The Ca-activated Cl channel and its control in rat olfactory receptor neurons

Abstract

Odorants activate sensory transduction in olfactory receptor neurons (ORNs) via a cAMP-signaling cascade, which results in the opening of nonselective, cyclic nucleotide-gated (CNG) channels. The consequent Ca2+ influx through CNG channels activates Cl channels, which serve to amplify the transduction signal. We investigate here some general properties of this Ca-activated Cl channel in rat, as well as its functional interplay with the CNG channel, by using inside-out membrane patches excised from ORN dendritic knobs/cilia. At physiological concentrations of external divalent cations, the maximally activated Cl current was approximately 30 times as large as the CNG current. The Cl channels on an excised patch could be activated by Ca2+ flux through the CNG channels opened by cAMP. The magnitude of the Cl current depended on the strength of Ca buffering in the bath solution, suggesting that the CNG and Cl channels were probably not organized as constituents of a local transducisome complex. Likewise, Cl channels and the Na/Ca exchanger, which extrudes Ca2+, appear to be spatially segregated. Based on the theory of buffered Ca2+ diffusion, we determined the Ca2+ diffusion coefficient and calculated that the CNG and Cl channel densities on the membrane were approximately 8 and 62 micro m-2, respectively. These densities, together with the Ca2+ diffusion coefficient, demonstrate that a given Cl channel is activated by Ca2+ originating from multiple CNG channels, thus allowing low-noise amplification of the olfactory receptor current.

Figure 1.

Run-down of the Ca-activated Cl current in excised patches. (A) Immediately after excision, an inside-out patch comprising membrane from the dendritic knob (and often cilia) was exposed for 3 s to saturating concentrations of Ca²⁺ (67 μM) and cAMP (100 μM) consecutively to induce maximal activation of Ca-activated Cl current and CNG current, respectively. The patch was held at −40 mV in symmetrical LiCl solutions. Repeated applications of the double-pulse protocol produced a gradual decline of the Ca-activated Cl current (“run-down”), whereas the CNG current remained constant. The time next to each trace denotes the beginning of each sweep after patch excision. (B) The maximal Ca-activated Cl and CNG currents from A are plotted against time after patch excision. (C) Individual patches showed a different degree of current run-down. Each open symbol represents one patch. On average, the current ratio, I_Cl/I_CNG, decreased from 2.11 ± 0.35 to 1.01 ± 0.23 (12 patches, ▪, mean ± SEM). (D) Run-down of Ca-activated Cl current at both sub-saturating (2.4 μM) and saturating (67 μM) Ca²⁺ concentrations. (E) Similar time courses of Cl current run-down at sub-saturating and saturating Ca²⁺ concentrations. Data from D. (F) The ratio of sub-saturating to saturating currents, I_sub/I_sat, remained constant during run-down (5 patches; ▴ is the patch in D and E), demonstrating that the current run-down was not associated with a quantitative change in sensitivity to Ca²⁺.

Figure 2.

Anion selectivity of Ca-activated Cl current and its sensitivity to other divalent cations. (A) Current-voltage relation from an inside-out patch at 67 μM Ca²⁺ and the effect of replacing bath NaCl with equimolar Na-methanesulfonate. Numbers next to the traces indicate NaCl concentration in mM. Pipette solution contained NaCl. The trace labeled “160” actually represents two hardly distinguishable I-V curves, obtained at the beginning and the end of the experiment, respectively. (B) The dependence of the measured reversal potential (V_Rev) on [Cl⁻]_i (□) shows only a small deviation from the Nernstian prediction (•). Mean ± SD from nine patches. (C) Current-voltage relations from an inside-out patch at 67 μM Ca²⁺ with bath solution containing 140 mM Cl⁻, Br⁻ or I⁻. The trace labeled “Cl⁻” actually comprises two literally identical I-V relations obtained at the beginning and the end of the experiment. The F⁻ permeability was measured separately in an inside-out patch (inset) with 140 mM F⁻ in the pipette solution and 140 mM NaCl solution and 67 μM Ca²⁺ in the bath solution. (D) Relative permeabilities (P_x/P_Cl) plotted against the hydration energy of each halide. Average of 10 (Cl⁻, Br⁻ and I⁻) and 6 (F⁻) patches, respectively. (E) Exposure of an inside-out patch to four different divalent cations at 1 mM for 3 s. (F) Currents activated by divalent cations were normalized to the Ca-activated Cl current. Besides Ca²⁺, only Sr²⁺ effectively activated the Cl⁻ channels. Mean ± SD of six patches.

Figure 3.

Inactivation/desensitization of Ca-activated Cl current. (A) Inactivation/desensitization depended on the presence of Ca²⁺, but not on current flow through Cl channels. The current recorded at −40 mV during a saturating Ca²⁺ step showed the typical inactivation (black trace). Applying the Ca²⁺ step at 0 mV (near the Cl⁻ reversal voltage) induced no current, but inactivation proceeded at the same rate (gray trace). Repeating the step at −40 mV completely restored the original current (“after” trace). This experiment illustrates that inactivation is reversible and not associated with Cl⁻ permeation or Cl⁻ depletion. (B) CNG-gated channels are inactivated by Ca²⁺-CaM. A patch was exposed to 10 μM cAMP and 67 μM Ca²⁺ (and 300 μM niflumic acid to block the Cl current) before and during application of CaM. 1 μM CaM quickly inactivated the cAMP-gated current. Holding potential was −40 mV. (C) Ca²⁺-CaM did not alter the Ca-activated Cl current. Subsaturating Ca²⁺ steps at −40 mV induced currents with similar amplitude and time course in the presence or absence of CaM. Different patch from B.

Figure 4.

Quantitative dependence of Cl current on Ca²⁺. (A and C) Cl current from an inside-out patch was activated by Ca²⁺ at increasing concentration at −40 mV (A) and +40 mV (C) after the rapid phase of the run-down was complete. Maximal currents recorded at the beginning (black traces) and at the end (magenta traces) had similar amplitudes, demonstrating stability of the Cl current (no further run-down) during the experiment. Upon application of Ca²⁺, currents rose to a peak value and, thereafter, declined steadily. (B) Dose–response relations at +40 mV and −40 mV. At each voltage, the relations measured at the initial peak and at 10 s are almost indistinguishable. Data points show the mean ± SD of six patches at −40 mV and seven patches at +40 mV, normalized before averaging. Smooth curves are the Hill equation fitted to the data (see text for values). (D) The ratio of Cl current at 10 s versus transient peak is plotted as a function of the Ca²⁺ concentration at −40 mV (▪) and +40 mV (○). The ∼40% current reduction over 10 s was relatively constant over the applied Ca²⁺ concentration. Numbers of patches are indicated with each mean value.

Figure 5.

Relative magnitudes of Ca-activated Cl current and CNG current in the presence of external divalent cations. Same experiment as in Fig. 1, but with 2 mM Ca²⁺ and 1 mM Mg²⁺ in the pipette solution. (A) An inside-out patch was exposed for 3 s sequentially to 67 μM Ca²⁺ and 100 μM cAMP. Otherwise symmetrical LiCl solutions. Holding potential was −40 mV. The CNG current was strongly suppressed by external Ca²⁺ and Mg²⁺, but the Ca-activated Cl current remained substantial. (B) Run-down of the Ca-activated Cl current persisted in the presence of external divalent cations, but the small CNG current stayed constant. (C) The mean ratio of Cl current to CNG current in the presence of external divalent cations was 33.34 ± 8.97 before Cl current run-down and 22.54 ± 7.43 (16 patches, ▪, mean ± SEM) after run-down. Open symbols represent individual patches, with I_Cl/I_CNG ranging from 5.3 to 124.

Figure 6.

Extraction of unitary CNG and Ca-activated Cl currents from variance analysis. (A) An inside-out patch was exposed for 2 s to 100 μM cAMP in divalent-free symmetrical NaCl solution at −40 mV. The patch showed a slow decline in current after the removal of cAMP, probably because of poor solution exchange in the patch. High-pass filtering at f_c = 2 Hz (trace labeled “hi-pass”) revealed an increased current variance during slow washout of cAMP, reflecting enhanced gating activity of channels at submaximal open probability (P_o). The sinusoidal waveform at the start of the trace is an artifact of the high-pass filtering. (B) A Ca-activated Cl current was recorded at −40 mV, which again was high-pass filtered at 2 Hz. Different patch from A. (C) Plot of variance against amplitude of the cAMP-activated current for data from A. The fitted parabolic function is σ² = iI − I²/N (see text), yielding N = 385 and i = 0.49 pA. (D) Variance analysis for the Ca-activated Cl channel in B. Fit parameters were N = 999 and i = 0.07 pA.

Figure 7.

Functional interaction between Cl channel and CNG channel. (A) Ca²⁺ influx through CNG channels induced a Cl current in an inside-out patch. The pipette solution contained CholCl and 1 mM Ca²⁺, and the holding potential was −40 mV. Application of a saturating (67 μM) Ca²⁺ concentration produced a maximal Ca-activated Cl current (black and magenta traces, representing the first and last traces in the experiment). At low Ca-buffer capacity (0.2 mM HEDTA), 100 μM cAMP caused sufficient Ca²⁺ influx to induce a steady activation of Ca-activated Cl current (green trace). Activation of Cl current to 70% of maximum indicates a steady local Ca²⁺ concentration of ∼2 μM. The Cl current is completely suppressed by 300 μM niflumic acid (dark-blue trace). Increasing the Ca-buffer capacity to 1 mM HEDTA prevented the activation of Cl channels (light-blue trace). (B) Model for the spatial arrangement of Ca-activated Cl channels (green) and CNG channels (red) in the ciliary membrane. The 4:1 ratio of channels reflects the observation that 4-times more active Cl channels can be detected in excised patches after complete run-down. The symbol d denotes the spacing between Cl channels, and r₁ = d/ is the distance from a given Cl channel to the nearest CNG channel. The distances to the 11 nearest CNG channels considered for the local free Ca²⁺ concentration are marked r₁ − r_6. (C) Contribution of increasingly distant CNG channels to the local sensed by a Cl channel. The values r₁ and D_Ca obtained for a given number of channels were used to calculate and expressed as a percentage of calculated for 15 CNG channels. This panel illustrates that 11 CNG channels belong to the catchment area of a Cl channel at 0.2 mM HEDTA, and have to be considered as Ca²⁺ sources in the model. (D) Determination of distance r₁ between Cl channels and CNG channels. For 1 mM HEDTA the graph shows the difference between the experimentally determined and the calculated Ca²⁺ concentrations while for 0.2 mM HEDTA the same function only with opposite sign, the difference between the calculated and the experimentally determined [Ca²⁺]_free is shown. The use of mirrored functions facilitates the accurate determination of D_Ca such that both functions have the same zero-crossover, i.e., intersect exactly at the abscissa (see materials and methods). This was achieved with r₁ = 126.5 nm and D_Ca = 89.5 μm²/s (solid lines), while with even a small deviation of D_Ca from this value the two curves do not have the same zero-crossover (dotted line is the cytosolic D_Ca of 220 μm²/s, Allbritton et al., 1992). The experimentally determined were 2.65 and 1.1 μM at 0.2 and 1 mM HEDTA, respectively (average of eight patches).

Figure 8.

Functional interaction between Cl channel and Na/Ca exchanger. When driven in reverse mode, the Na/Ca exchanger transports Ca²⁺ from the external solution to the cytoplasmic face of the patch membrane, where Ca-dependent Cl channels were activated. The pipette contained CholCl and 2 mM Ca²⁺, the holding potential was +40 mV. The maximal Ca-activated Cl current was determined by applying 67 μM Ca²⁺ (black and magenta traces, representing the first and last traces in the experiment) in CholCl solution. Upon replacing CholCl with NaCl (red trace, Ca-buffer capacity was reduced to 25 μM HEDTA), reverse Na/Ca exchange activated a constant Cl current that corresponded to 60% of the maximal current. This current was blocked by 300 μM niflumic acid (green trace) and did not occur at elevated Ca-buffer capacity (1 mM HEDTA, dark-blue trace). It also depended strictly on Na⁺ and could not be generated by Li⁺ (light-blue trace). A small pedestal-like current was observed when the solution was changed from Chol⁺ to Na⁺ or Li⁺ -containing solution, the origin of which is not fully understood, but might originate from current through CNG channels open even in the absence of cAMP (Kleene, 2000).