Adult American bullfrogs (Rana catesbeiana) were obtained from a local vendor. Bullfrogs in the weight range of 200–300 g were used for the study. Before being pithed and decapitated, the bullfrogs had been anesthetized in an ice bath for 20 min. Then, the tissues of AP and BP were carefully dissected out (refer to Figure 1) in an oxygenated extracellular solution containing (in mM): 95 NaCl, 2 KCl, 2 CaCl₂, 1 MgCl₂, 3 D-glucose, 1 creatine, 1 Na-pyruvate, and 10 HEPES. For outward K⁺ current recording, an additional 2 mM CdCl₂ was included. The osmolality was adjusted to 230 mOsm with NaCl and pH 7.40 with NaOH. Then, they were transferred and settled into a recording chamber containing about 1 ml of extracellular solution for the following experiment. Chemicals and salts were purchased from Sigma.

The recording chamber, along with the tissue, was placed under an upright microscope (Olympus) equipped with a 60 × water-immersion objective. Whole-cell patch-clamp recordings were performed at room temperature (∼23°C) within 2 h of dissection, using an EPC10/2 patch-clamp amplifier (HEKA Electronics, Lambrecht Pfalz, Germany) and Patchmaster (HEKA Electronics) was used to generate a voltage- or current-clamp commands to drive the amplifier. The hair cells were held at −90 mV. The liquid junction potential of −10 mV was corrected offline and data were corrected by subtracting 10 mV from all potentials.

Recording micropipettes were made from borosilicate glass tubes (BF150-86-10, Sutter Instruments, Novato, CA, USA) using a two-stage vertical pipette puller (Narishige, Tokyo, Japan) to tip diameters of ∼1 μm. The electrodes were coated with dental wax to reduce stray capacitance and enhance C-fast compensation. A Cs + -based internal solution was used to block K⁺ current and isolate Ca²⁺ current. Pipettes were filled with an internal solution containing (in mM): 80 Cs-gluconate, 20 CsCl, 10 TEA-Cl, 2 EGTA, 3 Mg-ATP, 0.5 Na-GTP, and 10 HEPES (230 mOsm, pH 7.30). The internal solution for outward K⁺ current recording contained the following: 80 K D-gluconate, 30 KCl, 2 EGTA, 3 Mg-ATP, 0.5 Na-GTP, and 10 HEPES. Pipettes were pulled to resistances of 4 to 7 MΩ. Series resistances (Rs) during recordings typically ranged from 10 to 30 MOhm, continuously compensated using the amplifier’s compensation circuitry. For K⁺ current recording, voltage steps were delivered from the holding potential of −90 mV to potentials ranging from −125 to + 45 mV in 10 mV increments, in order to obtain the current-voltage relationship (I-V curve). To verify the pharmacological identity of different potassium currents, we applied the more positive holding potential (−65 mV), with the depolarization voltage steps ranging from −125 to + 5 mV in 10 mV increments.

For Ca²⁺ current recording, a voltage ramp of 300 ms from −90 to + 70 mV was applied to hair cells to evoke Ca²⁺ currents, and the peak of this Ca²⁺ current (I_Ca) was determined. We then fitted the current-voltage relationship (I-V curve) with a Boltzmann equation as below to obtain the half activation potential (V_half) and the slope factor (k), which reflects the steepness of voltage dependence in Ca²⁺ current activation. Where V is the command membrane potential, V_rev is the reversal potential and G_max is the maximum chord conductance.

For membrane capacitance (C_m) measurement, the “Sine + DC” technique (Lindau and Neher, 1988) in Patchmaster was applied under a voltage clamp, using a software lock-in amplifier (Lindau and Neher, 1988; Gillis, 2000). Once forming a tight seal, each hair cell was held at −90 mV and driven with sine waves of 1 kHz (25 mV in amplitude) to determine the C_m (Li et al., 2009). The real-time changes in averaged C_m [ΔC_m = C_{m (response)}–C_{m (baseline)}], evoked by cell membrane depolarization, were used to generally assess and quantify the total synaptic vesicle exocytosis from hair cells.

For the exocytosis dynamic assay, ΔC_m was measured at variant stimulation lengths from 2, 5, 10, 20, 50, 100, 200, 500 to 1,000 ms and plotted against stimulation time (t) recorded from individual hair cell, and the data points were fitted to a combination formula, consisting of a single exponential function for the release of the readily releasable pool (RRP) of synaptic vesicles (C_m,RRP, τ_RRP) and a linear function for the sustained release of synaptic vesicles (R_sustained, SRR):

Then, the numbers of synaptic vesicles were estimated with the capacitance values, using a conversion factor of 37 aF/vesicle (Lenzi et al., 1999; Johnson et al., 2009).

For synaptic vesicle replenishment assessment, a paired-pulse of 100 ms was executed with various interpulse intervals (100, 200, 500, 1,000, and 2,000 ms). The average ratio of ΔC_m (ΔC_m2/ΔC_m1) was calculated to evaluate the C_m recovery.

Data analysis was performed using Igor Pro (WaveMetrics, USA) and Prism Software (GraphPad, USA) with built-in macros and functions. Statistical comparisons of means were assessed with the two-tailed paired or unpaired Student’s t-test for single measurements. For multiple measurements, two-way analysis of variance (ANOVA), followed by the Bonferroni post-hoc test for multiple comparisons, was used to compare the data sets between AP and BP groups. Data were expressed as mean ± SEM with p-values of < 0.05 which is considered statistically significant. For the main results, Cohen’s d was used to accompany the reporting of t-test and ANOVA results, which is an appropriate effect size for the comparison between two means. Cohen’s d can be calculated as the difference between the means divided by the pooled SD, and some minimal guidelines are that d = 0.2 indicates a small effect, d = 0.50 indicates a medium effect, and d = 0.80 indicates a large effect.

Membrane depolarization from −125 to +45 mV is capable of activating a combination of I_Ca, I_K(Ca), I_K, I_A, and I_K(IR) (Hudspeth and Lewis, 1988) in standard saline (Figure 2A). In our study, the external addition of 2 mM Cd²⁺ can completely block the inward I_Ca and therefore largely and rapidly eliminate the calcium-dependent outward potassium current I_K(Ca) (Smotherman and Narins, 1999a). TEA can block the voltage-dependent slowly activating potassium current I_K, and 4-AP serves as an A-type and delayed rectifier-type channel blocker (Nenov et al., 1997), has been reported to specifically eliminate I_A in hair cells (Lewis and Hudspeth, 1983), and probably affects K⁺ channels other than I_K(Ca) (Ricci et al., 2000).

First of all, we recorded the control potassium current without any blockers and noticed that the amplitude of potassium in medial AP hair cells was significantly larger than that in BP [F(1,22) = 5.16, p = 0.03, two-way ANOVA followed by Bonferroni post-hoc test; Figure 2B]. Besides, the steady-state (averaged over the last 100 ms of pulse) current amplitudes of a depolarizing step to + 45 mV differed significantly between AP and BP hair cells, with the means of 2.264 ± 0.350 pA and 0.864 ± 0.227 (p = 0.003, Cohen’s d = 1.331; Figure 2C, left panel), respectively. Considering the variance in cell size, we measured the capacitance membrane of hair cells from both AP and BP and found that AP hair cells have a significantly larger C-slow value than BP hair cells (17.2 ± 1.59 pF vs. 7.55 ± 0.995 pF, n = 13 and 12 for the two groups; unpaired Student’s t-test, p < 0.001). Then, we calculated the density of potassium current and found no significant difference in the two cell types (AP: 0.140 ± 0.027 nA/pF, BP: 0.139 ± 0.035 nA/pF, p = 0.97, Cohen’s d = 0.015; Figure 2C, right panel).

To further identify the different components of potassium current, we then applied various blockers. Given I_K(Ca) was found to be sensitive to the external calcium concentration, we first added CdCl₂ to the recording solution and observed that 2 mM CdCl₂ routinely reduced the amplitude of the net outward current in AP, with the amplitude decreased averagely from 1.66 ± 0.416 nA to 0.192 ± 0.041 nA (p = 0.02, paired t-test; Cohen’s d = 3.130), approximately 88.43% of reduction. The sensitivity to external cadmium suggests that the slowly non-inactivating outward current is a calcium-dependent potassium current, I_K(Ca), and it accounts for the majority of the outward current in medial AP hair cells. In BP hair cells, the external CdCl₂ reduced the average amplitude from 0.736 ± 0.191 nA to 0.145 ± 0.042 nA (p = 0.03, paired t-test; Cohen’s d = 4.469), approximately 80.32% of reduction of amplitude, which is comparable to medial AP hair cells.

In the next place, owing to the outward current appearing in the presence of cadmium which is I_K, we subsequently applied TEA to verify its pharmacological effect on it. We found that 10 mM TEA significantly reduced the non-inactivating component in AP hair cells, with the steady-state amplitude decreased from 0.215 ± 0.056 nA to 0.147 ± 0.039 nA (p = 0.04, paired t-test; Cohen’s d = 0.705) on average, with approximately 31.63% of reduction. In contrast, TEA had no effect on the amplitude of the non-inactivating outward current in BP hair cells, with the average amplitudes before and after adding TEA, which was 0.31 ± 0.106 nA and 0.204 ± 0.053 nA, respectively (p = 0.27, paired t-test).

Finally, the effect of 1 mM 4-AP was assessed on the fast inactivating component of the outward current (I_A). We noticed that of all medial AP hair cells in the study (n = 23), the majority (n = 20) showed the non-inactivating potassium currents (Figure 2D), and only three cells showed the typical inactivating potassium currents (Figure 2E). Whereas in all tested BP hair cells (n = 13), there was a rapidly inactivating outward current (Figures 2F, G), which could be blocked by 4-AP, confirming the universal presence of A-type potassium current (I_A).

Consistent with the previous findings (Smotherman and Narins, 1999a,b), our results revealed that in medial hair cells of AP, the outward current is predominantly I_K(Ca), which is rapidly activating and calcium-dependent and the majority also have a slowly activating, outwardly delayed rectifying and voltage-dependent potassium current (I_K). Additionally, the earlier reports on frog crista ampulla that different types of potassium channels exist in the hair cells located in different regions and cells from the central region showed little or no I_A (Masetto et al., 1994; Russo et al., 1995), which is similar to our results about medial AP hair cells. While in the majority of BP hair cells tested, a rapidly activating and inactivating potassium A type current (I_A) was identified and I_K(Ca) is similarly the dominant component in the outward potassium current.

Depolarization of hair cells causes a Ca²⁺ entry, triggering transient C_m increments and elevations of global Ca²⁺ influx, which is controlled by the activation of the voltage-dependent Ca²⁺ current. We applied ramp stimulation and conducted whole-cell patch-clamp recordings (Figure 3A). Peak I_Ca is the maximal amplitude with all the calcium channels open, and it was significantly larger in medial AP hair cells (376 ± 23 pA, n = 30) than that of BP (95.4 ± 5 pA, n = 30; p < 0.001, unpaired Student’s t-test; Cohen’s d = 1.664, Figure 3B). Similarly, AP hair cells are almost three times larger than BP in size (C-slow value: 15.4 ± 0.718 pF vs. 5.97 ± 0.394 pF, n = 30; unpaired Student’s t-test, p < 0.001; Cohen’s d = 1.654, Figure 3E). We whereafter individually divided a peak I_Ca by the C-slow value to estimate the density of the current, which is correlated with the Ca²⁺ channel density in the single hair cell membrane. Not surprisingly, we found that the average density of I_Ca in each BP hair cell (17.1 ± 1.3 pA/pF, n = 30) was notably lower than that of AP (25.3 ± 1.72 pA/pF, n = 30) (unpaired Student’s t-test, p < 0.001; Figure 3F).

To characterize the functional properties of hair cells from the two auditory organs more comprehensively, V_half and k were figured out to depict the steepness of voltage dependence in Ca²⁺ channel activation. Concretely speaking, V_half describes the membrane potential at which the conductance is half activated while the slope factor (k) reveals the voltage sensitivity of activation. Noting I_Ca in AP hair cells has a more negative V_half (−40.5 ± 0.901 vs. −37.3 ± 0.579 mV, n = 30; unpaired Student’s t-test, p = 0.004; Cohen’s d = 0.772, Figure 3C) and a steeper activation slope (5.34 ± 0.172 vs. 7.51 ± 0.204 mV, n = 30; unpaired Student’s t-test, p < 0.001; Cohen’s d = 1.450, Figure 3D), suggesting that Ca²⁺ influx in AP hair cells is more sensitive to depolarization.

During exocytosis, synaptic vesicles fuse with the presynaptic membrane, increasing the whole-cell capacitance (Moser and Beutner, 2000). We, therefore, studied the exocytosis of AP and BP hair cells by monitoring the changes in their whole-cell capacitance (ΔC_m) before and after voltage steps to elicit Ca²⁺ influx and exocytosis. A representative diagram of I_Ca and the corresponding ΔC_m in AP and BP hair cells of duration 50 ms (left) and 200 ms (right) is shown in Figure 4A, respectively. The induced ΔC_m was found to be significantly larger in AP than in BP hair cells [F(1,30) = 105, p < 0.001; Table 1 and Figure 4B]. Then, the pooled data for Ca²⁺ charge for all the stimulation durations were assessed, and we found that Q_Ca in BP hair cells was remarkably smaller [F(1,25) = 31.9, p < 0.001, two-way ANOVA followed by Bonferroni post-hoc test; Table 1 and Figure 4C], consistent with the findings shown in Figure 3B. The Ca²⁺ efficiency of exocytosis (qualitatively judged to be ΔC_m/Q_Ca) was found no significant difference for any pairwise comparison of means in response to all the stimulations from 2 to 1,000 ms (Table 1 and Figure 4D), indicative of a similar efficiency of Ca²⁺ influx in triggering exocytosis from AP and BP hair cells.

Ribbon synapses in hair cells can release the synaptic vesicles rapidly and continuously. When stimulation is turned on, hair cells release a small pool of synaptic vesicles rapidly, representing the exocytosis of the readily releasable pool (RRP) of synaptic vesicles docked at the active zones (Moser and Beutner, 2000; Beutner and Moser, 2001). Subsequently, hair cells are capable of releasing the synaptic vesicles continuously for as long as the stimulation is on, owning to the fast replenishment of synaptic vesicles from a refilling pool that is located further from the Ca²⁺ channels (Von Gersdorff et al., 1996; Von Gersdorff and Matthews, 1999).

To examine the dynamics of exocytosis, we varied the length of voltage steps from 2 ms to 1 s and used curve fitting to extract RRP and sustained release rate (SRR, refer to the R_sustained in section “Materials and methods”; Figure 5A). Assuming a single vesicle capacitance of 37 aF (Lenzi et al., 1999), we found that the RRP consisted of 705 ± 92 (n = 12) and 156 ± 18 (n = 14) synaptic vesicles for AP and BP hair cells (Figure 5C), respectively. The time constants (τ) for the depletion of this fast pool were 17.7 ± 4.71 and 14.2 ± 4.92 ms for the responses of AP and BP hair cells (Figure 5B), respectively, which showed no significant difference (p = 0.61, unpaired Student’s t-test). Subsequently, an estimation of the sustained release rate (SRR) was obtained by fitting the linear range of the data. We found that the average SRR was 5,615 ± 800 SVs/s and 1,063 ± 200 SVs/s in AP and BP hair cells (Figure 5E), respectively, from which it can be learned the release rate of the latter that is about five-folds greater than the former. No significant difference was observed for τ to release RRP while the other two parameters (RRP and SRR) were both significantly reduced in BP hair cells, suggesting that its function of exocytosis was comparatively weak. We then divided the vesicles in the total RRP by Ca²⁺ current, and the results represented the number of RRP vesicles triggered per unit Ca²⁺ influx, showing no significant difference in these two auditory hair cells (1.79 ± 0.431 SVs/pA vs. 2.18 ± 0.517 SVs/pA, n = 12 and 14 for AP and BP hair cells; p = 0.57; Figure 5D). However, it existed a significant difference in SRR/I_Ca between AP (18.5 ± 3.27 SVs/s⋅pA) and BP (11 ± 1.35 SVs/s⋅pA) hair cells (unpaired Student’s t-test, p = 0.04; Figure 5F).

To study the kinetics of vesicle replenishment after voltage-gated Ca²⁺ entry, we measured the recovery of the RRP from the depletion by using paired stimuli (two times 100 ms depolarizations to −10 mV) separated by various intervals (from 100 to 2,000 ms) (Figure 6B). Relatively short stimuli reveal that the RRP and longer steps induce sustained release (Parsons et al., 1994; Schnee et al., 2005). We depolarized hair cells for 100 ms, designing to ensure depleting RRP vesicles. Noting that it showed a larger first than second ΔC_m, which might be due to Ca²⁺ current inactivation, vesicle depletion, or postsynaptic glutamate receptor desensitization.

The ratio of the capacitance increases (ΔC_m2/ΔC_m1) in response to the paired step depolarizations was used to assess the synaptic vesicle replenishment. As shown in Figure 6B, the synaptic vesicle replenishment in BP was significantly slower with an interval as short as 100 and 200 ms (100 ms: AP: 0.712 ± 0.039, BP: 0.527 ± 0.037; 200 ms: AP: 0.712 ± 0.036, BP:0.544 ± 0.037; n = 11 and 10; two-way ANOVA followed by Bonferroni post-hoc test, p < 0.05), and no significant difference was observed between the two groups for the other intervals, despite the average ratio of AP showed a larger tendency. We then fitted the recovery curves with a one-phase association equation and obtained the recovery time constants (AP: tau = 1.247 s, BP: tau = 1.746 s). The complete recovery time for capacitance from paired-pulse depression in BP hair cells was greatly delayed compared to AP [F(1,19) = 13.7, p = 0.002, two-way ANOVA followed by Bonferroni post-hoc test]. Furthermore, we calculated the Ca charge ratio to evaluate the Ca²⁺ influx during the process of paired stimulus and found no significant difference in both cell types [F(1,20) = 0.787, p = 0.39, two-way ANOVA followed by Bonferroni post-hoc test], which excluded the consequence resulted from the difference in Ca²⁺ influx. Given the quick capacitance recovery in AP hair cells, it is probably ascribed to the high rate of vesicle replenishment, and the underlying mechanism needs further investigation.