Calcium Influx through L-type Channels Generates Protein Kinase M to Induce Burst Firing of Dopamine Cells in the Rat Ventral Tegmental Area*

Enhanced activity of the dopaminergic system originating in the ventral tegmental area is implicated in addictive and psychiatric disorders. Burst firing increases dopamine levels at the synapse to signal novelty and salience. We have previously reported a calcium-dependent burst firing of dopamine cells mediated by L-type channels following cholinergic stimulation; this paper describes a cellular mechanism resulting in burst firing following L-type channel activation. Calcium influx through L-type channels following FPL 64176 or (S)-(–)-Bay K8644 induced burst firing independent of dopamine, glutamate, or calcium from the internal stores. Burst firing induced as such was completely blocked by the substrate site protein kinase C (PKC) inhibitor chelerythrine but not by the diacylglycerol site inhibitor calphostin C. Western blotting analysis showed that FPL 64176 and (S)-(–)-Bay K8644 increased the cleavage of PKC to generate protein kinase M (PKM) and the specific calpain inhibitor MDL28170 blocked this increase. Prevention of PKM production by inhibiting calpain or depleting PKC blocked burst firing induction whereas direct loading of purified PKM into cells induced burst firing. Activation of the N-methyl-d-aspartic acid type glutamate or cholinergic receptors known to induce burst firing increased PKM expression. These results indicate that calcium influx through L-type channels activates a calcium-dependent protease that cleaves PKC to generate constitutively active and labile PKM resulting in burst firing of dopamine cells, a pathway that is involved in glutamatergic or cholinergic modulation of the central dopamine system.

Dopaminergic (DA) 2 projections from the ventral tegmental area (VTA) constitute the mesolimbocortical system that underlies drug abuse and schizophrenia, primarily as the result of increased DA transmission (1)(2)(3). The strength of DA transmission is regulated by the interplay between spike-dependent DA release, DA reuptake, and autoreceptor-mediated negative feedback. Essentially, more intense spiking releases more DA at the terminal, which in turn activates the negative feedback mechanism to shut the system off. For enduring DA transmission at greater intensity, it would require a higher accumulation of DA at the synapse without triggering the negative feedback machinery. Burst firing is one such mode of discharge that has been shown to enhance DA transmission (4 -7) by saturating reuptake transporters and reducing autoreceptor inhibition (8).
DA cells in the VTA are capable of both pacemaker-like and burst firing (9), the latter being shown to signal novelty and salient stimuli in whole animal studies (10). DA cells in slices predominantly display pacemaker-like firing; burst firing can be induced by surrogate synaptic stimulation with bath application of glutamatergic or cholinergic agonists (11,12) that mobilize Ca 2ϩ . Ca 2ϩ entry has been shown to be pivotal in regulating firing patterns of DA neurons. Intracellular administration of Ca 2ϩ evokes burst firing, while intracellular Ca 2ϩ chelators block it (13). Our previous study shows that carbachol, a general cholinergic agonist, induces burst firing primarily by promoting Ca 2ϩ influx through L-type channels (12). These channels are responsible for approximately one third of total Ca 2ϩ currents of DA neurons (14 -16) and contribute preferentially to whole-cell Ca 2ϩ currents evoked by small depolarizations (16,17). In line with this, L-type channels have been shown to be involved in spontaneous and burst firing (12, 18 -20).
Activation of L-type channels has been shown to modulate synaptic strength in DA cells (21). Additionally, L-type channels gate Ca 2ϩ -release from internal Ca 2ϩ stores, activate plasma membrane Ca 2ϩ -dependent K ϩ channels, as well as several Ca 2ϩ -dependent kinases such as protein kinase C (PKC), Ca 2ϩ -calmodulin-dependent kinase II (CaMKII), and protein kinase A (PKA) (22). These kinases are capable of phosphorylating a variety of ion channels to regulate the excitability of neurons (23). Finding the mechanism that controls the firing mode of DA cells would provide a vital means of modulating the system in both normal and disease conditions. Here, we present results that Ca 2ϩ influx through L-type channels activates a Ca 2ϩ -dependent protease, which in turn cleaves PKC to generate a labile fragment that is constitutively active (termed protein kinase M, PKM) to induce burst firing in DA neurons.

EXPERIMENTAL PROCEDURES
Slice Preparation-All procedures involving animal handling and tissue harvesting were in accordance with guidelines set by the Institutional Animal Care Committee at the Memorial University of Newfoundland. Sprague-Dawley rat pups (9 -21 days old) of either sex were deeply anesthetized with halothane and killed by chest compression. The skull was quickly opened to expose the brain, which was cooled in situ with ice-cold, carbogenated artificial cerebrospinal fluid (ACSF, composition: 126 mM NaCl, 2.5 mM KCl, 1.2 mM NaH 2 PO 4 , 1.2 mM MgCl 2 , 2.4 mM CaCl 2 , 18 mM NaHCO 3 , and 11 mM glucose, pH 7.4, when bubbled with 95% O 2 and 5% CO 2 ). The brain was removed, and a block containing the midbrain was cut on a Leica vibratome (VT 1000, Heidelberger, Germany). Tissue slices were allowed to recover at room temperature (22°C) in carbogenated ACSF for at least 1 h prior to recording. Slices were further trimmed to fit into a recording chamber and continuously perfused with carbogenated ACSF at a rate of 2-3 ml min Ϫ1 at room temperature. For PKC depletion experiments, one of the two VTA slices from the same animal was incubated in ACSF with or without 1-2 M phorbol 12-myristate 13-acetate (PMA) for 20 -24 h at room temperature in a partially sealed beaker continuously bubbled with carbogen.
Patch Clamp Recording-All recordings were made from the VTA identified under a dissecting microscope (Leica MZ6). Patch electrodes were prepared from KG-33 glass micropipettes (OD 1.5 mm, Garner Glass CO., Claremont, CA) on a P-97 Brown-Flaming micropipette puller (Sutter Instruments, Novato, CA). For nystatin-perforated patch clamp recording, glass electrodes were filled to the tip with intracellular solution (120 mM potassium acetate, 40 mM HEPES, 5 mM MgCl 2 , and 10 mM EGTA with pH adjusted to 7.35 using 0.1 N KOH) and then back-filled with the same solution containing 450 g ml Ϫ1 nystatin and Pluronic F127, yielding a tip resistance of 4 -8 M⍀.
For conventional whole-cell recording, electrodes were filled with a solution containing 126 mM potassium gluconate, 10 mM KCl, 0.2 mM EGTA, 10 mM HEPES, 4 mM MgATP, and 0.3 mM Na 3 GTP with pH adjusted to 7.35 and osmolarity to 295 mosM, yielding a resistance of 2-4 M⍀. Gigaohm seals were made using a Warner PC-505B (Warner Instruments Inc., Hamden, CT) or a MultiClamp 700B (Axon Instruments, Foster City, CA) amplifier. Signals were sampled at 5 kHz and digitized by DigiData 1320A using pCLAMP (versions 8 and 9) software (Axon Instruments).
Selection of nystatin-perforated cell recordings in current clamp mode was determined by the size of the action potential, since many VTA cells were spontaneously active. After adequate partitioning of nystatin into the membrane, action potentials overshot 0 mV and measured at least 50 mV. Quality of conventional whole-cell recordings was assessed by a brief voltage step (Ϫ20 mV, 10 ms) from the holding potential (Ϫ55 mV). Only cells that had an access resistance of Ͻ30 M⍀ and an input resistance of Ͼ200 M⍀ were included. Cells whose access resistance increased significantly during the course of recording (Ͼ20%) were discarded. Episodic protocols were used to induce hyperpolarization activated current (I h ) and derive passive characteristics of the cell such as current-voltage relation-ship and input resistance. Current pulses for I h induction were of 1 s duration, and the intervals between pulses were 8 s to allow complete recovery of I h channels. Cells that displayed a prominent I h and an apparent DA-induced hyperpolarization were identified as putative DA cells (11,12,24).
Components of extracellular and intracellular solutions were purchased from bulk distributors Fisher Scientific (Nepean, Ontario, Canada) and VWR International (Missisauga, Ontario, Canada). All other chemicals were obtained from Sigma and Tocris (Ellisville, MO). Chemicals were dissolved in deionized water or Me 2 SO as required. Aliquots of stock solutions were kept at Ϫ30°C. Prior to application, an aliquot was diluted to working concentration and applied to the ACSF bath. PKM (Sigma) was kept at Ϫ80°C and was diluted to 1 unit ml Ϫ1 immediately before use in the internal solution for conventional whole-cell recording. DA solution was made fresh daily with an equimolar concentration of the antioxidant disodium metabisulfite.
Western Blotting-Phosphorylated PKC and PKM were detected by a phospho-PKC (pan) antibody (Cell Signaling) that recognizes PKC␣, -␤I, -␤II, -␦, -⑀, -, and -isoforms only when phosphorylated at a carboxyl-terminal residue homologous to Ser660 of PKC␤II. Ser-660 is in the catalytic domain of PKC and allows simultaneous visualization of phosphorylated PKC and PKM. VTA slices from the same animal were hemisected to give equivalent halves for control and experimental groups. Slices were trimmed to contain the VTA and part of the substantia nigra and were homogenized in RIPA (1ϫ phosphatebuffered saline, 1% Triton X-100, 50 mM Tris-HCl, pH 7.4, 0.5% deoxycholate, 50 mM ␤-glycerophosphate, 50 mM sodium fluoride, 5 mM EDTA, 0.1% sodium orthovanadate, 0.1% SDS, 75 ng/ml phenylmethylsulfanyl fluoride and 1ϫ complete protease inhibitor (Roche Diagnostics, Quebec, Canada)). Insoluble materials were removed by centrifugation (15,000 ϫ g for 10 min), and protein concentration for each sample was determined by BCA TM assay using bovine serum albumin standard (Pierce). Equivalent mass of total protein from slice homogenates were separated on an 8.5% acrylamide gel and transferred to a polyvinylidene fluoride membrane. The blots were blocked with Blotto (5% w/v nonfat milk in TBST (15 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.1% (v/v) Tween 20); 10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.1% Tween 20) for 1 h at room temperature, then incubated with phospho-PKC (pan) antibody (1: 1,000) diluted in Blotto, rocking overnight at 4°C. Blots were washed five times in TBST and followed by goat anti-rabbit horseradish peroxidase-conjugated IgG diluted in Blotto (Cell Signaling, 1:15,000) for 90 min at room temperature. After five washes in TBST, blots were incubated in enhanced SuperSignal West Pico chemiluminescent substrate (Pierce) to visualize specific immune complexes using hyperfilm x-ray film (Amersham Biosciences). Blots were then stripped in a solution of 2% SDS, 63 mM Tris-HCl, pH 6.8, and 0.1 M ␤-mercaptoethanol for 30 min at 50°C. Stripping solution was removed by two washes with TBST at room temperature, followed by blocking with Blotto and incubation with anti-␤-actin antibody (Sigma, 1:4,000) to act as loading controls. Quantification was obtained from densitometric measurements of immunoreactive bands using Gel Logic100 imaging system with Gel Logic200 Software (Eastman Kodak Co.).
Data Analysis-Electrophysiological data were analyzed offline with Mini Analysis (Synaptosoft Inc., Decatur, GA) and pCLAMP software. Basal firing frequencies were averaged values of at least 5 min stable baseline recording. I h was measured as the difference in current or voltage between instantaneous and steady-state readings. Analysis of firing behavior was based on interspike intervals (ISIs) measured with the Mini Analysis program. Averaged as well as instantaneous firing frequencies were derived from those intervals. Coefficient of variance (CV) was calculated as the mean of ISIs over a 1-min period divided by their standard deviation. To compare CV values between cell groups, they were normalized against the mean CV value of the first 5 min. Relative density of ISIs in 2-s bins were plotted to reveal the distribution of a given ISI series, and the resulting histogram was fitted to a Lowess function. Burst firing was defined as two spikes or more in each bursting cycle at a frequency higher than non-bursting periods and separated by a post-burst hyperpolarization. For Western blotting analysis, density of phospho-PKM and PKC bands were normalized against the density of ␤-actin from the same sample.
Data were expressed as means and S.E. Statistical comparisons of electrophysiological data were performed using two-tailed unpaired Student's t test. Western blotting data were compared using paired t test since each control and experimental pair contained hemisected slices from the same animal. Values were considered significant when p Ͻ 0.05.

RESULTS
Except when indicated, experiments were done using the nystatinperforated whole-cell recording method at room temperature. Dialysis of PKM into recorded cells was done using conventional whole-cell recording. Only DA cells in the VTA identified according to criteria outlined under "Experimental Procedures" were included. In nystatinperforated recording, the average hyperpolarization following a brief application of 50 M DA (within 90 s) was Ϫ8.29 Ϯ 0.57 mV (n ϭ 71, excluding cells used for PKC depletion test). Most cells (47 of 71, 66%) were spontaneously active with single spike firing at a low basal firing frequency of 0.41 Ϯ 0.04 Hz; the remainder (24 of 71, 34%) were quiescent during base-line recordings.
Opening of L-type Calcium Channels Converts Firing Patterns-Bath application of 1-4 M FPL 64176, a benzoylpyrole site L-type calcium channel opener, for 5-10 min converted firing patterns from quiescent state or single spiking to burst firing in 80.3% of treated cells (57 of 71), application of 1 M FPL 64176 induced burst firing in 63.5% cells (33 of 52). The responses were spikedependent. Percentage of converted burst firing in spontaneously firing cells (41 of 47, 87.2%) was much higher than that in quiescent cells (16 of 24, 66.7%), a response that appeared to be related to their different resting membrane potentials (Ϫ49.35 Ϯ 0.54 mV for spiking cells versus Ϫ52.52 Ϯ 0.92 mV for quiescent cells, p Ͻ 0.05). The responses were dose-dependent: 2 M FPL 64176 induced burst firing in 9 of 17 cells that did not respond to 1 M and similarly, 4 M induced burst firing in 4 of 10 cells that did not respond following a dose of 2 M.
In cells that were spontaneously firing, FPL 64176 first induced a membrane depolarization (2.00 Ϯ 0.38 mV, n ϭ 41) accompanied by a 48.8 Ϯ 9.3% increase in firing rate in the first 5 min of drug application; burst firing started after a varying period of latency. The lag between the start of drug application and burst firing ranged between 5 and 25 min with an average of 13 min (Fig. 1, A and F). Within a burst firing cycle, action potentials were fired with increasing frequencies followed by a pronounced post-burst hyperpolarization (Fig. 1B). The average intra-burst firing frequency (1.12 Ϯ 0.13 Hz, n ϭ 41) was much higher than basal tonic frequency (0.43 Ϯ 0.05 Hz). The development of burst firing in quiescent cells (n ϭ 16) was slightly different: they responded to FPL 64176 with a membrane depolarization (1.38 Ϯ 0.38 mV) followed by a sudden appearance of burst firing (Fig. 1D) or irregular single spiking that evolved into burst firing (Fig. 1E). The L-type channelmediated burst firing was long lasting even after prolonged washout up to 3 h; however, it was readily inhibited by L-type channel blocker nifedipine at a range of doses (1-10 M, n ϭ 29). Density plots of ISIs in 2-s bins showed an ISI distribution that fits to a single Gaussian distribution under control conditions and to a mixed Gaussian distribution following FPL 64176 application (Fig. 1C). The leftward shift of the main peak indicates higher firing frequencies and the second peak at much lower frequency represents the long pauses of firing between adjacent bursts.
Bath application of 5 M (S)-(Ϫ)-Bay K8644, a dihydropyridine site L-type channel activator, also induced burst firing in three of five cells tested (Fig. 1G). The time course of the response was similar to that following FPL 64176, a depolarization and an increase in firing rates followed by a conversion of firing patterns 10 -25 min after the start of application.
FPL 64176 Induces Burst Firing Independent of an Intermediate Transmitter-L-type channels have been shown to enhance a slow NMDA current in DA neurons (21), suggesting that L-type channel opening could induce burst firing by promoting glutamate transmission. Because FPL 64176-induced burst firing was long lasting, we examined the involvement of synaptic mechanisms in two ways: whether induced burst firing persisted in the presence of a mixture containing 100 M APV, 10 M 6-cyano-7-nitroquinoxaline-2,3-dione, and 100 M picrotoxin (blocking NMDA, AMPA, and GABA A receptors, respectively) and whether pretreatment with this mixture altered the ability of FPL 64176 to induce burst firing. FPL 64176-induced burst firing was robust following the application of the mixture for 10 -30 min (n ϭ 7, Fig. 2A). Similarly, cells that were treated with the mixture for 5-10 min prior to FPL 64176 (1 M) application in the presence of the mixture still responded with burst firing (n ϭ 3). These results suggest that L-type channel opening does not induce burst firing either directly through increased glutamatergic transmission or indirectly by way of GABAergic interneurons in the VTA.
Activation of L-type channels may release DA from the soma and dendrites, since this release has been shown to be Ca 2ϩ -dependent (25,26). DA acting at somatodendritic autoreceptors forms the short loop negative feedback to regulate the excitability of DA cells, we therefore examined whether D 2 receptors were involved in burst firing. The D 2 receptor antagonist sulpiride (1-10 M) applied for 10 -15 min to 5 burst firing cells induced by FPL 64176 had no effect on firing patterns (Fig. 2B). Two cells to which 10 M sulpiride was applied for 8 min prior to FPL 64176 application in the presence of sulpiride displayed burst firing similar to that induced by FPL 64176 alone. These data suggest that L-type channel activation does not result in somatodendritic DA release to induce burst firing.
Internal Ca 2ϩ Stores Are Not Involved-Ca 2ϩ entry through L-type channels gates ryanodine receptors on internal Ca 2ϩ stores to mediate Ca 2ϩ -induced Ca 2ϩ release (22), we tested whether this was involved in burst firing. Cyclopiazonic acid (CPA) and thapsigargin were used to inhibit the sarco-endoreticulum Ca 2ϩ -dependent ATPase (SERCA) that refills internal Ca 2ϩ stores. After burst firing was induced by FPL 64176, CPA (20 -30 M) applied for 20 -60 min (n ϭ 4) or thapsigargin (1-2 M) applied for 50 -70 min (n ϭ 3) did not alter the induced burst firing (Fig. 2C). In three cells that were pretreated with CPA (20 M) for 30 min, FPL 64176 (1 M) was still able to induce strong burst firing. These results indicate that internal Ca 2ϩ stores are not necessary for L-type channelmediated burst firing.
Ca 2ϩ -dependent Protein Kinase Mediates Burst Firing-Increased intracellular Ca 2ϩ activates Ca 2ϩ -sensitive protein kinases such as PKC, PKA, and CaMKII that can phosphorylate ion channels including L-type channels themselves and regulate the excitability of neurons (27)(28)(29)(30). After burst firing was induced by FPL 64176, bath application of the PKA inhibitor H-89 (5 or 10 M) for 40 -130 min (n ϭ 4), the CaMKII inhibitor KN-93 (1-10 M) for 25-100 min (n ϭ 6), or both for 60 -70 min (n ϭ 3) did not block burst firing (Fig. 3A). The substrate site PKC inhibitor, chelerythrine (40 M), applied for 15-44 min (n ϭ 6) completely blocked FPL 64176-induced burst firing and the accompanying membrane potential oscillation. This blockade was reversible since burst firing reappeared after washing out chelerythrine for 18 -33 min (Fig. 3B). To establish the role of PKC in burst firing, we further used an inhibitor that binds to the diacylglycerol site of the regulatory domain. Calphostin C (1-2 M) applied for 40 -180 min (n ϭ 4) was largely ineffective except in one cell where burst firing slowed marginally but persisted (Fig. 3C). These results suggest that the catalytic subunit of PKC itself or a kinase with a similar substrate site is involved in burst firing.
Proteolytic Cleavage of PKC Mediates Burst Firing-To solve the apparent inconsistency between the two PKC inhibitors, we considered other modes of PKC activation. Ca 2ϩ entry has been shown to activate the protease calpain that proteolytically cleaves PKC to form PKM which has been shown to be constitutively active with a very short half-life within the cell (31-34). We therefore tested whether proteolytic cleavage of PKC could explain the inconsistency in response to the two PKC inhibitors. Cells were induced to burst fire by To further validate the role of PKM in L-type channel-induced burst firing, we compared the levels of phospho-PKC and PKM by Western blotting total protein lysates from slices that were treated with (S)-(Ϫ)-Bay K8644 or FPL 64176 alone and with MDL 28170 followed by FPL 64176. As shown in Fig. 4B   Burst Firing Cannot Be Induced after PKC Depletion-Since PKM is generated by cleavage of PKC (33,34), it is an obvious test to determine whetherf burst firing can be induced following PKC depletion. Persistent activation by phorbol esters such as PMA, leads to degradation of PKC isoforms that have the diacylglycerol site and is used experimentally to produce a PKC-depleted cell (35,36). Slices were incubated with or without PMA (1-2 M) for 20 -24 h followed by FPL 64176 (4 M) to induce burst firing. Regardless of incubation conditions, cells from these slices were no longer spontaneously active although they were capable of firing action potentials (Fig. 5A). There was no difference in resting membrane potential (control: Ϫ50.05 Ϯ 1.36 mV, n ϭ 6; PMA-treated: Ϫ48.25 Ϯ 1.38 mV, n ϭ 7, p Ͼ 0.05) or input resistance (control: 403.24 Ϯ 56.12 M⍀, n ϭ 6; PMA-treated: 293.58 Ϯ 42.93 M⍀, n ϭ 7, p Ͼ 0.05) between control and PMA-treated groups. PMA-treated cells had a significantly smaller DA-induced hyperpolarization (control: 12.03 Ϯ 3.43 mV; PMA-treated: 2.99 Ϯ 0.67 mV, p Ͻ 0.05). Western blots showed a disappearance in PMA-treated slices of a band around 82 kDa and the concomitant decrease of the PKM band at 45 kDa following FPL 64176 (Fig. 5B).
There was a clear difference in FPL 64176-induced firing between the two groups. Of 6 cells from control slices, FPL 64176 (4 M) induced burst firing in 3 (Fig. 5C1), large mem-brane potential oscillations in 2 and no change in one cell. Both burst firing and membrane potential oscillation could be blocked by nifedipine (5 M). Of seven cells from PMA-treated slices, FPL 64176 (4 M) only induced an average membrane depolarization of 0.62 Ϯ 0.26 mV with no burst firing or membrane potential oscillation in any of the cells tested (Fig. 5C2).
Direct Loading of PKM Induces Burst Firing-If PKM links L-type channels and burst firing, it would be expected that direct loading of PKM (1 unit/ml pipette solution) into the cells through the conventional whole-cell recording pipette should induce similar firing mode switching. Out of seven cells, three had single spike firing initially that became irregular and clustered by 3-5 min, with burst firing appearing at 9 -12 min. Burst firing was strong and stable for 20 -30 min (Fig. 6A) and was qualitatively similar to burst firing induced by FPL 64176. When burst firing was prevalent, density plots of ISIs showed a typical two-peak ISI distribution (Fig. 6C2). In another 2 cells, single spike firing became burst-like characterized by clustered spikes followed by a pause without a post-burst hyperpolarization. Only in two cells was no response seen. The control experiments were done using the same internal pipette solution without PKM, and all six cells tested maintained regular firing during the 40 min recording period (Fig. 6B). Plotting normalized ISI coefficient of variance values against the average value of the first 5 min (Fig. 6D)  Activation of NMDA and Cholinergic Receptors Increases PKM Expression-To test whether PKM is involved in the physiological regulation of burst firing in DA neurons, glutamate, or cholinergic agonists known to induce burst firing in DA cells were used to stimulate VTA slices and the amount of PKM was detected by Western blotting. Hemisected slices from the same animal were treated with or without NMDA (20 M) or carbachol (20 M) for 20 min to test whether it increased the levels of phosphorylated PKM. As shown in Fig. 7A, application of either glutamate agonist NMDA or cholinergic agonist carbachol caused a considerable increase in phospho-PKM expression. Densitometry of the bands corresponding to phospho-PKM of six independent experiments (Fig. 7B) showed that both NMDA (159 Ϯ 26% relative to untreated values, p Ͻ 0.05) and carbachol (154 Ϯ 41% relative to untreated values, p Ͻ 0.05) significantly increased phospho-PKM levels. Since glutamate and cholinergic inputs are major synaptic modulators of DA cells, these data indicate that PKM is involved in the synaptic regulation of firing behavior of DA cells.

DISCUSSION
Burst firing is an important property of DA cells in the VTA to signal novelty and salience that are associated with normal or abnormal expression of motivation and reward. We have previously reported that the cholinergic agonist carbachol induces a calcium-dependent burst firing of DA cells by promoting Ca 2ϩ influx through L-type channels, this paper studies how L-type channel activation induces burst firing. In this study, we observed that direct activation of L-type channels converted firing patterns of VTA DA cells independent of glutamate, dendritically released DA, or Ca 2ϩ released from internal stores. Ca 2ϩ influx through L-type channels did not activate PKA or CaMKII to induce burst firing. We have shown that Ca 2ϩ influx through L-type channels activates calpain proteases that in turn cleave PKC, releasing the short lived, active PKM fragment to produce burst firing. Disruption at any point along this chain of events results in an inability of causing burst firing of DA neurons in the VTA. We have also shown that PKM is involved in regulation of DA cells under physiological stimulations where the NMDA type glutamate or cholinergic receptors were activated.
L-type Channel-induced Burst Firing Does Not Require an Intermediate Transmitter-L-type channels are primarily expressed in the soma and dentrites of DA cells in the VTA area (15) although they have shown to modulate synaptic transmission to the VTA (21). Glutamate acting at the NMDA receptors mediates a slow excitatory synaptic transmission to VTA DA cells (37) that can be enhanced by L-type channel activation (21), suggesting that the effects of L-type channel activation could be secondary to increased glutamate transmission. This is particulary relevant since NMDA is one of the pharmacological tools used to induce burst firing in DA neurons (11,20). Our results that combined synaptic blockade at the NMDA, AMPA, and GABA A site before or after L-type channel activation did not alter burst firing suggest that glutamate transmission is not involved in L-type channel-induced burst firing. Alternatively, activation of L-type channels increases firing (19) and subsequently releases DA from somatodendritic site to terminate firing by D 2 -mediated autoinhibition (26). Autoinhibition wanes as DA is taken back by DA transporters (DAT) and a new cycle starts. This mode of action appears logical since all the proposed mechanisms have been shown to work in DA cells. However, L-type channel activation induced strong burst firing that was not blocked by the D 2 receptor antagonist sulpiride. Blocking the D 2 receptor before application of L-type channel opener did not alter burst firing either, arguing against a role for somatodendritic D 2 receptors in L-type channel-induced burst firing.
We have shown previously that burst firing of DA cells depends on Ca 2ϩ oscillation (12), a phenomenon that could be supported by the so-called Ca 2ϩ -induced Ca 2ϩ release from internal stores. The L-type channel has been shown to directly gate Ca 2ϩ releasing channels on the membrane of sarcoplasmic reticulum (22). The inositol 1,4,5-trisphosphate channels have a bell shape sensitivity relationship with intracellular Ca 2ϩ levels (22) so that their conductance increases initially with higher Ca 2ϩ concentrations; further increases in Ca 2ϩ reduces their conductance capabilities. Also, Ca 2ϩ release from internal stores has been found to mediate a slow inhibition in DA cells taking hundreds of milliseconds to develop (38). This is on a similar time scale to burst firing cycles we observed in slice preparations. These mechanisms can possibly produce a biphasic response that fits the membrane oscillation we observed. In our experiments, burst firing induced by the L-type channel opener FPL 64176 persisted when internal stores were depleted by prolonged inhibition of the refilling enzyme SERCA, suggesting that internal stores do not play an obligatory role in burst firing.
Ca 2ϩ Influx Generates PKM to Induce Burst Firing-The opening of L-type channels did not induce burst firing within the first 5 min following FPL 64176 application although Ca 2ϩ influx had already changed the excitability of the cells in the form of membrane depolarization and increased firing frequency. Burst firing emerged 5-25 min following the start of drug application and lasted for hours. The lag seems to indicate that Ca 2ϩ entry activated a signaling cascade that takes time to develop or a second messenger must migrate between the intracellular compartments. Since protein kinase activation requires multiple stages of phosphorylation and translocation (39), the involvement of a Ca 2ϩ -dependent protein kinases presents a reasonable mechanism underlying Ca 2ϩ -dependent burst fir- ing. For example, CaMKII has been shown to become persistently active during Ca 2ϩ oscillation to enhance L-type channels (40). Similarly, PKA has long been associated with adrenergic enhancement of L-type channel activity (41,42). The involvement of these kinases could explain why L-type channels continue operating despite high intracellular Ca 2ϩ levels. ChannelphosphorylationhasbeensuggestedtoreducetheCa 2ϩdependent inactivation of L-type channels (43). That burst firing persisted in the presence PKA and CaMKII blockers supports the concept that they are not necessary for Ca 2ϩdependent burst firing.
The striking result that chelerythrine, a PKC catalytic domain inhibitor, reversibly blocked burst firing, whereas calphostin C, a regulatory domain inhibitor did not, raised the possibility that Ca 2ϩ influx through L-type channels activates a PKC-like kinase or PKC operates in an atypical fashion. Elevated intracellular Ca 2ϩ has been shown to activate calpain (44), a protease that cleaves PKC catalytic domain from the regulatory domain (31)(32)(33) to generate PKM. Our results support this mode of PKC action and its regulation of burst firing through: 1) L-type channel openers induced burst firing accompanied by increased levels of phosphorylated PKM; 2) inhibition of the Ca 2ϩ -dependent protease calpain reduced the level of phosphorylated PKM and concomitantly blocked burst firing; 3) PKM is generally thought to be a proteolytic product of PKC (31)(32)(33) and depletion of PKC isoforms following prolonged incubation with PMA prevented L-type channel-induced burst firing; 4) direct loading of the cell with purified PKM through the recording pipette induced burst firing in a similar manner to Ca 2ϩ influx, with the temporal development of burst firing being consistent with diffusion of PKM into the cell (Fig. 8). The evidence collectively indicates that Ca 2ϩ influx through the L-type channel activates proteolytic cleavage of PKC to generate PKM that facilitates the development and maitenance of burst firing of DA cells.
PKC Is Functionally Important for DA-related Conditions-PKM was found in the rat brain in 1970s (45) but its function remained obscure until recent reports associating it with learning and memory (46,47). Several PKC isoforms have been found to give rise to proteolytic PKM counterparts, and it is reasonable to assume that all members of the PKC family are capable of generating their respective PKMs because of their similar structure of the hinge region that joins the regulatory and catalytic domains. Our results did not identify the parent PKC isoform that was cleaved to form PKM; it may involve several PKC isoforms, since many are expressed in DA neurons (48). The PKM used for cellular loading is derived from enzymatic digestion of total PKC, but from the PKC depletion experiments, it appears that the PKM involved in burst firing is generated by the isoforms with the diacylglycerol site. PKC has been implicated in addiction and motivation. Injection of PKC inhibitors into the VTA reduces cocaine-induced DA release in the nucleus accumbens (49) and delays the onset of behavioral sensitization (50,51). Repeated administration of cocaine increases overall PKC activity in the VTA, which may initiate behavioral sensitization (52). Being persistently active, PKM could fulfill all these behaviors involving PKCs. More importantly, PKM could achieve all this by its role in burst firing reported here as burst firing elevates terminal DA more effectively and all these behaviors involve increased DA transmission. In addition, addiction is a learned behavior and PKM has been  shown to facilitate learning and memory in rats (46,47). Taken together, PKC and its cleaved product PKM might play a significant role in central DA transmission and its related pathologies such as drug abuse. It is especially encouraging that activation of the NMDA and cholinergic receptors increased PKM because these receptors not only mediate synaptic regulation of DA cells, but they are also known to induce burst firing in DA cells.
Mechanisms of PKM-induced Burst Firing-There are many factors contributing to the strength of DA transmission. Besides the efficiency of the release machinery, it is usually agreed that DA reuptake and autoreceptor-mediated inhibition are important regulators of DA transmission. Released DA is rapidly taken back to the terminal by DAT, whereas D 2 autoreceptors at both terminal and somatodendritic sites respond to increased DA levels by a negative feedback loop that inhibit DA cell activity. PKC has been shown to modulate both processes. Activation of PKC leads to a decrease in DAT capacity (53,54) due to accelerated internalization (55,56), reduced recycling (57), and degradation (58). Additionally, PKC activation enhances phosphorylation, desensitization, and trafficking of D 2 receptors (59) resulting in dampened negative feedback. In PMA-treated slices, the responses of the somatodendritic autoreceptors to DA were blunted, and it remains to be tested whether these counterparts in the terminal are similarly affected. The effects of PKC on DAT and autoreceptor inhibitory feedback and our results that the proteolytic product of PKC (PKM) induces burst firing, a firing mode that is more effective in increasing DA release, suggest a tantalizing possibility that PKC serves as a signaling hub, integrating different aspects of DA transmission to boost its functionality over a long period of time. This postulate further supports the findings that implicate PKC in addiction, a chronic and debilitating condition that has long been thought to be due to enhanced DA transmission.