Introduction

The TTX-resistant (TTXr) voltage-gated sodium channels Nav1.8 and Nav1.9 are predominantly expressed in spinal sensory neurons [1, 13, 30]. Electrophysiological studies on cultured dorsal root ganglion (DRG) neurons from adult rodents and humans have revealed two distinct TTXr sodium currents, a slowly inactivating current (ITTXr-Slow) and a persistent current (ITTXr-Per) [8, 9, 11, 14, 24]. Studies on transgenic mice lacking Nav1.8 and Nav1.9 respectively have revealed the molecular determinant for ITTXr-Slow as Nav1.8 [2, 9] and Nav1.9 for ITTXr-Per [20]. The expression and functional properties of both ITTXr-Slow and ITTXr-Per suggest that they can substantially contribute to the electro-responsiveness of nociceptive neurons and thus they are both considered to be attractive targets for the development of new analgesic drugs [4, 12, 18, 23, 33].

Nav1.8 and Nav1.9 are the only sodium channel α subunits that have a serine residue in the pore loop of domain I (IP-loop) at the position known to be crucial for TTX and STX sensitivity of voltage-gated sodium channels (Fig. 1) [1, 3, 13, 27]. The sensitivities to other sodium channel blockers can also be influenced by the identity of the amino acid residue at this position. The cardiac channel Nav1.5, which has a cysteine (C374) at the corresponding position in the IP-loop, is sensitive to block by micromolar levels of cadmium and to external application of the membrane-impermeant quaternary derivative of lidocaine QX-314 [26, 29]. When externally applied, QX-314 can only reach the local anesthetic binding site, which involves specific S6 residues of different domains and is located in the inner portion of the pore, by passing through the sodium channel pore [21, 29, 32]. The cysteine residue located at the TTX-binding site of wild-type Nav1.5 and Nav1.4-Y401C channels is crucial for the ability of extracellular QX-314 to pass through the selectivity pore and block these sodium channels [29]. Substituting this residue in Nav1.5 with tyrosine, which is the corresponding residue in several TTX-sensitive channels including Nav1.4 (Y401), greatly decreases the block of Nav1.5 by cadmium and external QX-314 while increasing the block by STX and TTX. By contrast, substituting Y401 in Nav1.4 with cysteine (Y401C), alanine (Y401A) or aspartate (Y401D) increases block by external QX-314 [29], while decreasing block by STX [15].

Fig. 1
figure 1

Schematic showing the location of the critical residue for sodium channel sensitivity to guanidinium toxins. The residues surrounding the critical residue (Y401 in Nav1.4, shown in bold) in the IP-loop are shown for the known voltage-gated sodium channels Nav1.1–1.9

Full size image

The TTX sensitivity of ITTXr-Per (IC50~40 μM; 9) is similar to that of ITTXr-Slow (IC50~100 μM; 24) and recombinant Nav1.8 channels (IC50~60 μM; 1). However, relatively little is known about the sensitivity of ITTXr-Per to other agents that can block sodium channels. Interestingly, it has been purported that Nav1.9 generates a TTX-resistant but saxitoxin (STX)- sensitive neurotrophin-evoked depolarizing current within CNS neurons [5]. In the present study we asked whether ITTXr-Slowand ITTXr-Per share further pharmacological properties that might be determined by the serine pore residue of Nav1.8 and Nav1.9. We compared the sensitivity of ITTXr-Perand ITTXr-Slow to block by STX, extracellular cadmium and QX-314. We also determined whether the serine residue is a critical determinant for the sensitivity to these blockers by using site-directed mutagenisis. Since both Nav1.8 and Nav1.9 generally express very poorly in heterologous expression systems, we examined wild-type Nav1.4 (Nav1.4-WT) channels and mutants where tyrosine 401 was replaced by a serine (Nav1.4-Y401S) or a cysteine (Nav1.4-Y401C).

Our data indicate that the pore serine residue plays an important role in determining the sensitivity of ITTXr-Slow and ITTXr-Per to STX and TTX but not to cadmium and QX-314.

Material and methods

Construction of mammalian expression vectors encoding rat Nav1.4

The construction of wild-type Nav1.4-RBG4 construct for expression in mammalian cells has been described previously [31]. Nav1.4-Y401C and Nav1.4-Y401S cDNAs in pBluescript SK+ [15] were gifts from Edward Moczydlowski and Laurent Schild, respectively. The Nav1.4 coding sequence was cut out of these plasmid with the restriction endonuclease EcoR1, purified and ligated into the RBG4 expression vector at the EcoR1 site.

Transfection of HEK 293 Cells

Transfections were carried out using the calcium phosphate precipitation method as described previously [10]. HEK293 cells were grown under standard tissue culture conditions (5% CO2, 37°C) in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum. The calcium phosphate-DNA mixture (channel constructs and a green fluorescent protein reporter plasmid) was added to the cell culture medium for 1–2 h, after which the cells were washed with fresh medium. Cells exhibiting green fluorescent protein fluorescence were selected for whole-cell patch-clamp recordings after 1–2 days in culture.

Culture of DRG neurons

DRG cells were studied after short-term culture (<24 h). Cultures of DRG neurons were prepared as previously described [11]. Briefly, the L4 and L5 DRG ganglia were harvested from adult male Sprague Dawley rats. The DRG tissue was treated with collagenase A (1 mg/ml) for 25  min and collagenase D (1 mg/ml) and papain (30 U/ml) for 25 min. The suspension was dissociated in DMEM and Ham’s F12 medium supplemented with 10% fetal bovine serum, and DRG neurons were plated on glass coverslips coated with Poly-L-Lysine and Laminin. Electrophysiological recordings were made within 24 h of dissociation.

Whole-cell patch-clamp recordings

Whole-cell patch-clamp recordings were conducted at room temperature (~21°C) with an EPC-9 amplifier. Data were acquired on a PC with the pulse program (v 8.1, HEKA Electronics, Lambrecht, Germany). Fire-polished electrodes (0.8–2 MΩ) were fabricated from 1.7 mm capillary glass, using a Sutter P-97 puller (Novato, CA, USA). Cells were not considered for analysis if the initial seal resistance was <2 GΩ, if they had high leakage currents (holding current >0.1 nA at −80 mV for HEK 293 cells; >0.5 nA for DRG neurons), or an access resistance >4 MΩ . The average access resistance was 1.9±0.1 MΩ for HEK 293 cells (n=99) and 2.0±0.1 MΩ for DRG neurons (n=59) (mean±S.E). Voltage errors were minimized using 80–90% series resistance compensation, and the capacitance artifact was cancelled by using the computer-controlled circuitry of the patch-clamp amplifier. Linear leak subtraction, based on resistance estimates from four to five hyperpolarizing pulses applied before the depolarizing test potential, was used for all voltage-clamp recordings. Membrane currents were usually filtered at 2.5 kHz and sampled at 10 kHz. The pipette solution contained (in mM) 140 CsF, 1 EGTA, 10 NaCl and 10 HEPES, pH 7.3. The standard bathing solution was (in mM) 140 NaCl, 3 KCl, 1 MgCl2, 1 CaCl2, and 10 HEPES, pH 7.3. For experiments on DRG neurons, 250 nM TTX was added to the bathing solution to block all TTX-sensitive currents. The liquid junction potential for these solutions was <8 mV; data were not corrected to account for this offset. The osmolarity of all solutions was adjusted to 310 mOsm (Wescor 5500 osmometer, Logan, UT). The offset potential was zeroed before the cells were patched.

Toxin solutions and bath application

TTX was obtained from Alomone Laboratories Jerusalem, Israel, STX and QX-314 were obtained from Sigma-Aldrich St. Louis, MO, USA. Toxins were applied through a gravity- driven system allowing a rapid perfusion of the recording chamber. Toxins were washed off in specific experiments using the same perfusion system. The IC50 was calculated based on the single-site Langmuir inhibition isotherm using the following function: IC50=(Itoxin/I0)×[toxin]/(1-Itoxin/I0), where I0 and Itoxin are the peak sodium currents measured before and during application of toxin, respectively, and [toxin] is the concentration of toxin. The sensitivity of sodium currents to specific blockers was usually tested with several different blocker concentrations. However, for estimation of the IC50, a single concentration was chosen. The IC50s estimated with this approach are in good agreement with IC50s calculated from more complete dose-response curves. For example, Backx et al. [3] determined that the IC50 of Nav1.4 for TTX is 16 nM, the IC50 of Nav1.4-Y401C for TTX is >50 μM and the IC50 of Nav1.4-Y401C for cadmium is 29 μM, which are very close to our estimates (Table 1). Results are presented as mean ± SEM.

Table 1 Functional and Pharmacological properties of investigated sodium channels
Full size table

Data analysis

Data were analyzed with Pulsefit (HEKA El ectronics) and Origin (Microcal Software, Northampton, MA, USA) software programs. Unless otherwise noted, statistical significance was determined at P< 0.05, using an unpaired Student’s t test. Results are presented as mean ± SEM and error bars in the figures represent SE. The curves in the figures are drawn to guide the eye unless otherwise noted.

Results

ITTXr-Slow and ITTXr-Per are TTX- and STX-resistant

The majority of small (<25 μm in diameter) and medium-size (25–35 μm in diameter) DRG neurons from adult rat displayed two TTXr sodium currents, ITTXr-Slow and ITTXr-Per, with distinct kinetic properties (Fig. 2A-C). Both of these TTXr currents are known to be highly resistant to TTX: the IC50 of TTX for ITTXr-Slow is ~60 μM [1, 24] and for ITTXr-Per ~40 μM [9]. We asked whether ITTXr-Slow and ITTXr-Per also have similar sensitivities to STX. To examine ITTXr-Slow, cells were held at −80 mV and depolarizing pulses to 0 mV were applied (Fig. 2D). The IC50 value of 6.0±0.9 μM (n=4) for STX-induced block of ITTXr-Slow is in agreement with a previous report investigating Nav1.8 expressed in Xenopus oocytes [1]. To examine the STX sensitivity of ITTXr-Per, cells were held at −120 mV, and test pulses to −50 mV were applied (Fig. 2E). ITTXr-Per displayed resistance to micromolar concentrations of STX (IC50 13.7±3.4 μM, n=5) similar to that of ITTXr-Slow (P>0.05). By contrast, the cardiac sodium channel (Nav1.5) displays an intermediate sensitivity to STX (IC50 =91 nM; 26) and TTX sensitive channels such as Nav1.2 and Nav1.4 display high sensitivities to STX (IC50 ~2 nM; 15, 17). Thus, both of the TTXr sodium currents in DRG neurons display high resistance to STX as well as to TTX, a unique pharmacology not described for any other known sodium channel isoform.

Fig. 2
figure 2

Representative current traces of ITTXr-Slow and ITTXr-Perin DRG neurons. Currents were recorded in presence of 250 nM TTX and activated by 100 ms test pulses from −100 mV to +40 mV in steps of 10 mV. When held at −120 mV, cells exhibited both ITTXr-Slowand ITTXr-Per(A). When the holding potential was reduced to −80 mV, only ITTXr-Slowwas activated (B). ITTXr-Per(C) is revealed by subtraction of ITTXr-Slow(B) from the total current (A). (For functional parameters, see Table 1). (D, E) Representative DRG TTXr current traces from experiments performed with STX (5 μM). Holding the cells at −80 mV, ITTXr-Slowwas activated by test pulses from to 0 mV (D). Holding the cells at −120 mV, ITTXr-Perwas activated by 100 ms test pulses to −50 mV (E). Because of a different voltage-dependence of activation of ITTXr-Slowand ITTXr-Per,this protocol allows isolation of ITTXr-Per. Whereas ITTXr-Peris fully activated by the step to −50 mV, ITTXr-Slowis activated only at more depolarized potentials (Table 1, 9, 28). Both ITTXr-Slowand ITTXr-Perwere resistant to STX

Full size image

The role of the serine residue of Nav1.8 and Nav1.9 in determining the sensitivity to TTX and STX was examined on Nav1.4-WT channels and the mutant channels Nav1.4-Y401S and Nav1.4-Y401C. Robust, fast-inactivating inward currents were recorded from HEK293 cells transfected with all three constructs (Fig. 3). Cells were held at −100 mV and functional parameters were determined. All three constructs displayed nearly identical voltage dependence of activation and steady-state inactivation (Fig. 3D, E; Table 1). To examine the pharmacology, cells were held at −100 mV and depolarizing pulses from −100 mV to 0 mV were applied. Nav1.4-WT was blocked by nanomolar concentrations of both TTX (IC50 =13.5±1.0 nM, n=4) and STX (IC50=1.8±0.2 nM, n= 4) (Fig. 4a, D). In contrast, TTX was much less effective (P< 0.05) at blocking the two mutant constructs Nav1.4-Y401S (Fig. 4b; IC50>100 μM, n=4) and Nav1.4-Y401C (Fig. 4e; IC50=48±13 μM, n=4). Nav1.4-Y401S displayed a ~3000-fold lower sensitivity to STX compared to Nav1.4-WT channels (Fig. 4c; IC50=5.6±0.4 μM, n= 4), similar to the sensitivities determined for ITTXr-Slow and ITTXr-Per. The sensitivity of Nav1.4-Y401C to STX was ~200-fold lower when compared to Nav1.4-WT (Fig. 4f; IC50=0.4±0.06 μM, n= 4), but higher than for Nav1.4-Y401S. Thus, the serine residue in the IP-loop of sodium channels confers a pronounced resistance to both of the guanidinium toxins TTX and STX.

Fig. 3
figure 3

Representative current traces from Nav1.4-WT (n= 9) (A), Nav1.4-Y401S (n= 20) (B) and Nav1.4-Y401C (n= 14) (C) channels expressed in HEK293 cells. Cells were held at −100 mV. Current traces were activated by 100 ms test pulses from −100 mV to +50 mV in steps of 10 mV. All three constructs displayed nearly identical voltage dependence of activation (D). To examine the steady-state inactivation relations, currents were activated by 20 ms test pulses to −10 mV after 500 ms prepulses to potentials from −130 mV to −10 mV. Steady-state inactivation curves were identical for all three constructs (E). (For functional parameters, see Table 1)

Full size image
Fig. 4
figure 4

Representative Nav1.4 current traces from experiments performed with STX and TTX. Holding cells at −100 mV, the Nav1.4-constructs were activated by 100 ms voltage steps to 0 mV. Nav1.4-WT was blocked by nanomolar concentrations of both TTX (A) and STX (D). The two mutant Nav1.4 channels were insensitive to both toxins (Nav1.4-Y401S: B and E; Nav1.4-Y401C: C and F) (For IC50-values, see Table 1)

Full size image

ITTXr-Slow and ITTXr-Per sodium currents have different sensitivities to cadmium

We determined the sensitivity of ITTXr-Slow and ITTXr-Per in DRG neurons to extracellular cadmium, using the same pulse-protocols as described above (Fig. 5 a, b). In the presence of 250-nM TTX in the extracellular solution, which was needed to block all TTX-sensitive currents, ITTXr-Slow (Fig. 5a) displayed a significantly (P< 0.005) higher sensitivity to cadmium (IC50=0.3±0.03 mM, n= 6) compared to ITTXr-Per (Fig. 5b; IC50=1.2±0.2 mM μM, n=10).

Fig. 5
figure 5

Current traces from representative experiments performed with Cd++ . All currents were examined as described in Figures 1 and 4. ITTXr-Slow(A) was more sensitive to block by Cd++ than ITTXr-Per(B). The Cd++ -sensitivities for Nav1.4-WT (C) and Nav1.4-Y401S (D) were similar, but increased for Nav1.4-Y401C (E). (For IC50-values, see Table 1)

Full size image

The blocking effect of extracellular cadmium on the Nav1.4-WT, Nav1.4-Y401C and Nav1.4-Y401S was investigated. One-way ANOVAs were used to test for differences between the cadmium sensitivity of the currents generated by the Nav1.4 constructs, ITTXr-Slow and ITTXr-Per. ANOVA indicated that there were significant differences between the groups. The difference between pairs of individual data sets was tested by Fisher‘s least significant difference (LSD) multiple comparison test and P values <0.05 were considered to be significant. While Nav1.4-WT channels were relatively insensitive to cadmium (Fig. 5c; IC50=1.0±0.1 mM, n=6), Nav1.4-Y401C channels displayed a 100-fold enhanced cadmium sensitivity (Fig. 5e; IC50=10.5±3.0 μM, n=4; P<0.05, Fisher’s LSD). This enhanced sensitivity of the Nav1.4-Y401C construct confirms previous reports that a cysteine in the critical TTX-binding position in the IP-loop has a significant effect on the sensitivity to cadmium [17, 26]. In contrast, the cadmium sensitivity of Nav1.4-Y401S channels (Fig. 5d; IC50=1.1±0.2 mM, it n=4) was not different (P>0.05) from that of Nav1.4-WT channels, showing that the serine residue at this position in the pore does not alter the cadmium sensitivity of this sodium channel isoform. Because the cadmium-induced block of sodium channels is thought to occur through an interaction with pore residues near the TTX-binding site, it is possible that the presence of 250 nM TTX in these experiments might have altered the cadmium IC50 values for one or both of these currents. In order to estimate the influence of TTX on the cadmium-induced block of ITTXr-Slow and ITTXr-Per, the Cd++ sensitivities of Nav1.4-Y401S and Nav1.4-Y401C were also investigated in the presence of 250 nM TTX in the extracellular solution. The cadmium sensitivity of Nav1.4-Y401C was slightly decreased in the presence of TTX (IC50=17.3±5.0 μM, n=5). Similarly, TTX caused a small reduction of cadmium-induced block of Nav1.4-Y401S (IC50=2.1±0.4 mM, n=6). These results indicate that the presence of 250 nM TTX slightly attenuates the resistance of ITTXr-Slow and ITTXr-Per to cadmium. However, irrespective of TTX, pair-wise comparisons revealed that while the cadmium sensitivity of ITTXr-Per was not different from that of Nav1.4-WT or Nav1.4-Y401S currents, the cadmium sensitivities of ITTXr-Slow, ITTXr-Per and Nav1.4-Y401S currents were significantly different from those of Nav1.4-Y401C currents and the cadmium sensitivities of ITTXr-Per and Nav1.4-Y401S currents were significantly different from those of ITTXr-Slow. This indicates that, unlike cysteine in the IP loop, the serine pore residue of Nav1.8 and Nav1.9 does not enhance the cadmium sensitivity of these isoforms. The difference between the cadmium sensitivities of Nav1.8 and Nav1.9 currents is likely to be caused by other pore residues.

Intracellular, but not extracellular, block by QX-314 of ITTXr-Slow and ITTXr-Per

We determined whether or not a serine residue in the pore allows QX-314 to pass through the pore and block ITTXr-Slowor ITTXr-Per. First, QX-314 was applied extracellularly to determine the accessibility of QX-314 to Nav1.8 and Nav1.9. The extracellular solution contained 250 nM TTX in order to block all TTX-sensitive currents. Currents were activated every 30 s for 12 min. During this protocol 1 mM QX-314 induced very little reduction of the ITTXr-Slow peak amplitude (to 97±5%, n=4 with QX-314 versus 101±5%, n=4 under control conditions; P>0.05) (Fig. 6a). The effects of extracellular QX-314 on ITTXr-Per were hard to measure because this current exhibited substantial run-down in the whole-cell mode when activated every 30 s for 12 min without QX-314 (Fig. 6c), making it impossible to determine whether ITTXr-Per is blocked by extracellular QX-314. The ability of QX-314 to block ITTXr-Slow and ITTXr-Per when applied intracellularly was investigated using 200 μM QX-314 in the pipette solution. Experiments were started 3 min after whole-cell mode had been established and currents were activated at 0.75 Hz for 45 s. As shown in Fig. 6b, intracellular QX-314 blocks ITTXr-Slow (to 71±3%, n=5 with QX-314 versus 95±1%, n=5 under control conditions; P<0.001). Intracellular QX-314 also seemed to block ITTXr-Per compared to the control experiments (to 89±7%, n=6 with QX-314 versus 98±10%, n=8 under control conditions) (Fig. 6d). However the ITTXr-Per was relatively unstable under control conditions and the extent of current inhibition was relatively small and not significant (P>0.05).

Fig. 6
figure 6

Block of TTXr currents by extracellular (A,C,E) and intracellular (B,D,F) application of QX-314. (A) ITTXr-Slow was not blocked by extracellular QX-314 (1 mM). Currents were activated with test pulses from −80 mV to 0 mV every 30 s for 12 minutes. (B) ITTXr-Slow was sensitive to intracellular QX-314 (200 μM). Currents were activated after 3 min in the whole-cell mode with test pulses from −80 mV to 0 mV applied at 0.75 Hz for 45 s. (C) ITTXr-Per, activated with test pulses from −120 mV to −50 mV every 30 s for 12 minutes, displayed a substantial rundown under control conditions making it difficult to examine the effect of extracellular QX-314. (D) ITTXr-Per was relatively insensitive to block by intracellular QX-314 (200 μM). Currents were activated after 3 min in the whole-cell mode with test pulses from −120 mV to −50 mV applied at 0.75 Hz for 45 s. (E) Nav1.4-Y401C, but not Nav1.4-WT or Nav1.4-Y401S, channels were blocked by 1mM extracellular QX-314. Currents were evoked with test pulses from −100 mV to 0 mV every 30 s. In contrast, the three constructs were identically blocked by 200 μM intracellular QX-314 (F). Currents were evoked with test pulses from −100 mV to 0 mV at 0.75 Hz

Full size image

To determine the role of a serine at the crucial site of TTX binding on the QX-314 sensitivity of sodium channels, QX-314 experiments were also performed with the Nav1.4-WT, Nav1.4-Y401S and Nav1.4-Y401C constructs. As with ITTXr-Slow and ITTXr-Per, extracellular block of 1 mM QX-314 was determined on currents activated every 30 s for 12 min. In these experiments TTX was not added to the extracellular solution. Whereas the peak current amplitudes of both Nav1.4-WT (97±2%, n=4), and Nav1.4-Y401S (102±6%, n=4) were not significantly reduced by extracellular QX-314, the Nav1.4-Y401C current was reduced by extracellular QX-314 (to 56±8%, n=4) (Fig. 6e). While these data are in good agreement with a recent report from Sunami et al. [29], demonstrating that the cysteine residue allows extracellular QX-314 to block sodium channels, they also show for the first time that a serine residue at this site in the pore does not allow extracellular QX-314 to travel through the pore. We wanted to estimate whether the presence of TTX in extracellular solution might contribute to the insensitivity of ITTXr-Slowto extracellular QX-314. Therefore, the same QX-314 experiments were performed for Nav1.4-Y401C with 250 nM TTX in the extracellular solution. Nav1.4-Y401C was still blocked by extracellular QX-314 (~60%, n=2; data not shown), indicating that the lack of QX-314-induced extracellular block of ITTXr-Slow was not caused by the presence of 250 nM TTX in the extracellular solution. We also asked whether the serine residue in the TTX-binding site plays a role in determining the resistance of ITTXr-Slow and ITTXr-Per to intracellular block by QX-314. As demonstrated in Fig. 6f, all three Nav1.4 constructs were equally blocked by 200 μM intracellular QX314 (Nav1.4-WT: to 41±2%, n=4; Nav1.4-Y401S: to 38±2%, n=4; Nav1.4-Y401C: to 42±2%, n=5). Thus, this pore residue does not seem to interfere with the intracellular binding of QX-314.

One-way ANOVA was used to test for differences between the sensitivity of the currents generated by the Nav1.4 constructs, ITTXr-Slow and ITTXr-Per to intracellular QX-314. ANOVA indicated that there were significant differences between the groups. The difference between pairs of individual data sets was tested by Fisher’s least significant difference (LSD) multiple comparison test and P values <0.05 were considered to be significant. Pair-wise comparisons revealed that although there was no significant difference between the QX-314 sensitivity of the currents generated by the Nav1.4 constructs, ITTXr-Slow and ITTXr-Per were significantly less sensitive to intracellular QX-314. This analysis also indicated that ITTXr-Per is less sensitive to intracellular QX-314 than ITTXr-Slow.

Discussion

The signature of the TTX-binding residue in the IP loop is known to be critical for the sensitivity of sodium channels to several sodium channel blockers [3, 26, 29]. In this study we present novel data showing that a serine at this position in voltage-gated sodium channels 1) imparts a unique high resistance to block by STX and TTX, 2) does not account for the enhanced sensitivity of ITTXr-Slow to cadmium, and 3) does not allow extracellular QX-314 to pass through the pore.

ITTXr-Slowand ITTXr-Per are resistant to guanidinium toxins

When expressed in Xenopus oocytes, Nav1.8 displays TTX- and STX resistance similar to that of ITTXr-Slow from DRG neurons [1, 24, 27]. By contrast, mutant Nav1.8 channels where a phenylalanine was substituted for the serine residue at the critical position in the IP loop produce currents with high sensitivity to TTX and STX [27]. The high resistance of ITTXr-Slow and ITTXr-Per in DRG neurons to block by TTX and STX is a pharmacology not described for any other sodium channel isoform. ITTXr-Slow, ITTXr-Per and Nav1.4-Y401S channels displayed strikingly similar resistance to TTX and STX (Table 1). Because Nav1.8 and Nav1.9 are the only known sodium channel isoforms where the critical TTX-binding residue in the IP loop is serine, our data here comparing the pharmacology of ITTXr-Slow and ITTXr-Per with Nav1.4-Y401S currents are in agreement with previous studies suggesting that Nav1.8 and Nav1.9 are the molecular substrates for ITTXr-Slow and ITTXr-Per [2, 9, 20]. Interestingly, Nav1.9 has been proposed to act as a ligand-mediated sodium rather than as a voltage-gated sodium channel. Together with trkB receptors, Nav1.9 might be the molecular determinant for rapid BDNF or NT4/5 evoked depolarizing sodium currents observed in CNS neurons [5]. This neurotrophin-evoked sodium current was insensitive to TTX but was completely blocked by 10 nM STX. The authors of this study argued that because of the serine in the IP loop, Nav1.9 carries a typical sequence motif for a TTX-resistant/STX-sensitive sodium channel. This is contradicted by the present study and other reports [1, 15, 24, 27] which present clear evidence that the serine pore residue is accompanied with a high resistance to both TTX and STX. It was recently demonstrated that ITTXr-Per is abolished in Nav1.9-null mutant mice [20]. Therefore, we feel it is very likely that Nav1.9 generates the TTX- and STX-resistant persistent current described in the present study, and it is not clear what accounts for the STX-sensitive current attributed to Nav1.9 in the report from Blum et al. [5].

The serine pore residue does not determine the sensitivity to cadmium and QX-314

Sodium channels can be differentially blocked by cadmium [16, 17]. Because of an interaction with the cysteine residue in the IP loop, cadmium is a high-affinity blocker of Nav1.5. This is in agreement with site-directed mutagenesis experiments where cysteine substitution for phenylalanine or tyrosine strongly enhanced cadmium-sensitivities (this study and 3, 15, 26). A previous study on DRG neurons demonstrated that the cadmium sensitivity of slow TTXr currents is higher than the cadmium sensitivity of TTX-sensitive currents [24]. Interestingly, despite the common serine in their pore, ITTXr-Slow was more sensitive to cadmium block than ITTXr-Per. Moreover, we show that the cadmium-sensitivity of Nav1.4-Y401S was not altered compared to Nav1.4-WT channels, suggesting that the serine pore residue of Nav1.8 and Nav1.9 does not control the cadmium sensitivity of these isoforms. The higher cadmium sensitivity described for ITTXr-Slow currents in DRG neurons (24; this study) cannot be explained by the serine pore residue but may be due to the cysteine residue at position 833 (C833) in the IIP loop of Nav1.8. Rat Nav1.9 has an aspartic acid (D740) and Nav1.4 has an asparagine (N739) at the corresponding site. Carbonneau et al. [7] recently reported that a cysteine substitution at a nearby tryptophan residue (W736C) in the amino-terminus of the IP loop of Nav1.4 channels increased the cadmium sensitivity of Nav1.4 (IC50 ~150 μM). Interestingly, while both rat and mouse Nav1.8 have a cysteine residue at position 833, human and dog Nav1.8 have histidines at this location. The sensitivity of these nonrodent Nav1.8 channels to cadmium has not yet been tested.

When compared to TTX-sensitive sodium currents, ITTXr-Slow is relatively resistant to commonly used sodium channel blockers such as lidocaine, mexiletine and carbamazepine [2, 6, 11, 24, 25]. The sensitivity of ITTXr-Per to these agents is not known. Local anesthetics block sodium channels by binding to specific residues in the S6 segments that form part of the inner pore of the channel [19, 22, 32]. The quaternary lidocaine derivative QX-314 is membrane-impermeable and can only reach the inner pore by passing through the pore of the channel [21, 29]. Nav1.5 and cardiac sodium currents are sensitive to extracellular QX-314. Qu et al. [21] reported that a threonine residue in S6 of domain IV of Nav1.5 was an important determinant of sensitivity to extracellular QX-314 [21]. This threonine is conserved in S6 of domain IV of Nav1.8. Sunami et al. [29] reported that the size of the residue at the crucial TTX-binding site in domain I was an even more crucial determinant of sensitivity to extracellular QX-314. They reported that large residues such as phenyalanine and tyrosine prevented QX-314 from passing through the pore, but that alanine, cysteine and even a negatively charged but relatively small amino acid such as aspartate allowed external QX-314 to pass through the pore and block the channel. Since a serine is even smaller than cysteine, we expected that ITTXr-Slow, ITTXr-Per and Nav1.4-Y401S channels would be sensitive to extracellular QX-314. Surprisingly, we report in this study that while Nav1.4-Y401C channels are efficiently blocked by extracellular QX-314, Nav1.4-Y401S channels are not. External QX-314 did not produce significant block of the ITTXr-Slow in DRG neurons. Due to a substantial rundown of ITTXr-Per during electrophysiological recordings under control conditions (see Fig. 6c), it was not possible to examine the effect of external QX-314 on ITTXr-Per. Our finding that the serine pore residue in the IP loop does not allow extracellular QX-314 to pass through the pore helps explain why ITTXr-Slow was poorly blocked by extracellular QX-314. However, both ITTXr-Slow and ITTXr-Per were also relatively poorly blocked by intracellularly applied QX-314 compared to the Nav1.4 constructs. The fact that block of Nav1.4-Y401S by intracellular QX-314 was identical to that of Nav1.4-WT and Nav1.4-Y401C suggests the resistance of ITTXr-Slow and ITTXr-Per to local anesthetics cannot be explained by the serine pore residue in the IP loop of Nav1.8 and Nav1.9. It is not clear what determines this relative resistance of ITTXr-Slow and ITTXr-Per to local anesthetics, but it may be related to changes in residues in the S6 regions involved in local anesthetics binding or to differences in inactivation properties of these currents.

Our data show that ITTXr-Slow and ITTXr-Per display unique pharmacological properties. Functional properties of both ITTXr-Slow and ITTXr-Per suggest that they can play critical roles in action potential electrogenesis of nociceptive neurons. These isoforms are both attractive targets for the development of new analgesic drugs. Therefore, the present results may be valuable for the development of selective blockers of Nav1.8 and Nav1.9.