Characterization of PKD2L1 channel activity in CSFcNs from GATA3 mice.

Aa. Confocal image of a sagital spinal cord slice (the dissection plane passes through the CC) from a Gata3eGFP animal showing the distribution of CSFcNs in the anterior-posterior (A-P) direction. Ab, c and d. Confocal image of a coronal spinal cord slice from a Gata3eGFP animal showing the distribution of CSFcNs (b, green) around the CC (dotted line) and the immunoreactivity against the PKD2L1 channel (c, magenta). The overlay of the 2 channels plus the DNA (blue) are shown in d. In b, D is dorsal, V is ventral, L is left and R is right. Brightness and contrast were adjusted for display purposes. Ba. Spontaneous activity of a CSFcNs recorded at -60 mV at 34 °C. The green, horizontal dotted lines represent the different states of the channel: c = closed; o1 = 1 channel open; o2 = 2 channels open; o3 = 3 channels open. Bb. Histogram from the whole-cell current shown in a, with the 4 peaks corresponding to c, o1 and o2 (and o3) indicated with arrowheads. Bc. Boxplot showing the unitary current amplitude measured from 38 different neurons (either somatic, n = 13, or ApPr, n = 25, whole-cell recordings). Middle horizontal line shows the median value (−15.6 pA), upper and lower horizontal lines the 75th and 25th percentiles, respectively, and top and low whisker the 90th and 10th percentiles, respectively. Bd. Boxplot showing the apparent open probability for a single channel measured from 38 different neurons (either somatic, n = 13, or ApPr, n = 25, whole-cell recordings). Middle horizontal line shows the median value (0.04), upper and lower horizontal lines the 75th and 25th percentiles, respectively, and top and low whisker the 90th and 10th percentiles, respectively. Ca. Spontaneous activity recorded at different holding potentials. Cb. IV relationship contructed from recordings performed at -52 (n = 9), -72 (n = 16), -92 (n = 14) and -102 mV (n = 7) holding potentials. Black symbols show averages ± SDs, and gray diamonds show individual values. The dotted line is a linear fit to the average data. From this fit an average unitary conductance of 222 ± 8.0 pS was calculated, with an extrapolated reversal potential of -2.2 mV. The Vm values have been corrected for a calculated liquid junction potential of -12 mV. In B, open diamonds correspond to individual neurons and filled circles to the mean ± SD.

PKD2L1 channel activity mediates both phasic and tonic currents.

Aa. Recording of a CSFcN in control condition and during pressure application of dibucaine hydrochloride (100 µM). Ab. Single channel event frequency calculated in control conditions and during dibucaine application. The average frequency was reduced from 176.9 ± 163 to 5 ± 8.3 Hz (n = 12, p = 5 × 10-4). Ba. Voltage-clamp recording showing how dibucaine application blocks the spontaneous events and reduces the holding current from -23 to -10 pA. The horizontal dotted lines indicate the average baseline current before and during dibucaine application. The bottom trace shows the average current value calculated from 100 ms time-periods, where the reduction in the holding current can be readily appreciated. Bb. Histograms of the recorded current during the control (black) and the dibucaine (orange) time periods. The data has been adjusted with a gaussian function (continuous lines). The mean values of the Gaussian fits correspond to the holding current values shown in Ba (dotted lines; -23 and -10 pA, control and dibucaine, respectively). Bc. Current-clamp recording showing the hyperpolarisation induced by dibucaine application, from -60.4 to -95 mV in this example. a and c correspond to different neurons. Ca. Effect of dibucaine pressure application on the HC: -18.3 ± 9.5 in control to -10.5 ± 6.6 pA in dibucaine (n = 13, = 2 × 10-3). Cb. Effect of dibucaine application on the resting membrane potential: -66.8 ± 11.0 in control to -89.0 ± 14.0 mV in dibucaine (n = 11, p = 1 × 10-3). Cc. Effect of dibucaine pressure application on the HC in the presence of voltage-gated and ionotropic channels blockers: -18.1 ± 8.1 pA in control to -10.8 ± 5.4 pA in dibucaine (n = 11, p = 1 × 10-3). Cd. Effect of dibucaine pressure application on the HC in the presence of 10 mM BAPTA in the internal solution of the recorded neurons: -13.0 ± 5.7 pA in control to -7.8 ± 2.5 pA in dibucaine (n = 11, p = 1 × 10-3). Ce. Input resistance (IR) values calculated from somatic (1.8 ± 0.46 GΩ, n = 14), ApPr (2.3 ± 0.48 GΩ, n = 29) and isolated ApPr (4.4 ± 0.12 GΩ, n = 8) recordings. p = 0.002, soma vs whole-ApPr; p = 6 × 10-6, iApPr vs soma and p = 4 × 10-7, iApPr vs whole-ApPr. In Ab and C, diamonds correspond to individual neurons and circles to the mean ± SD. Statistical comparison between groups was performed with a Wilcoxon signed-rank test for paired data (Ab and Ca to d) and a Wilcoxon–Mann–Whitney for unpaired data (Ce).

pH sensitivity of PKD2L1-mediated phasic and tonic currents.

Aa. Top. Spontaneous PKD2L1 channel activity recorded in a CSFcN at a -60 mV holding potential. A pH 6.5 solution was pressure-applied during 10 s (starting at 3 s, red area). Middle. Holding current calculated from the recording shown on the top. Bottom. Membrane charge calculated from the recording shown on the top. Inset. Segments 7.4 and 6.5 in Aa are shown with expanded time and amplitude scales in order to see the decrease in the spontaneous single channel activity during the application of the acidic solution, without any change in the single channel current. Ab. Mean membrane charge ± SD calculated (gray surface) from 10 neurons tested in the same conditions as the cell shown in a. The dotted line corresponds to a linear fit to the control period (first 3 seconds of the recording). The application of the acidic solution produces a clear decrease in the slope of the membrane charge, indicating a decrease in the spontaneous openings of the channels. B. Same experiment as in a, but the spontaneous activity was recorded in I-clamp. The application of the acidic solution produces a hyperpolarization of the RMP (from -57.2 to -80 mV in this example). Ca. The single channel open time during 500 ms decreased from 46 ± 47 ms in control conditions to 7 ± 7 ms during the pressure application of a pH 6.5 solution (ctrl vs pH 6.5, n = 11, p= 0.001). Cb. Histograms of the nmax values calculated during 500 ms spontaneous, control recordings (black curve) and during the pressure application of a pH 6.5 solution (red curve). The mean nmax decreased from 1.7 ± 0.60 to 1 ± 0 (ctrl vs pH 6.5, n = 11, p = 0.016). A two sample Kolmogorov-Smirnov test also indicated a significant difference between the 2 distributions (p = 0.013). Cc. Effect of a pH 6.5 solution application on the holding current: -16.8 ± 6.7 in control conditions vs. -12.0 ± 5.2 pA in the pH 6.5 solution (n = 10, p = 0.001). Cd. Effect of a pH 6.5 solution application on the resting membrane potential: -64.5 ± 5.2 in control conditions vs. -76.0 ± 3.4 mV in the pH 6.5 solution (n = 8, p = 0.008). Ce. Normalized apparent po1 (apparent po1 test/apparent po1 at pH 7.4) as a function of pH. N = 10 for pH 6.5, 25 for pH 7.4, 8 for pH 8.4 and 7 for pH 10.4. The dotted line shows the fitting of the data to a Hill equation that yields a half pH value of 8.1 ± 0.04. Error bars are SDs. Cf. Normalized holding current (HC test/HC at pH 7.4) as a function of pH. N = 5 for pH 6, 5 for 6.5, 6 for pH 8.4 and 7 for pH 10.4. The dotted line shows the fitting of the data to a Hill equation that yelds a half pH value of 7.5 ± 0.02. Error bars are SDs. In C, diamonds correspond to individual neurons and circles to the mean ± SD. D. Relationship between Vm and HC differences recorded in 28 individual neurons. The dotted gray line is a constrained fit (passing through the origin) with a linear function (ax + b, where b = 0), which shows a slope of 2.3 ± 0.3 GΩ. The 2 variables are strongly correlated (Pearson linear correlation coefficient = -0.84; r2 = 0.71). In C, statistical comparison between groups was performed with a Wilcoxon signed-rank test.

Proton photolysis induces an off-current in CSFcNs.

A. Schematic of the experiment. Pictures showing the eGFP fluorescence (left), the Alexa 594 fluorescence (middle) and the merging of the 2 channels (right). The yellow star indicates the recorded cell, and the objective the location of the targeted compartment. The yellow dotted line shows the approximate boundaries of the central canal. B. Schematic of the photolysis reaction shown for the 2 caged compounds used: MNI-Glutamate (top) and MNI-γLGG (bottom). C. 2 second-long recordings showing the typical response of a CSFcN (shown in A) to the photolysis of MNI-Glutamate on the ApPr. The photolysis (magenta arrowhead; 2 mW, 1 ms duration) was repeated 5 times with 10 seconds intervals. The inset shows sweep #1 in an expanded scale in order to appreciate the fast kinetics of the AMPAR-mediated current. Vertical, magenta arrowhead and dotted line indicate the laser pulse (1 ms, 4.3 mW; which is measured with a photodiode in the laser path). D. Top. The black trace shows the current evoked by the photolysis of MNI-γLGG with a 0.5 ms, 4.3 mW laser pulse on the ApPr, and the gray trace the spontaneous current recording in the same CSFcN. During the spontaneous recording the maximum number of channels opened simultaneously (nmax) was 1, and during the 500 ms window after the photolysis nmax was 3. The total open time in a time window of 500 ms was 17 ms during the spontaneous recording and 196 ms after the photolysis. Bottom. Membrane charge calculated from the above recordings. The vertical, magenta arrowhead indicates the laser pulse. E. Normalized (to the 2 s value) membrane charge as a function of time. Black trace shows the average, gray area the ± SD and dotted traces individual experiments (n = 13). The magenta continuous line represents the fit of the average curve with the sum of a linear + an exponential function representing the increase evoked by the photolysis and the linear increase due to the spontaneous channel openings, respectively (see methods). The τ of the exponential function was 258 ± 2 ms. The x-span of the magenta area represents 500 ms (≈ 2 τs). The vertical, magenta arrowhead indicates the laser pulse. Fa. Histograms of the nmax values calculated during 500 ms time windows for spontaneous recordings (black curve) and after photolysis (magenta curve). The mean nmax was 1.16 ± 0.37 in spontaneous recordings and 2.8 ± 0.63 after photolysis (n = 19, p = 4 × 10-6). A two sample Kolmogorov-Smirnov test also indicated a significant difference between the 2 distributions (p = 1 × 10-6). Fb. Total open time for a single channel during a 500 ms time window (spontaneous vs photolysis). The mean single channel open time increased from 30 ± 36 to 207 ± 51 ms (500 ms spontaneous recording vs 500 ms after the photolysis pulse; n = 19; p = 4 × 10-6). Ga. Example of laser-evoked currents (left, black traces) and the corresponding idealized events (right, green traces). MNI-γLGG was used in this experiment. The blue arrowheads below each idealized trace indicates the timing of double events (where 2 channels opened simultaneously). Vertical, magenta arrowheads and dotted lines indicate the laser pulse (1 ms, 4.3 mW). Gb. Latency distribution of double events (95 photolysis repetitions from 20 neurons) shown with 2 different time resolutions (the graph on the right corresponds to the time indicated by the dotted rectangle on the graph on the left). The rising phase of the histogram has been fitted with an exponential function (green trace) that has a τ of 30 ± 12 ms. The magenta arrowheads indicate the timing of the laser pulse. Ha. The black traces show the responses of a CSFcN to the photolysis of MNI-γLGG on the ApPr with a train of 5 laser pulses (2 mW, 0.5 ms duration) at 180 Hz. The photolysis train was repeated 6 times (#1 to #6) with 10 seconds intervals. The brown curve shows the membrane charge calculated from the average of the 6 sweeps. b. Sweep #2 and its corresponding membrane charge are shown on an expanded time scale in order to better appreciate the onset of the response, that happens after the end of the train. In F, statistical comparison between groups was performed with a Wilcoxon signed-rank test.

PKD2L1 channels are segregated to the ApPr.

Aa. Top. Schematic drawing showing the different photolysis locations. Bottom. Superimposition of pictures showing the eGFP (green) and the Alexa 594 (red, which corresponds to the recorded cell fluorescence. Ab. Representative sweeps showing the membrane currents recorded when the photolysis of MNI-γLGG was performed either on the ApPr (black traces) or on the soma (magenta traces; laser pulse duration = 700 µs; power 5.3 mW). Scale bars are the same for black and magenta traces. Ac. Average membrane charge calculated from recordings corresponding to the photolysis of MNI-γLGG or MNI-glutamate in different locations. Shaded areas correspond to ± SDs. Bottom graph shows the same traces but in a different amplitude scale. The fact that no PKD2L1-mediated activity can be triggered when the photolysis is done a few microns away from the ApPr (5 to 7 µm) is compatible with the high spatial resolution of the photolysis technique and the restricted proton diffusion. Ba. Top. Schematic drawing of experimental design. Whole-cell recordings were performed from iApPrs. Bottom. Pictures showing the different recording configurations: left in cell-attached, where the eGFP green signal can be seen inside the recording pipette as the ApPr membrane goes into the pipette when the positive pressure used for patching is released; midddle in whole-cell a few seconds after break-in, where it can be seen that the green eGFP fluorescence has already washed-out; right in whole-cell, where the morphology of the isolated ApPr can be appreciated by the red Alexa 594 fluorescence. Bb. Whole-cell recordings from the isolated ApPr shown in a at two different holding potentials, -50 and -90 mV. A clear PKD2L1-dependent spontaneous activity can be seen. HP: holding potential. Scale bars are the same for the currents recorded at the two holding potentials. Ca. Top. Schematic drawing of experimental design. The photolysis was produced on an isolated ApPr. Bottom. Pictures showing the different recording configurations. Same as Ba. Cb. Whole-cell recordings from the isolated ApPr shown in a and its response to photolysis. The black traces are in control conditions and the bottom, blue trace in the presence of 100 µM dibucaine (laser pulse duration = 1 ms; power 5.1 mW). Cc. Average membrane charge calculated from 4 individual sweeps in control and 4 individual sweeps in dibucaine. D. Recordings from an ApPr in the whole-cell configuration (a) and in the outside-out configuration (b). The traces on the left show the current responses to a 50 ms depolarization to -10 mV from a -60 mV holding potential. This voltage pulse induces an unclamped sodium spike in whole-cell but no response in outside-out, confirming that the outside-out patch has completely detached from the ApPr (as isolated ApPr do not have sodium currents). PKD2L1 activity is still present in the outside-out recording. Ea. Maximum intensity projection of 41 deconvolved optical sections spaced by 130 nm. CSFcNs expressing eGFP (green, left panel), which also express PKD2L1 receptors (magenta, middle panel.). The overlay of the 2 channels is shown on the right panel. The dotted white line in the left, GFP panel, indicates the distance along which the density plots shown in b were constructed. Eb. Density plots depicting anti-PKD2L1 mean fluorescence intensity (trace) ± SDs (shade) measured at the soma (green) and in the ApPr (black). n = 6 for each compartment. The fluorescence mas measured along the dotted green and black lines shown in the scheme. The maximal mean intensity is 84 for the ApPr and 12.5 for the soma. Brightness and contrast were adjusted for display purposes. Unsaturated images were used for quantification. In A and C, the vertical dotted lines and magenta arrowheads correspond to the timing of the photolysis pulse.

The coupling between the ApPr and the soma measured with paired recordings.

A. Paired current-clamp recordings between the ApPr and the soma. The cartoons on top of the traces show the recording configurations. Aa. The current pulses are applied to the soma, and voltage measured both in the soma and the ApPr (left). The CC (0.94) is calculated by plotting the membrane voltage responses in the ApPr vs the membrane voltage responses in the soma (right). Ab. Same as in a, but now the current pulses are applied to the ApPr. The CC is 0.96. Ac. Spontaneous events recorded simultaneously in the ApPr and the soma. Resting membrane potential was -60 mV in both compartments. Ad. Coupling coefficient measured in 3 conditions: somatic current injection (0.9 ± 0.09), ApPr current injection (0.92 ± 0.09) and for spontaneous activity (soma/ApPr: 0.88 ± 0.05). B. Paired voltage-clamp recordings between the ApPr and the soma. Ba. Voltage pulses (−110 to -20 mV in 10 mV steps) are applied either to the soma (left) or to the ApPr (right). Bb. Currents recovered in the non-pulsed compartments. Bc. The currents shown in b as a function of the voltage steps. The slope of each curve represents the coupling conductance between the 2 compartments (12.85 and 12.4 nS in this example). The average coupling conductance was 13.7 ± 0.62 (pulses applied to the ApPr) and 13.6 ± 0.62 nS (n = 7). All the recordings shown in this figure correspond to the same CSFcN.

Calmidazolium effect on PKD2L1-mediated tonic and phasic currents.

Aa. Top. Spontaneous PKD2L1 channel activity recorded in a CSFcN at a -60 mV holding potential. The recording lasted 35 s, and calmidazolium was pressure-applied at a concentration of 10 µM during 10 s (starting at 3 s, green area). Middle. Membrane charge calculated from the recording on the top. Bottom. Mean membrane charge ± SD calculated from 10 neurons tested in the same conditions as the cell shown in the top panel. The dotted line corresponds to a linear fit to the control period (first 3 seconds of the recording). The application of calmidazolium produces an increase in the slope of the membrane charge. Ab. Holding current from the neuron shown in “a”. Ac. Effect of calmidazolium application on the holding current: -13.3 ± 5.3 pA in control conditions vs -16.0 ± 4.7 pA in calmidazolium (n = 10, p = 0.008). Ba. Spontaneous activity of the CSFcN shown in A, recorded in current-clamp. Calmidazolium application is indicated with the green area. The bottom traces show segments i and ii with different time and amplitude scales in order to see the increase in spontaneous activity during calmidazolium. Bb. Effect of calmidazolium application on the resting membrane potential. Notice the shift from -62.3 ± 3.3 in control conditions vs -50.0 ± 5.2 mV in the presence of calmidazolium (n = 10, p = 0.002). In Ac and Bb, diamonds correspond to individual neurons and circles to the AV ± SD. Cmz: calmidazolium. In Ac and Bb, statistical comparison between groups was performed with a Wilcoxon signed-rank test.

Puffing acidic solutions and proton photolysis in the soma can induce ASIC-mediated currents.

Aa. Top. Spontaneous PKD2L1 channel activity recorded in a CSFcN at a -60 mV holding potential. A pH 6.5 solution was pressure-applied during 4 s (starting at 2 s, magenta area). Bottom. Membrane charge calculated from the recording on the top. The dotted line corresponds to a linear fit during the control period (first 2 seconds of the recording). The application of the acidic solution produces a short-lasting inward current that is probably due to the activation of somatic ASIC channels. This is manifested as a sudden increase in the membrane charge (black arrowhead) that is followed by a subsequent decrease. Ab. Same experiment as in a, under current-clamp. The application of the acidic solution produces a short-lasting depolarization that is probably due to the activation of somatic ASIC channels, followed by hyperpolarization. Vm: membrane potential. B. Spontaneous activity recorded in a CSFcN upon the somatic uncaging of MNI-Glutamate. The black traces show individual repetitions (7 uncagings at 10 sec intervals) and the red trace the corresponding average (AV). Vertical magenta arrowhead and line show the timing of the laser pulse (1 ms, 3 mW). Somatic photolysis does not evoke any PKD2L1-dependent activity, but does produce a small inward current that is probably mediated by ASIC channels.

Effect of alkaline ACSF on PKD2L1-mediated currents.

Aa. Top. Spontaneous PKD2L1 channel activity recorded in a CSFcN at a -60 mV holding potential. A pH 8.4 solution was pressure-applied during 10 s (starting at 3 s, yellow area). Middle. Holding current calculated from the recording shown on the top. Bottom. Membrane charge calculated from the recording shown on the top. B. Chosen segments of the recording shown in A with different scales. C. Same experiment as in A, but the spontaneous activity was recorded in current-clamp. The application of the basic solution produces a depolarization of the RMP (from -64.5 to -47.6 mV in this example). Da. Effect of the pH 8.4 solution on the total open time of a single channel during a 500 ms time window. The mean single channel open time increased from 16 ± 11 (ctrl) to 126 ± 73 ms (pH 8.4; n = 8, p = 0.078). Db. Histograms of the nmax values calculated during 500 ms time windows for spontaneous recordings (black curve) and during the application of a pH 8.4 solution (magenta curve). The mean nmax increased 1.25 ± 0.46 (ctrl) to 2.75 ± 0.70 (pH 8.4; n = 8, p = 0.016). A two sample Kolmogorov-Smirnov test also indicated a significant difference between the 2 distributions (p = 0.011). Dc. The application of a pH 8.4 solution produced a shift of the holding current from -11.7 ± 7.3 to -15.2 ± 7.9 pA (n = 8, p = 0.0078). Dd. A pH 8.4 solution induced a shift of the resting membrane potential from -75.2 ± 7.0 vs -66.2 ± 10 mV (n = 6, p = 0.03). In D, statistical comparison between groups was performed with a Wilcoxon signed-rank test.

Currents induced by proton photolysis in the ApPr depend on PKD2L1 and not ASIC channels

A. Top. Example of laser-evoked currents in control conditions (black trace) and in the presence of 100 µm dibucaine (gray trace). Bottom. Membrane charge calculated from the recordings shown above. The vertical, magenta arrowhead indicates the laser pulse (1 ms, 4.3 mW). B. Average, normalized membrane charge calculated from recordings in control conditions (same data as in 4E) and in the presence of dibucaine. The vertical, magenta arrowhead indicates the laser pulse. C. Single channel current calculated from spontaneous current recordings (spont) and during a 500 ms time window after the photolysis. Gray symbols correspond to individual neurons. Mean ± SD (black symbols): -16.7 ± 2.0 (spont) to 17.0 ± 2.0 (phot) pA, n = 21 neurons, p = 0.56. Wilcoxon signed-rank test. D. Top. Example of laser-evoked currents induced by the photolysis of MNI-Glutamate without NBQX (black trace) and without NBQX + dibucaine (gray trace). The inset shows the current traces in expanded scales to better illustrate the AMPA-mediated current (yellow arrowheads). Bottom. Membrane charge calculated from the above recordings. The vertical, magenta arrowhead indicates the laser pulse (1 ms, 5.2 mW). E. Top. Example of laser-evoked currents induced by the photolysis of MNI-Glutamate without (black trace) and with ASIC channel blockers (magenta traces). Middle. Membrane charge calculated from sweeps number 1 (control) and 4 (ASIC blockers). Bottom. Normalized (to the 2 s value) membrane charge as a function of time in control conditions (black trace, n = 13; same data as in Figure 4E) and in the presence of ASIC blockers (magenta trace, n = 7). The dotted red line represents the fit to the data with the same function as in Figure 4E. The τ value in the presence of ASIC blockers was 190 ms, close to the 258 ms value for the control curve. Shaded areas represent ± SDs. The vertical, magenta arrowhead indicates the laser pulse (500 µs, 5.8 mW). In C, statistical comparison between groups was performed with a Wilcoxon signed-rank test.