KCa1.1 channels contribute to optogenetically driven post-stimulation silencing in cerebellar molecular layer interneurons

Abstract

Using cell-attached recordings from molecular layer interneurons (MLI) of the cerebellar cortex of adult mice expressing channel rhodopsin 2, we show that wide-field optical activation induces an increase in firing rate during illumination and a firing pause when the illumination ends (post-stimulation silencing; PSS). Significant spike rate changes with respect to basal firing rate were observed for optical activations lasting 200 ms and 1 s as well as for 1 s long trains of 10 ms pulses at 50 Hz. For all conditions, the net effect of optical activation on the integrated spike rate is significantly reduced because of PSS. Three lines of evidence indicate that this PSS is due to intrinsic factors. Firstly, PSS is induced when the optical stimulation is restricted to a single MLI using a 405-nm laser delivering a diffraction-limited spot at the focal plane. Secondly, PSS is not affected by block of GABA-A or GABA-B receptors, ruling out synaptic interactions amongst MLIs. Thirdly, PSS is mimicked in whole-cell recording experiments by step depolarizations under current clamp. Activation of Ca-dependent K channels during the spike trains appears as a likely candidate to underlie PSS. Using immunocytochemistry, we find that one such channel type, KCa1.1, is present in the somato-dendritic and axonal compartments of MLIs. In cell-attached recordings, charybdotoxin and iberiotoxin significantly reduce the optically induced PSS, while TRAM-34 does not affect it, suggesting that KCa1.1 channels, but not KCa3.1 channels, contribute to PSS.

Figure 1.

Excitation–inhibition sequence during optical MLI activation. (A1) Upper trace: Loose-seal cell-attached recording from an MLI soma. Lower traces: Raster plots for three consecutive runs at 1 min intervals. Spike rate increases during the 1-s-long light pulse (purple rectangle). After the light pulse there is a period with no spike activity. (A2) Black: PSTH for the three runs. The blue trace, with corresponding y axis on the right, displays the integral of the PSTH. The yellow line is an extrapolation of a linear fit of the blue trace during the pre-stimulus period, where blue and yellow traces superimpose. The distance between blue and yellow traces, representing spike number excess, increases during the light pulse and decreases thereafter. (B1 and B2) Same experiment, but using a 200 ms light pulse, which results in a similar increase in spike rate and a shorter post-stimulus silent period. (C) The biphasic effect of optical stimulation on spike rate was observed for 24 MLIs tested with 1-s-long light pulses. Pre denotes the average spike rate during the period preceding the pulse, Stim corresponds to the spike rate averaged over the 1-s-long pulse, and Post denotes the spike rate during the 1 s subsequent to the light pulse. Wilcoxon paired test P < 1.2e7 for both changes. (D) Pooled data on a subset of experiments in which 200 ms and 1-s-long optical stimulations were performed (three to seven repetitions per cell, alternating durations). Significant increases in the spike rate are observed during the light pulse for both durations (left) as well as a significant decrease in the spike rate compared to pre-stimulus values during the post-stimulation period (right). (E) Spike number excess is significantly larger at the end of the light pulse (labeled “peak”) compared to steady state (SS), both for 0.2 s long pulses (Wilcoxon paired test P = 0.01) and for 1 s long pulses (Wilcoxon paired test P = 0.004).

Figure 2.

Train of optostimulations induces PSS. (A1) Raster plots for the spike activity recorded from an MLI during a 1-s-long train of 10 ms light pulses delivered at 50 Hz. (A2) Corresponding PSTH. (B1 and B2) Raster plots and PSTH from the same recording during 1-s-long light pulses. (C) Pooled data (N = 6) on the ratio of spike rate during the stimulation to basal rates, using either continuous light pulse or a 50 Hz train lasting 1 s. The red line shows a fit of the data by a linear function constrained to passing through the origin, with a slope of 0.85. (D) From the same group of cells, pooled data on the ratio of post-stimulation to basal rates. Slope of linear fit, 1.12.

Figure 3.

Local optical stimulation induces PSS. (A1) Upper trace: Loose-seal cell-attached recording from an MLI soma. Lower traces: Raster plots for three consecutive runs of 200 ms long light pulses confined to the recorded MLI using a 405 nm laser. (A2) PSTH for the three runs. (B) Pooled data on the ratio of spike rate during the stimulation to basal rates obtained with 200-ms-long LED-based light pulses (same data pool as in Fig. 1 D) and 200 ms 405 nm laser-based light pulses (N = 6). (C) Pooled data on the ratio of post-stimulation to basal rates obtained with 200-ms-long LED-based light pulses (same data pool as in Fig. 1 D) and 200 ms 405 nm laser-based light pulses (N = 6).

Figure 4.

GABAergic synapses do not contribute to PSS. (A) Raster plots for the spike activity recorded from an MLI during 1 s optical activation under control conditions (upper panel) and after the addition of 30 μM of the GABA-A R antagonist SR95531 to the bathing solution (lower panel, starting 4 min after the drug was added to the bath). (B) Raster plots from a different MLI subject to 1 s optical activation under control conditions (upper panel) and after the addition of 40 μM of the GABA-B R antagonist CGP 55845 to the bathing solution (lower panel, starting 7 min after the drug was added). (C) Pooled data on the ratio of the spike rate during the 1 s period after the end of the light pulse over the spike rate preceding the light.

Figure 5.

Individual MLI depolarization mimics the optically induced excitation–inhibition sequence. (A1) Sample voltage trace from a tight-seal whole cell recording of an MLI maintained in current-clamp configuration. 18 pA steady current was injected to set the basal firing rate at around 6 Hz. A current pulse (25 pA amplitude, 1 s duration; timing indicated in upper trace) increased the firing rate to 50 Hz. The increase was followed by a silent period upon the end of the stimulation. (A2) Temporal evolution of the peak spike amplitude (upper trace) and of the minimum of the AHP (bottom trace) starting 5 s before the stimulation and ending with the stimulation. The red lines correspond to the mean values for pre-stimulation (peak: −13.8 mV, AHP: −76.1 mV) and during stimulation (peak: −12.2 mV, AHP: −73.6 mV). (A3) PSTH from five repetitions. (B1) Voltage recording from the same MLI, in response to a 1 s long optical stimulation. (B2) Temporal evolution of peak spike amplitude and AHP minimum, as described in A2. Mean maximum and minimum AP values are shown for pre-stimulation (peak: −12 mV, AHP: −75.6 mV) and during stimulation (peak: −12.4 mV, AHP: −73.7 mV). (B3) PSTH from five repetitions. (C) Pooled data on spike rates from five MLIs, before the step of current injection (Pre) during the current step (I step) and during the 1 s following the stimulation (Post; cell injected with the same current as in the Pre period). For each MLI, three to seven repetitions were performed. (D) Ratios for the increase in spike rate during the current step (left) and for the decrease observed after the current step.

Figure 6.

Expression pattern of KCa1.1 channels in MLIs. (A) Confocal immunofluorescence image of a sagittal cerebellar slice from a nNOS-ChR2 BAC mouse stained with an antibody targeting Slo1 KCa1.1 channels. The signal is present at the PC soma as well as throughout the molecular layer. (B–F) Maximum intensity projection images (37 planes at 0.4 um interval) from dual GFP/GCaMP3 and Slo1 staining following the ckit:cre sparse GCaMP3 expression strategy. The green channel displays the GFP/GCaMP3 signals while the red channel displays the Slo1 signals. (B) Superposition of the two channels. (C) Slo1 signal extracted by applying a digital mask from the GFP channel to the Slo1 channel. (D) Zoom of the axon initial segment, indicated by white arrows. (E) Thresholding applied to the Slo1 signal suggests channel clustering, as indicated by white arrows in F. Calibration bars were 20 μm in A, 10 μm in B, C, and E, and 5 μm in D and F.

Figure 7.

KCa1.1 channels are involved in PSS. (A1) Left: Spike activity recorded from an MLI during 1 s optical activation under control condition (upper panel) and 9 min after the addition of 100 nM ChTx (lower panel). (A2) Expanded view for average of spikes detected in the last 200 ms preceding optical stimulation (a and d), in the initial 100 ms of photostimulation (b and e) and in the last 100 ms of photostimulation (c and f), as indicated by arrowheads in A1. The 40-pA calibration bar applies to all traces. (A3) Time course for the ChTx-induced change in the spike rate ratio for the MLI shown in A1. The optical stimulation protocol was performed at intervals of 0.7–1 min. (B) Pooled data from six MLIs on the amplitudes of the depolarized (left) and hyperpolarized (right) components of the spike waveform before optical stimulation (bars a), in the initial 100 ms of the photostimulation (bars b), and in the last 100 ms of the photostimulation (bars c). Data have been normalized to pre-stimulus values in control saline (a). They are shown in control conditions (Ctl; bars a, b, and c) and in the presence of ChTx (bars d, e, and f). (C) Pooled data for the effect of ChTx (N = 6; Wilcoxon paired test P = 0.03), IbTx (N = 6; Wilcoxon paired test P = 0.03), and TRAM-34 (N = 5; Wilcoxon paired test P = 0.62) on the light-induced PSS.