Differential Expression of Sodium Channel β Subunits in Dorsal Root Ganglion Sensory Neurons*

Background: Auxiliary β subunits regulate the voltage-gated sodium channels of dorsal root ganglion (DRG) neurons. Results: β subunits are differentially expressed in subpopulations of DRG neurons and regulate Nav1.7 channels in an isoform-specific manner. Conclusion: Differential β subunit expression and isoform-specific regulation have important implications for the sodium currents of DRG neurons. Significance: β subunits are important determinants of sodium channel function and sensory neuron excitability. The small-diameter (<25 μm) and large-diameter (>30 μm) sensory neurons of the dorsal root ganglion (DRG) express distinct combinations of tetrodotoxin sensitive and tetrodotoxin-resistant Na+ channels that underlie the unique electrical properties of these neurons. In vivo, these Na+ channels are formed as complexes of pore-forming α and auxiliary β subunits. The goal of this study was to investigate the expression of β subunits in DRG sensory neurons. Quantitative single-cell RT-PCR revealed that β subunit mRNA is differentially expressed in small (β2 and β3) and large (β1 and β2) DRG neurons. This raises the possibility that β subunit availability and Na+ channel composition and functional regulation may differ in these subpopulations of sensory neurons. To further explore these possibilities, we quantitatively compared the mRNA expression of the β subunit with that of Nav1.7, a TTX-sensitive Na+ channel widely expressed in both small and large DRG neurons. Nav1.7 and β subunit mRNAs were significantly correlated in small (β2 and β3) and large (β1 and β2) DRG neurons, indicating that these subunits are coexpressed in the same populations. Co-immunoprecipitation and immunocytochemistry indicated that Nav1.7 formed stable complexes with the β1–β3 subunits in vivo and that Nav1.7 and β3 co-localized within the plasma membranes of small DRG neurons. Heterologous expression studies showed that β3 induced a hyperpolarizing shift in Nav1.7 activation, whereas β1 produced a depolarizing shift in inactivation and faster recovery. The data indicate that β3 and β1 subunits are preferentially expressed in small and large DRG neurons, respectively, and that these auxiliary subunits differentially regulate the gating properties of Nav1.7 channels.

The small-diameter (<25 m) and large-diameter (>30 m) sensory neurons of the dorsal root ganglion (DRG) express distinct combinations of tetrodotoxin sensitive and tetrodotoxinresistant Na ؉ channels that underlie the unique electrical properties of these neurons. In vivo, these Na ؉ channels are formed as complexes of pore-forming ␣ and auxiliary ␤ subunits. The goal of this study was to investigate the expression of ␤ subunits in DRG sensory neurons. Quantitative single-cell RT-PCR revealed that ␤ subunit mRNA is differentially expressed in small (␤ 2 and ␤ 3 ) and large (␤ 1 and ␤ 2 ) DRG neurons. This raises the possibility that ␤ subunit availability and Na ؉ channel composition and functional regulation may differ in these subpopulations of sensory neurons. To further explore these possibilities, we quantitatively compared the mRNA expression of the ␤ subunit with that of Na v 1.7, a TTX-sensitive Na ؉ channel widely expressed in both small and large DRG neurons. Na v 1.7 and ␤ subunit mRNAs were significantly correlated in small (␤ 2 and ␤ 3 ) and large (␤ 1 and ␤ 2 ) DRG neurons, indicating that these subunits are coexpressed in the same populations. Co-immunoprecipitation and immunocytochemistry indicated that Na v 1.7 formed stable complexes with the ␤ 1 -␤ 3 subunits in vivo and that Na v 1.7 and ␤ 3 co-localized within the plasma membranes of small DRG neurons. Heterologous expression studies showed that ␤ 3 induced a hyperpolarizing shift in Na v 1.7 activation, whereas ␤ 1 produced a depolarizing shift in inactivation and faster recovery. The data indicate that ␤ 3 and ␤ 1 subunits are preferentially expressed in small and large DRG neurons, respectively, and that these auxiliary subunits differentially regulate the gating properties of Na v 1.7 channels.
The sensory neurons of the dorsal root ganglia (DRG) 2 give rise to nerve fibers that convey information about thermal, mechanical, and chemical stimulation from peripheral tissues to the central nervous system. These neurons express a unique combination of tetrodotoxin-sensitive (TTX-S) and tetrodotoxin-resistant (TTX-R) Na ϩ currents that produce the rapid rising phase of the action potentials. Much of what is currently known about Na ϩ channel expression in sensory neurons has been derived from electrophysiological studies of cultured DRG neurons (1)(2)(3). The small-diameter neurons (Ͻ25 m) isolated from the DRG represent the cell bodies of unmyelinated nociceptors and preferentially express TTX-R Na ϩ current, whereas the large-diameter neurons (Ͼ30 m), typically associated with low threshold mechanoreceptors, predominately express TTX-S Na ϩ current. DRG sensory neurons express at least six distinct Na ϩ channel isoforms that display properties similar to the endogenous TTX-S (Na v 1.1, Na v 1.2, Na v 1.6, and Na v 1.7) and TTX-R (Na v 1.8 and Na v 1.9) Na ϩ currents observed in these neurons (4 -7).
In vivo, voltage-gated sodium channels form complexes with auxiliary ␤ subunits that regulate the trafficking, gating properties, and kinetics of the endogenous Na ϩ channels (8 -12). ␤ subunits are relatively small proteins (33-36 kDa) composed of a single membrane-spanning ␣ helix, a short intracellular C terminus, and a large extracellular N terminus incorporating an immunoglobulin-like fold similar to that found in adhesion molecules (8,13). Immunocytochemistry and in situ hybridization indicate that all four isoforms of the ␤ subunit (␤ 1 -␤ 4 ) are expressed in sensory neurons (12,14,15).
In this study, we employed a combination of single-cell RT-PCR, immunocytochemistry, immunoprecipitation, and electrophysiology to further investigate ␤ subunit expression in DRG sensory neurons. The data indicate that small and large DRG neurons express different complements of ␤ subunits. The functional consequences of ␤ subunit expression were evaluated by examining their regulation of Na v 1.7, a TTX-S Na ϩ channel widely expressed in sensory neurons and an important contributor to pain sensation (19,20). The ␤ 3 and ␤ 1 subunits differentially regulated heterologously expressed Na v 1.7 channels. The preferential expression of ␤ subunits in small (␤ 2 and ␤ 3 ) and large (␤ 1 and ␤ 2 ) neurons, coupled with the isoform-specific ␤ subunit regulation of Na v 1.7 activation (␤ 3 ) and inactivation (␤ 1 ), predicts substantial differences in the TTX-S currents of DRG sensory neurons.

EXPERIMENTAL PROCEDURES
Preparation of DRG Neurons-Postnatal day 7 Sprague-Dawley rats (P7) were anesthetized with isoflurane before decapitation, and the DRG were harvested from all accessible levels. The ganglia were incubated for 30 min at 37°C in 2 ml of Hanks' balanced salt solution/HEPES containing 1.5 mg/ml collagenase (Sigma-Aldrich), followed by 1 mg/ml trypsin (Sigma-Aldrich) for an additional 30 min. Trypsin was removed, and the ganglia were transferred to Leibovitz's L-15 medium supplemented with 1% fetal bovine serum (Invitrogen), 2 mM glutamine, 2% penicillin/streptomycin (Invitrogen), and 50 ng/ml nerve growth factor (Sigma-Aldrich). The ganglia were disrupted using fire-polished Pasteur pipettes, and dissociated neurons were plated onto polylysine-coated glass coverslips and placed into 35-mm dishes containing supplemented Leibovitz's medium. Neurons were suitable for single-cell harvesting and electrophysiology for up to 8 h after plating. Animal protocols were approved by the Animal Care and Use Committee of Thomas Jefferson University.
Single-cell RT-PCR-Detailed methods for performing single-cell RT-PCR with dissociated DRG neurons were published recently (7). Small-diameter (Ͻ25 m) and large-diameter (Ͼ30 m) DRG neurons were individually harvested by drawing them into a large bore pipette (30 -50-m diameter) containing sterile bath solution. The neurons were osmotically lysed by 10-fold dilution with sterile water and rapidly frozen. The mRNA present in the cell lysates was reverse-transcribed using random hexamer primers (Stratagene) in a standard 25-l Moloney murine leukemia virus reverse transcription reaction (Fisher). Aliquots of the transcription reaction (1-2 l) were quantitatively analyzed using a SYBR Green reaction mixture on an Mx30005P real-time PCR machine (Agilent Technologies). ␤-Actin was quantitatively measured in each sample and used to normalize for differences in cellular mRNA expression. The absolute number of mRNA copies of each transcript was determined by comparing the threshold cycle (C t ) of the single-cell lysates with known cDNA standards assayed in parallel reactions. PCR primers were designed to span exon/ intron borders to eliminate the detection of genomic DNA, and concentrations (50 -200 nM) were optimized to achieve high amplification efficiency without the formation of primer dimers (Proligo, Sigma-Aldrich). The specificity of the realtime detections was assessed using melting curve analysis, and the identity of the amplified DNA was determined by sequencing.
Na v 1.7 Stable Cell Line-Rat Na v 1.7 cDNA was subcloned into the pcDNA3 expression vector (Invitrogen) and transfected into HEK293 cells using a standard calcium phosphate precipitation method (Invitrogen). After 2 weeks of selection for neomycin resistance (800 g/ml), the remaining colonies were isolated and transferred to separate culture plates for expansion. Na v 1.7 expression was verified using RT-PCR and electrophysiology to measure Na ϩ currents. The HEK293 cell line stably expressing Na v 1.7 was maintained under standard culture conditions in DMEM supplemented with 10% FBS, 2 mM L-glutamine, 100 units/ml penicillin, 10 mg/ml streptomycin, and 400 g/ml neomycin (Invitrogen).
Electrophysiology-Macroscopic Na ϩ currents of HEK293 cells stably expressing the Na v 1.7 channel were recorded using the whole-cell patch clamp technique. The pipette solution contained 5 mM NaCl, 135 mM CsF, 10 mM EGTA, and 10 mM HEPES (pH 7.4). The bath solution contained 150 mM NaCl, 2 mM KCl, 1.5 mM CaCl 2 , 1 mM MgCl 2 , and 10 mM HEPES (pH 7.4). Patch electrodes were fashioned from Corning 8161 borosilicate glass and coated with Sylgard (Dow Corning Corp. ) to minimize pipette capacitance. Recording pipettes had low access resistances (Ͻ1 megohm), and the residual series resistance was 80% compensated. A correction for the liquid junction potential between the pipette and the bath solutions (Ϫ7 mV) was applied to the holding potential before the formation of gigohm seals. After establishing the whole-cell configuration, the cells were dialyzed for 10 min at room temperature (22°C) prior to recording Na ϩ currents. Voltage pulses were generated, and currents were recorded using pCLAMP and an Axopatch 200 amplifier (Molecular Devices). Whole-cell currents were filtered at 5 kHz and digitized at 10 kHz with a Digidata 1440A system (Molecular Devices).
Current-voltage relationships were obtained by plotting the current density (picoamperes/picofarad) versus the test voltage. Normalized Na ϩ conductance (G Na ) was calculated from the peak Na ϩ current (I Na ) at each test potential (V): G Na ϭ I Na / (V Ϫ E Na ), where E Na is the measured Na ϩ ion reversal potential. The steady-state inactivation was determined by normalizing the peak Na ϩ current (I) measured after conditioning prepulses (Ϫ130 to Ϫ10 mV for 500 ms) to the maximal Na ϩ current amplitude (I max ) measured after prepulses to Ϫ140 mV and plotted against the conditioning voltage. The activation and steady-state inactivation were fitted to Boltzmann functions: G/G max (I/I max ) ϭ 1/(1 ϩ exp(V 0.5 Ϫ V)/k v ), where V 0.5 is the midpoint, and k v is the slope factor. The predicted window currents were calculated from the product of the activation and steady-state inactivation curves as described previously (21). Recovery from inactivation was determined using depolarizing prepulses (Ϫ30 mV/20 ms) before returning to Ϫ100 mV for variable intervals (0 -1200 ms). Standard test pulses (Ϫ30 mV/20 ms) were used to assess availability. The recovery time course was fitted to the sum of two exponentials, yielding estimates of the fast ( f ) and slow ( s ) time constants.
Immunoprecipitation and Western Analysis-Rat DRG were harvested and immediately placed in ice-cold Hanks' balanced salt solution. The ganglia were washed with ice-cold Hanks' balanced salt solution and pelleted by low speed centrifugation at 4°C. Hanks' balanced salt solution was replaced with ice-cold lysis buffer (50 mM Tris, 1.0 mM EDTA, 1.0 mM EGTA, 150 mM NaCl, and 1.0% Triton X-100) supplemented with protease inhibitors (Sigma-Aldrich). The samples were homogenized on ice and centrifuged at 15,000 rpm for 20 min at 4°C. The supernatant was recovered and assayed for protein concentration using the Bradford method (Bio-Rad). Lysates (1 mg) were incubated overnight at 4°C in 1 ml of lysis buffer containing either 10 g of control mouse IgG or 10 g of mouse anti-Na v 1.7 monoclonal antibody N68/6 (UC Davis/NIH Neuro-Mab Facility). Anti-Na v 1.7 antibody N68/6 does not cross-react with other Na ϩ channel isoforms or channel proteins extracted from adult rat brain. Protein G-agarose resin (Thermo Scientific) was added (100 l), and the lysates were incubated for 6 h at 4°C before washing with ice-cold lysis buffer. Proteins were eluted from the protein G-agarose by the addition of 50 l of 0.2 M glycine buffer (pH 2.5). The pH was neutralized by adding 10 l of 1 M Tris buffer (pH 9.0), mixed with 3ϫ sample buffer, and separated on 12% SDS-polyacrylamide gels. Proteins were transferred to Protran nitrocellulose membranes (Whatman); blocked with 5% BSA; washed with Tris-buffered saline with 0.1% Tween 20 (TBS/Tween); and incubated overnight with rabbit anti-SCN1B (Cell Applications), rabbit anti-SCN2B (Sigma-Aldrich), or rabbit anti-SCN3B (Abcam) polyclonal antibody in TBS/Tween containing 5% BSA. These commercial antibodies (SCN1B, SCN2B, and SCN3B) are highly specific and do not display cross-reactivity with other members of the ␤ subunit family. The membranes were incubated with HRP-conjugated goat anti-rabbit secondary antibody (Thermo Scientific) for 1 h at room temperature, and labeled proteins were detected by chemiluminescence (Thermo Scientific). We routinely failed to observe Na v 1.7 or ␤ subunit precipitation from cell lysates preincubated with control IgG, further supporting the specificity of the Na v 1.7 immunoprecipitations. The low level expression of the ␤ 4 subunits in DRG neurons (see Fig. 1) combined with the poor quality of available anti-␤ 4 antibodies prevented detailed analysis of this protein.
Immunocytochemistry-Dissociated DRG neurons were plated onto polylysine-coated glass coverslips and fixed in PBS containing 4% paraformaldehyde for 10 min. Cells were permeabilized with 0.1% Triton X-100 in PBS for 5 min before several washes with PBS. Nonspecific antibody binding was reduced by incubating the cells with 5% BSA and 5% goat serum in TBS/Tween for 60 min. Permeabilized cells were incubated with mouse anti-Na v 1.7 monoclonal antibody or rabbit anti-SCN1B, rabbit anti-SCN2B, or rabbit anti-SCN3B polyclonal antibody (1:500 dilution) for 60 min before adding Alexa Fluor 488-conjugated anti-mouse or Alexa Fluor 594-conjugated anti-rabbit fluorescent secondary antibody (Invitrogen) for 60 min . After several washes with PBS, the coverslips were dried overnight and mounted onto glass slides with Mowiol 4.88 (Calbiochem). The slides were imaged on a Zeiss LSM 510 META confocal microscope equipped with FITC and rhodamine filter sets at the Kimmel Cancer Center at Jefferson Medical College.

RESULTS
The expression of ␤ subunits was investigated in acutely dissociated DRG sensory neurons isolated from 7-day-old neonatal rats. Neurons were individually harvested, and the mRNA present in the cell lysates was quantitatively measured (mRNA copies/neuron) using real-time PCR. Fig. 1 compares the expression of the ␤ subunit transcripts in small-diameter (Ͻ25 m) and large-diameter (Ͼ30 m) DRG neurons. The data indicate that small neurons preferentially expressed the ␤ 2 and ␤ 3 isoforms (2000 -4000 copies/neuron). Although ␤ 1 was also detected in these neurons, the mRNA copy number was 5-fold lower (Ͻ400 copies/neuron). This contrasts with large-diameter neurons, which highly expressed ␤ 1 and ␤ 2 mRNAs (Ϸ4500 copies/neuron), whereas ␤ 3 was present at lower levels (Ͻ2000 copies/neuron). The ␤ 4 subunit was expressed at comparatively low levels in both the small (Ͻ500 copies/neuron) and large (Ͻ2000 copies/neuron) neurons. The data indicate that small (␤ 2 and ␤ 3 ) and large (␤ 1 and ␤ 2 ) DRG neurons express different complements of auxiliary ␤ subunits. To investigate the relationship between Na v 1.7 and ␤ subunits, the mRNAs encoding these subunits were quantitatively measured in small and large DRG neurons. Fig. 2 plots the number of Na v 1.7 mRNA copies versus the ␤ subunit mRNA measured from the same neurons. The data were statistically evaluated using Pearson product-moment correlation analysis to determine the strength of mRNA coexpression in these neurons. The Na v 1.7-␤ 2 and Na v 1.7-␤ 3 mRNAs were found to be significantly correlated, with Pearson coefficients (r) of 0.777 and 0.775, respectively (p Ͻ 0.001). Despite the low expression of ␤ 1 mRNA (344 copies/neurons), these subunits were significantly correlated with Na v 1.7 (r ϭ 0.537, p Ͻ 0.01), although the physiological relevance of this association is not clear. The Na v 1.7 and ␤ 4 mRNAs were not associated in these neurons (r ϭ 0.193). These data indicate that ␤ 2 and ␤ 3 subunit transcripts are abundantly expressed in small DRG neurons and are significantly correlated with Na v 1.7 mRNA. Fig. 2 also shows the correlation of Na v 1.7 and ␤ subunit mRNAs in large neurons. Na v 1.7 expression was significantly correlated with the ␤ 1 (r ϭ 0.732) and ␤ 2 (r ϭ 0.680) subunits (p Ͻ 0.001). This contrasted with the ␤ 3 (r ϭ 0.357, p ϭ 0.112) and ␤ 4 (r ϭ 0.342, p ϭ 0.152) subunits, which were not correlated with Na v 1.7. The data indicate that the Na v 1.7, ␤ 1 subunit, and ␤ 2 subunit mRNAs are coexpressed in the same population of large-diameter neurons.
Na v 1.7-␤ interactions were further investigated using co-immunoprecipitation and Western blotting (Fig. 3). Fig. 3A shows a Western blot of DRG homogenates isolated from postnatal day 7 animals probed with the anti-Na v 1.7 antibody. The anti-Na v 1.7 antibody labeled a single high molecular mass protein (Ϸ270 kDa) that is characteristic of Na v 1.7 channels. Immunoprecipitated Na v 1.7 complexes were separated on acrylamide gels, transferred to nitrocellulose membranes, and probed with ␤-specific antibodies. Fig. 3 (B-D) shows that the anti-␤ subunit antibodies labeled low molecular mass proteins (32-34 kDa) that are slightly smaller than what was previously reported for the ␤ 1 (36 kDa) and ␤ 2 (33 kDa) subunits of adult rats (24). ␤ subunits are highly glycosylated proteins containing 30 -36% carbohydrate by weight (24,25). Differences in the carbohydrate content of these subunits account for variations in the molecular mass of the ␤ 1 subunit expressed in skeletal muscle (25). We speculate that the lower molecular masses observed in P7 animals (Ϸ1-2 kDa) may represent partially glycosylated ␤  subunits (26). The immunoprecipitation data show that the ␤ 1 -␤ 3 subunits formed stable complexes with Na v 1.7 channels isolated from the DRG and are therefore candidates for regulating these channels in vivo. Unfortunately, it is impossible to associate the Na v 1.7-␤ interactions detected using immunoblotting techniques with specific subpopulations of small and large DRG neurons.
Potential Na v 1.7-␤ subunit interactions were further investigated by immunocytochemistry. Fig. 4 shows the confocal imaging of small neurons labeled with Na v 1.7-and ␤-specific antibodies. The cytoplasm of these neurons displayed diffuse labeling for Na v 1.7 and the ␤ 1 and ␤ 2 subunits. Merged images revealed some overlap of Na v 1.7 with the ␤ 1 and ␤ 2 subunits, predominately within the intracellular compartment. By contrast, the majority of the ␤ 3 immunofluorescence was localized along the cell periphery, consistent with the labeling of membrane-bound proteins. The merged images display considerable overlap of Na v 1.7 and ␤ 3 around the cell periphery, consistent with the co-localization of these proteins near the plasma membrane.
Initial attempts to investigate the ␤ subunit regulation of endogenous Na v 1.7 channels in dissociated DRG neurons were complicated by the variable expression of Na v 1.7 and ␤ subunits and the presence of multiple overlapping components of TTX-S Na ϩ current in these neurons. We therefore conducted heterologous expression studies to further investigate the ␤ subunit regulation of Na v 1.7 channels. HEK293 cells stably expressing Na v 1.7 were transiently transfected with ␤ subunits. Fig. 5 shows examples of whole-cell Na ϩ currents recorded from cells expressing Na v 1.7 alone or with coexpressed ␤ 1 or ␤ 3 subunits. In the absence of ␤ subunits, the Na v 1.7 channels produced rapidly gating Na ϩ current. Coexpressing ␤ subunits (␤ 1 -␤ 4 ) had no effect on the current kinetics or peak Na ϩ current amplitudes.
To investigate potential changes in voltage-dependent gating, the Na ϩ conductance was calculated from the peak currents and plotted versus the test voltage (Fig. 6A). Coexpressing the ␤ 3 subunit produced a significant hyperpolarizing shift (Ϫ9 mV) in Na v 1.7 activation. Steady-state inactivation was determined using 500-ms prepulses to voltages between Ϫ130 and Ϫ5 mV. ␤ 1 induced a depolarizing shift (ϩ5 mV) in the midpoint of Na v 1.7 inactivation (Fig. 6A). By contrast, coexpressing the ␤ 2 or ␤ 4 subunits did not alter the activation or the steadystate inactivation of the channels.
Recovery from inactivation was determined by applying depolarizing prepulses (Ϫ20 mV/30 ms) before returning to Ϫ100 mV for varying intervals (0 -1200 ms). The recovery time course of Na v 1.7 channels was biexponential, with f and s time constants of 26 and 153 ms, respectively (Fig. 6B). Coexpressing ␤ 1 significantly reduced both f (14 ms) and s (67 ms), consis- tent with a more rapid recovery from inactivation (Fig. 6B). The remaining ␤ subunits (␤ 2 -␤ 4 ) had no effect on recovery from inactivation ( Table 1).
The overlap of activation and steady-state inactivation of Na ϩ channels defines a range of voltages (i.e. window) where Na ϩ channels can be partially activated but are not fully inactivated. Na ϩ channels within this hyperpolarized range of voltages may become persistently activated, resulting in inward Na ϩ currents that could potentially depolarize the resting membrane potential and increase neuronal excitability. ␤ subunit-induced increases in the overlap of Na ϩ channel activation and inactivation tend to expand this window and, consequently, the fraction of persistently activated channels. The ␤ 1 subunit produced a ϩ5-mV depolarizing shift in steady-state inactivation, whereas ␤ 3 produced a Ϫ9-mV shift in Na v 1.7 activation ( Table 1) that could potentially increase the window currents. Fig. 6C shows the predicted window currents of Na v 1.7 channels coexpressed with either the ␤ 1 or ␤ 3 subunits. Despite acting by different mechanisms, the ␤ 1 and ␤ 3 subunits produced similar 2-3-fold increases in the Na v 1.7 window current.
To gain a better understanding of the mechanism of ␤ subunit regulation, chimeras were generated by exchanging the structural domains of the ␤ 1 subunit that shifted steady-state inactivation and accelerated recovery from inactivation with the homologous domains of the ␤ 2 subunit that had no effect on Na v 1.7 gating ( Table 1). The extracellular N-terminal, intracellular C-terminal, and membrane-spanning domains of ␤ 1 were systematically replaced with those of ␤ 2 and transiently expressed in HEK293 cells stably expressing Na v 1.7 channels. Chimeras that retained the extracellular N-terminal domain of ␤ 1 (␤ 112 and ␤ 11⌬ ) fully recapitulated the hyperpolarizing shift in steady-state inactivation and faster recovery observed with the wild-type ␤ 1 subunit (Table 1). Conversely, substitutions that replaced the N terminus of ␤ 1 (␤ 211 and ␤ 221 ) completely abolished Na v 1.7 regulation. C-terminal deletions of the ␤ 1 subunit (␤ 11⌬ ) retained full activity, indicating that the intracellular domain is not essential. The data indicate that the extracellular N-terminal domain of ␤ 1 is critical for the functional regulation of Na v 1.7 channels.

DISCUSSION
The goal of this study was to investigate the expression of auxiliary ␤ subunits in DRG neurons and to characterize the ␤ subunit regulation of Na v 1.7, a TTX-S Na ϩ channel widely expressed in sensory neurons. Single-cell analysis demonstrated that ␤ subunit mRNAs were differentially expressed in small (␤ 2 and ␤ 3 ) and large (␤ 1 and ␤ 2 ) DRG neurons (Fig. 1). Comparisons of Na v 1.7, ␤ 2 , and ␤ 3 mRNAs measured in individual small neurons showed that the expression of these subunits was significantly correlated (Fig. 2), indicating that these transcripts are coexpressed in the same neurons. By contrast, the Na v 1.7 mRNA of large neurons was found to be significantly correlated with the ␤ 1 and ␤ 2 subunits. These data indicate that the Na v 1.7 channels present in small and large DRG neurons are coexpressed with different complements of auxiliary ␤ subunits.
Interactions between Na v 1.7 and ␤ subunits were further explored by co-immunoprecipitation of Na v 1.7 channels. Na v 1.7 coprecipitated with the ␤ 1 -␤ 3 subunits (Fig. 3), indicating that these subunits form stable complexes in vivo. Despite supporting a direct physical interaction between Na v 1.7 and the ␤ 1 -␤ 3 subunits, it is impossible to ascertain the neurons in which these interactions occurred (i.e. small versus large) using immunoprecipitation techniques. However, immunofluorescence imaging showed that Na v 1.7 and ␤ 3 co-localized near the periphery of the small DRG neurons (Fig. 4). Although ␤ 2 subunits are also highly expressed in small neurons, they failed to display obvious co-localization with Na v 1.7 channels. The combination of Na v 1.7-␤ 3 mRNA correlation (Fig. 2), co-immuno-precipitation (Fig. 3), and co-localization near the plasma membrane (Fig. 4) supports the idea that ␤ 3 subunits partner with Na v 1.7 channels. Although these data do not preclude Na v 1.7 interaction with other ␤ subunits, they suggest an important contribution of Na v 1.7-␤ 3 channels to the TTX-S Na ϩ currents of small DRG neurons.
Previous studies of ␤ subunit regulation of heterologously expressed Na v 1.7 channels have produced conflicting data. Initial studies of Na v 1.7 channels expressed in Xenopus oocytes indicated that the ␤ 1 and ␤ 2 subunits failed to alter the expression or gating properties of Na v 1.7, suggesting that these channels may be not regulated by these auxiliary subunits (27,28). Subsequent work, also in oocytes, found that coexpressing ␤ 1 accelerated inactivation and recovery kinetics and produced a hyperpolarizing shift in Na v 1.7 activation (29). The regulation of Na v 1.7 channels by the ␤ 3 and ␤ 4 subunits has not been investigated.
In this study, HEK293 cells stably expressing Na v 1.7 channels were employed to further investigate the functional consequences of Na v 1.7-␤ interactions. Coexpressing ␤ subunits (␤ 1 -␤ 4 ) did not alter the peak Na ϩ current densities or Na v 1.7 current kinetics. However, ␤ 1 produced a depolarizing shift in steady-state inactivation and a faster recovery from inactivation (Table 1). At voltages near the resting membrane potentials of DRG neurons (ϷϪ60 mV), depolarizing shifts in inactivation would tend to increase the fraction of Na v 1.7 channels available to open in response to depolarization. Similar increases in availability along with the associated increase in Na ϩ current density are well known to reduce the threshold for initiating action potentials (30 -32). The rate of Na ϩ channel recovery from inactivation is an important determinant of the absolute and relative refractory periods of action potentials. The faster recovery of Na v 1.7-␤ 1 channels predicts rapid repriming at hyperpolarized voltages that may reduce the duration of the refractory periods, thereby enabling increased firing frequency in large-diameter neurons highly expressing the Na v 1.7-␤ 1 combination.
The ␤ 3 subunit produced a Ϫ9-mV shift in Na v 1.7 activation, causing the channels to open at more hyperpolarized voltages (Table 1). Such shifts in activation and the accompanying increase in Na ϩ current at more hyperpolarized voltages are predicted to increase neuronal excitability and could potentially reduce the threshold for firing action potentials in smalldiameter neurons. This mechanism is consistent with studies showing that Na ϩ channels with low activation thresholds are critical determinants of action potential initiation at the axon initial segment (33,34).
Na v 1.7-␤ subunit interactions that induce hyperpolarizing shifts in activation (␤ 3 ) or depolarizing shifts in inactivation (␤ 1 ) tend to increase the overlap of activation and inactivation gating (Fig. 6C). At voltages within this overlap region, Na ϩ channels are partially activated but not fully inactivated, increasing the potential of persistent window currents (35). At Ϫ50 mV, the peak window current probability predicts that a small percentage (0.1%) of Na v 1.7 channels will be persistently activated. Coexpressing the ␤ 1 or ␤ 3 subunits increased the probability of persistent activation by 2-3-fold. Persistent activation of Na v 1.7-␤ 1 and Na v 1.7-␤ 3 channels and the resulting inward Na ϩ current at resting membrane potentials could depolarize the neuron, leading to increased excitability of DRG neurons. Similar mechanisms are believed to underlie the increased excitability of sensory neurons harboring inherited human pain disorder mutations that produce shifts in Na v 1.7 activation and inactivation of similar polarity and magnitude as those observed for the Na v 1.7-␤ 1 and Na v 1.7-␤ 3 channels (36 -38).
Previous studies have employed chimeras, deletion analysis, and mutations to define the structural domains of ␤ subunits that are critical for Na ϩ channel regulation (21, 39 -42). The findings indicate that the extracellular N-terminal domain of ␤ 1 FIGURE 6. ␤ subunits shift activation and inactivation of Na v 1. 7 channels. A, the normalized conductance was determined from the peak Na ϩ currents and is plotted versus the test potential. Also plotted is the steady-state inactivation obtained using 500-ms prepulses to voltages between Ϫ130 and Ϫ5 mV. The smooth curves are fits of the activation and inactivation data to Boltzmann functions with the parameters listed in Table 1. Data are the means Ϯ S.E. of 14 (Na v 1.7), 26 (␤ 1 ), 9 (␤ 2 ), 21 (␤ 3 ), and 8 (␤ 4 ) determinations. B, Na ϩ channels were inactivated by a brief depolarization (Ϫ30 mV/20 ms), and the recovery time course (0 -1200 ms) was measured at Ϫ100 mV. The smooth curves are biexponential curve fits with the fast and slow time constants listed in Table 1. Data are the means Ϯ S.E. of 15 (Na v 1.7), 22 (␤ 1 ), 10 (␤ 2 ), 17 (␤ 3 ), and 8 (␤ 4 ) determinations. C, window current probabilities predicted from the activation and steady-state inactivation of the Na v 1.7 channels expressed alone or with either the ␤ 1 or ␤ 3 subunits.
is essential for the functional regulation of neuronal and skeletal muscle Na ϩ channels. This contrasts with the ␤ 1 regulation of cardiac Na ϩ channels, where the membrane-spanning domain was found to be critical for the increased expression and accelerated recovery of Na v 1.5 channels (43). These data imply that different structural domains and therefore different molecular interactions are responsible for ␤ 1 regulation of neuronal and cardiac Na ϩ channels. ␤ 1 mRNA is highly expressed in large DRG neurons (Fig. 1), where it is significantly correlated with Na v 1.7, indicating that these subunits are coexpressed in the same population of largediameter neurons (Fig. 2). ␤ subunit chimeras were employed to identify the structural domains of ␤ 1 required to produce the observed depolarizing shift in steady-state inactivation and the accelerated recovery of Na v 1.7 channels (Table 1). Chimeras incorporating the extracellular N-terminal domain of ␤ 1 (␤ 112 ) retained the shift in inactivation and faster recovery, whereas replacing the extracellular domain (␤ 211 ) completely eliminated these effects. These data indicate that the N-terminal domain of the ␤ 1 subunit is required for Na v 1.7 regulation. ␤ 1 subunits with a truncated C terminus (␤ 11⌬ ) retained full functional regulation, indicating that the intracellular domain is nonessential. Interactions between the N terminus of ␤ 1 and extracellular loops of Na v 1.7 may be important for the functional regulation of these channels, similar to what has been described previously for other neuronal Na ϩ channels (40,41).
Recent work employed a similar approach to investigate the ␤ 1 regulation of Na v 1.8, a TTX-R channel that produces the majority of the inward Na ϩ current in small-diameter DRG neurons (21). Substitution of the extracellular N-terminal domain of ␤ 1 had no effect on the expression or gating properties of Na v 1.8 channels. Rather, the intracellular C-terminal domain of ␤ 1 was found to be the critical determinant of Na v 1.8 regulation. These data indicate that the N and C termini of the ␤ 1 subunit differentially regulate the gating properties of Na v 1.7 and Na v 1.8 channels.
Much of what is currently known about ␤ subunit expression in the DRG has been derived from immunocytochemistry and in situ hybridization (4,12,(15)(16)(17)(18)44). These studies indicate that all four isoforms of ␤ subunits (␤ 1 -␤ 4 ) are present in the DRG and that these subunits are differentially expressed in subpopulations of sensory neurons (12,15). ␤ 3 subunits are prom-inently expressed in small and medium neurons, whereas ␤ 1 and ␤ 4 are preferentially expressed in large neurons (4,16,17). ␤ 2 appears to be widely expressed in the DRG and does not show a clear preference for neuronal size (15,45). These findings are in good agreement with our single-cell analysis of gene expression and are consistent with the conclusion that ␤ subunits are differentially expressed in subpopulations of DRG neurons. Unfortunately, histological approaches do not provide quantitative assessments of ␤ subunit expression levels or insight into the functional regulation of Na ϩ channels by ␤ subunits. Our data indicate that the differential expression of ␤ subunits in DRG neurons combined with isoform-specific ␤ subunit regulation of Na v 1.7 activation (␤ 3 ) and inactivation (␤ 1 ) predicts substantial differences in the predominant TTX-S Na ϩ currents of small and large sensory neurons.
Previous work investigated the role of the ␤ 1 and ␤ 2 subunits in sensory neurons using Scn1b and Scn2b null mice (45,46). Whole-cell recordings from DRG neurons isolated from the ␤ 1 knock-outs revealed small changes in the amplitudes and gating properties of TTX-S and TTX-R Na ϩ currents (46). The relatively subtle effects of the Scn1b knock-out on DRG Na ϩ currents coupled with the low level expression of ␤ 1 subunits in small-diameter sensory neurons suggest that these subunits may not be important regulators of the Na ϩ channels expressed in nociceptors. Neurons from the Scn2b null mice displayed reductions in TTX-S Na ϩ current amplitude and Na ϩ channel mRNA and protein (46). Although the underlying mechanism is unclear, the Scn2b knock-out appears to reduce TTX-S Na ϩ currents by decreasing Na ϩ channel mRNA and protein expression. Based on the comparison of Na ϩ currents recorded from control and Scn2b null mice, the ␤ 2 subunits were proposed to increase Na ϩ channel expression (Na v 1.1, Na v 1.6, and Na v 1.7), produce hyperpolarizing shifts in activation, and accelerate the kinetics of the endogenous TTX-S Na ϩ currents (46). These effects were not recapitulated in our heterologous expression studies of Na v 1.7-␤ 2 channels, where no changes in Na ϩ current density, voltage dependence, or current kinetics were observed. Rather, our findings are consistent with previous work showing that the ␤ 2 subunit has no effect on the expression or gating properties of the Na v 1.3, Na v 1.6, and Na v 1.8 channels (21,47). The reasons for the apparent discrepancy between in vivo knockdown and heterologous expression The parameters were obtained from curve fits of Na v 1.7 activation, inactivation, and recovery from inactivation (Fig. 6). The data were tested for significant differences by analysis of variance (p Ͻ 0.001), followed by Dunnett's post hoc test at a significance level of p Ͻ 0.05. For Dunnett's test, the effects of ␤ subunits were compared with values measured for Na v 1.7 channels expressed alone. Data are the means Ϯ S.E. of between 8 and 30 experiments.