INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Degenerin/epithelial Na⁺ channel (DEG/ENaC) proteins are a family of voltage-insensitive cation channels that have a variety of expression patterns and functional roles (Alvarez de la Rosa et al. 2000; Mano and Driscoll 1999). Although sequence similarity is limited to a few small regions of the protein, DEG/ENaC subunits share a common topology. Family members have relatively short intracellular N- and C-termini, two membrane-spanning sequences that are probably  helices, and a large extracellular loop that contains 14 conserved cysteine residues (Canessa et al. 1994a; Renard et al. 1994; Snyder et al. 1994). Individual DEG/ENaC subunits assemble as homomultimers or heteromultimers to form a channel. Reports on the precise number of subunits that form a channel vary, suggesting that four to nine are involved (Eskandari et al. 1999; Firsov et al. 1998; Kosari et al. 1998; Snyder et al. 1998). Amiloride inhibits current by blocking the channel pore, although the sensitivity to amiloride varies substantially for different DEG/ENaC channels (Garty and Palmer 1997).

Of the nine mammalian genes for DEG/ENaC proteins, three produce H⁺-activated channels: brain Na⁺ channel 1 (BNC1)¹ (García-Añoveros et al. 1997; Price et al. 1996; Waldman et al. 1996), acid-sensing ion channel (ASIC) (García-Añoveros et al. 1997; Waldman et al. 1997b), and dorsal root acid sensing ion channel (DRASIC) (Waldman et al. 1997a). Although protons are the only known ligands for these channels, the neurotransmitters FMRFamide (Phe-Met-Arg-Phe-amide) and neuropeptide FF potentiate H⁺-gated current from ASIC and DRASIC by slowing the rate of desensitization in the continued presence of an acid stimulus (Askwith et al. 2000). A reduction in extracellular Ca²⁺ or Mg²⁺ concentration also enhances H⁺-gated currents by shifting the threshold activation to higher pH values (Immke and McCleskey 2001; Waldman et al. 1997b; Zhang and Canessa 2001). Zn²⁺ has also been reported to potentiate acid activated currents generated by BNC1a (Baron et al. 2001).

The DRASIC channel has attracted interest for several reasons. It is expressed in both large- and small-diameter dorsal root ganglion (DRG) sensory neurons (Price et al. 2001; Waldman et al. 1997a); in general, large-diameter DRG neurons tend to detect innocuous stimuli such as light touch, and small-diameter DRG neurons tend to detect noxious mechanical, thermal, and chemical stimuli (Lawson 1992). Moreover, DRASIC localizes both to specialized cutaneous mechanosensory structures, such as Merkel cell/neurite complexes, Meissner corpuscles, and lanceolate fibers, and to small free nerve endings associated with nociception (Price et al. 2001). In addition, disruption of the DRASIC gene in mice alters both mechanosensory and nociceptive responses as assessed by single-fiber recording in a skin-nerve preparation (Price et al. 2001).

Expression of DRASIC in heterologous cells generates acid-activated currents that rapidly desensitize in the continued presence of the stimulus, although there is a small component of sustained current (Babinski et al. 2000; Lingueglia et al. 1997; Sutherland et al. 2001; Waldman et al. 1997a). Moreover, the properties of the currents generated by DRASIC differ from those produced by BNC1 and ASIC (Sutherland et al. 2001; Waldman and Lazdunski 1998). Some properties of the currents generated by heterologous expression of DRASIC and other acid-activated DEG/ENaC channels are similar to those of the acid-activated currents that have been reported in sensory neurons for many years (Akaike and Ueno 1994; Krishtal and Pidoplichko 1981b; Petruska et al. 2000). This similarity suggests the hypothesis that DEG/ENaC channels are responsible for H⁺-gated DRG currents. Preliminary studies of DRG neurons from DRASIC null mice supported this hypothesis (Price et al. 2001).

To better understand the contribution of DRASIC, we compared transient H⁺-activated currents in DRG neurons from wild-type mice and mice in which the DRASIC gene was disrupted. However, H⁺-gated DEG/ENaC proteins are not the only channels activated by acid; the VR1 cation channel is also H⁺ gated (Caterina et al. 1997; Tominaga et al. 1998). VR1 is expressed primarily in small-diameter DRG neurons where it generates sustained acid-evoked currents (Caterina et al. 1997, 2000; Davis et al. 2000; Tominaga et al. 1998). Moreover, VR1 plays an important role in nociception (Caterina et al. 2000; Davis et al. 2000). Therefore to focus specifically on H⁺-gated DEG/ENaC channels rather than VR1, we studied large-diameter DRG neurons and examined the transient H⁺-gated currents.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell preparation

Mice with a disruption of the DRASIC gene were generated as previously described (Price et al. 2001). DRASIC +/+ and -/- littermates were anesthetized with halothane. Following decapitation, the spine was opened, the spinal cord rapidly removed, and the thoracic and lumbar DRGs were dissected. All procedures were reviewed and approved by the local Institutional Animal Care and Use Committee.

DRG neurons were cultured as described with some modifications (Price et al. 2000). In short, dissected ganglia were digested with collagenase IV (Worthington, Lakewood, NJ) followed by trypsin (GIBCO BRL, Rockville, MD) in 37°C Hank's balanced buffer. Following repeated washing and trituration, cells were placed in Petri dishes that were coated with poly-L-lysine and laminin (Sigma, St. Louis, MO).

We studied DRG neurons with a diameter of 30-35 µm and focused specifically on cells with transient acid-evoked currents. Cells were obtained from 20 DRASIC +/+ and 22 DRASIC -/- animals; most of the animals were littermates. Transient H⁺-gated currents were observed in 44% (n = 156) of DRASIC +/+ neurons and 34% (n = 180) of DRASIC -/- neurons.

Whole cell patch-clamp recording

Glass pipettes (Fisher Scientific, Palatine, IL) were prepared (3-6 M) with a Flaming/Brown micropippette puller (Sutter Instrument, model P-97/VF, Novato, CA). Whole cell recordings were made using an Axopatch 200 (Axon Instruments, Union City, CA) amplifier. pH stimuli were applied by a fast changing of solutions with a series of capillaries closely positioned to the cells. Digital recordings were captured by TL-1 DMA Interface and pClamp6.0 software (Axon Instruments). A brief depolarizing voltage step test was applied before experiments, and cells that did not yield voltage-gated Na⁺ currents were not included in the study. Membrane voltage was maintained at 70 mV. Series resistance was compensated (~60%) in some but not all studies; because the acid-induced current changes are relatively slow, series resistance compensation did not change the results. Experiments were performed at room temperature (20-23°C). Most cells were studied within 1 day after seeding, although some were studied on day 2.

Cells were superfused in buffer solutions containing (in mM) 128 NaCl, 5 MgCl₂, 1.8 CaCl₂, 20 HEPES, 5.4 KCl, 5.55 glucose, adjusted to pH 8 with N-methyl-D-glucamine (NMDG). The pipette solution contained (in mM) 100 KCl, 10 NaCl, 2 MgCl₂, 10 EGTA, 20 HEPES, 1 Na₂ATP, adjusted to pH 7.4 with KOH. For buffer solutions with pH lower than 6, HEPES was replaced with equal concentration of MES. In several studies we used the response to a pH 5 solution for purposes of comparison. We elected not to use the response to a pH 4 solution for such purposes because we found the responses to be much more variable and a pH 4 stimulus at times caused irreversible changes in current, especially in -/- neurons that showed substantial rundown (see RESULTS).

Data analysis

Data are means ± SE. Igor software (Wave Metrics) was used for curve fitting. Desensitization of proton-gated currents were fit to either single or double exponential functions using the following equations

where Y is current amplitude at time t and K₀ is the amplitude of the sustained component. For the single exponential function  is the time constant. For a double exponential function, _f is the time constant for the fast component and _s is the time constant for the slow component. K_f and K_s indicate the contributions to current amplitude from the fast and slow components, respectively. We compared fits of the desensitization curves with one or two exponentials using an F-test (Munson and Rodbard 1980). A Student's t-test was used to compare amplitudes and time constants of pH-gated currents; a paired t-test was used when comparing data from the same cell. A value of P < 0.05 was considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Acid-evoked currents from DRG of DRASIC +/+ and -/- mice

We isolated cells from DRG and examined acid-evoked currents in large-diameter sensory neurons. In wild-type neurons, application of a pH 5 solution elicited a current that rapidly activated and then desensitized in the continued presence of the stimulus (Fig. 1A). This response is similar to that previously reported (Akaike and Ueno 1994; Benson et al. 1999; Krishtal and Pidoplichko 1981a; Price et al. 2000). We also observed H⁺-gated currents in DRASIC -/- neurons; the persistence of current in the absence of DRASIC suggests that other channels also contribute to the acid-evoked current. However, the appearance of the current was much different in DRASIC -/- neurons (Fig. 1A). The most striking effect of losing DRASIC was slowing of desensitization (Price et al. 2001). The peak current amplitude also increased (Fig. 1B). These data indicate that DRASIC contributes to pH-gated current in large-diameter DRG neurons.

View larger version (22K): [in this window] [in a new window]   Fig. 1. Whole cell pH-gated currents in wild-type and dorsal root acid sensing ion channel (DRASIC) null dorsal root ganglion (DRG) sensory neurons. A: example of currents evoked by pH 5 in wild-type and DRASIC -/- DRG neurons. Bar indicates application of pH 5 solution. B: peak amplitude of pH 5 evoked currents from DRASIC +/+ and -/- neurons. DRASIC +/+, n = 49, diameter = 33.3 ± 0.5 µm and DRASIC -/-, n = 52, diameter = 32.1 ± 0.4 µm. Asterisk indicates P < 0.01.

To assess the kinetic consequences of losing DRASIC, we fit the current desensitization during a pH 5 stimulus to exponential functions (Fig. 2). We compared the fit of the data with one or two exponentials using an F-ratio test (Munson and Rodbard 1980). In 9 of 10 wild-type neurons tested, the time course of desensitization was fit significantly better by two than by one exponential function (P < 0.001). Most of the desensitizing current was accounted for by the fast time constant (_f, Table 1). In contrast, in 10 of 12 DRASIC -/- neurons tested, desensitization was fit by a single exponential function (Fig. 2); using two exponentials did not produce a statistically better fit to the data. The time constant () for desensitization in -/- neurons (1.51 ± 0.10 s) was significantly slower than the _f (0.24 ± 0.02 s) that accounted for the majority of the desensitization in +/+ neurons (Table 1). In addition to the desensitizing current, there was a small contribution of sustained current in wild-type neurons; the amplitude of this current was estimated by K₀ in the curve-fitting (Table 1). Although the average value tended to decrease in DRASIC -/- neurons, the change was not statistically significant. We reached the same conclusion if we used only a single exponential to fit data from +/+ neurons (Table 1).

View larger version (10K): [in this window] [in a new window]   Fig. 2. Desensitization of acid-evoked current in whole cell patches of wild-type and DRASIC null DRG neurons. Currents evoked by a pH 5 solution are shown in black, fit of data to a single exponential function It = K0 + K1 * exp(t/) is in green, and fit to a double exponential function It = K0 + Kf * exp(t/f) + Ks * exp(t/s) is in red. Current amplitudes were scaled to be equal. Desensitization of wild-type current was better fit by a double exponential function, whereas DRASIC -/- current was well fit by a single exponential function. Note that the red and green lines are superimposed in the DRASIC -/- neuron.

The increased peak amplitude and slow kinetics of the H⁺-gated current in DRASIC -/- neurons suggested that the net charge flow during the transient phase of the acid-evoked response would be increased. To test this, we integrated the current for the first 5 s of acid application. The charge entering DRASIC -/- neurons (11.94 ± 0.08 nC, n = 11) was significantly increased compared with wild-type cells (5.64 ± 0.05 nC, n = 7, P < 0.01). The increased cation influx might lead to a greater depolarization of DRASIC -/- neurons and thus presumably an enhanced excitatory response.

Recovery from desensitization

During the course of our experiments using DRASIC -/- neurons, we found that with repeated stimuli the amplitude of pH-gated currents progressively declined (i.e., rundown). To evaluate this, we applied repeated 2-s pH 5 stimuli at an interval of 22 s; this interval is sufficient for DRG currents (see following text) and heterologously expressed DRASIC and ASIC to recover from desensitization (Sutherland et al. 2001). In wild-type neurons, we observed little rundown during more than 5 min recording (Fig. 3). In other experiments carried out for longer than 10 min, we also saw little rundown (not shown). However, in DRASIC -/- neurons the current progressively decreased with repeated pH 5 stimuli.

View larger version (22K): [in this window] [in a new window]   Fig. 3. Rundown of whole cell pH-gated currents in wild-type and DRASIC null DRG neurons with repeated acid stimulation. Each pH 5 stimulation was ~2 s duration with an interval of 22 s. Peak amplitude of currents were normalized to current evoked by the 2nd pH 5 stimulus. DRASIC +/+, n = 6; DRASIC -/-, n = 5. All values after the 3rd stimulus were different between the 2 genotypes.

Earlier studies of DEG/ENaC subunits expressed in heterologous cells suggested that DRASIC currents recover from desensitization faster than currents from either BNC1 or ASIC subunits (Sutherland et al. 2001). To test the recovery from acid-induced desensitization, we exposed neurons to a pH 5 stimulus for 10-15 s, a duration sufficient to completely desensitize transient proton-gated current. Neurons were then bathed in a pH 8 solution for defined short intervals before they were again challenged with a pH 5 solution. The amplitude of the subsequent pH-gated current was then compared with that of the first. Figure 4 shows that 5 s after returning to a pH 8 solution, H⁺-gated currents had completely recovered in wild-type neurons. In DRASIC null neurons, current recovered at a similar rate, but failed to reach prestimulation values; 20 s following desensitization, current had recovered only to 84 ± 3% of the initial response (Fig. 4). The failure of the current to completely recover is consistent with the rundown shown in Fig. 3.

View larger version (18K): [in this window] [in a new window]   Fig. 4. Recovery from desensitization of whole cell pH-gated currents in wild-type and DRASIC null DRG neurons. An initial 10- to 15-s stimulation with pH 5 solution was followed by a 2nd pH 5 stimulation at indicated intervals. The peak amplitude of the 2nd pH-gated current was compared with that of the initial one in wild-type (n = 7-9) and DRASIC null DRG (n = 3-8). Because the rundown of current in DRASIC null neurons was greatest during the 1st 3 stimuli with pH 5, the experiments in both wild-type and DRASIC null neurons were begun after the 3rd pH application. Asterisks indicate P < 0.05.

pH sensitivity of current in DRG neurons

When expressed in heterologous cells, DRASIC, BNC1a, ASIC, and ASIC each show a different dependence on pH (Babinski et al. 1999, 2000; Bassilana et al. 1997; Champigny et al. 1998; Chen et al. 1998; de Weille et al. 1998; Lingueglia et al. 1997; Sutherland et al. 2001; Waldman et al. 1997a,b); of these, DRASIC is the most sensitive, especially for pH ranging from 7.1 to 6.5. Figure 5 shows the effect of pH on current normalized to that obtained at pH 5. The loss of DRASIC significantly reduced the sensitivity of DRG neurons to pH. For example, challenge with a pH 6.9 solution increased current 9.4 ± 2.4% in wild-type neurons but generated little current (1.6 ± 0.4%) in DRASIC null neurons. The reduced pH sensitivity is consistent with the loss of the DRASIC subunit, which is the most sensitive to pH, and consequently an increased contribution from other less sensitive subunits to the proton-gated currents.

View larger version (17K): [in this window] [in a new window]   Fig. 5. pH sensitivity of whole cell pH-gated currents in wild-type and DRASIC null DRG neurons. Currents were evoked by changing pH of extracellular solution from pH 8 to indicated value. To minimize the effect of rundown, all solution changes to the test pH were bracketed by changes to a pH 5 solution; the current amplitude at the test pH was then compared with the average amplitude of the 2 currents elicited by pH 5. Peak amplitudes of pH-gated currents were normalized to currents evoked by pH 5. This procedure was followed for neurons of both genotypes. DRASIC +/+, n = 7-12; DRASIC -/-, n = 7-16; asterisks indicate P < 0.05.

Blockers and stimulators of proton-gated current

Most DEG/ENaC channels are inhibited by amiloride, although the sensitivity varies considerably. Earlier studies indicated that, of the pH-sensitive DEG/ENaC channels, DRASIC is the least sensitive (Bassilana et al. 1997; Champigny et al. 1998; Chen et al. 1998; de Weille et al. 1998; Lingueglia et al. 1995; Sutherland et al. 2001; Waldman et al. 1997a,b). Amiloride inhibited current evoked by a pH 5 solution in large DRG sensory neurons of both genotypes (Fig. 6). However, at 10 µM amiloride, the loss of DRASIC almost doubled the inhibition. These data suggest an increased sensitivity to amiloride consistent with the loss of the relatively amiloride-resistant DRASIC subunit.

View larger version (19K): [in this window] [in a new window]   Fig. 6. Sensitivity to amiloride of whole cell pH-gated currents in wild-type and DRASIC null DRG neurons. A: example of pH 5 evoked currents that were blocked by 10 µM amiloride. Bars indicate duration of application. B: peak currents evoked by pH 5 solution in presence of indicated concentration of amiloride. Currents were normalized to currents evoked by pH 5. DRASIC +/+, n = 7-14; DRASIC -/-, n = 8-11; asterisk indicates P < 0.05.

Although the contribution of VR1 to H⁺-gated current is predominately in small-diameter DRG neurons (Caterina et al. 1997, 2000; Davis et al. 2000; Tominaga et al. 1998) rather than the large-diameter neurons studied here, we tested the effect of capsaicin, an activator of VR1. We applied both a pH 5 and a capsaicin (1 µM) stimulus to 32 wild-type DRG neurons (31 ± 6 µm diam); 16 (50%) showed a transient H⁺-gated current. Of these 16 neurons, 6 also showed a capsaicin response. The transient H⁺-gated current amplitude was similar in cells with (1,327 ± 348 pA) and without (1,430 ± 416 pA) a response to capsaicin (547 ± 213 pA). We obtained similar results in DRASIC -/- neurons (32 ± 5 µm); of 36 DRG neurons 14 (39%) showed transient H⁺-gated currents, and of these, 4 also showed a capsaicin current response (485 ± 314 pA). The transient H⁺-gated current amplitude was similar in cells with (1,019 ± 380 pA) and without (1,103 ± 317 pA) a response to capsaicin; in these studies, the values for H⁺-gated currents were influenced by rundown in the DRASIC -/- animals. Thus capsaicin-induced VR1 currents were similar in cells expressing transient H⁺-gated currents and did not appear to alter the transient acid-evoked currents. Moreover, the VR1 inhibitor capsazepine (10 µM) did not alter the transient response to a pH 5 stimulus in either DRASIC +/+ or -/- neurons (Fig. 7).

View larger version (21K): [in this window] [in a new window]   Fig. 7. Effect of the VR1 channel blocker capsazepine on whole cell pH-gated currents and response to capsaicin. Data are examples of currents evoked by pH 5, pH 5 in the presence of capsazepine (10 µM), and capsaicin (1 µM).

Response of DRG neurons to FMRFamide and FRRFamide

FMRFamide (Phe-Met-Arg-Phe-amide) and related peptides are a family of neuropeptides that act as neurotransmitters and neuromodulators in invertebrates. The FMRFamide-activated Na⁺ channel (FaNaCh) is a DEG/ENaC family member that serves as the receptor for FMRFamide in the mollusk Helix aspersa (Lingueglia et al. 1995). We previously showed that FMRFamide-related neuropeptides do not activate DEG/ENaC channels, but they slow desensitization of H⁺-gated current and induce sustained current in heterologous cells expressing DRASIC and ASIC, but not BNC1 (Askwith et al. 2000). Moreover, DRASIC and ASIC showed different responses to FMRFamide and FRRFamide. For example, ASIC responded to FMRFamide with the addition of a sustained current, whereas DRASIC showed a slower desensitization. These observations raised the question of how the loss of DRASIC would influence the response of DRG neurons to these peptides. If in DRASIC -/- neurons BNC1 remained as the only pH-gated DEG/ENaC protein, then we expected that the response to these peptides would be eliminated. However, if ASIC also contributed, we expected that the response would be present, but altered by the loss of DRASIC. Application of FMRFamide prior to acid stimulation altered the response to a pH 5 solution in wild-type DRG (Fig. 8), consistent with previous results (Askwith et al. 2000). FRRFamide produced an even more marked response, slowing the rate of desensitization. DRASIC null DRG neurons showed different results (Fig. 8). FMRFamide had no appreciable effect and FRRFamide slowed desensitization, although less strikingly than in wild-type neurons. Table 2 shows the effect of the two peptides on the kinetics of desensitization. In wild-type neurons, FMRFamide increased the nondesensitizing current (K_o), whereas FRRFamide prolonged both _s and _f (Table 2). In DRASIC -/- neurons, FMRFamide had no statistically significant effect, whereas FRRFamide produced a small but significant prolongation of the desensitization rate.

View larger version (20K): [in this window] [in a new window]   Fig. 8. Effect of FMFRamide and FRRFamide on H+-gated currents in wild-type and DRASIC null DRG neurons. FMRFamide and FRRFamide were present during times indicated by cross-hatched bars.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our data indicate that the DRASIC channel subunit makes an important contribution to the transient H⁺-gated currents that have been reported in sensory neurons for many years (Akaike and Ueno 1994; Krishtal and Pidoplichko 1981b; Petruska et al. 2000). This conclusion is consistent with earlier observations that DRG contain DRASIC transcripts (Waldman et al. 1997a) and that DRASIC protein is present in the soma as well as the peripheral nerve endings of large-diameter DRG sensory neurons (Price et al. 2001).

Loss of DRASIC alters pH-gated currents in large-diameter DRG neurons

Large-diameter DRG neurons from DRASIC null animals retained acid-evoked currents, indicating that DRASIC is not the sole DEG/ENaC subunit responsible for these currents. However, when DRASIC was absent, the biophysical properties were altered. Our data invite a comparison to biophysical data obtained when DRASIC and other DEG/ENaC subunits are expressed in heterologous cells.

First, loss of DRASIC markedly slowed the desensitization rate following a pH stimulus. Previous studies have indicated that after extracellular pH falls, DRASIC currents desensitize faster than those of any other subunit or combination of other subunits (Sutherland et al. 2001; Waldman and Lazdunski 1998). For example, at pH 6 the desensitization time constant for DRASIC was 0.32 s, whereas that for ASIC was 3.5 s and that for ASIC was 1.7 s (Sutherland et al. 2001). Our data suggest that DRASIC plays an important role in determining the fast desensitization of acid-evoked currents in the DRG, and that in its absence other subunits set the desensitization rate. However, we cannot exclude the additional possibility that in the absence of DRASIC, posttranslutional modifications of the remaining subunits contributed to the altered kinetics.

Second, DRASIC null DRG neurons showed a reduced pH sensitivity between pH 7 and 6. Of the H⁺-activated DEG/ENaC channels, DRASIC is the most sensitive to acid (Babinski et al. 1999; Bassilana et al. 1997; Champigny et al. 1998; Chen et al. 1998; de Weille et al. 1998; Lingueglia et al. 1997; Sutherland et al. 2001; Waldman et al. 1997a,b). For example, the pH values that induced half-maximal activation of heterologously expressed DRASIC, ASIC, and ASIC were 6.7, 6.4, and 5.9, respectively (Sutherland et al. 2001). BNC1 was the least sensitive to pH (Bassilana et al. 1997; Champigny et al. 1998; Lingueglia et al. 1997). Thus the reduced pH sensitivity when DRASIC is missing from DRG sensory neurons is consistent with its properties in heterologous systems.

Third, amiloride sensitivity increased in DRASIC null DRG neurons. DEG/ENaC currents show a wide range of sensitivity to amiloride block (Alvarez de la Rosa et al. 2000). ENaC channels are the most sensitive to amiloride, with 50% inhibition (IC₅₀) at ~100 nM (Canessa et al. 1994b). By comparison, DRASIC is the least sensitive with an IC₅₀ of 60-100 µM (Sutherland et al. 2001; Waldman et al. 1997a). Between these extremes, ASIC had an IC₅₀ of ~10 µM, ASIC an IC₅₀ of ~20 µM, and BNC1 an IC₅₀ of ~30 µM (Bassilana et al. 1997; Champigny et al. 1998; Chen et al. 1998; Sutherland et al. 2001; Waldman et al. 1997b). Thus finding an enhanced amiloride sensitivity in DRASIC null DRG neurons is consistent with the conclusion that DRASIC makes a major contribution to H⁺-stimulated currents.

Fourth, FMRFamide-like neuropeptides alter pH-stimulated DRASIC currents, and FMRFamide and FRRFamide generate a different pattern of response (Askwith et al. 2000). Both peptides slowed desensitization of the transient desensitizing current, although equivalent concentrations of FRRFamide had a greater effect. In wild-type DRG neurons we observed a similar response. Moreover, the lack of DRASIC eliminated the response to FMRFamide and attenuated the response to FRRFamide. These results indicate that DRASIC plays a prominent role in the response to these peptides.

Fifth, the rate of recovery from desensitization is also a distinguishing characteristic of DRASIC because it recovers faster than either BNC1 or ASIC (Sutherland et al. 2001). Thus we predicted a slowed recovery from desensitization in DRASIC null neurons. However, the loss of DRASIC did not alter the rate of recovery. It is possible that without DRASIC a combination of other H⁺-gated DEG/ENaC subunits recover from desensitization with a time course faster than that of any individual subunit alone. Alternatively, proteins associated with DRASIC or posttranslational changes in the complex might alter the rate of recovery from desensitization. It is also possible that the concurrent rundown in DRASIC null neurons might have masked differences in recovery between the two genotypes. Although our data do not reveal the mechanism for the striking rundown in DRASIC -/- neurons, these findings offer the opportunity to discover the regulation of the H⁺-gated DRG currents.

VR1 also generates H⁺-gated currents in DRG neurons (Caterina et al. 1997, 2000; Davis et al. 2000; Tominaga et al. 1998). However, those currents occur predominantly in the small-diameter nociceptive DRG neurons, and the acid-stimulated VR1 currents are sustained. This contrasts with the transient H⁺-gated currents we studied in large-diameter neurons. Moreover, our studies with capsaicin and capsazepine indicate that VR1 makes no appreciable contribution to the transient H⁺-gated currents in these large neurons.

DRASIC contributes to DRG acid-evoked currents as a heteromultimer

Previous studies have shown that both large- and small-diameter DRG neurons contain the transcripts and/or protein of several DEG/ENaC channel subunits, including DRASIC, BNC1b, ASIC, ASIC, ENaC, and ENaC (Chen et al. 1998; Drumond et al. 2000; Garcia-Anoveros et al. 2001; Lingueglia et al. 1997; Olson et al. 1998; Price et al. 2000, 2001; Waldman et al. 1997a). Biochemical studies indicate that some DEG/ENaC subunits can coassemble, with the interaction mediated at least in part via the N-terminus and M1 (Adams et al. 1997; Babinski et al. 2000; Bassilana et al. 1997). Moreover coexpression of more than one subunit in heterologous cells can alter the functional characteristics of the current (Babinski et al. 2000; Bassilana et al. 1997; Lingueglia et al. 1997; Zhang and Canessa 2001), suggesting that at least some subunits coassemble to generate heteromultimers. These studies raised the question of how DRASIC contributes to DRG H⁺-gated currents. DRASIC might participate as a subunit in a heteromultimeric complex with other DEG/ENaC channel subunits. Alternatively, DRASIC might form homomultimeric channels in neurons that also contained other DEG/ENaC channels.

Our data are most readily explained if DRASIC forms heteromultimeric channels in combination with one or more other DEG/ENaC subunits. If two separate channels each contributed to the current and one of the channels was removed by deleting DRASIC, then we would have expected the amplitude of current to decrease in DRASIC -/- neurons. However, this was not the case; the current increased suggesting a heteromultimeric construction in which the loss of DRASIC altered the channel properties. An alternative explanation would be increased expression of ASIC in DRASIC -/- animals. However, we found that disrupting the DRASIC gene did not increase transcripts of BNC1 or ASIC (Price et al. 2001). Thus DRASIC probably forms heteromultimeric proton-gated channels in combination with other DEG/ENaC subunits. Of course we cannot exclude the possibility of a small population of DRASIC homomultimeric channels or of two populations of heteromultimeric channels.

Our findings raise a question about the identities of the other H⁺-gated DEG/ENaC subunits that contribute to DRG acid-evoked current. DRASIC and ASIC, but not BNC1 currents respond to FMRFamide-like neuropeptides (Askwith et al. 2000). Therefore retention of a response to these peptides, albeit an altered one, in DRASIC -/- neurons suggests that ASIC contributes to the H⁺-gated currents. Our data do not directly assess the presence of BNC1. However, BNC1 protein is present in large-diameter neurons, as well as their peripheral sensory structures (Garcia-Anoveros et al. 1997; Price et al. 2000). Thus BNC1 might also play a role.

Physiological implications

Members of the DEG/ENaC family of cation channels appear to play an important role in mechanosensation. In C. elegans, the loss of MEC-4 or MEC-10 disrupts the normal response to touch (Driscoll and Chalfie 1991; Huang and Chalfie 1994), and UNC-8 and UNC-105 have been implicated in proprioception and detection of muscle stretch (Liu et al. 1996; Tavernarakis et al. 1997). In Drosophila melanogaster, Pickpocket is localized in multiple dendritic neurons at the site of mechanotransduction (Adams et al. 1998). In mammals, DRASIC and BNC1 have been detected in the large-diameter DRG neurons that respond to innocuous mechanical stimuli and in their specialized cutaneous extensions (Drumond et al. 2000; Garcia-Anoveros et al. 2001; Price et al. 2000, 2001). In addition, disrupting the mouse genes for DRASIC and BNC1 altered the stimulus-response characteristics of cutaneous mechanosensors (Price et al. 2000, 2001). Our results indicate that DRASIC makes an important contribution to H⁺-gated currents in these neurons. This observation raises the question of why the peripheral extensions of these neurons are not sensitive to acid, rather they are activated by innocuous mechanical stimuli (Lewin and Stucky 2000; Steen et al. 1992). Perhaps in the periphery of these mechanosensitive neurons DRASIC is tethered to extracellular matrix proteins that confer or enhance mechanosensitivity but mask pH-sensitive sites. Thus we suggest that H⁺-gated current may be a signature of DRASIC function in large mechanosensory neurons where acid is not the physiological stimulus.