Identification and Functional Characterization of a Voltage-gated Chloride Channel and Its Novel Splice Variant in Taste Bud Cells^*

Isolation of ClC-4 and Splice Variant ClC-4A cDNAs from Mouse Taste Bud Cells

Isolation of ClC-4 and ClC-4A cDNAs from taste bud cells was accomplished in two steps. 1) A 3′-end cDNA fragment was isolated from a single taste cell cDNA library by differential screening with cDNAs from gustatory and non-gustatory lingual epithelia. Construction and differential screening of single taste cell cDNA libraries were described previously (20, 27). Briefly, taste papillae were separated from the rest of a mouse tongue and enzymatically dissociated into individual cells. Taste bud cells were identified by their characteristic bipolar shape and transferred individually into Eppendorf tubes. First-strand cDNAs were synthesized from single cells with oligo(dT) primers, tailed with dATP and terminal transferase, and amplified by PCR. The PCR products were ligated into the λZapII vector to construct single taste cell cDNA libraries. Individual phage plaques were picked, and their insert cDNAs were amplified by PCR with vector-specific primers. The amplified products were size-fractionated by electrophoresis, transferred onto a nylon membrane, and screened with ³²P-labeled cDNAs prepared from non-gustatory lingual epithelium devoid of taste buds. Insert cDNAs that were not hybridized with the non-gustatory probe were presumed to be expressed selectively in taste cells, and their sequences were analyzed and searched against genome and expressed sequence tag data bases. One of the clones matched the 3′-end of human ClC-4. 2) Isolation of full-length cDNAs was carried out with mouse taste tissue cDNA. PCR primers were designed to encompass the entire mouse ClC-4 coding region (sense, 5′-AGGAGGATGATCTAGGACGCTGTC-3′; and antisense, 5′-TCTCAAAATAATGCCCATCTTATTGCT-3′). Two fragments were obtained and subcloned into the pCR-BluntII-TOPO vector. DNA sequence analysis and sequence alignment showed that the long fragment was the same as mouse ClC-4 (GenBank™ accession number XM_193014), whereas the short fragment was a splice variant of ClC-4, which we designated ClC-4A (GenBank™ accession number DQ186662). To determine whether these two isoforms are expressed in all three types of gustatory lingual papillae, a new set of PCR primers (sense, 5′-AGGAGGATGATCTAGGACGCTGTC-3′; and antisense, 5′-CATGCACCAGTGGGCCCTCTTTG-3′) was designed and synthesized to encompass the differentially spliced region, and PCRs were performed with first-strand cDNAs from circumvallate, foliate, and fungiform papillae and non-gustatory lingual epithelium.

In Situ Hybridization

Digoxigenin-labeled RNA probes (ClC-4, 2.2 kb) were used for in situ hybridization on fresh frozen sections (14 μm) as described (52). An alkaline phosphatase-conjugated anti-digoxigenin antibody in the presence of nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate was used for detection.

Immunohistochemistry

Polyclonal antiserum against a keyhole limpet hemocyanin-conjugated 13-amino acid peptide near the C termini of mouse, rat, and human ClC-4 and ClC-4A was raised in rabbits (Alpha Diagnostic International, Inc.). Frozen sections (10 μm) of murine lingual tissue (previously fixed in 4% paraformaldehyde and cryoprotected in 20% sucrose) were blocked in 3% bovine serum albumin, 0.3% Triton X-100, 2% goat serum, and 0.1% sodium azide in phosphate-buffered saline for 1 h at room temperature and then incubated overnight at 4 °C with the polyclonal antiserum (1:1000 dilution). The secondary antibody used was Cy3-conjugated goat anti-rabbit Ig (Jackson ImmunoResearch Laboratories, Inc.).

To determine the type of cells expressing ClC-4/ClC-4A, double immunostaining was carried out on taste sections with the rabbit polyclonal antibody against ClC-4/ClC-4A and the mouse monoclonal antibodies against IP₃R3 (1:50 dilution; BD Biosciences) and SNAP-25 (1:1000 dilution; Sternberger Monoclonals Inc., Lutherville, MD). The secondary antibodies used were fluorescein isothiocyanate-conjugated anti-mouse and Cy3-conjugated anti-rabbit antibodies. Fluorescent signals were viewed under a Leica confocal microscope.

Heterologous Expression in Xenopus Oocytes and Electrophysiological Recordings

Murine ClC-4 and ClC-4A cDNAs and human ClC-4 (GenBank ™ accession number AB019432) were subcloned into the pCR-BluntII-TOPO expression vector. Capped sense cRNA was synthesized from the linearized expression constructs by T7 RNA polymerase using the mMessage mMachine in vitro transcription kit and tailed with a poly(A) tailing kit (Ambion Inc.). The synthesized cRNA products were phenol-extracted, ethyl alcohol-precipitated, and then dissolved in nuclease-free water at ~0.5 ng/nl for injection. Dumont stage V or VI oocytes were obtained from adult female laboratory-bred Xenopus laevis, and their follicles were removed by collagenase digestion. Oocytes were injected with 50 nl of 0.5 ng/nl cRNA and maintained at 18 °C for 4–6 days in modified Barth’s solution supplemented with 5 mm sodium pyruvate (53). Membrane currents were recorded using the two-electrode voltage-clamp technique in ND96 solution (96 mm NaCl, 2 mm KCl, 1.8 mm CaCl₂, 1 mm MgCl₂, and 5 mm NaHEPES (pH 7.4)). To determine anionic selectivity, 80 mm Cl⁻ was substituted with equivalent amounts of Br⁻, I⁻, or $N{O}_3^{-}$. In the study on the effect of pH on channel conductance, 5 mm HEPES (for pH 7.4) was replaced with 5 mm MES (for pH <7.0). For pharmacological analyses, inhibitors of niflumic acid (0.3 mm) or NPPB (at 0.1 mm) were dissolved in ND96 solution.

Recording and current-passing micropipettes with tip resistances <5 megaohms when filled with 3 m KCl were pulled from glass capillaries (A-M Systems, Inc., Carlsborg, WA) using a horizontal puller (Sutter P-80/PC). Currents were recorded with a GeneClamp 500 amplifier, digitized with a Digidata 1200 A/DD/A system, and stored on a computer running pCLAMP 6 software (all from Axon Instruments, Foster City, CA). Currents were low pass-filtered at 2 kHz and are shown without subtraction of leakage currents.

Isolation of ClC-4 cDNAs

To identify genes that are selectively expressed in taste bud cells, we isolated individual taste bud cells and constructed single cell cDNA libraries with the λZapII vector (20, 27). Individual phage plaques were picked, and their insert cDNAs were amplified by PCR with vector-specific primers. The amplified products were size-fractionated by electrophoresis and transferred onto a nylon membrane. Southern hybridization of the nylon membranes showed that many inserts from the single cell cDNA libraries could hybridize with the radiolabeled cDNAs prepared from non-gustatory lingual epithelium devoid of taste buds. The few that were not hybridized were considered taste bud-specific (Fig. 1); insert DNAs of these clones were sequenced and bioinformatically analyzed. Among ~200 taste bud-specific clones analyzed, one clone (GA5508) with a 906-bp insert cDNA was 83% identical to the 3′-end sequence of the human voltage-gated, pH-sensitive chloride channel ClC-4 cDNA (GenBank™ accession number NM_001830.2).

A subsequent search against the mouse genome data bases with the GA5508 insert sequence and the human ClC-4 cDNA sequence identified a corresponding mouse genomic sequence. Its putative cDNA sequence was predicted using GenScan software, encompassing partial mouse ClC-4 cDNA sequences deposited in the GenBank™ Data Bank. PCR primers were designed to cover the entire coding region of mouse ClC-4.

Surprisingly, PCR amplification with taste circumvallate papilla cDNA yielded two fragments of 2.4 and 2.23 kb. DNA sequencing showed that the former is the same as the previously described mouse ClC-4 chloride channel with 747 amino acid residues (GenBank™ accession number XM_193014). However, the 2.23-kb fragment is a novel splice variant, lacking 155 bp near the 5′-end of ClC-4, including the presumed ClC-4 start codon (ATG). The amino acid sequence of the short form was deduced from the next in-frame start codon, which is probably the true start codon, there being no other start codon. A stop codon is present ~140 bp upstream. Thus, ClC-4A putatively encodes a protein of 687 amino acid residues, lacking the N-terminal 60 amino acid residues of ClC-4. Sequence analysis indicated that the deletion of 155 bp is a result of exon skipping and that the new N terminus of the ClC-4A protein starts within a presumed helix domain of ClC-4 (helix A), which may be expected to significantly affect the channel’s functions (Fig. 2). To determine whether these two isoforms are expressed in all three types of gustatory lingual papillae, a new set of PCR primers was designed and synthesized specifically for amplification of the differentially spliced region of the cDNAs. Reverse transcription-PCR detected both forms of the cDNA in all three types of taste papillae (circumvallate, foliate, and fungiform), but not in non-gustatory lingual epithelium (Fig. 3).

Localization of the ClC-4/ClC-4A RNA Transcripts and Proteins to Taste Bud Cells

To localize the RNA transcripts to taste bud cells, in situ hybridization was carried out with a 2.2-kb probe common to both ClC-4 and ClC-4A. The results demonstrated that the ClC-4/ClC-4A transcripts were selectively expressed in taste bud cells, but were absent from the surrounding non-gustatory lingual epithelial cells (Fig. 4). Sense probe controls showed no nonspecific hybridization to lingual tissue.

To determine the expression of the ClC-4/ClC-4A proteins in taste receptor cells, we used immunohistochemistry with antiserum to a peptide near the C termini of the ClC-4/ClC-4A proteins on sections of murine lingual tissue. This antibody was able to recognize both ClC-4 and ClC-4A. The immunostaining results indicated that the ClC-4/ClC-4A proteins were present on the cytoplasmic membrane. Some spotty staining seen in cells within the body of the bud suggested that the proteins could also be present in vesicles such as endosomes and synaptic vesicles (Fig. 5).

To determine the type of cells expressing ClC-4/ClC-4A, double immunostaining was carried out on taste bud-containing sections with the anti-ClC-4/ClC-4A antibody (produced from rabbit) and mouse monoclonal antibodies against IP₃R3 and SNAP-25, which are cellular markers for Type II and III taste bud cells, respectively. Confocal laser scanning microscope images showed that nearly all IP₃R3-expressing cells also expressed ClC-4/ClC-4A (Fig. 6). Interestingly, many, if not all, SNAP-25-expressing cells also expressed these ClC channels. Because Type II cells are taste receptor cells and because Type III cells make synapses with afferent neurons, the expression of ClC-4/ClC-4A in these two types of cells suggests that these chloride channels may play a major role in bitter, sweet, and umami taste signal transduction, modulation, and transmission.

Functional Characterization of the ClC-4 and ClC-4A Channels

To characterize the function of ClC-4 and ClC-4A, we subcloned their cDNAs into the pCR-BluntII-TOPO expression vector, synthesized the capped sense cRNA with in vitro transcription, and tailed the cRNA with poly(A). The cRNA was then purified and injected into Xenopus oocytes. Membrane currents were recorded using the two-electrode voltage-clamp technique 4–6 days after injection. Strong outward currents (which were absent in the control oocytes) were recorded in the ClC-4 cRNA-injected oocytes.

The regulation of human ClC-4 activity by external pH has been reported, although the results appear contradictory to each other (54, 55). To examine the effect of external pH on these chloride channels in our system and to determine whether pH exerts any differential effect on mouse ClC-4 and ClC-4A, we perfused the oocytes with the bathing solutions buffered to different pH values. Extracellular acidification markedly reduced ClC-4-mediated currents (Fig. 7, left panel), indicating that mouse ClC-4 channels were open in the pH range of 7.5 to 6.5 and began to close as the pH fell from 6.5. In contrast, currents in mouse ClC-4A-injected oocytes at pH 7.5 and 6.5 were close to the basal level as recorded from the control oocytes, and strong outward currents were recorded at pH 6.0 and 5.5, indicating that ClC-4A was closed at pH 7.5 and 6.5, but open at the lower pH (Fig. 7, middle panel). We could not reliably record the chloride currents below pH 5.5 or 5.0 because the endogenous currents of the Xenopus oocytes became more active in these more acidic media.

Our mouse ClC-4 pH sensitivity was in full agreement with previously reported human ClC-4 results from two groups (55, 56), but not with the results from a third group (54). To confirm that this is not due to experimental variation among different laboratories, we expressed the human ClC-4 cDNA isolated from a human taste cDNA library and found that its characteristics are identical to those reported by the first two groups (55, 56): open at neutral pH and closed at acidic pH (Fig. 7, right panel).

To determine the ion selectivity of ClC-4 and ClC-4A, chloride was replaced with other anions (Fig. 8). The conductance sequences of murine ClC-4 and ClC-4A were I⁻ = $N{O}_3^{-}$ > Br⁻ > Cl⁻ (Fig. 8A) and I⁻ ≫ Br⁻ = Cl⁻ > $N{O}_3^{-}$ (Fig. 8 B), respectively. These ion selectivities are quite different from those reported for human ClC-4 (55, 56). To confirm our recording system, we expressed and recorded from human ClC-4 and found that the conductance sequence we measured was nearly identical to that reported by the two groups (55, 56): $N{O}_3^{-}$ > Cl⁻ > Br⁻ > I⁻ (Fig. 8C). These data indicate that all three chloride channels have unique ion selectivities and that the 2% difference in amino acid sequence between human and mouse ClC-4 is critical in conferring anion selectivity.

To pharmacologically characterize these ion channels, we tested the effect of the chloride channel inhibitors niflumic acid and NPPB at the optimal pH for each channel. The results indicated that niflumic acid had only a slight inhibitory effect on human and mouse ClC-4 at pH 7.4 (Fig. 9, left and middle panels), but significantly inhibited the conductance of ClC-4A at pH 5.5 (right panels). NPPB had no apparent effect on the conductance of human and mouse ClC-4 channels, but did slightly inhibit ClC-4A (Fig. 10).