An Alternative Splicing Product of the Murine trpv1 Gene Dominant Negatively Modulates the Activity of TRPV1 Channels*

Transient receptor potential vanilloid 1 (TRPV1), or vanilloid receptor 1, is the founding member of the vanilloid type of TRP superfamily of nonselective cation channels. TRPV1 is activated by noxious heat, acid, and alkaloid irritants as well as several endogenous ligands and is sensitized by inflammatory factors, thereby serving important functions in detecting noxious stimuli in the sensory system and pathological states in different parts of the body. Whereas numerous studies have been carried out using the rat and human TRPV1 cDNA, the mouse TRPV1 cDNA has not been characterized. Here, we report molecular cloning of two TRPV1 cDNA variants from dorsal root ganglia of C57BL/6 mice. The deduced proteins are designated TRPV1α and TRPV1β and contain 839 and 829 amino acids, respectively. TRPV1β arises from an alternative intron recognition signal within exon 7 of the trpv1 gene. We found a predominant expression of TRPV1α in many tissues and significant expression of TRPV1β in dorsal root ganglia, skin, stomach, and tongue. When expressed in HEK 293 cells or Xenopus oocytes, TRPV1α formed a Ca2+-permeable channel activated by ligands known to stimulate TRPV1. TRPV1β was not functional by itself but its co-expression inhibited the function of TRPV1α. Furthermore, although both isoforms were synthesized at a similar rate, less TRPV1β than TRPV1α protein was found in cells and on the cell surface, indicating that the β isoform is highly unstable. Our data suggest that TRPV1β is a naturally occurring dominant-negative regulator of the responses of sensory neurons to noxious stimuli.

Transient receptor potential vanilloid 1 (TRPV1), or vanilloid receptor 1, is the founding member of the vanilloid type of TRP superfamily of nonselective cation channels. TRPV1 is activated by noxious heat, acid, and alkaloid irritants as well as several endogenous ligands and is sensitized by inflammatory factors, thereby serving important functions in detecting noxious stimuli in the sensory system and pathological states in different parts of the body. Whereas numerous studies have been carried out using the rat and human TRPV1 cDNA, the mouse TRPV1 cDNA has not been characterized. Here, we report molecular cloning of two TRPV1 cDNA variants from dorsal root ganglia of C57BL/6 mice. The deduced proteins are designated TRPV1␣ and TRPV1␤ and contain 839 and 829 amino acids, respectively. TRPV1␤ arises from an alternative intron recognition signal within exon 7 of the trpv1 gene. We found a predominant expression of TRPV1␣ in many tissues and significant expression of TRPV1␤ in dorsal root ganglia, skin, stomach, and tongue. When expressed in HEK 293 cells or Xenopus oocytes, TRPV1␣ formed a Ca 2؉ -permeable channel activated by ligands known to stimulate TRPV1. TRPV1␤ was not functional by itself but its coexpression inhibited the function of TRPV1␣. Furthermore, although both isoforms were synthesized at a similar rate, less TRPV1␤ than TRPV1␣ protein was found in cells and on the cell surface, indicating that the ␤ isoform is highly unstable. Our data suggest that TRPV1␤ is a naturally occurring dominant-negative regulator of the responses of sensory neurons to noxious stimuli.
Homologues of Drosophila transient receptor potential (TRP) 1 protein form a rapidly growing family of non-selective cation channels known as the TRP superfamily. The TRP channels are involved in a large variety of cellular functions including receptor and store-operated Ca 2ϩ entry (1,2), Ca 2ϩ transport (3,4), temperature sensation (5,6), and trace metal detection (7). The temperature-sensing TRP channels consist of at least 6 members with temperature thresholds ranging from as low as Ͻ17°C for extreme cold to Ͼ53°C for extreme heat (reviewed in Refs. 6 and 8). Among them, TRPV1 has received a great deal of attention because it was the first cloned channel that responded to pain-producing heat (Ͼ43°C) and acid stimuli and it is the receptor for capsaicin, the pungent ingredient of hot chili pepper (5,9). These functional features and the predominant expression of TRPV1 mRNA in the small diameter neurons of the rat dorsal root ganglia (DRG) provide support for TRPV1 as a key player in nociception of primary afferent neurons. Subsequent studies showed that TRPV1 is activated by the endogenous cannabinoid receptor ligand anandamide, N-arachidonyldopamines, and several lipoxygenase products, such as 15-hydroxyeicosatetraenoic acid (10 -13) and its activity is strongly potentiated by inflammatory mediators (14), suggesting a role of TRPV1 besides nociception. Indeed, TRPV1 is expressed in non-sensory tissues (15). Detailed studies have established that such signaling steps involved in inflammatory responses as phosphorylation by protein kinase A (16), protein kinase C (14), and breakdown of phosphatidylinositol 4,5-bisphosphate (17), all greatly enhance TRPV1 activity through different mechanisms. Therefore, TRPV1 is a polymodal detector of both noxious pain stimuli of the mammalian somatosensory system and a mediator of inflammatory pain.
Knockout of the trpv1 gene in mice showed severe deficiency of sensory neurons to vanilloid compounds, protons, or heat (Ͼ43°C) and impaired thermal hypersensitivity in response to inflammation (18,19). However, these studies were done without a characterization of the mouse trpv1 gene. Expression data for TRPV1 in mouse is also lacking. Because mice are commonly used for genetic manipulations, we decided to clone the mouse TRPV1 cDNA and characterize its protein function in heterologous systems. Here we show that at least two splice variants of TRPV1 cDNAs are present in several tissues. One of them, TRPV1␣, is equivalent to the widely studied rat and human TRPV1. The other one, TRPV1␤, encodes a dominantnegative subunit of the TRPV1 channel.

EXPERIMENTAL PROCEDURES
Molecular Cloning of Murine TRPV1␣ and TRPV1␤-DRG from four C57BL/6 mice were dissected, pooled, and placed in 1.2 ml of TRIzol reagent (Invitrogen) after the animals were sacrificed by cervical disarticulation followed by decapitation. The DRG were homogenized in the TRIzol solution in a glass-glass Dounce homogenizer and the homogenate was passed through a 20-guage and then a 25-guage syringe needle several times. Total DRG RNA was extracted following the manufacturer's protocol (Invitrogen). First strand cDNA was prepared from 15 g of the total RNA using SuperScript II reverse transcriptase (Invitrogen) and random hexamers (Amersham Biosciences). PCR primers for murine TRPV1 cDNA were designed according to the genomic sequences of the murine trpv1 gene (Celera scaffold: GA_x5J8B7W5FBR:1500001.2000000 and GenBank TM accession number AL663116), which were obtained by BlastN search using the se-quence of rat TRPV1 cDNA. Two fragments of murine cDNA were amplified from the first strand DRG cDNA using primer pairs S1, 5Ј-ATGGAGAAATGGGCTAGCTTAG-3Ј and A1, 5Ј-CTAGGCGATCAC-CTCCAGCAC-3Ј for fragment F1 and S2, 5Ј-CTGTCCAGGAAGT-TCACTGAATG-3Ј, and A2, 5Ј-GTGTCCCTCATTTCTCCCCT-3Ј for fragment F2 by PCR in 50-l reactions containing 1ϫ Pfu buffer (10 mM KCl, 10 mM (NH 4 ) 2 SO 4 , 2 mM MgSO 4 , 20 mM Tris-Cl, pH 8.75, 0.1% Triton X-100, 0.1 g/ml bovine serum albumin), 0.2 mM each dATP, dGTP, dCTP, and dTTP, 0.2 M each of the sense and antisense primers, 0.25 unit Pfu polymerase, and 1 l of the first strand cDNA. The reactions were carried out by heating to 95°C for 5 min followed by 30 cycles of 95, 56, and 72°C each for 1 min and a final extension at 72°C for 10 min. The PCR products were subcloned to pAGA3 vector (Gen-Bank TM accession number AY452085), which was modified from pAGA2 (20) by replacing the polycloning site with the sequence 5Ј-CCATG-GATATCGATGGATCCCGGGTCGACGGTACCTCTAGATAACTAG-3Ј. The sequences of all PCR clones were determined by the Rightmire DNA Sequencing Facility of the Ohio State University. The full-length murine TRPV1␣ and TRPV1␤ were assembled in pcDNA3 (Invitrogen) and pAGA3 vector by joining the EcoRI site contained in the F1 and F2 fragments using standard subcloning techniques.
Distribution of TRPV1␣ and TRPV1␤ in Mouse Tissues-Total RNA and first strand cDNA were prepared from mouse tissues as described above. The presence of TRPV1␣ and TRPV1␤ messengers was determined by PCR using primers: S3, 5Ј-CTGTCCAGGAAGTTCACT-GAATG-3Ј and A3, 5Ј-CTAGTAGAAGATGCGCTTGAC-3Ј. The quality and amount of cDNA prepared from each tissue were verified by amplification of glyceraldeyde-3-phosphate dehydrogenase using primers: GS, 5Ј-TCCTGCACCACCAACTGCTTAGC-3Ј and GA, 5Ј-CACCAC-CCTGTTGCTGTAGC-3Ј. The PCR conditions were 95°C for 1 min followed by 35 cycles of 1 min at 58°C, 1 min at 72°C, and 45 s at 95°C. A 10-min extension at 72°C was added at the end of the last cycle. Aliquots of the PCR products were subjected to agarose gel (2%) electrophoresis and visualized by ethidium bromide staining.
Southern blotting was used to distinguish the products of TRPV1␣ and -␤. Full-length cDNA for TRPV1␣ and -␤ was cut with BsrGI and included in the agarose gels as positive controls for hybridization conditions of the ␣and ␤-specific probes. DNA from the agarose gel was transferred to a Hybond nylon membrane (Amersham Biosciences) and Southern hybridization was performed in a Hybaid hybridization oven using end-labeled oligonucleotide probes: P␣, 5Ј-GTGATCGCCTACAG-TAGCAG-3Ј for TRPV1␣, and P␤, 5Ј-AGTGCTGGAGAACCGCCACG-3Ј for TRPV1␤, at 55°C for 2 h. The membranes were washed three times for 15 min each in 0.1% SDS and 2ϫ SSC solution at 55°C, sealed in plastic bags, and exposed to x-ray film.
Expression Constructs and Transfection in HEK 293 Cells-Both full-length TRPV1␣ and TRPV1␤ were inserted between the BamHI and EcoRI sites of pcDNA3 and the ClaI and XbaI sites of pAGA3. For bicistronic expression with enhanced green fluorescence protein (EGFP), the inserts plus a part of the cytomegalovirus promoter were excised from the pcDNA3 vector and subcloned into pIRES2-EGFP (Clontech) at the NdeI and EcoRI sites. For expression of TRPV1␣-TRPV1␣ (␣␣) or TRPV1␤-TRPV1␣ (␤␣) concatemers, the stop codon of the TRPV1 cDNA was eliminated by replacing the XhoI/EcoRI fragment coding for its C-terminal end with a PCR fragment that incorporated a sequence 5Ј-TGTCAACAGCAACAATTG-3Ј immediately after the codon for the last amino acid, Lys. The resulting full-length TRPV1␣ and TRPV1␤ were subcloned into pEGFP-N1 (Clontech) at the NheI and XhoI sites. The stop codon-removed full-length TRPV1␣ and TRPV1␤ were excised from the EGFP fusion constructs using ClaI and SalI and subcloned into a pAGA3 construct that contains the coding sequence for hemagglutinin epitope HA-pAGA3. The resulting plasmids were then cut open with SalI and XbaI and ligated with the unmodified TRPV1␣ excised from TRPV1␣/pcDNA3 using ClaI and XbaI. Both SalI and ClaI sites were filled in by treatment with Klenow fragment of DNA polymerase I prior to digestion by XbaI. The resulting constructs encode ␣␣ and ␤␣ concatemers with a short sequence, MYPYDVPDYAMDID, fused to the N-terminal ends and the sequence, CQQQQFCS-RAQASNSAVDD, added between the two TRPV1 coding sequences. For the ␣␣␣␣ and ␣␣␤␣ concatemers, a non-stopping ␣␣ concatemer was first made in HA-pAGA3 using the same strategy as described above with the second TRPV1␣ generated from TRPV1␣-EGFP. This concatemer was then joined with the ␣␣ and ␤␣ concatemers described above using again the same strategy involving the ligation between the filled SalI and ClaI ends at one side and XbaI at the other side. All concatemers were shuttled back to the pcDNA3 vector using ClaI/XbaI sites. Similar methods were used to make the ␣␤␤␣ and ␣␤␣␤ concatemers.
HEK 293 cells were grown in Dulbecco's modified Eagle's medium containing 4.5 mg/ml glucose, 10% heat-inactivated fetal bovine serum, 50 units/ml penicillin, and 50 g/ml streptomycin. For biochemical and electrophysiological experiments, cells were seeded in either 60-or 35-mm culture dishes and transfected with the desired DNA constructs using LipofectAMINE 2000 (Invitrogen) following the protocol provided by the manufacturer. For intracellular Ca 2ϩ measurements, cells were transfected with the desired DNA constructs in the wells of 96-well plates without pre-seeding using LipofectAMINE 2000. To prevent cell loss from subsequent washing, the wells were treated with 20 g/ml polyornithine (M r Ͼ 30,000, Sigma) for Ͼ15 min and rinsed once with Hank's balanced salt solution without Mg 2ϩ and Ca 2ϩ . For each well, the plasmid DNA (25 ng) and LipofectAMINE 2000 (0.4 l) were mixed in 50 l of Opti-MEM (Invitrogen) and added to the well before the addition of 100,000 cells resuspended in 100 l of the medium without antibiotics. The cells were incubated for 24 -28 h without medium change. The transfection efficiency was about 70% as determined using the pEGFP-N1 vector. Intracellular Ca 2ϩ Measurements and Whole Cell Recordings of HEK 293 Cells-Transiently transfected HEK 293 cells in 96-well plates were loaded with Fluo4/AM, and intracellular Ca 2ϩ was monitored using a fluid handling integrated fluorescence plate reader, Flex Station (Molecular Devices) as described previously (21). The extracellular solution contained 140 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 1.8 mM CaCl 2 , 10 mM glucose, 0.1% bovine serum albumin, and 15 mM Hepes, pH 7.4. All experiments were performed at 32°C. Whole cell recordings were made 24 h after replating HEK 293 cells transfected with TRPV1␣ or TRPV1␤ in pIRES2-EGFP vector in 35-mm dishes as described previously (21).
For electrophysiological recordings of TRPV1 and TRPV3 activities, oocytes were placed in a 50-l chamber, which was perfused with a bath solution containing (in mM): 140 NaCl, 5 KCl, 2 CaCl 2, 1 MgCl 2 , 10 glucose, and 10 Hepes, pH 7.40. The cell was impaled with two intracellular glass electrodes filled with 3 M KCl (for voltage monitor) and 3 M potassium acetate (for current injection) connected to a CA-1a amplifier (Dagan Co., Minneapolis, MN). The oocytes were clamped at either Ϫ80 or Ϫ40 mV and currents were continuously recorded using a chart recorder (Astro Med, Inc., West Warwick, RI). The pH 5.0 and 5.5 solutions were made of (in mM) 140 NaCl, 5 KCl, 2 CaCl 2, 1 MgCl 2 , 10 glucose, and 10 MES, whereas capsaicin, capsazepine, and ruthenium red were dissolved in the bath solution at the final concentrations and applied to the cells by perfusion. CaCl 2 was omitted in some experiments to minimize the Ca 2ϩ -dependent channel inactivation and the endogenous Ca 2ϩ -activated Cl Ϫ conductance.
Immunodetection of TRPV1 in HEK 293 Cells and in Xenopus Oocytes-TRPV1␣and TRPV1␤-transfected (0.5 g of DNA) HEK 293 cells seeded in polyornithine-treated 60-mm dishes were washed three times with ice-cold phosphate-balanced saline (mM: 136 NaCl, 1.4 KCl, 10 Na 2 HPO 4 , 1.7 KH 2 PO 4 , pH 8.0) 1 day after transfection and then incubated in 1.0 ml of phosphate-buffered saline containing 0.5 mg/ml Sulfo-NHS-LC-biotin (Pierce) on ice for 30 min. For Xenopus oocytes, the biotinylation was performed on 50 oocytes at 19°C 4 days after cRNA injection. The labeling was then terminated by the addition of 50 mM glycine and cells were washed three times with phosphate-buffered saline containing 50 mM glycine on ice. The washed cells were harvested immediately in 0.5 ml (or 1.5 ml for oocytes) of ice-cold RIPA buffer containing 150 mM NaCl, 50 mM Tris-Cl, pH 8.0, 0.5% sodium deoxycholate, 1% IGEPAL, 0.1% SDS, 5 mM EDTA, and protease inhibitors (1 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, 10 mM NaF). All procedures for streptavidin precipitation were carried out at 4°C. Aliquots of crude lysates (0.4 ml for HEK cells and 0.5 ml for oocytes) were incubated with agarose beads for 30 min and centrifuged at 5,000 rpm for 2 min. The supernatant was transferred to a new Eppendorf tube containing streptavidin-agarose beads. After incubation for 120 min, the beads were centrifuged at 5,000 rpm for 2 min, washed three times with the RIPA buffer, and resuspended in 20 l of 2ϫ SDS-PAGE loading buffer (1ϫ ϭ 62.5 mM Tris-Cl, 1% SDS, 10% glycerol, 10% ␤-mercaptoethanol, pH 6.8). For Western blotting of biotinylated proteins, one-half of the streptavidin-agarose-precipitated proteins were subjected to 8% SDS-PAGE and the proteins were detected using anti-TRPV1 (Oncogene) or anti-EGFP antibodies (BD Biosciences). For total proteins, small aliquots (usually 10 l or otherwise indicated) of crude lysates were subjected to Western blotting. Metabolic labeling and immunoprecipitation were performed essentially as previously described (22) except that anti-TRPV1 antibody was used.

Molecular Cloning of Two Splice Variants of TRPV1-Based
on the known cDNA sequences for rat and human TRPV1 and the mouse genomic sequence, we isolated the cDNA for murine TRPV1 by RT-PCR using total RNA prepared from the DRG of C57BL/6 mice. The open reading frame of the murine TRPV1 was predicted to contain 2517 base pairs, coding for a polypeptide of 839 amino acids. Several clones isolated by the RT-PCR matched exactly the predicted sequence, designated here as TRPV1␣ (GenBank TM accession number AY452083). However, a subset of clones lacked a 30-nucleotide region at position 1195-1224, coding for a putative TRPV1 variant of 829 amino acids, designated here as TRPV1␤ (GenBank TM accession number AY452084). A further analysis indicated that the ␤ variant arose from an alternative use of an intron recognition site (gt-) within exon 7 of the trpv1 gene (Fig. 1A). In TRPV1␣, this site is part of the "gtg" codon for Val 399 and the coding sequence continues to Pro 408 before the start of intron 7. In TRPV1␤, intron 7 starts 30 nucleotides earlier than the ␣ variant, resulting in the deletion of 10 amino acids corresponding to Val 399 -Pro 408 of the ␣ isoform from the encoded polypeptide. The deleted region is located near the end of the cytoplasmic N terminus, about 25 residues before the start of the transmembrane region (Fig. 1B).
To determine the expression patterns of the ␣ and ␤ isoforms of TRPV1, we isolated total RNA from 11 mouse tissues and performed RT-PCR. Primers were chosen to amplify a region encompassing the missing nucleotides in TRPV1␤ so that the products for the ␣ and ␤ isoform were 216 and 186 base pairs, respectively. As shown in Fig. 1C, all tissues tested expressed TRPV1␣ to varying degrees with DRG and liver being the most and the least abundant, respectively. The product for TRPV1␤ was barely detectable from DRG by ethidium bromide staining. To better distinguish the PCR products of the ␣ and ␤ isoforms, we performed Southern blot hybridization using oligonucleo-tide probes that specifically recognize either isoform. The specificity of the probes was confirmed by including the plasmid containing either TRPV1␣ or TRPV1␤, cut with BsrGI, on the same blot. Fig. 1D shows that the ␣ probe hybridized mainly with TRPV1␣ (upper), whereas the ␤ probe hybridized only with TRPV1␤ (lower). Whereas the Southern result for TRPV1␣ is consistent with that shown by ethidium bromide staining, the result for TRPV1␤ revealed its expression in brain, DRG, skin, stomach, and tongue, but not other tissues. Interestingly, although the relative levels of TRPV1␤ were low as compared with that of TRPV1␣, the ratios between the two splice variants were different in these tissues. Even though the highest levels of ␣ and ␤ variants were both found in DRG, stomach and tongue appeared to express a higher proportion of the ␤ variant than other tissues. Skin also appeared to express a relatively high proportion of TRPV1␤ because much less TRPV1␣ messenger was detected in skin than in skeletal muscle and spinal cord, but only skin showed a detectable level of TRPV1␤.
Functional Measurements of the Mouse TRPV1 Variants-We expressed TRPV1␣ and TRPV1␤ individually in HEK 293 cells and measured fluorescence changes in response to capsaicin in Fluo4-loaded cells at 32°C using a fluorescence microplate reader. In cells expressing TRPV1␣, capsaicin induced a dose-dependent rise in intracellular Ca 2ϩ concentration, as indicated by the increase in Fluo4 fluorescence (Fig. 2, A and  B). The EC 50 for capsaicin is 0.26 Ϯ 0.06 M (mean Ϯ S.E. of three experiments, performed in triplicate samples for each concentration), a value similar to the rat and human TRPV1 expressed in mammalian cells (23,24). The capsaicin-induced response was dose dependently blocked by ruthenium red (Fig.  2C), a known blocker of the TRPV channels (25,26). The remaining Ca 2ϩ transient seen in the presence of 10 M ruthenium red probably represents TRPV1-mediated Ca 2ϩ release resulting from the presence of the expressed channels on intracellular membranes (27), which still responded to the membranepermeable agonist but were not affected by the membraneimpermeable inhibitor. Similarly, addition of resiniferatoxin, anandamide, and N-arachidonyldopamine, as well as lowering the extracellular pH to 5.6, induced an increase in intracellular Ca 2ϩ in TRPV1␣-transfected cells (Fig. 2E), demonstrating that the mouse TRPV1␣ formed a Ca 2ϩ -permeable channel like its rat and human counterparts. In contrast, no change was induced by the drug treatments in cells that expressed either TRPV1␤ or the vector DNA (Fig. 2, B and E). Reducing the extracellular pH to 5.6 caused a similar transient rise in intracellular Ca 2ϩ in both TRPV1␤-and vector-transfected cells (Fig. 2E). This transient intracellular Ca 2ϩ increase was also consistently seen in untransfected HEK 293 cells (not shown) and thus represents the endogenous response to extracellular acidification. We have shown recently that 2-aminoethoxydiphenyl borate (2APB) activates rat TRPV1 and mouse TRPV1␣ (21). However, 2APB had no effect on cells that expressed TRPV1␤ (Fig. 2E).
In whole cell patch clamp recordings, application of 5 M capsaicin to TRPV1␣-transfected cells elicited an outwardly rectifying current, which reached its peak values within 5 s of the drug application and then rapidly declined (Fig. 2D). The current-voltage relationship of the capsaicin-induced activity, as revealed by the voltage ramp (Fig. 2D, inset), is typical of those for heterologously expressed rat and human TRPV1 channels (5,27,28). In contrast, 5 M capsaicin did not elicit any current in TRPV1␤-transfected cells even after more than 40 s of continued drug presence.
The functionality of TRPV1␣ and TRPV1␤ was further confirmed in Xenopus oocytes after injection of the cRNA for each isoform. Capsaicin-induced current was only seen in oocytes injected with the cRNA of TRPV1␣ but not that of TRPV1␤ (Fig. 3A). The EC 50 for capsaicin was 1.52 Ϯ 0.09 M (n ϭ 5) for the TRPV1␣-injected oocytes. The larger EC 50 value observed in the oocytes as compared with that in mammalian cells had also been noted for heterologously expressed rat and human TRPV1 (24). Lowering extracellular pH to 5.0 elicited a small increase in inward current (Ͻ100 nA at Ϫ80 mV) in both the uninjected and TRPV1␤-injected oocytes. However, the same treatment caused a large increase in inward current in TRPV1␣-injected cells (Fig. 3E). Therefore, when expressed individually only the ␣ but not the ␤ isoform of the mouse TRPV1 formed functional channels. We have shown that while the capsaicin-induced TRPV1␣ current was blocked by both ruthenium red and capsazepine, the 2APB-induced current was abolished by ruthenium red but only partially blocked by capsazepine (21). Like in the case of 2APB, 10 M capsazepine decreased the acid-induced mouse TRPV1␣ current to 63 Ϯ 5% (n ϭ 7) of its original value, whereas 1 M ruthenium red reduced the activity to 8 Ϯ 3% (n ϭ 7) of the original response (Fig. 3F). Capsazepine has been known to be a poor inhibitor of acid-and heat-induced responses for rat, but not human, TRPV1 channels (24).
Coinjection of TRPV1␣ and TRPV1␤ cRNAs in the ratios of 1:1 and 1:2 significantly decreased the capsaicin-induced current in oocytes by 85 and 97%, respectively (Fig. 3, A and B), indicating that coexpression of TRPV1␤ with TRPV1␣ reduced the amount of functional TRPV1 channels. As a control, coinjection of mouse TRPV3 and TRPV1␤ cRNAs at a 1:3 ratio did not change the activity elicited by a TRPV3 activator, 1 mM 2APB (21) (Fig. 3, D and E), indicating that the dominantnegative effect of TRPV1␤ is specific for the TRPV1 channels.
Protein Expression of the Mouse TRPV1 Variants-The lack of activity of TRPV1␤ could be because of either a functional or a trafficking and stability defect. In either case, coassembly of the ␣-␤ complex would be responsible for the dominant-negative effect of the ␤ isoform because functional TRPV channels are composed of four subunits (29). To better understand the ␤ subunit-induced inhibition of TRPV1 channels, we examined the expression levels of TRPV1␣ and TRPV1␤ by Western blotting using an anti-TRPV1 antibody in oocytes injected with 5 or 15 ng/cell TRPV1␣ cRNA, 15 ng/cell TRPV1␤ cRNA, or a mixture of TRPV1␣ and -␤ cRNA at 5 and 15 ng/cell, respectively. As shown in Fig. 4A (left), the level of TRPV1␤ protein in the total cell lysate was much lower than that of TRPV1␣ even when only one-third of TRPV1␣ cRNA was injected. The expression level of TRPV1 proteins in the coinjected cells was also lower than that injected with the cRNA for ␣ subunit alone. To learn the relative levels of TRPV1 proteins expressed on the plasma membrane, we performed cell surface biotinylation experiments (Fig. 4A, right). Significant enrichments of glycosylated TRPV1 proteins, as shown by the larger molecular weight bands, were noted in the fractions that were precipitated by streptavidin-agarose, indicating that the biotinylated samples represent the more completely processed proteins targeted to the plasma membrane. As a negative control, the biotinylated samples were also precipitated with glutathione-agarose beads. No protein was detected by the anti-TRPV1 antibody (not shown). Biotinylated TRPV1 proteins in oocytes expressing TRPV1␤ were also much lower than that in cells expressing TRPV1␣. However, the reduced TRPV1␤ expression on the cell surface appeared to result from the decrease in total TRPV1␤ protein, but not a defect in trafficking, because a similar amount of biotinylated proteins were detected when the TRPV1 proteins in the total lysates were adjusted to be equivalent (Fig.  4B). Therefore, TRPV1␤ protein appeared to be unstable as compared with TRPV1␣. TRPV1␤ also decreased the stability of TRPV1␣ when the two isoforms were coexpressed. However, this decrease cannot account entirely for the loss of TRPV1 channel activity when the two isoforms were coexpressed, which was 85 and 97% at the even lower ␣:␤ ratios of 1:1 and 1:2, respectively (Fig. 2B). Similar results were obtained from HEK 293 cells transiently transfected with the cDNAs for TRPV1␣, -␤, or ␣ and ␤ at a 1:3 ratio (Fig. 4C). Even with 3 times the amount of cDNA, the level of TRPV1␤ protein was much lower than that of TRPV1␣. The biotinylation experiments also revealed the reduced TRPV1␤ protein on the cell surface. However, this residual amount represents a significant association of TRPV1␤ with the plasma membrane as EGFP, a cytosolic protein, expressed in the same cells was not detected in the streptavidinprecipitated sample (Fig. 4C, right). To be certain that ␣ and ␤ subunits are expressed in the same cells and to maintain the desired ratios of ␣ and ␤ in the assembled tetrameric TRPV1 channels, we made concatemers composed of two or four ␣ and ␤ subunits in a single polypeptide in different combinations. As shown in Fig. 4D, the tetramer containing one ␤ subunit (␣␣␤␣) was expressed to a lower level than the tetramer with four ␣ subunits (␣␣␣␣). Tetramers with two ␤ subunits (␣␤␤␣ and ␣␤␣␤) were expressed at even lower levels. Similarly, the ␤␣ dimer, which was assumed to form channels composed of ␤␣␤␣, was expressed at a much lower level than the ␣␣ dimer. Thus, the presence of ␤ subunit reduced the stability of the assembled channels in a dose-dependent fashion. To learn the rate of protein synthesis for the two TRPV1 isoforms, we performed metabolic labeling and immunoprecipitation in TRPV1␣and ␤-transfected cells. The results showed that TRPV1␤ was synthesized at a similar rate as TRPV1␣ (Fig. 4E), ruling out the possibility that the decreased TRPV1␤ expression was because FIG. 3. Functional studies of TRPV1␣ and TRPV1␤ in Xenopus oocytes. Mouse TRPV1␣, TRPV1␤, and TRPV3 were expressed in Xenopus oocytes by injection of the corresponding cRNAs. Capsaicin, 2APB, and proton-activated currents were studied by two-electrode voltage clamp. A and B, 10 M capsaicin elicited inward current in cells expressing TRPV1␣ but not TRPV1␤. Coexpressing TRPV1␤ with TRPV1␣ greatly reduced capsaicin-elicited current. Currents were recorded at Ϫ80 mV in the presence of 2 mM Ca 2ϩ . A, representative traces from oocytes injected with the cRNAs for TRPV1␣, ␤, and ␣:␤ ratios as indicated. B, averages Ϯ S.E. of peak current amplitudes from 5 oocytes for each condition. *, p Ͻ 0.001, significantly different as determined by Student's t test. C and D, coexpression with TRPV1␤ did not alter the activity of TRPV3. Oocytes were injected with cRNA for TRPV3 or TRPV3 and TRPV1␤ at a 1:3 ratio. The response to 1 mM 2APB was measured at Ϫ40 mV in a nominally Ca 2ϩ -free bath solution. C, representative traces. D, averages Ϯ S.E. of peak currents for 7 oocytes each. E, acid, pH 5.0, induced current in TRPV1␣-but not TRPV1␤-expressing oocytes. Cells were held at Ϫ80 mV and stimulated in the presence of 2 mM Ca 2ϩ . F, sensitivity of proton-induced current in TRPV1␣-expressing cells to 10 M capsazepine (CPZ) and 1 M ruthenium red (RR). The cell was held at Ϫ40 mV and measured in a nominally Ca 2ϩ -free solution.
of reduced transcription and translation efficiencies, impaired RNA stability, or lowered transfection efficiency.
In Xenopus occytes, the expression level of the ␤␣ dimer was also much lower than that of the ␣␣ dimer (Fig. 5A). Coinjection of 3 times the amount of ␤␣ cRNA with the ␣␣ cRNA did not significantly change the expression level of the total TRPV1 protein, which should include both ␤␣ and ␣␣ concatemers. When voltage-clamped at Ϫ80 mV, occytes expressing the ␣␣ dimer responded to capsaicin (10 M) or protons, pH 5.0, just like those expressing TRPV1␣ alone. The ␤␣ dimer-expressing cells showed no response to capsaicin but a small and delayed response to protons (Fig. 5B). The differential loss of responses to vanilloids and to protons has been reported for a number of rat TRPV1 mutants (30,31). At the ␣␣ and ␤␣ ratios of 1:1 and 1:3, the response to capsaicin was reduced 12 and 66%, respectively, as compared with ␣␣ alone. To acid stimulation, the response was reduced 22 and 40%, respectively (Fig. 5C). Thus, the coinjection of the ␤␣ and the ␣␣ cRNAs at the 3:1 ratio significantly reduced the activity of the assembled TRPV1 channels even though no significant changes in the TRPV1 proteins were detectable. This, together with the results showing that a significant, although very low, level of ␤ subunit was found on the cell surface and that the functional loss was more pronounced than the decrease in cell surface expression in oocytes that coexpressed monomeric TRPV1␣ and -␤, suggests that in addition to being unstable, there are other mechanisms that prevent TRPV1␤ from forming functional channels.

DISCUSSION
The vanilloid receptor type 1 or TRPV1 has been the subject of many studies because of its significant roles in nociception and inflammatory pain (5,18,19). The recent cloning of the rat TRPV1 gene represents a major advancement in the understanding of the structure and function of this channel. Whereas the TRPV1 cDNAs from rat, human, chicken, and guinea pig are available for heterologous studies, the mouse cDNA sequence was only derived from the genomic data base despite the fact that the TRPV1 knockout mice have been made in two laboratories (18,19). In the current study, we isolated mouse TRPV1␣ cDNA from the DRG and confirmed its sequence as well as its function after heterologous expression in HEK 293 cells and in Xenopus oocytes. As predicted, the TRPV1␣ forms a Ca 2ϩ -permeable channel that is activated by capsaicin, protons, resiniferatoxin, anandamide, and N-arachidonyldopamine and blocked by ruthenium red and capsazepine. These features are similar to that of the rat and human TRPV1 channels. Moreover, the current study also revealed the existence of TRPV1␤ in the brain, DRG, skin, stomach, and tongue. The ␤ isoform is a product of an alternative splicing of the trpv1 gene and it differs from the ␣ isoform by missing 10 amino acids S-labeled TRPV1 proteins, representing de novo synthesis, were revealed by autoradiography following separation of precipitated samples by electrophoresis.
near the end of the cytoplasmic N terminus.
Our functional examinations indicated that TRPV1␤ did not form a functional channel when it was expressed alone, but instead, it exerted a dominant-negative effect on TRPV1␣ when they were co-expressed. At the protein level, although the two isoforms are synthesized at a similar rate, TRPV1␤ appears to be less stable than TRPV1␣. The coassembly of TRPV1␤ with the ␣ subunit also affected the stability of TRPV1␣, making less TRPV1 proteins available at the plasma membrane. However, no apparent defect in TRPV1␤ translocation to the plasma membrane was found in the current study because the remaining proteins were labeled by biotin from the extracellular side as efficiently as TRPV1␣. Because the residual amount of TRPV1␤ on the plasma membrane was not activated by factors known to stimulate TRPV1, two other possibilities exist. Either the residual proteins are not assembled into tetrameric channels properly or channels that contain ␤ subunits cannot be opened.
A number of splice variants for rat TRPV1, namely SIC (32), VR.5Јsv (33,34), and VR1L2 (34) have been reported. SIC and VR.5Јsv were not activated by capsaicin, protons, or heat. SIC was shown to be mechanosensitive; however, this feature could be attributed to its unique C-terminal sequence that was identical to that of TRPV4. Both VR.5Јsv and SIC lack a large part of the cytoplasmic N terminus. In addition, VR.5Јsv also lacks the entire exon 7. Interestingly, the lack of exon 7 also gave rise to VR1L2, which is otherwise identical to the full-length TRPV1. Exon 7 encodes 60 amino acids that include the later half of the third ankyrin-like repeat. The deletion of exon 7 does not change the reading frame of the subsequent sequence. It has been suggested that exon 7 of the trpv1 gene is predisposed to alternative splicing (34). The 5Ј splice site joining exon 7 and intron 7 of the trpv1 gene was Tgt and Cgt for rat and mouse, respectively. Both are considered atypical based on the Ggt-agG consensus sequence for intron recognition (35,36). The atypical 5Ј-3Ј splice sites might be responsible for exon 7 skipping in VR.5Јsv and VR1L2. A number of the splice variants encompassing exon 7 might also exist in human TRPV1 (37). The mouse TRPV1␤ reported here misses only the last 10 amino acids encoded by exon 7 and does not affect the ankyrinlike repeat. With Ggt being the 5Ј slice site, mouse trpv1␤ fits the consensus splicing sequence better than trpv1␣.
The functional consequence of the loss of exon 7 remains to be elucidated. The expression and functional data for VR1L2 have not yet been reported. Our data on the mouse TRPV1␤ is the first to show that the lost of the last 10 amino acids encoded by exon 7 makes the protein unstable and the channel inactive. Previous studies have shown that critical vanilloid binding sites of rat TRPV1 include a stretch of ϳ8 amino acids near the third transmembrane segment on the cytosolic side as well as Arg 114 and Glu 761 in the N and C termini, respectively (38,39). A number of negatively charged residues in or near the putative pore loop have also been implicated in the capsaicin or proton sensitivity of this channel (30,40). However, the 10 amino acids missing in TRPV1␤ are not known to be involved in binding to vanilloids or sensitivity to extracellular acidification. The instability of TRPV1␤ suggests that the loss of the 10 cytoplasmic N-terminal residues may lead to misfolding of the TRP protein, which in turn deviates the protein into the degradation pathway. The misfolding may also prevent the proper assembly of multimeric channels or, alternatively, it renders the channel inactive by affecting its gating mechanisms.
The cloning of mouse TRPV1␣ and TRPV1␤ adds an additional repertoire to the growing list of TRPV constructs that are currently used to examine the molecular and functional properties of native capsaicin receptors and thermal-sensitive channels. The dominant-negative feature of the TRPV1␤ may not be unique as a dominant-negative rat TRPV1 was created by changing three residues, NML, in the sixth transmembrane segment to FAP (31). Similarly, VR.5Јsv was indicated to have a dominant-negative effect on the full-length TRPV1 (34). However, there are a number of differences among these clones. First, TRPV1␤ is a native product found in a number of tissues and therefore, its presence and modulatory roles on the TRPV1 channel are physiologically relevant. Second, the NML676FAP mutant displayed no response to capsaicin but a partial response to protons. TRPV1␤, on the other hand, was insensitive to either capsaicin or protons. Third, TRPV1␤ contains a more complete TRPV1 sequence than VR.5Јsv and hence most of the structural and functional motifs of the functional TRPV1␣. The preserved motifs will allow TRPV1␤ to interact with other subunits of the native capsaicin receptors and modulate their function. This feature, for example, may be useful for studying heteromultimerization of the TRPV proteins. The current study, however, suggests that TRPV1 and TRPV3 do not form heteromultimers, contrary to a previous report (41).
As compared with TRPV1␣, TRPV1␤ expression was very low in tissues tested. However, the ratios between the splice FIG. 5. Relative activities of TRPV1 ␣-␤ concatemers expressed in Xenopus oocytes. The ␣␣ and ␤␣ dimers of TRPV1 were expressed in the oocytes either individually or in combination in ␣␣:␤␣ ratios of 1:1 and 1:3. A, Western blots showing the levels of TRPV1 dimer proteins in total lysates and streptavidin-agarose precipitates (SA-ag Prec., lower right panel). 1ϫ represents 5 ng/cell cRNA injected. B and C, currents elicited by 10 M capsaicin or acidification to pH 5.0. Cells were held at Ϫ80 mV and stimulated in the presence of 2 mM Ca 2ϩ . B shows sample traces for each combination of cRNA injection. C shows averages Ϯ S.E. of peak current amplitudes normalized to those obtained for ␣␣ in response to capsaicin (left) or pH 5.0 (right). The numbers of oocytes measured for each condition are shown in parentheses. *, p Ͻ 0.001; **, p Ͻ 0.01, significantly different than the ␣␣ control by Student's t test.
variants were not consistent across all tissues, suggesting that the expressions of the two TRPV1 isoforms are differentially regulated. Native capsaicin receptors undergo rapid desensitization after treatment with vanilloid ligands (5,42,43). A number of mechanisms may account for this effect. In the short term, calmodulin binds to the channel in a Ca 2ϩ -dependent manner and causes inactivation (44). Calcineurin has also been implicated to play a role in capsaicin receptor desensitization (45) and protein kinase A has been shown to reverse this process (16). In the long term, the influx of Ca 2ϩ and/or mitochondrial dysfunction is thought to cause apoptosis of vanilloidsensitive neurons and thus cell death (43,46). Because of its dominant-negative effect, up-regulation of TRPV1␤ could be a protective mechanism for the desensitization/down-regulation of capsaicin receptors in native neurons without causing cell death. It remains to be determined whether the expression of TRPV1␤ is altered after capsaicin treatment and/or under conditions of chronic neurogenic pain and inflammation.