PKCγ-deficient mice (Abeliovich et al., 1993a) were provided by Dr. Masanobu Kano (University of Tokyo, Japan). The PKCγ^fl/fl mice were generated in our previous study (Watanave et al., 2022). All mice used in this study were maintained on a C57BL/6J genetic background in our breeding colony at the Institute of Experimental Animal Research, Gunma University Graduate of Medicine, Gunma, Japan. Homozygous PKCγ-deficient and PKCγ^fl/fl mice were obtained by crossing respective fertile heterozygous animals and were genotyped by PCR. RGS8-EGFP mice (Kaneko et al., 2018) were crossed with mice from the PKCγ-deficient or PKCγ^fl/fl mouse lines to obtain mice carrying both genotypes, which was confirmed by PCR. All procedures regarding the care and treatment of animals were carried out according to the Japanese Act on the Welfare and Management of Animals, and the experimental protocol was approved by the Institutional Committee of Gunma University (approval numbers 23-018 and 21-065).

We used the expression plasmid (pAAV-L7-6-minCMV-PKCγ-mCherry-WPRE, pAAV-L7-6-minCMV-GFP-P2A-Cre-WPRE or pAAV-L7-6-minCMV-mCherry-P2A-Cre-WPRE) for AAV9 or AAV PHP.eB (Chan et al., 2017) vector production. The AAV vectors were designed to express GFP (or mCherry) and PKCγ (or Cre) under the control of L7 promotor (Sawada et al., 2010; Nitta et al., 2017) with minCMV (Matsuzaki et al., 2014). The woodchuck hepatitis virus posttranscriptional regulatory element (WPRE) sequence was inserted following PKCγ (or Cre).

Recombinant single-strand AAV vectors were produced by transfection of HEK293T cells (Thermo Fisher Scientific, Waltham, MA, USA) with pAAV/L7-minCMV- PKCγ-mCherry-WPRE (or pAAV-L7-6-GFP-P2A-Cre-WPRE or pAAV-L7-6-mCherry-P2A-Cre-WPRE), pAAV2/9 (provided by Dr. J. Wilson; or pAAV-PHP.eB), and a helper plasmid (Stratagene, La Jolla, CA, USA) using an ultracentrifuge purification method described previously (Konno and Hirai, 2020). The genomic titers of the purified AAV vectors were determined by quantitative real-time PCR using Power SYBR Green Master Mix (Thermo Fisher), according to the manufacturer’s instructions, with the primers 5’-CTGTTGGGCACTGACAATTC-3’ and 5’-GAAGGGACGTAGCAGAAGGA-3’ targeting the WPRE sequence. Expression plasmid vectors were used as standards.

To eliminate PKCγ expression from mature PKCγ^fl/fl mouse PCs, AAV9 vectors expressing Cre recombinase together with GFP or mCherry under the PC-specific L7-6 promoter [10 μl, 1.0 × 10¹¹ vector genomes (vg) /ml] were injected into the cerebellum of 3–5-weeks old PKCγ^fl/fl mice. For the PKCγ rescue experiment, intravenous injection of the PC-targeting AAV PHP.eB vectors expressing mCherry PKCγ (100 μl, 5.0 × 10¹² vg/ml) was administered to 3-week-old systemic PKCγ-null mice [postnatal day (P)21–P25].

For direct cerebellar injection of AAV9 vectors, mice were deeply anesthetized with a combination of ketamine (100 mg/kg body weight) and xylazine (10 mg/kg body weight), and placed in a stereotactic frame. The skin covering the occipital bone was cut, and a burr hole was created 2 mm caudal from the lambda. The tip of a Hamilton syringe (33 gauge) (Hamilton Company, Reno, ND, USA) with an attached micropump (UltraMicroPump II; World Precision Instrument, Sarasota, FL, USA) was inserted 1.6 mm below the pia mater of the cerebellar vermis. AAV9 vector suspension (10 μl) was injected at a rate of 300 nl/min using a microprocessor-based controller (Micro4; World Precision Instrument). The syringe was left in place for 2 min following the injection. After closing the scalp, the mice were returned to standard home cages.

For intravenous injection, 100 μl of the PHP.eB vector suspension was injected into the retro-orbital sinus of deeply anesthetized mice using a 0.5-mL syringe with a 30-gauge needle (08277; Nipro, Osaka, Japan).

Parasagittal cerebellar slices (250 μm in thickness) from P49–P69 mice were prepared as previously described (Watanave et al., 2018, 2019). The slices were perfused in an extracellular solution containing (in mM): 125 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 26 NaHCO₃, 2 CaCl₂, 1 MgCl₂, 10 glucose, and 0.1 picrotoxin, bubbled continuously with a mixture of 95% O₂ and 5% CO₂ at room temperature during the recordings. For some experiments, the extracellular solutions contained 20 μM ZD7288 to block HCN channels. PCs were visualized using a 40× water-immersion objective attached to an upright microscope (Axioskop; Carl Zeiss). The resistance of the patch pipette was 3–6 MΩ when filled with an intracellular solution containing (in mM): 122.5 K-gluconate, 17.5 KCl, 8 NaCl, 2 Mg ATP, 0.3 NaGTP, 10 HEPES, and 0.2 EGTA (pH 7.2, adjusted with KOH, 290-310 Osm). All whole-cell data was obtained using an EPC8 amplifier (HEKA Electronik), controlled with pClamp 10 software (Molecular Device), and analyzed using Clampfit software (Molecular Device). No bridge balance compensation was performed. We discarded the data if the holding current when the cell was held at −70 mV was smaller than −500 pA. The intrinsic properties of PCs were recorded in the current–clamp mode. Negative currents (<−500 pA) were injected to adjust the baseline to approximately −70 mV, and various sizes of depolarizing current steps (500 ms; 50 to 500 pA, in 50 pA increments) were applied to evoke action potentials. The spikes obtained from each current injection were counted. The lowest injected currents which generated action potential were corrected as rheobase currents. The thresholds of the action potentials were obtained at the time point where dV/dt = 30. Input resistances were simultaneously obtained in each recording by applying negative currents (500 ms;−50 pA) to PCs. Inter spike interval (ISI) and adaptation index were obtained from the action potential traces which showed 10–15 action potentials. Adaptation indexes were calculated by dividing 1st ISI by last ISI of each trace. Sag voltage and rebound depolarization were recorded by injecting negative step currents (500 ms; −50 to −250 pA, in 50 pA increments). Sag voltage was calculated by subtracting mean potential at 450–500 ms from the negative peak potentials. Rebound depolarization was measured by subtracting the peak potential of the positive direction at 0–500 ms (after the negative current injection) from baseline potentials. Liquid junction potential was not corrected in this study.

The mice 4 weeks after the viral injections were deeply anesthetized and transcardially perfused with PBS and 4% PFA (in 0.1 M phosphate buffer), and 50-μm-thick cerebellar slices were obtained. The slices were treated with the following primary antibodies: rat monoclonal anti-GFP (1:1,000; Cat. No. 04404-84, Nacalai Tesque, Kyoto, Japan), rabbit monoclonal anti-PKCγ (1:1,000; Cat. No. AB_2571824; Nittobo Medical, Tokyo, Japan), rabbit polyclonal anti-calbindin 28K (1:500; Cat. No. C2724; Sigma-Aldrich St. Louis, MO, USA), goat polyclonal anti Aldolase-C (1:200; Cat. No. AldolaseC-Go-Af800; Nittobo Medical, Tokyo, Japan) and the following secondary antibodies: Alexa Fluor 488 donkey anti-rat IgG (1:1,000; Thermo Fisher Scientific), Alexa Fluor Plus 488 donkey anti-goat IgG (1:2,000; Thermo Fisher Scientific), Alexa Fluor Plus 555 donkey anti-rabbit IgG (1:2,000; Thermo Fisher Scientific), Alexa Fluor 568 donkey anti-rabbit IgG (1:1,000; Thermo Fisher Scientific). Fluorescence images were acquired using a fluorescence microscope (BZ-X800, Keyence, Osaka, Japan).

The sample size of the recording was determined as referring to resembling studies. After we had checked the normality of the data, significant differences were analyzed using Welch’s t-test, one-way or two-way repeated measure analysis of variance (ANOVA) followed by Bonferroni’s post-hoc test. Statistical analyses were performed using GraphPad Prism 9 (GraphPad Software, San Diego, CA, USA). Data are expressed as the mean ± SEM. The detailed statistics of two-way ANOVA, including F value, degree of free and individual post-hoc p-values are compiled in Supplementary Table 1.

Analysis of conventional PKCγ-deficient mice have suggested that PKCγ plays a critical role in brain development (Chen et al., 1995; Kano et al., 1995). Systemic PKCγ-knockout (PKCγ-KO) mice show motor deficits, which may be due to the developmental defects. Therefore, roles of PKCγ in adult mouse PCs have not been fully clarified yet. To assess the physiological significance of PKCγ in mature mouse PCs, we used PKCγ^fl/fl mice that had lost PKCγ expression in a Cre recombinase-dependent manner (Watanave et al., 2022). At 3 weeks of age, the PKCγ^fl/fl mice and their wild-type (WT) littermates were administered a cerebellar injection of AAV9 vectors expressing Cre and GFP under the control of a PC-specific L7-6 promoter (Nitta et al., 2017; Figure 1A). Immunohistochemistry 4 weeks after the AAV injection confirmed conditional knockout of PKCγ specifically in GFP-labeled (Cre-expressing) PKCγ^fl/fl mouse PCs (PKCγ-cKO PCs, hereafter), in contrast with the clear PKCγ immunolabeling in GFP-negative PCs (Figure 1B). To clarify the effect of PKCγ deletion, we examined electrophysiological properties of (GFP-expressing) PKCγ-cKO PCs using acute cerebellar slices of cerebellar vermis. We first recorded the intrinsic excitability of PKCγ-cKO PCs randomly selected from the cerebellar vermis in the current clamp mode. Negative currents were injected to adjust holding potentials to −70 mV, and the spikes were evoked by current injections ranging from 100 to 500 pA in 100 pA increments. PKCγ-cKO PCs showed higher intrinsic excitability than WT mouse PCs (Figure 1C; WT: n = 34 from eight mice, PKCγ-cKO: n = 33 from seven mice, p < 0.01 at 200 pA, p < 0.05 at 300 pA by Bonferroni’s post-hoc test following 2-way ANOVA). Also, to obtain more detailed intrinsic properties, we have analyzed input resistances (Ri), action potential threshold, and the rheobase currents of those data. Although a significant difference was not detected in Ri and the thresholds between WT and cKO PCs, the rheobase currents were significantly lower in cKO PCs, which could reflect higher intrinsic excitability in cKO PCs (Supplementary Figure 1). These results suggested that PKCγ in mature mouse PCs suppresses the intrinsic excitability of PCs.

Although we randomly chose PCs for electrophysiological analysis in Figure 1, many PCs examined were from middle lobules (IV–VII), which contained both Z− and Z+ PCs, since large portion of GFP-expressing PCs were observed in lobules IV–VII (Figure 1B). To clarify the association of zebrin expression profiles with intrinsic excitability, we recorded the intrinsic excitability of PCs, focusing on lobules I–III and IX–X, containing mostly Z− and Z+ PCs, respectively (Figure 2A). PKCγ-cKO mouse PCs in lobules I–III showed enhanced intrinsic excitability compared with WT mouse PCs (Figure 2B; WT: n = 7 from three mice, PKCγ-cKO: n = 10 from four mice, p < 0.01 at 200 pA, p < 0.05 at 300 pA by Bonferroni’s post-hoc test following 2-way ANOVA). In contrast, the intrinsic excitability of PCs in lobules IX–X did not differ significantly between genotypes (Figure 2C; WT: n = 9 from four mice, PKCγ-cKO: n = 7 from five mice, p = 0.515 by 2-way ANOVA). Other intrinsic properties are shown in Supplementary Figure 2. Although it is not significant, Ri tends to be lower in cKO PCs than in WT PCs at lobules I–III (Supplementary Figure 2A). The thresholds were comparable between the genotypes at lobules I–III (Supplementary Figure 2B). The rheobase current was significantly lower in cKO PCs than in WT PCs at lobules I–III (Supplementary Figure 2C). No significant differences between WT and cKO PCs were observed at lobules IX–X (Supplementary Figures 2D–F). These results suggested that PKCγ suppresses the intrinsic excitability selectively in lobules I–III.

To investigate whether PKCγ modulates the intrinsic firing patterns, we have analyzed firing patterns of PCs present in lobules I–III. Both WT and PKCγ-cKO PCs exhibited 2 firing patterns, tonic firing and initial burst firing (Supplementary Figure 3A). The ratio of each firing type was almost comparable between WT and PKCγ-cKO PCs (Supplementary Figure 3B), suggesting that PKCγ does not affect the firing pattern at least in PCs of lobules I–III.

To prove that the PKCγ-mediated modulation of intrinsic excitability is present only in Z− PCs, we recorded the intrinsic excitability of Z− and Z+ PCs in the same lobule. To visualize the zebrin expression in PCs, we used RGS8-EGFP mice, that robustly express GFP in Z− PCs (Viet et al., 2022). PKCγ^fl/fl mice were crossed with RGS8-EGFP mice to obtain PKCγ^fl/fl/RGS-EGFP mice. In the following experiments, we used PKCγ^fl/fl mice and their WT littermates on an RGS-EGFP mouse background to differentiate Z− and Z+ PCs. The PKCγ^fl/fl and WT mice (on RGS-EGFP background) were administered cerebellar injections of AAV9 expressing mCherry-P2A-Cre under the PC-specific L7-6 promoter (Figure 3A). The intrinsic excitability was recorded from PCs in lobules IV and V, which contained mixed PCs with strong (Z−) and faint (Z+) GFP fluorescence (Figure 3B and Supplementary Figure 4). Z− PKCγ-deficient PCs (labeled with both GFP and mCherry; PKCγ-cKO) showed significantly enhanced intrinsic excitability compared with Z− WT (PKCγ-expressing) mouse PCs (similarly labeled with both GFP and mCherry) (Figure 3C; WT: n = 7 from three mice, PKCγ-cKO: n = 9 from four mice, p < 0.01 at 300 and 400 pA, p < 0.05 at 200 and 500 pA by Bonferroni’s post-hoc test following 2-way ANOVA), which was confirmed with lower rheobase in cKO PCs (Supplementary Figure 5). In contrast, the intrinsic excitability for Z+ PCs (labeled clearly with mCherry, but only faintly with GFP) did not differ significantly between PKCγ-deficient mice and control WT mice (Figure 3D; WT: n = 11 from five mice, PKCγ-cKO: n = 9 from four mice, p = 0.634 by 2-way ANOVA).

At this point, to evaluate the effect of PKCγ to the timing of the intrinsic firing, we analyzed inter-spike interval (ISI) and spike adaptation of Z− and Z+ PCs. Although the 1st ISI in Z− PCs are comparable between WT and cKO groups (Supplementary Figure 6A), the adaptation ratio tends to be lower in cKO PCs than in WT PCs (Supplementary Figure 6B). In contrast, both 1st ISI and the adaptation ratio in Z+ PCs are comparable between WT and cKO PCs (Supplementary Figures 6C, D). Although the difference does not reach statistically significant level, PKCγ might affect timing of the intrinsic firing.

To confirm that the intrinsic excitability is regulated by PKCγ exclusively in Z− PCs, we conducted a rescue experiment in which PKCγ was virally restored to PKCγ-KO mouse PCs. Briefly, systemic PKCγ-KO mice (on an RGS-EGFP mouse background) were intravenously injected with blood-brain barrier-penetrating AAV-PHP.eB expressing PKCγ-mCherry under the PC-specific L7-6 promoter (Figure 4A). Four weeks after the viral injection, the intrinsic excitability was recorded from Z− and Z+ PCs labeled with mCherry (KO + PKCγ) in lobules IV and V. Compared with WT mouse PCs, systemic PKCγ-KO mice showed enhanced intrinsic excitability in Z− PCs (Figure 4B; WT: n = 8 from three mice, KO: n = 12 from three mice, WT vs. KO: p < 0.05 at 400 and 500 pA by Bonferroni’s post-hoc test following 2-way ANOVA), but not in Z+ PCs (Figure 4C; WT: n=8 from three mice, KO: n=7 from three mice, p = 0.980 by 2-way ANOVA). AAV-mediated PKCγ expression restored the enhanced intrinsic excitability in Z− PKCγ-KO mouse PCs to a level comparable to that in WT mouse PCs (Figure 4B; KO + PKCγ: n = 8 from three mice, KO vs. KO + PKCγ: p < 0.05 at 400 and 500 pA by Bonferroni’s post-hoc test following 2-way ANOVA). In contrast, re-expression of PKCγ in Z+ PKCγ-KO mouse PCs did not exert any significant influence on the intrinsic excitability (Figure 4C; KO + PKCγ: n = 8 from three mice, p = 0.980 by 2-way ANOVA). The other intrinsic properties of PCs are shown in Supplementary Figure 7. These results further suggested that PKCγ negatively regulates the intrinsic excitability exclusively in Z− PCs.

A recent study showed that HCN channel-mediated currents are larger in lobule III PCs (mostly Z− PCs) than in lobule X PCs (mostly Z+ PCs) (Beekhof and Schonewille, 2023). Moreover, HCN currents were shown to be modulated by PKC in cultured mammalian cells (Reetz and Strauss, 2013). Therefore, we examined whether PKCγ could modulate HCN currents, and consequently, contribute to regulation of the intrinsic excitability in Z− PCs.

In the subsequent experiments (Figures 5, 6), the HCN currents were recorded in Z+ and Z− PCs in lobules IV and V from mice on RGS8-EGFP background. WT and PKCγ^fl/fl mice that were treated with AAV expressing mCherry and Cre under the control of the L7-6 promoter (as described in Figure 3A). Z+ and Z− PCs were identified by the intensity of GFP fluorescence. The HCN channels were activated by injecting negative currents for 500 ms (−50 to −250 pA, in 50 pA increments) to mCherry-labeled PCs present in lobules IV and V. Because HCN currents, in response to injection of negative currents, generate sag voltage and rebound depolarization in neurons (Han et al., 2017), we compared these voltage changes (in response to injection of negative currents) in WT and PKCγ-cKO mouse PCs, as illustrated in Figure 5A.

Compared to WT mice, PKCγ-cKO mice showed significantly larger sag voltage for Z− PCs (Figure 5B; WT: n = 11 from six mice, PKCγ-cKO: n = 12 from eight mice, p < 0.05 at −100, −150, and −250 pA by Bonferroni’s post-hoc test following 2-way ANOVA). In contrast, sizes of sag voltage in Z+ PCs were comparable between PKCγ-cKO and WT mice (Figure 5C; WT: n = 11 from eight mice, PKCγ-cKO: n = 12 from seven mice, p = 0.755 by 2-way ANOVA). Rebound depolarization was calculated by subtracting the baseline potential from the peak amplitude of rebound potentiation following injection of the hyperpolarizing current. Because some of the PKCγ-cKO PCs generated action potentials during the rebound potentiation after injection of −200 to −250 pA current (Figure 5D), rebound depolarization amplitudes were measured from traces obtained by injection of −50 to −200 pA current. Z− PKCγ-cKO PCs showed larger rebound potentiation than Z− WT mouse PCs (Figure 5E; WT: n = 11 from seven mice, PKCγ-cKO: n = 11 from seven mice, p < 0.05 at −150 pA, p < 0.01 at −200 pA by Bonferroni’s post-hoc test following 2-way ANOVA), whereas the amplitude in Z+ PCs was almost comparable between PKCγ-cKO and WT mice (Figure 5F; WT: n = 10 from seven mice, PKCγ-cKO: n = 7 from four mice, p = 0.456 by 2-way ANOVA). Thus, conditional deletion of PKCγ significantly enhanced sag voltage and rebound potentiation, namely, the HCN currents in Z− PCs, suggesting that PKCγ negatively modulates the HCN currents exclusively in Z− PCs.

To investigate whether the enhanced HCN currents in PKCγ-cKO PCs in the Z− module underlay the enhanced intrinsic excitability, the HCN channel was blocked by bath-application of 20 μM ZD7288, a blocker of the HCN channels. Both sag voltage and rebound potentiation in Z− PCs of lobules IV and V, which are elicited during and after injection of negative currents, respectively, were almost completely eliminated in the presence of ZD7288 (Figures 6A–C). However, the intrinsic excitability in PKCγ-cKO PCs in the Z− module was not affected by application of ZD7288 (Figure 6D; PKCγ-cKO without ZD7288: n = 9 from three mice, PKCγ-cKO with ZD7288: n = 9 from three mice, p = 0.930 by 2-way ANOVA). The other intrinsic properties of WT and PKCγ-cKO PCs in the presence of ZD7288 are shown in Supplementary Figure 8. These results suggested that enhanced intrinsic excitability in PKCγ-cKO PCs in the Z− module was not due to increase in the HCN currents.

Notably, we found that application of ZD7288 significantly enlarged intrinsic excitability in WT Z− PCs (Figures 6E, F; WT without ZD7288: n = 7 from three mice, WT with ZD7288: n = 8 from three mice, p < 0.05 at 200 and 300 pA by Bonferroni’s post-hoc test following 2-way ANOVA) together with the increase in the Ri and decrease in the rheobase current (Supplementary Figure 9).