D1 dopamine receptor mediates dopamine-induced cytotoxicity via the ERK signal cascade.

Postsynaptic striatal neurodegeneration occurs through unknown mechanisms, but it is linked to high extracellular levels of synaptic dopamine. Dopamine-mediated cytotoxicity of striatal neurons occurs through two distinct pathways: autoxidation and the D1 dopamine receptor-linked signaling pathway. Here we investigated the mitogen-activated protein kinase (MAPK) signaling pathways activated upon the acute stimulation of D1 dopamine receptors. In SK-N-MC neuroblastoma cells, endogenously expressing D1 dopamine receptors, dopamine caused activation of phosphorylated (p-)ERK1/2 and of the stress-signaling kinases, p-JNK and p-p38 MAPK, in a time- and dose-dependent manner. Selective stimulation of D1 receptors with the agonist SKF R-38393 caused p-ERK1/2, but not p-JNK or p-p38 MAPK activation, in a manner sensitive to the receptor-selective antagonist SCH 23390, protein kinase A inhibition (KT5720), and MEK1/2 inhibition (U0126 or PD98059). Activation of ERK by D1 dopamine receptors resulted in oxidative stress and cytotoxicity. In cells transfected with a catalytically defective mutant of MEK1, the upstream ERK-specific kinase, both dopamine- and SKF R-38393-mediated cytotoxicity was markedly attenuated, confirming the participation of the ERK signaling pathway. Cell fractionation studies showed that only a small amount of p-ERK1/2 was translocated to the nucleus, with the majority retained in the cytoplasm. From coimmunoprecipitation studies, p-ERK was found to form stable heterotrimeric complexes with the D1 dopamine receptor and beta-arrestin2. In cells transfected with the dominant negative mutant of beta-arrestin2, the formation of such complexes was substantially inhibited. These data provide novel mechanistic insights into the role of ERK in the cytotoxicity mediated upon activation of the D1 dopamine receptor.

Striatal neurodegeneration is found in several human disorders involving postsynaptic dopamine neuronal dysfunction, such as the L-dopa-unresponsive parkinsonism subtype of multiple system atrophy (1), methamphetamine-induced neurotoxicity (2,3), and secondary dopamine dysfunction in Huntington's disease (4), in which dopaminergic transmission is interrupted by progressive loss of the striatal neurons bearing the postsynaptic D1 and D2 dopamine receptors (5). Although all of these diseases are characterized by elevated dopamine levels in the synapse, the underlying pathological mechanism(s) promoting striatal degeneration remains enigmatic. Recent studies suggest a role for the participation of both reactive oxygen species and reactive nitrogen species in the degeneration of striatal neurons (6 -9). Reactive oxygen species and reactive nitrogen species may be produced through autoxidation of extracellular dopamine in the synapse and may be especially important in the striatal degeneration seen in the L-dopa-unresponsive parkinsonism subtype of multiple system atrophy (1,10).
Dopamine in the synapse causes activation of the postsynaptic D1 and D2 subtypes of the dopamine receptor. Accumulating evidence in primates indicates that blockage of D1 dopamine receptors with selective antagonists has strong neuroprotective and anti-parkinsonian effects (11,12), supporting a coparticipatory role of the D1 receptors in the maintenance and/or pathogenesis of postsynaptic neurodegeneration. In both rat striatal primary cultures (13) and in SK-N-MC neuroblastoma cells (14), which endogenously express a functional D1 dopamine receptor (15), we have shown recently that the chronic treatment of these cells with dopamine promotes increased expression of the nitric-oxide synthases, neuronal nitric-oxide synthase and inducible nitric-oxide synthase, accompanied by increased nitric oxide production, oxidative stress, and cytotoxicity. These effects of dopamine were only partially mimicked by the oxidant H 2 O 2 and by direct stimulation of the D1 dopamine receptor with the selective agonist SKF R-38393. Dopamine effects were only partly blocked by the antioxidant sodium metabisulfite (SMBS) 1 and by the D1 receptor antagonist SCH 23390. However, together these compounds completely ablated the actions of dopamine (13,14). These data suggest that dopamine-mediated responses are mediated by at least two distinct and nonoverlapping pathways: through extracellular autoxidation of the compound and through chronic activation of the D1 dopamine receptor.
The current studies were undertaken to examine further the procytotoxic signaling pathways activated by D1 dopamine receptors. We examined the role of mitogen-activated protein kinases (MAPKs), the downstream mediators of signal transduction from the cell surface receptors to the nucleus. MAPKs have been implicated in a wide variety of cellular processes such as proliferation, differentiation, and apoptotic cell death. In mammals, three major groups of MAPKs have been identified: extracellular signal-regulated kinases 1 and 2 (ERK1/2) (16), p38 MAPK (17), and c-Jun N-terminal kinase (JNK) (18).
The ERK-linked signaling pathways are stimulated by receptor tyrosine kinases (19) and G protein-coupled receptors (GPCRs) (20) and generally lead to a mitogenic and proliferative response (21). JNK and p38 MAPK are stimulated by cellular stresses, such as free radicals and inflammatory agents, leading to apoptotic cell death (22). In the present studies, the activation of MAPKs was assessed by examining the expression levels and function of the phosphorylated MAPKs: p-JNK, p-p38 MAPK, and p-ERK1/2. Instead of observing an elevation of the archetypical proapoptosis MAPKs, p-JNK and p-p38 MAPK, we surprisingly found that D1 dopamine receptors selectively stimulated the ERK1/2 signaling cascade. Moreover, the activated p-ERK1/2 was primarily retained in the cytoplasm and formed a heterotrimeric complex with the D1 receptor and ␤-arrestin2, with only low levels seen in the nuclei, suggesting that the failure of p-ERK to be translocated into the nucleus may be linked to the subsequent cytotoxic response. Our findings underscore a novel and selective mechanism by which stimulation of D1 dopamine receptors promotes oxidative stress and cytotoxicity while providing a heuristic view in which to evaluate alternate roles of p-ERK1/2 in mediating these cellular responses.
Cell Culture, Drug Treatment, and Cell Viability-Human neuroblasto- ma-derived SK-N-MC cells were grown in 12-well culture plates (seeding density, 1.0 ϫ 10 5 /well) in RPMI 1640 medium without phenol red, supplemented with 10% (v/v) Nu-serum (Collaborative Biomedical Products, Bedford, MA), antibiotics, and 2 mM L-glutamine in a humidified atmosphere of 95% air, 5% CO 2 at 37°C until 90% confluent. Cells were then serum starved overnight with serum-free RPMI medium and incubated with various drugs for 16 h or the indicated times. Control cells were treated with an equal concentration of solvent (0.2% H 2 O). After incubation, 0.3 ml of medium was removed to measure nitrite concentration. Cell viability in 12-well plates was evaluated by counting viable cells in a Neubauer cell using the trypan blue exclusion test, as described by the manufacturer's (Sigma) protocol, whereby viable cells exposed to trypan blue for no more than 15 min exclude the dye. Values from each treatment were expressed as a percentage of cell survival relative to control.
Plasmids and Transfection-Wild type MEK1 and catalytically defective MEK1 mutant (MEK1-K97M) cDNA were kind gifts from M. Cobb, University of Texas Southwestern Medical Center, Dallas, TX. The MEK1-K97M cDNA construct was subcloned into vector pRSET (Invitrogen), and the wild type MEK1 cDNA construct was subcloned into vector pCMV5 (Invitrogen). ␤-Arrestin1 dominant negative mutant (␤-arrestin1 V53D) and ␤-arrestin2 dominant negative mutant (␤-ar-restin2 V54D) cDNA constructs (gifts from M. G. Caron, Duke University, Durham, NC) were subcloned into the mammalian expression vector pcDNA1AMP (Invitrogen). SK-N-MC neuroblastoma cells were transiently transfected (1 g of DNA/10 5 cells) by LipofectAMINE 2000 (Invitrogen). Briefly, cells were seeded in 12 wells (10 5 cells seeded/well) and grown to 80% confluence on the day of transfection. DNA-Lipo-fectAMINE 2000 was prepared as follows. Diluted DNA in Opti-MEM (without serum) was brought to a concentration of 1.6 g/100 l. Then diluted LipofectAMINE 2000 reagent in Opti-MEM (without serum) was brought to a concentration of 4.0 l/100 l, mixed gently, and incubated for 5 min at room temperature. After a 5-min incubation, the diluted DNA was combined with the diluted LipofectAMINE 2000, mixed gently, and incubated for 20 min at room temperature to allow DNA-LipofectAMINE 2000 complexes to form. 200 l of the DNA-LipofectAMINE 2000 complex was added to each well containing cells and Dulbecco's modified Eagle's medium (without serum and antibiotics), mixed gently, and incubated at 37°C and 5% CO 2 for 5 h. After the 5-h incubation, the medium was replaced with Dulbecco's modified Eagle's medium containing 20% fetal bovine serum and used 48 h after transfection. Mock transfected cells were transfected with a pcDNA3 plasmid (Invitrogen) that lacked a DNA insert.
Measurement of Nitrite Concentration-The nitrite concentration was measured as described previously (13,14). Briefly, 0.3 ml of Griess reagent (1 part of 0.1% naphthylethylenediamine dihydrochloride in H 2 O and 1 part of 1% sulfanilamide in 5% H 3 PO 4 ) and 0.3 ml of culture medium from treated cells (see above) were mixed. After a 30-min incubation at 45°C, the absorbance was determined at 550 nm. The nitrite concentration was measured from a standard curve using NaNO 2 at a range of 0 -10 M. Results were expressed as nM/1.0 ϫ 10 5 cells or -fold increase of the nitrite level from drug treatment over a distilled water-treated control.
Subcellular Fractionation-For the separation of nuclear and cytosolic pools of endogenous p-ERK1/2, serum-starved cells were stimulated with dopamine or SKF R-38393. Monolayers were washed twice with ice-cold phosphate-buffered saline and solubilized with 2 ml of ice-cold lysis buffer. Cells were incubated on ice for 10 min to allow lysis. The cellular extract was centrifuged at 500 ϫ g for 5 min to pellet nuclei. The supernatants contained the cytosolic fraction. Pellets containing cell nuclei were washed with lysis buffer and pelleted again at 500 ϫ g for 10 min. Both cytosolic and nuclear fractions were solubilized in 2 ϫ Laemmli loading buffer, and p-ERK1/2 levels were determined by immunoblotting. The purity of nuclear and cytosolic fractions was verified by immunoblotting with antibodies to Elk-1.
Analysis of Data-All statistical analyses values were accomplished using Instat Statistical Software (GraphPad, Sorrento Valley, CA). Statistical comparisons were performed with one-factor analysis of variance followed by Fisher's least squares difference test. Values of p Ͻ 0.05 were considered statistically significant. receptor selective agonist, SKF R-38393, causes profound cytotoxicity, as indexed by increased oxidative stress and accelerated cell death (14). SK-N-MC neuroblastoma cells, a postsynaptic striatal cell model endogenously expressing the D1 receptors (15,27), were treated with increasing concentrations of dopamine in the presence of SCH 23390, a D1-selective antagonist, or SMBS, a potent antioxidant. In the presence of either of these two compounds, the cytotoxic effects of dopamine were reduced sharply by ϳ50%, as indexed by a decrease in both nitrite production (Fig. 1A) and cell death (Fig. 1B). When SCH 23390 and SMBS were used together, the cytotoxic effects of dopamine were completely prevented, indicating that both autoxidation of dopamine and direct activation of the D1 receptors by dopamine contribute to mediation of dopamine cytotoxicity.

Activation of D1 Dopamine Receptors Enhances Oxidative
The direct participation of D1 dopamine receptors in inducing cytotoxicity was tested by stimulating the receptor with the  (Fig. 1C) and cell death (Fig. 1D). SMBS did not modulate the SKF R-38393mediated effects, consistent with lack of oxidation of this stable compound. By contrast, SCH 23390 attenuated the effects of SKF R-38393 with regard to both nitrite production and cell death.
Dopamine Enhances Phosphorylation of the ERK1/2 MAP Kinases-To gain insight into the mechanistic pathways by which dopamine mediates its toxic effects, we analyzed the early signaling cascades that are activated by examining the phosphorylation of the three major groups of MAP kinases: ERK1/2, JNK, and p38 MAPK, in SK-N-MC cells treated with dopamine or H 2 O 2 . Dopamine caused a strong enhancement of p-ERK1/2 in SK-N-MC cells in a dose-dependent manner, and an increase of 116 Ϯ 14.9% was observed at the highest concentration of dopamine (100 M) used in these studies ( Fig. 2A). By contrast, H 2 O 2 enhancement of p-ERK1/2 was weak, and even at 100 M, a modest increase of only 33% was seen. On the other hand, dopamine caused only a weak enhancement (33%) of p-JNK, whereas H 2 O 2 increased p-JNK levels by 65% (Fig.   2B). Dopamine and H 2 O 2 activated p-p38 MAPK to similar levels, and increases of 143 and 133%, respectively, were observed at concentrations of 100 M (Fig. 2C).
The time course of activation of p-ERK1/2, p-JNK, and p-p38 MAPK by 100 M dopamine was investigated. Maximal activation of p-ERK1/2 and p-p38 MAPK was achieved within 45 min of exposure to dopamine (Fig. 2D), with a slight decrease after 60 and 120 min of exposure. The weak activation of p-JNK by dopamine was seen only after 60 min and was sustained up to 120 min of exposure to dopamine. These data indicate that dopamine strongly enhances the phosphorylation of the ERK1/2 and p38 MAPKs, with weak activation of JNK.
D1 Dopamine Receptor-mediated Cytotoxicity Proceeds through the p-ERK1/2 Signaling Pathway-Because the ability of dopamine to enhance phosphorylation of ERK1/2 was much higher than that of H 2 O 2 , we speculated that the increase in p-ERK1/2 levels was not the result of the oxidant properties of these compounds, but was a direct consequence of D1 dopamine receptor stimulation by dopamine. To test for this, we examined p-ERK1/2, p-JNK, and p-p38 MAPK levels after stimulating SK-N-MC cells with the D1 receptor-selective agonist, SKF R-38393. Treatment of cells with increasing doses of SKF FIG. 5-continued R-38393 selectively enhanced p-ERK1/2 levels (Fig. 3A), with no corresponding increase in either p-JNK or p-p38 MAPK levels (Fig. 3B). Moreover, the increase in p-ERK1/2 was blocked by the D1-selective antagonist SCH 23390, protein kinase A inhibitor KT5720, and by U0126, a selective inhibitor of MEK1/2 (Fig. 3A).
One feature of the ERK signaling cascade is that ERK1/2 phosphorylation is exclusively regulated by MEK1/2, the upstream dual specificity kinase. This dual phosphorylation is necessary and sufficient for ERK activation. We therefore also measured p-MEK1/2 levels by stimulating SK-N-MC cells with SKF R-38393 and found a dose-dependent increase in p-MEK1/2 which was blocked completely by SCH 23390 (Fig. 3C).
To assess further the participation p-ERK1/2 in mediating dopamine and D1 receptor cytotoxicity, SK-N-MC cells were treated for 16 h with dopamine or SKF R-38393 (50 M each) in the presence or absence of selective inhibitors of the various MAPKs, and nitrite levels and cell viability were measured. Cells were also treated with 50 M H 2 O 2 to measure separately the role of MAPKs in the oxidative pathway. 10 M SB203580, a selective inhibitors of p38 MAPK, reduced nitrite production by both dopamine (47%) and H 2 O 2 (62%), with a significant (p Ͻ 0.05) increase in cell viability (Fig. 4A). When cells were treated with SKF R-38393, SB203580 caused a modest decrease (21%) in nitrite production, without any significant changes in cell viability (Fig. 4A). By contrast, inhibition of p-JNK activity with its selective inhibitor SB600125 had virtually no significant effect on SKF R-38393-mediated nitrite production or cell death, while significantly reducing nitrite levels produced by dopamine and H 2 O 2 (29 and 44%, respectively) (Fig. 4B).
U0126 and PD98059, another selective inhibitor of MEK1/2 (10 M each), blocked (41 and 48%, respectively) dopaminemediated nitrite production, causing a significant (p Ͻ 0.05) increase in cell viability (Fig. 4, C and D). Moreover, both U0126 and PD98059 almost completely blocked an SKF R-38393-mediated increase in nitrite production (by 86 and 84%, respectively), while significantly (p Ͻ 0.05) restoring cell viability to near control levels (Fig. 4, C and D). By contrast, H 2 O 2 -induced nitrite production was much less affected by the presence of the MEK1/2 inhibitors, with inhibition of only 15-22% with U0126 and PD98059, respectively, and without significant changes in cell viability (Fig. 4, C and D). Together, these data suggest that the cytotoxicity observed is the result of the direct stimulation of D1 receptors and is dependent on activation of p-ERK1/2, but not p-JNK, with only a modest involvement of p-p38 MAPK in these events.
D1 Dopamine Receptor-mediated Cytotoxicity Involves MEK1-To confirm the involvement of MEK1 in D1 dopamine receptor-mediated oxidative stress and cytotoxicity, we transfected SK-N-MC cells with wild type MEK1 and its catalytically defective MEK1 mutant, MEK1-K97M. Compared with wild type MEK1-transfected SK-N-MC cells, MEK1-K97Mtransfected cells had decreased nitrite production and cell death upon treatment with both dopamine and SKF R-38393 ( Fig. 5A) (38 and 56%, respectively; p Ͻ 0.05), without affecting H 2 O 2 -induced nitrite production and cytotoxicity (Fig. 5A). To ascertain whether MEK1 was involved in the direct receptor stimulation pathway of dopamine action or through production of free radicals, MEK1-K97M-transfected cells were treated with dopamine (Fig. 5B), SKF R-38393 (Fig. 5C), and H 2 O 2 (Fig. 5D) in the absence or presence of SMBS and SCH 23390. In wild type MEK1-transfected cells, dopamine-induced nitrite production and cytotoxicity were significantly (p Ͻ 0.05) and equally ablated by both SMBS and SCH 23390 (Fig. 5B). How-ever, in MEK1-K97M-transfected cells, dopamine-induced nitrite production and cytotoxicity were only blocked upon cotreatment with the antioxidant SMBS, but not by the D1 receptor-selective antagonist SCH 23390 (Fig. 5B). Meanwhile, in wild type MEK1-transfected cells, 50 M SKF R-38393 caused a significant (p Ͻ 0.05) increase in nitrite production and cytotoxicity, which were blocked by the D1 receptor-selective antagonist SCH 23390, but not by SMBS (Fig. 5C). In contrast to MEK1-transfected cells, SKF R-38393 had no significant effect on either nitrite production or cytotoxicity in MEK1-K97M-transfected SK-N-MC cells (Fig. 5C). On the other hand, in both wild type MEK1-and MEK1-K97M-transfected SK-N-MC cells, H 2 O 2 -induced nitrite production and cytotoxicity were only blocked by cotreatment with the antioxidant SMBS (Fig. 5D), but not by the D1 receptor-selective antagonist SCH 23390 (Fig. 5D). These results are consistent with our finding that MEK1/2 is selectively activated by D1 dopamine receptors but not through the oxidative stress pathway of dopamine autoxidation.
We examined the p-ERK1/2 expression levels after transfecting cells with wild type MEK1 and the MEK1 mutant. In wild type MEK1-transfected SK-N-MC cells, dopamine and SKF R-38393 caused a strong enhancement of p-ERK1/2 in a dosedependent manner. Maximal increases of 172.7 Ϯ 17.6% and 94.7 Ϯ 13.5% were observed at the highest concentration of dopamine (100 M) and SKF R-38393 (100 M), respectively, used in these studies (Fig. 5, E and F). However, in MEK1-K97M-transfected cells, dopamine-and SKF R-38393-induced p-ERK1/2 expressions were significantly (p Ͻ 0.05) reduced by about 45 and 73% at the highest concentration of dopamine (100 M) and SKF R-38393 (100 M), respectively (Fig. 5, E and  F). In addition, total MEK1/2 expression was detected with no substantial difference in wild type MEK1-and MEK1-K97Mtransfected cells, indicating that the reduction in p-ERK1/2 levels was not caused by changes in MEK1/2 protein levels (Fig.  5, E and F). These data clearly indicate the participation of MEK1/2 and thereby the p-ERK1/2 pathway upon activation of the D1 dopamine receptor.
p-ERK Is Retained in the Cytoplasm in a Heterotrimeric Complex-p-ERK elicits a wide range of biological functions through phosphorylation of nuclear and cytoplasmic substrates (28), and the translocation of ERK from the cytoplasm to the nucleus is mandatory for efficient regulation of cell survival and proliferation (29,30). Because our results showed paradoxically that p-ERK promoted a cytotoxic response, we conducted studies to determine whether p-ERK was translocated in an appropriate manner into the nucleus. Accordingly, SK-N-MC cells were treated with either dopamine or SKF R-38393 in a dose-and time-dependent manner, the cells fractionated into cytoplasm and nucleus fractions, and levels of p-ERK1/2 in these fractions were assessed by Western blots. Dopamine and SKF R-38393 (100 M each) caused a time-dependent increase in p-ERK1/2 levels in both the cytoplasm and the nuclei fractions, reaching maximal levels in 60 min (Fig. 6A). However, the relative increase in p-ERK1/2 in the nuclei was significantly (p Ͻ 0.05) below that seen in the cytoplasm at all times. Only 35% of p-ERK1/2 was translocated to the nuclei, whereas 65% of p-ERK1/2 was retained in the cytoplasm after a 60-min treatment with dopamine or SKF R-38393 (Fig. 6A). Similarly, both dopamine and SKF R-38393 caused an increase in p-ERK1/2 in a dose-dependent manner, but there were clearly much higher levels (ϳ65%) of p-ERK1/2 in the cytoplasm compared with the nuclei (Fig. 6B). As a positive control, SK-N-MC cells were treated with forskolin, a cAMP-elevating agent, which caused a dose-dependent increase in p-ERK1/2 levels. In contrast to our findings with dopamine and SKF R-38393, in forskolin-treated cells 72% of p-ERK1/2 was translocated to the nuclei, and only 28% of p-ERK1/2 was retained in the cytoplasm after a 30-min 10 M forskolin treatment (Fig. 6C).
Upon activation, GPCRs are phosphorylated and then coupled to ␤-arrestin, resulting in endocytosis of the p-GPCR. Once internalized, ␤-arrestin acts as a scaffolding protein, binding to p-ERK to form a stable complex, leading to the cytosolic retention of the p-GPCR (31)(32)(33). We wondered whether stimulation of D1 receptors also results in formation of such heterotrimeric complexes comprising the D1 dopamine receptor, ␤-arrestin, and p-ERK. Using selective antibodies, we found that these SK-N-MC cells express only the ␤-arrestin2 protein and not the ␤-arrestin1 subtype (Fig. 7A). We next treated SK-N-MC cells with 100 M dopamine or 100 M SKF R-38393 for different time periods and conducted coimmunoprecipitation studies using antibodies against pan-␤-arrestin. p-ERK1/2 was found to be associated with ␤-arrestin2 within 60 min of treatment with the agonist, followed by a decrease at 120 min of treatment (Fig. 7B). This result is consistent with the time course of dopamine-induced p-ERK1/2 expression (Fig. 2D) and with the time course of dopamine-and SKF R-38393-induced cytosolic retention of p-ERK (Fig. 6A). In all immunopellets, precipitated ␤-arrestin2 was detected at similar levels (Fig. 7B).
To verify the participation of ␤-arrestin2 in the formation of the heterotrimeric complex, we examined the formation of this complex in SK-N-MC cells transfected with a dominant nega-tive mutant of ␤-arrestin2 (␤-arrestin2 V54D), ␤-arrestin1 dominant negative mutant (␤-arrestin1 V53D), or mock (pcDNA3). In mock and ␤-arrestin1 V53D-transfected cells, the formation of the heterotrimeric complex was increased substantially after 100 M SKF R-38393 and 100 M dopamine treatment as shown in coimmunoprecipitation studies (Fig. 8,  A-C); whereas, in ␤-arrestin2 V54D-transfected cells, the formation of the heterotrimeric complex induced by 100 M SKF R-38393 and 100 M dopamine was ablated (Fig. 8, A-C). This is in consistent with the finding that stimulation of D1 recep- tors with agonists forms a stable heterotrimeric protein complex consisting of D1 dopamine receptors, ␤-arrestin2, and p-ERK1/2, and ␤-arrestin2 acts as a scaffolding protein. DISCUSSION The MAPKs constitute an important family of signaling proteins (17,34) that regulate neuronal survival and death, proliferation, differentiation, and plasticity. In particular, ERK is involved in neuronal development, memory formation, survival, and adaption (16,(35)(36)(37)(38)(39). Yet, recent studies suggest that activation of the ERK signaling cascade may also induce a cytotoxic neuronal response to stress (40 -47), suggesting that ERK signaling may have an alternate function mediating neurotoxic events.
The results presented here provide evidence for the selective participation of ERK in the early events underlying the genesis of striatal cytotoxicity upon stimulation of D1 dopamine receptors. Thus, instead of observing D1 receptor-mediated increases in levels of p-p38 MAPK and p-JNK, the prototypical MAPKs activated in response to oxidative stress, we paradoxically observed a selective increase in p-ERK1/2. The increase in p-ERK1/2 expression was blocked by the D1-selective antag-onist, and when SK-N-MC cells were treated with dopamine in the presence of SMBS, the antioxidant failed to affect p-ERK1/2 levels (data not shown). However, inhibition of MEK1/2 by U0126 blocked the SKF R-38393-mediated increase in p-ERK1/2 levels. Moreover, suppression of p-ERK1/2 expression with catalytically defective MEK1 mutant MEK1-K97M also blocked the SKF R-38393-and the dopamine-induced increases in p-ERK1/2 levels and relieved the D1 dopamine receptor-mediated cytotoxicity. Together, these combined data indicate that stimulation of the D1 dopamine receptor leads to p-ERK1/2 activation.
Dopamine, however, appears to activate other MAPKs, in addition to p-ERK1/2. Thus, dopamine also caused a robust increase in p-p38 MAPK, with a modest increase in p-JNK levels, and in the presence of SB203580 or SB600125, nitrite production was significantly attenuated, with partial restoration of cell viability. Moreover, the p-ERK1/2 inhibitors and MEK1-K97M and also caused only partial blockade of nitrite production and cell death. Interestingly, none of these effects was observed with the D1 dopamine receptor-selective agonist SKF R-38393. These results indicate that dopamine-mediated  cytotoxicity occurs, at least in part, through the p38 MAPK and JNK, in addition to the ERK, signaling cascades. These findings are in congruence with those obtained using H 2 O 2 , which also increased the levels of both p38 MAPK and p-JNK but caused only very modest changes in p-ERK1/2. Inhibition of these kinases, p38 MAPK and p-JNK, also caused a large reduction in nitrite levels and in cell death, after treatment with H 2 O 2 . Together, these findings suggest that the oxidative stress pathway induced by both dopamine and H 2 O 2 occurs through the activation of the stress-linked MAPKs, p38 MAPK and JNK. A schematic representation summarizing our findings is presented in Fig. 9, where we show how dopamine can activate the various MAPKs through two distinct pathways: the direct stimulation of the D1 dopamine receptors and through an oxidative pathway, arising from its autoxidation. The D1 dopamine receptors activate p-ERK1/2, whereas the oxidative pathway predominantly activates p-p38 MAPK and p-JNK. However, the possibility that some kinases of the autoxidative pathway of dopamine also activate p-ERK1/2 cannot be entirely eliminated. Indeed, some studies suggest that MEKK1 (MAPK kinase kinase kinase 1), which functions as a stress-activated protein kinase, primarily activating JNK and p38 MAPK, may also induce modest ERK activation (48 -50). Thus, our studies do not entirely eliminate the possibility that some coparticipation of the oxidative pathway of dopamine, activating through MEKK1, could also cause modest activation of ERK (Fig. 9).
Activation of the ERK cascade by dopamine receptors is a complex process that may involve multiple mechanisms (51)(52)(53)(54). D1 dopamine receptors are known to exert their physiological actions by stimulating cellular adenylyl cyclase via G s (55). Recent studies show that activation of ERK1/2 occurs via a G s /adenylyl cyclase/protein kinase A-dependent pathway that causes activation of the small GTPase, Rap1 (56,57). Our results also indicate the participation of the cAMP/protein kinase A-dependent pathway in these processes, because inhibition of protein kinase A completely blocked D1 receptorinduced ERK activation in SK-N-MC cells.
In resting cells, ERK is kept in the cytoplasm by the microtubule cytoskeleton, which serves as a major docking matrix for up to 35% of cellular ERK1 and ERK2 (58). Other putative cytosol-anchoring proteins of ERK are the MAP kinase phosphatases (MKPs), in particular MKP3 (29,59). ERK is also retained in the cytosol by its association with MEK1/2 (60). After phosphorylation by MEK1/2, p-ERK1/2 is detached from this cytosolic anchor and rapidly translocated into the nucleus (61), in a process that requires dimerization and phosphate incorporation into the regulatory Thr and Tyr residues of at least one of the ERKs in the dimer (62). Several sites on ERK have been identified to be important in both its nuclear translocation (residues 321-327) and its cytosolic retention (residues 312-320), with the three acidic residues Asp-316, Asp-319, and Glu-320 being especially important (63). Once in the nucleus, p-ERK1/2 can phosphorylate substrates, such as Elk-1 (64), thereby transmitting signals originally received by cell surface receptors to the nucleus.
In our studies, we found that the majority of p-ERK1/2 was retained in the cytoplasm, with only a modest amount translocated into the nucleus. The failure of p-ERK1/2 to translocate into the nucleus, coupled with its retention in the cytoplasm, provides some evidence for a possible mechanism by which ERK activation may trigger a cytotoxic, rather than a mitogenic, response upon activation of the D1 dopamine receptor. In our coimmunoprecipitation studies, we have shown that p-ERK1/2 forms a stable heterotrimeric complex with the D1 dopamine receptor and ␤-arrestin2. The arrestin proteins enable the internalization and trafficking of certain phosphorylated GPCRs, including the D1 dopamine receptor (65), away from the plasma membrane, terminating receptor-dependent signals by precluding receptor-G protein coupling (66). Blockade of the expression of ␤-arrestin2 with its dominant negative mutant almost completely prevented the formation of the heterotrimeric complex, while simultaneously blocking the phosphorylation of pERK1/2.
A similar observation has also been made for a limited number of GPCRs. Thus, activation of the protease-activated receptor type 2 (67), vasopressin V2 (68), and the angiotensin AT1a receptor (31) leads to the cytosolic retention of p-ERK, with low transcription activity and lack of a mitogenic response. In all of these studies, the cytosolic pool of p-ERK was shown to form stable complexes with the activated receptors and ␤-arrestin (69). In other studies with vasopressin V2 receptors, it was shown that the stability of the receptor-␤-arrestin complex, and not the specificity of the G protein coupling, was paramount in determining the subcellular distribution of ERK and its retention in the cytosol (31,70). Interestingly, in human embryonic kidney-293 cells, the rat variant of D1 dopamine receptors was found to form only transient complexes with ␤-arrestin, whereas both the AT1a and the vasopressin V2 receptors were shown to form stable complexes (32). Yet, our data clearly show that at least in SK-N-MC cells, the D1 dopamine receptor is able to form stable complexes with both ␤-arrestin2 and p-ERK1/2.
If abnormally retained in the cytoplasm, ␤-arrestin-bound p-ERK1/2 may phosphorylate multiple plasma membrane, cytoplasmic, and cytoskeletal substrates (34), resulting in altered, and perhaps inappropriate, series of cellular events. The significance of cytosolic retention of p-ERK1/2 is underscored by recent findings in postmortem tissue of neurodegenerative diseases, where neuronal inclusion bodies were found to contain substantially high levels of aggregated p-ERK1/2. Thus, from immunohistochemical analyses, granular precipitates of p-ERK1/2 were seen in the cytoplasm of neurons exhibiting early tau deposition in Alzheimer's disease, in neurons with Pick bodies in Pick's disease, and in neurons in progressive supranuclear palsy and corticobasal degeneration (71). Moreover, p-ERK1/2, but not p-p38 MAPK or p-JNK, was found in Lewy bodies in Parkinson's disease and in dementia with Lewy bodies (72), which were shown to be located primarily in the cytoplasm and not in the nucleus (73). Therefore, elucidation of the precise mechanisms that cause cytosolic retention of p-ERK1/2 is likely to provide important insights into the etiology of multiple neurodegenerative diseases.