dopamine receptor-mediated ca2 signaling in striatal ... · dars are divided into the d1 class (d1r...

Dopamine Receptor-mediated Ca2� Signaling in Striatal MediumSpiny Neurons*

Received for publication, July 1, 2004, and in revised form, August 2, 2004Published, JBC Papers in Press, August 2, 2004, DOI 10.1074/jbc.M407389200

Tie-Shan Tang and Ilya Bezprozvanny‡

From the Department of Physiology, University of Texas Southwestern Medical Center, Dallas, Texas 75390-9040

Inositol 1,4,5-trisphosphate (InsP3) and cAMP are thetwo second messengers that play an important role inneuronal signaling. Here, we investigated the interac-tions of InsP3- and cAMP-mediated signaling pathwaysactivated by dopamine in striatal medium spiny neu-rons (MSN). We found that in �40% of the MSN, applica-tion of dopamine elicited robust repetitive Ca2� tran-sients (oscillations). In pharmacological experimentswith specific agonists and antagonists, we found thatthe observed Ca2� oscillations were triggered by activa-tion of D1 class dopamine receptors (DARs). We furtherdemonstrated that activation of phospholipase C wasrequired for induction of dopamine-induced Ca2� oscil-lations and that maintenance of dopamine-evoked Ca2�

oscillations required both Ca2� influx and Ca2� mobili-zation from internal Ca2� stores. In “priming” experi-ments with a type 2 5-hydroxytryptamine receptor ago-nist, we have shown a likely role for calcyon in couplingD1 class DARs with Ca2� oscillations in MSN. In exper-iments with the DAR-specific agonist SKF83959, we dis-covered that phospholipase C activation alone could notaccount for dopamine-induced Ca2� oscillations. We fur-ther demonstrated that direct activation of protein ki-nase A by 8-bromo-cAMP or inhibition of protein phos-phatase-1 (PP1) or calcineurin (PP2B) resulted inelevation of basal Ca2� levels in MSN, but not in Ca2�

oscillations. In experiments with competitive peptides,we have shown an importance of type 1 InsP3 receptorassociation with PP1� and with AKAP9�protein kinase Afor dopamine-induced Ca2� oscillations. In experimentswith MSN from DARPP-32 knock-out mice, we demon-strated a regulatory role of DARPP-32 in dopamine-in-duced Ca2� oscillations. Our results indicate that, fol-lowing D1 class DAR activation, InsP3 and cAMPsignaling pathways converge on the type 1 InsP3 recep-tor, resulting in Ca2� oscillations in MSN.

Dopamine is an important transmitter and neuromodulatorin the brain. The cellular mechanisms by which dopamineaffects neuronal function are only beginning to be elucidated (1,2). Striatal medium spiny neurons (MSN)1 express multiple

subtypes of dopamine receptors (DARs) (3–5). On the basis oftheir molecular structure and pharmacological properties,DARs are divided into the D1 class (D1R and D5R) and D2 class(D2R, D3R, and D4R) (6). D1 class DARs are coupled to Gs/olf,activation of adenylyl cyclase, and cAMP production (3). Acti-vation of D2 class DARs has dual effects of inhibiting cAMPproduction (3) and activating phospholipase C� (PLC�) (7). Aputative D1 class DAR subtype coupled to PLC activation andphosphatidylinositol 4,5-diphosphate hydrolysis, but not tocAMP production, has been postulated (8–10), but has not yetbeen isolated or cloned. Recently, a specific agonist for thisreceptor (SKF83959) was identified (11). The D1R-binding pro-tein calcyon has been isolated by yeast two-hybrid methods(12). Association of D1/D5 receptors with calcyon enables cou-pling of D1/D5 receptors with Gq/11, resulting in PLC activationand inositol 1,4,5-trisphosphate (InsP3) generation (12, 13).

Cross-talk between cAMP and Ca2� signaling pathwaysplays an important role in dopaminergic signaling in the neos-triatum (1). Activation of D1 class DARs enhances L-type Ca2�

channel activity (14–16) and currents via the �-amino-3-hy-droxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor (17)and the N-methyl-D-aspartic acid (NMDA) receptor (18, 19). Incontrast, activation of D2 class DARs reduces L-type Ca2�

currents (7) and NMDA receptor activity (20). D1/D5 receptor-mediated facilitation of L-type Ca2� channels and AMPA andNMDA receptors results from increased phosphorylation ofthese channels by protein kinase A (PKA) (16, 21) and de-creased dephosphorylation of these channels by protein phos-phatase-1 (PP1) (17, 22, 23). DARPP-32 (dopamine- and cAMP-regulated phosphoprotein of Mr 32,000) (24, 25) is partlyresponsible for inhibition of PP1 activity following activation ofD1/D5 receptors (26). DARPP-32 phosphorylated by PKA at asingle threonine residue (Thr-34) is transformed into a potentinhibitor of PP1, which in turn regulates the phosphorylationstate of many neurotransmitter receptors and voltage-gatedion channels (1).

The type 1 inositol 1,4,5-trisphosphate receptor (InsP3R1) isa predominant InsP3 receptor isoform in the brain (27).InsP3R1 plays an important role in neuronal Ca2� signaling(28). Neuronal InsP3R1 is one of the major substrates of PKAphosphorylation in the brain (29, 30). PKA phosphorylatesInsP3R1 at two sites, Ser-1589 and Ser-1755 (31–36). PKAphosphorylation activates InsP3R1 by increasing the sensitiv-ity of InsP3R1 to activation by InsP3 (35, 37–39). In previousbiochemical studies, we discovered direct association of neuro-

* This work was supported by the Robert A. Welch Foundation, theHereditary Disease Foundation, and National Institutes of HealthGrant R01 NS38082 (to I. B.). The costs of publication of this articlewere defrayed in part by the payment of page charges. This article musttherefore be hereby marked “advertisement” in accordance with 18U.S.C. Section 1734 solely to indicate this fact.

‡ To whom correspondence should be addressed: Dept. of Physiology,UT Southwestern Medical Center, 5323 Harry Hines Blvd., K4.112,Dallas, TX 75390-9040. Tel.: 214-648-6737; Fax: 214-648-2974;E-mail: [email protected].

1 The abbreviations used are: MSN, medium spiny neuron(s); DAR,dopamine receptor; D1R, dopamine type 1 receptor; PLC, phospholipaseC; InsP3, inositol 1,4,5-trisphosphate; AMPA, �-amino-3-hydroxy-5-

methyl-4-isoxazolepropionic acid; NMDA, N-methyl-D-aspartic acid;PKA, protein kinase A; PP, protein phosphatase; InsP3R1, type 1 ino-sitol 1,4,5-trisphosphate receptor; DIV, days in vitro; ACSF, artificialcerebrospinal fluid; FITC, fluorescein isothiocyanate; LIZ, leucine/iso-leucine zipper; EGFP, enhanced green fluorescent protein; Br, bromo;TTX, tetrodotoxin; CNQX, 6-cyano-7-nitroquinoxaline-2,3-dione; 5-HT2,type 2 5-hydroxytryptamine receptor; SERCA, sarco/endoplasmic retic-ulum Ca2� ATPase.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 279, No. 40, Issue of October 1, pp. 42082–42094, 2004© 2004 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org42082

nal InsP3R1 with PP1� (39) and showed that theInsP3R1�AKAP9�PKA complex is formed in the brain (40). Inexperiments with striatal slices, we demonstrated transientphosphorylation of striatal InsP3R1 by PKA in response todopamine and proposed that InsP3R1 may participate in cross-talk between cAMP and Ca2� dopaminergic signaling path-ways in the striatum (39). In this study, we used Ca2� imagingtechniques to investigate dopamine-induced Ca2� signals inprimary cultures of striatal MSN.

EXPERIMENTAL PROCEDURES

DARPP-32 Knock-out Mice—The generation and breeding ofDARPP-32 knock-out mice (C57BL/6; a kind gift of Dr. Paul Greengard)were described previously (26, 41). The brains of day 14.5–15.5 embryosfrom wild-type (D32�/�) and DARPP-32 knock-out (D32�/�) mice werecollected, and the striata were dissected. The striatal lysates of wild-type and D32�/� embryos were analyzed by Western blotting withanti-DARPP-32 monoclonal antibodies (Cell Signaling Technologies).

Primary Cultures of Rat and Mouse MSN—Primary cultures of ratMSN were established from day 17–18 rat embryos as described previ-ously (42). The mouse MSN cultures were established with some mod-ifications of the protocol used for rat MSN culture. Briefly, using land-marks described previously (43), striata were dissected from brains ofD32�/� and D32�/� embryonic mice in ice-cold dissection solution (1�divalent-free Hanks’ balanced salt solution, 15 mM HEPES, 10 mM

NaHCO3, and 100 units/ml penicillin/streptomycin (Invitrogen), pH7.2). The striata from mice with identical genotypes were pooled andtreated with 0.25% trypsin for 7 min at 37 °C. After addition of 10%heat-inactivated fetal bovine serum (Invitrogen) in Dulbecco’s modifiedEagle’s medium (Invitrogen), the tissue was dissociated with tritura-tion solution (1� divalent-free Hanks’ balanced salt solution and 1.0%DNase I, pH 7.2) (44) by repetitive pipetting. The cells were washed;plated at a density of 1 � 106 cells/ml on 12-mm round coverslipsprecoated with poly-D-lysine (Mr � 30,000–70,000, 0.01% final concen-tration) in plating medium containing 60% Dulbecco’s modified Eagle’smedium, 30% Neurobasal medium, 10% fetal bovine serum, and 100units/ml penicillin/streptomycin; and maintained at 37 °C in 5% CO2.24 h later, the cultures were transferred to culture medium (65%Dulbecco’s modified Eagle’s medium, 30% Neurobasal medium, 1� B27(Invitrogen), 5% fetal bovine serum, and 100 units/ml penicillin/strep-tomycin). 4 �M cytosine arabinoside (Sigma) was added at 2–4 days invitro (DIV) to inhibit glial cell growth. The cultures were fed with freshculture medium every 7 days. The identity of established cultures wasconfirmed in immunostaining experiments with anti-GAD65 mono-clonal antibodies (Chemicon International, Inc.).

Ca2� Imaging Experiments—Ca2� imaging experiments with 18–20DIV rat and 16–18 DIV mouse MSN were performed as describedpreviously (42). Briefly, MSN were loaded with 5 �M Fura-2/AM (Mo-lecular Probes, Inc.) in artificial cerebrospinal fluid (ACSF; 140 mM

NaCl, 5 mM KCl, 1 mM MgCl2, 2 mM CaCl2, and 10 mM HEPES, pH 7.3)for 45 min at 37 °C. For imaging experiments, the coverslips weremounted on a recording/perfusion chamber (RC-26G, Warner Instru-ments) maintained at 37 °C (PH1, Warner Instruments), positioned onthe movable stage of an Olympus IX-70 inverted microscope, and per-fused with ACSF by gravity flow. For some experiments (see below), theculture was washed extensively with Ca2�-free ACSF (CaCl2 was omit-ted from ACSF, which was supplemented with 100 �M EGTA). Follow-ing Fura-2 loading, the MSN cells were intermittently excited by 340-and 380-nm UV light (DeltaRAM illuminator, PTI) using a Fura-2dichroic filter cube (Chroma Technologies) and a 60� UV light-grade oilimmersion objective (Olympus). The emitted light was collected by anIC-300 camera (PTI), and the images were digitized by ImageMasterPro software (PTI). Base-line (1–3 min) measurements were obtainedprior to bath application of dopamine dissolved in ACSF or Ca2�-freeACSF. Others drugs were applied to the recording chamber as describedbelow. The dopamine and drug solutions were prewarmed to 37 °Cbefore application to MSN. Images at 340- and 380-nm excitation wave-lengths were captured every 5 s, and 340/380 nm ratio image traceswere recorded. Background fluorescence was determined according tothe manufacturer’s (PTI) recommendations and subtracted.

R9 Peptide Loading Experiments—R9 (RRRRRRRRR) and R9-IC(RRRRRRRRRGHPPHMNVNPQQPA) peptides were chemically syn-thesized (University of Texas Southwestern Protein Chemistry Tech-nology Center), coupled to fluorescein isothiocyanate (FITC) at the Nterminus, and dissolved in phosphate-buffered saline. In loading exper-iments, the R9 peptides were added to rat MSN for 10 min at 50 �M.

Following R9 peptide loading, neurons were washed and incubated inculture medium for �2 h prior to Ca2� imaging experiments. As judgedby FITC fluorescence, �90% of the MSN were loaded with the R9 andR9-IC peptides in these experiments.

Enhanced Green Fluorescent Protein (EGFP) and EGFP-RT1-Leucine/Isoleucine Zipper (LIZ) Transfections—The EGFP-RT1-LIZconstruct was generated by subcloning the PCR-amplified LIZ region ofrat InsP3R1 (amino acids 1251–1287) (40) into the pEGFP-C3 expres-sion vector (Clontech) and verified by sequencing. The pEGFP-C3 ex-pression vector without an insert was used as a negative control(EGFP). The rat MSN cultures at 19–20 DIV were transfected with theEGFP or EGFP-RT1-LIZ plasmid by the calcium phosphate method asdescribed previously (42, 45). 48 h after transfection, MSN were loadedwith 5 �M Fura-2/AM and used in Ca2� imaging experiments as de-scribed above. Prior to Ca2� imaging experiments, EGFP- and EGFP-RT1-LIZ-transfected MSN were identified by GFP imaging as describedpreviously (42).

Drugs—SKF83959 was provided by the National Institute of MentalHealth Synthesis Program (Menlo Park, CA). Dopamine hydrobromide,atropine sulfate, U73122, U73343, thapsigargin, 8-bromo (Br)-cAMP, ca-lyculin A, cyclosporin A, and okadaic acid were from Calbiochem. (�)-SKF38393 hydrobromide ((�)-1-phenyl-2,3,4,5-tetrahydro-(1H)-3-benzaz-epine-7,8-diol hydrochloride), R(�)-SCH23390 hydrochloride, spiperonehydrochloride, (�)-MK801 maleate, (�)-quinpirole dihydrochloride, tetro-dotoxin (TTX), ketanserin tartrate, prazosin hydrochloride, �-methyl-5-hydroxytryptamine, 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), andnifedipine were purchased from Tocris (Ellisville, MO).

RESULTS

Dopamine Induces Ca2� Oscillations in MSN—To test theeffects of dopamine stimulation on Ca2� signaling, we per-formed Ca2� imaging experiments with primary cultures ofMSN established from embryonic day 18 rats as describedpreviously (42, 46). To monitor intracellular Ca2� dynamics,MSN were loaded with the ratiometric Ca2� imaging dyeFura-2, and the intracellular free Ca2� concentrations wereestimated from the ratio of Fura-2 emission at 340- and 380-nmexcitation wavelengths (340/380 nm ratio). On average, thebasal levels of 340/380 nm ratios in unstimulated MSN wereequal to 0.52 � 0.05 (n � 29). Approximately 40% of thecultured MSN (20 DIV) in our experiments displayed repetitiveCa2� transients (oscillations) in response to application of 400�M dopamine (Fig. 1, A and B). On average (n � 29 MSN), theseCa2� oscillations started 5.1 � 2.2 min after dopamine appli-cation, had an amplitude of 0.89 � 0.1, occurred with a fre-quency of 12 � 6 spikes/20 min, and lasted at least 20–30 min.

D1 Class Receptors Specifically Mediate Dopamine-evokedCa2� Oscillations in MSN—Striatal MSN abundantly expressmultiple subtypes of DARs (3–5, 47). Based on pharmacologicaland molecular properties, DARs are divided into two classes:D1 and D2. Which DAR class mediates the dopamine-inducedCa2� oscillations observed in our experiments (Fig. 1)? Toanswer this question, we performed a series of experimentswith D1 and D2 class receptor-specific antagonists and ago-nists. We found that the dopamine-induced Ca2� oscillations(Figs. 1 and 2A) in MSN were completely blocked in the pres-ence of 5 �M SCH23390 (D1 class DAR antagonist) and 5 �M

spiperone (D2 class DAR antagonist) (Fig. 2B). We furtherfound that blockade of D2 class DARs by 5 �M spiperone hadonly a minor effect on dopamine-induced Ca2� oscillations (Fig.2C), whereas blockade of D1 class DARs by 5 �M SCH23390resulted in almost complete suppression of dopamine-inducedCa2� oscillations (Fig. 2D). In complementary experiments, wefound that the D1 class DAR-specific agonist SKF38393 (50 �M)induced Ca2� oscillations in MSN (Fig. 2E), whereas the D2class DAR-specific agonist quinpirole (10 �M) was much lesseffective (Fig. 2F). Characteristically, application of quinpiroleresulted in an instant Ca2� spike (Fig. 2F), whereas applicationof SKF38393 resulted in Ca2� oscillations with a delay ofseveral minutes (Fig. 2E), as observed after application of do-pamine (Figs. 1B and 2A). Taken together, these data suggest

Dopamine and Ca2� Signaling in MSN 42083

that activation of either D1 or D2 class DARs leads to Ca2�

responses in MSN, but that D1 class DARs are coupled to Ca2�

oscillations more efficiently and via a different mechanismthan D2 class DARs.

Classical D1 class DARs (D1R and D5R) are coupled to cAMPproduction and not to PLC activation and InsP3-induced Ca2�

release. What is an explanation for the Ca2� signals observedin our experiments (Figs. 1 and 2)? One possibility is that 400�M dopamine may facilitate action potential firing in MSNcultures, which will lead to Ca2� influx via voltage-gated L-type Ca2� channels. To test this possibility, we performedexperiments in the presence of 2 �M TTX, a specific voltage-gated Na� channel blocker that inhibits action potential firingin striatal neurons (48, 49). Preincubation of MSN with 2 �M

TTX had no effect on dopamine-induced Ca2� oscillations (Fig.3A), indicating that the observed responses were not due toaction potential firing in MSN cultures. Another possibility isthat 400 �M dopamine can nonspecifically activate other PLC-linked neurotransmitter receptors expressed in MSN, such as5-hydroxytryptamine type 2 (5-HT2) receptors, �1-adrenorecep-tors, and muscarinic type acetylcholine receptors (50–52). Todetermine whether activation of any of these receptors is in-volved in dopamine-induced Ca2� oscillations, we repeated ex-periments in the presence of 20 �M ketanserin (5-HT2 receptorantagonist), 10 �M prazosin (�1-adrenoreceptor antagonist),and 5 �M atropine (muscarinic type acetylcholine receptor an-tagonist). To rule out activation of D2 receptors (also coupled toPLC), we also included 5 �M spiperone (D2 class DAR antago-nist). Preincubation of MSN with a TTX/spiperone/ketanserin/prazosin/atropine mixture had no significant effect on dopam-ine-induced Ca2� oscillations (Fig. 3B), confirming thatdopamine-induced Ca2� oscillations are mediated primarily by

D1 class DARs, with a minor contribution from D2 class DARs(Fig. 2F).

In additional control experiments, we studied Ca2� signalsin MSN induced by 5-HT2 receptor activation. We found thatapplication of 100 �M �-methyl-5-hydroxytryptamine (5-HT2

receptor-specific agonist) in the presence of 2 �M TTX inducedone large Ca2� spike, which was occasionally followed by one totwo spikes of much smaller amplitude (Fig. 3C). These re-sponses were completely eliminated in the presence of 20 �M

ketanserin (5-HT2 receptor antagonist) (Fig. 3D). Thus, MSNCa2� signals mediated by 5-HT2 receptors differ from dopam-ine-induced responses because 1) Ca2� transients in responseto dopamine occurred with a lag time of 3–7 min, and the firstCa2� transient in response to �-methyl-5-hydroxytryptamineoccurred instantly (Fig. 3C); 2) multiple Ca2� transients ofequal amplitude (oscillations) were observed in response todopamine, but only one large transient and one to two muchsmaller transients were observed in response to �-methyl-5-hydroxytryptamine (Fig. 3C); and 3) dopamine-induced Ca2�

transients were not sensitive to ketanserin (Fig. 3B), but�-methyl-5-hydroxytryptamine-induced Ca2� transients werecompletely blocked by ketanserin (Fig. 3D). These results indi-cate that the dopamine-induced Ca2� oscillations in our exper-iments were mediated predominantly by D1 class DARs, butnot by 5-HT2 receptors.

Ca2� Influx Is Required for Maintenance of Dopamine-in-duced Ca2� Oscillations in MSN—The dopamine-induced in-crease in the intracellular free Ca2� concentration observed inour experiments (Figs. 1–3) could result from Ca2� releasefrom internal Ca2� stores and/or extracellular Ca2� influx. Totest the role of extracellular Ca2� influx in dopamine-evokedCa2� oscillations, we shifted the MSN from 2 mM Ca2� in the

FIG. 1. Dopamine-induced Ca2�

transients in rat MSN. A, dopamine in-duces Ca2� transients in MSN. Repre-sentative images show Fura-2 340/380nm ratios in rat MSN before (�1 min) andafter (0–28 min) application of 400 �M

dopamine. The images were collected inACSF containing 2 �M TTX. Dopaminewas applied to the bath at 0 min. Fura-2340/380 nm ratio images in MSN areshown in pseudocolor for each minute ofthe experiment. The pseudocolor calibra-tion scale for 340/380 nm ratios is shownon the right. B, a single 340/380 nm ratiotrace in MSN. A region of interest waschosen in the soma of a single MSN (indi-cated by the arrowheads in A), and the340/380 nm ratio in the selected region ofinterest was plotted versus time in theexperiment. The basal 340/380 nm ratio(before application of dopamine) for thecell shown was 0.55. The time of 400 �M

dopamine application (300 s) is indicatedby the arrow. For the cell shown, Ca2�

oscillations started 360 s after dopamineapplication (at 660 s) and had an averageamplitude of 0.9 and a frequency of 14spikes/20 min. Data from the same exper-iment were used to generate A and B.

Dopamine and Ca2� Signaling in MSN42084

extracellular medium to Ca2�-free extracellular medium dur-ing dopamine-induced oscillations. We found that the dopam-ine-induced Ca2� oscillations quickly ceased in Ca2�-free me-dium, but restarted again following return to the extracellularmedium containing 2 mM Ca2� (Fig. 4A). These experimentsdemonstrate that extracellular Ca2� influx is required for themaintenance of dopamine-induced Ca2� oscillations in MSN.

What was the source of extracellular Ca2� influx in theseexperiments? Activation of D1 class DARs enhances L-typeCa2� currents (14–16) and activates AMPA receptors (17) andNMDA receptors (16, 18, 19, 53). To evaluate a possible role ofthese channels in dopamine-induced Ca2� oscillations, we useda combination of specific blockers of L-type Ca2� channels (10�M nifedipine), AMPA receptors (20 �M CNQX), and NMDAreceptors (10 �M (�)-MK801). We found that dopamine-inducedCa2� oscillations quickly ceased after application of a (�)-MK801/CNQX/nifedipine mixture (Fig. 4B). When MSN werepreincubated with a (�)-MK801/CNQX/nifedipine mixture for10 min, application of dopamine resulted in greatly attenuatedCa2� oscillations (Fig. 4C). Preincubation with a (�)-MK801/CNQX, CNQX/nifedipine, or (�)-MK801/nifedipine mixturehad a partial inhibitory effect on dopamine-induced Ca2� os-cillations (Fig. 4, D–F). From these experiments, we concludedthat L-type Ca2� channels and AMPA and NMDA receptorscontribute jointly to maintenance of dopamine-induced Ca2�

oscillations in MSN, but are not required for initiation of theseoscillations (Fig. 4C).

PLC Activation and Intracellular Ca2� Mobilization Are Re-

quired for the Initiation and Maintenance of Dopamine-inducedCa2� Oscillations in MSN—The experiments described above(Fig. 4) indicate that Ca2� influx via L-type Ca2� channels andAMPA and NMDA receptors is necessary for maintenance ofdopamine-induced Ca2� oscillations, but is not required fortheir initiation. Thus, in the next series of experiments, wefocused on mobilization of Ca2� from intracellular Ca2� stores.Activation of PLC leads to hydrolysis of phosphatidylinositol4,5-diphosphate and generation of diacylglycerol and InsP3.InsP3 activates InsP3R1 and releases Ca2� from intracellularCa2� stores. In MSN, activation of PLC-coupled class I metabo-tropic glutamate receptors efficiently evokes intracellular Ca2�

mobilization (42, 54). To test whether activation of PLC isinvolved in dopamine-induced Ca2� oscillations in MSN, weperformed experiments with U73122, a selective PLC inhibitor.We found that preincubation with 10 �M U73122 for 10 minresulted in a complete block of dopamine-induced Ca2� oscilla-tions (Fig. 5A). In contrast, preincubation with the same con-centration of the inactive analog U73343 had no effect ondopamine-induced Ca2� oscillations (Fig. 5B). These experi-ments demonstrate an essential role of PLC in initiation ofdopamine-induced Ca2� oscillations in MSN. To test the im-portance of Ca2� mobilization from intracellular Ca2� stores,we performed experiments with the SERCA Ca2� pump-spe-cific inhibitor thapsigargin. We found that acute application of10 �M thapsigargin quickly stopped dopamine-induced Ca2�

oscillations (Fig. 5C). Preincubation of MSN with 10 �M thap-sigargin for 10 min resulted in almost a complete block of

FIG. 2. D1 class DARs mediate dopamine-evoked Ca2� transients in MSN. The images were collected in ACSF and analyzed as describedin the Fig. 1 legend. A, dopamine-induced Ca2� transients in a single MSN. The data are representative of n � 48 MSN. B, a combination of a D1class DAR antagonist (5 �M SCH23390) and a D2 class DAR antagonist (5 �M spiperone) prevents dopamine-induced Ca2� oscillations in MSN (n �46). C, a combination of a D2 class DAR antagonist (5 �M spiperone) and an NMDA receptor blocker (10 �M (�)-MK801) does not preventdopamine-induced Ca2� oscillations in MSN (n � 27). D, a combination of a D1 class DAR antagonist (5 �M SCH23390) and an NMDA receptorblocker (10 �M (�)-MK801) suppress dopamine-induced Ca2� oscillations in MSN (n � 43). E, the D1 class DAR-specific agonist SKF38393 (50 �M)induces Ca2� oscillations in MSN (n � 10). F, the D2 class DAR-specific agonist quinpirole (10 �M) is less potent in inducing Ca2� oscillations inMSN (n � 23).


dopamine-induced Ca2� oscillations (Fig. 5D). The experimentsdescribed here (Fig. 5) led us to conclude that PLC activationand InsP3R1-mediated Ca2� mobilization from intracellularCa2� stores play a critical role in initiation and maintenance ofdopamine-induced Ca2� oscillations.

A Role of the Putative PLC-linked D1 Class DAR in Dopam-ine-induced Ca2� Oscillations—Classical D1 class DARs (D1Rand D5R) are coupled to cAMP production and not to PLCactivation. What is an explanation for the critical role of PLC(Fig. 5) in D1 class receptor-mediated Ca2� oscillations? Incontrol experiments, we ruled out involvement of other PLC-coupled receptors expressed in MSN, such as 5-HT2 receptors,�1-adrenoreceptors, and muscarinic type acetylcholine receptorreceptors (Fig. 3). Recent data suggested the existence of adistinct D1 class DAR subtype that is coupled to PLC activation(8–10). This receptor has not been purified or cloned, but re-cently, the benzazepine compound SKF83959 has been identi-fied as a specific agonist for this putative PLC-linked D1 classDAR subtype (11). To test a potential role of this putative D1class receptor subtype in the observed Ca2� signals, we per-formed experiments with SKF83959. We found that applicationof 400 �M SKF83959 induced an instant single Ca2� transientin MSN (Fig. 6A). A similar response to 400 �M SKF83959 wasobserved in Ca2�-free medium (Fig. 6B). The response to 400�M SKF83959 was eliminated in the presence of the PLC in-hibitor U73122 (Fig. 6C). These results confirm the existence ofa PLC-coupled DAR activated by SKF83959 in MSN. However,application of SKF83959 never resulted in Ca2� oscillations,indicating that activation of these receptors is not sufficient tosupport dopamine-induced Ca2� oscillations in MSN.

A Role of Calcyon in Dopamine-induced Ca2� Oscillations—D1R and D5R are coupled to the heterotrimeric G protein �subunit Gs/olf, and this activation results in activation of ad-enylyl cyclase and cAMP production. Recently, the D1R-bind-ing protein calcyon has been identified by yeast two-hybrid

screen (12, 13). Calcyon is a single pass transmembrane proteinof 24 kDa that binds to the C-terminal tail of D1R. Coexpres-sion of D1R or D5R with calcyon in a heterologous systemenables coupling of these receptors to the Gq/11 subunit, result-ing in activation of PLC�, generation of InsP3, and Ca2� re-lease (12). In neurons, association of D1 class receptors withcalcyon is promoted by “priming” resulting from activation ofother types of Gq/11-coupled receptors (12, 55). To evaluate apossible role of calcyon in the dopamine-induced Ca2� tran-sients observed in our experiments, we “primed” MSN with 50�M �-methyl-5-hydroxytryptamine (5-HT2 receptor-specific ag-onist) in the presence of 2 �M TTX. As described above (Fig.3C), application of �-methyl-5-hydroxytryptamine to MSN re-sulted in an instant single Ca2� spike, consistent with tran-sient activation of Gq/11 proteins and PLC (Fig. 7A). Applicationof 100 �M dopamine to “naı̈ve” MSN was not sufficient to resultin any Ca2� responses (Fig. 7B). In contrast, application of thesame concentration of dopamine to MSN primed with 50 �M

�-methyl-5-hydroxytryptamine resulted in an instant singleCa2� spike, followed by repetitive Ca2� transients (oscillations)after a 3–7-min delay (Fig. 7C). The frequency and amplitude ofthe observed Ca2� oscillations were similar to those of oscilla-tions induced by 400 �M dopamine in experiments with naı̈ve(non-primed) MSN (Fig. 1). These experiments indicate thatpriming of MSN with 50 �M �-methyl-5-hydroxytryptamineincreased the potency of dopamine to induce Ca2� oscillations,consistent with an involvement of calcyon (12, 13).

A Role of PKA Phosphorylation in Dopamine-induced Ca2�

Oscillations in MSN—The experiments described above (Figs.5 and 6) indicate that activation of PLC is necessary but notsufficient to result in Ca2� oscillations in MSN. To explain thisresult, we reasoned that increases in both cAMP and InsP3

might be required to support dopamine-induced Ca2� oscilla-tions. To test this hypothesis, we investigated the effects of aPKA activator (8-Br-cAMP) and of protein phosphatase inhib-

FIG. 3. D2 class receptors, 5-HT2 receptors, muscarinic type acetylcholine receptors, and �1-adrenoreceptors are not essential fordopamine-induced Ca2� oscillations in MSN. The images were collected in ACSF containing 2 �M TTX. A, dopamine-induced Ca2� transientsin a single MSN. The data are representative of n � 31 MSN. B, a combination of antagonists for 5-HT2 receptors (20 �M ketanserin),�1-adrenoreceptors (10 �M prazosin), muscarinic type acetylcholine receptors (5 �M atropine), and D2 class DARs (5 �M spiperone) does not preventdopamine-evoked Ca2� oscillations in MSN (n � 12). C, application of the 5-HT2 receptor agonist �-methyl-5-hydroxytryptamine (100 �M) to MSNinduces an instant single Ca2� spike, followed by one to two much smaller spikes (n � 49). D, the 5-HT2 receptor antagonist ketanserin (20 �M)blocks Ca2� responses to the 5-HT2 receptor agonist �-methyl-5-hydroxytryptamine (100 �M) in MSN (n � 27).


itors on Ca2� signals in MSN. We found that application of 500�M 8-Br-cAMP to MSN resulted in elevation of basal Ca2�

levels in MSN (Fig. 8A), most likely due to PKA-induced phos-phorylation and activation of InsP3R1 (39). Similar effects wereobserved in response to application of 10 nM calyculin A (PP1/PP2A inhibitor) (Fig. 8B) or 10 nM cyclosporin A (PP2B inhib-itor) (Fig. 8C). In contrast, application of 1 nM okadaic acid, aPP2A-specific inhibitor at this concentration, had a less pro-nounced and more delayed effect on basal Ca2� levels in MSN(Fig. 8D). The relative potencies of 8-Br-cAMP, calyculin A,cyclosporin A, and okadaic acid in these experiments correlatewell with the relative potency of these drugs to shift striatalInsP3R1 into the PKA-phosphorylated state that was comparedin our previous experiments (39).

Persistent activation of PKA or blockade of PP1 or PP2Bresults in Ca2� elevation, most likely due to hyperphosphory-lation and activation of InsP3R1 (39). Is it possible to generateCa2� oscillations in MSN by activating both the PLC and PKApathways? The answer this question, we applied a mixture of400 �M SKF83959 and 500 �M 8-Br-cAMP to MSN. We foundthat application of an SKF83959/8-Br-cAMP mixture (Fig. 8E)

resulted in an initial small Ca2� transient (similar toSKF83959 application) (Fig. 6A), followed by a persistent ele-vation of Ca2� levels (similar to 8-Br-cAMP application) (Fig.8A). Thus, SKF83959 and 8-Br-cAMP appear to affect Ca2�

levels in an independent manner, and not lead to Ca2� oscilla-tions as observed in response to dopamine. We further exploredthe connection between InsP3 and cAMP signaling pathways byexamining the effects of calyculin A, cyclosporin A, and okadaicacid on dopamine-induced Ca2� oscillations. We found thatpreincubation with calyculin A suppressed dopamine-inducedCa2� oscillations, presumably due to the rise in basal Ca2�

levels (Fig. 8F). Preincubation with cyclosporin A initially in-creased the frequency of Ca2� oscillations and eventually sup-pressed oscillations, presumably due to the rise in basal Ca2�

levels (Fig. 8G). In contrast to calyculin A and cyclosporin A,okadaic acid did not have a significant effect on dopamine-induced Ca2� oscillations (Fig. 8H). From these results, weconclude that continuous hyperphosphorylation of InsP3R1 byPKA disrupts the in vivo regulatory mechanism required fordopamine-induced Ca2� oscillations in MSN.

FIG. 4. Ca2� influx is required for maintenance (but not initiation) of dopamine-induced Ca2� oscillations in MSN. The images werecollected in ACSF and analyzed as described in the Fig. 1 legend. A, dopamine-induced Ca2� oscillations quickly ceased in Ca2�-free medium andrestarted after re-addition of 2 mM Ca2� to the extracellular medium (n � 43). Changes in extracellular Ca2� concentration are shown above the340/380 nm ratio trace. B, addition of a mixture of an L-type Ca2� channel blocker (10 �M nifedipine), an AMPA receptor blocker (20 �M CNQX),and an NMDA receptor blocker (10 �M (�)-MK801) stopped dopamine-induced Ca2� oscillations in MSN (n � 44). C, preincubation of MSN witha mixture of an L-type Ca2� channel blocker (10 �M nifedipine), an AMPA receptor blocker (20 �M CNQX), and an NMDA receptor blocker (10 �M

(�)-MK801) greatly attenuated dopamine-induced Ca2� oscillations in MSN (n � 36). D–F, preincubation of MSN with a mixture of an AMPAreceptor blocker (20 �M CNQX) and an NMDA receptor blocker (10 �M (�)-MK801) (n � 40), an L-type Ca2� channel blocker (10 �M nifedipine)and an AMPA receptor blocker (20 �M CNQX) (n � 28), and an L-type Ca2� channel blocker (10 �M nifedipine) and an NMDA receptor blocker (10�M (�)-MK801) (n � 20), respectively, had a partial inhibitory effect on dopamine-induced Ca2� oscillations in MSN.


A Role of InsP3R1-PP1� Association in Dopamine-inducedCa2� Oscillations—In a previous study, we described a directassociation between the InsP3R1 C terminus and PP1� (39). Isformation of an InsP3R1�PP1� complex physiologically rele-vant? To address this question, we introduced the IC peptide(amino acids 2736–2749 of InsP3R1), which corresponds to aminimal PP1�-binding site in the InsP3R1 sequence (39), intoMSN. We reasoned that the IC peptide would displace PP1�from the complex with InsP3R1 and therefore may have adominant-negative effect on the ability of PP1� to dephospho-rylate InsP3R1. To introduce the IC peptide into MSN, we tookadvantage of a recently developed protein delivery technology(56, 57). In our experiments, we used the R9 signal (58) todeliver the FITC-labeled C-terminal R9-IC peptide and controlR9 peptide into cultured MSN (see “Experimental Procedures”for details). 2 h following FITC-labeled R9 peptide loading,MSN were washed, loaded with Fura-2, and used in Ca2�

imaging experiments. Using FITC fluorescence, we estimatedthat �90% of the MSN were loaded with the R9 and R9-ICpeptides in our experiments.

Application of 400 �M dopamine induced Ca2� oscillations inboth R9 (Fig. 9A) and R9-IC (Fig. 9B) peptide-loaded MSN.However, Ca2� oscillations were more frequent and had in-creased amplitude in R9-IC peptide-loaded neurons (Fig. 9B).Statistical analysis revealed that the average basal Ca2� levelswere similar (p � 0.05) in control, R9 peptide-loaded, andR9-IC peptide-loaded MSN (Table I). The average latency fromdopamine application to the first Ca2� spike was shorter inR9-IC peptide-loaded neurons than in control or R9 peptide-loaded neurons (Table I), but the difference did not reach sta-tistical significance (p � 0.05, unpaired t test). The averageamplitude of Ca2� transients was equal to 0.89 � 0.1 (n � 29)in control MSN, 0.86 � 0.09 (n � 18) in R9 peptide-loadedMSN, and 1.1 � 0.15 (n � 29) in R9-IC peptide-loaded MSN(Table I). The average frequency of Ca2� transients was equalto 12 � 6 spikes/20 min (n � 29) in control MSN, 11 � 6spikes/20 min (n � 18) in R9 peptide-loaded MSN, and 20 � 8spikes/20 min (n � 29) in R9-IC peptide-loaded MSN (Table I).From these results, we determined that loading of MSN withthe R9-IC peptide results in statistically significant (p � 0.05)

TABLE IStatistical analysis of dopamine-induced Ca2� oscillations in MSN

The quantitative parameters of Ca2� oscillations are shown as means � S.D. The number of cells used in the analysis is shown in the first column(n).

MSN n Basal Ca2� (R340/380) Lag time Amplitude (R340/380) Frequency (spikes/20 min)

min

Control 29 0.52 � 0.05 5.1 � 2.2 0.89 � 0.1 12 � 6R9 18 0.51 � 0.03 4.7 � 2.1 0.86 � 0.09 11 � 6R9-IC 29 0.49 � 0.04 4.3 � 2.2 1.1 � 0.15a 20 � 8a

EGFP 17 0.57 � 0.03 3.9 � 2.8 0.82 � 0.1 15 � 4EGFP-RT1-LIZ 15 0.57 � 0.05 16.2 � 6.3a 0.76 � 0.04 7 � 3a

Mouse D32�/� 38 0.51 � 0.03 4.5 � 2.1 0.87 � 0.08 13 � 6Mouse D32�/� 154 0.53 � 0.04 6.1 � 2.5a 0.87 � 0.11 9 � 4a

a Significant (p � 0.05, unpaired t test) differences from control MSN.

FIG. 5. PLC activation and intracellular Ca2� mobilization are required for the initiation and maintenance of dopamine-inducedCa2� oscillations in MSN. The images were collected in ACSF and analyzed as described in the Fig. 1 legend. A, preincubation with a PLCinhibitor (10 �M U73122) prevented dopamine-induced Ca2� oscillations in MSN (n � 40). B, preincubation with 10 �M U73343 (inactive analogof U73122) had no effect on dopamine-induced Ca2� oscillations in MSN (n � 41). C, application of the SERCA Ca2� pump inhibitor thapsigargin(10 �M) stopped dopamine-induced Ca2� oscillations in MSN (n � 37). D, preincubation of MSN with the SERCA Ca2� pump inhibitor thapsigargin(10 �M) for 10 min suppressed dopamine-induced Ca2� oscillations in MSN (n � 22).


increases in the average spike amplitude and spike frequency(Table I). These effects appear to be specific for IC peptidesequence, as loading with the control R9 peptide had no signif-icant effects (p � 0.05) on the main properties of dopamine-induced Ca2� oscillations in MSN (Table I). Thus, we concludethat disruption of InsP3R1-PP1� association by the IC peptidehas a significant potentiating effect on the amplitude and fre-quency of dopamine-induced Ca2� oscillations in MSN.

A Role of InsP3R1-AKAP9-PKA Association in Dopamine-induced Ca2� Oscillations—PKA phosphorylation increasesInsP3R1 sensitivity to InsP3 (35, 37–39). In recent biochemicalexperiments, we demonstrated the formation of aInsP3R1�AKAP9�PKA ternary complex in the brain (40). Wefound that InsP3R1-AKAP9 association is mediated via the LIZmotif in the InsP3R1 coupling domain and the fourth LIZ motifin AKAP9 (40). We further showed that InsP3R1-AKAP9 asso-ciation is disrupted in the presence of the recombinant LIZfragment of InsP3R1 (RT1-LIZ) (40). To evaluate the functionalconsequences of InsP3R1-AKAP9-PKA association for dopam-ine-induced Ca2� signaling, we transiently expressed theEGFP-RT1-LIZ construct in MSN. As a negative control, MSNcultures were transfected with the EGFP plasmid. MSN trans-fected with EGFP or EGFP-RT1-LIZ were identified by GFPimaging as we described previously (42).

We found that 400 �M dopamine induced Ca2� oscillations inMSN transfected with either EGFP (Fig. 9C) or EGFP-RT1-LIZ(Fig. 9D). However, compared with EGFP-transfected MSN,Ca2� oscillations in EGFP-RT1-LIZ-transfected MSN startedafter a longer delay and occurred with a reduced frequency(Fig. 9, C and D). Statistical analysis revealed that the averagebasal Ca2� levels were similar (p � 0.05) in control (untrans-fected), EGFP-transfected, and EGFP-RT1-LIZ-transfectedMSN (Table I). The average latency from dopamine applicationto the first Ca2� spike was equal to 5.1 � 2.2 min (n � 29) incontrol (untransfected) MSN, 4 � 2.8 min (n � 11) in EGFP-transfected MSN, and 16.2 � 2.1 min (n � 9) in EGFP-RT1-LIZ-transfected MSN (Table I). The amplitude of Ca2� tran-sients was lower in EGFP-RT1-LIZ-transfected neurons thanin control (untransfected) or EGFP-transfected neurons, butthe difference did not reach a level of statistical significance(p � 0.05, unpaired t test) (Table I). The average frequency ofCa2� transients was equal to 12 � 6 spikes/20 min (n � 29) incontrol (untransfected) MSN, 15 � 4 spikes/20 min (n � 11) inEGFP-transfected MSN, and 7 � 3 spikes/20 min (n � 9) inEGFP-RT1-LIZ-transfected MSN (Table I). Thus, expression of

FIG. 6. Role of the putative PLC-linked D1 class DAR in do-pamine-induced Ca2� oscillations. The images were collected inACSF and analyzed as described in the Fig. 1 legend. A, application of400 �M SKF83959 (specific agonist for the putative PLC-linked D1 classDAR subtype) (11) induced an instant Ca2� transient in MSN (n � 91).B, application of 400 �M SKF83959 induced a Ca2� transient in MSN inCa2�-free ACSF (n � 33). C, preincubation with the PLC inhibitorU73122 (10 �M) prevented SKF83959-evoked Ca2� responses in MSN(n � 58).

FIG. 7. Dopamine-induced Ca2� oscillations are primed by ac-tivation of 5-HT2 receptors. The images were collected in ACSFcontaining 2 �M TTX and analyzed as described in the Fig. 1 legend. A,application of 50 �M �-methyl-5-hydroxytryptamine (5-HT2 receptor-specific agonist) caused an instant single Ca2� spike in rat MSN (n �24). B, 100 �M dopamine was not sufficient to induce any Ca2� responsein naı̈ve MSN (n � 78). C, 100 �M dopamine induced Ca2� oscillationsin MSN primed with 50 �M �-methyl-5-hydroxytryptamine 5 min priorto dopamine application (n � 38).


EGFP-RT1-LIZ in MSN resulted in a significantly (p � 0.05)longer latency duration and reduced spiking frequency (TableI). Expression of EGFP alone had no significant effects (p �0.05) on the main properties of dopamine-induced Ca2� oscil-lations in MSN compared with control (untransfected) cells(Table I), indicating that the observed effects are specific forthe RT1-LIZ sequence.

A Role of DARPP-32 in Dopamine-induced Ca2� Oscilla-tions—DARPP-32 (59) is a regulatory phosphoprotein that hasbeen suggested to play a key role in dopaminergic signaling inthe striatum by regulating PP1 activity (1). In contrast to othercomponents of the dopaminergic signaling pathway, no phar-macological blockers of DARPP-32 function are currently avail-able. To evaluate the importance of DARPP-32 in dopamine-induced Ca2� signals, we took advantage of DARPP-32 knock-out mice (26, 41). Western blot experiments with anti-DARPP-32 monoclonal antibodies confirmed the presence ofthe DARRP-32 protein in striatal lysates prepared from thewild-type (D32�/�) mouse pups, but not from the knock-out(D32�/�) mouse pups (Fig. 10, A and B).

We established primary MSN cultures from the wild-type(D32�/�) and knock-out (D32�/�) embryonic brains and ana-lyzed dopamine-induced Ca2� signals in 16–18 DIV MSN by

Ca2� imaging with the Fura-2 indicator. Application of 400 �M

dopamine induced repetitive Ca2� transients (oscillations) inwild-type mouse MSN (Fig. 10C) and in D32�/� MSN (Fig.10D), similar to rat MSN (Fig. 1). Compared with wild-type(D32�/�) mouse MSN, oscillations in D32�/� mouse MSN had asimilar amplitude (Fig. 10, C and D). However, compared withwild-type MSN, oscillations in D32�/� mouse MSN startedwith a longer delay after dopamine application and had areduced frequency of oscillations (Fig. 10, C and D). Statisticalanalysis revealed that the main parameters of dopamine-in-duced Ca2� spikes were not significantly different (p � 0.05)between wild-type mouse MSN and rat MSN (Table I). Fur-thermore, the average basal Ca2� levels and the average am-plitude of Ca2� transients were similar (p � 0.05) in D32�/�

and D32�/� mouse MSN (Table I). In contrast, the averagelatency from dopamine application to the first Ca2� spike wasequal to 4.5 � 2.1 min (n � 38) in wild-type MSN and 6.1 � 2.5min (n � 154) in D32�/� mouse MSN (Table I). Also, theaverage frequency of Ca2� transients was equal to 13 � 6spikes/20 min (n � 38) in wild-type MSN and 9 � 4 spikes/20min (n � 154) in D32�/� mouse MSN (Table I). Thus, weconclude that genetic ablation of DARPP-32 leads to a statis-tically significant (p � 0.05) increase in the lag time between

FIG. 8. Effects of PKA phosphorylation on Ca2� signaling in MSN. The images were collected in ACSF and analyzed as described in theFig. 1 legend. A–D, changes in basal Ca2� levels in MSN induced by application of 500 �M 8-Br-cAMP (membrane-permeable cAMP analog) (n �64), 10 nM calyculin A (PP1 inhibitor) (n � 34), 10 nM cyclosporin A (PP2B inhibitor) (n � 33), and 1 nM okadaic acid (OA; PP2A inhibitor) (n �33), respectively; E, Ca2� responses in MSN induced by application of a SKF83959/8-Br-cAMP mixture (n � 39); F–H, effects of 10 nM calyculinA (n � 39), 10 nM cyclosporin A (n � 40), and 1 nM okadaic acid (n � 23), respectively, on dopamine-induced Ca2� oscillations in MSN.


dopamine application and the first Ca2� spike and reduction inthe frequency of dopamine-induced Ca2� spikes in MSN, con-sistent with a regulatory role played by DARPP-32 in MSN.

DISCUSSION

Cross-talk between cAMP and Ca2� signaling pathwaysplays an important role in dopaminergic signaling in the neos-triatum (1). In this study, we have investigated dopamine-evoked Ca2� signals in cultured striatal MSN. Our results haveelucidated a connection between D1 class DAR activation andcAMP/Ca2� signaling in MSN and provided insights into cel-lular mechanisms of dopamine function in striatum. The dataobtained in our study are consistent with the model shown inFig. 11. We propose that dopamine acts on D1/D5 receptors,leading to activation of adenylyl cyclase, production of cAMP,and activation of PKA (3). PKA phosphorylates DARPP-32 andInhibitor-1 proteins, converting them into potent inhibitors ofPP1 (1). Activation of PKA and inhibition of PP1 lead to in-creased phosphorylation and activation of L-type Ca2� chan-nels, NMDA receptors, and AMPA receptors (16, 17, 21–23)and InsP3R1 (39). We recently discovered an association ofneuronal InsP3R1 with AKAP9�PKA, which is mediated by aLIZ motif in the InsP3R1 sequence (40). EGFP-RT1-LIZ trans-fection experiments (Fig. 9, C and D) demonstrated that thePKA associated with InsP3R1�AKAP9 plays a major role indopamine-induced InsP3R1 phosphorylation in MSN. PKAphosphorylation of Ca2� influx channels and InsP3R1 is neces-sary but not sufficient to result in Ca2� oscillations. As exper-iments with the PLC antagonist U73122 demonstrated (Fig. 5),activation of PLC is also required. Previous studies demon-strated coupling of D2 class DARs to InsP3 production and

Ca2� release in striatal neurons (7). However, dopamine-in-duced Ca2� oscillations in our experiments were mediated byD1 class (cAMP-coupled), not by D2 class (PLC-coupled), DARs(Figs. 2 and 3).

We propose two potential solutions to this apparent contra-diction (see Fig. 11). One possibility is that dopamine acts on aputative PLC-linked D1 class DAR (9, 10, 60). Indeed,SKF83959, a specific agonist for this putative PLC-linked D1class DAR (11), induced Ca2� release from intracellular storesin MSN (Fig. 6). Another possibility is that some fraction ofD1/D5 receptors in MSN is associated with calcyon, whichenables coupling of these receptors to Gq/11 and PLC (12, 13). Insupport of the calcyon hypothesis, priming of MSN with a5-HT2 receptor agonist facilitated dopamine-induced Ca2� os-cillations in our experiments (Fig. 7C). It has been previouslyreported that priming is not effective in striatal neurons (55).One potential explanation for this discrepancy is different agesof MSN cultures: 18–20 DIV in our experiments and 4–10 DIVin the experiments of Lezcano and Bergson (55). Indeed, incontrol experiments, we found that responsiveness of MSN todopamine increased with maturity of culture (data not shown).It remains to be determined whether D1/D5 receptors com-plexed with calcyon (12, 13) and a putative PLC-linked D1 classDAR (8–11) are the same or distinct (as shown on Fig. 11)molecular entities.

Simultaneous activation of cAMP and InsP3 signaling path-ways leads to activation of InsP3R1 (see Fig. 11). Interestingly,dopamine-induced Ca2� oscillations started with a delay of �5min (Fig. 1). In contrast, activation of metabotropic glutamatetype 1 and 5 receptors (42), D2 class DARs (Fig. 2F), 5-HT2

FIG. 9. Effects of the R9-IC peptide and EGFP-RT1-LIZ on dopamine-induced Ca2� oscillations in MSN. The images were collected inACSF containing 2 �M TTX and analyzed as described in the Fig. 1 legend. A and B, representative dopamine-induced Ca2� oscillations in MSNloaded with the R9 and R9-IC peptides, respectively. The sequences of the R9 and R9-IC peptides are shown above the traces. C and D,representative dopamine-induced Ca2� oscillations in MSN transfected with EGFP and EGFP-RT1-LIZ, respectively.


receptors (Fig. 3C), and the putative PLC-coupled D1 classDAR (Fig. 6A) leads to an instant Ca2� transient in MSN. Thedelay between application of dopamine and initiation of Ca2�

oscillations is similar to the delay between application of 8-Br-cAMP (PKA activator) and calyculin A (PP1 inhibitor) andchanges in basal Ca2� levels in MSN (Fig. 8, A and B). Toexplain these results, we reasoned that phosphorylation ofInsP3R1 by PKA is necessary for initiation of Ca2� oscillations.PKA phosphorylation activates InsP3R1 by increasing the sen-sitivity of InsP3R1 to activation by InsP3 (35, 37–39), andapparently, InsP3 levels in dopamine-stimulated MSN are suf-ficient for activation of phosphorylated InsP3R1, but not foractivation of unphosphorylated InsP3R1.

Following InsP3R1 activation and release of Ca2� from in-tracellular stores, the cytosolic Ca2� concentration rises.InsP3R1 is under feedback regulation by cytosolic Ca2�: lowCa2� concentrations (�300 nM) activate InsP3R1, and higherCa2� concentrations (�300 nM) inhibit InsP3R1 (61–63). Thus,release of Ca2� is expected to proceed in a highly cooperativemanner and to terminate quickly due to Ca2� inactivation ofInsP3R1. An increase in cytosolic Ca2� will also lead to activa-tion of the Ca2�-dependent phosphatase calcineurin (PP2B)(Fig. 11). Activated calcineurin dephosphorylates DARPP-32and Inhibitor-1, resulting in disinhibition of PP1 (1), which inturn dephosphorylates and “turns off” InsP3R1 (39) and Ca2�

influx channels (16, 17, 21–23). Our model predicts thatInsP3R1 must be phosphorylated by PKA again for the nextCa2� spike to occur. Thus, in contrast to most models of Ca2�

oscillations, which are centered on biphasic modulation ofInsP3R1 by Ca2� (64), dopamine-induced Ca2� oscillations inMSN rely on an intricate interplay between Ca2� and cAMPsignaling pathways, leading to repetitive rounds of InsP3R1phosphorylation by PKA and dephosphorylation by PP1.

This hypothesis is supported by the effects of cyclosporin A(calcineurin inhibitor) on the phosphorylated state of striatalInsP3R1 (39) and dopamine-induced Ca2� oscillations in MSN(Fig. 8G). The proposed role of DARPP-32 is consistent with theanalysis of dopamine-induced Ca2� oscillations in MSN fromDARPP-32 knock-out mice (Fig. 10). An increase in the lag timeand a reduction in the frequency of Ca2� oscillations (Fig. 10 (Cand D) and Table I) are expected in the absence of DARPP-32,a condition that should favor PP1-mediated dephosphorylationof target proteins, including InsP3R1. The relatively mild ef-fects observed in MSN from DARPP-32 knock-out mice mostlikely result from the redundant functions of DARPP-32 andInhibitor-1, both of which are expressed in the striatum (65,66). The experiments with the competitive R9-IC peptide andEGFP-RT1-LIZ indicate that InsP3R1-associated PP1� (39)and InsP3R1�AKAP9-associated PKA (40) play a major role inthe control of the phosphorylated state of InsP3R1 (Fig. 9).Consistent with impaired PP1 function, MSN loaded with theR9-IC peptide displayed increased amplitude and frequency ofdopamine-induced Ca2� oscillations (Fig. 9 (A and B) and TableI). In contrast, disruption of the InsP3R1�AKAP9�PKA complexby overexpressed EGFP-RT1-LIZ resulted in delayed Ca2� os-cillations and reduced spike frequency (Fig. 9 (C and D) andTable I).

Previous studies suggested an important role of L-type Ca2�

channels, NMDA receptors, and AMPA receptors in the cross-talk between cAMP and Ca2� signaling in striatal neurons (16,17, 21–23). We also observed that maintenance of dopamine-induced Ca2� oscillations requires Ca2� influx from the extra-cellular space (Fig. 4A) and established a role of L-type Ca2�

channels, NMDA receptors, and AMPA receptors in mediatingCa2� influx (Fig. 4, C–F). If Ca2� influx is blocked, Ca2� oscil-lations can be initiated, but cannot be maintained (Fig. 4, B and

FIG. 10. Role of DARPP-32 in dopamine-induced Ca2� oscillations in MSN. The images were collected in ACSF containing 2 �M TTX andanalyzed as described in the Fig. 1 legend. A and B, Western blotting of striatal lysates from the wild-type (D32�/�) and DARPP-32 knock-out(D32�/�) pup brain lysates, respectively, with anti-DARPP-32 monoclonal antibodies. The lysate from each pup (lanes 1–6 for D32�/� and lanes1–7 for D32�/�) was loaded on a separate lane. The expected positions of DARPP-32 on the gels are indicated. C and D, representativedopamine-induced Ca2� oscillations in wild-type (D32�/�) and DARPP-32 knock-out (D32�/�) mouse MSN, respectively.


C). These results indicate that the most likely role of Ca2�

influx via L-type Ca2� channels, NMDA receptors, and AMPAreceptors is to prevent depletion of intracellular Ca2� stores.Indeed, application of the SERCA pump inhibitor thapsigarginresulted in rapid cessation of dopamine-induced Ca2� oscilla-tions (Fig. 5, C and D), indicating that the Ca2� stores in MSNcan be easily depleted. Similar conclusions have been reachedearlier for Ca2� oscillations induced by application of gonado-tropin-releasing hormone to cultured gonadotrophs (67). Thus,it appears that for both MSN and gonadotrophs, Ca2� oscilla-tions are driven by InsP3-mediated Ca2� release from the en-doplasmic reticulum, but that Ca2� influx via plasma mem-brane Ca2� channels is necessary for endoplasmic reticulumrefilling to maintain the oscillations.

In conclusion, using a Ca2� imaging technique, we havedescribed dopamine-induced Ca2� oscillatory responses inMSN. The dopamine-induced Ca2� oscillations are mediatedprimarily by D1 class DARs and require an intricate interplaybetween cAMP, InsP3, and Ca2� signaling pathways, whichconverge on InsP3R1 regulation (Fig. 11). Direct associationof InsP3R1 with PP1� (39) and formation of theInsP3R1�AKAP9�PKA complex (40) are important for fidelity ofcAMP, InsP3, and Ca2� cross-talk. Future experiments will beneeded to further test our model (Fig. 11) and to understandthe possible physiological relevance of dopamine-induced Ca2�

oscillations for striatal function.

Acknowledgments—We are grateful to Paul Greengard for providingDARPP-32 knock-out mice; to Paul Greengard, James Surmeier, andAnita Aperia for helpful discussions, experimental advice, and com-ments on the manuscript; and to Zhengnan Wang, Zheng Yan, andPhyllis Foley for expert technical and administrative assistance.

REFERENCES

1. Greengard, P., Allen, P. B., and Nairn, A. C. (1999) Neuron 23, 435–4472. Nicola, S. M., Surmeier, D. J., and Malenka, R. C. (2000) Annu. Rev. Neurosci.

23, 185–2153. Missale, C., Nash, S. R., Robinson, S. W., Jaber, M., and Caron, M. G. (1998)

Physiol. Rev. 78, 189–2254. Vallone, D., Picetti, R., and Borrelli, E. (2000) Neurosci. Biobehav. Rev. 24,

125–1325. Surmeier, D. J., Song, W. J., and Yan, Z. (1996) J. Neurosci. 16, 6579–65916. Jackson, D. M., and Westlind-Danielsson, A. (1994) Pharmacol. Ther. 64,

291–3707. Hernandez-Lopez, S., Tkatch, T., Perez-Garci, E., Galarraga, E., Bargas, J.,

Hamm, H., and Surmeier, D. J. (2000) J. Neurosci. 20, 8987–89958. Undie, A. S., and Friedman, E. (1990) J. Pharmacol. Exp. Ther. 253, 987–9929. Friedman, E., Jin, L. Q., Cai, G. P., Hollon, T. R., Drago, J., Sibley, D. R., and

Wang, H. Y. (1997) Mol. Pharmacol. 51, 6–1110. Panchalingam, S., and Undie, A. S. (2001) Neuropharmacology 40, 826–83711. Jin, L. Q., Goswami, S., Cai, G., Zhen, X., and Friedman, E. (2003) J. Neuro-

chem. 85, 378–38612. Lezcano, N., Mrzljak, L., Eubanks, S., Levenson, R., Goldman-Rakic, P., and

Bergson, C. (2000) Science 287, 1660–166413. Bergson, C., Levenson, R., Goldman-Rakic, P. S., and Lidow, M. S. (2003)

Trends Pharmacol. Sci. 24, 486–49214. Hernandez-Lopez, S., Bargas, J., Surmeier, D. J., Reyes, A., and Galarraga, E.

(1997) J. Neurosci. 17, 3334–334215. Surmeier, D. J., Bargas, J., Hemmings, H. C., Jr., Nairn, A. C., and Greengard,

P. (1995) Neuron 14, 385–39716. Cepeda, C., Colwell, C. S., Itri, J. N., Chandler, S. H., and Levine, M. S. (1998)

J. Neurophysiol. 79, 82–94

FIG. 11. Model of D1 class DAR-mediated cAMP, InsP3, and Ca2� signaling in MSN. Application of dopamine activates D1/D5 receptors,leading to activation of adenylyl cyclase, production of cAMP, and activation of PKA. Actions of dopamine can be mimicked by the D1receptor-specific agonist SKF38393 and prevented by the D1 receptor-specific antagonist SCH23390, but not by the D2 receptor-specific antagonistspiperone. Activated PKA phosphorylates DARPP-32 and Inhibitor-1, converting them into potent inhibitors of PP1 (1). Activation of PKA andinhibition of PP1 lead to increased phosphorylation and activation of Ca2� influx channels (L-type Ca2� channels (CaCh), NMDA receptors(NMDAR), and AMPA receptors (AMPAR)) and InsP3R1. InsP3R1 is associated with AKAP9�PKA via the LIZ motif (40) and with PP1� via theC-terminal IC peptide region (39). Dopamine also acts on a putative PLC-linked D1 class DAR (activated by SKF83959) (11) and/or a PLC-coupledD1/D5 receptor-calcyon complex (12). Formation of a D1/D5 receptor-calcyon complex is promoted by priming of MSN with 5-HT2 receptor agonist(not shown). It remains to be determined whether D1/D5 receptors complexed with calcyon (12, 13) and a putative PLC-linked D1 class DAR (8–11)are the same or distinct (as shown) molecular entities. Simultaneous activation of cAMP and InsP3 signaling pathways leads to activation ofInsP3R1 (39), which is under biphasic regulation by Ca2� (61–63). An increase in cytosolic Ca2� activates calcineurin (PP2B), which dephospho-rylates DARPP-32 and Inhibitor-1, resulting in disinhibition of PP1 (1), which in turn dephosphorylates InsP3R1 and Ca2� influx channels. Ca2�

influx via L-type Ca2� channels, NMDA receptors, and AMPA receptors and SERCA pump activity are necessary to prevent depletion ofintracellular Ca2� stores during oscillations. The proposed model indicates that dopamine-induced Ca2� oscillations in MSN rely on an intricateinterplay between Ca2� and cAMP signaling pathways, leading to repetitive rounds of InsP3R1 phosphorylation by PKA and dephosphorylationby PP1. The model is supported by pharmacological experiments (U73122, PLC blocker; calyculin A (CalA), PP1 blocker; cyclosporin A (CsA), PP2Bblocker; thapsigargin, SERCA blocker; (�)-MK801, NMDA receptor blocker; CNQX, AMPA receptor blocker; and nifedipine, L-type Ca2� channelblocker), effects of competitive R9-IC and EGFP-RT1-LIZ peptides, and analysis of DARPP-32 knock-out (KO) mice.


17. Yan, Z., Hsieh-Wilson, L., Feng, J., Tomizawa, K., Allen, P. B., Fienberg, A. A.,Nairn, A. C., and Greengard, P. (1999) Nat. Neurosci. 2, 13–17

18. Levine, M. S., Altemus, K. L., Cepeda, C., Cromwell, H. C., Crawford, C.,Ariano, M. A., Drago, J., Sibley, D. R., and Westphal, H. (1996) J. Neurosci.16, 5870–5882

19. Flores-Hernandez, J., Cepeda, C., Hernandez-Echeagaray, E., Calvert, C. R.,Jokel, E. S., Fienberg, A. A., Greengard, P., and Levine, M. S. (2002)J. Neurophysiol. 88, 3010–3020

20. Kotecha, S. A., Oak, J. N., Jackson, M. F., Perez, Y., Orser, B. A., Van Tol,H. H., and MacDonald, J. F. (2002) Neuron 35, 1111–1122

21. Blank, T., Nijholt, I., Teichert, U., Kugler, H., Behrsing, H., Fienberg, A.,Greengard, P., and Spiess, J. (1997) Proc. Natl. Acad. Sci. U. S. A. 94,14859–14864

22. Rajadhyaksha, A., Leveque, J., Macias, W., Barczak, A., and Konradi, C.(1998) Dev. Neurosci. 20, 204–215

23. Snyder, G. L., Fienberg, A. A., Huganir, R. L., and Greengard, P. (1998)J. Neurosci. 18, 10297–10303

24. Ouimet, C. C., Miller, P. E., Hemmings, H. C., Jr., Walaas, S. I., andGreengard, P. (1984) J. Neurosci. 4, 111–124

25. Walaas, S. I., and Greengard, P. (1984) J. Neurosci. 4, 84–9826. Fienberg, A. A., Hiroi, N., Mermelstein, P. G., Song, W., Snyder, G. L., Nishi,

A., Cheramy, A., O’Callaghan, J. P., Miller, D. B., Cole, D. G., Corbett, R.,Haile, C. N., Cooper, D. C., Onn, S. P., Grace, A. A., Ouimet, C. C., White,F. J., Hyman, S. E., Surmeier, D. J., Girault, J., Nestler, E. J., andGreengard, P. (1998) Science 281, 838–842

27. Furuichi, T., Kohda, K., Miyawaki, A., and Mikoshiba, K. (1994) Curr. Opin.Neurobiol. 4, 294–303

28. Berridge, M. J. (1998) Neuron 21, 13–2629. Walaas, S. I., Nairn, A. C., and Greengard, P. (1986) J. Neurosci. 6, 954–96130. Supattapone, S., Danoff, S. K., Theibert, A., Joseph, S. K., Steiner, J., and

Snyder, S. H. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, 8747–875031. Danoff, S. K., Ferris, C. D., Donath, C., Fischer, G. A., Munemitsu, S., Ullrich,

A., Snyder, S. H., and Ross, C. A. (1991) Proc. Natl. Acad. Sci. U. S. A. 88,2951–2955

32. Ferris, C. D., Cameron, A. M., Bredt, D. S., Huganir, R. L., and Snyder, S. S.(1991) Biochem. Biophys. Res. Commun. 175, 192–198

33. Haug, L. S., Jensen, V., Hvalby, O., Walaas, S. I., and Ostvold, A. C. (1999)J. Biol. Chem. 274, 7467–7473

34. Pieper, A. A., Brat, D. J., O’Hearn, E., Krug, D. K., Kaplin, A. I., Takahashi, K.,Greenberg, J. H., Ginty, D., Molliver, M. E., and Snyder, S. H. (2001)Neuroscience 102, 433–444

35. Wagner, L. E., II, Li, W. H., and Yule, D. I. (2003) J. Biol. Chem. 278,45811–45817

36. Soulsby, M. D., Alzayady, K., Xu, Q., and Wojcikiewicz, R. J. (2004) FEBS Lett.557, 181–184

37. Nakade, S., Rhee, S. K., Hamanaka, H., and Mikoshiba, K. (1994) J. Biol.Chem. 269, 6735–6742

38. Wojcikiewicz, R. J., and Luo, S. G. (1998) J. Biol. Chem. 273, 5670–567739. Tang, T.-S., Tu, H., Wang, Z., and Bezprozvanny, I. (2003) J. Neurosci. 23,

403–41540. Tu, H., Tang, T.-S., Wang, Z., and Bezprozvanny, I. (2004) J. Biol. Chem. 279,

19375–1938241. Fienberg, A. A., and Greengard, P. (2000) Brain Res. Brain Res. Rev. 31,

313–31942. Tang, T.-S., Tu, H., Chan, E. Y., Maximov, A., Wang, Z., Wellington, C. L.,

Hayden, M. R., and Bezprozvanny, I. (2003) Neuron 39, 227–23943. Howe, A. R., and Surmeier, D. J. (1995) J. Neurosci. 15, 458–46944. Goslin, K., Asmussen, H., and Banker, G. (1998) in Culturing Nerve Cells

(Banker, G., and Goslin, K., eds) pp. 339–370, MIT Press, Cambridge, MA45. Maximov, A., and Bezprozvanny, I. (2002) J. Neurosci. 22, 6939–695246. Mao, L., and Wang, J. Q. (2001) Brain Res. Mol. Brain Res. 86, 125–13747. Gerfen, C. R. (1992) Trends Neurosci. 15, 133–13948. Nisenbaum, E. S., and Wilson, C. J. (1995) J. Neurosci. 15, 4449–446349. Wilson, C. J., and Kawaguchi, Y. (1996) J. Neurosci. 16, 2397–241050. Hoyer, D., Hannon, J. P., and Martin, G. R. (2002) Pharmacol. Biochem.

Behav. 71, 533–55451. Harrison, J. K., Pearson, W. R., and Lynch, K. R. (1991) Trends Pharmacol.

Sci. 12, 62–6752. Caulfield, M. P. (1993) Pharmacol. Ther. 58, 319–37953. Cepeda, C., and Levine, M. S. (1998) Dev. Neurosci. 20, 1–1854. Morikawa, H., Khodakhah, K., and Williams, J. T. (2003) J. Neurosci. 23,

149–15755. Lezcano, N., and Bergson, C. (2002) J. Neurophysiol. 87, 2167–217556. Schwarze, S. R., Hruska, K. A., and Dowdy, S. F. (2000) Trends Cell Biol. 10,

290–29557. Becker-Hapak, M., McAllister, S. S., and Dowdy, S. F. (2001) Methods 24,

247–25658. Wender, P. A., Mitchell, D. J., Pattabiraman, K., Pelkey, E. T., Steinman, L.,

and Rothbard, J. B. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 13003–1300859. Walaas, S. I., Aswad, D. W., and Greengard, P. (1983) Nature 301, 69–7160. Pacheco, M. A., and Jope, R. S. (1997) J. Neurochem. 69, 639–64461. Iino, M. (1990) J. Gen. Physiol. 95, 1103–112262. Bezprozvanny, I., Watras, J., and Ehrlich, B. E. (1991) Nature 351, 751–75463. Finch, E. A., Turner, T. J., and Goldin, S. M. (1991) Science 252, 443–44664. Keizer, J., and Deyoung, G. W. (1992) Biophys. J. 61, 649–66065. Gustafson, E. L., and Greengard, P. (1990) Exp. Brain Res. 79, 447–45866. Gustafson, E. L., Girault, J. A., Hemmings, H. C., Jr., Nairn, A. C., and

Greengard, P. (1991) J. Comp. Neurol. 310, 170–18867. Stojilkovic, S. S., Kukuljan, M., Iida, T., Rojas, E., and Catt, K. J. (1992) Proc.

Natl. Acad. Sci. U. S. A. 89, 4081–4085


dopamine receptor-mediated ca2 signaling in striatal ... · dars are divided into the d1 class (d1r...

Documents