Neuromedin U Type 1 Receptor Stimulation of A-type K+ Current Requires the βγ Subunits of Go Protein, Protein Kinase A, and Extracellular Signal-regulated Kinase 1/2 (ERK1/2) in Sensory Neurons*

Background: Regulation of IA has an important role in determining the electrical properties of dorsal root ganglion neurons. Results: Neuromedin U (NMU) increases IA via NMUR1 that couples to the βγ subunits of Go protein, PKA, and ERK. Conclusion: NMUR1 stimulation of IA contributes to neuronal hypoexcitability. Significance: Selective regulation of IA by G protein-coupled receptors will contribute to neuronal excitability and subsequent transmission of the nociceptive electrical signals. Although neuromedin U (NMU) has been implicated in analgesia, the detailed mechanisms still remain unclear. In this study, we identify a novel functional role of NMU type 1 receptor (NMUR1) in regulating the transient outward K+ currents (IA) in small dorsal root ganglion (DRG) neurons. We found that NMU reversibly increased IA in a dose-dependent manner, instead the sustained delayed rectifier K+ current (IDR) was not affected. This NMU-induced IA increase was pertussis toxin-sensitive and was totally reversed by NMUR1 knockdown. Intracellular application of GDPβS (guanosine 5′-O-(2-thiodiphosphate)), QEHA peptide, or a selective antibody raised against the Gαo or Gβ blocked the stimulatory effects of NMU. Pretreatment of the cells with the protein kinase A (PKA) inhibitor or ERK inhibitor abolished the NMU-induced IA response, whereas inhibition of phosphatidylinositol 3-kinase or PKC had no such effects. Exposure of DRG neurons to NMU markedly induced the phosphorylation of ERK (p-ERK), whereas p-JNK or p-p38 was not affected. Moreover, the NMU-induced p-ERK increase was attenuated by PKA inhibition and activation of PKA by foskolin would mimic the NMU-induced IA increase. Functionally, we observed a significant decrease of the firing rate of neuronal action potential induced by NMU and pretreatment of DRG neurons with 4-AP could abolish this effect. In summary, these results suggested that NMU increases IA via activation of NMUR1 that couples sequentially to the downstream activities of Gβγ of the Go protein, PKA, and ERK, which could contribute to its physiological functions including neuronal hypoexcitability in DRG neurons.

Although neuromedin U (NMU) has been implicated in analgesia, the detailed mechanisms still remain unclear. In this study, we identify a novel functional role of NMU type 1 receptor (NMUR1) in regulating the transient outward K ؉ currents (I A ) in small dorsal root ganglion (DRG) neurons. We found that NMU reversibly increased I A in a dose-dependent manner, instead the sustained delayed rectifier K ؉ current (I DR ) was not affected. This NMU-induced I A increase was pertussis toxinsensitive and was totally reversed by NMUR1 knockdown. Intracellular application of GDP␤S (guanosine 5-O-(2-thiodiphosphate)), QEHA peptide, or a selective antibody raised against the G␣ o or G␤ blocked the stimulatory effects of NMU. Pretreatment of the cells with the protein kinase A (PKA) inhibitor or ERK inhibitor abolished the NMU-induced I A response, whereas inhibition of phosphatidylinositol 3-kinase or PKC had no such effects. Exposure of DRG neurons to NMU markedly induced the phosphorylation of ERK (p-ERK), whereas p-JNK or p-p38 was not affected. Moreover, the NMU-induced p-ERK increase was attenuated by PKA inhibition and activation of PKA by foskolin would mimic the NMU-induced I A increase. Functionally, we observed a significant decrease of the firing rate of neuronal action potential induced by NMU and pretreatment of DRG neurons with 4-AP could abolish this effect. In summary, these results suggested that NMU increases I A via activation of NMUR1 that couples sequentially to the downstream activities of G␤␥ of the G o protein, PKA, and ERK, which could contribute to its physiological functions including neuronal hypoexcitability in DRG neurons.
Neuromedin U (NMU), 2 a highly conserved peptide, is expressed widely not only in the central nervous system (CNS) (1) but also in peripheral tissues (2). This peptide acting through its receptor has been implicated in regulating a variety of physiological functions, including stress response, energy homeostasis, and nociception (2). Two G protein-coupled receptors including NMU type 1 receptor (NMUR1) and NMUR2 have been identified as the endogenous receptors for NMU (3,4). NMUR1 is expressed more abundantly in peripheral tissues, whereas NMUR2 is localized more centrally. These distribution patterns suggest that each receptor subtype may have different but specific physiological roles (5). For example, NMU has recently been reported to be involved in spinal nociceptive reflexes (5,6), and this CNS effect was thought to be mediated by NMUR2, but not NMUR1. In contrast, NMUR1 was shown to decrease the neuronal excitability of peripheral dorsal root ganglion (DRG) (7), rather than the spinal cord neurons (8). Interestingly, NMUR1 was reported to be expressed in small to medium diameter neurons in the peripheral neuronal system (7,8), whereas NMUR2 was not detectable. These data together suggest that NMUR1 could be involved in NMU-mediated peripheral analgesic effects. Nevertheless, the detailed mechanisms underlying NMUR1 anti-nociception in peripheral neuronal system are still largely unknown and as such require elucidation.
Nociceptive neurons, which are small to medium size (9, 10), consist of several classes that differ in their expression patterns of various ion channels that lead to a functional diversity in sensory neuron signaling (11,12). Abnormal excitability of neurons is attributable to alterations in the expression and functional characteristics of voltage-gated ion channels such as Na ϩ , Ca 2ϩ , and K ϩ (13,14). In contrast to Na ϩ and Ca 2ϩ channels, K ϩ channel expression in DRG neurons has received attention only recently (15,16). Voltage-gated K ϩ channel (Kv) currents are critical determinants of neuronal excitability in sensory neurons and are divided into two major categories: the transient A-type K ϩ currents (I A ) and the sustained delayed rectifier-type K ϩ currents (I DR ) (10,17). I A activates at subthreshold membrane potentials, inactivate rapidly, and rapidly recover from inactivation: these are important characteristics to influence neuronal excitability and to modulate synaptic plasticity (18,19). For example, down-regulation of I A in the nociceptors may increase pain sensation (20). As I A has an important role in determining the electrical properties of DRG neurons, the regulation of I A will contribute to neuronal excitability (14) and subsequent transmission of the nociceptive electrical signals (20,21).
In this study, we investigated the effects of NMU on I A and elucidated the underlying mechanisms in small DRG neurons in mice. Based on pharmacological manipulation of the NMUinduced I A increase, we report that NMU increased I A via the ␤␥ subunits of G o protein in a PKA-dependent ERK pathway. This I A increase induced by NMU was mediated by activation of the NMUR1 receptors. One of the immediate outcomes is the induction of neuronal hypoexcitability in the DRG neurons with possible implications in pain processing.

EXPERIMENTAL PROCEDURES
Pharmacological Agents-All drugs were obtained from Sigma, unless otherwise indicated. Stock solutions of NMU, GDP␤S, pertussis toxin (PTX), cholera toxin (CTX), PKI 6-22, H-89, 4-aminopyridine (4-AP), CdCl 2 , and chelerythrine chloride were prepared in distilled deionized water. Stock solutions of U73122, LY294002, U0126, PD98059, forskolin, calphostin C, and GF109203X were prepared in dimethyl sulfoxide. The concentration of dimethyl sulfoxide in the bath solution is expected to be less than 0.01%, and had no functional effects on I A . An antibody raised against the G␣ o protein was a generous gift from Dr. Q. Ma (Institute of Neuroscience, Soochow University). This antibody has been previously shown to specifically recognize the G o protein but not the G i subunit (22). The QEHA peptide and SKEE peptide (22) were synthesized by GenScript (NJ).
Isolation of Dorsal Root Ganglion Neurons-All procedures have been approved by the Animal Care and Use Committee of Soochow University and were in compliance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Neurons were isolated from lumbar (L 4 -L 6 ) dorsal root ganglia (DRG) of adult mice (ICR, 4 -6 weeks) using our previously described methods (7,23). Briefly, the freshly removed ganglia were minced and enzymatically digested with collagenase I for 30 min (37°C) and then with trypsin (2.5 mg/ml) for 15 min (37°C). They were then gently triturated with a fire-polished pipette to obtain a suspension with single cells. Isolated cells were collected by centrifugation and suspended in minimal essential medium supplemented with FBS (10%), B27 (2%), Glutamax (1%), and penicillin/streptomycin (20 units/ml, 0.2 mg/ml). Dissociated neurons were plated on polylysine (0.1 mg/ml)-coated glass coverslips and kept in 95% air and 5% CO 2 incubators at 37°C until recording. Isolated DRG neurons were recorded within 24 -48 h after plating.
Electrophysiological Recordings-For electrophysiological recording, only small-sized DRG neurons (soma diameter 20 -27 m) were selected. Whole cell patch-clamp recordings were performed at room temperature (22-24°C) with a Multi-Clamp 700B amplifier (Molecular Devices, USA). Recording pipettes (WPI) had ϳ3 M⍀ resistance when filled with internal solution. pClamp 10.2 was used to acquire and analyze data. A-type K ϩ currents were recorded in voltage-clamp mode. Signals were filtered at 1 kHz and digitized at 10 kHz. Series resistance (Rs) and capacitance (Cm) values were taken directly from readings of the amplifier after electronic subtraction of the capacitive transients. Series resistance was compensated to the maximum extent possible (at least 75%). Current traces were corrected with on-line P/6 trace subtraction. Multiple independently controlled glass syringes served as reservoirs for a gravity-driven local perfusion system. Solution exchange was accomplished by constant suction through a glass capillary tube at the opposite end of the recording dish. To record K V currents, Na ϩ in control external solution was replaced with equimolar choline and the Ca 2ϩ concentrations were reduced to 0.03 mM to suppress Ca 2ϩ currents and prevent Ca 2ϩ channels becoming Na ϩ conducting (24). The reduced external Ca 2ϩ would also be expected to suppress Ca 2ϩ -activated K ϩ current. For currentclamp and Na ϩ current recordings, the pipette solution contained the following (in mM): KCl 110, NaCl 10, EGTA 2, HEPES 25, Mg-ATP 4, and Na 2 GTP 0.3, adjusted to pH 7.3 with KOH. The external solution contained (in mM): NaCl 128, KCl 2, CaCl 2 2, MgCl 2 2, glucose 30, HEPES 25, pH adjusted to 7.4 with NaOH. We blocked the sustained delayed rectifier K ϩ current (I DR ) and voltage-gated Ca 2ϩ channels, respectively, by application of 25 mM triethanolamine (25) and 100 M CdCl 2 in the external solution for current clamp recordings. Electrophysiological analysis was performed as previously described (26,27). In experiments in which neurons were dialyzed with compounds (antibody or peptides), current measurements were started at least 5 min after breaking the patch.
Protein Kinase A (PKA) Activity Assay-Cells were seeded in 12-well plates and pretreated with either vehicle (0.1% dimethyl sulfoxide) or NMU for 10 min. The cells were washed with ice-cold phosphate-buffered saline and placed on ice. 200 l of lysis buffer (20 mM MOPS, 50 mM ␤-glycerolphosphate, 50 mM sodium fluoride, 1 mM sodium vanadate, 5 mM EGTA, 2 mM EDTA, 1% Nonidet P-40, 1 mM dithiothreitol, 1 mM benzamidine, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml of leupeptin, and 10 g/ml of aprotinin) was added. After a 10-min incubation on ice, the cells were scraped off and transferred to microcentrifuge tubes. The cell lysates were centrifuged for 15 min and aliquots of the supernatants containing 0.2 g of protein were assayed for PKA activity according to the manufacturer's instructions using an enzyme-linked immunosorbent assay (ELISA) kit (JC Bio-Tec).
siRNA Transfection in DRG Neurons-The cells were seeded at high density onto laminin-polyornithine-coated coverslips. The NMUR1 siRNAs (sense, 5Ј-GCCUCAGUCAACUAACU-GUTT-3Ј, antisense, 5Ј-ACAGUUAGUUGACUGAGGCTT-3Ј) were purchased from Genepharm (Shanghai, China). The siRNA sequences were subjected to BLAST analysis to exclude potential off-target effects. The negative control siRNA (5Ј-UUCUCCGAACGUGUCACGUTT-3Ј) was used as control. DRG neurons were transfected with 0.6 g of siRNA (100 nM final concentration, chemical modification with 6-FAM) using Oligofectamine (Invitrogen). After 24 h post-transfection, small DRG neurons that fluoresced green as observed under the inverted fluorescence microscope (Ti-2000, Nikon) were selected for whole cell patch clamp analysis.
Immunofluorescent Staining of NMUR1-DRG neurons were grown on polylysine-coated coverslips, and fixed in phosphatebuffered saline (PBS) containing 4% sucrose and 4% paraformaldehyde (PFA) for 20 min at 4°C. The fixed cells were washed three times with PBS before permeabilization in PBS containing 0.1% Triton X-100 for 5 min. Nonspecific staining was blocked by 4% goat serum in PBS for 1 h. Slides were incubated with primary antibody rabbit anti-mouse NMUR1 (Santa Cruz, 1:500) for 1 h at room temparature. After washing three times with PBS, Alexa Fluor 488-conjugated chicken anti-rabbit IgG (green) was applied to the samples at a dilution of 1:500. The slides were examined by using a laser-scanning confocal microscope (Fluoview BX61, Olympus). Negative controls, omitting each primary antibody, were performed in each case and no significant staining was observed (data not shown).
Data Analysis-All data are presented as mean Ϯ S.E., and GraphPad Prism software was used for data plotting. Treatment effects were statistically analyzed by one-way analysis of variance followed by Bonferroni's multiple comparison test. Paired or two-sample t test was used when comparisons were restricted to two means. Error probabilities of p Ͻ 0.05 were considered statistically significant. Concentration-response curves were fitted by the sigmoidal Hill equation: I/I control ϭ 1/(1 ϩ 10 (log IC50-X) n H ), where X is the decadic logarithm of the concentration used, IC 50 is the concentration at which the halfmaximum effect occurs, and nH is the Hill coefficient. The membrane conductance (G) at each command potential (V m ) was determined by dividing the measured membrane current (I) by the driving force as follows: G ϭ I/(V m Ϫ E K ), where E K is the equilibrium K ϩ potential and was calculated to be Ϫ76 mV. Activation data (G-V curve) were fitted by the following modified Boltzmann equation: where G max is the fitted maximal conductance, V1 ⁄ 2 is the membrane potential for half-activation, and k is the slope factor. Steady-state inactivation of I A was fitted with the following negative Boltzmann equation:

NMU Selectively Increases I A in Small DRG Neurons-There
were two main types of outward voltage-gated K ϩ channel (Kv) currents described in nociceptive DRG neurons: the transient A-type K ϩ current (I A ) and the sustained delayed rectifier K ϩ current (I DR ) (24,28). To begin dissecting the roles of NMU in modulation of I A in small (soma diameter 20 -27 m) DRG neurons, we first isolated these two kinetically different Kv currents in these neurons. With Na ϩ and Ca 2ϩ currents eliminated (see "Experimental Procedures"), a large outward current was evoked by a command potential of ϩ40 mV from a holding potential of Ϫ80 mV in small DRG neurons (Fig. 1A). The typical current profile observed in these neurons exhibited a rapidly inactivating component along with a sustained component. A 150-ms prepulse to ϩ10 mV allowed the transient channels to inactivate, leaving only the I DR (Fig. 1A, left). Subtraction of I DR from the total current yields a transient I A (Fig. 1A, right). The I A could be blocked by application of 5 mM 4-AP (increase % ϭ 82.4 Ϯ 2.3%, n ϭ 7), which further confirmed effective I A isolation.
Bath application of 1 M NMU increased I A by 26.2 Ϯ 1.8% in small DRG neurons (Fig. 1, B and D), whereas I DR was not significantly affected (Fig. 1, C and D). Upon washout of NMU, the amplitude of I A partially returned to the pre-treatment level within 3 min (Fig. 1B). Using the magnitude of the effect that NMU has on I A elicited by depolarization to ϩ40 mV, it is clear that NMU increased I A in a dose-dependent manner (Fig. 1E). The relationship between the concentration of NMU used and the degree of increase observed is described by a logistic equation where the concentration of NMU producing the halfmaximal increase (IC 50 ) is 0.32 M, the apparent Hill coefficient is 0.91, and the maximal incremental effect is 28.8 Ϯ 2.6% (Fig.  1E). To further identify the selective I A increase induced by NMU, we investigated whether 4-AP, a classic I A blocker, would abrogate NMU-mediated response. Indeed, application of 5 mM 4-AP abrogated the NMU-induced I A increase in small DRG neurons (Fig. 1F). Notably, application of NMU during the maximum 4-AP-induced response failed to produce any further inhibition (Fig. 1F).
NMU Rightward Shifts Steady-state Inactivation Curve-As a dose-dependent increase in the peak current density of I A was evident, we next determined whether the biophysical properties of I A were affected by NMU. The current-voltage (I-V) curves showed that 1 M NMU significantly increased I A at each tested voltage from Ϫ10 mV, and peaked at ϩ40 mV where the current density increased from 209.3 Ϯ 22.8 to 288.1 Ϯ 25.6 pA/pF (n ϭ 9, p Ͻ 0.01, Fig. 2, A and B). We next analyzed the voltage dependence of activation and inactivation potentials and first observed a slight but statistically insignificant shift of 1.3 mV in the hyperpolarized direction of the activation potential (V1 ⁄ 2 from Ϫ9.7 Ϯ 1.3 to Ϫ10.2 Ϯ 1.6 mV, and k value from 20.5 Ϯ 1.8 to 19.9 Ϯ 1.4, n ϭ 7) (Fig. 2C). In contrast, NMU at 1 M significantly shifted the steady-state inactivation potentials of I A channels in the depolarizing direction by ϳ12 mV (V1 ⁄ 2 from Ϫ49.5 Ϯ 1.8 to Ϫ37.6 Ϯ 1.1 mV, p Ͻ 0.05; and k value from 11.7 Ϯ 0.7 to 16.4 Ϯ 1.1, p Ͻ 0.05) (Fig. 2D), which suggest that the increase in I A observed upon application of NMU in mouse DRG neurons could be due to a decreased proportion of channels remaining in the inactivated state.
NMUR1 Knockdown Blocked NMU-induced I A Increase-Previous reports including ours have clearly shown the localization of NMUR1, but not NMUR2, in both small-and medium-sized DRG neurons (7,29). To obtain further evidence that the NMU-induced I A increase was via NMUR1, first we examined the subcellular expression of NMUR1 in small DRG neurons. Fig. 3A clearly showed the membrane localization of NMUR1 in small-sized DRG neurons. Negative controls omitting the primary antibody showed no background (not shown).
To determine whether the NMU-induced I A increase was mediated via NMUR1, we used a siRNA knockdown approach to examine the effect of NMU on I A in NMUR1-silenced small DRG neurons. Western blot analysis showed that expression of NMUR1 was significantly reduced in cells transfected with NMUR1 siRNA compared with cells transfected with the control siRNA (Fig. 3B, ctrl siRNA). As shown in Fig. 3C, knockdown of NMUR1 in small DRG neurons resulted in a near complete elimination of NMU enhancement of I A (Fig. 3C), whereas NMU still robustly increased I A in the control siRNAtransfected cells (Fig. 3, C and D). These results together suggested that NMU increased I A via activation of NMUR1 in small DRG neurons in mice.
NMUR1-mediated I A Increase Requires G␤␥ Subunits of G o Protein-NMUR1 belongs to a large family of G protein-coupled receptors (2,30,31). To investigate whether heterotrimeric G proteins are involved in the NMUR1-mediated I A response, we dialyzed small DRG neurons with GDP␤S (1 mM), a nonhydrolysable GDP analog. GDP␤S completely abolished the increase of I A induced by 1 M NMU (increase % ϭ 3.1 Ϯ 0.8, Fig. 4, A and B), indicating that G protein activation was required for NMU action. We further determined which isoform of G␣ was involved in this inhibition. First, we examined the effects of NMU on I A in the presence of CTX, which inactivates G␣ s by ADP-ribosylation. After pretreatment of DRG neurons with CTX (0.5 g/ml), NMU (1 M) still robustly increased I A (increase % ϭ 23.6 Ϯ 2.1, Fig. 4, A and B), which indicated that the increase of I A was G␣ s independent. However, pretreatment of the cells with PTX (0.2 g/ml), which catalyzes the ADP-ribosylation of G␣ i/o , abolished the stimulatory effect of NMU (increase % ϭ 3.7 Ϯ 1.0, Fig. 4, A and B). The increase of I A induced by NMU was sensitive to PTX, but not CTX, indicating that G␣ i/o , but not G␣ s , was involved. Furthermore, an antibody that specifically binds G o but not the G i subunit (27) was used to determine which subtype of G protein was involved in the NMU response. This antibody abolished the increase of NMU on I A (increase % ϭ 1.5 Ϯ 1.7, Fig. 4, A and B) FIGURE 1. NMU selectively increased I A in small DRG neurons. A, isolation of I A . Na ϩ currents, Ca 2ϩ currents, and Ca 2ϩ -activated K ϩ currents have been eliminated (see "Experimental Procedures"). Left, the membrane voltage was held at Ϫ80 mV and I A was isolated by a two-step voltage protocol. Right, the remaining current after off-line subtraction of the noninactivating portion of current remaining after the prepulse to Ϫ10 mV. This protocol is used for isolation of I A in all figures. B and C, representative current traces and the time course showed that I A (n ϭ 7, B), but not sustained delayed rectifier K ϩ current (I DR ) (n ϭ 9, C), was significantly increased during exposure to 1 M NMU. D, summary data of current density showed that 1 M NMU selectively increased I A . E, dose-response curve for the stimulatory effects of NMU on I A . The line represents the best fit of the data points to the sigmoidal Hill equation. Number of cells tested at each concentration of NMU is indicated in parentheses. F, time course and exemplary traces showed no effect of 1 M NMU on I A in the presence of 4-AP (n ϭ 5). ***, p Ͻ 0.001 versus control.
without significantly altering the current amplitude in the absence of agonist. The denatured G o -specific antibody (increase % ϭ 26.8 Ϯ 2.3, Fig. 4B) and a G i -specific antibody (increase % ϭ 23.9 Ϯ 3.7, not shown) had no such effect. These results suggested that a G o -, but not G i -like protein, was involved in the NMU-mediated response. We further determined the implication of the native G protein ␤␥ subunit (G␤␥) of G o in the NMU-mediated response. These subunits were quenched in two ways. First, a synthetic peptide, QEHA (200 M) (26,32), which competitively binds G␤␥ and blocks G␤␥-mediated signaling, was introduced into the recording pipette (Fig. 4B). Second, a G␤ subunit-specific antibody was applied in the pipette (Fig. 4B). We found that either anti-G␤ or QEHA, but not its scrambled peptide SKEE (Fig. 4B), abolished the stimulatory effect of NMU on I A , which suggested that the G␤␥ subunit of G o was necessary for the NMUR1-mediated I A increase.
PKA Was Involved in NMUR1-mediated I A Increase-Because NMU appeared to increase I A via the ␤␥ subunits of the G o protein, we further determined the downstream intracellular signaling molecules. Previous studies suggested that the immediate downstream mediator of G␤␥ was phosphatidylinositol 3-kinase (PI3K) (32). However, we found that the PI3K inhibitor LY294002 failed to affect the I A increase induced by NMU (increase % ϭ 25.5 Ϯ 2.3, Fig. 5A). G o protein activation was reported to activate PKC via PLC in cerebellar Purkinje neurons (32). We therefore pretreated DRG neurons with the PLC inhibitor U73122 (3 M) and found that the NMU-mediated stimulatory effect was not affected by the PLC inhibitor (increase % ϭ 24.9 Ϯ 2.5, Fig. 5A). As it has also been reported that PKC could be PLC independently activated (32), we investigated whether the stimulatory effects of NMU were mediated by a PKC-dependent pathway. Preincubation of mouse DRG neurons with PKC inhibitors, GF109203X (1 M, increase % ϭ 26.1 Ϯ 3.4, Fig. 5A) or chelerythrine chloride (1 M, increase % ϭ 24.6 Ϯ 3.1, Fig. 5A), did not significantly affect the I A increase induced by NMU. A similar result was obtained with another PKC inhibitor, calphostin C (50 nM, increase % ϭ 27.6 Ϯ 2.9, Fig. 5A). Previous studies have shown that G␤␥ could stimulate PKA activity via regulation of adenylate cyclase types II and V (33,34). Therefore, we determined whether PKA was involved in the NMU-induced I A increase and found that small DRG neurons pretreated with a PKA inhibitor, H89 (1 M), almost completely abolished the NMU-induced I A increase (increase % ϭ 1.7 Ϯ 0.5, Fig. 5A). In contrast, H85 (1 M), a structurally related but inactive analog, had no such effect (not shown). The PKA dependence of the NMU-induced I A increase was further validated by dialyzing the cells with a pipette solution containing PKI 6-22, another PKA inhibitor. Our results showed that intracellular application of PKI 6-22 (1 M) via the recording pipette solution blocked the NMU-induced I A increase (increase % ϭ 1.3 Ϯ 1.1, Fig. 5A). These results suggest that G␤␥-dependent PKA was involved in the NMUR1-mediated I A increase. As a complementary test of our hypothesis, we further investigated whether forskolin, a potent PKA activator, would also induce the I A increase. Indeed, application of 20 M forskolin to DRG neurons mimicked the NMU-induced I A increase (increase % ϭ 17.9 Ϯ 1.3, Fig. 5B). In addition, effects of NMU on the PKA activity in DRG neurons were also detected. As shown in Fig. 5C, treatment with 1 M NMU for 10 min resulted in a ϳ2.3-fold increase of PKA activity in DRG neurons. Knockdown of NMUR1 in DRG neurons resulted in a near complete elimination of the NMU-stimulated PKA activity (Fig. 5C). These findings further support the conclusion that the NMUR1-mediated I A increase in small DRG neurons occurs through a G o protein ␤␥ subunit-mediated PKA pathway.
ERK Signaling Was Involved in NMUR1-mediated I A Increase-ERK/mitogen-activated protein kinase (MAPK) pathway has been demonstrated to play a crucial role in pain regulation. Activation of the MAPK pathway was reported to regulate I A in dorsal horn neurons (35). It was therefore of interest to examine whether the MAPK signaling pathway is involved in the NMUinduced I A increase. Western blot analysis showed that exposure of DRG neurons to NMU (1 M) markedly increased the expression of phosphorylation of ERK (p-ERK) (Fig. 6A), whereas p-JNK (Fig. 6B) and p-p38 (Fig. 6C) remain unchanged, which suggested that ERK could be involved in a NMU-induced response such as I A regulation. As a complementary test of our hypothesis, knockdown of NMUR1 inhibited the NMU-induced p-ERK expression increase. In contrast, the NMU-induced phosphorylation of ERK was not affected by control siRNA treatment (Fig. 6, D and E). In addition, pretreatment of cells with both the PKA antagonist as well as the MAPK/ERK kinase (MEK) inhibitor U0126 (20 M) eliminated NMU-induced ERK activation (Fig. 6, D and E). We next investigate whether ERK activation was exactly involved in the NMU-in-  7). To determine the steady-state inactivation, conditioning prepulses ranging from Ϫ100 to ϩ40 mV were applied at 5-s intervals in ϩ10 mV increments for 150 ms followed by a voltage step to ϩ40 mV for 500 ms. *, p Ͻ 0.05 versus control, **, p Ͻ 0.01 versus control.    duced I A increase. We applied U0126, the inhibitor of mitogenactivated ERK kinase (the upstream activator of ERK), to small DRG neurons. Preincubation of the cells with U0126 abrogated the I A increase induced by NMU (increase % ϭ 3.3 Ϯ 0.5, Fig.  6F), whereas U0124, a negative control analog of U0126, elicited no such effects (not shown). Similar results were obtained with another MAPK inhibitor, PD98059 (increase % ϭ 2.6 Ϯ 0.9, Fig. 6F). These results together suggest that activation of NMUR1 increase I A via activation of ERK in mouse DRG neurons.
NMU Decreases Neuronal Excitability in Small DRG Neurons-Kv currents including I A and I DR in primary sensory neurons are important for regulating neuronal excitability (20,36).
To further address the functional implication of the I A increase induced by NMUR1, we tested the effects of NMU on membrane excitability in small DRG neurons. We first investigated whether voltage-gated Na ϩ channels would be regulated by NMU and found that NMU failed to affect Na ϩ current density in small DRG neurons (Fig. 7, A and B). As shown in Fig. 7, C and D, the current-voltage (I-V) curves were not shifted. Therefore, by using an external solution including CdCl 2 to block the voltage-gated Ca 2ϩ channels and triethanolamine to block the I DR , we found that 1 M NMU significantly decreased the rate of firing to 51.56 Ϯ 6.53% (Fig. 7, E and F). Upon washout of NMU, the rate of firing was restored (Fig. 7, E and F). In addition, NMU at 1 M significantly decreased the AP amplitude and prolonged the first spike latency. Other membrane properties of neuronal excitability including threshold and resting membrane potential were not significantly changed by NMU (1 M) in small DRG neurons (Table 1). To further verify that this NMU-induced hypoexcitability was via the I A increase, we examined the effects of NMU on neuronal excitability with application of 4-AP in the external solution (n ϭ 17 neurons, Fig. 7, G and H) and found that pretreatment of cells with 4-AP completely abolished NMU-induced hypoexcitability in small DRG neurons (Fig. 7H). Other membrane properties of neuronal excitability are shown in Table 1. These results together demonstrated that activation of NMUR1 by NMU induces neuronal hypoexcitability via an increase of I A in small DRG neurons.

DISCUSSION
Our present study adds a new piece of information to the NMUR1 signaling pathway by demonstrating that activation of NUMR1 stimulates A-type K ϩ currents (I A ) via the ␤␥ subunits of the G o protein and PKA-dependent ERK1/2 pathway and leads to a decrease in neuronal excitability in mouse peripheral sensory DRG neurons, whereas I DR remains unchanged.
Unlike G i , which inhibits adenylyl cyclase, the main function of G o can be interpreted through the actions of a common pool of G␤␥ dimers (37,38). Consistently, we have found that the G␤␥ subunits of G o are involved in the NMUR1-mediated I A increase, because: 1) the response is abolished by dialyzing cells with an anti-G o antibody; 2) intracellular application of an antibody raised against G␤ or a G␤␥ blocking peptide, QEHA, abolishes the NMUR1-mediated response. A known target of G␤␥ is PI3K, which has been shown to regulate voltage-gated Ca 2ϩ channels via PI3K (39). Our previous study has found that in hippocampus neurons the inhibition of L-type Ca 2ϩ channel currents by NMUR1 signaling is mediated by the PI3K-dependent novel PKC pathway (22). However, in the present study, both PI3K inhibitor LY294002 (3 M), and PKC inhibitors including calphostin C (50 nM) or GF109203 (1 M) or chelerythrine chloride (1 M) have no significant effect in DRG neurons, whereas both PKA inhibitors H89 (1 M) and PKI 6-22 (1 M) show remarkable suppression of the NMU-induced I A increase. The difference of downstream pathways utilized by NMUR1 in hippocampal neurons and DRG neurons is likely due to the different origins of the neuron tissues, which have quite different functions. This type of difference has been documented previously. For example, stimulation of dopamine D1 receptor suppresses the slowly inactivating K ϩ current via PKA in rat medial prefrontal cortex pyramidal neurons (40), whereas activation of the dopamine D1 receptor inhibits T-type Ca 2ϩ channel currents via direct G␤␥ binding (41). In addition, the differences in tissue-specific alternative splice variants of ion channels could lead to the activation of different signaling pathways and result in different responses according to the different tissue functions and microenvironment. For example, mutually exclusive splicing patterns in the Cav2.2 gene modulate N-type channel function in nociceptors, leading to a change in morphine analgesia (42). Furthermore, alternative splicing of the Cav1.3 channel IQ domain generates a molecular switch for Ca 2ϩ -dependent inactivation within auditory hair cells (43).
PKC-mediated responses are also shown to be tissue specific. For example, activation of PKC by PMA inhibits I A in dendrites of hippocampal CA1 pyramidal neurons (44), but enhances I A amplitude in rat cerebellar granule cells (45). In addition, the regulation of I A could be PKC independent, as reported in murine proximal colonic myocytes (46). In the present study, we have found that in mouse DRG neurons the stimulatory effects of NMU on I A are PKC independent, but rather are mediated by PKA followed by ERK/MAPK. In pyramidal neuron dendrites, activation of ERK leads to reduction of Kv4.2encoded I A via phosphorylation of Thr-607 in the C-terminal of Kv4.2 (47). However, we have found that in mouse small DRG neurons I A increases after NMUR1 activation and subsequent ERK phosphorylation. This effect is NMUR1 and ERK dependent, because it is blocked by NMUR1 siRNA or MAPK inhibitor PD98059. Although the discrepancy remains to be further clarified, this could be possibly explained by a distinct ERK phosphorylation site or a different Kv subtype encoding I A in DRG neurons, in which Kv1.4, Kv3.4, Kv4.2, and Kv4.3 channels are endogenously expressed (48). An alternative hypothesis is that ERK phosphorylates an intermediate protein, that in turn up-regulates the I A currents in DRG neurons. Furthermore, different KCHIP1 splice variants are able to generate different, even opposite modulation of Kv4 channel currents (49).
A-type K ϩ channels, the key components controlling membrane excitability (50), have been implicated in the delay of the spike onset (first spike latency) and the decrease in the firing frequency of central neurons (51). Changes in membrane excitability of DRG neurons can directly influence pain symptoms such as hyperalgesia and allodynia in vivo (52). An important consequence of the I A current is to delay the first-spike firing in response to small depolarization (53). Consistent with the NMU-induced I A increase, activation of NMUR1 in small DRG neurons does lead to an inhibition of neuronal excitability including decreasing spike frequency and increasing first-spike latency, which are the two major parameters that determine the timing of neurotransmitter release, and hence pain transmission (54). Our current study is consistent with our previous reports that NMU induced the membrane hypoexcitability in peripheral sensory neurons. In DRG neurons, both T-type Ca 2ϩ channel currents (7) and I A are regulated by G␤␥-dependent PKA activation, downstream of NMUR1 signaling. The stimulation of I A and inhibition of T-currents both lead to

Membrane properties of small DRG neurons in mice induced by 1 M NMU in the absence (؊) or presence (؉) of 4-AP
Cm, capacitance; R in , input resistance; RMP, resting membrane potential; AHP, afterhyperpolarization potential; FSL, First spike latency; n, number of cells (*, p Ͻ 0.05 versus pre-NMU).
hypoexcitability. The divergent signals are not contradictory to each other; rather this may imply an additive or even a synergistic effect of NMUR1 in controlling the membrane excitability of DRG neurons. This view is supported by the evidence that T-type Ca 2ϩ channels and Kv4 channels can form a signaling complex in rat cerebellar stellate cells that efficiently couples Ca 2ϩ influx to KChIP3 to modulate the Kv4 function (55). Although whether the Cav3-Kv4 signaling complex exists in mouse small DRG neurons remains to be further investigated, this interaction may be critical for allowing Kv4 channels encoding I A to function in the subthreshold membrane potential range to regulate neuronal firing properties.
In summary, this study discovers a novel function of NMU in regulation of I A in mouse small DRG neurons, whereas the I DR remains unchanged. Based on pharmacological manipulation of the NMU-induced I A increase, we have identified a second messenger pathway initiated by NMUR1 that couples sequentially to the ␤␥ subunits of the G o protein, PKA, and ERK. The NMUR1-mediated I A increase contributes to its physiological functions including neuronal hypoexcitability. These results suggest that NMUR1 signaling could be a potential therapeutic target for clinical management of pain. However, as several subtypes of Kv channels could underlie I A in small DRG neurons (56), further studies by using anatomical and genetic approaches are necessary to identify the exact Kv subunits that underlie I A , and mediate the NMUR1-dependent regulation of I A in DRG neurons.