Molecular Cloning of a Non-inactivating Proton-gated Na+ Channel Specific for Sensory Neurons*

We have cloned and expressed a novel proton-gated Na+ channel subunit that is specific for sensory neurons. In COS cells, it forms a Na+ channel that responds to a drop of the extracellular pH with both a rapidly inactivating and a sustained Na+ current. This biphasic kinetic closely resembles that of the H+-gated current described in sensory neurons of dorsal root ganglia (1). Both the abundance of this novel H+-gated Na+ channel subunit in sensory neurons and the kinetics of the channel suggest that it is part of the channel complex responsible for the sustained H+-activated cation current in sensory neurons that is thought to be important for the prolonged perception of pain that accompanies tissue acidosis (1, 2).


MATERIALS AND METHODS
Cloning of DRASIC-We used an anchored PCR approach to identify the sequences upstream and downstream of the expressed sequence tag (W62694). An double stranded adapter (anchor) was prepared by annealing the oligonucleotides GATTTAGGTGACACTATAGAATCGA-GGTCGACGGTATCCAGTCGACGAATTC and PO 4 -GAATTCGTCGA-CTG-NH 2 . The shorter oligonucleotide was protected with a 3Ј NH 2 group to avoid extension during the PCR reaction. This adapter was ligated to double stranded rat brain cDNA resulting in a cDNA with known sequences (the anchor) on both extremities. The so prepared anchored cDNA was used to amplify the 5Ј and the 3Ј end of the coding sequence by PCR. This was done using either the primer GATTTAG-GTGACACTATAGAA or TAGAATCGAGGTCGACGGTATC, which are identical to parts of the longer of the two adapter oligonucleotides together with either the sense primer (CACTACACGCTATGCCAAGG, for amplification of the 3Ј end) or the antisense primer (CCCAG-CAACTCCGACACTTC, for amplification of the 5Ј end) complementary to the expressed sequence tag (W62694). The PCR products were subcloned into Bluescript, and five clones each for the 5Ј PCR and for the 3Ј PCR were sequenced. The anchored PCR allowed us to identify the sequences upstream of the first ATG codon and downstream of the stop codon. However all clones isolated from brain contained introns with in frame stop codons and code for truncated proteins lacking the second transmembrane domain that was found to be essential for channel function (10). Analysis of the tissue distribution showed that high levels of the mRNA are only found in DRG. Primers flanking the coding sequence (sense: ACGAATTCTCCCTGGTCCAGCCATGAAAC, antisense: CCTCGAGCTAGAGCCTTGTGACGAGGTAA) that contained an EcoRI site (sense) or an XhoI site (antisense) were used to amplify the full-length coding sequence from DRG cDNA. The PCR product was digested with EcoRI and XhoI and subcloned into the EcoRI/SalIdigested PCI expression vector. One clone was sequenced on both strands, and two independent clones were sequenced on one strand using an Applied Biosystems sequencer. Unlike in brain, the three clones isolated from DRG code for full-length proteins.
Expression of the ASIC Subunits and Electrophysiology-COS7 cells were co-transfected with DRASIC cDNA in the PCI expression vector and an expression vector containing the CD8 receptor cDNA using DEAE-dextran. 3 days later, cells binding CD8 antibody-coated beads (11) were used for experiments. Ion currents were recorded using either the whole cell or the patch-clamp technique. The pipette solution contained (in mM): KCl 120, NaCl 30, MgCl 2 2, EGTA 5, HEPES 10 (pH 7.2). For the "0 sodium" solution NaCl was replaced by KCl. The bath solution contained in mM: NaCl 140, KCl 5, MgCl 2 2, CaCl 2 2, HEPES 10 (pH 7.3). Rapid changes in extracellular pH were induced by opening an outlet of a microperfusion system at a distance of ϳ50 m from the cell. Test solutions having a pH of less then 6 were buffered with 10 mM MES rather than HEPES. Experiments were carried out at room temperature (20 -24°C).
Northern Blot and in Situ Hybridization-4 g of total RNA from dorsal root ganglia of 7-day-old rats and 4 g of poly(A ϩ ) RNA from adult rat brain were separated on a 1% formaldehyde-agarose gel and subsequently transferred to nylon membranes. The blots were hybridized with a random prime 32 P-labeled fragment of the DRASIC cDNA corresponding to nucleotide 141-1145 in 6 ϫ SSC, 10 ϫ Denhardt's solution, 0.1% SDS, 100 g/ml herring sperm DNA, washed with 0.1 ϫ SSC, 0.1% SDS at 70°C, and subsequently exposed to a Fuji phosphoimager screen. For the in situ hybridizations on frozen fixed 10-m brain sections from adult Wistar rats, we used a 33 P-random primelabeled fragment of DRASIC corresponding to nucleotide 141-1145. Brain sections from adult rats were hybridized with the 33 P-end-labeled probes overnight at 37°C in 50% formamide, 2 ϫ SSC, and subsequently washed at room temperature in 1 ϫ SSC. Sections (6 m) and primary cultures of rat dorsal root ganglia were hybridized with doublestranded DNA fragments labeled by PCR with DIG-dUTP (sections), or fluorescein-12-dUTP (primary cultures). The probes used correspond to nucleotide 141-1145. Probe labeling, sample preparation, hybridization, and visualization of DIG nucleic acids with alkaline phosphataseconjugated anti-DIG antibodies was carried out following the protocols from Boehringer Mannheim. Primary cultures of DRG neurons from * This work was supported by the Centre National de la Recherche Scientifique (CNRS) and the Association Française contre les Myopathies (AFM). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Computer Analysis-The sequence alignment was computed with the GCG (Genetics Computer Group, Madison, WI) software package. All comparisons of sequences with data bases were done using the Blast network server at the National Center for Biotechnology Information (NCBI).

RESULTS AND DISCUSSION
Comparison of the ASIC protein sequence with the data base of expressed sequence tags identified one novel member of this family of ion channels. We used anchored PCR to clone the complete coding sequence from rat DRG. The DRASIC cDNA has an open reading frame of 1599 base pairs preceded by stop codons and codes for a protein of 533 amino acids. DRASIC belongs to the amiloride-sensitive Na ϩ channel (12-18)/degenerin (19 -21) family of ion channels and shares 53% sequence identity with its closest relative ASIC (Fig. 1). A DRA-SIC transcript of Ϸ2.6 kilobases was detected in total RNA of DRG (Fig. 2a). In brain poly(A ϩ ) RNA where ASIC mRNA is abundant (7), no DRASIC transcript was detectable. Furthermore a mouse multitissue Northern blot (CLONTECH) with poly(A ϩ ) RNA from brain, heart, spleen, lung, liver, skeletal muscle kidney, and testis did not give any signal (not shown) with the probe that labeled the DRASIC mRNA in total RNA from DRG, indicating that DRASIC is specific for sensory neurons. In situ hybridization confirmed those results (Fig. 2, b-d). DRASIC is expressed in DRG neurons and absent in brain. The small sensory neurons are thought to carry the nociceptive signals from polymodal sensory nerve endings and interestingly small neurons are intensely labeled. The specific expression in sensory neurons suggests that the DRASIC channel has properties required for a specific function of this type of neuron. Expression of DRASIC in COS cells induced a H ϩ -gated cation channel with properties clearly distinct from those of ASIC (7). A rapid decrease of the extracellular pH from pH 7.4 to pH 4 induces a fast rising, rapidly inactivating current followed by a much slower activating sustained inward current (Fig. 3a). Surprisingly, expression of DRASIC can induce both a rapidly and a slowly activating current. The kinetics of the DRASIC current very closely resemble the biphasic H ϩ -gated cation current described in sensory neurons (1). Both the transient and the sustained DRASIC current reverse at ϩ32 Ϯ 3 mV (n ϭ 5), which is close to the Na ϩ equilibrium potential of ϩ40 mV in the experimental conditions concerned (Fig. 3b). This indicates that the two components are highly selective for Na ϩ (gNa ϩ /gK ϩ ϭ 13.5). Unitary currents were recorded from outside-out patches in the absence of Na ϩ in the pipette (Fig. 3, c  and d). The slope conductance of DRASIC is with 12.6 Ϯ 0.2 picosiemens (n ϭ 3) (Fig. 3d), close to that reported for ASIC (14.3 picosiemens) (7). The unitary current has a reversal potential of ϩ62 mV (Fig. 3d), indicating an 11.5-fold higher selectivity of the channel for Na ϩ over K ϩ . Amiloride inhibits the transient current with a K 0.5 of 63 Ϯ 2 M (Fig. 3, e and f). The effect of amiloride on the sustained DRASIC current is complex. In the presence of 200 M amiloride where the transient current is inhibited by 68 Ϯ 5% (Fig. 3, e and f), the sustained current is higher than in the absence of amiloride (Fig. 3e). A closer examination of the pH dependence of the DRASIC current shows that the transient and the sustained phase can be clearly separated (Fig. 3, g-i). The transient current is activated when the pH drops only slightly (halfmaximal activation at pH 6.5 when stepping from pH 7.3; Fig.  3h) but requires an initial pH above 7 for full activation (Fig.  3i). On the contrary, the sustained current needs more important acidification (below pH 4) for activity (Fig. 3h) but may still be activated if the resting pH is far below pH 7 (Fig. 3i). The situation is similar in sensory neurons (1) where a slight acidification activates only the transient current, while both the transient and the sustained current are activated after more important drops of the extracellular pH. A H ϩ -gated cation channel capable of mediating a prolonged sensation of pain during tissue acidosis should not only be activated when the pH drops rapidly but also when the pH decreases slowly, since this is likely to happen during the onset of a tissue acidosis. Unlike ASIC, that requires a rapid (Ͻ Ͻ1 s) drop of the pH (not shown), DRASIC responds to slow decreases of the pH (Fig. 3j). If the extracellular pH is decreased gradually by approaching the cell slowly with the perfusion outlet, the first transient current disappears, while the sustained component still develops to its full size (Fig. 3j). The kinetics of the DRA-SIC channel and the fact that DRASIC mRNA is only present in sensory neurons, where it is abundant, suggest that DRASIC  (25) from Helix aspersa, and ␣ENaC is the epithelial amiloride-sensitive Na ϩ channel ␣ subunit (14). is part of the channel complex responsible for the sustained H ϩ -gated current in sensory neurons. However, there are important differences between the non-inactivating DRASIC current and the sustained current described in sensory neurons (1). To activate the sustained DRASIC current, the pH has to become very acidic (pH 4; Fig. 3h), while a tonic response in sensory neurons is already obtained at pH 6 (1). Furthermore, both the rapidly inactivating and the sustained phase of the DRASIC current are highly selective for Na ϩ , while in sensory neurons a transient Na ϩ -selective current is followed by a sustained current that discriminates only poorly between Na ϩ and K ϩ (1). Those differences between the DRASIC current and the native current indicate that more than just DRASIC is required to form the non-inactivating H ϩ -gated cation channel in sensory neurons. H ϩ -gated sustained Na ϩ -selective currents were never reported in sensory neurons, where DRASIC is well expressed, suggesting that DRASIC in sensory neurons has indeed properties distinct from the DRASIC channel expressed in COS cells. This might be due either to a specific posttranslational modification, such as phosphorylation, or to an association with other subunits. Heteromultimeric association of homologous subunits is commonly found with ion channels (22,23) and might be the link between the DRASIC subunit and the sustained non-selective current recorded in sensory neurons. Furthermore, relatives of DRASIC, the amiloride-sensitive Na ϩ channel (13,15,16,18) and the degenerins of Caenorhabditis elegans (19), even require several homologous subunits for correct function, and it would be surprising if this would not be the case for the H ϩ -gated cation channels. ASIC, that is also expressed in sensory neurons (7), is not the missing partner of a, whole-cell current in response to a rapid drop in pH from 7.3 to 4 recorded at Ϫ60 mV. b, current-voltage relations of the peak and the sustained current recorded in the whole-cell configuration. Reversal potentials of both components are ϩ32 Ϯ 3 mV (n ϭ 5). Na ϩ equilibrium potential at 40.1 mV. c, unitary currents recorded from an outside-out patch held at Ϫ50 mV and exposed to a drop in pH from 7.3 to 4. d, current-voltage relation of the unitary sustained current. The slope conductance is 12.6 Ϯ 0.2 picosiemens, the current reverses at ϩ62 Ϯ 2 mV. Data are means from three patches. e, whole-cell currents in response to a drop in pH from 7.3 to 4, in the presence and absence of 200 M amiloride. f, peak whole-cell current amplitude as a function of the amiloride concentration. pH steps were from pH 7.3 to pH 4. g, whole-cell currents recorded after a drop of the pH from different resting pH values to pH 4. h, peak and sustained whole-cell current amplitude as a function of test pH. Ordinate is expressed as a fraction of the saturation level of the Bolzmann fit. Peak current pH 0.5 ϭ 6.5, sustained current pH 0.5 ϭ 3.5. i, peak and sustained whole-cell current amplitude as a function of resting pH. Steps were to pH 4. Peak current pH 0.5 ϭ 6.9, sustained current pH 0.5 ϭ 4.9. The points in f, h, and i represent the mean Ϯ S.E. from at least four cells. j, comparison between the responses to a slow or a rapid shift in pH. The slow shift, indicated by the dotted ramp, was obtained by slowly approaching the cell with the outlet of the perfusion tube from a distance of ϳ0.5 cm, while the rapid shift was induced by suddenly opening the perfusion tube at a distance of ϳ50 m from the cell. Currents were recorded from DRASIC-transfected COS cells using either the whole-cell suction pipette technique with 140 mM extracellular Na ϩ and 30 mM intracellular Na ϩ or the outside-out patch-clamp technique with 140 mM Na ϩ in the bath solution 0 mM Na ϩ in the pipette. DRASIC since co-expression of both subunits yields currents that can be explained by two independent channels (not shown).
A diversity of H ϩ -gated cation channels is described in both sensory neurons (1, 4 -6) and in neurons of the central nervous system (6 -9). It is therefore likely that new members of this ion channel family will be discovered in the near future. The localization of their mRNAs and proteins should allow studies about the interaction of different H ϩ -gated cation channel subunits expressed in the same type of neuron and might lead to the identification of subunit combinations with properties identical to the native H ϩ -gated channels. The identification of subunits that associate with DRASIC is of particular interest because of the potential importance of sustained H ϩ -gated cation currents for the prolonged sensation of pain caused by acids. The development of blockers that are selective for a H ϩ -gated cation channel specific for sensory neurons, such as DRASIC, might lead to the discovery of new non-addictive analgesics.