Phospholipase C (cid:1) 4 and Protein Kinase C (cid:2) and/or Protein Kinase C (cid:1) I Are Involved in the Induction of Long Term Depression in Cerebellar Purkinje Cells* Activation of the type-1 metabotropic glutamate

] i , intracellular Ca 2 (cid:1) ; PKC, protein kinase C; ACSF, artificial cerebrospinal fluid; EPSC, excitatory postsynaptic current; GABA, (cid:2) -aminobutyric acid; GFAP, glial fibrillary acidic protein; CJ, conjunctive stimulation protocol; DHPG, ( RS )-3,5-dihydroxyphenylgly- cine; TPA, 12- O -tetradecanoylphorbol-13-acetate. Localization PLC (cid:3) Intracellular Ca 2 Elevation—

As predicted, mGluR1-deficient mutant mice exhibit impaired cerebellar LTD (6,7), however, there is no disruption of LTD in PKC␥-deficient mice (8). These results raise the possibility that disruption of one of the intermediate molecules in the mGluR1 signaling pathway may disrupt LTD. Of the four isoforms of PLC, PLC␤1-4 (9, 10), two are abundant in the cerebellum, PLC␤3 and PLC␤4 (11)(12)(13). PLC␤3 is expressed predominantly in the caudal half of cerebellar Purkinje cells, whereas PLC␤4 is distributed throughout cerebellar Purkinje cells. Fly homologue of PLC␤4 has been implicated in transduction of visual information in Drosophila photoreceptors (14,15), however, the role of PLC␤4 in the cerebellum remains unknown. The PLC␤4-deficient mice were viable but had a higher mortality rate than wild-type mice, and the body weight of PLC␤4-deficient mice was generally less than that of wildtype mice in the early stages of postnatal development, as reported previously (16,17). The body weight of the PLC␤4deficient mice gradually increased to match wild-type mice 8 weeks after birth. Using a light microscope, no differences were detected in the size of whole cerebellum, lobe size, or Purkinje cell size between PLC␤4-deficient and wild-type mice. Anatomical alterations are minimal in mGluR1-deficient mutant mice (6) and cerebellar architecture is also normal in glial fibrillary acidic protein (GFAP)-deficient mutant mice (18). Only one abnormality in the cerebellar anatomy of PLC␤4-deficient mice has been reported so far; persistent multiple climbing fiber innervation of Purkinje cells (19), which has also been reported in mGluR1-, GluR␦2-, and PKC␥-deficient but not GFAP-deficient mice (18, 20 -22). Eight PKC isozymes (␣, ␤I, ␤II, ␥, ␦, ⑀, , and ) are expressed in the cerebellum, of which six (␣, ␤I, ␥, ␦, ⑀, and ) are found in cerebellar Purkinje cells (23)(24)(25). Selective expression of a pseudosubstrate PKC inhibitor, PKC inhibitor peptide (Arg 19 -Val 31 ), in Purkinje cells completely blocked cerebellar LTD (26). Therefore, using PLC␤4-deficient mice in the present study, we examined the effects of disruption of PLC␤4 on cerebellar LTD and determined which PKC isozymes were essential for the induction of LTD.

EXPERIMENTAL PROCEDURES
Generation of PLC␤4-deficient Mice-Mice with a disruption of the PLC␤4 gene were generated in the laboratory of M. Katsuki according to standard methods (27). A genomic clone encoding the PLC␤4 catalytic region (denoted the Y region) was isolated to construct a targeting vector in which exons that encode amino acid residues 539 -646 were replaced with a neomycin-resistant gene cassette, and a diphtheria toxin fragment A gene was attached to the 3Ј-end of the targeting vector for negative selection. Embryonic stem cells were transfected with the targeting vector by electroporation and selected with G418 (250 g/ml) for 8 days. G418-resistant colonies were isolated, and the targeted clones were selected using genomic Southern blot analysis with a probe as illustrated in Fig. 1A. Chimeric mice were generated from frozen C57BL/6J blastocytes injected with the embryonic stem cells after warming (28). Male chimeric mice were mated with C57BL/6J female mice. The tail DNA of offspring was analyzed using Southern blot analysis (Fig. 1B) to identify the genotype or amplified using polymerase chain reaction.
Phospholipase C Assay-PLC enzymatic activity was quantified in 200 l of assay mixture containing 150 M PIP 2 . The mixture contained 20,000 cpm [ 3 H]PIP 2 , 1 mM EGTA, 10 mM CaCl 2 , 0.1% sodium deoxycholate, 1 mg/ml bovine serum albumin, and 50 mM HEPES, pH 6.8. The reaction mixture was incubated at 37°C and centrifuged (10,000 ϫ g for 30 min) to precipitate the cerebellar homogenate, and the reaction was terminated as previously described (29).
Western Blot Analysis and Immunohistochemistry-Each lobe of the vermis of the cerebellum from wild-type and PLC␤4-deficient mice was homogenized, and 3 g of protein was separated using 7.5% SDSpolyacrylamide gel electrophoresis. Separated proteins were transferred to a nitrocellulose membrane. The membrane was incubated with anti-PLC␤1, -␤2, -␤3, or -␤4 antibodies (1/1000 dilution, Santa Cruz Biotechnology, Santa Cruz, CA) and then with an alkaline-phosphatase-labeled secondary antibody (1/5000 dilution, Promega, Madison, WI). Immunoreacted bands were visualized using ProtoBlot Western blot AP Systems (Promega).
At the third or fourth postnatal week, mice were deeply anesthetized with pentobarbital (4 mg/100 g) and transcardially perfused with 4% phosphate-buffered paraformaldehyde (4°C, pH 7.4). The brains were immersed in the same fixative for half a day and then embedded in paraffin. Sagittal or coronal paraffin-embedded sections (3-5 m thick) were prepared for immunohistochemical visualization using a streptavidin-peroxidase reaction (Nichirei Co. Ltd., Japan (30,31)). As a blocking step, the sections were incubated with 3% H 2 O 2 in distilled water for 10 min and then 10% normal goat serum for 1 h. Affinity-purified rabbit polyclonal primary antibodies against either mouse PLC␤3 (1/500), PLC␤4 (1/50), PKC␣ (1/100, Life Technologies, Inc., Rockville, MD), PKC␤I (1/500, Life Technologies, Inc.), PKC␤II (1/500, Life Technologies, Inc.), or PKC␥ (1/500, Life Technologies, Inc.) were applied to brain sections overnight at 4°C. Subsequently, sections were incubated with biotin-conjugated goat anti-rabbit immunoglobulin G for 1 h at room temperature (23-26°C). Sections were then incubated with peroxidaseconjugated streptavidin for 1 h at room temperature. Between each incubation step, the sections were rinsed twice in 0.01 M phosphatebuffered saline, pH 7.4, for 5 min each. The final peroxidase reaction was performed using 0.05% diaminobenzidine and 0.005% H 2 O 2 . The same sections were stained with cresyl violet for Nissl staining. For immunohistochemical analysis of PKC isozymes, a fluorescein isothiocyanate-conjugated secondary antibody was used and the immunostained sections were examined using fluorescence microscopy.
Ca 2ϩ Imaging-Sagittal slices (180 -200 m thick) of cerebellar vermis were prepared from 3-to 5-week-old wild-type and PLC␤4-deficient mice using a microslicer (DTK-1000, Dosaka, Japan) and maintained at room temperature in artificial cerebrospinal fluid (ACSF), which consisted of 138.6 mM NaCl, 3.35 mM KCl, 21 mM NaHCO 3 , 0.6 mM NaH 2 PO 4 , 9.9 mM glucose, 2.5 mM CaCl 2 , and 1 mM MgCl 2 and was gassed with a mixture of 95% O 2 and 5% CO 2 (pH 7.4). The Ca 2ϩ indicator fura-2 (1 mM, Dojin, Japan) was injected into Purkinje cells for 25-45 min through patch pipettes or cerebellar slices were incubated in 10 M fura-2 AM (Dojin) for 1 h with 0.001% Cremophore EL. The slices were then maintained in ACSF for at least 30 min and transferred to the stage of an Axioplan 2 microscope (Zeiss, Germany). Fluorescence Ca 2ϩ ratio imaging was carried out by excitation of the indicator at 340:380 nm, and paired emission images were acquired using a cooled charge-coupled device camera (C4880, Hamamatsu Photonics, Japan) at 510 nm. Fluorescence images were acquired using a 60ϫ water immersion objective (LUMPlanFI, numerical aperture (NA) 0.90, Olympus, Japan) that efficiently passed 340-nm light, and ratios were determined using a digital image acquisition system and image-processing software (ARGUS 50/CA, Hamamatsu Photonics, Japan).
Pharmacological Stimulation-The experimental protocols for LTD were performed as described previously with slight modification (32,33). Briefly, sagittal slices (400 m thick) of cerebellar vermis were prepared from 3-to 5-week-old wild-type and PLC␤4-deficient mice using a microslicer and maintained at room temperature in ACSF, including 0.5 M tetrodotoxin and 1 M BAPTA-AM, and saturated with 95% O 2 /5% CO 2 . To stimulate the cells, each slice was then transferred to a cylinder chamber ( 35 mm) in medium containing 50 mM KCl and 100 M glutamate. Five minutes after stimulation, the slices were washed with ACSF for 5 min, followed by fixation with 4% paraformaldehyde. Immunohistochemistry was performed as described under "Western Blot Analysis and Immunohistochemistry." Statistics-Data were analyzed using one-way analysis of variance, and statistical significance was determined using a Student's t test or Mann-Whitney U test. Differences were considered significant when P was less than 0.05.
During the course of the present study, the care of the animals conformed to the guidelines established by the Institutional Animal Investigation Committee at the University of Tokyo.

Biochemical and Histological Characterization of Cerebellar
PLC␤-The total activity of membrane-associated PLC was examined using [ 3 H]PIP 2 as a substrate in cerebellar slices from wild-type (n ϭ 4) and PLC␤4-deficient mice (n ϭ 5). As shown in Fig. 1C, the total PLC activity in PLC␤4-deficient mice was less than 30% of control values in rostral cerebellum and less than 40% in caudal cerebellum. These data suggest that PLC␤4 activity in rostral and caudal cerebellum was 70 and 60% of the total PLC activity, respectively. Total PLC activity was found to be 5.6 nmol/mg/min in rostral cerebellum and 4.5 nmol/mg/min in caudal cerebellum.
Immunohistochemical analysis was performed using an anti-PLC␤4 antibody (Fig. 2, B and C). Each of the lobes in the cerebellar slices was numbered from 1 to 10 as shown in Fig.  2C. PLC␤4 was expressed uniformly in Purkinje cells in rostral (lobes 1-6) and caudal (lobes 7-10) cerebellum from wild-type mice (Fig. 2C), whereas PLC␤3 is more abundant in Purkinje cells in caudal cerebellum from wild-type mice (13, 19; Fig. 2, D-F). No morphological changes were observed in the cerebellum of PLC␤4-deficient mice when examined using light microscopy (data not shown).
Normal PF-Purkinje Cell Synaptic Transmission in PLC␤4deficient Mouse Cerebellum-To examine PF-Purkinje cell syn-aptic function in PLC␤4-deficient mice, we measured the rise and decay time constants of EPSCs, which were calculated using a single-exponential fit (34) and paired-pulse facilitation in acute cerebellar slices. The mean rise time constant was 1.23 Ϯ 0.06 ms (n ϭ 30) and 1.21 Ϯ 0.05 ms (n ϭ 35) in Purkinje cells from wild-type and PLC␤4-deficient mice, respectively. The mean decay time constant was 14.1 Ϯ 0.5 ms (n ϭ 30) in wild-type versus 12.8 Ϯ 0.5 ms (n ϭ 35) in PLC␤4-deficient Purkinje cells. There was no significant difference in either the rise or decay time constants between wild-type and PLC␤4deficient mice (p Ͼ 0.05; Fig. 3, A and B). The PF responses exhibited paired-pulse facilitation (35), which decreased with increasing interpulse intervals in a similar manner in wildtype and PLC␤4-deficient mice (Fig. 3C). Therefore, short term plasticity in PF-Purkinje cell synapses appeared normal in PLC␤4-deficient mice. Furthermore, no significant difference was found in the resting membrane potentials (Ϫ55.5 Ϯ 1. can be replaced by depolarizing Purkinje cells to allow calcium influx through voltage-gated calcium channels (36,37). We recorded PF-EPSCs from Purkinje cells in cerebellar lobes from wild-type and PLC␤4-deficient mice using whole-cell patch clamp and a conjunctive stimulation protocol (CJ) composed of 300 PF stimuli in conjunction with a depolarizing pulse (200 ms, Ϫ60 to ϩ20 mV) repeated at 1 Hz. In 21 of 25 Purkinje cells from wild-type mice (lobes 1-10), CJ stimulation depressed the amplitude of PF-EPSCs, and this depression persisted over 30 min after the onset of the stimulation (Fig. 4A). The mean PF-EPSCs amplitude, measured 25-30 min after CJ stimulation, was reduced to 75.8% Ϯ 3.6% (n ϭ 17 from 13 mice, two cells studied blind) of the original baseline EPSC amplitude. Depression could be induced even after 40 min in whole-cell recording configuration, indicating that cell dialysis had no significant effect on LTD induction. In PLC␤4-deficient mice, Purkinje cells exhibited reduced LTD after CJ stimulation in rostral cerebellum (lobes 1-6; Fig. 4B), whereas LTD was intact in caudal cerebellum (lobes 7-10; Fig. 4C). The mean amplitude of PF-EPSCs in rostral cerebellum recorded 25-30 min after CJ stimulation was 90.1% Ϯ 5.5% of control (n ϭ 16 from 11 mice, two cells studied blind). The difference between the wild-type and PLC␤4-deficient mice was significant (Mann-Whitney U test, p Ͻ 0.05) in rostral cerebellum, whereas LTD from caudal cerebellum in PLC␤4-deficient mice (67.5% Ϯ 2.5%; n ϭ 11 from 10 mice) was comparable to LTD in wild-type mice (Mann-Whitney U test, p Ͼ 0.05).
To exclude the possibility that the lack of Ca 2ϩ release in the mutant mice was an artifact of slice preparation, we examined AMPAR-induced Ca 2ϩ release after DHPG stimulation. Application of AMPA evoked a large Ca 2ϩ transient in Purkinje cells in wild-type cerebellum (Fig. 5K). As shown in Fig. 5 (E and J), large Ca 2ϩ responses were also obtained in Purkinje cells in rostral and caudal cerebellum from PLC␤4-deficient mice following application of AMPA. There was an additional slow phase of the AMPA-induced Ca 2ϩ response in the dendrite (Fig.  5, E, J, and K), which may be due to Ca 2ϩ signals traveling from distal parts of the dendrite. In the soma, however, the two phases overlapped. These results suggest that classic PKC isozymes were not activated in rostral cerebellum from PLC␤4deficient mice.
To investigate possible colocalization of classic PKC isozymes with PLC␤4, we examined the distribution of classic PKC isozymes using antibodies against each isozyme. Immunostaining with antibodies were done as described under "Experimental Procedure." As shown in Fig. 6 (A-D), PKC␣, ␤I, and ␥ were expressed uniformly in Purkinje cells, whereas PKC␤II was not detected. These data are consistent with data obtained previously by several authors (23,25). To investigate the PKC isozymes coupled to PLC␤4 and PLC␤3, we examined the translocation of PKC isozymes during LTD induction using immunohistochemistry. Fluorescence-labeled secondary antibodies were used in this experiment, because fluorescent images showed a relatively large difference between wild-type and PLC␤4-deficient mice with high contrast. 400-m cerebellar slices from wild-type (n ϭ 8 from four mice) and PLC␤4deficient mice (n ϭ 8 from four mice) were incubated for 5 min in ACSF with (n ϭ 4 of each mice) or without (n ϭ 4 of each mice) 100 M glutamate and 50 mM KCl. After stimulation, samples were rinsed for 5 min, followed by fixation. From 10 to 15 sections (5-m thickness) from each slice were stained with antibodies. In wild-type mice, there was strong staining for PKC␣ in the dendrites of Purkinje cells (Fig. 6F), indicating that PKC␣ is translocated. In contrast, no stain was detected in dendrites in PLC␤4-deficient mice (Fig. 6G). PKC␤I immunoreactivity was very strong in Purkinje cell dendrites and soma in all lobes of wild-type mice (Fig. 6I), whereas the fluorescent signal was observed only in cell somas in rostral part of PLC␤4deficient mice (Fig. 6J). No difference in staining for PKC␥, however, was detectable between wild-type and PLC␤4-deficient mice (Fig. 6, L and M).
Unfortunately, we could not determine the coupling selectivity between PLC (␤3 and ␤4) and PKC (␣ and ␤I), because imaging of PKC in caudal part is not clear (data not shown). To overcome this difficulty, a real time imaging of GFP-labeled PKC in living cells under LTD condition is desirable, but it is impossible at present stage. Therefore, we concluded that, at the lowest estimate, both PKC␣ and PKC␤I were translocated during LTD induction, but PKC␥ was not.

DISCUSSION
In the present study, the mGluR1-mediated Ca 2ϩ response and LTD induction was greatly reduced in the rostral cerebellum from PLC␤4-deficient mice, an area in which PLC␤1 and PLC␤3 were also not expressed strongly in these mutant mice. In the caudal cerebellum, however, the residual PLC␤3 activity was sufficient to generate Ca 2ϩ elevation and LTD induction. These results suggest that there was a minimum level of PLC␤3 and PLC␤4 required to generate the mGluR1-mediated Ca 2ϩ response and LTD. We also showed that LTD induction in rostral and caudal cerebellum required activation of classic PKC isozymes. previous reports of expression of the corresponding PLC␤ isoform mRNA (11)(12)(13). In PLC␤4-deficient mice, Although PLC␤1 was expressed in rostral cerebellar Purkinje cells, Purkinje cells in rostral cerebellum from PLC␤4-deficient mice lacked the mGluR1-mediated Ca 2ϩ response. These results indicate that (i) PLC␤1 is not involved in the mGluR1-mediated signaling pathway in cerebellar Purkinje cells and does not have a role in the induction of cerebellar LTD and (ii) mGluR1mediated responses in caudal cerebellar Purkinje cells from PLC␤4-deficient mice were produced by activation of PLC␤3 alone. These results suggest that PLC␤4 is a link between the activation of mGluR1 and the induction of LTD in rostral cerebellar Purkinje cells.

Involvement of PKC Isozymes in the Formation of LTD-
The results of the present study showing that LTD induction was greatly reduced in PLC␤4-deficient mice is consistent with the lack of LTD in cerebellum from mGluR1-deficient mice (6, 7) but does not appear to be consistent with the intact LTD induction observed in PKC␥-deficient mice (8) if PLC␤4 activates PKC␥. Recent evidence using the expression of a PKC inhibitor in Purkinje cells indicates that PKC is required for LTD induction (26). PKC␣, ␤I, and ␥ were expressed strongly and uniformly in cerebellar Purkinje cells, whereas PKC␤II was not expressed in Purkinje cells as shown in Fig. 6 (A-D) (25). PKC␦, ⑀, and were also expressed in cerebellar Purkinje cells (23); however, these isozymes are Ca 2ϩ -independent (for review, see Ref. 38), thus, the contribution of these isozymes to LTD induction is likely to be small. Therefore, the remaining isozymes, PKC␣ and/or PKC␤I, may compensate for the lack of PKC␥ in rostral cerebellum of PKC␥-deficient mice.
In PLC␤4-deficient mice, there did not appear to be any compensation for the lack of PLC␤4 by PLC␤1 in the rostral cerebellum. Thus, evidence suggests that, although compensation for deletion of protein isoforms in the signaling pathway downstream of PLC␤ occurs, there is no compensatory mechanism for the deletion of PLC␤4 itself.
Select PKC Translocation during LTD Induction-Translocation of PKC isozymes after 12-O-tetradecanoylphorbol-13acetate (TPA) stimulation has been clearly observed in several cell systems (40 -44) but not with stimulation sufficient for LTD induction. As described above, the combination of PLC (␤3 and ␤4) activation and translocation of PKC (␣ and ␤I) is very likely. Translocation of PKC␥ has been observed after stimulation used to induce long term potentiation in neurons in the CA1 region of hippocampus (45,46) and TPA stimulation in COS-7 cells (43) but not by LTD-forming conditions in Purkinje cells in the present study. This is consistent with previous results from PKC␥-deficient mice (8). This result further indi- . PKC␣ immunoreactivity appeared in Purkinje cell dendrites and soma in wild-type (ϩ/ϩ) mice 5 min after stimulation (F), whereas only cell somas were stained in PLC␤4-deficient (Ϫ/Ϫ) mice (G). PKC␤I immunoreactivity was observed in Purkinje cell dendrites in wild-type (ϩ/ϩ) mice (I) with LTD stimulation, whereas no immunoreactivity was observed in dendrites in PLC␤4-deficient (Ϫ/Ϫ) mice (J). Fluorescence images of Purkinje cells were not different between wild-type (L) and PLC␤4-deficient mice (M) when using anti-PKC␥ antibody. Bar ϭ 100 m in A-D and 50 m in E-M.
cates that PKC␥ was not activated by the signaling pathway through PLC␤3 and PLC␤4 in cerebellar Purkinje cells.
It has been reported that Purkinje cells in rostral cerebellum from PLC␤4-deficient mice form persistent multiple synapses with climbing fibers (19). This difference may underlie the lack of LTD induction in PLC␤4-deficient mice, however, Chen et al. (8) report that, in PKC␥-deficient mice also, each climbing fiber forms multiple synapses with Purkinje cells and generates multiple spikes that resemble complex spikes, and these mice do exhibit LTD. Thus, the persistent multiple innervation of Purkinje cells by climbing fibers in rostral cerebellum of PLC␤4-deficient mice does not appear to be involved in LTD induction. Moreover, eye blink conditioning is impaired in PLC␤4-deficient mice (47). The results from the present study support the idea that induction of LTD has a role in eye blink conditioning, but the developmental shift from multiple to mono-innervation of Purkinje cells by climbing fibers does not have a role in either LTD induction or eye blink conditioning. These ideas are expressed in Fig. 7 as a molecular linkage of mGluR1-G q -PLC␤4-PKC␣ and/or PKC␤I.
Taken together, the results obtained in the present study provide strong support for the idea that cerebellar LTD involves PKC activation. Further studies are needed to determine if the signaling pathway involves more specific combinations between signaling molecules, such as mGluR1-G q -PLC␤4-PKC␣ or mGluR1-G q -PLC␤3-PKC␤I.