Functional interactions between ligand-gated, voltage-, and Ca2+-activated ion channels are essential to the properties of excitable cells and thus to the working of the nervous system. The outer hair cells in the mammalian cochlea receive efferent inputs from the brain stem through cholinergic nerve fibers that form synapses at their base. The acetylcholine released from these efferent fibers activates fast inhibitory postsynaptic currents mediated, to some extent, by small-conductance Ca2+-activated K+ channels (SK) that had not been cloned. Here we report the cloning, characterization, and expression of a complete SK2 cDNA from the mouse cochlea. The cDNAs of the mouse cochlea α9 and α10 acetylcholine receptors were also obtained, sequenced, and coexpressed with the SK2 channels. Human cultured cell lines transfected with SK2 yielded Ca2+-sensitive K+ current that was blocked by dequalinium chloride and apamin, known blockers of SK channels. Xenopus oocytes injected with SK2 in vitro transcribed RNA, under conditions where only outward K+ currents could be recorded, expressed an outward current that was sensitive to EGTA, dequalinium chloride, and apamin. In HEK-293 cells cotransfected with cochlear SK2 plus α9/α10 receptors, acetylcholine induced an inward current followed by a robust outward current. The results indicate that SK2 and the α9/α10 acetylcholine receptors are sufficient to partly recapitulate the native hair cell efferent synaptic response.
The sensory organs of hearing and balance consist of highly differentiated epithelia containing hair cells. In the mammalian cochlea, there are two distinct types of hair cells. The inner hair cells contain the machinery that transduces changes in sound pressure stimuli into afferent signals, which are transmitted to the brain. The outer hair cells (OHCs), in contrast, mostly receive feedback modulation from the brain stem through cholinergic fibers that form synapses at their basal ends (Slepecky 1996). This efferent control, in the mammalian cochlea, regulates frequency selectivity, enhances sensitivity in the presence of background noise and plays a role in protection from noise overexposure.
The acetylcholine receptors (AChRs) of the OHCs are thought to be heteroligomers composed of α9 and α10 subunits (Elgoyhen et al. 2001; Maison et al. 2002). The expression of α9 subunits have been studied in the chick basilar papilla and in rat, human, and guinea pig cochlea (Elgoyhen et al. 1994; Hiel et al. 2000; Luebke 1996; Park et al. 1997; Simmons and Morley 1998). The α10 subunit was cloned from a rat cochlea cDNA library and transcripts were detected in the hair cells of the rat cochlea and cristae (Elgoyhen et al. 2001). Coexpression of the α9 and α10 AChRs subunits resulted in acetylcholine (ACh)-sensitive channels that differ from α9 homomeric receptor and that were similar to the AChRs expressed in the hair cells (Elgoyhen et al. 2001).
In vitro studies using whole cell recordings in isolated OHCs and cochlea sensory epithelia explants suggested that the Ca2+ influx through the AChRs subsequently gates small-conductance Ca2+-activated K+ channels (SK), initiating inhibitory hyperpolarization of the OHCs (Blanchet et al. 1996; Fuchs and Murrow 1992a,b; Housley and Ashmore 1991; Oliver et al. 2000; Yuhas and Fuchs 1999). Additionally, SK currents have been studied in a wide range of other tissues (Kohler et al. 1996; Panofen et al. 2002; Pennefather et al. 1985; Ro et al. 2001; Strobaek et al. 2000). SK channels have been cloned from rat and human brain (Kohler et al. 1996), from mouse colon smooth muscle (Ro et al. 2001), and from the trout CNS (Panofen et al. 2002), although no SK channel was cloned from the cochlea. SK channels are voltage-independent and gated solely by intracellular Ca2+. These channels are heteromeric complexes containing pore-forming α-subunits and calmodulin (CaM) (Keen et al. 1999; Schumacher et al. 2001; Xia et al. 1998). Although in vitro studies indicate that the ACh-activated hyperpolarizing conductance of the OHCs is mediated through SK channels, in vivo physiological characterization of the efferent influences on the OHCs suggested that efferent effects are not completely sensitive to apamin, a selective SK channel blocker. Thus other channels besides SK may also be involved (Yoshida et al. 2001). Nonetheless, the very fast activation of the SK channels by Ca2+ influx through AChRs indicated that the SK channels and the AChRs must be closely associated at the base of the OHC (Oliver et al. 2000). The cloning of the cochlea SK channels and the study of their functional expression and coexpression with α9/α10 AChRs are required to further understand their contribution to the native efferent synaptic response.
PCR amplification of cDNA fragments of small-conductance potassium channels SK1, SK2,and SK3 from adult mice inner ears
Total RNA was purified from 12-wk-old mice inner ears using Trizol reagent following the protocol suggested by the manufacturer (Invitrogen, Palo Alto, CA). RNA (1 μg) was then used as template to synthesize cDNA using SuperScript II reverse transcriptase. Aliquots of this cDNA were used to amplify specific SK isoform fragments using primers based on the published rat SK1, SK2,and SK3 sequences (Kohler et al. 1996). For SK1 amplification, 5′ primer CAGGCCCAGCAGGAGGAGTT and 3′ primer GGCGGCTGTGGTCAGGTG;for SK2 amplification, 5′ primer GGCATGGCAGCAGCAGCGGCACTA and 3′ primer CACATGCTTTTCTGCTTTGGTAA;and for SK3 amplification, 5′ primer GCAACTGCTTGAACTTGTGTA and 3′ primer GCAACTGCTTGAACTTGTGTA were used.
The amplified fragments were recovered after agarose gel separation and subcloned into pCRII-TOPO (Invitrogen). Complete DNA sequence analysis was performed in both directions and confirmed that the cloned fragments corresponded to the expected SK1, SK2, and SK3 regions.
Cloning of the full-length cochlea SK2 cDNA
The cochlea SK2 fragment obtained by PCR was labeled by random priming with digoxigening-UTP (Roche Molecular Biochemicals, Indianapolis, IN) and used as a probe in hybridization experiments to screen a mouse cochlea λZAP-cDNA library (1–2 days old, DBA mice) that was kindly provided by Dr. Guy Richardson (University of Sussex, Brighton, UK). The cDNA library was plated at 105 plaques/plate into 16 plates; a total of 1.6 million clones were screened using the SK2 probe. The sequence alignments were obtained using CLUSTAL W 1.8 (Thompson et al. 1994).
ACh α9 and α10 mouse cochlea cDNAs
The α9 AChR cDNA is a 1,608-bp RIKEN clone (F930004H14) obtained from a subtracted adult inner ear mouse cDNA library (http://genome.rtc.riken.go.jp/ Unigene Library number: Lib.9974). A representative α10 AChR cDNA was obtained by RT-PCR using a modified 3′ RACE approach (Barritt et al. 1999). Oligonucleotide primers were designed based on a deduced mouse ACh α10 transcript sequence that was generated by in silico identification of overlapping mouse ESTs using the rat ortholog (GenBank Accession number: NM_022639). The primers used to generate the 2,221-bp cDNA fragment were the 5′ primers GGTTTGAGGTCGTGTAAGAGGAG and ATATGGAAAGGGACGGAAGATGG and the 3′ primer GGGAGCACT CATTTATTATGGATC. Resulting amplification products were initially cloned into pCRII-TOPO and sequence-verified.
Construction of expression plasmids and expression in heterologous systems
The full-length SK2 cDNA was obtained from the excised phagemid by digesting with Eco RI and Xho I and subcloned into Eco RI and Xho I digested pIRES (Clontech, Palo Alto, CA). Using similar methods, the α9 and α10 AChRs' complete cDNAs were subcloned into pIRES. Each of the resulting expression plasmids contained either SK2 or α9 or α10 AChRs' complete cDNAs under the control of the CMV promoter. Additionally, a plasmid containing the green fluorescence protein was used as a reporter gene in all transfections. Plasmid DNA was isolated using Qiagen plasmid kits (Qiagen, Valencia, CA) and was transfected into HEK-293 cells by Ca2+ phosphate precipitation, following procedures suggested by the supplier (Invitrogen).
Whole cell currents and current–voltage relationship were obtained from green HEK-293 cells, 48 h after transfection.
For oocyte expression a vector, pNLE, was constructed by subcloning the 5′ and 3′ untranslated regions of a Xenopus β-globin gene from pSP64T (Krieg and Melton 1984) into pGEM9Zf (Promega, Madison, WI). The SK2, α9, and α10 AChRs' full-length cDNAs were then subcloned into pNLE. From these expression plasmids, RNA was in vitro transcribed using T7 RNA polymerase and injected into Stage V–VI oocytes as described (Calvo et al. 1994).
Oocytes were voltage-clamped using a two-microelectrode amplifier (Warner Instruments, Hamden, CT). Macroscopic K+ currents were recorded in bath solution containing (in mM) NaCl, 100; KCl, 2–30 (the concentration of NaCl was reduced on an equimolar basis); MgCl2, 0.5; HEPES (N-[2-hydroxyethyl]piperazine-N′-[2-ethanesulfonic acid]), 5 (pH 7.6 with NaOH). To block endogenous Ca2+-activated Cl– currents in oocytes, the bath solution contained 1 mM niflumic acid (Collier et al. 1996). The pipettes contained 3 M KCl. Because the channel is activated by Ca2+, we injected approximately 100 nl of 20 μM Ca2+ with a third pipette into the oocytes. Oocytes expressing robust apamin-sensitive currents were transferred onto an inverted microscope and inside-out patches excised to record single-channel activity. Single-channel data were acquired from this expression system using only oocytes with large currents at the whole cell level.
Standard patch-clamp recording techniques (Hamill et al. 1981) were used to record whole cell and single-channel currents from HEK-293 using Axopatch 200B amplifier (Axon Instruments, Union City, CA). The same amplifier was used to record single-channel currents from Xenopus oocytes expressing robust Ca2+-activated K+ currents. In all experiments, the cell capacitance was calculated as the ratio of total charge (the integrated area under the current transient) to the magnitude of the pulse (20 mV). The series resistance was compensated electronically. For single-channel recording, the amplitude histogram at a given test potential was generated. Leak-subtracted current recordings were idealized using a half-height criterion. Idealized records were used to construct ensemble-average currents, open probability, and to generate histograms for the distributions of open intervals. Single- and biexponential probability density functions were fitted to all open intervals. The number of channels in a patch were estimated using binomial analysis or from the stacking of the unitary events (Sigworth 1995). Peak open probability was determined from the ensemble current using the measured single-channel current amplitude and the estimated number of channels in the patch. Dequalinium chloride (DQ) was purchased from Tocris (Ballwin, MO). In experiments using HEK 293 cells the intracellular calcium concentration was fixed with a pipette Ca2+ concentration of 5 μM. The currents conferred by transfection of SK2 channels into HEK 293 cells were Ca2+ dependent. The bath solution contained (in mM): N-methyl-d-glucamine (NMG), 140; KCl, 4; CaCl2, 2; MgCl2, 1; HEPES, 10, pH 7.4 with methane sulfonic acid; and the internal solution contained (in mM): K-glutamic acid, 130; KCl, 10; EGTA, 10; HEPES, 10, pH 7.2 with KOH; CaCl2 with calculated free [Ca2+] of 500 nM. In these experiments sodium (Na+) was replaced by K+ (high external K+) to ensure that the cell's resting potential was clamped to approximately 0 mV to infer the correct holding and step potentials.
Niflumic acid, apamin, and all other chemicals were purchased from Sigma-Aldrich (St. Louis, MO). All experiments were performed at room temperature (22–23°C). Data are expressed as means ± SD. For single-channel analysis a custom-written software (Q-program) was used, as previously described (Rodriguez-Contreras and Yamoah 2001; Rodriguez-Contreras et al. 2002).
Cloning of the mouse cochlea SK2 channel
SK channel cDNA fragments were obtained using PCR and primers specific for the SK1, SK2, and SK3 channels based on the published sequences of these channels cloned from rat brain (Kohler et al. 1996). Using inner ear RNA as template in RT-PCR experiments, PCR products of the expected size for each of the SK1, SK2, and SK3 fragments were obtained. However, a second round of amplification was required to purify and subclone the SK1 and SK3 cDNA fragments. Sequence comparisons revealed homologies with the corresponding SK fragments previously cloned from rat brain (Kohler et al. 1996). Using the SK2-amplified fragment as a probe to screen a mouse cochlea cDNA library, a full-length SK2 cDNA was cloned and completely sequenced. The SK2 cDNA was 2,145 bp long and contained 75 nucleotides of a 5′-untranslated sequence with a partially conserved Kozak sequence, 1,725 nucleotides of the coding region, and 348 nucleotides of a 3′-untranslated sequence including a polyadenylation signal (the sequence was deposited with GenBank, Accession number: AY 123778). The cDNA encoded a protein consisting of 574 amino acids, which contained the functional domains described previously in the SK channels (Kohler et al. 1996). These domains include transmembrane segments S1–S6 and the pore region, characteristic of the K+ channel family of proteins, as well as the CaMBD. The protein sequence of the cochlea SK2 was identical to the one cloned from the mouse colon [not shown (Ro et al. 2001)] and to the SK2 channel cloned from the mouse heart [not shown (Xu et al. 2003)]. It also showed high homology to the SK2 channels from rat brain, human brain, chicken basilar papilla, and trout CNS (Fig. 1). The conserved amino acids among all the species are shown in white with black background, highlighting the high degree of SK2 sequence conservation among the species. Eighty percent of all the amino acids in the mouse SK2 sequence are identical in the 5 species shown, including trout. Sequence divergence is clustered almost exclusively to the cytoplasmic amino end in front of S1 and at the carboxy end, after the CaMBD moiety of the channel. However, within these regions, there are also subdomains that are completely conserved in all the sequences.
Mouse cochlea AChRs
Full-length α9 and α10 AChR cDNAs, which included the associated polyadenylation signal, were obtained. The α9 AChR is a RIKEN clone (F930004H14) obtained from a subtracted adult mouse inner ear cDNA library (Unigene Library number: Lib.9974), and contains a 1,608-bp cDNA representing the complete cochlea mouse α9 transcript. Sequence and BLAST analyses showed a 99.7 and a 99.9% homology with mouse cDNA (GenBank Accession number: AK010496) and mouse genomic exons (Accession number: XM_132045) sequences, respectively.
A representative α10 AChR cDNA was obtained by RT-PCR as described in methods (Accession number: NM_022639). BLAST searches showed that the mouse and rat α10 AChR sequences were 92.4% homologous at the nucleotide level and 97.8% at the amino acid level. More recently, a bacterial artificial chromosome clone (clone RP23-3D10; Accession number AC098723) was identified with a genomic fragment containing the Chrna10 gene (α10 AChR) and was placed on mouse chromosome 7 50.0 cM [7 (90631260–90636033 bp)] using the ENSEMBL BLAST (http://www.ensembl.org/Mus_musculus/blastview). Only a single nucleotide differed between the amplified α10 cDNA sequence and the associated Chrna10 exons and this nucleotide change did not alter the deduced amino acid sequence.
Figure 2 shows the best alignments of the deduced protein sequences of mouse cochlea α9 (Fig. 2A) and of the α10 AChRs (Fig. 2B), to the corresponding rat, human, and chicken AChRs. Deduced sequences of the mouse cochlea AChRs, as expected, contain cytoplasmic amino and carboxy termini and the characteristic nicotinic transmembrane segments, TM1–TM4. Overall the α9 AChRs sequences are 66% conserved and the α10 sequences 63% conserved. The TM3–TM4 cytoplasmic loop is longer and more divergent among the α9 (135 amino acids) than among the α10 AChRs (102 amino acids).
Heterologous expression of SK2 channels and coexpression with α9/α10 AChR
Injection of SK2 RNA into Xenopus oocytes resulted in the expression of a Ca2+-dependent, apamin-sensitive current.
Two types of experiments were performed. First, as described in methods, single calcium injections were used. After calcium injection, the current was turned on for about 5 min. In these experiments the amplitude of the currents were seen to decrease gradually as a function of time, presumably as a result of the buffering of the injected calcium by the oocyte; however, the traces shown in Figs. 3, 4, and 5 were all recorded immediately after the injections, and before the magnitude of the current began to decrease.
Second, experiments were also performed by providing calcium constantly, with a third pipette. In those experiments the K+ currents persisted. Approximately 70% of the current remained after 5 min if the “calcium/third pipette” remained impaled in the oocytes, perhaps because there was a steady-state leakage of calcium. However, these experiments were not done routinely because the “third electrode” occasionally made the ooctyes leaky.
Figure 3 shows that Xenopus oocytes injected with SK2 channels expressed macroscopic currents under conditions where only K+ currents could be recorded. The mean peak current recorded at a step voltage of 60 mV was 0.50 ± 0.02 μA (n = 11). Application of 10 nM apamin completely blocked the Ca2+-sensitive outward current (peak current at a step potential of 50 mV after application of 10 nM apamin = 0.007 ± 0.001 μA, n = 10: Fig. 3, A and B). These currents were voltage-insensitive. However, the magnitude of the current increased with increasing voltage, as expected, because of its dependency on the driving force. No currents were recorded in experiments with uninjected oocytes under the same conditions.
Whole cell currents and current–voltage relationship obtained from SK2-transfected HEK-293 cells, 48 h after transfection, are shown in Fig. 3, C and D. Cells were held at a holding potential of –80 mV and stepped to increasingly positive voltages. Using the experimental protocol, the channel activity showed an apparent voltage-independence. The currents were sensitive to apamin (data not shown) and to DQ (Fig. 3C), whereas the IC50 obtained for apamin was approximately 0.6 nM (Fig. 3E) and the IC50 for DQ was approximately 6 μM (Fig. 3F). In control experiments using cells transfected with GFP only, under the same conditions, no current was generated.
Single-channel recordings were obtained from SK2-injected oocytes, which had been shown to express large apamin- and Ca2+-sensitive currents at the whole cell level. As shown in Fig. 4, the resulting outward current was sensitive to the Ca2+ chelator, EGTA (Fig. 4, A, B, and E), and DQ (Fig. 4F). An example of the amplitude histograms used to determine the unitary current amplitude is shown in Fig. 4C. The estimated single-channel conductance from the regression line of the current–voltage relation in Fig. 4D was 18 ± 3 pS (n = 5, experiments on different oocytes).
When exposed to ACh, HEK-293 cells expressing α9/α10 AChRs yielded outward or inward current, depending on the holding potentials. Figure 5 shows traces from a HEK-293 cell expressing α9/α10 AChRs and exposed to ACh held at 50 mV (Fig. 5A) and –70 mV (Fig. 5B), respectively. No current was recorded from nontransfected cells after exposure to the same concentration of ACh (Fig. 5C). The amplitude of the currents induced in the transfected cells was dependent on the ACh concentration (Fig. 5D). Figure 5 also shows a representative response of a cell cotransfected with cochlear SK2 channels plus α9/α10 AChRs. In this cell, ACh application induced an inward current at a holding potential of –40 mV, followed by an outward current, indicated with an arrow (IK,Ca) in Fig. 5E. The inward current superimposes on the inward current of the trace without apamin, showing that the AChR current is not affected by the presence of apamin; only the outward current is blocked. Because the current trace was recorded from the same cell, with and without apamin, and because identical analog and digital filtrations were used, the current superimposes well with the inward current of the trace without apamin. Note that our rapid exchange system introduced some degree of noise into the recordings. The AChR current is not affected by the presence of apamin. The separate currents, without apamin (top) and in the presence of 2.5 μM apamin (bottom) are also shown in reduced size in Fig. 5E. In 10 of the 15 cells cotransfected with cochlear SK2 channels plus α9/α10 AChRs that were examined, ACh application induced inward followed by outward currents at a holding potential of –30 mV, whereas the other 5 cells had no response. The time-to-peak of the ACh response was 121 ± 34 ms (n = 5), whereas the time-to-peak of the outward current in coexpression experiments was 235 ± 76 ms (n = 4).
This study is the first report of the cloning and functional expression of a mouse cochlear SK2 channel and it provides the evidence that the activation of SK2 channels by Ca2+ through α9/α10 AChRs from the same species can be reconstituted in heterologous expression systems for detailed structure–function analysis. The coexpression of these channels partly recapitulates the ACh-induced OHC response.
Expression of SK channels in the mouse inner ear
The expression of SK2 channels in the cochlea was previously established biochemically. Oliver et al. (2000) demonstrated that the cochlea SK2 was localized at the base of the OHCs using a polyclonal antibody raised to a peptide in the carboxyl intracellular domain of the rat SK2 channel. We report the cloning and the molecular characterization of an SK2 channel from the cochlea. Although similarities between the cochlea SK2 and the brain and colon isoforms were anticipated, the results are reassuring given the unique role of SK2 in efferent modulation of OHC function (Oliver et al. 2000; Yuhas and Fuchs 1999). The SK2 channel reported here was cloned from a cochlea cDNA library; complementary PCR experiments described were performed using cDNA prepared from cochlea RNA. Our PCR experiments indicate that alternative spliced forms of the SK2 channel may be expressed in the cochlea (data not shown). However, the SK2 clone described in this manuscript was the only one obtained in the hybridization experiments and is likely to be the most highly expressed form.
Also, our results were consistent with previous reports that demonstrated in situ hybridization localization of the SK2 transcript in the OHCs of the rat cochlea. Moreover, it was shown that no hybridization signal could be detected in the adult rat cochlea when an SK3-specific probe was used (Dulon et al. 1998). Our results do not exclude the possibility of other channels of the SK family, such as SK4, being expressed in the cochlea. When expression of SK2, and SK3 was examined using PCR, SK2 was the one readily amplified in the mouse inner ear, SK3 appeared to be expressed at much lower levels. It is possible that SK3 transcripts were derived from the cochlea vasculature, given that these channels are known to be expressed in erythrocytes and smooth muscle (Grygorczyk and Schwarz 1983; Ro et al. 2001; Schwarz et al. 1989). Traces of blood are present in the cochlear tissue because perfusions were not performed before the dissections.
The SK2 protein sequences are highly conserved across species including fish, particularly the transmembrane and functional domains but also thoughout the whole protein in general. The high level of sequence conservation in the intracellular domains points to possible interaction with other proteins that may be required for the localization and functioning of these channels.
Expression of the cloned SK2 channel in Xenopus oocytes as well as in HEK-293 cells resulted in outward K+ currents that had the expected pharmacological profile of SK2 channels. The currents were highly sensitive to apamin block. Apamin sensitivity is determined by the presence of aspartate and asparagine on opposite sides of the pore (Ishii et al. 1997). Consistently, both of these residues were found in the cochlea channel. The bath solution contained niflumic acid to block Ca2+-activated Cl– channels and the external and internal (pipette) solution contained Cl– concentrations such that the Erev of Cl– was positive to 60 mV. Thus at –30- to 50-mV step voltages, all outward currents are K+ currents. The SK2 single-channel conductance was 18 ± 3 pS (n = 5). This is comparable to a single-channel conductance of 19 pS obtained in chick hair cells (Yuhas and Fuchs 1999). However, a smaller conductance of approximately 9 pS was measured in some of the patches (n = 3 patches). Previous studies showed that the rat brain SK2 channels expressed in oocytes yielded a single-channel conductance of approximately 11 pS, with a reversal potential at –6 mV (Kohler et al. 1996). Variations in the magnitude of the conductance of the SK2 channel may result from technical difficulties in obtaining data with sufficient signal-to-noise ratio and/or the possibility that the channel may exhibit different and stable conductance levels.
The activation of K+ conductances at the base of the OHCs by Ca2+ entry through AChR is very fast. The SK channels and the AChRs were inferred to be colocalized closely within approximately 10-nm domains to account for the fast kinetics of the onset of the K+ conductance (Oliver et al. 2000). In HEK-293 cells cotransfected with SK2 channels and α9/α10 AChRs, ACh induced inward currents followed by outward currents, that were blocked by apamin. However, the onset of the outward currents was not as fast as recorded in the OHCs, suggesting perhaps the requirement of other protein(s) that may serve as auxiliary subunits in close association with these channels at the base of the OHCs. The heterologous coexpression of SK2 and α9/α10 AChRs partly recapitulates the cochlea efferent feedback response. Not only are the findings important in understanding the synaptic regulation in the cochlea, but they suggest a requirement for other, not yet identified, molecular components for proper functional interaction between the AChRs and SK2 and establish a system in which they may be further investigated.
We thank Dr. Guy Richardson, University of Sussex, Brighton, UK, for donating the mouse cochlea cDNA library.
This work was supported by National Institute on Deafness and Other Communication Disorders Grants DC-03826 and DC-04215 to E. N. Yamoah; DC-04279 to K. W. Beisel; and DC-05578 and DC-06421 to A. E. Vázquez.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 by the American Physiological Society