Sodium channel Nav1.8 produces a slowly inactivating, tetrodotoxin-resistant current, characterized by recovery from inactivation with fast and slow components, and contributes a substantial fraction of the current underlying the depolarizing phase of the action potential of dorsal root ganglion (DRG) neurons. Nav1.8 C-terminus carries a conserved calmodulin-binding isoleucine–glutamine (IQ) motif. We show here that calmodulin coimmunoprecipitates with endogenous Nav1.8 channels from native DRG, suggesting that the two proteins can interact in vivo. Treatment of native DRG neurons with a calmodulin-binding peptide (CBP) reduced the current density of Nav1.8 by nearly 65%, without changing voltage dependency of activation or steady-state inactivation. To investigate the functional role of CaM binding to the IQ motif in the Nav1.8 C-terminus, the IQ dipeptide was substituted by DE; we show that this impairs the binding of CaM to the IQ motif. Mutant Nav1.8IQ/DE channels produce currents with roughly 50% amplitude, but with unchanged voltage dependency of activation and inactivation when expressed in DRG neurons from Nav1.8-null mice. We also show that blocking the interaction of CaM and Nav1.8 using CBP or the IQ/DE substitution causes a buildup of inactivated channels and, in the case of the IQ/DE mutation, stimulation even at a low frequency of 0.1 Hz significantly enhances the frequency-dependent inhibition of the Nav1.8 current. This study presents, for the first time, evidence that calmodulin associates with a sodium channel, Nav1.8, in native neurons, and demonstrates a regulation of Nav1.8 currents that can significantly affect electrogenesis of DRG neurons in which Nav1.8 is normally expressed.
Voltage-gated sodium channels (VGSCs) are major contributors to action potential electrogenesis, and their modulation can regulate the function of excitable cells. Within the VGSC family, Nav1.8 channels are abundantly expressed in small neurons of dorsal root ganglion (DRG) (Akopian et al. 1996), most of which are nociceptive (Djouhri et al. 2003). Recent studies have provided strong evidence for an important role of Nav1.8 in inflammatory and neuropathic pain (Wood and Waxman 2005; Wood et al. 2004).
Nav1.8 produces a slowly inactivating sodium current, characterized by depolarized voltage dependency of activation (Akopian et al. 1996, 1999; Sangameswaran et al. 1996), and rapid recovery from fast inactivation (Cummins and Waxman 1997; Elliott and Elliott 1993). Nav1.8 channels produce the majority of the inward current during the action potential upstroke in the DRG neurons in which they are present (Blair and Bean 2002; Renganathan et al. 2001) and contribute to the adaptation of firing induced by sustained stimulation, such as prolonged exposure to capsaicin (Blair and Bean 2003). Mechanisms that regulate current density and biophysical properties of Nav1.8 are therefore a major focus in the investigations of this channel's role in nociception.
Calmodulin (CaM) binds directly to ion channel target proteins by the isoleucine–glutamine (IQ) motif (Rhoads and Friedberg 1997), in a Ca2+-independent manner (Erickson et al. 2001; Kim et al. 2004; Mori et al. 2000; Xia et al. 1998), and has been shown to regulate current density and biophysical properties of sodium channels (Deschenes et al. 2002; Herzog et al. 2003b; Kim et al. 2004; Tan et al. 2002; Young and Caldwell 2005). The C-terminus of all VGSC contains a conserved IQ motif (Herzog et al. 2003b; Mori et al. 2000; Rhoads and Friedberg 1997). CaM has been shown to regulate current densities and gating properties of several sodium channels (Nav1.2, Nav1.4, Nav1.5, and Nav1.6) but, importantly, this channel modulation is isoform dependent (Deschenes et al. 2002; Herzog et al. 2003b; Kim et al. 2004; Tan et al. 2002; Young and Caldwell 2005). For example, the substitution of IQ with glutamic acid residues (EE) in Nav1.6, which blocks CaM binding, causes a hyperpolarizing shift of steady-state inactivation and a twofold slower fast inactivation, whereas these properties remain unchanged in Nav1.4IQ/EE (Herzog et al. 2003b). Additionally, the effect of CaM binding on Nav1.4 varies in HEK293 and CHO cell lines, suggesting a contribution of cell type–specific factors (Young and Caldwell 2005). Therefore it is important to investigate the effects of CaM on sodium channels in their native neurons.
Genetic assays previously showed that CaM can bind to the C-terminus of Nav1.8 in yeast (Malik-Hall et al. 2003). In this study, we investigated the ability of CaM to bind Nav1.8 by the IQ motif in the channel's C-terminus and to the full-length channel in vivo, and to modulate the properties of this channel in native DRG neurons. We report here for the first time that CaM can associate with Nav1.8 channels in native neurons and show that CaM regulates the current density and the ability of this channel to respond to high-frequency depolarizations in DRG neurons.
Plasmids and antibodies
The plasmid pRK-Nav1.8 was a gift from Dr. John Wood (University College, London) and the CaM bacterial expression plasmid pSGC02, a generous gift from Drs. Geoffrey S. Pitt and Smita Ghosh (Columbia University, New York). The plasmid pEGFP, which encodes green fluorescent protein, was purchased from Clontech (Palo Alto, CA). The I1859Q1860/DE substitution in the IQ motif of Nav1.8 was introduced by site-directed mutagenesis using the QuickChange XL system (Stratagene, La Jolla, CA). Nav1.8-specific polyclonal antibody was purchased from Alamone Labs (Jerusalem, Israel). We verified the specificity of the Nav1.8 antibody on immunoblots of brain, spinal cord, and DRG tissues. A single Nav1.8 immunoreactive protein was observed in DRG samples, which was absent from brain tissue and was reduced in samples of axotomoized DRG compared with control (data not shown). K58/35 pan sodium channel antibody was purchased from Sigma (St. Louis, MO). Monoclonal calmodulin antibody was purchased from Upstate (Lake Placid, NY). Secondary antibody goat anti-mouse HRP conjugated immunoglobins (IgGs) was purchased from Dako (Hamburg, Germany). Rabbit polyclonal anti-6x-His was from Santa Cruz Biotechnology (Santa Cruz, CA).
Lumbar (L2–L6) DRGs were collected from two young adult male Sprague–Dawley rats (250–275 g) and homogenized in 0.5 ml buffer (200 mM NaCl, 20 mM Tris, pH 7.4) supplemented with 1% Triton X-100 and complete protease inhibitor cocktail containing 1 mM EDTA (Roche, Indianapolis, IN). The lysate was incubated on ice for 30 min and the Triton X-100 soluble proteins were collected in the supernatant after centrifugation for 20 min at 15,000 × g at 4°C. The insoluble material was resuspended and again subjected to homogenization and extraction as above. Supernatants were combined and precleared using 40 μl of a 50% slurry of Protein-A agarose (Upstate) for 1 h at room temperature. The DRG-cleared lysate was collected and incubated with 2 μg of Nav1.8-specific polyclonal antibody (Alamone Labs) for 2–3 h before 40 μl of a 50% slurry of Protein-A agarose was added and the mixture continued to incubate for an additional 1.5 h at 4°C. The beads were collected, washed twice in (0.4 ml) each of lysis buffer (except 0.1% Triton X-100 was used), followed by two washes (0.4 ml) each in lysis buffer without Triton X-100. Protease inhibitors and 1 mM EDTA were included throughout the washes. The beads were immediately suspended in 40 μl of 2 × SDS sample buffer and denatured for 30 min at 37°C. All reagents used in protein electrophoresis were designed around the NuPAGE electrophoresis system (Invitrogen, Carlsbad, CA). The samples used for sodium channel analysis were subjected to SDS–PAGE on 4–12% gradient gels (1.5 h at 200 mV) using MOPS-based buffers, whereas the samples used to detect calmodulin were subjected to 12% gels using MES-based buffers. The proteins were transferred to nitrocellulose membranes according to manufacturer's recommendations (2 h at 30 mV). Membranes were blocked in 10% powdered milk overnight at 4°C and washed the following day in TBST (10 mM Tris, pH 7.4, 150 mM NaCl, 0.5% bovine serum albumin, and 0.2% Tween-20) before addition of 1:1,000 primary monoclonal antibodies [monoclonal anti-calmodulin (Upstate) or K58/35 pan sodium channel antibody (Sigma)] in TBST. The membranes were incubated for 2 h in primary antibody and washed extensively for 45 min in TBST and incubated to secondary antibody (1:10,000 goat anti-mouse HRP conjugated immunoglobins; Dako). The membranes were washed extensively in TBST and the immunoreactive proteins were detected using chemiluminescence (NEN-PLUS, Perkin–Elmer Life Sciences, Boston, MA) and autoradiography.
Purification of Nav1.8 C-terminus 6X-His-tagged fusion proteins
A 6X-His fusion protein was constructed from the C-terminus of Nav1.8, amino acids 1724 to 1956 (GenBank accession number U53833) and subcloned in pET-15b vector (Novagen, Madison, WI) using the Nde I and Bam H1 restriction sites. Substitution of the IQ dipeptide by DE in the consensus CaM binding sequence was accomplished using QuickChange XL site-directed mutagensis (Stratagene) and confirmed by sequence analysis. Cotransfection with a CaM expression plasmid pSGC02 increased the production of wild-type channel C-terminus fusion protein as previously reported (Kim et al. 2004). BL21 cells (Stratagene) were transformed in the presence of carbenicillin (100 μg/ml) and chloramphenicol (100 μg/ml) to select for both expression plasmids. Large-scale growths were inoculated using an overnight starter culture (1:100) and grown to an OD595 of 0.6 at 37°C in 2XYT media plus antibiotics. The cells were chilled on ice to <10°C and protein expression induced using 1 mM IPTG. After 4 h at room temperature, the bacteria were collected by centrifugation at 2,500 × g for 10 min, the pellets washed in ice-cold PBS, and frozen at −80°C. Bacterial pellets were resuspended in 20 ml of ice-cold lysis buffer (10 mM Tris, pH 8, 0.1 mM EDTA, 200 mM NaCl) supplemented with protease inhibitor cocktail (Roche) and lysed using a Microfluidics device (Newton, MA). The lysate was incubated on ice for 45 min in 1% Tween-20, briefly sonicated, and centrifuged at 10,000 × g for 20 min to remove insoluble material. The supernatant was diluted with 10 volumes of 50 mM phosphate buffer (pH 8.0 plus 300 mM NaCl and 20 mM imidazole) and applied to NTA-agarose for 2 h at 4°C (Invitrogen). Bound proteins were washed with 10 column volumes of phosphate buffer before elution (250 mM imidazole in phosphate buffer). Positive protein fractions were identified using Bradford and pooled and concentrated using Amicon Ultra4 (10,000 molecular weight cutoff concentrators; Millipore, Bedford, MA). The purified fusion proteins were applied to a phenyl-Sepharose CL-4B in the presence of Ca2+ to remove any CaM that copurified with the fusion proteins (Ohya et al. 1987). The proteins were concentrated and desalted (20 mM Tris, 200 mM NaCl, and 0.1 mM EDTA) using an Amicon Ultra4 concentrator. Fusion proteins were quantified using Bradford (BSA as a standard) and equal protein levels were visually inspected using Coomassie Blue staining.
In vitro binding assays
CaM binding to the purified Nav1.8 C-tail wild-type and IQ/DE mutation was determined using CaM-Sepharose (Stratagene). CaM-Sepharose beads were blocked using 1% BSA and 0.2% I-block (Tropix, Bedford, MA) and equilibrated in binding buffer (20 mM Tris, 200 mM NaCl, 0.1 mM EDTA, 5 mM EGTA, 0.1%Tween-20, and 0.1% BSA). Total CaM binding was measured using a batch-based pull-down assay. Briefly, similar levels of fusion proteins (4 μg) where added to 750 μl of binding buffer (20 mM Tris, 200 mM NaCl, 0.1 mM EDTA, 5 mM EGTA, 0.1% Tween-20, and 0.1% BSA) containing 20 μl of CaM-Sepharose beads (Stratagene). This binding reaction was incubated at room temperature on a rotating platform for 30 min, the beads collected by low-speed centrifugation, and washed twice in 1 ml of binding buffer minus BSA. The bound fusion proteins were eluted from the CaM-Sepharose using 2 × SDS sample buffer and heated to 90°C for 5 min. Electrophoresis was performed using the Novagen gel/transfer system (12% Tris-Bis gels) and MES as the running buffer (Invitrogen). The proteins were transferred to PVDF immobolon membrane (Invitrogen) and the membrane blocked overnight with 5% milk in TBST. After blocking, the membranes were rinsed and exposed to primary rabbit polyclonal anti-6x-His (Santa Cruz,) at 1:1000 in TBST plus 0.1% BSA for 2 h at room temperature. Membranes were washed six times for 10 min each and an anti-rabbit goat secondary antibody-coupled to HRP (Dako) was used at 1:10,000 for 1 h. Membranes were washed as above and the proteins visualized using chemiluminescence (Perkin–Elmer) and autoradiography. CaM binding was quantified using densitometric measurement of band intensity using 1D Image Analysis Software (Eastman Kodak, Rochester, NY). Multiple exposure times ensured linearity in the chemiluminescent intensity. One-way t-tests were performed to identify differences between the means. Comparison analyses were conducted using Origin 6.1 (Microcal Software, Northampton, MA).
DRG cultures follow a protocol described previously (Rizzo et al. 1994). Briefly, adult C57BL/6 mice or Sprague–Dawley rats (1–2 mo old) were decapitated, and L4 and L5 DRG were quickly removed and desheathed in sterile complete saline solution (CSS), pH 7.2. The tissue was then enzymatically digested at 37°C for 25 min with collagenase A (1 mg/ml; Roche) in CSS and for 25 min with collagenase D (1 mg/ml; Roche) and papain (30 U/ml; Worthington, Lakewood, NJ) in CSS at 37°C. The treated tissues were gently centrifuged (100 × g for 3 min), and the pellets were triturated in DRG media (1:1 DMEM/F12, 10% FCS, 100 U/ml penicillin, and 0.1 mg/ml streptomycin) containing 1.5 mg/ml BSA (Fraction V; Sigma) and 1.5 mg/ml trypsin inhibitor (Sigma). Cells were then plated on polyornithine–laminin-coated glass coverslips, flooded with DRG media after 1 h, and incubated at 37°C in a humidified 95% air–5% CO2 incubator.
Electroporation of Nav1.8 channels into Nav1.8-null DRG neurons
DRG neurons were transfected with Nav1.8 or its mutant derivative IQ/DE using the Nucleofector system (Amaxa, Gaithersburg, MD) as described previously (Dib-Hajj et al. 2005). Briefly, adult Nav1.8-null (Nav1.8−/− in C57/BL6 genetic background) mice (Akopian et al. 1999) were decapitated and 10 pairs of DRG were quickly removed and desheathed in sterile calcium and magnesium-free Hank's buffered saline solution (HBSS), pH 7.2. The tissue was treated at 37°C for 50 min with an enzyme mixture of dispase (5 mg/ml; Roche), collagenase A (2 mg/ml; Roche), and DNase I (0.1 mg/ml; Roche) in HBSS, then triturated in DRG media (DMEM supplemented with GlutaMax I, 10% FCS, 2 mM l-glutamine, 16.5 mM NaHCO3, 6 g/l glucose, 100 U/ml penicillin, and 0.1 mg/ml streptomycin). Cells were centrifuged for 5 min at 80 × g and the pellet was resuspended in mouse neuron Nucleofector solution (Amaxa) to a final concentration of 1 × 106 cells/100 μl. WT Nav1.8 or IQ/DE mutant plasmid (10 μg) and 2 μg of EGFP DNA (Clontech) were mixed with DRG cells. Cells were electroporated using the preset program O-003. Cells were then plated on polyornithine–laminin-coated glass coverslips, flooded with DRG media after 1 h, and incubated at 37°C in a humidified 95% air–5% CO2 incubator. Electrophysiological studies were conducted 24–52 h after transfection, and most of the cells that expressed GFP also expressed slowly inactivating tetrodotoxin-resistant (TTX-R) sodium currents. Because these currents are not observed in untransfected Nav1.8-null neurons or Nav1.8-null neurons transfected with GFP alone, this confirmed that most of the cells that expressed GFP also had been successfully cotransfected with the recombinant TTX-R channel constructs.
Whole cell patch-clamp recordings of sodium currents in acutely dissociated mouse (C57/BL6) or rat (Sprague–Dawley) DRG neurons (2–12 h after plating) were performed to test the effect of the calmodulin-binding peptide (CBP, calmodulin-dependent protein kinase II 290–309 peptide) on biophysical properties of the slowly inactivating Nav1.8 TTX-R currents. DRG neurons displayed only short (<10 μm) axonal processes during the brief period of culture, facilitating the voltage clamp. Conventional whole cell patch-clamp recordings were made from small DRG neurons (≤25 μm diameter), using Axopatch 200B amplifiers (Axon Instruments, Foster City, CA). For currents >20 nA, we switched to the 50-MΩ feedback resistor (β of 0.1), which can pass ≤200 nA.
Micropipettes (0.6–0.9 MΩ) were pulled from patch-clamp capillary glass (PG10165-4; World Precision Instruments, Sarasota, FL) with a Flaming Brown P80 micropipette puller (Sutter Instruments, Novato, CA), and polished on a microforge. The average series resistance was 0.82 ± 0.05 MΩ (n = 79). Capacity transients were cancelled using computer-controlled circuitry, and series resistance was compensated (>85%) in all experiments. The pipette solution contained (in mM): 140 CsF, 1 EGTA, 10 NaCl, and 10 HEPES, pH 7.3 (adjusted to 310 mOsm/l with sucrose). To investigate the effect of intracellular free calcium ion on Nav1.8 currents in DRG neurons, the Ca2+-free pipette solution contained (in mM): 140 CsF, 5 BAPTA, 10 NaCl, and 10 HEPES, pH 7.3, and to investigate the effect of high Ca2+ on the Nav1.8 currents, the 5 mM BAPTA in the pipette solution was replaced with 1 mM BAPTA and 1 mM CaCl2 (free calcium concentration ≈ 10 μM). The full Na+ bath solution contained (in mM): 140 NaCl, 3 KCl, 1 MgCl2, 1 CaCl2, 20 TEA-Cl, 10 glucose, and 10 HEPES, pH 7.3 (adjusted to 320 mOsm/l with sucrose). Endogenous Ca2+ currents were blocked by 0.1 mM CdCl2. Tetrodotoxin (TTX, 300 nM) was included in the bath solution to inhibit endogenous tetrodotoxin-sensitive (TTX-S) Na+ currents, which are completely blocked by this toxin concentration. The pipette potential was zeroed before seal formation and voltages were not corrected for liquid junction potential. Leakage current was digitally subtracted on-line using hyperpolarizing control pulses, applied after the test pulse (P/N subtraction). Whole cell currents were filtered at 5 kHz and acquired at 50 kHz using Clampex 8.2 software (Axon Instruments). All recordings were started 10 min after establishing whole cell configuration to allow currents to stabilize and minimize the contamination of residual persistent TTX-R Nav1.9 currents. For current density measurements, membrane currents were normalized to membrane capacitance and calculated as the integral of the transient current in response to a 5-ms hyperpolarizing pulse from the holding potential of −70 to −80 mV. The cell capacitance of the DRG neurons did not vary significantly (P > 0.05) under the experimental conditions of this study (data not shown). All experiments were performed at room temperature (21–25°C).
Experimental protocols and data analysis
Wild-type DRG neurons were held at −70 mV to minimize the contamination of the recordings by the persistent TTX-R Nav1.9 currents (Akopian et al. 1999; Cummins et al. 1999). Activation curves were constructed with the membrane potential held at −70 mV and the application of a series of 40-ms test pulses to voltages that ranged from −70 to +60 mV in 10-mV increments. The peak value of INa at each membrane potential (Vm) was plotted. The relationship of peak INa versus Vm was fitted with the following equation where Vrev is the reversal potential of INa and G is the voltage-dependent sodium current conductance. G was fitted using the following Boltzmann distribution equation where Gmax is the maximum conductance, V1/2 is the membrane potential at half-maximal conductance, and k is the slope factor.
The voltage-dependent steady-state inactivation was estimated by measuring the peak current amplitude elicited by a 40-ms test pulse to −10 mV after a 500-ms prepulse to potential over the range of −80 to +10 mV with a 20-s interpulse period. The normalized curves (I/I0) were fitted using the following Boltzmann distribution equation where I0 is the peak Na+ current at tested pulse measured from the most negative preconditioning pulse potential, Vm is the preconditioning pulse potential, V1/2 is the membrane potential at half-maximal I, and k is the slope factor.
The fast-inactivation decay of Nav1.8 current was fitted with an exponential function where I is the current, A is the percentage of the channels inactivating with time constant τ, t is time, and C is the steady-state asymptote.
An ANOVA was conducted, with a subsequent Student–Newman–Keuls test for identifying statistically significant differences between the mean values of the four experimental conditions (Ca2+-free, high Ca2+, Ca2+-free + CBP, and high Ca2+ + CBP). Because of small sample sizes, we also conducted parallel, nonparametric analyses (a Kruskal–Wallis test with follow-up Mann–Whitney U tests), adjusting for the number of comparisons (n = 6) using the Bonferroni correction factor, and identical conclusions were found (results not shown). Elsewhere, Student's unpaired t-tests were used, as indicated, with the criterion for statistical significance set at 0.05. Descriptive data are presented as means ± SE. Data were analyzed using Clampfit 8.2 software (Axon Instruments) and Origin 6.1 (Microcal Software, Northampton, MA). The ANOVA and post hoc analyses were conducted using SPSS 10.1.3.
Calmodulin coimmunoprecipitates with Nav1.8 channels from rat DRG
Calmodulin (CaM) has been shown to bind to the distal part of the Nav1.8 C-terminus, which carries an IQ motif in a yeast two-hybrid assay (Malik-Hall et al. 2003; our unpublished data). To demonstrate that native Nav1.8 and CaM can associate in DRG neurons, we investigated the ability of anti-Nav1.8 antibodies to coimmunoprecipitate (co-IP) the channel and CaM from rat DRG lysate. Because the CaM-IQ motif interaction is Ca2+ independent (Erickson et al. 2001; Kim et al. 2004; Mori et al. 2000; Xia et al. 1998), we carried out the co-IP experiments using Ca2+-free solutions containing 1 mM EDTA. DRG lysates were incubated with anti-Nav1.8 antibodies, the immunoprecipitated complex was split in half, and the constituent proteins were separated on two different gel systems to optimally resolve the high molecular weight channel proteins and the low molecular weight CaM protein. Pan sodium channel (Fig. 1A) and CaM (Fig. 1B) antibodies were used to detect the presence of sodium channels and CaM in the co-IP precipitate. Polyclonal anti-Nav1.8 antibodies specifically immunoprecipitated a protein that cross-reacted with the monoclonal pan sodium channel antibody (Fig. 1A, lane 1), whereas control IgG antibodies failed to do so (Fig. 1A, lane 2). Similarly, the polyclonal anti-Nav1.8 antibody specifically immunoprecipitated a protein that reacted with the CaM-specific monoclonal antibody (Fig. 1B, lane 1) and co-migrated with purified recombinant CaM (Fig. 1B, lane 3). Together with the results from in vitro binding assays (see following text) and yeast two-hybrid assays showing the interaction of CaM with the C-terminus polypeptide of Nav1.8 (Malik-Hall et al. 2003), our data demonstrate that CaM and native Nav1.8 channels can form a complex in DRG neurons in the absence of Ca2+, in agreement with the view that this interaction is mediated by the apoCaM/IQ-motif binding.
Calmodulin antagonist reduces Nav1.8 currents in small mouse DRG neurons
We used a CaM antagonist to investigate the role of CaM in regulating Nav1.8 currents in native DRG neurons. The CaM-binding peptide (CBP), which is a Ca2+-dependent competitive inhibitor of substrate binding to CaM with subnanomolar affinity (Payne et al. 1988; Waxham et al. 1998), was used in electrophysiological recordings of the Nav1.8 slowly inactivating TTX-R currents from acutely isolated small-diameter mouse DRG neurons (soma diameter ≤25 μm). Data were collected 10 min after establishment of whole cell configuration to allow optimal dialysis of pipette solution into the cytoplasm of the patched cell.
Because CaM cannot bind the CBP inhibitor peptide in the absence of Ca2+ (Hanley et al. 1987; Tan et al. 2002), we first tested the effect of Ca2+ concentration in the pipette solution on native Nav1.8 currents. Figure 2A (top) shows representative families of Nav1.8 currents recorded in Ca2+-free and high Ca2+ pipette solutions. Figure 2B shows that Nav1.8 current density in high (10 μM) Ca2+ concentration (638.51 ± 107.52 pA/pF; n = 13) did not differ from that in Ca2+-free (5 mM BAPTA) solution (669.62 ± 85.42 pA/pF; n = 13). As expected, Nav1.8 current amplitude was not affected by CBP in Ca2+-free solution (Fig. 2B). The current density of Nav1.8 in the presence of CBP in Ca2+-free solution (632.29 ± 125.90 pA/pF; n = 10) did not differ (P > 0.05) from that in Ca2+-free or high Ca2+ buffers. In contrast, the peak current amplitude of Nav1.8 was significantly reduced (P < 0.05) in the presence of CBP in high Ca2+ concentration (Fig. 2A, bottom); the current density of Nav1.8 (Fig. 2B) was reduced by nearly 65% under these conditions to 221.78 ± 70.05 pA/pF (n = 13). Extracellular (bath) application of 10 μM trifluoroperazine (TFP), a cell-permeable CaM antagonist, caused a similar reduction in the peak current of Nav1.8 (data not shown). Thus the Nav1.8 current density in these assays is regulated by CaM binding to the channel in a Ca2+-independent manner, consistent with our IP data (Fig. 1) and with previous reports that suggest that CaM binds to the same determinants in the C-terminus of Nav1.2 channels when Ca2+ free or Ca2+ loaded (Kim et al. 2004).
Reduction in current density of Nav1.8 does not arise from altered voltage-dependent activation or inactivation properties
To determine whether reduced Nav1.8 current density might have resulted from a shift in voltage dependency of the channel, the current–voltage relationship of Nav1.8 was investigated under the three experimental conditions (Fig. 3A). The V1/2 of activation of Nav1.8 channels did not differ (P > 0.05) under the experimental conditions: Ca2+-free solution: −16.51 ± 0.46 mV, k = 6.24 ± 0.32 (n = 14); high Ca2+ solution: −16.48 ± 0.40 mV, k = 5.36 ± 0.30 (n = 18); CBP plus high Ca2+: −15.60 ± 0.43 mV, k = 4.84 ± 0.31 (n = 14). Thus CaM modulation of the current density of Nav1.8 does not appear to result from changes in the voltage dependency of activation of the channels.
Steady-state inactivation (Fig. 3B) was also investigated under the three experimental conditions used in these studies. The V1/2 was measured with 500-ms prepulses to potentials over the range of −80 to +10 mV: Ca2+-free solution: −43.94 ± 0.85 mV, k = 6.46 ± 0.44 (n = 14); high Ca2+ solution: −43.05 ± 0.44 mV, k = 7.36 ± 0.12 (n = 10); CBP plus high Ca2+: −40.53 ± 0.78 mV, k = 7.54 ± 0.38 (n = 8). The V1/2 did not significantly differ (P > 0.05) under these experimental conditions. Thus our results suggest that CaM modulation of the current density of Nav1.8 is not a result of changes in the voltage dependency of steady-state inactivation of the channels.
Kinetic properties of activation and inactivation of Nav1.8 channel in the presence of CBP, with or without 10 μM free calcium, were also investigated. The time-to-peak (Fig. 3C) and decay (Fig. 3D) of Nav1.8 in 10 μM Ca2+ or in the presence of CBP in 10 μM Ca2+ were not different from those in Ca2+-free solution at all tested voltages. These data demonstrate that CaM does not regulate the kinetics or voltage-dependent gating of Nav1.8.
The IQ-motif in the C-terminus of Nav1.8 is important for CaM regulation of current density in DRG neurons
All voltage-gated sodium channels contain a CaM-binding IQ motif: IQXXXRXXXXR in their C-terminus (Herzog et al. 2003b; Mori et al. 2000; Rhoads and Friedberg 1997). CBP, which tightly binds CaM in the presence of Ca2+ and prevents CaM binding to its targets, caused a significant reduction in the current density of Nav1.8 (Fig. 2). To determine whether CaM regulation of Nav1.8 current density is dependent on the direct binding of CaM to the IQ motif in the C-terminus of the channel, we investigated the effect of substituting aspartic acid (D) and glutamic acid (E) for the I1859Q1860 residues, respectively. Substitution of these conserved residues to charged amino acids blocks CaM binding to the IQ motifs of sodium channels (Herzog et al. 2003b; Young and Caldwell 2005) and calcium channels (Zuhlke et al. 1999). Immunoprecipitation of the native Nav1.8 channels from DRG neurons indicated that CaM is closely associated with the Nav1.8 sodium channel (see Fig. 1). Thus impairment of this interaction could lead to changes in the biophysical properties of the Nav1.8 channel in a manner similar to that of the competitive inhibition of this interaction by CBP.
To demonstrate direct binding of CaM to the C-terminus of Nav1.8, we constructed a 6X-His tagged fusion protein spanning amino acids 1724–1956 of the C-terminus of Nav1.8 and a mutant fusion protein whereby the IQ residues at positions 1859/1860 were mutated to D and E, respectively. Wild-type (WT) and IQ/DE recombinant proteins were purified to homogeneity (Fig. 4A). Pull-down binding assays using CaM-Sepharose beads were used to measure CaM binding to these purified proteins. The binding reaction contained 5 mM EGTA to ensure that any interaction would be the result of apoCaM binding (see functional data, Fig. 2). The signal of the bound WT and IQ/DE C-terminus protein was quantitated by densitometry. The average signal from four assays was 371.5 ± 47.0 arbitrary units (AU) for WT C-terminus and 44.4 ± 17.5 (AU) for the IQ/DE mutant C-terminus (Fig. 4, B and C). Mutation of the IQ residues to DE reduced CaM binding to the C-terminus of Nav1.8 compared with WT by more than eightfold (P < 0.001).
We studied the effect of the IQ/DE mutation on functional Nav1.8 channels using patch-clamp methods after expression in Nav1.8−/− DRG neurons. WT and mutant IQ/DE plasmids were electroporated into Nav1.8−/− mouse DRG neurons (Akopian et al. 1999) to record this current in its native environment. DRG neurons of Nav1.8−/− mice completely lack the slowly inactivating TTX-R sodium current produced by Nav1.8 (Akopian et al. 1999) and thus provide a robust expression system for Nav1.8 (Liu et al. 2005; this study) and other sodium channels that have been rendered TTX-R by site-directed mutagenesis, permitting the sodium current of interest to be recorded in isolation (Cummins et al. 2001; Herzog et al. 2003a,b; Rush et al. 2005). Figure 5A shows representative traces of WT and IQ/DE mutant Nav1.8 currents expressed in transfected Nav1.8−/− DRG neurons. Figure 5B shows that the current density of IQ/DE mutant (194.29 ± 36.08 pA/pF; n = 29) was significantly smaller (P < 0.05) than that of WT Nav1.8 channels (376.67 ± 76.65 pA/pF, n = 33). The voltage dependency of channel activation of IQ/DE mutant channels (V1/2 = −24.21 ± 0.30 mV, k = 6.91 ± 0.20; n = 25) did not differ (P > 0.05) from that of WT Nav1.8 channels (V1/2 = −22.29 ± 0.42 mV, k = 6.75 ± 0.30; n = 27) (Fig. 5C).
The voltage dependency of steady-state inactivation of WT and IQ/DE mutant channels was also investigated. As shown in Fig. 5D, the V1/2 of steady-state inactivation curve of the IQ/DE mutant (−42.38 ± 0.46 mV, k = 7.08 ± 0.21; n = 11) and WT (−38.97 ± 0.31 mV, k = 5.76 ± 0.14; n = 14) channels do not differ (P > 0.05), but the slope (k) of IQ/DE mutant was significantly larger (P < 0.05, t-test) than that of the WT channel.
IQ/DE mutant channels exhibit slower recovery from inactivation
To investigate the effect of direct binding of CaM to the IQ motif in the C-terminus of Nav1.8 on recovery from inactivation, we compared recovery from inactivation of WT Nav1.8 and the IQ/DE mutant channels using a double-pulse protocol (Fig. 6). Data were fitted with a double-exponential function as shown in Fig. 6. Figure 6 (bottom) shows recovery from inactivation of wild-type and IQ/DE mutant channels where channels displayed two components of recovery from inactivation: one rapid and the other much slower. Recovery from inactivation of wild-type channels was well fitted by a double-exponential function with fast and slow components of recovery time constants of 5.94 ± 0.88 and 1,930 ± 220 ms, respectively. Similarly, recovery from inactivation of IQ/DE mutant channel was also well fitted by a double-exponential function with fast and slow components of recovery time constants of 6.28 ± 1.06 and 3,650 ± 300 ms, respectively. The slower recovery component of the IQ/DE mutant (3,650 ± 300 ms) was significantly different (P < 0.05, t-test) from that of the wild-type channel (1,930 ± 220 ms), whereas the faster components were not significantly different (P > 0.05). Surprisingly, IQ/DE mutant channels do not fully recover from inactivation even after 10 s of interpulse period of recovery (Fig. 6), suggesting that the change of the slope of steady-state inactivation curve could be caused by accumulation of inactivated channels.
IQ/DE mutant channels demonstrate strong frequency-dependent current reduction
Slow recovery from inactivation of the IQ/DE mutant channels suggested that these channels may undergo frequency-dependent inhibition. Figure 7 shows the frequency-dependent inhibition of WT and IQ/DE mutant channels. Twenty repetitive 100-ms depolarizing pulses to −10 mV, from a holding potential of −70 mV, were applied at 0.1- and 1-Hz frequencies (Fig. 6A). The peak current amplitude of WT Nav1.8 channel at the 20th pulse did not change, compared with that at the first pulse, at a depolarization frequency of 0.1 Hz (0.98 ± 0.03, n = 6), whereas the peak current amplitude at the 20th pulse at 1 Hz decreased to 0.69 ± 0.04 (n = 11) of that of the first pulse. As expected, the peak amplitude of IQ/DE mutant channels at the 20th pulse, compared with the first pulse, was decreased even at 0.1 Hz (0.80 ± 0.03, n = 5), with more inhibition at 1 Hz (0.46 ± 0.05, n = 14). Higher frequencies produced an even greater reduction in the peak current amplitude; compared with the peak current amplitude at the first pulse, the residual peak current amplitude of the IQ/DE mutant channels at the 20th depolarizing pulse at 2 Hz was 0.37 ± 0.05 (n = 8) and at 5 Hz was 0.18 ± 0.08 (n = 5). Current amplitudes of wild-type channels were also reduced at higher frequencies to 0.56 ± 0.08 (n = 8) at 2 Hz and 0.48 ± 0.14 (n = 4) at 5 Hz (Fig. 6B). The frequency-dependent inhibition of the wild-type channels was significantly smaller than that of the IQ/DE mutant channels at all tested frequencies (P < 0.05).
The current density reported in Fig. 2 was measured after a series of depolarizations to construct the I–V curve of Nav1.8 in the presence of CaM antagonists and for the IQ/DE mutant channels. Because the recording protocol could have led to the accumulation of inactivated channels leading to a reduced current density, we further investigated the mechanism underlying the reduction of Nav1.8 current density by CaM antagonists, by measuring the peak current amplitude of the first pulse to −10 mV, 10 min after establishing whole cell configuration. The peak current density of Nav1.8 in the presence of CBP was significantly decreased (P < 0.05, t-test) to 160.31 ± 35.72 pA/pF (n = 15) compared with control conditions (398.48 ± 104.08 pA/pF, n = 12). Although accumulation of inactivated channels may contribute to the reduction in current density, this finding shows that it does not fully account for the entire reduction.
CBP increases the frequency-dependent inhibition of endogenous Nav1.8 currents in native rat DRG neurons
Two populations of slowly inactivating TTX-R currents, TTX-R1 and TTX-R2, were described in rat DRG neurons with different degrees of frequency-dependent inhibition (Rush et al. 1998). We hypothesized that the association, or lack thereof, of Nav1.8 channels and CaM might underlie the two frequency-dependent states. To test this possibility, we investigated the recovery from inactivation and frequency-dependent inhibition of Nav1.8 currents at 1 Hz in rat small-diameter DRG neurons in the absence and the presence of CaM antagonist CBP. Figure 8A shows recovery from inactivation of rat Nav1.8 channels in native DRG neurons. Data were fitted with a double-exponential function as shown in Fig. 6. Recovery from inactivation of Nav1.8 channels in the absence of CBP was well fitted by a double-exponential function with fast and slow components of recovery time constants of 5.2 ± 0.6 and 3,396 ± 259 ms (n = 11), respectively. Recovery from inactivation of Nav1.8 channels in the presence of CBP was also well fitted by a double-exponential function with fast and slow components of recovery time constants of 5.7 ± 0.7 and 4,442 ± 221 ms (n = 10), respectively. Similar to the effect of IQ/DE mutation in recombinant Nav1.8 channels in transfected DRG neurons, the slow component of recovery from inactivation of native Nav1.8 in the presence of CBP was significantly slower than that in the absence of CBP (P < 0.05). Interestingly, the slow component of recovery from inactivation of Nav1.8 channels in native neurons is slower than that in neurons transfected with WT Nav1.8 channel, suggesting that the frequency-dependent inhibition of Nav1.8 channels in acutely isolated native neurons should be stronger than that in transfected neurons.
Nav1.8 channels in native neurons show similar frequency-dependent inhibition in Ca2+-free and high Ca2+ pipette solutions (Fig. 8B) with average residual currents at the 20th pulse of 48.9 ± 3.9% (n = 15) in Ca2+-free and 48.6 ± 4.1% (n = 11) in high Ca2+ solutions, compared with the current amplitude at the first pulse. However, the presence of the CBP peptide in high Ca2+ buffer in the pipette solution significantly reduced (P < 0.02) the average of the frequency-dependent inhibition of the peak current to 33.1 ± 2.8% (n = 8) of the peak current at the first pulse.
Consistent with the differences in the recovery from inactivation of Nav1.8 channels in native and transfected DRG neurons, the frequency-dependent inhibition of Nav1.8 currents in native neurons (49%, Fig. 8B) is much stronger than that of recombinant WT Nav1.8 in transfected neurons (69%, Fig. 7B). One major difference in these assays is time in culture: 2–12 h for the acutely isolated native DRG neurons and 24–52 h for neurons transfected with Nav1.8 channels. To investigate possible effects of time in culture on the frequency-dependent inhibition of Nav1.8 channels in native neurons, we compared this inhibition in acutely isolated native neurons and after 24 h in culture. Acutely tested DRG neurons show a frequency-dependent reduction of the Nav1.8 current at 1-Hz stimulation (0.51 ± 0.32; n = 29), whereas neurons after 24 h in culture show less frequency-dependent inhibition of the current (0.69 ± 0.02, n = 31) (Fig. 8C). Frequency-dependent inhibition of Nav1.8 in native neurons after 24 h in culture (0.69 ± 0.02, n = 31) is not significantly different from that in neurons transfected with recombinant WT Nav1.8 channels (0.69 ± 0.04; Fig. 7B).
This study shows that sodium channel Nav1.8, which contributes most of the current for the action potential (Blair and Bean 2002; Renganathan et al. 2001) in nociceptive DRG neurons where it is normally expressed (Akopian et al. 1996; Black et al. 1996; Djouhri et al. 2003), associates with, and is modulated by, calmodulin (CaM). The voltage dependency and kinetics of activation and inactivation of Nav1.8 were unchanged in the IQ/DE mutant, which impairs CaM binding, or after the application of a CaM antagonist [calmodulin-binding peptide (CBP)]. CBP in the recording pipette causes a reduction in Nav1.8 current density and in the ability of this channel to follow high-frequency depolarizations. Both of the CaM-mediated effects on Nav1.8 are Ca2+ independent, consistent with direct binding of apoCaM to the IQ motif in the C-terminus of the channel.
CaM interacts with the IQ motif in the C-terminus of several sodium channels and was previously shown to influence the properties of Nav1.4, Nav1.5, and Nav1.6 currents in an isoform-specific manner (Deschenes et al. 2002; Herzog et al. 2003b; Tan et al. 2002). Although we were not able to show direct binding of GST-Nav1.8 C-terminus and CaM in our earlier studies (Herzog et al. 2003b), recent studies on the interaction of Nav1.5 C-terminus and CaM (Kim et al. 2004) suggested that misfolding of the bacterially expressed channel C-terminus polypeptides might account for the lack of binding of GST-Nav1.8 and CaM in our previous assays. Adopting the coexpression protocol developed by Kim et al. (2004) as well as a 6X-His tag, a much smaller affinity tag than GST, we observed CaM binding to the C-terminus of Nav1.8 in direct binding assays in this study, in agreement with a previous report of their interaction in a yeast two-hybrid assay (Malik-Hall et al. 2003). This interaction appears to involve apoCaM to the IQ motif in the C-terminus of the channel because it is not Ca2+ dependent, and is significantly reduced by more than eightfold by the substitution of the IQ residues with acidic residues DE. Importantly, we now show in this study that CaM coimmunoprecipitates with endogenous full-length Nav1.8 channels from native DRG neurons and modulates the Nav1.8 current within these cells. These data provide strong evidence that CaM is a functional partner of sodium channels in native neurons.
We investigated regulation of Nav1.8 by CaM in native DRG neurons because cell background has been shown to influence biophysical properties of other sodium channels, such as Nav1.3 and Nav1.6 currents (Cummins et al. 2001; Herzog et al. 2003a). Consistent with a role of cell type–specific regulation of sodium currents, the effects of CaM on Nav1.4 channels was recently shown to significantly differ in HEK293 and CHO cell lines (Young and Caldwell 2005). The use of DRG neurons as an expression system is also important because it permits direct comparison of the effect of CaM on Nav1.8 (this study) and Nav1.4 and Nav1.6 (Herzog et al. 2003b), as discussed in the following text. Thus the effects of CaM on Nav1.8 channels that we describe here are biologically relevant and are predicted to significantly impact the electrogenic properties of DRG neurons.
CaM binding to the C-terminus of sodium channels has been shown to regulate the current densities of several channels. Mutations of IQ motifs of Nav1.4, Nav1.6 (Herzog et al. 2003b), and Nav1.8 (this study) reduce sodium current density to different levels. Although most of the IQ motif in Nav1.4, Nav1.6, and Nav1.8 channels is conserved, it is not identical (Herzog et al. 2003b); thus the differential sensitivity of various channels to the IQ mutations could be caused by altered affinity of CaM binding to the mutant motifs. The N and C lobes of CaM contribute to the binding to the channel; the C lobe has been shown to bind the proximal region of the IQ motif, whereas the N terminal lobe may bind other sequences in the C-terminus, other cytosolic parts of the channel, or other channel protein partners (Young and Caldwell 2005). Thus the local environment of the IQ motif in the C-terminus and perhaps isoform-specific sequences in other cytosolic regions of the channel may influence regulation of sodium channels by CaM.
Nav1.8 (this study) and Nav1.5 (Young and Caldwell 2005) are regulated by CaM in a Ca2+-independent manner. In contrast, Herzog et al. (2003b) showed that CaM regulation of inactivation kinetics of the TTX-S channel Nav1.6 were Ca2+ sensitive. More recently, Young and Caldwell (2005) reported that CaM modulation of inactivation of another TTX-S channel, Nav1.4, was also Ca2+ sensitive. However, neither the current density nor gating properties of Nav1.8 were different under Ca2+-free or 10 μM Ca2+ conditions. Also, frequency-dependent inhibition of Nav1.8 was regulated by CaM but in a Ca2+-independent manner. The lack of Ca2+ modulation of CaM regulation of Nav1.5 and Nav1.8 suggests that TTX-R channels, which are genetically linked, might have evolved a Ca2+-independent form of CaM regulation, unlike that which evolved for the TTX-S channels.
The reduction of the current density of Nav1.8 was not accompanied by significant changes in activation and inactivation properties of the channel. Removal of channels from the cell surface by endocytic pathways might account for the reduction of the current density. Regulation of channel density of Nav1.2 at the cell surface by clathrin-mediated endocytosis has been linked to a conserved motif in the C-terminus of sodium channels including Nav1.8 (Garrido et al. 2001). Recently, Nav1.8 has been shown to interact by the membrane-proximal portion of its C-terminus with a novel protein, CAP-1A, which can act as an adaptor molecule linking Nav1.8 and clathrin (Liu et al. 2005). Overexpression of CAP-1A and Nav1.8 in Nav1.8−/− mouse DRG neurons caused a reduction of the current density by clathrin-mediated endocytosis (Liu et al. 2005). Thus CaM binding to the IQ motif in the membrane-distal portion of the C-terminus might interfere with the binding of components of the endocytic machinery to the channel, resulting in the stabilization of channels at the cell surface, thus increasing the current density. Although removal of channels from the cell surface by endocytosis can underlie a reduction in current density, it does not account for the slowed recovery from inactivation of Nav1.8IQ/DE mutant channel.
Alternatively, the lack of an effect of CaM anatagonists or the IQ/DE mutations on the voltage dependency of activation and inactivation of Nav1.8 suggests that CaM may regulate single-channel properties, leading to a reduction in current density and thus negatively influencing frequency-dependent inhibition of the channel. In agreement with this view, it has recently been reported that DRG neurons contain two populations of Nav1.8-like channels with single-channel conductances of 9.2 and 14.9 pS, and with different frequency-dependent inhibition at 1-Hz stimulation frequency (Cardenas et al. 2004). The CBP-mediated and IQ/DE-mediated reduction in the current density and the CBP-mediated frequency-dependent inhibition of Nav1.8 are consistent with a CaM binding effect on single-channel properties. Additional experiments are needed to determine the contribution of these alternative mechanisms to the current density reduction of Nav1.8, which is caused by blocking CaM binding to the channel.
In addition to the reduction in Nav1.8 current density by >50%, the application of CBP to native DRG, or the introduction of the IQ/DE mutation, hindered the slow component of recovery from inactivation of this channel, an effect that may underlie the greater frequency-dependent inhibition of Nav1.8 that we observed. The net outcome of this change was a dramatic frequency-dependent inhibition of the current even at a frequency as low as 0.1 Hz. Interestingly, two populations of Nav1.8 TTX-R currents with different frequency-dependent inhibition profiles have been described in DRG neurons (Cardenas et al. 2004; Rush et al. 1998). These previous reports and our current data are in agreement with those of Blair and Bean (2003), who reported a 35–50% variability range in use-dependent inhibition of Nav1.8 in native rat DRG neurons. What is remarkable is that the application of CBP to native DRG neurons reduced the variability of frequency-dependent inhibition of Nav1.8 currents, consistent with the possibility of channels in two states: free or bound to CaM.
Nav1.8 contributes most of the current underlying the upstroke of the action potential in the DRG neurons where it is expressed (Blair and Bean 2002; Renganathan et al. 2001) and is essential for repetitive firing of DRG neurons (Renganathan et al. 2001). Recently, Blair and Bean (2003) suggested that slow inactivation of Nav1.8 may limit the response of DRG neurons to prolonged stimulation. We speculate that one possible mechanism for the adaptation of these neurons could be the release of CaM from the channel complex and its sequestration by other CaM-binding proteins on sustained depolarizations of the neuron. The interaction of Nav1.8 and CaM in vivo may modulate the ability of nociceptive DRG neurons where this channel is normally expressed, enabling these cells to respond to high-frequency stimulation such as that occurring after trauma or inflammation. Thus CaM regulation of the Nav1.8 channel's current density, recovery from inactivation and frequency-dependent inhibition can have a significant impact on the electrogenesis of DRG neurons.
This work was supported in part by grants from the National Multiple Sclerosis Society and the Rehabilitation Research Service and Medical Research Service, Department of Veterans Affairs. The Center for Neuroscience and Regeneration Research is a collaboration of the Paralyzed Veterans of America and the United Spinal Association with Yale University.
We thank Dr. Anthony Rush for valuable discussions, L. Tyrrell and R. Blackman for technical assistance, and K. Hudman for statistical advice.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 by the American Physiological Society