Abstract
Na channel NaN (Nav1.9) produces a persistent TTX-resistant (TTX-R) current in small-diameter neurons of dorsal root ganglia (DRG) and trigeminal ganglia. Nav1.9-specific antibodies react in immunoblot assays with a 210 kDa protein from the membrane fractions of adult DRG and trigeminal ganglia. The size of the immunoreactive protein is in close agreement with the predicted Nav1.9 theoretical molecular weight of 201 kDa, suggesting limited glycosylation of this channel in adult tissues. Neonatal rat DRG membrane fractions, however, contain an additional higher molecular weight immunoreactive protein. Reverse transcription-PCR analysis did not show additional longer transcripts that could encode the larger protein. Enzymatic deglycosylation of the membrane preparations converted both immunoreactive proteins into a single faster migrating band, consistent with two states of glycosylation of Nav1.9. The developmental change in the glycosylation state of Nav1.9 is paralleled by a developmental change in the gating of the persistent TTX-R Na+ current attributable to Nav1.9 in native DRG neurons. Whole-cell patch-clamp analysis demonstrates that the midpoint of steady-state inactivation is shifted 7 mV in a hyperpolarized direction in neonatal (postnatal days 0–3) compared with adult DRG neurons, although there is no significant difference in activation. Pretreatment of neonatal DRG neurons with neuraminidase causes an 8 mV depolarizing shift in the midpoint of steady-state inactivation of Nav1.9, making it indistinguishable from that of adult DRG neurons. Our data show that extensive glycosylation of rat Nav1.9 is developmentally regulated and changes a critical property of this channel in native neurons.
- spinal sensory neurons
- ion channel
- tetrodotoxin resistant
- persistent Na current
- desialidation
- voltage clamp
Voltage-gated Na channels are multimers that consist of the pore-forming α-subunit and auxiliary β-subunits (Catterall, 2000). Ten distinct α-subunits have been identified in rat, with homologs from mammalian species, including humans (Goldin et al., 2000; Goldin, 2001). Na channel α-subunits are expressed in a tissue-specific and developmentally specific manner (Beckh et al., 1989; Akopian et al., 1996; Felts et al., 1997; Dib-Hajj et al., 1998; Schaller and Caldwell, 2000).
Many Na channels are heavily glycosylated (Barchi et al., 1980; Miller et al., 1983; Messner and Catterall, 1985; Schmidt and Catterall, 1986,1987). The carbohydrate moiety accounts for 15–30% of the mass of rat brain and skeletal muscle and eel α-subunits (Miller et al., 1983;Messner and Catterall, 1985) but only 5% of rat cardiac α-subunit (Cohen and Barchi, 1993). Cotranslational glycosylation is important for α-subunit folding and interaction with auxiliary subunits (Schmidt and Catterall, 1987). Subsequent post-translational addition of sialic acid accounts for the bulk of the carbohydrate glycocalyx of brain, skeletal muscle, and eel channels (Miller et al., 1983; Messner and Catterall, 1985). Inhibition of sialic acid addition does not affect the assembly of functional channels in the cell membrane (Schmidt and Catterall, 1987); however, such inhibition or enzymatic desialidation modifies the gating properties of Na channels (Recio-Pinto et al., 1990; Bennett et al., 1997; Zhang et al., 1999).
Glycosylation has been shown to modulate the gating properties of eel Na channels reconstituted in lipid bilayers (Recio-Pinto et al., 1990) and recombinant skeletal (Nav1.4) and cardiac (Nav1.5) channels expressed in mammalian cell lines (Bennett et al., 1997; Zhang et al., 1999). Bennett et al. (1997)documented a depolarizing shift of activation and inactivation after desialidation, whereas Zhang et al. (1999) documented a depolarizing shift of activation but a smaller hyperpolarizing shift of inactivation. None of these studies, however, investigated the effect of deglycosylation within native cells. The physiological properties of Na channels can vary, depending on the cell type in which they are expressed, and can differ significantly for a channel expressed in mammalian cell lines versus neurons (Cummins et al., 2001). Voltage-gated K+ channels Kv1.1 and Kv1.2 from brain possess a different glycocalyx from recombinant channels expressed in the COS-1 cell line (Shi and Trimmer, 1999). Thus, the effects of glycosylation of voltage-gated channels in native tissues may differ from those in cell lines.
Nv1.9/NaN is expressed preferentially in small dorsal root ganglion (DRG) and trigeminal ganglion neurons and their axons (Dib-Hajj et al., 1998; Fjell et al., 2000). Nav1.9 channels are resistant to tetrodotoxin (TTX-R) and produce Na currents that are persistent at −70 to −60 mV, with wide overlap between activation and steady-state inactivation (Cummins et al., 1999). As predicted from these properties, Nav1.9 appears to contribute to setting the resting membrane potential and to subthreshold electrogenesis (Herzog et al., 2001).
This study shows that Nav1.9 is present in two glycosylated states at early neonatal ages but in only a lightly glycosylated state in adults. Heavier glycosylation may contribute to a hyperpolarized shift in steady-state inactivation of Nav1.9 in neonatal compared with adult neurons. Pretreatment of cultured neurons by neuraminidase, which removes sialic acid residues, abolishes this shift. These findings document the first case of developmental regulation of glycosylation of a Na channel and demonstrate a functional correlate of differential glycosylation of channels in their native environment.
MATERIALS AND METHODS
Animals. Adult, timed-pregnant and postnatal [postnatal day 0 (P0), P2, P3, P4, P7, and P21] Sprague Dawley rats were used to harvest tissue. Adult Sprague Dawley female rats were used to investigate the effect of axotomy on Nav1.9 protein levels. Axotomy was performed as described previously (Dib-Hajj et al., 1996, 1998). Briefly, Sprague Dawley female rats were anesthetized with ketamine (40 mg/kg, i.p.) and xylazine (2.5 mg/kg, i.p). Sciatic nerves at midthigh level were exposed on the right side, ligated with 4–0 sutures proximal to the pyriform ligament, transected, and placed in a silicon cuff to prevent regeneration (Fitzgerald et al., 1985). Fourteen days after axotomy, the rats were anesthetized, and control (contralateral) and injured (ipsilateral) L4/L5 DRG were removed for analysis. Experiments were performed in accordance with NIH guidelines for the care and use of laboratory animals.
Preparation of membrane fraction. The fourth and fifth lumber DRG (L4 and L5), trigeminal ganglia, spinal cord, and liver tissues were dissected from Sprague Dawley rats and either processed immediately or snap frozen in liquid nitrogen and kept at −80°C for future processing. Tissues were homogenized in a glass dounce in ice-cold lysis buffer at 30 μl/mg tissue. The lysis buffer (0.3m sucrose, 10 mm Tris, pH 8.1, and 2 mm EDTA) was supplemented with protease inhibitors: 1 mm PMSF, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 1 mm DTT, 1 mm benzamidine, 1 mmpepstatin, 8 μg/ml calpain 1, and 8 μg/ml calpain 11). Homogenates were kept on ice for 1 hr before centrifugation at 1000 ×g (low-speed spin) for 7 min at 4°C to remove nuclei and intact cells. The pellet was rehomogenized and spun again under the same conditions. The supernatants from the two low-speed spins were combined and centrifuged at 120,000 × g for 1 hr at 4°C. The pellet, containing the total membrane fraction, was suspended in 0.2 m KCl and 10 mm HEPES, pH 7.4.
To solubilize the membrane fraction, an equal volume of 5% Triton X-100 and 10 mm HEPES, pH 7.4, was added to the sample, and the suspension was kept on ice for 1 hr. The unsolubilized material was pelleted by centrifugation at 10,000 × g for 10 min at 4°C, and the soluble material in the supernatant was collected for additional processing. Protein content was determined using Bio-Rad (Hercules, CA) DC assay for high detergent samples.
Antibodies. An anti-Nav1.9 polyclonal antibody was raised in rabbits against the C-terminal 18 amino acid peptide (CNGDLSSLDVAKVKVHND) and affinity purified over the specific peptide column (Fjell et al., 2000). Anti-Nav1.9 antibody was used at a final concentration of 0.2 μg/ml. A generic Na channel antibody against an 18 amino acid, highly conserved peptide (TEEQKKYYNAMKKLGSKK) in the cytoplasmic loop connecting domains 3 and 4 (L3) of the channel was obtained from Upstate Biotechnology (Lake Placid, NY) and used at a final concentration of 2 μg/ml.
Immunoblot assay. Samples (10–20 μg) were denatured in Laemmli's sample buffer for 20 min at 37°C. Proteins were fractionated by SDS-PAGE using either 5 or 4–15% gradient Tris-HCl Ready gels (Bio-Rad) and then electrotransferred to Immobilon-P membrane (Millipore, Bedford, MA) overnight at 22 mV and 4°C. Blots were blocked with 10% dried milk in TBS for 1 hr at room temperature before incubation for 2 hr at room temperature with the primary antibody diluted in 5% BSA in TBS. Blots were washed extensively in TBST (TBS plus 0.2% Tween 20). Immunoreactive proteins were detected by incubating with a 1:10,000 dilution in 1.25% BSA of a goat anti-rabbit IgG secondary antibody conjugated to horseradish peroxidase (Dako, Glostrup, Denmark) for 1 hr at room temperature. The signal was detected by Renaissance chemiluminescence according to the recommendations of the manufacturer (NEN, Boston, MA).
Deglycosylation of membrane fractions. The membrane fraction of L4 and L5 DRG from three to five rats at different postnatal ages were prepared as described above. The membrane pellet was suspended in 10 μl of 0.2 m KCl and 10 mm HEPES, pH 7.4, SDS was added to a final concentration of 0.3%, and the sample was heated to 37°for 15 min. After cooling on ice, 4 vol of a buffer containing 40 mm sodium phosphate, pH 7.0, 10 mm EDTA, 0.6% Triton X-100, and 1% β-mercaptoethanol were added, followed by 1 μl (500 U) of the N-glycosidase PNGase F (New England Biolabs, Beverly, MA) and incubated at 37°C for 1 hr. This enzyme cleaves the glycosidic bond between the N-acetylglucosamine (GlcNAc) group and the asparagine residue of N-linked glycoproteins. A control sample was prepared without the addition of PNGase. Ten microliters of a 6× sample buffer was added to the reaction, and the sample was denatured and the proteins were fractionated on a gradient gel as described above.
DRG primary cultures. Neonatal and adult animals were decapitated, and L4 and L5 DRG were quickly removed and desheathed in sterile complete saline solution (CSS), pH 7.2. The DRG were then enzymatically digested at 37°C for 20 min with collagenase A (1 mg/ml; Roche, Indianapolis, IN) in CSS and for 15 min with collagenase D (1 mg/ml; Roche) and papain (30 U/ml; Worthington, Lakewood, NJ) in CSS at 37°C. The DRG 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 mg/ml BSA (Fraction V; Sigma, St. Louis, MO) and 1 mg/ml trypsin inhibitor (Sigma). The cells were then plated on poly-ornithine–laminin-coated glass coverslips, flooded with DRG media after 1 hr, and incubated at 37°C in a humidified 95% air–5% CO2 incubator.
Desialidation of Na channels. Neonatal (P0–P3) and adult DRG cultures were enzymatically treated to remove sialic acid residues from the carbohydrate moiety of membrane proteins. Neuraminidase treatment was performed as described previously (Zhang et al., 1999). Briefly, DRG cultures were treated with 0.3 U/ml (800 U/mg protein) neuraminidase type X (Sigma) for 3–5 hr at 37°C before the electrophysiological recordings were made. Drugs were washed out immediately before recording with the bath solution listed below. Sister cultures that were not treated by neuraminidase served as controls.
Electrophysiological recordings. This study focused on the small-diameter C-type DRG neurons that produce the persistent TTX-R Na+ currents attributable to Nav1.9 (Cummins et al., 1999; Dib-Hajj et al., 1999a). Na current properties in DRG neurons were investigated 2–8 hr after plating. DRG neurons displayed only short (<10 μm) axonal processes during the short period of culture, facilitating the voltage clamp.
Coverslips carrying cultured DRG neurons were mounted in a small flow-through chamber on the microscope stage and were continuously perfused with bath external solution (see below) with a push–pull syringe pump (World Precision Instruments, Saratoga, FL). Cells were voltage clamped via the whole-cell configuration with an Axopatch-200B amplifier (Axon Instruments, Foster City, CA) using standard techniques. For currents >20 nA, we switched to the 50 MΩ feedback resistor (β of 0.1), which can pass up to 200 nA. Micropipettes (0.4–0.6 MΩ) were pulled from borosilicate glasses (Boralex) with a Flaming Brown P80 micropipette puller, polished on a microforge, and coated with a mixture prepared of three parts of finely shredded parafilm and one part each of light and heavy mineral oil (Sigma) to reduce the pipette capacitance. The average series resistance was 0.74 ± 0.04 MΩ (n = 72). Capacity transients were cancelled, and series resistance was compensated (90%) as necessary. The pipette solution contained (in mm): 140 CsF, 1 EGTA, 10 NaCl, and 10 HEPES, pH 7.3 (adjusted to 310 mOsm/l with glucose). The bath solution contained (in mm): 140 NaCl, 3 KCl, 1 MgCl2, 1 CaCl2, 0.1 CdCl2, and 20 HEPES, pH 7.3 (adjusted to 320 mOsm/l with glucose). CdCl2 was used to block Ca2+ currents. 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 before the test pulse, of one-sixth test pulse amplitude (−P/6 procedure). Whole-cell currents were filtered at 5 kHz and acquired at 50 kHz using Clampex 8.1 software (Axon Instruments). For current density measurements, membrane currents were normalized to membrane capacitance, which was calculated as the integral of the transient current in response to a brief hyperpolarizing pulse from −120 (holding potential) to −130 mV. All experiments were performed at room temperature (21–25°C).
Separation of slow and persistent TTX-R Na+ currents using prepulse inactivation. Prepulse inactivation takes advantage of the differences in the inactivation properties of the slow and persistent TTX-R Na+ currents (Cummins et al., 1999). TTX at 300 nm was included in the bath solution to isolate slow and persistent TTX-R currents from fast TTX-sensitive (TTX-S) Na+ currents (which are completely blocked by this TTX concentration). Na+ currents were evoked from a holding potential of −130 mV to test pulses ranging from −100 to +60 mV in 5 mV steps. Persistent TTX-R Na+ current was obtained by subtracting the current obtained after a −50 mV prepulse (500 msec duration), which elicits only slow Na+ current, from the current obtained with the more hyperpolarized prepulse (−130 mV), which elicits both slow and persistent TTX-R Na+currents.
Conductance was determined asIp/(VR− V), where Ip is the peak inward current, VR is the reversal potential, and V is the test pulse voltage.VR was determined by fitting the normalized Na+ current voltage (I/Imax) relationships to the following Boltzmann equation: where V1/2 is the voltage for half-maximal activation in millivolts, V is the test pulse voltage, k is the corresponding slope factor, andGi is a scaling factor with the dimensions of a conductance.
Normalized conductance (G/Gmax) was fit with a single Boltzmann relationship of the following form: where V is the test pulse voltage,V1/2 is the voltage for half-maximal activation in millivolts, and k is the corresponding slope factor.
Steady-state voltage-dependent inactivation curves were measured using 500 msec prepulses to the indicated potentials, followed by a test pulse to −50 mV, in which no activation of slow Na+ current occurs. Peak test pulse current was plotted as a function of prepulse potential, normalized, and fit with a single Boltzmann function: where Vpp is the prepulse potential, Vh is the midpoint potential, and kh is the corresponding slope factor.
Data analysis. Patch-clamp data were analyzed using a combination of Clampfit 8.1 software (Axon Instruments) and Origin (Microcal Software Inc., Northampton, MA). Student's unpairedt test was used for statistical analysis of the data. Allp values were significant at the 0.05 level or better. Data are presented as mean ± SEM.
RESULTS
Characterization of rNav1.9-specific antibody
A polyclonal antibody was raised against the C-terminal peptide sequence of Nav1.9, and the affinity-purified antisera was used successfully to localize this channel to rat small-diameter IB4+ DRG sensory neurons (Fjell et al., 2000). We further characterized this antibody in immunoblot assays (Fig. 1). An immunoreactive protein of an apparent molecular weight of ∼210 kDa, which is in close agreement with the predicted molecular weight of Nav1.9 (Dib-Hajj et al., 1998), is detected in membrane fractions of adult rat DRG and trigeminal ganglia but not from liver or spinal cord (Fig. 1A). A smaller protein of ∼100 kDa is detected in most of the samples but is also detectable when the primary antibody was omitted from the assay (data not shown), indicating that it is a nonspecific product. To confirm the presence of proteins in the spinal cord sample, the blot was stripped and reprobed with a generic Na channel antibody (Fig. 1B). Multiple immunoreactive bands are detected in the DRG and trigeminal ganglia (Fig. 1B), consistent with the expression of multiple Na channels in these tissues (Akopian et al., 1996; Black et al., 1996; Dib-Hajj et al., 1998; Kim and Chung, 1999). A major immunoreactive band, which migrates much slower than the 209 kDa marker, is observed in the spinal cord (Fig. 1B), consistent with the presence of brain-type channels (Felts et al., 1997), which are heavily glycosylated (Messner and Catterall, 1985). Immunoblots of membrane fractions of 24 hr DRG cultures show the presence of a single ∼210 kDa immunoreactive band (Fig.1C). The presence of the 210 kDa immunoreactive protein in cultured DRG neurons, at a time in culture when a substantial persistent TTX-R Na+ current attributable to Nav1.9 is detected (Cummins et al., 1999) and immunostaining of small DRG neurons with this antibody has been shown (Fjell et al., 2000; Sleeper et al., 2000), are consistent with the conclusion that it represents Nav1.9 in both the cultured DRG and native tissue. Interestingly, the nonspecific band at ∼100 kDa is not detected in the DRG culture sample, further supporting the conclusion that it is unrelated to Nav1.9.
Transcripts of Nav1.9 as well as the persistent TTX-R Na current that is attributable to this channel are downregulated in DRG neurons after transection of their peripheral projections in the sciatic nerve (Dib-Hajj et al., 1998; Sleeper et al., 2000). Recently, we showed that there is a reduction in the Nav1.9 immunostaining of axotomized DRG neurons using this Nav1.9-specific antibody (Sleeper et al., 2000). We now report a similar finding in an immunoblot assay of DRG tissue 14 d after axotomy, using this Nav1.9-specific antibody. Figure2 shows an immunoblot assay using the membrane fractions of rat DRG tissues contralateral and ipsilateral to axotomy. The 100 kDa band serves as a convenient internal control to demonstrate equal loading of protein per lane. The intensities of the Nav1.9 bands (Fig. 2) were determined by densitometry and show that injury resulted in a 31 and 38% reduction of the Nav1.9 signal in two experiments in which axotomized and control DRG were processed in parallel (see Materials and Methods).
Immunoblot analysis of Nav1.9 reveals developmental regulation of the glycosylation of this channel
Immunoblot analysis of the membrane fraction shows that P3 DRG and trigeminal ganglia contain two immunoreactive bands compared with a single band in adult tissue (Fig.3A). The immunoreactive band in adult tissue comigrates with the smaller of the two bands in the P3 tissue. Because the peptide used in the production of this antibody consists of the C-terminal 18 amino acid residues of Nav1.9, a protein with additional sequence encoding a longer N-terminal polypeptide, interdomain cytoplasmic loops, or an insertion into the C-terminal polypeptide could account for the higher molecular weight immunoreactive protein. To examine these possibilities, we performed reverse transcription-PCR analysis on adult and P3 DRG and trigeminal ganglia templates, which did not show a difference in the length of the amplicons that encode these regions of the channel (data not shown). Therefore, a developmentally regulated post-translational modification of Nav1.9 is the most likely source of the higher molecular weight immunoreactive protein in the P3 tissue.
Membrane fractions from DRG and trigeminal ganglia at different developmental stages were analyzed by a similar immunoblot assay using the Nav1.9-specific antibody. Immunoblot analysis of membrane fraction from DRG shows a 210 kDa immunoreactive protein at embryonic day 17 (E17), in addition to a higher molecular weight protein that becomes prominent at P0 but declines by P7 and is not detectable by P21 (Fig. 3B). A similar pattern is seen in trigeminal ganglia at these developmental stages (Fig.3B).
Na channels from adult brain and skeletal and cardiac muscles are known to be glycosylated. The apparent molecular weight (∼210 kDa) of Nav1.9 is ∼5% higher than the theoretical molecular weight of 201 kDa (Dib-Hajj et al., 1998). The additional mass could be attributable to glycosylation of this channel. We reasoned that the increased mass of the additional immunoreactive protein in the neonatal tissue is attributable to a heavily glycosylated form of Nav1.9. To test this hypothesis, membrane fractions of DRG were treated with the glycosidase enzyme PNGase F and analyzed by immunoblot assay. PNGase F cleaves the glycosidic bond of N-acetylglucosamine that is linked to the aspargine (N) side chain of N-glycosylated proteins. Figure4A shows the effect of the enzymatic treatment of the membrane fractions from P0, P2, and P4 on the size of Nav1.9 protein in these tissues. The enzymatic treatment converted both immunoreactive proteins to a faster migrating band compared with the untreated sample. This clearly shows that even the 210 kDa protein is itself glycosylated. Figure4B shows the mobility shift of the immunoreactive protein after the treatment of adult DRG fractions with PNGase F, also indicating that the adult isoform is lightly glycosylated.
Isolation of Nav1.9 Na+ currents in adult and neonatal DRG neurons
Slow and persistent TTX-R Na+currents produced by Nav1.8/SNS and Nav1.9/NaN channels, respectively (Cummins et al., 1999; Dib-Hajj et al., 1999a), are observed in adult (Fig.5A) and neonatal (Fig.5D) DRG neurons. These two types of currents can be separated by a prepulse protocol (Cummins et al., 1999). The mean peak persistent TTX-R Na+ current density normalized to capacitance for adult and neonatal DRG neurons are 0.79 ± 0.1 (n = 30) and 0.67 ± 0.07 (n = 29) nA/pF, respectively. The decrease in neonatal persistent TTX-R Na+ current density is not statistically significant (p > 0.05). The capacitance of adult and neonatal DRG neurons are 19.55 ± 1.83 and 17.22 ± 1.32 pF, respectively, which are also not significantly different (p > 0.05). The adult persistent TTX-R Na+ current density is consistent with the results obtained in other studies (Cummins et al., 1999, 2000; Renganathan et al., 2000a,b).
Activation and steady-state inactivation of persistent TTX-R Na+ current in adult and neonatal neurons
The current–voltage relationships for the persistent TTX-R Na+ currents in adult (open circles) (n = 30) and neonatal (filled circles) (n = 29) DRG neurons (Fig. 6A) are similar and suggest that the additional glycosylation in neonatal Nav1.9 channels does not have an effect on this relationship. The persistent TTX-R Na+current in adult and neonatal neurons activates between −80 and −70 mV, peaks at approximately −40 mV, and reverses at +50 mV. The midpoint voltages (V1/2) and slope for activation of adult and neonatal persistent TTX-R Na+ currents were obtained from fitting the conductance–voltage curve with the Boltzmann equation (Fig.6B). The V1/2 and slope for activation of adult persistent TTX-R Na+ current are −57.3 ± 1.1 mV and 6.2 ± 0.5 mV/e-fold (open circles) (n = 30) and, for neonatal persistent TTX-R Na+ current, are −54.2 ± 0.9 mV and 6.6 ± 0.4 mV/e-fold (filled circles) (n = 29), respectively. The 3 mV difference in theV1/2 did not reach statistical significance (p > 0.05). These results suggest that the presence of negative charges attributable to sialidation does not influence activation properties of the persistent TTX-R Na+ current.
We next compared the steady-state inactivation of adult and neonatal persistent TTX-R Na+ current (Fig.6C). The V1/2 and slope for steady-state inactivation of the adult TTX-R persistent Na+ current are −48.2 ± 1.3 mV and 6.0 ± 0.3 mV/e-fold (open circles) (n = 30), respectively, whereas those of the neonatal persistent TTX-R Na+ current are −55.2 ± 1.3 mV and 6.5 ± 0.4 mV/e-fold, respectively (filled circles) (n = 29). The V1/2 but not the slopes of neonatal and adult persistent TTX-R Na+currents are significantly different (p < 0.001). Thus, V1/2 of the steady-state inactivation of the persistent TTX-R Na+current of neonatal neurons is shifted by ∼7 mV in the hyperpolarizing direction relative to that of adult DRG neurons. These results suggest that the presence of negative charges attributable to the terminal sialic acid residues on the glycocalyx influences the steady-state inactivation of the neonatal persistent TTX-R Na+ current.
Enzymatic desialidation of neonatal persistent TTX-R Na+ channels does not affect activation
Sialic acid is a prominent constituent of glycosylation of neuronal and muscle Na channels (Miller et al., 1983; Messner and Catterall, 1985; Roberts and Barchi, 1987). Therefore, we hypothesized that the additional glycosylation of neonatal Nav1.9 channel protein might carry extra negative charges. We investigated the effect of different degrees of glycosylation on the voltage dependence of Nav1.9 channel gating by comparing the voltage dependence of activation and steady-state inactivation in neonatal and adult neurons. To test whether sialic acid residues affect the voltage dependence of Nav1.9 gating, we pretreated DRG neurons in culture with neuraminidase to remove the sialic acid from Na channels. Removal of negatively charged sugar residues on the extracellular face of Na channels by desialidation reduces the surface potential and has been shown to affect Nav1.4 and Nav1.5 gating (Bennett et al., 1997; Zhang et al., 1999).
The current–voltage relationship for the persistent TTX-R Na+ current for neuraminidase-treated neonatal DRG neurons is shown in Figure 6A(filled squares) (n = 15), and it is similar (p > 0.05) to that for control nontreated neonatal DRG neurons (filled circles) (n = 36). The persistent TTX-R Na+ current density and capacitance of neonatal DRG neurons after the neuraminidase treatment are 0.63 ± 0.08 nA/pF and 15.34 ± 1.3 pF, respectively. The current density and cell capacitance are similar between treated and nontreated neonatal control neurons, suggesting that desialidation does not affect the production of functional, membrane-bound Nav1.9 channel complexes.
The activation curves for neuraminidase-treated neonatal DRG neurons are shown in Figure 6B (filled squares) (n = 15). The meanV1/2 values and slopes obtained from Boltzmann fits of steady-state activation data are −57.5 ± 2.2 mV and 5.84 ± 0.8 mV/e-fold, respectively, and are similar to the values −54.2 ± 1.2 mV and 6.62 ± 0.4 mV/e-fold obtained for nontreated neonatal DRG neurons (Table 1). These results further validate the conclusion (see above) that the increased glycosylation of neonatal Nav1.9 does not affect the steady-state activation properties of the persistent TTX-R Na+ current in native neurons.
Enzymatic desialidation of neonatal persistent TTX-R Na+ channels causes a depolarizing shift in steady-state inactivation
Our experiments show that steady-state inactivation of the persistent TTX-R Na+ current in neonatal neurons is shifted by ∼7 mV in a hyperpolarizing direction compared with adult neurons, possibly attributable to the more extensive glycosylation of Nav1.9 channels. To test this hypothesis, we examined the effects of neuraminidase treatment on the voltage dependence of steady-state inactivation of the persistent TTX-R Na+ current in neonatal DRG neurons. The mean V1/2 values and slopes obtained from neonatal DRG neurons after enzymatic desialidation were −47.1 ± 1.0 mV and 5.0 ± 0.6 mV/e-fold, respectively. Enzymatic treatment produced a significant (p < 0.001) depolarizing shift (8 mV) in theV1/2 for steady-state inactivation of the persistent TTX-R Na+ current (Fig.6C, filled squares; Table 1). Desialidation, however, had no significant effects (p > 0.05) on the slope for steady-state inactivation of neonatal TTX-R persistent Na+ current.
Enzymatic desialidation does not shift steady-state activation and inactivation of the persistent TTX-R Na+ current in adult DRG neurons
Because adult Nav1.9 channels are glycosylated but to a lesser extent than the neonatal channel, we wanted to test whether desialidation affects channel gating. Neuraminidase treatment of adult DRG cultures did not result in a significant change of the V1/2 and the slope values for activation and steady-state inactivation (n = 16) (Fig. 6B,C, open squares) from the values obtained in control neurons (Fig.6B,C, open circles; Table 1). The persistent TTX-R Na+ current density and capacitance of the adult DRG neurons after the neuraminidase treatment are 0.62 ± 0.1 nA/pF and 21.14 ± 1.6 pF, respectively, which are similar to those obtained in adult control neurons.
DISCUSSION
We show in this study that Na channel Nav1.9 exists in two glycosylated states during late embryonic and early postnatal stages but only in the less glycosylated state subsequently. Using patch clamp, we show that the current density, current–voltage relationship, and voltage dependence of activation of the persistent TTX-R Na+ current are similar in neonatal and adult DRG neurons. The voltage dependence of steady-state inactivation, however, is hyperpolarized by ∼7 mV in neonatal compared with adult DRG neurons. The difference in steady-state inactivation of the persistent TTX-R Na+ current is attributable to sialidation of the Nav1.9 channel at neonatal stages. Enzymatic desialidation of neonatal (P0–P3) DRG neurons converts the persistent TTX-R Na+ current to the adult type.
Adult DRG neurons express six Na channels (Black et al., 1996; Dib-Hajj et al., 1998), including the TTX-R Nav1.8/SNS (Akopian et al., 1996) and Nav1.9/NaN (Dib-Hajj et al., 1998). A number of TTX-S Na channels undergo developmentally regulated and mutually exclusive alternative splicing of exon 5 without a change in the protein size (Sarao et al., 1991; Gustafson et al., 1993; Belcher et al., 1995; Plummer et al., 1997) or of exon 18, which produces a truncated two-domain protein (Plummer et al., 1997). In contrast, Scn11a, the gene encoding Nav1.9, does not contain alternative exons 5N and 18N (Dib-Hajj et al., 1999b). Neonatal and adult Nav1.9 transcripts encoding the variable regions of this channel, which include the N and C termini and interdomain cytoplasmic loops, have similar lengths and sequences. The presence of two Nav1.9 immunoreactive proteins, therefore, indicates post-translational modification of the channel.
We show in this study that Nav1.9 channel undergoes developmentally regulated post-translational processing. Na channels from rat brain and skeletal and cardiac muscle are glycosylated (Barchi et al., 1980; Miller et al., 1983; Messner and Catterall, 1985; Schmidt and Catterall, 1986, 1987; Cohen and Barchi, 1993). Immunoblot analysis shows that Nav1.9 is present in two glycosylated forms around and shortly after birth. The observed molecular weight of Nav1.9 in adult neurons is ∼5% higher than the predicted molecular weight of 201 kDa (Dib-Hajj et al., 1998). The extra mass of adult Nav1.9 is attributable to limited glycosylation that is evident by the change in its mobility after PNGase treatment. The exact nature of the sugar group on this protein is not known, and, if sialic acid is present, it does not produce a detectable effect on gating of Nav1.9 in adult neurons.
The ∼210 kDa protein is the only detectable form of the channel in adult DRG neurons, and an immunoreactive protein of similar size is also present in neonatal tissue. Limited glycosylation of Nav1.9 in adult tissue may modulate the stability and surface expression of the channel complex. Cotranslational glycosylation has been shown to be important for the proper folding, subsequent channel modification, and interaction with auxiliary subunits of rat brain Na channel α-subunits (Schmidt and Catterall, 1987). The heavier neonatal form of Nav1.9 indicates more extensive processing of the glycocalyx. Sialic acid residues account for the bulk of the sugar content of eel and rat brain and skeletal muscle Na channels (Miller et al., 1983; Messner and Catterall, 1985). Metabolic inhibition of sialic acid addition or enzymatic desialidation does not affect the assembly or surface expression of rat brain Na channels (Schmidt and Catterall, 1987); rather, it modifies their gating properties (Recio-Pinto et al., 1990;Bennett et al., 1997; Zhang et al., 1999). Consistent with the presence of sialic acid residues in the neonatal form of Nav1.9 is the finding that enzymatic desialidation of neonatal DRG cultures changes the properties of the persistent TTX-R Na+ current.
Heavier glycosylation of neonatal Nav1.9 is consistent with the presence of sialic acid that affects gating properties of this channel. Neonatal persistent TTX-R Na+ current inactivates at a more hyperpolarized potential compared with that in adult neurons. Treatment of DRG cultures with neuraminidase shifts steady-state inactivation of persistent TTX-R Na+ current by ∼8 mV in a depolarized direction in neonatal but not adult neurons. The effect of neuraminidase treatment indicates that sialic acid contributes to the heavier glycosylation of the neonatal channel and is responsible for the shift in inactivation. The observation that desialidation does not shift the voltage dependence of steady-state inactivation for adult persistent TTX-R Na+ current is consistent with the limited glycosylation of adult Nav1.9, including a reduced sialic acid component. Alternatively, sialic acid residues may be present in the Nav1.9 Na channels but in a locus that is not proximal to the inactivation gate.
Desialidation did not cause a depolarizing shift in activation of the neonatal persistent TTX-R Na+ current. In contrast, previous studies on rat Nav1.4 and human Nav1.5 showed that neuraminidase pretreatment caused a depolarizing shift (∼10 mV) ofV1/2 for activation of these channels (Bennett et al., 1997; Zhang et al., 1999). The reasons for this discrepancy are not clear. These previous studies reported different effects of desialidation on the steady-state inactivation of Nav1.4. Whereas Bennett et al. (1997) reported an ∼10 mV depolarizing shift in the steady-state inactivation of Nav1.4, Zhang et al. (1999) reported an ∼4 mV hyperpolarizing shift. Desialidation of human Nav1.5 expressed in HEK 293 cell lines and mouse Nav1.5 in native cardiac myocytes produces different effects on the voltage dependence of steady-state inactivation (Zhang et al., 1999; Ufret-Vincenty et al., 2001); desialidation causes an ∼10 mV depolarizing shift in human Nav1.5 but an ∼8 mV hyperpolarizing shift in mouse Nav1.5. These differences are likely attributable to the cell background rather than a species difference.Cummins et al. (2001) demonstrated that the electrophysiological properties of a Na channel can differ depending on the cell type in which it is expressed. The disagreement between the results of Bennett et al. (1997), Zhang et al. (1999), and Ufret-Vincenty et al. (2001)points to the potential risk of extrapolating results from cell lines to expression in the native environment.
Alternatively, the inability of desialidation to change voltage dependence of activation of Nav1.9 may reflect isoform-specific differences. Although Nav1.9 and Nav1.5 have a similar distribution of N-glycosylation motifs, their gating properties are significantly different (Cummins et al., 1999). Changing external calcium concentration causes nonequivalent shifts on the voltage dependence of channel gating (Frankenhaeuser and Hodgkin, 1957). If sialic acid residues are heterogeneously distributed and if different channel gating domains lie in different parts of the resulting nonuniform transmembrane electric field, differential effects on particular gating behaviors may occur that could depend on the degree or pattern of sialic acid in the glycocalyx. Negative charges in the vicinity of different S4 voltage sensors may have a differential effect on the gating of the channel. Reduced positive charges in the S4 segments of domains II and III of Nav1.9 may underlie the significantly different gating properties of this channel compared with Nav1.4 and Nav1.5 (Cummins et al., 1999; Dib-Hajj et al., 1999a). Also, whereas Nav1.4 currents can be modeled by a Hodgkin–Huxley type equation using three activation particles and one inactivation particle, Nav1.9 currents are best fit with one activation and one inactivation particle (Herzog et al., 2001). It is not unreasonable, therefore, to suggest that desialidation will have a different effect if sialic residues in neonatal Nav1.9 are clustered around the inactivation gate but not the activation gate.
Nav1.9 produces a persistent TTX-R current with wide overlap between activation and steady-state inactivation (Cummins et al., 1999; Dib-Hajj et al., 1999a). Our biophysical studies suggest that, as a result of its voltage dependence and persistent kinetics, Nav1.9 plays an important role in setting the membrane resting potential and in subthreshold electrogenesis (Herzog et al., 2001). The present results demonstrate that the heavily glycosylated neonatal channel inactivates at more hyperpolarized potentials compared with adult channels. Because neonatal Nav1.9 has less of an overlap between its activation and inactivation, neonatal Nav1.9 might be expected to contribute a smaller depolarizing influence on resting potential and a smaller amplification of depolarizing inputs. Different properties of neonatal and adult Nav1.9 channels may thus cause sensory neurons to respond differently to similar stimuli.
Footnotes
This work was supported in part by grants from the National Multiple Sclerosis Society and the Rehabilitation Research and Development Service and Medical Research Services, Department of Veterans Affairs, and by gifts from the Paralyzed Veterans of America and Eastern Paralyzed Veterans Association. We also thank the Blinded Veterans of America for their support. We thank Dr. Joel A. Black and William N. Hormuzdiar for providing tissues and cultures, Dr. Ted Cummins for helpful discussions, and Bart Toftness for technical assistance.
Correspondence should be addressed to Dr. Sulayman D. Dib-Hajj, Paralyzed Veterans of America/Eastern Paralyzed Veterans Association Neuroscience Research Center (127A), Veterans Administration Medical Center, Building 34, 950 Campbell Avenue, West Haven, CT 06516. E-mail:sulayman.dib-hajj{at}yale.edu.