Effects of Strontium on the Permeation and Gating Phenotype of Calcium Channels in Hair Cells

Abstract

To minimize the effects of Ca²⁺ buffering and signaling, this study sought to examine single Ca²⁺ channel properties using Sr²⁺ ions, which substitute well for Ca²⁺ but bind weakly to intracellular Ca²⁺ buffers. Two single-channel fluctuations were distinguished by their sensitivity to dihydropyridine agonist (L-type) and insensitivity toward dihydropyridine antagonist (non-L-type). The L- and non-L-type single channels were observed with single-channel conductances of 16 and 19 pS at 70 mM Sr²⁺ and 11 and 13 pS at 5 mM Sr²⁺, respectively. We obtained K_D estimates of 5.2 and 1.9 mM for Sr²⁺ for L- and non-L-type channels, respectively. At Ca²⁺ concentration of ∼2 mM, the single-channel conductances of Sr²⁺ for the L-type channel was ∼1.5 and 4.0 pS for the non-L-type channels. Thus the limits of single-channel microdomain at the membrane potential of a hair cell (e.g., −65 mV) for Sr²⁺ ranges from 800 to 2,000 ion/ms, assuming an E_Ca of 100 mV. The channels are ≥4-fold more sensitive at the physiological concentration ranges than at concentrations >10 mM. Additionally, the channels have the propensity to dwell in the closed state at high concentrations of Sr²⁺, which is reflected in the time constant of the first latency distributions. It is concluded that the concentration of the permeant ion modulates the gating of hair cell Ca²⁺ channels. Finally, the closed state/s that is/are altered by high concentrations of Sr²⁺ may represent divalent ion-dependent inactivation of the L-type channel.

INTRODUCTION

Ca²⁺ channels are multi-ion pores that allow the passive flow of several divalent, and under restricted conditions, monovalent cations (Hess et al. 1986; Hille 2001). Among the permeant ions, Sr²⁺, by virtue of its similar charge and size to Ca²⁺, produced currents that are akin to Ca²⁺ currents (Bourinet et al. 1996; Byerly et al. 1985; McNaughton and Randall 1997). The magnitude and time constants of activation and deactivation of whole cell Sr²⁺ currents are similar to the properties of Ca²⁺ currents of Ca_v2.2 channels. The properties of Sr²⁺ currents are not identical to Ca²⁺ currents, however, and the differences rely mainly on the magnitude and extend inactivation (McNaughton and Randall 1997). Moreover, the effects of Sr²⁺ on the elementary properties of Ca²⁺ channels remain uncertain. Instead, much attention has focused on studies of the effects of Ba²⁺ on the macroscopic and microscopic phenotype of Ca²⁺ channels, partly because Ba²⁺ currents are ≥2-fold larger than Ca²⁺ currents, and Ba²⁺ currents are moderately stable compared with the fast rundown of Ca²⁺ currents (Fox et al. 1987; Rodriguez-Contreras et al. 2002; Wakamori et al. 1998; Yue and Marban 1990). Ba²⁺ markedly alters not only the voltage- and time-dependent activation and deactivation kinetics but also abolishes inactivation in Ca²⁺ channels that rely on Ca²⁺ and calmodulin in conferring the decay of the current (Imredy and Yue 1994; Mori et al. 2004). Apparently, whereas Ba²⁺ is a poor substitute for Ca²⁺ in various biological processes, Sr²⁺ can replace Ca²⁺ in several Ca²⁺-dependent reactions (Falke et al. 1994; Xu-Friedman and Regehr 1999).

In addition to traversing Ca²⁺ channels, Sr²⁺ mediates neurotransmitter release in the squid giant synapse (Augustine and Eckert 1984), the neuromuscular junction (Bain and Quastel 1992; Dodge et al. 1969), hippocampal synapse (Abdul-Ghani et al. 1996), and cerebellar synapses (Falke et al. 1994; Xu-Friedman and Regehr 1999). Invariably, the effects of Sr²⁺ on neurotransmitter release are desynchronized and reduced compared with Ca²⁺, suggesting that Sr²⁺ is less effective than Ca²⁺ in triggering exocytotic transmitter release (Falke et al. 1994; Xu-Friedman and Regehr 1999).

Ohmori (1985) showed that, in vestibular hair cells of the chick, Sr²⁺ maintained mechanoelectrical transducer currents. In addition, activation of Ca²⁺-activated K⁺ channels (BK channels) can be supported by Sr²⁺ in goldfish saccular hair cells (Sugihara 1998). Because hair cells in lower vertebrates rely on the interplay between the activation of voltage-gated Ca²⁺ channels and the activation of BK channels to confer electrical tuning (Armstrong and Roberts 1998; Art et al. 1993; Fuchs and Evans 1988; Lewis and Hudspeth 1983), and Sr²⁺ can replace Ca²⁺ to induce membrane electrical resonance, we surmised that understanding the effects of Sr²⁺ on the elementary properties of Ca²⁺ channels in hair cells would provide clues to the mechanisms of hair cell functions. Indeed, to prevent cross-talk between different Ca²⁺-dependent pathways in hair cells, the macro-domain Ca²⁺ handling is tightly regulated by mobile Ca²⁺ buffers (Hall et al. 1997). Thus incoming Ca²⁺ at the micro-domains of Ca²⁺ channels is a major determinant of Ca²⁺-dependent functions. Additionally, because intracellular Ca²⁺ buffers bind weakly with Sr²⁺ (Yawo 1999), we rationalized that Sr²⁺ flux through Ca²⁺ channels represents an unadulterated means to study the throughput of the channel. Thus we determined the unitary current phenotypes of Ca²⁺ channels of bullfrog saccular hair cells using Sr²⁺ as the charge carrier.

METHODS

American bullfrogs (Rana Castebeina) were purchased from West Jersey Biological Supply Co. (Wenonah, NJ) and Nebraska Scientific (Omaha, NB) and kept in the laboratory in freshwater circulating containers. The methods for dissociating single hair cells from the saccule of bullfrogs and for the set-up of patch-clamp experiments (Hamill et al. 1981) have been described previously (Rodriguez-Contreras and Yamoah 2001; Rodriguez-Contreras et al. 2002; Yamoah and Gillespie 1996). All chemicals were obtained from Sigma Chemical (St. Louis, MO) unless otherwise noted. Bullfrogs were killed, and inner ears were removed and placed quickly in oxygenated, low Ca²⁺ solution (LCS; in mM): 110 NaCl, 2 KCl, 3 d-glucose, 0.1 CaCl₂, and 5 HEPES, pH 7.4, with NaOH. To disrupt the macular tight junctions, the preparation was exposed to 4 mM EGTA for 15 min. After washing the preparation with fresh saline, the saccular macula was isolated and incubated in frog saline containing 50 μg/ml protease (type XXIV) for 20 min. The tissue was washed and transferred to frog saline containing 1 mg/ml bovine serum albumin (Sigma) and 2 mg/ml DNAse I (Worthington, Lakewood, NJ) for 10 min. This procedure has been used previously (Chabbert 1997) and differs from more severe enzymatic treatments that can affect ionic conductances (Armstrong and Roberts 1998). The preparation was washed and placed on the recording chamber at 4°C before recording at room temperature. The otolithic membrane was removed mechanically, and hair cells were dissociated from the macula using an eyelash. Hair cells were allowed to settle onto the bottom of the recording chamber (coated with the lectin concavalin-A) for 20 min before patch-clamp recording.

Patch pipettes were pulled from borosilicate and quartz glass capillaries (World Precision Instruments Sarasota, FL and Sutter Instruments, San Rafael, CA) on a Flaming-Brown and laser microelectrode pullers (Sutter Instruments). For single-channel recordings, the borosilicate glass electrodes were coated with Sylgard 184 (Dow-Corning, Midland, MI) ≤100 μm from the tip and fire-polished before use. Bath solution contained (in mM) 90 NaCl, 25 tetraethylammonium chloride (TEACl), 5 4-aminopyridine (4-AP), 5 BaCl₂/CaCl₂/SrCl₂, 3 glucose, and 5 HEPES (pH 7.4). Ca²⁺ currents were recorded using perforated patch electrodes (resistances 2–4 MΩ) whose tips contained (in mM) 120 CsCl and 5 HEPES (neutralized to pH 7.3 with CsOH). To gain electrical assess to hair cells and to minimize wash-out of intracellular molecules, patch electrodes were backfilled with solution containing (in mM) 130 CsCl, 1 CaCl₂, 5 HEPES, and amphotericin 200 μg/ml (neutralized to pH 7.3 with CsOH) (Korn et al. 1991). To ensure that recordings are made in the perforated-patch mode instead of whole cell mode, the backfilled solution of the patch electrode contained 1 mM Ca²⁺. A switch from the perforated to whole cell mode resulted in cell death caused by Ca²⁺ toxicity. Series resistance (5–10 MΩ) was compensated (nominally 50–60%). Liquid-junction potentials were not corrected because they were <2 mV. Voltage-clamped Ba²⁺/Ca²⁺/Sr²⁺ currents were amplified with Axopatch 200B amplifiers (Molecular Devices, Sunnyvale, CA). Cell capacitance was calculated by integrating the area under an uncompensated capacitive-transient elicited by a 20-mV hyperpolarizing pulse from a holding potential of −80 mV. Cell capacitance and series resistance was compensated as much as possible almost to the point of ringing. In general, 60–80% of the series resistance was compensated. Current records were filtered at 2–5 kHz with a low-pass Bessel filter and digitized at 10 kHz with a Digidata interface controlled by custom-made software.

For single-channel recordings, patch electrodes were filled with a Sr²⁺ solution (5–70 mM) containing (in mM) 20 TEACl, 5 4-AP, and 5 HEPES at pH 7.4 (adjusted with TEAOH). N-methyl-d-glucamine (NMG) was used to substitute for divalent Sr²⁺ and to maintain an osmolarity of ∼280 mmosmol. Quartz glass electrodes were used to record single-channel fluctuations when the Sr²⁺ concentration was <20 mM. To identify L-type channels, experiments were conducted in the presence of the channel's agonist Bay K 8644 (Calbiochem, La Jolla, CA). Stock solutions of Bay K 8644 (100 mM) were made in DMSO, and a final concentration of 5 μM was used. The bath solution contained (in mM) 80 KCl, 3 d-glucose, 20 TEACl, 1 CaCl₂, 5 4AP, and 5 HEPES, pH 7.4, with TEAOH, to shift the resting potential to ∼0 mV (Rodriguez-Contreras and Yamoah 2003). Here, liquid junction potentials were measured and corrected as described previously (Rodriguez-Contreras et al. 2002). Experiments were carried out at room temperature (∼21°C).

Data analysis

Whole cell Ba²⁺/Ca²⁺/Sr²⁺ current amplitudes at varying test potentials were measured at the peak and steady-state levels using a peak and steady-state detection routine. For single channel records, leakage and capacitative transient currents were subtracted by fitting a smooth template to null traces. Leak-subtracted current recordings were idealized using a half-height criterion. Transitions between closed and open levels were determined using a threshold detection algorithm, which required that two data points exist above the half mean amplitude of the single-unit opening. The computer-detected openings were confirmed by visual inspection, and sweeps with excessive noise were discarded. Amplitude histograms at a given test potential were generated and fitted to a single Gaussian distribution using a Levenberg-Marquardt algorithm to obtain the mean and SD. At least five voltage steps and their corresponding single-channel currents were used to determine the unitary conductance. Single-channel current-voltage relations were fitted by linear least-square regression lines and single-channel conductances were obtained from the slope of the regression lines. Idealized records were used to construct ensemble-averaged currents and open probability (P_o), as well as to generate histograms for the distributions of open and closed time intervals. P_o, voltage, and recording time surface plots were generated using a built-in matrix function from Origin software (MicroCal, Northampton, MA). To ensure that the recordings were made on single L-type channels, only patches containing single channel events that were blocked by nimodipine (5 μM) were analyzed. Additionally, kinetic analyses were performed on patches, which contained only one channel. The criteria consisted of quantitative determination followed by visual inspection of the data. The patches contained one channel, as there was no stacking of events. Furthermore, direct transition from fast to slow kinetics with similar current amplitude, and vice versa, were observed in the selected recordings, indicating that the gating modes were derived from a single channel. The cumulative first latency histograms were determined as described by (Zei and Aldrich 1998). Curve fits and data analyses were performed using Origin software (MicroCal). All averaged and normalized data are presented as means ± SE. Frogs were housed and killed using a protocol approved by the University of California, Davis, IACUC Committee.

RESULTS

We determined the similarities and differences in the magnitude and profile of whole cell Ba²⁺, Ca²⁺, and Sr²⁺ currents in hair cells. Whereas replacing Ca²⁺ with Ba²⁺ in the recording bath solution strongly affected the amplitude and decay of the whole cell inward currents in saccular hair cells, the effects of Sr²⁺ were relatively modest (Fig. 1). Inward currents were elicited from a holding potential of −60 mV and a step potential to −10 mV. Outward K⁺ currents were blocked with external 4-AP and TEA as well as internal Cs⁺. On switching the bath solution, which contained 5 mM Ca²⁺, to one that contained the same concentrations of Ba²⁺, the inward current increased by ∼1.4-fold. The mean amplitudes of the current (in pA) for Ca²⁺ and Ba²⁺ were 778 + 56 and 992 ± 101 (n = 7; P < 0.01), respectively. Additionally, Ba²⁺ abolished the slow decay of the current profile. In contrast, in 5 mM Sr²⁺, the inward current mirrored roughly the Ca²⁺ current in magnitude, activation, and deactivation. The mean amplitudes of the current (in pA) for Sr²⁺ and Ca²⁺ were 755 ± 62 and 811 ± 45 (n = 7; P = 0.08), respectively. Inactivation kinetics differed slightly, which is shown in the inset in Fig. 1. Although some of the macroscopic features of Sr²⁺ and Ca²⁺ currents appeared similar, single-channel currents should confirm these parallels and reveal the distinct properties of Sr²⁺ currents.

FIG. 1.

Whole cell Sr²⁺ current compared with Ca²⁺ and Ba²⁺ currents from hair cells. Examples of inward current traces recorded using a depolarizing voltage step from a holding potential of −60 mV to a step potential of −10 mV. The current traces were obtained from the same cell after changing the external solution in the order from Sr²⁺ to Ca²⁺ to Ba²⁺ (5 mM). Shown in the inset are normalized current traces to show the effects of the charge carrier on the activation and inactivation of the current traces. For comparison, the Sr²⁺ current trace is shown in solid, and the Ca²⁺ and Ba²⁺ currents are depicted with dotted lines. The mean amplitude of the current (in pA) for Sr²⁺ and Ca²⁺ were 755 ± 62 and 811 ± 45 (n = 7; P = 0.08) and Ca²⁺ and Ba²⁺ were 778 + 56 and 992 ± 101 (n = 7; P < 0.01), respectively.

Inward Ca²⁺ currents from hair cells in the saccule are generated by two distinct channels: nimodipine-sensitive (L-type) and -insensitive (non-L-type) channels (Rodriguez-Contreras and Yamoah 2001). Figure 2, A–D, shows examples of single-channel Sr²⁺ currents recorded from hair cells isolated from the frog saccule. Data in Fig. 2, A and B, represent Bay K 8644–modified currents. In contrast, the non-L-type channel currents were recorded with patch-pipettes containing 10 μM nimodipine (Fig. 2, C and D). We examined the permeation phenotype of the two-channel subtypes using different concentrations of patch-pipette Sr²⁺. For the traces shown, the amplitude histograms at different step potentials were generated to determine the mean single-channel amplitude as shown in Fig. 2E, and the resultant current-voltage relations yielded single-channel conductances (γ, in pS) of 17, and 8 for 70 and 5 mM, respectively for the L-type channel, whereas 18 and 13 pS were obtained for the same concentrations of Sr²⁺ for the non-L-type channel. The mean values for the single-channel conductances for Sr²⁺ followed in descending sequence from 70 to 5 mM for the L-type and non-L-type channels (Rodriguez-Contreras and Yamoah 2001). Respectively, they are 16.2 ± 1.9 and 18.7 ± 1.7 for 70 mM (n = 6) and 11.2 ± 1.8 and 13.3 ± 2.6 for 5 mM (n = 5).

FIG. 2.

Permeation of single L-type and non-L-type channels by Sr²⁺. Under outside-out cell-attached configuration, hair cell membrane patches were held at −70 mV and stepped to the depolarizing potentials, indicated beside the traces. The potentials have been corrected for liquid junction potential errors. A: typical consecutive inward single-channel current traces using 70 mM Sr²⁺ and 5 mM (B) as charge carriers are shown for the L-type channel. The L-type channel activity was recorded in the presence of 5 μM BayK 8644 to promote long openings. The solid lines indicate the closed levels. C and D: non-L-type current traces recorded using patch pipettes containing 10 μM nimodipine using 70 and 5 mM Sr²⁺, respectively, as the charge carrier. E: examples of the amplitude histograms from current traces at the step potentials indicated, which were used to generate the I-V plots. F: the scattered points of the I-V relation were fitted with a linear regression to determine to conductance of the channels. For the examples shown, the single channel conductance values for L-type and non-L-type channels in 70 and 5 mM Sr²⁺ are (in pS; legend = L) as follows: L₇₀, ○ = 17; L₅, • = 5; non-L₇₀, ▪ = 18; non-L₅, □ = 13.

To determine the sensitivity range of Ca²⁺ channels in hair cells for Sr²⁺, we examined the relation between the single-channel conductances versus Sr²⁺ concentrations. Using a Langmuir isotherm, a fit to the data accrued from recordings from the L-type channel (Fig. 3A) and the non-L-type channel (Fig. 3B) showed two important features of the channels: 1) the apparent dissociation constants of L- and non-L-type channels for Sr²⁺ ions were (in mM) 5.2 ± 1.9 and 1.9 ± 1.5 (n = 7), respectively; 2) meanwhile, the maximum conductances for the two channels were (in pS) 16.6 ± 1.4 and 19.2 ± 1.9 (n = 7) for the L- and non-L-type channels, respectively, and were reached at external Sr²⁺ concentrations above ∼20 mM. Similar ranges of results have been reported for Ca²⁺ and Ba²⁺ currents (Rodriguez-Contreras et al. 2002). The ascending limbs of the conductance versus concentration curves for the two channel subtypes suggest that the sensitivity range of the channel for the permeant ion lies between 1 and 5 mM. Thus optimum assessment of Ca²⁺ channel functions should be evaluated within these narrow ranges of concentrations.

FIG. 3.

Conductance vs. concentration curves. The conductance of the L- and non-L-type channels were determined using different concentrations of Sr²⁺ (2.5–70 mM). The affinity of Sr²⁺ (○) for L- (A) and non-L-type (□). (B) Ca²⁺ channels in hair cells in the frog saccule was estimated from these plots. The solid lines represent fits with a Langmuir isotherm of the form γ = γ_max/(1 + K_D/[Sr²⁺]), where γ and γ_max represent the single channel conductance and maximum conductance, respectively. The dotted lines show the conductance-concentration curves for Ca²⁺ (Rodriguez-Contreras and Yamoah 2003). The estimated K_Ds and γ_maxs of the 2-channel subtypes for Sr²⁺ were as follows: L-type, 5.2 ± 1.9 and 16.6 ± 1.4 (n = 7); non-L-type, 1.9 ± 1.5 and 19.2 ± 1.9 (n = 7), respectively.

To ascertain whether the concentrations of Sr²⁺ ions have noticeable effects on the voltage dependence and kinetics of the channels, we examined the properties of the L-type channel at nonsaturating (5 mM) and saturating (70 mM) concentrations. To ensure that the recordings were made on single L-type channels, only patches containing single channel events that were blocked by nimodipine (5 μM) were analyzed. Additionally, kinetic analyses were performed on patches, which contained only one channel. The criteria consisted of quantitative determination followed by visual inspection of the data. The patches contained one channel, because there was no stacking of events. Furthermore, direct transition from fast to slow kinetics with similar current amplitude, and vice versa, were observed in the selected recordings, indicating that the gating modes were derived from a single channel. Figure 4 is a layout of cell-attached recordings of a family of consecutive single channel current traces, recorded using 5 (A) and 70 mM (B) Sr²⁺ as the charge carrier. With 5 mM Sr²⁺, the dwell times of the openings were frequent and long compared with recordings in 70 mM. Another visible difference was that, at low concentrations of the permeant ion, the openings were sustained, which was reflected in the persistent ensemble-averaged currents shown below the corresponding family of traces. In contrast, in 70 mM Sr²⁺, the channel had frequent openings predominantly only at the early phase of the step voltage, which is readily visible in the well-resolved decay of the ensemble-averaged currents. Notwithstanding the presence of 5 μM Bay K, an agonist of L-type channels, in the bath and pipette solutions, the probability of channel opening (P_o) was invariably <1 (Fig. 5). The rightward shift in the voltage dependence of activation as the pipette Sr²⁺ concentrations were raised from 5 to 70 mM is in keeping with surface charge screening effect of divalent cations (Rodriguez-Contreras and Yamoah 2003; Rodriguez-Contreras et al. 2002; Zhou and Jones 1995). The inference from the finding that the P_o was less than ∼0.5 is that the channel has multiple nonconducting states. Indeed, by increasing the Sr²⁺ concentrations (70 mM), fluctuations of the open probability (P_o) plummeted in comparison to reduced concentrations (5 mM; Figs. 5 and 6, A and B), suggesting that the dwell times in the closed states may be current or ion concentration dependent. For graphic illustrations, we depict the voltage and time dependence of the normalized P_o of recordings performed at Sr²⁺ concentrations of 70 (Fig. 6A) and 5 mM (Fig. 6B). The gating of the channel was noticeably altered in several ways: 1) at saturating concentrations of Sr²⁺ (70 mM), nonconducting states are more favored than open states; 2) although channel openings are stochastic in nature, the P_o of the channel remains infinitesimally low at negative step potentials (−60 to −40 mV), and even at step potentials greater than −40 mV, the increase in P_o is relatively modest compared with recordings done at Sr²⁺ concentrations (5 mM) within the sensitivity range of the channel; 3) finally, at the steeper phase on the sensitivity curve of the channel, the P_o of the channel was high at negative voltages and remained so at more positive voltages, indicating that, at low concentrations of the permeant ion, the channel has increased propensity to exit from the nonconducting states. The shift in the activation of the channel using 5 and 70 mM Sr²⁺ is shown in the normalized P_o versus voltage plot (Fig. 6C). The midpoints of activation (in mV) were −24.1 ± 1.2 and −13.7 ± 1.5 (n = 7), and the slope factor of the curves were virtually unchanged (in mV) at 4.3 ± 0.7 and 4.4 ± 0.5 (n = 7) for 5 and 70 mM Sr²⁺, respectively. Moreover, the slope of the relation between the midpoint of activation and Sr²⁺ concentrations (0.20 VM⁻¹; Fig. 6D) was similar to the value obtained for Ca²⁺ as the permeant ion (Rodriguez-Contreras and Yamoah 2003). Although the ∼10-mV rightward shift in the activation curves as the external divalent cation concentration was increased from 5 to 70 mM was consistent with surface charge screening effects, the value fell short of the predicted shift assuming the absence of permeant ion binding interaction with surface charges (∼25 mV) (Rodriguez-Contreras and Yamoah 2003; Zhou and Jones 1995). Thus Sr²⁺ binding to the surface charge may dictate the extent of the shift in the activation curve (Rodriguez-Contreras and Yamoah 2003).

FIG. 4.

Family of single channel [dihydropyridine-sensitive (L-type)] traces recorded in 5 and 70 mM Sr²⁺. A: exemplary consecutive single channel traces were generated at different test potentials (indicated on top of each column of traces). The holding potential was −70 mV and the concentration of Sr²⁺ in the patch pipette was 5 mM. The ensemble averaged current traces shown below the single channel current traces were derived from ∼150 consecutive sweeps. At 5 mM Sr²⁺, the ensuing ensemble averaged currents were sustained in nature. The scale bars (horizontal = 100 ms and vertical = 1 pA). B: similar consecutive sweeps of single channel current traces obtained when the patch pipette contained 70 mM Sr²⁺. The vertical scale bar = 2 pA; the horizontal bar = 100 ms. The holding potential was −70 mV and the step potentials are indicated. Note that the ensemble averaged currents derived from ∼150 traces showed a slow decay.

FIG. 5.

Probability (P_o) of L-type channel openings. The P_o was determined for Sr²⁺ currents using records from ∼200 400-ms consecutive sweeps at different step potentials using 5 (•) and 70 mM (○) in the patch pipettes. Note that the maximum P_o was <0.5. The solid curves are fits to a Boltzmann equation (also see Fig. 6).

FIG. 6.

Changes from high to low Sr²⁺ concentrations A: average P_o of individual single channel fluctuations recorded using 70 mM Sr²⁺ in the patch pipette plotted as a function of time and voltage. B: similar data were collected in 5 mM Sr²⁺. The surface plots show distinct modal gating in the L-type channel at different Sr²⁺ concentrations. C: shift in activation curves with changes in concentrations of Sr²⁺. Normalized nP_o values (n = number of channels in a given patch) at different test potentials and Sr²⁺ concentrations [5 (•) and 70 mM (○)]. The solid lines represent fits to the Boltzmann equation P_o = 1/[1+(e^{(V − V}_1/2⁾)/k], where V_1/2 is the half-activation potential and k is the slope of the e-fold change. The V_1/2s (in mV) were −24 ± 1.2 and −13.7 ± 1.5 (n = 7) and ks (in mV) were 4.3 ± 0.7 and 4.4 ± 0.5 for 5 and 70 mM, respectively. D: the shift in V_1/2 as function of Sr²⁺ concentration (□). The solid line represents a linear regression through the experimental data, and the dotted line is data from Ca²⁺ currents (Rodriguez-Contreras and Yamoah 2003).

If indeed the gating of the channel is altered by the permeant ion concentrations, and that at saturating levels the transition rate constants are altered, we would expect the waiting time to first opening or the first latency distribution to change in recordings performed at the two different spectrums of the sensitivity range of the channel. Figure 7 shows the normalized cumulative first latency distribution at 70 (A) and 5 mM Sr²⁺ (B) in the patch pipette solution. In accord with a voltage-dependent channel, the time constant of the first latency (τ_f) was faster at depolarized than at hyperpolarized voltage steps. However, in keeping with the notion that the transition between closed states may be ion concentration dependent, τ_fs were starkly faster in recordings performed in pipette Sr²⁺ concentrations that lay within the most sensitive phase of the channel's conductance–concentration relations (Fig. 3) than at saturating concentrations. Figure 7C compares the τ_fs of Ba²⁺, Ca²⁺, and Sr²⁺ to assess the divalent cations that may promote multiple closed states or reluctance to transition from the closed states to first openings.

FIG. 7.

Changes in first latency distribution at different concentrations of Sr²⁺ A: cummulative 1st latency distribution (solid lines) at different step potentials indicated. The data were collected using 70 mM Sr²⁺ as the permeant ion. The dotted lines represent exponential fits to the distribution. As expected for a voltage-dependent current, the time constants of the 1st latency (F) were brief at depolarized potentials. B: similar data collected at low concentrations of the permeant ion (Sr²⁺ at 5 mM). The inset compares the time constants at 5 and 70 mM after liquid junction potential and surface charge screening effects have been considered. C: comparisons of the data obtained in Sr²⁺ (solid line) to 2 other divalent cations Ba²⁺ and Ca²⁺ (dotted lines) (Rodriguez-Contreras and Yamoah 2003).

Another direct way to determine the effects of charge carrier concentration on the kinetics of Sr²⁺ currents is to obtain the open and closed time distributions from the single-channel records. Figure 8 shows open and closed time distributions at Sr²⁺ concentrations of 70 (Fig. 8, A and B) and 5 mM (Fig. 8, C and D). At 70 mM Sr²⁺, channel gating could be described by two open and two closed states, whereas at 5 mM Sr²⁺ (Fig. 8, C and D), an additional open state was included in the fit of open time distributions. As shown, patch depolarization increased the duration of the open states, whereas it had the opposite effect on the duration of the closed states. In addition, the time constant of longer duration in both closed and open time distributions was more frequent than the time constant of brief duration as judged from the relative areas of each component. We also show the relation between open and closed state duration and the membrane potential at 70 (Fig. 9, A and B) and 5 mM Sr²⁺ (Fig. 9, C and D). Comparison of open and closed times suggests that, at 5 mM, when the channel is at its sensitivity range, not only does the dwell time of the open state increase, but also the dwell time of the closed state plummets. It is the opposite case for recordings at saturating concentrations of Sr²⁺ (70 mM).

FIG. 8.

Open and closed time distributions. Exponential distributions of open and closed time intervals were observed at all membrane potentials. Dwell time histograms were binned logarithmically (time in log scale) using 10 bins per decade and plotted with a square root transformation of the number of events. Data at 70 mM Sr²⁺ (A and B) were described by 2 open and 2 closed states. As shown in the insets of each panel, the duration of open states increased, whereas the durations of closed states decreased with membrane depolarization. The areas of each component, which represent approximately the time spent in the respective states, are shown in parentheses. At 5 mM (C and D), open and closed times were best described by 2 states at most potentials but at some potentials the open time distributions were best fit with 3 time constants.

FIG. 9.

Open and closed times are voltage dependent. Time constants of open and closed states were plotted at different membrane potentials for Sr²⁺ current recordings at 70 (A and B) and 5 mM (C and D). At 70 mM, brief open states and long closed states dominated. However, at 5 mM, long open times and brief closed times were the norm. A 3rd time constant is shown for open times (C) to show another component that was detected at more depolarized potentials (> −20 mV).

DISCUSSION

Here, we report on the first extensive single-channel analysis of Sr²⁺ currents in native hair cells. These recordings showed elementary properties that are usually impossible to identify from reports that use saturating concentrations of divalent cations to study Ca²⁺ channels. By determining the sensitivity ranges of the channels and by using saturating and nonsaturating concentrations of Sr²⁺, nuances of the permeation and gating characteristics of the channels that are prone to both permeant ion concentration and surface charge screening effects, are shown. First, we showed that macroscopic Sr²⁺ currents essentially mirror the profile of the current of the physiological charge carrier Ca²⁺ in magnitude and activation. The inactivation of whole cell Sr²⁺ currents is much more similar to Ca²⁺, although subtle differences are apparent, which is in keeping with reports that Sr²⁺ can substitute for Ca²⁺ in several biological processes (Abdul-Ghani et al. 1996; Xu-Friedman and Regehr 1999). Second, these findings show that, for Ca²⁺ channel subtypes, such as the Ca_v1.3 channels in hair cells that are subjected to Ca²⁺-dependent inactivation, mediated by increase in intracellular Ca²⁺ within a few angstroms of the channel via calmodulin binding (Peterson et al. 2000; Rodriguez-Contreras and Yamoah 2003; Shen et al. 2006; Yang et al. 2006), outcomes of using high concentrations of the permeant ion should be interpreted with the following caveats. Specifically, analyses of the open probability, time constants of openings and closures, and the kinetics of first latency of openings can be influenced considerably by the concentration of the permeant ion. Additionally, these findings should be substantiated with the underpinnings of the key features of the handling of the divalent cation used in the study (Carter et al. 2002; Xu-Friedman and Regehr 1999). Third, increased P_o, number of open states, and reduced dwell times of closed states of the channels in the physiologically important 1- to 5-mM range promotes the increase in whole cell current amplitude and duration. Thus in the functional context of the channels as biological Ca²⁺ transporters, the cellular physiological implications of these findings could be vast.

By virtue of its similar charge and size, Sr²⁺ substitutes well for Ca²⁺ in mediating synaptic transmission at the squid giant synapse (Augustine and Eckert 1984), neuromuscular junction (Bain and Quastel 1992; Dodge et al. 1969), and synapses in the brain (Tokunaga et al. 2004; Xu-Friedman and Regehr 1999), albeit at different kinetics. Indeed, Sr²⁺ can maintain mechanoelectrical transduction (Ohmori 1985) and activate Ca²⁺-activated K⁺ channels in hair cells (Sugihara 1998) and in other systems (Oberhauser et al. 1988). However, because intracellular Ca²⁺ buffers are less efficient in chelating Sr²⁺ than Ca²⁺ (Yawo 1999), replacing pipette Ca²⁺ with Sr²⁺ was predicted to reduce the influence of Ca²⁺-dependent modifications and intracellular Ca²⁺ buffers. Thus we surmised that Sr²⁺ entry through Ca²⁺ channels would be more suitable to examine the elementary properties of voltage-gated Ca²⁺ channels in hair cells than Ca²⁺ flux itself, which is confounded by Ca²⁺-dependent buffering and signaling.

Previous analyses of the concentration dependence of unitary current and conductance to determine the K_D of Ca²⁺ channels have shown that the concentration versus conductance curves provide a more reliable measure since they are independent of surface charge screening effects and voltage errors introduced by liquid junction potentials (Church and Stanley 1996; Rodriguez-Contreras and Yamoah 2003; Rodriguez-Contreras et al. 2002). From the calculated maximum conductances and the K_Ds of the two Ca²⁺ channel subtypes, at the equivalent physiological Ca²⁺ concentration of ∼2 mM, the conductance (in pS) of Sr²⁺ for the L- and non-L-type channels were 1.5 and 4.0, respectively. These conductance values dovetail well with extrapolated data from concentration dependence curves of Ca²⁺, which were ∼1.0 and 2.7 pS for the L- and non-L-type channels (Rodriguez-Contreras et al. 2002). Indeed, direct measurement of the physiological Ca²⁺ conductance of the L-type channel yielded ∼1.2 pS (Rodriguez-Contreras et al. 2002), which is within the range of these results. Thus at a membrane potential of hair cells (e.g., −65 mV), and a reversal potential for Sr²⁺/Ca²⁺ of ∼100 mV, the limits of the flux rate of Ca²⁺ ions ranges from 500 to 1,400 ions/ms and that of Sr²⁺ ions ranges from 800 to 2,000 ion/ms. For clustered channels in hair cells, these unitary Ca²⁺/Sr²⁺ domains are expected to increase by ≥100-fold (Roberts et al. 1990; Rodriguez-Contreras and Yamoah 2001). Another essential feature of the concentration dependence curves is the difference in the steepness at 1–5 mM (Sr²⁺: 0.36 pS/mM; Ca²⁺; 0.34 pS/mM) and 10–70 mM (Sr²⁺: 0.08 pS/mM; Ca²⁺; 0.04 pS/mM), suggesting that the channels are ∼8-fold and 4-fold more sensitive at physiological concentration ranges of Ca²⁺ and Sr²⁺, respectively. By retaining high selectivity and sensitivity, an optimal throughput of divalent cations is ensured at physiological concentration ranges.

A question that arises from this study is how can one evaluate the respective roles of subsaturating and saturating concentrations of permeant ions on 1) permeation, as it relates to means by which the permeating ions interact with binding sites within the pore of the channel (Kuo and Bean 1993); 2) allosteric effects of binding of divalent and/or monovalent ions to sites of the channel that mediate changes in protein conformational changes that produce alterations in gating mechanism (Zamponi and Snutch 1996); and 3) last, electrostatic actions of different concentrations of divalent ions on the membrane electric field (Hille 2001; Zhou and Jones 1995). Our Sr²⁺ current amplitudes and conductances (γ) at saturating concentrations are consistent with those reported previously for Ca_v1.2, 1.3, 2.2, and 2.3 channels, which follow the sequence (γ_Ba > γ_Sr > γ_Ca) (Fox et al. 1987; Rodriguez-Contreras and Yamoah 2003). Moreover, comparison of the conductance obtained at subsaturating concentrations of divalent ions was within the range of prior observations (Rodriguez-Contreras and Yamoah 2003). In particular, the overall sequence in unitary conductances at subsaturating concentrations (5 mM) was distinct from data obtained at saturating concentrations (70 mM) for both channels (at 5 mM L-type: γ_Ba = γ_Sr > γ_Ca; non-L-type: γ_Sr > γ_Ba = γ_Ca). Of note, the results cannot be accounted for by the shifts in the activation curves from high to low concentrations of the permeant ions.

Although it is difficult to disentangle the effects of surface charge screening from allosteric effects, it has been shown that the binding of Ca²⁺ to Ca_v2.1 channels produces functional allosteric mechanisms that alter channel gating (Zamponi and Snutch 1996). In this study, we showed that changing the concentration of Sr²⁺ from saturating to subsaturating levels profoundly alters the gating of the Ca_v1.3 channels. In particular, at subsaturating levels, the Ca²⁺ channel exhibits increased long openings and a reduced closed time. Because both charge screening and selective binding of Sr²⁺ (Hille 2001) to the channel cannot account for these findings, we suggest that an allosteric interaction of Sr²⁺ with the L-type channel in hair cells may mediate a state-dependent alterations of a closed state (Gilly and Armstrong 1982a,b), which is in keeping with the ensuing alterations of the first latency distributions (Fig. 7). Specifically, the time constant of the first latency distribution was shorter at subsaturating concentrations of Sr²⁺ than at saturating levels.

The physiological interest of this study lies in the increased realization that Ca²⁺ channels in hair cells may exhibit strong Ca²⁺-dependent inactivation (Rodriguez-Contreras and Yamoah 2003; Shen et al. 2006; Song et al. 2003; Yang et al. 2006) that may vary during different stages of development and regeneration (Levic et al. 2007; Yang et al. 2006) against the backdrop of previous reports that hair cells use an efficient Ca²⁺-buffering mechanism (Roberts 1993) that may mask the genuine phenotype of Ca²⁺ currents. It is also clear that the Ca²⁺ channels are extremely sensitive at subsaturating concentrations of the permeant divalent ions, and we can infer that Ca²⁺-dependent inactivation might actually have dramatic effects on Ca²⁺ currents at physiological Ca²⁺ concentrations at the resting potential because the channels have a sizable P_o at baseline. Analysis of the Ca²⁺ handling at the subnanometer Ca²⁺ domains of the Ca²⁺ channel in hair cells will show further details of the basic properties of the channel as a physiological Ca²⁺ transporter.