have been identified. The α subunit contains four homologous domains, with each subdivided into six transmembrane helices (Fig. 1A). In the cell membrane, the four domains assemble around a central pore through which sodium ions pass selectively (Catterall, 2000). The fourth segment of each domain contains positively charged residues at every third position (Stühmer et al. 1989). These ‘voltage sensors’ are highly conserved across isoforms, and they respond to potential changes with outward movements that trigger pore opening. Activation is quickly followed by inactivation, in which a conserved intracellular linker between domains III and IV blocks the pore (West et al. 1992; Patton et al. 1992).

image

Figure 1. Diagrams of voltage-gated sodium channel and chimeras A, schematic diagram of the sodium channel indicating the four homologous domains (I–IV), each consisting of six transmembrane segments. The regions that were swapped to construct the chimeric channels are shaded. B, the regions swapped between the Nav1.2 and Nav1.6 isoforms are diagrammed for each chimera. Numbers above and below the top bar denote residue positions for Nav1.2 and Nav1.6, respectively. The ‘N’ region includes the N-terminus and domain I up to the S3 segment. The ‘C’ region contains domain IV and the C-terminus. Clear bars represent Nav1.2-derived sequences and grey bars represent Nav1.6-derived sequences.

Download figure to PowerPoint

The Nav1.2 and Nav1.6 channels are two voltage-gated sodium channel isoforms found in adult CNS neurons. These isoforms differ in their electrophysiological properties, even though they share 75% sequence identity and 84% sequence conservation (Smith et al. 1998). The major regions involved in channel activation and inactivation such as the S4 voltage sensor segments, the III–IV linker inactivation particle, and the inactivation particle docking site are conserved between them. Therefore, other regions of the channel most likely confer isoform specificity.

The terminal domains of the channels may be important for the functional differences between these isoforms. The C-terminus can modulate inactivation of Nav1.5 and Nav1.8 (Cormier et al. 2002; Motoike et al. 2004; Choi et al. 2004; Glaaser et al. 2006). The C-terminus of Nav1.2 has been shown to affect the kinetics and voltage dependence of steady-state inactivation (Mantegazza et al. 2001), and the N-terminus of Nav1.2 may interact with intracellular elements that contribute to the voltage dependence of inactivation (Kamiya et al. 2004). In this report, we constructed chimeric channels to study the functional effects of the terminal domains of Nav1.2 and Nav1.6. Replacing the N-terminal 202 residues of Nav1.2 with the corresponding region from Nav1.6 conferred the donor channel's voltage dependence of activation on the host channel. Exchanging the C-terminal 436 amino acids of Nav1.6 and the corresponding region of Nav1.2 altered the voltage dependence and kinetics of inactivation, but the changes did not reflect a direct transfer of inactivation properties between the two isoforms. Finally, the N- and C-terminal domains from Nav1.6 demonstrated functional cooperation that was isoform specific.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References
  7. Appendix

Ethical approval

All experiments were performed according to guidelines established by and with the approval of the Institutional Animal Care and Use Committee of the University of California, Irvine.

Construction of chimeric channels

The Nav1.2-FLAG construct was previously characterized (Smith & Goldin, 1997). Briefly, the Scn2a cDNA encoding the rat Nav1.2 channel was tagged at the 5′ end with the synthetic FLAG sequence 5′-GACTATAA-AGACGATGACGATAAA-3′. The tagged construct was placed between a cytomegalovirus (CMV) promoter and a polyadenylation signal from the β-globin gene. A T7 promoter was included behind the CMV promoter to facilitate in vitro transcription. The cDNA clones encoding the β1 subunit and β2 subunit were also previously described (Smith et al. 1998).

The Scn8a cDNA encoding the mouse Nav1.6 channel was tagged with the c-myc sequence 5′-GAGCAAAA-GCTCATCTCAGAAGAGGATCTA-3′ at the 5′ end and placed into the MCS of the mammalian vector pRC/CMV(A) behind a T7 promoter. This Nav1.6 clone differed from the previously published sequence (Burgess et al. 1995) in the following places: V5L, K15R, I142T, N153T, E937Q, V958A, and the I–II linker. The E937Q mutation made the channel resistant to tetrodotoxin. The I–II linker in this clone contained 333 amino acids and represented the major splice variant found in rat (Dietrich et al. 1998; Smith & Goldin, 1998). The V5L, K15R, I142T and N153T mutations could be restored to wild-type residues with no significant change to the channel's voltage dependence of activation and inactivation or its inactivation kinetics. The effect of the V958A substitution was not tested. The c-myc tag did not significantly alter the voltage dependence of activation and inactivation or the inactivation kinetics.

Chimeras of Nav1.2 and Nav1.6 were created using unique restriction sites shared between the two constructs described. Because of the limited number of unique sites, the N-terminal chimera included part of domain I and the C-terminal chimera included part of domain IV. To generate the channel chimeras Nav1.6-N2 and Nav1.2-N6, the gene fragment before the NdeI restriction site (amino acid positions A202–Y203) from Nav1.2 was exchanged with the corresponding fragment from Nav1.6. The resulting ‘N-region’ chimeras switched channel isoform in the third segment of domain I (DIS3). To obtain the chimeras Nav1.2-C6 and Nav1.6-C2, the gene fragment after the BstEII restriction site (amino acid positions M1545–V1546) from Nav1.2 was exchanged with the corresponding fragment from Nav1.6. The resulting ‘C-region’ chimeras switched channel isoform in the first segment of domain IV (DIVS1). Finally, the double-swap chimeras Nav1.2-NC6 and Nav1.6-NC2 were constructed by exchanging the NdeI–BstEII fragments between Nav1.2 and Nav1.6. Figure 1B shows the regions exchanged between the Nav1.2 and Nav1.6 α subunits to generate the chimeras. The cDNA construct sequencing was performed by Laguna Scientific (Laguna Niguel, CA, USA).

Expression and electrophysiology

Plasmids containing the channel constructs were linearized at unique restriction sites NotI, SacII, or SmaI. Capped RNA transcripts were generated from linearized DNA templates using the T7 mMESSAGE mMACHINE transcription kit (Ambion, Austin, TX, USA). RNA yield was estimated by glyoxal gel electrophoresis. Stage V oocytes were removed from adult female Xenopus laevis frogs and prepared as previously described (Goldin, 1991). Approximately 0.02–2 ng of RNA was injected per oocyte to obtain current levels between 1 and 5 μA after 24–48 h. Injected oocytes were incubated in ND-96 solution containing 96 mm NaCl, 2 mm KCl, 1.8 mm CaCl2, 1 mm MgCl2, 0.1 mg ml−1 gentamicin, 0.55 mg ml−1 pyruvate, 0.5 mm theophylline, and 5 mm Hepes at pH 7.5 for 24–72 h at 20°C before analysis.

Sodium currents were recorded from oocytes at room temperature using the two-electrode voltage clamp OC-725 (Warner Instruments, Hamden, CT, USA) with DigiData 1320A interface (Molecular Devices, Sunnyvale, CA, USA) and pCLAMP 8 software (Molecular Devices). Oocytes were maintained in ND-96 solution (without gentamicin, pyruvate and theophylline) during recording. Transient capacitive and leak currents were corrected by P/4 subtraction.

The voltage dependence of activation was determined from the sodium currents elicited when the oocyte was step-depolarized from a holding potential of −100 mV to +30 mV in increments of 10 mV. Conductance (G) was calculated using the equation:

  • image

in which I is the current amplitude, V is the test potential, and Vr is the reversal potential. The reversal potential was extrapolated from fitting the I–V curve with the equation

  • image

in which z is the gating charge, V_1/2 is the half-maximal activity potential, and g is a factor related to the number of channels contributing to the observed current. Conductance values were normalized to the peak conductance and fitted with the two-state Boltzmann equation

  • image

The voltage dependence of inactivation was determined from the sodium currents elicited during a −5 mV test pulse immediately after the oocyte was depolarized from a holding potential of −100 mV to a range of conditioning potentials between −95 mV and +25 mV for 100 ms. Currents from the test pulses were normalized to the peak current and fitted with the two-state Boltzmann equation

  • image

in which I is the test pulse current, V is the conditioning potential, V_1/2 is the half-maximal inactivation potential, and a is the slope factor.

The kinetics of fast inactivation were determined from the inactivation phase of the current traces elicited during the conditioning prepulses in the two-step inactivation protocol. The traces were fitted with the double-exponential equation

  • image

in which I is the current, Afast and Aslow are the current fraction inactivating with the time constants τfast and τslow, K is the time shift, and C is the steady-state non-inactivating current. The time shift was selected as the point at which the current trace began to inactivate exponentially.

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References
  7. Appendix

Nav1.2 and Nav1.6 sodium channels show a number of significant differences in their voltage dependence and kinetics of gating, despite the fact that they share 75% sequence identity and 84% sequence conservation (Smith et al. 1998). Because the regions that have been identified as the major determinants in channel activation and inactivation are highly conserved between these two channels, differences in other regions of these isoforms are likely to be important in modulating activation and inactivation. To test the hypothesis that the carboxy- and amino-terminal regions are involved in the modulation of gating, we constructed chimeric channels in which those segments were swapped (Fig. 1B). The regions that were swapped included the amino terminus up to IS3, which will be referred to as the N-region, and the carboxy terminus plus domain IV, which will be referred to as the C-region. The chimeras were named based on the parental channel followed by the region that was substituted from the other isoform. For example, Nav1.2-N6 consists of the Nav1.2 channel with the Nav1.6 N-region, Nav1.2-C6 consists of the Nav1.2 channel with the Nav1.6 C-region, and Nav1.2-NC6 consists of the Nav1.2 channel with both N- and C-regions from Nav1.6 (Fig. 1B). The reverse chimeras, Nav1.6-N2, Nav1.6-C2 and Nav1.6-NC2, were also constructed. However, we were unable to detect any current through the Nav1.6-NC2 chimera, even though its entire cDNA was sequenced to verify that there were no unintended mutations in the channel that could have prevented its expression. The properties of these chimeric channels were analysed by expression in Xenopus oocytes in the absence and presence of the β1 and β2 subunits.

The N-region modulates the voltage dependence of activation

In the absence of the β subunits, Nav1.2 activated at more positive potentials than Nav1.6, with a V_1/2 of −6.6 mV for Nav1.2 compared to −14.0 mV for Nav1.6 (Fig. 2A and Table 1). Since the S4 voltage sensors are identical in these two isoforms, residues in other regions are most likely responsible for this difference. When the N-region of Nav1.2 was replaced with the corresponding region from Nav1.6 (Nav1.2-N6), the chimera demonstrated a V_1/2 of −13.2 mV, which was very similar to that of Nav1.6. In contrast, exchanging the C-region from Nav1.6 into Nav1.2 (Nav1.2-C6) did not significantly affect the V_1/2 of activation. Swapping both N- and C-regions from Nav1.6 into Nav1.2 (Nav1.2-NC6) resulted in a channel with a V_1/2 for activation of −14.9 mV, which was similar to that of Nav1.6 (Fig. 2A and Table 1). These results indicate that the N-terminal 205 amino acids of the Nav1.6 α subunit are important for determining the voltage dependence of activation for this isoform.

image

Figure 2. Voltage dependence of activation Oocytes were injected with RNA encoding each of the parental Nav1.2 and Nav1.6 isoforms and the chimeras as α subunits alone (A) or in the presence of β1 and β2 subunits (B). Sodium currents were recorded by depolarizations from −90 mV to +30 mV in 10 mV increments from a holding potential of −100 mV. Conductance values were calculated by dividing the peak current amplitude by the driving force at each potential and normalizing to the maximum conductance, as described in Methods. The values shown are averages and the error bars are standard deviations. The data were fitted with a two-state Boltzmann equation, and the parameters of the fits are shown in Table 1.

Download figure to PowerPoint

Table 1.  Parameters of voltage-dependent activation
Channelαα+β1 +β2
z (e0)V_1/2 (mV)nz (e0)V_1/2 (mV)n
  1. aStatistically significant difference from Nav1.2 α alone at P < 0.001; bstatistically significant difference from Nav1.6 α alone at P < 0.001; cstatistically significant difference from Nav1.2 α+β1 +β2 at P < 0.001; dstatistically significant difference from Nav1.6 α+β1 +β2 at P≤ 0.002.

Nav1.24.0 ± 0.1−6.6 ± 1.674.6 ± 0.6−13.8 ± 0.9 5
Nav1.63.4 ± 0.3−14.0 ± 2.6 124.7 ± 0.5−21.6 ± 1.4 3
Nav1.2-N63.7 ± 0.4−13.2 ± 3.1a124.8 ± 0.5−20.4 ± 1.8c4
Nav1.2-C64.1 ± 0.2−6.8 ± 1.494.8 ± 0.4−14.7 ± 1.7 11
Nav1.2-NC64.5 ± 0.4−14.9 ± 2.7a115.1 ± 0.6−22.0 ± 1.8c5
Nav1.6-N22.6 ± 0.5 −4.5 ± 2.0b33.7 ± 0.2−12.4 ± 1.7d5
Nav1.6-C23.2 ± 0.2−12.9 ± 1.9 93.7 ± 0.4−16.6 ± 1.9d9

To determine if the same region in Nav1.2 was involved in the Nav1.2 voltage dependence of activation, we replaced the N-terminal 205 residues of the Nav1.6 α subunit with the corresponding region from the Nav1.2 isoform (Nav1.6-N2). Unlike our other chimeras, the expression of Nav1.6-N2 in the membrane of Xenopus oocytes was very low, even though we injected up to 15 times more RNA compared to the maximum injected for the other chimeras. The cDNA of the entire chimera was sequenced to verify that there were no unintended mutations in the channel that might have decreased expression. On average, we observed 0.2–0.8 μA of current when 2.3–23 ng of RNA were injected, with more variability between oocytes than between the amounts injected. Nav1.6-N2 showed a V_1/2 of activation similar to that of Nav1.2 (Fig. 2A and Table 1), suggesting that the N-terminal 202 amino acids of the Nav1.2 α subunit were responsible for isoform-specific voltage dependence of activation. Exchanging the C-region from Nav1.2 into Nav1.6 (Nav1.6-C2) resulted in a channel with a V_1/2 of activation that was similar to that of the parental Nav1.6 channel, indicating that this region in either channel is not an important determinant of isoform-specific voltage dependence of activation for α subunit channels.

The CNS sodium channel α subunit isoforms normally associate with both β1 and β2 auxiliary subunits in 1: 1: 1 stoichiometry (Isom et al. 1994). Therefore, we examined whether the presence of the β subunits affected the ability of the N-region to convey isoform-specific voltage dependence of activation. The Nav1.2, Nav1.6 and chimeric channels were expressed with at least a 2: 2: 1 molar ratio of β1: β2: α subunits, except for Nav1.6-N2, which will be described later. In the presence of β1 and β2, the V_1/2 of activation was shifted 7.2 mV in the negative direction for Nav1.2 and 7.6 mV in the negative direction for Nav1.6, thereby maintaining the voltage difference between the two isoforms (Fig. 2B and Table 1). Swapping the N-region from Nav1.6 into Nav1.2 had a similar effect on the voltage dependence of activation in the presence of β1 and β2 as it did in their absence, since Nav1.2-N6 showed a V_1/2 similar to that of Nav1.6 in both cases. Nav1.2-NC6 also maintained the shift to Nav1.6-like voltage dependence in the presence of β1 and β2 (Fig. 2B and Table 1). These data indicate that the N-region of Nav1.6 also confers isoform-specific voltage dependence of activation in the presence of the β subunits.

Swapping the C-region of Nav1.2 into Nav1.6 (Nav1.6-C2) in the presence of β1 and β2 shifted the voltage dependence of activation towards that of Nav1.2 (Fig. 2B and Table 1). In contrast, the reciprocal chimera Nav1.2-C6 maintained the voltage dependence of the parental Nav1.2 channel. These results suggest that the difference between Nav1.2 and Nav1.6 in the presence of the β subunits depends at least partially on the C-region.

Because of the low surface expression of Nav1.6-N2, it was necessary to inject large quantities of α subunit RNA to obtain current amplitudes that could be analysed. For this reason, we could not co-inject molar ratios of the β subunits without killing the oocyte, so a 1: 1: 2 molar ratio of β1: β2: α was used. Assuming that low expression of Nav1.6-N2 was due to less α subunit protein in the oocyte membrane, then this ratio would still have resulted in a molar excess of protein for the β subunits. In support of this assumption, the current amplitude recorded from coexpression of Nav1.6-N2 and β subunits averaged 2 μA, which was four times the average current amplitude from the same amount of α subunit RNA expressed alone. Furthermore, decreasing the amount of β subunit RNA that was injected to one-tenth the original amount did not decrease the average current amplitude. The increase in current amplitude corresponded well with the current amplitude increase in oocytes shown previously for Nav1.2 α subunits coexpressed with an excess of the β subunits (Isom et al. 1992; Isom et al. 1995). When Nav1.6-N2 was injected as α+β subunits, this chimera demonstrated a V_1/2 of −12.4 mV, which was similar to that of Nav1.2 coexpressed with the β subunits (Fig. 2B and Table 1). Thus, the N-region of Nav1.2 still determined the isoform-specific voltage dependence of activation in the presence of the β subunits.

Both the N- and C-regions affect inactivation

In addition to differences in the voltage dependence of activation, the Nav1.2 and Nav1.6 isoforms showed even larger differences with respect to inactivation. In the absence of the β subunits, Nav1.2 had an inactivation V_1/2 of −33.8 mV, which was approximately 23 mV more positive than that of Nav1.6 (Fig. 3A and Table 2). However, in contrast to their effects on activation, neither the N- nor C-region from Nav1.6 shifted the voltage dependence of inactivation to be comparable to that of Nav1.6 when they were swapped into Nav1.2 individually. Rather, each substitution had a small but statistically significant effect in the opposite direction. Substitution of the N-region (Nav1.2-N6) shifted the V_1/2 by −4 mV, whereas substitution of the C-region (Nav1.2-C6) shifted it by approximately +2 mV. The effects were additive as indicated by the fact that substituting both regions from Nav1.6 into Nav1.2 (Nav1.2-NC6) resulted in a channel with the same inactivation V_1/2 as Nav1.2 (Fig. 3A and Table 2). Substituting the N-region from Nav1.2 into Nav1.6 (Nav1.6-N2) had a larger effect, shifting the V_1/2 by approximately +12 mV towards that of Nav1.2. Substituting the C-region from Nav1.2 into Nav1.6 (Nav1.6-C2) also had a larger effect, shifting the V_1/2 by approximately −10 mV further away from that of Nav1.2 (Fig. 3A and Table 2). Again, neither region could transfer isoform-specific inactivation of Nav1.2 to Nav1.6. These results indicate that the inactivation gating of Nav1.6 is altered to a greater degree than that of Nav1.2 by sequence changes in its terminal regions.

image

Figure 3. Voltage dependence of inactivation Oocytes were injected with RNA encoding each of the parental Nav1.2 and Nav1.6 isoforms and the chimeras as α subunits alone (A) or in the presence of β1 and β2 subunits (B). Oocytes were depolarized from a holding potential of −100 mV to a range of conditioning potentials between −95 mV and +25 mV for 100 ms, followed immediately by a test pulse to −5 mV. The peak current amplitude during each test pulse was normalized to the current amplitude of the first test pulse and plotted as a function of the conditioning pulse potential. The values shown are averages and the error bars are standard deviations. The data were fitted with a two-state Boltzmann equation, and the parameters of the fits are shown in Table 2.

Download figure to PowerPoint

Table 2.  Parameters of voltage-dependent inactivation
Channelαα+β1 +β2
a (mV)V_1/2 (mV)na (mV)V_1/2 (mV)n
  1. aStatistically significant difference from Nav1.2 α alone at P≤ 0.004; bstatistically significant difference from Nav1.6 α alone at P < 0.001; cstatistically significant difference from Nav1.2 α+β1 +β2 at P≤ 0.002; dstatistically significant difference from Nav1.6 α+β1 +β2 at P < 0.001.

Nav1.29.6 ± 0.5−33.8 ± 0.9 98.2 ± 0.7−47.9 ± 1.65
Nav1.68.9 ± 0.9−57.2 ± 2.1 126.0 ± 0.3−53.9 ± 1.67
Nav1.2-N68.6 ± 0.3−37.8 ± 2.5a117.2 ± 0.2−47.2 ± 1.07
Nav1.2-C69.1 ± 0.7−31.9 ± 1.2a67.3 ± 0.5 −43.5 ± 1.6c9
Nav1.2-NC68.6 ± 0.4−34.6 ± 0.8 79.0 ± 0.5 −43.0 ± 1.7c5
Nav1.6-N212.2 ± 0.8 −44.3 + 3.3b47.1 ± 0.3 −50.1 ± 0.5d6
Nav1.6-C28.0 ± 0.6−67.1 ± 1.4b97.1 ± 0.8 −59.7 ± 1.1d5

The presence of β1 and β2 had opposite effects on the voltage dependence of inactivation for Nav1.2 and Nav1.6, thus decreasing the difference between the two isoforms. The two accessory subunits shifted the V_1/2 of Nav1.2 by approximately −14 mV, and they shifted the V_1/2 for Nav1.6 by approximately +3 mV (Fig. 3B and Table 2). Substituting the N-region from Nav1.6 into Nav1.2 (Nav1.2-N6) in the presence of the β subunits did not have a significant effect, whereas substituting the C-region (Nav1.2-C6) caused an approximately +4 mV shift in the voltage dependence of inactivation. Substituting both regions (Nav1.2-NC6) caused an approximately +5 mV shift in the voltage dependence of inactivation (Fig. 3B and Table 2). These effects were greater than those observed in the absence of the β subunits. Substituting the N-region from Nav1.2 into Nav1.6 (Nav1.6-N2) in the presence of the β subunits produced a smaller voltage shift (+4 mV) than it did for the α subunit alone. Substituting the C-region from Nav1.2 into Nav1.6 (Nav1.6-C2) in the presence of the β subunits also produced a smaller voltage shift (−6 mV) than it did for the α subunit alone (Fig. 3B and Table 2). Therefore, the β subunits modulated the effects of the α subunit's N- and C-regions on the V_1/2 of inactivation.

The Nav1.2 and Nav1.6 channels differed in their kinetics of inactivation, with Nav1.6 being faster. The magnitude of the slow time constant for Nav1.6 was either similar to or smaller than that of Nav1.2 at all potentials (Fig. 4A), and the magnitude of the fast time constant for Nav1.6 was smaller than that of Nav1.2 at all potentials (Fig. 4B). There was also a significantly larger percentage of Nav1.6 current inactivating with the fast time constant at potentials ≤−5 mV (Fig. 4C). Swapping the N-region of Nav1.6 into Nav1.2 (Nav1.2-N6) did not affect the kinetics of inactivation, but swapping the N-region of Nav1.2 into Nav1.6 (Nav1.6-N2) slowed inactivation kinetics by decreasing the percentage of current inactivating with the fast time constant at all potentials (Fig. 4C) and increasing the magnitude of the fast time constant at negative potentials (Fig. 4B). Swapping the C-region had a significant effect on the kinetics of inactivation for both isoforms. Swapping the C-region of Nav1.2 into Nav1.6 (Nav1.6-C2) accelerated inactivation, and swapping the C-region of Nav1.6 into Nav1.2 (Nav1.2-C6) slowed inactivation (Fig. 5). However, each effect resulted from a different alteration. The C-region of Nav1.2 decreased the magnitude of the slow time constant compared to Nav1.6 (Fig. 4A), while the C-region of Nav1.6 decreased the percentage of current inactivating with the fast time constant at potentials ≥ 0 mV compared to Nav1.2 (Fig. 4C). Surprisingly, swapping both the N- and C-regions from Nav1.6 into Nav1.2 (Nav1.2-NC6) restored the inactivation kinetics to be comparable to Nav1.2 (Figs 4AC and 5), even though the N-region by itself had no effect.

image

Figure 4. Kinetics of fast inactivation Oocytes were injected with RNA encoding each of the parental Nav1.2 and Nav1.6 isoforms and the chimeras as α subunits alone (A, B and C) or in the presence of β1 and β2 subunits (D, E and F). Traces of sodium currents elicited by 100 ms depolarizations between −25 mV and +20 mV from a holding potential of −100 mV were fitted with the double exponential equation described in Methods. Time constants for the slow component (τslow) of fast inactivation are plotted on a logarithmic scale in panels A and D. Time constants for the fast component (τfast) of fast inactivation are plotted on a logarithmic scale in panels B and E. Panels C and F show the fraction of current inactivating with τfast. The values shown are averages and the error bars indicate standard deviations. Sample sizes were Nav1.2 (8), Nav1.6 (12), Nav1.2-C6 (6), Nav1.2-N6 (9), Nav1.6-C2 (9), Nav1.6-N2 (4), Nav1.2-NC6 (6), Nav1.2 +β1 +β2 (5), Nav1.6 +β1 +β2 (7), Nav1.2-C6 +β1 +β2 (9), Nav1.2-N6 +β1 +β2 (7), Nav1.6-C2 +β1 +β2 (5), Nav1.6-N2 +β1 +β2 (6), Nav1.2-NC6 +β1 +β2 (5).

Download figure to PowerPoint

image

Figure 5. Effects of the C-terminus on inactivation kinetics Sodium currents were recorded from oocytes expressing parental Nav1.2 and Nav1.6 isoforms and the chimeras as α subunits alone during 100 ms depolarization to 0 mV from a holding potential of −100 mV. Current traces were normalized to directly compare the kinetics of inactivation. The peak current amplitudes were Nav1.2 (1.3 μA), Nav1.6 (1.6 μA), Nav1.2-C6 (1.6 μA), Nav1.6-C2 (1.3 μA), Nav1.2-NC6 (2.2 μA).

Download figure to PowerPoint

Co-expression of the β1 and β2 subunits accelerated inactivation of both Nav1.2 and Nav1.6 (compare Fig. 4DF to Fig. 4AC). The β subunits decreased both the slow (Fig. 4D) and the fast (Fig. 4E) time constants for Nav1.2, but their effect on the percentage of Nav1.2 current inactivating with τfast was significant only at the negative potentials (Fig. 4F). On the other hand, the β subunits substantially increased the percentage of Nav1.6 current inactivating with τfast at all the potentials examined. The β subunits also altered the fast and slow time constants, although these effects were subtle. In the presence of β subunits, swapping the N-region from Nav1.6 into Nav1.2 (Nav1.2-N6) or the N-region from Nav1.2 into Nav1.6 (Nav1.6-N2) did not have any significant effect (Fig. 4DF). Swapping the C-region of Nav1.2 into Nav1.6 (Nav1.6-C2) lowered the percentage of current inactivating with the fast time constant between −10 and +10 mV potentials (Fig. 4F). Swapping the C-region from Nav1.6 into Nav1.2 (Nav1.2-C6) slightly increased the magnitudes of the slow and fast time constants (Fig. 4D and E). Swapping both the N- and C-regions from Nav1.6 into Nav1.2 (Nav1.2-NC6) also slightly increased the slow and fast time constants (Fig. 4D and E). Overall, swapping the terminal regions in the presence of the β subunits did not cause large shifts in the kinetics of inactivation like the ones associated with swapping the C-region from Nav1.6 into Nav1.2 (Nav1.2-C6) or the N-region from Nav1.2 into Nav1.6 (Nav1.6-N2) in the absence of the β subunits.

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References
  7. Appendix

In this study, we have shown that Nav1.2 has a more depolarized V_1/2 of activation and a more depolarized V_1/2 of inactivation than Nav1.6 both in the absence and in the presence of the β1 and β2 subunits. These differences are consistent with how the voltage-dependent properties of the two isoforms compare to each other in mammalian neurons in a previous study by Rush et al. (2005), suggesting that the comparative voltage-dependent differences between Nav1.2 and Nav1.6 in oocytes reflect their in vivo behaviours. We also observed differences between the isoforms in the τslow of inactivation that were not noted by Rush et al. (2005). We further demonstrated that the terminal regions of the Nav1.2 and Nav1.6 α subunits are important for the differences between the two isoforms. The N-region, which includes the N-terminus and the first half of domain I, affects the voltage dependence of activation in an isoform-specific manner. The C-region, which includes domain IV and the C-terminus, influences the voltage dependence and kinetics of inactivation for both isoforms, although it does not confer isoform specificity. Isoform-specific interaction may also be occurring between the N- and C-regions.

Both the N- and C-regions in the chimeras include a portion of a transmembrane domain along with the cytoplasmic terminus. We believe that the cytoplasmic N- and C-termini of Nav1.2 and Nav1.6 are more likely than domains I and IV to be responsible for the differences observed in channel gating because these regions are significantly more divergent than domains I and IV. For example, the domain I S1–S3 regions of Nav1.2 and Nav1.6 are 78.6% identical (94.0% conserved), whereas the N-termini are only 66.1% identical (82.7% conserved). Similarly, the domain IV regions of Nav1.2 and Nav1.6 are 87.2% identical (92.8% conserved), whereas the C-termini are only 65.2% identical (76.4% conserved). However, we cannot rule out the possibility that even conservative substitutions in domain I S1–S3 and domain IV might affect channel gating, based on the recent finding that the mutation I136V shifts the voltage-dependent activation of Nav1.7 (Cheng et al. 2008).

Our data indicate that the N-region can shift the V_1/2 of inactivation as well as impart isoform-specific voltage dependence of activation, suggesting that this region is likely to be involved in the conformational changes coupling activation to inactivation. This hypothesis is supported by changes in the slopes of the activation and inactivation curves of some N-region chimeras, which might be caused by altered allosteric interactions between the DI–S4 sensor segment responsible for initiating the activation process and the DIV–S4 sensor segment responsible for initiating the inactivation process (Chanda et al. 2004). Whether the movements of the S4 segments are affected by isoform-specific N-region differences in the transmembrane segments that alter the packing of S1–S3 around S4 in domain I or by differences in the N-terminus that change its interaction with sites anchoring the S4 segment cannot be determined from our current results.

The C-region can affect both the voltage dependence and kinetics of inactivation. Within this region, the Nav1.2 and Nav1.6 sequences differ the most in the distal halves of their C-termini, which contain 65 residue differences compared to only 19 in the proximal half. Deletions in the distal half of the Nav1.2 C-terminus have been shown to accelerate the inactivation rate and shift the voltage dependence in the negative direction, suggesting that this region inhibits inactivation (Mantegazza et al. 2001). It is possible that the C-terminus alters binding of the III–IV linker inactivation particle to the docking site, either by directly interacting with one of these regions or by steric interference. Previous studies have shown that the C-terminus of Nav1.5 modulates inactivation by interacting with the III–IV linker inactivation particle (Cormier et al. 2002; Motoike et al. 2004; Glaaser et al. 2006). Since the C-termini of Nav1.2 and Nav1.6 differ in length and sequence, they may interfere differentially with fast inactivation depending on their proximity to the inactivation gate. However, neither domain IV nor the C-terminus was sufficient to transfer isoform-specific inactivation into Nav1.2 or Nav1.6.

We have also shown that the N-terminal 205 amino acids from Nav1.6 can compensate for the effects of the Nav1.6 C-region. Since the N-region did not affect the inactivation kinetics of Nav1.2 by itself, it is likely that there was cooperativity between the two Nav1.6-derived terminal regions. The possibility that the N-terminus of a voltage-gated sodium channel interacts with another cytoplasmic factor has been suggested for Nav1.2. Kamiya et al. (2004) have shown that coexpression of Nav1.2 with its N-terminal 102 amino acid truncation peptide in HEK293 cells can alter the channel's voltage dependence of inactivation. Although we were unable to examine functional cooperation between the N- and C-terminal sequences of Nav1.2 due to difficulties in expressing the Nav1.6-NC2 chimera in Xenopus oocytes, our data from the Nav1.2-NC6 chimeric channel suggest that the terminal domains of the Nav1.6 channel isoform do interact.

These results suggest that the C-region of the Nav1.6 sodium channel can interfere with fast inactivation unless the Nav1.6 N-region is present to modulate the influence of the C-region. This modulation may occur through either direct or allosteric interactions between the termini, resulting in movement of the C-terminus away from the inactivation gate. Since this cooperativity did not occur when the terminal regions were from different isoforms (Nav1.2 and Nav1.6), it probably involves molecular determinants in these regions specific to Nav1.6. The interaction is likely to contribute at least partially to the isoform-specific inactivation properties of Nav1.6.

The observation that Nav1.6-N2 showed low expression in oocytes and Nav1.6-NC2 showed no expression suggests that the N-region of Nav1.6 may be important for its stable expression at the surface membrane. The fact that the concomitant loss of the native C-region of Nav1.6 exacerbated the effect of the loss of its native N-region suggests that both regions may contribute to the stability of the channel. We believe that the N-region's effect might reflect the structural folding of the cytoplasmic N-terminus domain. When a soluble fluorescent protein was attached before the N-terminus of Nav1.6-N2, the resulting Nav1.6 N2-ECFP channel had more than twice the average whole-cell current amplitude with 1/25th the amount of RNA injected compared to Nav1.6-N2 (A. Lee & A. L. Goldin, unpublished observations). One possible explanation is that the ectopically placed N-terminus from Nav1.2 misfolded, hindering access to a region that added stability to the protein, and the attachment of the fluorescent protein caused a structural refolding that exposed the site required for stability. We do not know if the N-termini of Nav1.2 and Nav1.6 actually affect channel stability in vivo. However, if this region is important for stability in vivo, then the surface expression of sodium channels may be regulated by post-translational modification or cofactor binding that alters the tertiary structure of the N-terminus.

Our results show that the β1 and β2 subunits influenced the effects of the α subunit N- and C-regions on channel inactivation more than the effects of the N-region on activation. Specifically, the effects of N- or C-region swap on the voltage dependence and kinetics of inactivation were diminished in the presence of the β subunits, suggesting that the regions of the α subunit involved in modulating inactivation were themselves modulated by the β subunits. Nav1.6-N2 was the one chimera that showed an obvious β subunit modulation of its N-region's effect on activation as well as inactivation. Nav1.6-N2 showed distinct slopes for its activation and inactivation curves that were made more similar to the slopes of the other chimeras by the presence of the β subunits. The changes to the slopes of both the activation and inactivation curves might have resulted from changes in allosteric interactions among the voltage sensors. This phenotype, along with the previously noted β subunit enhancement of Nav1.6-N2 functional expression, suggests that the association of β subunits can induce a global conformational change to the α subunit that affects a wide range of its physiological properties.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References
  7. Appendix

Appendix

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References
  7. Appendix

Acknowledgements

This work was supported by grants from the National Multiple Sclerosis Society (RG3405A) and NIH (NS48336).