The Activity of Spontaneous Action Potentials in Developing Hair Cells Is Regulated by Ca2+-Dependence of a Transient K+ Current

Spontaneous action potentials have been described in developing sensory systems. These rhythmic activities may have instructional roles for the functional development of synaptic connections. The importance of spontaneous action potentials in the developing auditory system is underpinned by the stark correlation between the time of auditory system functional maturity, and the cessation of spontaneous action potentials. A prominent K+ current that regulates patterning of action potentials is IA. This current undergoes marked changes in expression during chicken hair cell development. Although the properties of IA are not normally classified as Ca2+-dependent, we demonstrate that throughout the development of chicken hair cells, IA is greatly reduced by acute alterations of intracellular Ca2+. As determinants of spike timing and firing frequency, intracellular Ca2+ buffers shift the activation and inactivation properties of the current to more positive potentials. Our findings provide evidence to demonstrate that the kinetics and functional expression of IA are tightly regulated by intracellular Ca2+. Such feedback mechanism between the functional expression of IA and intracellular Ca2+ may shape the activity of spontaneous action potentials, thus potentially sculpting synaptic connections in an activity-dependent manner in the developing cochlea.


Introduction
Structured spontaneous action potentials (SAPs) play an instructive role in the survival and refinement of neuronal connections in developing sensory systems [1][2][3][4][5]. Indeed, SAPs with different firing patterns encode information that may translate into variable gene expression that contributes to proper functional development [3,[6][7][8]. Similarly, developing hair cells (HCs) fire SAPs before the onset of hearing [9][10][11][12][13][14], at a period when major synaptic refinement occurs within the cochlea and cochlear nucleus (CN) [9,13,[15][16][17]. Before the onset of hearing, auditory neurons at various levels of the auditory pathway show low, bursting and occasional rhythmic spontaneous activity. This rhythmic activity is robust in the cochlear ganglion cells of the pre-hatched chickens [18,19], and the signal is abolished at higher levels by silencing the cochlea with the tetrodotoxin (TTX) injections or cochlear ablation [20]. Thus, SAPs may arise in the cochlea and it has been predicted that they may be modulated by the release of ATP from supporting cells and potentially contribute towards the establishment of proper tonotopic maps along auditory axes [21][22][23]. Incidentally, SAPs reappear during HC regeneration in chicken basilar papilla [9].
Spontaneous activity in developing HCs is Ca 2+ -dependent [9,13,16,23], producing rhythmic changes in intracellular Ca 2+ (Ca 2+ i ) [24,25]. Moreover, phasic alterations of Ca 2+ i mediate the expression of K + channels that modulate SAPs, or influence the gene expression that may refine neuronal connections [26][27][28][29][30]. 4-AP-sensitive currents, presumably of A-type K + currents (I A ) regulate spike timing and firing frequency in neurons [31][32][33] and cardiac myocytes [34]. The unique properties of the underlying channels include rapid, transient activation in the sub-threshold potentials, followed by fast inactivation. Although I A is not known to be Ca 2+ -sensitive, in neurons and myocytes, the kinetics and expression have been associated with changes in Ca 2+ i via channel protein interactions with Ca 2+ -sensitive proteins [35][36][37][38][39]. The functional relevance of these changes in Ca 2+ homeostasis and I A expression are unknown but the correlation of the feedback between Ca 2+ i and Ca 2+ -sensitive processes, such as the expression of I A , can have important ramifications in Ca 2+ oscillations and wave propagation in developing HCs [40][41][42][43].
In order to better understand the mechanisms of these spontaneously generated action potentials, we investigated the functional expression of I A and its possible sensitivity to Ca 2+ i handling. Chelation of Ca 2+ i shifted the activation properties of the current to more positive potentials and reduced the expression of I A . Our findings provided direct evidence to demonstrate that I A is tightly regulated by Ca 2+ i , a feedback mechanism that may shape the patterning of SAPs and ultimately sculpt synaptic connections in the developing cochlea.

Isolation of the Chicken Basilar Papilla
The present investigation was approved in accordance with the guidelines of the Institutional Animal Care and Use Committee of University of California, Davis. The protocol number was 15544 under the institutional authorization code A3433-01. This study included chickens at different stages of embryonic development ranging from E6-E21 as well as post-hatched chickens. Fertilized eggs were incubated at 37uC in a Marsh automatic incubator (Lyon Electric). Before experiments, chicken embryos were killed and staged according to the following: from E8-E12: based on visceral arches, feather gems and eyelids: and after E12 based on the length of the beak [44]. Basilar papillae were isolated as described previously [9]. The preparations were dissected in oxygenated chicken saline containing (in mM) 155 NaCl, 6 KCl, 4 CaCl 2 , 2 MgCl 2 , 5 Hepes, and 3 glucose, pH 7.4. The tegmentum vasculosum and the tectorial membrane were removed without any prior enzymatic treatment using a fine minutia needle. Chicken basilar papillae were stored in a 37uC incubator in Minimum Essential Medium (Invitrogen) before recording from HCs in situ. All experiments were performed at room temperature (21-23uC) within 5-45 mins of isolation. All reagents were obtained from Sigma Chemicals, unless specified otherwise.

Electrophysiology
K + currents were recorded in a whole-cell voltage-clamp configuration, using 2-3 MV resistance pipettes. Currents were amplified with an Axopatch 200B amplifier (Molecular Devices, Union City, CA) and filtered at a frequency of 2-5 kHz through a Figure 1. The pattern of spontaneous electrical activity in developing HCs is altered by I A blocker and changes in Ca 2+ I . (A) Spontaneous electrical activity recorded from apical E12 HCs using 2-mM pipette BAPTA, for 80 s before (upper trace) and after (lower trace) application of 2.5 mM 4-AP. The interspike interval distribution histogram is shown beside (right) the recorded spontaneous action potentials (SAPs). (B) SAPs recorded from basal HCs at E12 using 2 mM BAPTA as in (A) for 80 s before (upper trace) and after (lower trace) bath application of 2.5 mM 4-AP. (C) SAPs recorded from a basal E12 HC using 10-mM pipette BAPTA for 80 s before (upper trace) and after (lower trace) addition of 2.5 mM 4-AP. Inhibition of the 4-AP-sensitve current promoted tonic firing. Moreover, the inhibitory effect of the 4-AP was attenuated using 10-mM pipette BAPTA. Thus, I A contributes towards the patterning of SAPs in developing HCs, and this effect may be Ca 2+ -dependent. doi:10.1371/journal.pone.0029005.g001 low-pass Bessel filter. The data was digitized at 5-500 kHz using an analog-to-digital converter (Digidata 1200; Molecular Devices). The sampling frequency was determined by the protocols used. No online leak current subtraction was made, and only recordings with holding currents less than 50 pA were accepted for analyses. The liquid junction potentials were measured (3.560.9 mV, n = 149) and corrected online [45]. The capacitative transients were used to estimate the capacitance of the cell as an indirect measure of cell size. Membrane capacitance was calculated by dividing the area under the transient current in response to a voltage step as described [9]. The capacitative decay was fitted with a single exponential curve to determine the membrane time constant. Series resistance was estimated from the membrane time constant, given its capacitance. This study included ,703 cells with a series resistance (Rs) within a 5-15 MV range. After 60-90% compensation of the mean residual, uncompensated Rs was 5.160.5 MV. The seal resistance was typically 5-20 GV.
Action potentials were amplified (1006) with an Axopatch 200B amplifier (Axon Instruments) and filtered at 2-5 kHz through a low-pass Bessel filter. The data were digitized at 5-500 kHz using an analog-to-digital converter (Digidata 1200; Axon Instruments). The sampling frequency was determined by the protocols used. Action potentials were recorded using extracellular solution containing (in mM) NaCl 145, KCl 6, MgCl 2 1, CaCl 2 0-2, Dglucose 10, Hepes 10, pH 7.3. For whole-cell recordings of action potentials we used pipette solution containing (in mM): KCl 130, Hepes 10, D-glucose, 5 KATP, 2-10, EGTA or BAPTA. Regarding perforated patch experiments, the tips of the pipettes were filled with the internal solution containing (in mM): KCl 150, Hepes 10, D-glucose 10, pH 7.3. The pipettes were front-filled with the internal solution and back-filled with the same solution containing 250-mg/ml amphotericin. The stock solutions of all toxins were made either in ddH2O or DMSO and stored at 220uC.

Data analysis
The number of cells (n) is given for each data set. Data were analyzed using pClamp8 (Molecular Devices), Origin7.0 (Origi-nLab Corp. Northampton, MA) and Excel (Microsoft). Time constants (ts) were obtained from fits using Origin software. Time constants were obtained by fitting multiple exponential terms to the activation and decay of the current. The equation was of the form: Where I 0 is the initial current magnitude, t 1 , t 2 …t n are the time constants, and A 1 , A 2 …A n , are the proportionality constants. Voltage-dependence of activation was examined from peak amplitude (measured at ,3-5 ms from the onset of voltage step) of tail currents at different developmental stages (embryonic, E) daysix to postnatal (P) day-two (E6-P2), then normalized steady-state curves were fitted with the Boltzmann distribution. Additionally, the steady-state inactivation curve was generated from normalized currents measured at a test potential following several conditioning pre-pulses. Pooled data were presented as mean 6 SD. Significant differences between groups were tested using the Student's t test, with p,0.05 or 0.01 indicating statistical differences.

Pharmacological elimination of I A profoundly alters the structure of SAPs in developing hair cells
To examine the roles of I A on SAPs in developing HCs, we employed pharmacological strategies [9,46]. We applied 2.5 mM 4-AP, voltage-dependent K + -current (I A ) blocker, to spontaneously active HCs from the developing basilar papilla. Shown in Figures 1A and B are the effects of 4-AP on the patterning of SAPs. An  Table 1. Summary data on the effects of position and intracellular Ca +2 buffering on spike width, and rate of change of voltage of the depolarizing and repolarizing phase of spikes at E12.     immediate observation is that 4-AP induces marked alterations in the frequency of AP firing, which is reflected in a patent reduction in the inter-spike intervals plotted on the right panels (Fig. 1). For example, 4-AP induced a 10-fold increase in AP firing in developing HCs at the apical aspects of the basilar papilla at E12 (pre-4-AP: 0.1160.05 Hz; post-4-AP: 1.160.2 Hz, p,0.01; n = 7). Similarly, basal HCs at E12 fire APs at the rate of ,0.5 Hz (0.560.2 Hz, n = 9). Moreover, after application of 4-AP, the frequency of firing increased to ,2 Hz (2.160.6 Hz, n = 9; p,0.05). These experiments were performed using 2 mM BAPTA in the pipette solution to mimic physiological buffering capacity [47]. Spontaneous action potentials in developing HCs are Ca 2+ -dependent [9], and to assess the role of intracellular Ca 2+ (Ca 2+ i ) in 4-AP-mediated changes in AP firing rates, we used 10 mM pipette-BAPTA. Surprisingly, the typical phasic and burst pattern seen in E12 apical HCs (Fig. 1A) was replaced with tonic AP firing in the presence of 10 mM BAPTA (Fig. 1C). Even more startlingly was the resulting lack of 4-APmediated changes in the presence of 10 mM BAPTA. Changes in frequency of firing versus pipette-BAPTA concentrations are shown (Fig. 2). The correlation between 4-AP-mediated effects and high concentrations of BAPTA suggests that the 4-AP-sensitive current (I A ) may be sensitive to Ca 2+ I availability. Additionally, the data hinted that I A regulates the firing pattern of developing HCs in a Ca 2+ I -dependent manner. The effects of BAPTA and 4-AP on individual action potentials are outlined in Table 1.
The structure of SAPs in developing HCs not only differs between apical and basal cells, but also varies at different developmental stages as shown in figure 3. Along the tonotopic axes, the general trend varied from phasic bursting apical cells to tonic bursting basal cells (Fig. 3). Thus, we reasoned that different kinetics and the magnitude of I A during development, and tonotopic expression of the current could, in part, be responsible for diverse spiking activity (Fig. 3).

I A is expressed at the earliest stages of HC development
Previous studies have shown that K v 4.2 and K v 1.x channels are the presumptive channels that generate I A in chicken HCs. These channels were expressed, at the level of mRNA, at the earliest stages of development [48]{Sokolowski, 2004 #192, [49]}. However, functional studies in mature HCs showed that the current was limited to basal HCs [50,51]. We determined the functional expression and kinetics of I A in HCs with respect to developmental age and position along the tonotopic axis. Figure 4 shows examples of whole-cell current profiles and current-voltage relations in HCs at the apical (Fig. 4A) and basal (Fig. 4B) aspects of the basilar papilla elicited from different holding potentials at E8-P2 ( Fig. 4C; summary data). The transient component of the current was remarkably sensitive to holding voltages throughout development. The differencecurrent between current traces generated at a holding potential of 290 and 230 mV was the main transient component. The magnitude of the transient current plummeted as HCs matured (Fig. 4C), and the transient outward current was virtually absent in mature apical HCs. To ensure that the transient current was indeed I A , we examined the sensitivity of the current to 4-AP [50,51]. A sizable portion (,50% at E10) of the sustained component of the outward K + current elicited from a holding potential of 290 mV was suppressed upon application of 20 mM TEA and 300 nM apamin in the bath, revealing a fast inactivating current that was sensitive to 2.5 mM 4-AP (Fig. 5A). Figure 5B illustrates the current-voltage relations at E10 and E18, and figure 5C shows the summary data, revealing a striking agreement between the magnitudes of the 4-AP-and holding potential-sensitive currents across different developmental stages of basal HCs. The result is in keeping with molecular biological evidence showing the expression of K v 4.2 channels in developing HCs [48]. Table 2. Summary data on the effects of development and tonotopic position of HC-A-current activation and inactivation (V 1/2 (mV), k (mV)) with 2 mM BAPTA in pipette solution. Although there is a small difference in voltage dependence of inactivation between apical and basal region of the papilla, this difference becomes significant at more mature stages (n = 7 cells). doi:10.1371/journal.pone.0029005.t002 Table 3. Voltage dependence of activation (V 1/2 (mV), k (mV)).

Changes in voltage-dependent activation and inactivation of I A during development
The voltage-dependence of the activation and inactivation of I A varied with age ( Figure 6, Tables 2, 3, 4). Differences in voltagedependent properties of I A between apical and basal HCs were stark at E18 compared to E12 (Fig. 6B). At E12, the V 1/2 of the activation curve was ,0.5 mV for apical and ,25.5 mV for basal cells. The half-activation voltage (V 1/2 ) of inactivation was ,256 mV for apical and ,259 mV for basal cells. By E18, the transient outward current in basal HCs inactivated at more negative potentials than apical HCs ( Figure 5B; Table 2). At E18, the V 1/2 of inactivation was ,247 mV for apical and 263 mV for basal cells. Consistent with previous reports, the data suggest that as HCs mature, I A is mainly confined to cells at the basal aspects of the cochlea [50][51][52].

Changes in Ca 2+ I buffering have a profound effect on I A
To examine the underlying mechanisms for the BAPTAinduced 4-AP-insensitivity to SAPs described in figure 1, we tested I A sensitivity to acute changes in Ca 2+ chelators. We used BAPTA-AM, a selective Ca 2+ chelator that is the cell-permeable analog of BAPTA. Application of 5 mM BAPTA-AM to basal HCs produced a marked reduction of the transient K + current (Fig. 7A). Moreover, raising the external Ca 2+ from 0 to 2 mM had no effect on the transient current (Fig. 7B), hinting that Ca 2+ entry via Ca 2+ channels may not be sufficient to affect I A . Rather, the release of Ca 2+ i may be the Ca 2+ source to alter I A . We probed the Ca 2+ -dependence of I A further by examining the voltage-and time-dependent properties of the current using varying concentrations of BAPTA (Tables 3, 4, 5). Figure 8 illustrates the effects of 2 and 10 mM intracellular BAPTA on I A in an apical E12 HC. Reducing the available Ca 2+ I by using 10 mM pipette BAPTA produced a significant shift in the voltage-sensitivity of activation. The V 1/2 s of the steady-state activation were (in mV) ,0 and 23 using (in mM) 2 and 10 BAPTA, respectively. Moreover the V 1/2 s of the steady-state inactivation were ,256 mV and 248 mV using (in mM) 2 and 10 BAPTA, respectively. Additionally, we examined the time-dependence of the development of inactivation after varying durations at 250 mV and 290 mV. The time constants of inactivation (t I ) were compared at different stages of development and with respect to different concentrations of Ca 2+ I buffer (Table 5). At all developmental stages tested (E12-E18), the kinetics of decay of the current were fitted with two ts. As shown in figure 9, the use of different concentrations of BAPTA had a marked effect on the amplitude and inactivation kinetics of I A . For example, for an apical E12 HC, the amplitude of the current measured at +40 mV was ,825 pA in 2 mM BAPTA and ,405 pA in 10 mM BAPTA. Moreover, inactivation time constants (t I s) of I A were ,5 ms and 75 ms, using 2 mM BAPTA. Using 10 mM BAPTA the ts were ,75 ms and 300 ms (Fig. 9A). Table 4. Voltage dependence of inactivation (V 1/2 (mV), k (mV)).  Similarly, the time dependence of development and recovery from inactivation were duly affected (Figs. 9B-D). Whereas the current shows complete inactivation within ,75 ms, using 2 mM BAPTA, the time constant of inactivation was prolonged by , 4-fold (,300 ms) in the presence of 10 mM BAPTA ( Figure 9C). Meanwhile, the current recovered from inactivation with a time constant of ,450 ms in 2 mM BAPTA and ,1200 ms in 10 mM BAPTA (Fig. 9D). The results clearly suggest that the 4-APsensitive current is regulated tightly by Ca 2+ i .

Discussion
The transient outward K + current (I A ) in the developing chicken HC was identified functionally by Murrow (1994) and Griguer and Fuchs (1996) [51,53] and the molecular characteristics were revealed later [48] {Rajeevan, 1999 #193, [52]}. Moreover, the short and long-term modulation of I A in developing chicken HC is unknown, despite the currents purported role in spike timing in SAPs. Also important, roles of I A and its regulation and the ensuing effects on the structure of SAPs in developing HCs are unclear. This study fills in major gaps in our understanding of the patterning of SAPs in HCs. Our findings include; 1) I A contributes towards the patterning of SAPs. Since the current dictates the spike timing and firing pattern, it is conceivable to suggest that it may regulate the release of neurotransmitters in the developing cochlea to sculpt synapse formation. 2) Expression of I A is regulated developmentally and tonotopically. At the apical aspects of the basilar papilla the expression of I A is prominent only in early development. By contrast, I A persists in HCs at the basal aspects of the matured basilar papilla. 3) Alterations of Ca 2+ I by Ca 2+ buffers produced marked effects on the functional expression, voltagedependence and kinetics of I A . Increased concentrations of the Ca 2+ I buffer, BAPTA shifted the voltage-dependence of activation and inactivation to more positive potentials and reduced the number of functional channels expressed. These findings reaffirm the tight interplay that occurs between Ca 2+ I , and K + currents. Suppression of I A visibly alters spike timing and frequency, transforming phasic action potential burst into tonic firing in spontaneously active chicken HCs. The results are reminiscent of the effects of a transient outward current in the developing mouse HC [16]. However, after increased Ca 2+ buffer, not only did the phasic patterning of spikes appear tonic, but also suppression of I A had no effect on the firing pattern, raising the possibility that reduced Ca 2+ I could stifle the availability of I A . The deduced amino acid sequence of the a-subunits of K v 1.2, K v 1.3 and K v 1.5 and the   [55,56]. Alternatively, it is conceivable that the presences of multiple potential Ca 2+ -dependent regulatory sites at the C-terminus of the channel and other interacting partners that form the channel complex could affect the localization, expression and specific functional properties of the native K + conductance, as observed in cardiac myocytes and neurons [57][58][59][60][61][62][63]. Ca 2+ handling in the developing chicken basilar papilla may undergo marked plasticity to account for varying demands for the expression of Ca 2+ -dependent processes and increased Ca 2+ influx through increased transduction channel numbers [64]. For (A) Normalized current traces elicited using ,1 s pulse for E12 cells using 2 mM (dark gray line) and 10 mM BAPTA (light gray line) in the pipette (Left). The fast components of decay were eliminated with increased BAPTA as well as with maturation. Holding potential was 290 mV and voltage steps were in 10 mV increments. Current decay could be best fitted with two different time constants (t) at all stages of development. The time constants were prolonged with an increase in intracellular BAPTA as well as with maturation [ts (in ms) for E12, using 2 mM BAPTA were: 563, and 75613 (n = 11 cells); ts for 10 mM BAPTA were 6763, and 498613 (n = 10 cells) and for E18: 52613, and 611625 (n = 7 cells)]. For the example shown at E12, the exponential fits yielded t 1 = 5 ms and t 2 = 70 ms for 2 mM BAPTA. The current traces illustrated from the E18 HCs were fitted with t 1 = 50 ms and t 2 = 600 ms. See Table 4  An example of such protocol on an E12 cell is shown (Left). The exponential fits to the data are shown on the right; ts (in ms) for E12: 5156161 using 2 mM BAPTA and 9976203 ms using 10 mM BAPTA in the pipette (n = 6 cells). (D) To estimate how quickly the current recovered from inactivation, we measured recovery kinetics after 1-second at the holding potential of 290 mV, using standard recovery time protocol. We measured the amplitude of the transient potassium currents by depolarizing to a fixed potential after variable time at 290 mV. Amplitude of these currents was normalized to the amplitude of the currents activated from 290 mV and plotted against the duration of steps to 290 mV. Examples of traces generated to examine the time course of the recovery from inactivation for E12 currents (Left). The right panel shows the time course and the exponential fits [Fits in solid lines, ts for E12: 446656 ms using 2 mM BAPTA and 1286678 ms using 10 mM BAPTA in the pipette (n = 5 cells)]. Amplitude of the currents was normalized to the amplitude of the currents activated from a holding potential of 290 mV and plotted against the duration of step to 290 mV. doi:10.1371/journal.pone.0029005.g009 example the expression of the mobile Ca 2+ buffer, calbindin, is regulated during development and along the tonotopic axis of the cochlear in accordance with the demands' of Ca 2+ regulation [64]. Similar assortments of Ca 2+ regulation can be seen in the apicobasal gradient of the basilar papilla [65]. The differential and increased expression of Ca 2+ buffers during development dovetails well with the expression pattern of Ca 2+ buffers [64]. In keeping with the expression pattern of Ca 2+ buffers in HCs of the developing basilar papilla and our findings, the kinetics and voltage-dependence of I A are expected to be altered. Voltage dependence of the activation and inactivation of I A are shifted to more positive potentials with increased Ca 2+ buffers. Throughout development, basal HCs express I A of faster kinetics of inactivation than at the apical aspects of the basilar papilla. Moreover, with maturation this difference is magnified as apical HCs lose the transient component of K + currents. At the basal aspects, the ratio of I A to the total outward K + current is diminished with maturation. Whereas these findings are in accordance with changes in Ca 2+ handling in the chicken basilar papilla, they are in stark contrast to reports on the expression of 4-AP sensitive current, presumably I A , in the developing mouse HC, where the expression of the current increases with development and persists in mature HCs [16,66]. Together, our data demonstrates that reduction of Ca 2+ I availability has profound effects not only on the kinetics of I A and the number of channels expressed in HCs, but also on the ensuing SAP, which is sculpted by the current.