Phylogenetic differences in calcium permeability of the auditory hair cell cholinergic nicotinic receptor

Abstract

The α9 and α10 cholinergic nicotinic receptor subunits assemble to form the receptor that mediates efferent inhibition of hair cell function within the auditory sensory organ, a mechanism thought to modulate the dynamic range of hearing. In contrast to all nicotinic receptors, which serve excitatory neurotransmission, the activation of α9α10 produces hyperpolarization of hair cells. An evolutionary analysis has shown that the α10 subunit exhibits signatures of positive selection only along the mammalian lineage, strongly suggesting the acquisition of a unique function. To establish whether mammalian α9α10 receptors have acquired distinct functional properties as a consequence of this evolutionary pressure, we compared the properties of rat and chicken recombinant and native α9α10 receptors. Our main finding in the present work is that, in contrast to the high (pCa²⁺/pMonovalents ∼10) Ca²⁺ permeability reported for rat α9α10 receptors, recombinant and native chicken α9α10 receptors have a much lower permeability (∼2) to this cation, comparable to that of neuronal α4β2 receptors. Moreover, we show that, in contrast to α10, α7 as well as α4 and β2 nicotinic subunits are under purifying selection in vertebrates, consistent with the conserved Ca²⁺ permeability reported across species. These results have important consequences for the activation of signaling cascades that lead to hyperpolarization of hair cells after α9α10 gating at the cholinergic–hair cell synapse. In addition, they suggest that high Ca²⁺ permeability of the α9α10 cholinergic nicotinic receptor might have evolved together with other features that have given the mammalian ear an expanded high-frequency sensitivity.

Nicotinic cholinergic receptors (nAChRs) are transmembrane allosteric proteins involved in the physiological responses to the neurotransmitter acetylcholine (ACh) (1). They are assembled from five identical (homopentamer) or different (heteropentamer) polypeptide chains arranged symmetrically around a central axis that lines a channel pore (1). The selective pressures for maintaining a wide diversity of nAChR subunits remain unknown (2, 3). The pentameric structure and function of muscle nAChR is well established; however, although much progress has been achieved, the possible combinatorial assemblies of neuronal subunits and their stoichiometry, distribution and function in the central nervous system remain subjects of ongoing investigation (4). Uniquely among these nonmuscle nAChRs, a clear function has been identified for the α9 and α10 nAChR subunits. These subunits assemble to form a pentameric receptor with a likely α9₂α10₃ stoichiometry, which serves cholinergic neurotransmission at the synapse between efferent fibers and hair cells of the auditory and vestibular end organs (5–8). Efferent function has been studied in cochlear hair cells, where it is thought to inhibit outer hair cell (OHC) electromotility and so reduce cochlear amplification (9). In contrast to all of the nicotinic receptors that serve excitatory neurotransmission, the activation of α9α10 nAChRs produces hyperpolarization of hair cells (see ref. 9 for a review). This is thought to be produced by entry of Ca²⁺ through the receptor and the subsequent activation of a small-conductance SK2 Ca²⁺-dependent potassium channel (10, 11). Thus, Ca²⁺ plays a major role in downstream signaling after activation of the α9α10 nAChR.

Analyzing the evolutionary history of α9 and α10 confers an additional degree of complexity to the diversity of nAChRs. Whereas α9 has been under purifying selection (i.e., selection against nonsynonymous substitutions at the DNA level) across all vertebrate species, α10 shows signatures of positive Darwinian selection only along the mammalian lineage (12). This observation suggests the possible acquisition of distinct functional properties for this receptor in mammals. To determine whether mammalian α9α10 nAChRs differ functionally from their nonmammalian vertebrate counterparts, we compared responses mediated by receptors of Rattus novergicus (rn, rat) with those of Gallus gallus (gg, chicken). We report that, in contrast to the high Ca²⁺ permeability reported for the recombinant and native rat and recombinant human α9α10 receptors (10, 13, 14), recombinant and native chicken α9α10 nAChRs have a much lower permeability to this cation, comparable to that of neuronal α4β2 receptors (15). These results have important physiological implications for the signaling cascades after α9α10 activation at the cholinergic–cochlear hair cell synapse and suggest that the high Ca²⁺ permeability of this nAChR might have accompanied the acquisition of high-frequency sensitivity in mammals.

Results

Responses to ACh.

To determine whether chicken α9 and α10 nAChR subunits assemble into functional channels, we injected in vitro transcribed cRNAs into Xenopus laevis oocytes and recorded responses to ACh under two-electrode voltage clamp. Injected oocytes responded to ACh in a concentration-dependent manner, with an apparent affinity similar to that of rat receptors (EC₅₀: rn, 12.6 ± 2.0 μM, n = 6; gg, 11.0 ± 3.3 μM, n = 6; P = 0.6831) (Fig. S1A). The desensitization kinetics of chicken and rat receptors were similar (Fig. S1B).

Calcium Permeability.

Rat and human recombinant α9α10 nAChRs (13, 14), as well as native α9α10 nAChRs of rat cochlear hair cells (10), have a high Ca²⁺ permeability [relative Ca²⁺ to monovalent permeability (pCa/pMono) close to or exceeding 10]. Moreover, the fractional Ca²⁺ current (P_f, the fractional Ca²⁺ current, estimated as Ca²⁺ fluorescence increase/total charge entering the cell) in rat α9α10- transfected GH4C1 cells (22%) is the largest value estimated to date for a ligand-gated ion channel (16). To assess the significance of Ca²⁺ flux through chicken recombinant α9α10 nAChRs, in a first set of experiments we analyzed the contribution of the ICl_Ca to ACh-evoked responses. ICl_Ca is an endogenous Cl⁻ current present in Xenopus oocytes that is activated by an increase in intracellular Ca²⁺ (17). Receptors with very high Ca²⁺ permeability, such as α7 (18) and α9α10 (7) nAChRs, have a more prominent contribution of the ICl_Ca to ACh-evoked responses compared with the modest contribution of receptors with lower Ca²⁺ permeability (19). Fig. 1 shows representative responses to 100 μM ACh at a V_hold of −70 mV before and after a 3-h incubation with the membrane permeant fast Ca²⁺ chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid-acetoxymethyl ester (BAPTA-AM) for rat α9α10 nAChRs (Fig. 1A) and α4β2 nAChRs (Fig. 1B), receptors with high (pCa/pMono, 9; ref, 13) and low (pCa/pNa, 1.6; ref. 15) Ca²⁺ permeability, respectively. Whereas ACh-evoked currents were markedly reduced after BAPTA-AM incubation in the case of rat α9α10 receptors (indicating a strong contribution of ICl_Ca and thus influx of Ca²⁺ through the receptor), the response in α4β2 receptors remained unaffected. A similar experiment performed for chicken α9α10 receptors produced results strikingly similar to those obtained with rat α4β2, but not rat α9α10, nAChRs (Fig. 1C). The percentage of remaining response was close to 100% for both chicken α9α10 (100.3% ± 14%; n = 7) and α4β2 (95.7% ± 3.4%; n = 5) receptors, and was 11.7% ± 2.7% (n = 10) for rat α9α10 receptors.

Fig. 1.

Chicken α9α10 nAChRs do not activate oocyte ICl_Ca. Representative traces of responses evoked by a 20-s application of 100 μM ACh in oocytes expressing rat α9α10 (A), rat α4β2 (B), or chicken α9α10 nAChR (C), before (Left) and after (Right) a 3-h incubation with the fast calcium chelator BAPTA-AM (V_hold = −70 mV).

To further estimate the relative Ca²⁺ to monovalent permeability of chicken α9α10 receptors, we evaluated the shift in the reversal potential E_rev as a function of increases in extracellular Ca²⁺ concentration as for rat (13) and human (14) α9α10 nAChRs. Current-voltage (I-V) curves were elicited by a 2-s voltage ramp protocol (−120 to +50 mV) at the plateau of the response to ACh from a V_hold of −70 mV in the presence of different Ca²⁺ concentrations. Fig. 2A shows representative I-V curves for only three Ca²⁺ concentrations for the sake of clarity; insets show a magnification near the E_rev. Fig. 2B shows averaged E_rev as a function of five external Ca²⁺ concentrations. Rat α9α10 receptors demonstrated a significant shift in E_rev to more depolarized potentials in response to each increase in extracellular Ca²⁺. A 10-fold increase in Ca²⁺ concentration from 0.2 to 2 mM resulted in a shift in E_rev of 40 ± 8.6 mV (n = 5) (Table S1). Moreover, the fit of the data to a straight line had a slope of 39.3 ± 0.4 (n = 5), which was significantly different from 0 (P = 0.0001), thus illustrating the high degree of dependency of E_rev on the extracellular Ca²⁺ concentration. In contrast, the shifts in reversal potential were negligible in the case of rat α4β2 and chicken α9α10 nAChRs (Table S1).

Fig. 2.

Chicken recombinant α9α10 nAChR has lower relative calcium permeability. (A) Representative I-V curves obtained on application of voltage ramps (−120 to +50 mV, 2 s) at the plateau of the response to 10 μM (α9α10) or 100 μM (α4β2) ACh in oocytes superfused with N-methyl glucamine–based solutions containing different Ca²⁺ concentrations for oocytes expressing rat α9α10 (Upper Left), rat α4β2 (Upper Right), or chicken α9α10 nAChRs (Lower Left). (Insets) Magnification near the E_rev. (B) Plots of E_rev values as a function of Ca²⁺ concentration for rat α9α10, rat α4β2, and chicken α9α10 nAChRs. Values are mean ± SEM of n = 5–8 per point.

The fit of the foregoing data to the extended Goldman–Hodgkin–Katz (GHK) equation resulted in a pCa/pMono of 10.3 ± 0.9 (n = 5) for rat α9α10 receptors, a result similar to that reported previously (13). In the case of chicken α4β2 and α9α10 receptors, the data from most oocytes did not fit the GHK equation, because of the insensitivity of E_rev to variations in extracellular Ca²⁺ concentration. Nonetheless, a fit was possible in three of eight oocytes expressing chicken α9α10 and in one of five oocytes expressing α4β2, for a pCa/pMono value close to 2. In summary, Ca²⁺ permeability of the chicken α9α10 nAChR is similar to that of the heteromeric neuronal α4β2 receptor and significantly lower than that of the rat α9α10 receptor.

Calcium Permeability of nAChRs in Chicken Hair Cells.

Earlier studies suggested that Ca²⁺ can permeate the chicken hair cell nAChR (20), but did not define permeability ratios. Tight-seal intracellular recordings were made from “short” hair cells in a region corresponding to ∼500–1,000 Hz along the tonotopically organized basilar papilla from late-stage chicken embryos (days 17–20 in ovo, hatching at day 21). These cells responded to ACh with a combination of ligand-gated cation current followed by Ca²⁺-dependent potassium current through apamin-sensitive (“SK-like”) potassium channels, as demonstrated previously (21).

The isolated ACh current through the α9α10 nAChR was measured as described in SI Materials and Methods. Under these conditions, application of ACh (1 mM) from a “puffer” pipette evoked an inward current at negative membrane potentials (Fig. S2A), which reversed in sign near 0 mV (Fig. S2 B and C for step and ramp protocols, respectively). The overall shape of the I-V curves obtained in chicken short hair cells was similar to that reported for rat inner hair cells and OHCs (10, 22).

To estimate the relative pCa/pMono of native chicken α9α10 receptors, we measured the reversal potential of ACh-evoked membrane current in five external Ca²⁺ concentrations ranging from 0.1 to 10 mM. Fig. 3A shows representative I-V curves at three Ca²⁺ concentrations, and Fig. 3B shows the E_rev as a function of external Ca²⁺. Because each individual I-V curve represents a different chicken short hair cell, the absolute magnitude of responses at different Ca²⁺ concentrations cannot be compared. As the external calcium concentration was increased, the reversal potential shifted to more positive membrane potentials. The variation of E_rev with log external Ca²⁺ concentration was fit (minimum least squared difference) with a straight line of slope 5 mV/log₁₀ [Ca²⁺]_o (n = 5–8 cells at each concentration) (Fig. 3B). From the extended GHK equation, the Ca²⁺ permeability relative to that of the dominant monovalent cation was calculated as 2.6. For the purpose of comparison, E_rev obtained from rat inner hair cells under similar recording conditions (10) are plotted on the same graph. The reversal potential of ACh-evoked currents of rat inner hair cells had a steeper slope (17 mV/log₁₀ [Ca²⁺]_o) (Fig. 3B), and a calculated pCa/pMono of 8 (10).

Fig. 3.

Chicken short hair cell nAChR has lower relative calcium permeability. (A) Representative I-V curves obtained by voltage ramp protocols used to define the reversal potentials in external calcium concentrations of 0.1, 1, and 3 mM. (Inset) Magnification near the E_rev. Each I-V curve represents a different hair cell, and thus absolute amplitudes cannot be compared. (B) Average (± SEM) E_rev as a function of external calcium concentration in chicken short hair cells (gray squares, 5–8 cells per point) compared with previously published data for rat inner hair cells (black triangles). (From ref. 9.)

These differences in calcium permeability might influence the subsequent activation of calcium-dependent potassium channels that hyperpolarize and thus inhibit auditory hair cells (10, 21). This question is further complicated by the specialized ultrastructure of the efferent synapse. A near-membrane postsynaptic cistern (23, 24) delimits a restricted diffusion space for ACh-evoked calcium signals and may operate as a source of additional calcium, as a fixed calcium buffer, or both. We examined this possibility using a combination of membrane voltage commands and a “calcium store inhibitor” to alter the ACh-evoked, calcium-dependent potassium current of inner hair cells from the immature rat cochlea and short hair cells of the chicken basilar papilla. ACh (100 μM, 300 ms) was “puff”-applied to rat and chicken hair cells voltage-clamped to −40 mV using ionic conditions to activate calcium-dependent potassium currents in the hair cells (5 mM internal EGTA, 1.3 mM external calcium) (Fig. 4). Near the peak of the evoked outward calcium-dependent potassium current, the membrane potential was then clamped to +40 mV to sharply reduce driving force for inward calcium flux (Fig. 4). As a result, ACh-activated potassium current decayed, reflecting its dependence on calcium influx, but did so at different rates in rat and chicken hair cells (Fig. 4 A and B). The average ± SEM decay time constant was 37 ± 5 ms in 12 rat inner hair cells and threefold longer, 124 ± 14 ms, in 12 chicken hair cells. Although various factors might influence the decay of ACh-activated potassium current, we uncovered a contribution from synaptic calcium stores by exposing hair cells to the sarcoplasmic/endoplasmic calcium ATPase (SERCA) inhibitor tert-benzo-hydroquinone (t-BHQ; 25 μM). t-BHQ had no significant effect on outward current decay in rat inner hair cells (decay time constant, 32 ± 3 ms; P = 0.382, paired t test), but reduced that of chicken hair cells to 74 ± 7 ms (P = 0.005, paired t test) (Fig. 4 C and D).

Fig. 4.

ACh-evoked potassium current in chicken and rat hair cells. (A) A puff of ACh (100 μM for 300 ms) caused an outward current at −40 mV lasting approximately 2 s in rat inner hair cells (gray). When the membrane potential was stepped to +40 mV, the initially large outward current rapidly decayed (black) to a level dominated by a voltage-dependent outward current. (B) ACh-evoked membrane current in a chicken hair cell lasting approximately 2 s at −40 mV (gray). With a step to +40 mV, the initially large outward current decayed, albeit relatively slowly, to a level reflecting the small voltage-dependent outward current (black). (C) ACh-evoked “difference current” after subtracting membrane current in the absence of ACh in a rat inner hair cell under control conditions (black) and after exposure to the SERCA inhibitor t-BHQ (gray). (D) Same as in C, but for a chicken short hair cell.

Molecular Evolution of α7 and α4β2 nAChRs.

The phylogenetic difference in Ca²⁺ permeability appears as a peculiar feature of the α9α10 nAChR and has not been reported for other nAChRs. For example, Ca²⁺ permeability is at the high end (pCa/pNa, ∼10–20) for both mammalian (18, 25, 26) and chicken (27) α7 nAChRs and 10 times lower both for mouse and chicken α4β2 receptors (pCa/pNa, 1.7) (28, 29). Thus, conservation of the molecular bases underlying Ca²⁺ permeability in α7 and α4β2 nAChRs during the evolution of birds and mammals would be expected. To study the evolutionary histories of vertebrate α7, α4, and β2 subunits, we implemented the branch-site test of positive selection, which detects the presence of sites under positive selection within a given gene in specific lineages of a phylogeny (30, 31). This test is based in the comparison of two evolutionary models. The branch-site model (H_A) allows sites in the foreground lineage to evolve under positive selection, whereas in the null model (H₀), sites are allowed to evolve only under either purifying selection (ω <1) or neutral evolution (ω = 1). We then used a likelihood ratio test (LRT) to assess whether the branch-site model fit the data better than the null hypothesis. Our results (Table 1) indicate that for the genes encoding the α7 (CHRNA7), α4 (CHRNA4), and β2 (CHRNB2) subunits, a model that allows sites to evolve positively in the lineage leading to mammals does not fit the data better than the null hypothesis. For the sake of comparison, we extended our previous analysis of the evolution of CHRNA10 to include additional species. As shown in Table 1, for CHRNA10 with mammals as the foreground branch, an LRT value of 9.39 was obtained (P < 0.0021), indicating positive selection in this lineage. Conducting the same tests but considering other vertebrate lineages (placental mammals or birds/reptiles) as evolving positively revealed no signatures of positive selection for any of the four genes. In addition, a phylogenetic tree including known vertebrate nAChR subunits from representative species shows that whereas α9, α7, α4, β2, and other nAChR subunits follow the main vertebrate taxa grouping, mammalian α10 subunits form a unique clade separate from other nonmammalian vertebrate α10 subunits that group with α9 subunits (Fig. S3). Thus, as shown in a collapsed phylogenetic tree (Fig. 5), all nAChRs have their individual unique branch, with the exception of α10, which has two separate branches, one for mammals and one for nonmammalian vertebrates. Taken together, these results indicate that, in contrast to the mammalian α10 nAChR subunit, α7, α4, and β2 were not shaped by positive selection and were under purifying selection in all vertebrate species.

Table 1.

Likelihood value estimates for the branch-site and null hypothesis models

	CHRNA4	CHRNB2	CHRNA7	CHRNA10
Foreground lineage: mammals
LnL HA	−19,249.50	−12,662.16	−16,620.05	−16,185.80
LnL H0	−19,250.54	−12,662.16	−16,621.27	−16,190.49
LRT	2.07	0	2.43	9.39
P (χ21)	0.150	1	0.119	0.002
Foreground lineage: placental mammals
LnL HA	−19,251.48	−12,662.20	−16,615.01	−16,213.78
LnL H0	−19,251.48	−12,662.20	−16,616.83	−16,214.76
LRT	0	0	3.64	1.96
P (χ21)	1	1	0.057	0.162
Foreground lineage: birds/reptiles
LnL HA	−19,251.48	−12,660.13	−16,621.28	−16,218.23
LnL H0	−19,251.48	−12,661.07	−16,621.28	−16,218.22
LRT	0	1.90	0	0
P (χ21)	1	0.169	1	1

CHRNA4, CHRNB2, CHRNA7, and CHRNA10 genes from 23 vertebrate species described in Table S2 were analyzed using the branch-site test for positive evolution. The analysis was conducted separately three times for each gene, using mammals, placental mammals and birds/reptiles as the foreground lineage. LRT = 2 × (LnLH_A − LnLH₀). H_A, branch-site model; H₀, null hypothesis; LnL, log-likelihood.

Fig. 5.

Phylogenetic tree of nAChR subunits. The tree represents the collapsed phylogenetic relationships among vertebrate nAChR subunits. The branches corresponding to the same subunits of different species were condensed up to the node at which one subunit separates from its closest neighbor. The complete tree is shown in Fig. S3. Numbers in branches indicate the bootstrap value obtained after 1,000 replicates. Note that the α10 subunit is the only one to show two separate branches; nonmammalian α10 subunits are closely related to α9 subunits, whereas mammalian α10 subunits form a separate clade.

Discussion

The ability of Ca²⁺ to perform its various local tasks within the cell depends on steep gradients in concentration encountered in the vicinity of open Ca²⁺-permeable channels (32). In contrast to voltage-gated Ca²⁺ channels, ligand-gated channels provide Ca²⁺ influx at voltages close to the resting membrane potential, as is the case for nAChRs. Based on their relative Ca²⁺ permeability, nAChRs have been subdivided into three categories that are conserved among vertebrate species. The muscle type is the least permeable of all nAChR family members (pCa/pNa, ∼0.2–0.4) (33–35). Heteromeric neuronal receptors have higher Ca²⁺ permeability (pCa/pNa, ∼1.5) (19, 35–37), and the homomeric α7 and α8 receptors have the highest relative Ca²⁺ permeability (pCa/pNa, ∼10–20) (18, 27). The main finding of the present work is that, in contrast to other nAChRs, the Ca²⁺ permeability of α9α10 nAChRs is not uniform across species. It is unexpectedly low and similar to that of heteromeric neuronal nAChRs in chicken α9α10 nAChRs (present results) and high and similar to that of homomeric α7- and α8-containing receptors in rat α9α10 nAChRs (10, 13, 14). Given that the α10 nAChR subunit has been under strong positive selection in mammals (ref. 12 and present results), we suggest that increased Ca²⁺ permeability might be the functional novelty resulting from that selective pressure. Although some degree of variability in Ca²⁺ permeability exists across species for other nAChRs, the large differences described here for α9α10 nAChRs have not been reported previously (38). Thus, in contrast to the findings for α10, the lack of evidence of positive selection of α7, α4, and β2 nAChR subunits across vertebrate species is in agreement with the observation that Ca²⁺ permeability ratios are conserved as well.

Several lines of evidence pinpoint the pore lining transmembrane region 2 (TM2) of nAChRs as responsible for providing the molecular determinants underlying Ca²⁺ permeability. Thus, mutations at positions −1′, 16′, and 17′ of TM2 (Fig. S4) abolish Ca²⁺ permeability of several nAChR receptors (15, 27). Alignment of the α10 TM2 region of several vertebrate species indicates that all of these residues are highly conserved in mammalian and nonmammalian vertebrates (Fig. S4), precluding their involvement in the observed permeability differences.

Efferent control is a common feature of all vertebrate hair cells (39), and ACh is thought to produce hyperpolarization of hair cells in both mammals and birds (as well as other vertebrates) through the activation of SK Ca²⁺-dependent K⁺ channels (see ref. 9 for a review). Nearby cytoplasmic stores may contribute additional calcium in hair cells (40). Although postsynaptic cisterns are omnipresent features of efferent synapses in auditory hair cells of birds and mammals (23, 24), the relative contributions of calcium influx and store release might vary between chicken and rat hair cells. Compared with rat hair cells, in chicken hair cells, cholinergic activation of calcium-dependent SK channels had a prolonged time course at a voltage at which calcium influx was minimized. This activation was sensitive to SERCA inhibition in chicken hair cells but not in rat hair cells, suggesting that in the former, the postsynaptic calcium signal owes its duration to release from stores. It also suggests that release of calcium from synaptic stores, at least under these experimental conditions, plays a lesser role in rat hair cells, consistent with greater calcium influx through these cells’ nAChRs. Of interest, the decay time constant of SK “tail current” in rat hair cells is similar to that reported previously for synaptic currents of inner (41, 42) and outer (11) cochlear hair cells. However, it should be noted that calcium stores might become more important during prolonged cholinergic inhibition of rat cochlear hair cells (40, 43).

Given that the lineages leading to modern birds and mammals separated at least 300 Mya (39, 40), the foregoing results might indicate that independent solutions to functional demands evolved in parallel for the avian basilar papilla and the mammalian cochlea. However, the increasingly calcium-permeable nAChR of rat cochlear hair cells might also indicate that stronger efferent inhibition is required for mammalian cochlear hair cells. It has been observed (44–46) and recently elaborated (47) that OHCs exhibit increasing membrane ionic conductance moving from apical, low-frequency positions to higher-frequency basal locations in the mammalian cochlea. This increased membrane conductance along with decreased capacitance (membrane area) confers shorter membrane time constants to high-frequency basal OHCs. Consequently, the prestin-dependent electromotility that provides cochlear amplification in mammals can extend to acoustic frequencies above 10 kHz (47). Because efferent synaptic inhibition acts as a parallel shunt, the inhibitory conductance must increase with “resting” membrane conductance to remain effective. Thus, larger synaptic contacts in the high-frequency basal region might be expected, as has been reported previously (46). Moreover, the fact that RT-PCR and in situ hybridization measurement of α9 expression shows both longitudinal and radial gradients, with the greatest expression over OHCs in basal regions (48), might indicate a higher density of nAChRs in the high-frequency region of the cochlea. Further inhibitory strength could be gained by increasing the flux of calcium that activates the hyperpolarizing K⁺ conductance. The recent report indicating that cholinergic inhibition of basal (high-frequency) rat OHCs specifically uses large-conductance BK channels in addition to the small-conductance SK channels that serve solely in apical OHCs lends additional credence to this hypothesis (46). Thus, it can be conjectured that increased Ca²⁺ permeability helps activate low–Ca²⁺-affinity, high-conductance BK channels (49) in mammalian basal OHCs, whereas Ca²⁺ influx provided by the nonmammalian nAChR suffices to activate the high–Ca²⁺-affinity, low-conductance SK channels (49).

In summary, the present work reports a substantial interspecies difference in Ca²⁺ permeability for a nAChR. In addition, it suggests that efferent inhibition of mammalian cochlear hair cells might especially benefit from a cholinergic receptor with enhanced calcium permeability.

Materials and Methods

All experimental protocols were carried out in accordance with National Institutes of Health guidelines for the care and use of laboratory animals, as well as guidelines of the Institutional Animal Care and Use Committees of The Johns Hopkins University and Instituto de Investigaciones en Ingeniería Genética y Biología Molecular.

Expression of Recombinant Receptors in X. laevis Oocytes.

The plasmids used for heterologous expression in X. laevis oocytes are described in SI Materials and Methods. Electrophysiological recordings from Xenopus oocytes were obtained as described previously (6, 50) and are described in more detail in SI Materials and Methods.

Chicken Basilar Papilla.

Patch-clamp recordings from short hair cells of the auditory organ (basilar papilla) were obtained as described in SI Materials and Methods.

Calcium Permeability.

The relative pCa/pMono values were evaluated as done previously for recombinant and native mammalian α9α10 receptors (10, 13, 14, 50), by analyzing the shift in the reversal potential (E_rev) as a function of the increase in the extracellular Ca²⁺ concentration and fitting the data to the GHK constant field voltage equation, assuming no anion permeability and extended to include divalent cations (51, 52). A detailed description of the protocol is provided in SI Materials and Methods.

Molecular Evolution Analysis.

All sequences were downloaded from GenBank and Ensembl databases. Alignments of sequences used for the evolutionary analysis are shown in Fig. S5, and accession numbers are listed in Table S2. To test for the presence of positive selection, the branch-site test of positive selection developed by Yang and coworkers (30, 31) was applied using the codeml program implemented in PAML, as discussed in SI Materials and Methods. Phylogenetic trees were built using MEGA4, as described in SI Materials and Methods.