Identification of tyrosine hydroxylase as a physiological substrate for Cdk5

Abstract

Cyclin-dependent kinase 5 (Cdk5) is emerging as a neuronal protein kinase involved in multiple aspects of neurotransmission in both post- and presynaptic compartments. Within the reward/motor circuitry of the basal ganglia, Cdk5 regulates dopamine neurotransmission via phosphorylation of the postsynaptic signal transduction pathway integrator, DARPP-32 (dopamine- and cyclic AMP-regulated phosphoprotein, M(r) 32,000). Cdk5 has also been implicated in regulating various steps in the presynaptic vesicle cycle. Here we report that Cdk5 phosphorylates tyrosine hydroxylase (TH), the key enzyme for synthesis of dopamine. Using phosphopeptide mapping, site-directed mutagenesis, and phosphorylation state-specific antibodies, the site was identified as Ser31, a previously defined extracellular signal-regulated kinases 1/2 (ERK1/2) site. The phosphorylation of Ser31 by Cdk5 versus ERK1/2 was investigated in intact mouse striatal tissue using a pharmacological approach. The results indicated that Cdk5 phosphorylates TH directly and also regulates ERK1/2-dependent phosphorylation of TH through the phosphorylation of mitogen-activated protein kinase kinase 1 (MEK1). Finally, phospho-Ser31 TH levels were increased in dopaminergic neurons of rats trained to chronically self-administer cocaine. These results demonstrate direct and indirect regulation of the phosphorylation state of a Cdk5/ERK1/2 site on TH and suggest a role for these pathways in the neuroadaptive changes associated with chronic cocaine exposure.

Fig. 1

Phosphorylation of TH by Cdk5 and phosphoamino acid and phosphopeptide map analysis. (a) Schematic diagram of TH with sites of in situ phosphorylation indicated (yellow) along with protein kinases. Positions of all serine residues followed by proline are indicated in red and labeled, except for Ser31. (b) Time-course of in vitro phosphorylation reaction. Radiolabeled (P-TH) and Coomassie stained (TH) TH phosphorylated by Cdk5 and quantified stoichiometry (bottom). (c) Phosphoamino acid analysis of ³²P-labeled TH phosphorylated by Cdk5 is shown, top left. The positions of comigrating phosphoserine, phosphothreonine and phosphotyrosine standards and the origin are indicated. Tryptic phosphopeptide maps of TH phosphorylated in vitro with the indicated protein kinases are shown in the four larger panels. The positions of comigrating phosphopeptides are indicated by numbers. The positions of the origins and phenol red markers are also indicated.

Fig. 2

Identification of Ser31 as the major site on TH phosphorylated by Cdk5 by site-directed mutagenesis and phosphorylation state-specific antibodies. (a) Time-course of in vitro phosphorylation of wild-type versus S31A TH by Cdk5. Coomassie stained and ³²P-labeled TH are shown in the top two panels with quantification of stoichiometry in the lower panel. (b) Immunoblots of reaction mixtures from in vitro phosphorylation of TH with Cdk5, ERK1/2, and PKA stopped at the indicated time points. Detection of phospho-Ser31, phospho-Ser40 and total TH are shown in the top, middle, and bottom panels, respectively.

Fig. 3

Intramolecular effects of phosphorylation at Ser40 and phospho-mimetic mutations at Ser40 and 19 on phosphorylation of TH by Cdk5. (a) Histogram depicting quantitation of V/K-values for phosphorylation of the indicated forms of TH by Cdk5. Error bars (SEM) were derived from two separate determinations for each reaction. (b) Quantitative kinetic values for phosphorylation of the indicated forms of TH by Cdk5. *nd, not determined as described in the text.

Fig. 4

Location of TH in the presynaptic compartment and phosphorylation by Cdk5 in the striatum. (a) Mouse striatal tissue was immunostained for TH (red) and postsynaptic DARPP-32 (green), and fluorescent images were collected by confocal laser scanning microscopy. Immunoreactivity for DARPP-32 occurs in somata and dendrites of caudatoputamen. In contrast, TH immunoreactivity is localized to neuropili and neural processes originating from nigral tracts. No overlapping signal (yellow) was detected in any images examined. Scale bar represents 25 μm. (b) Cdk5 phosphorylates TH in striatum. Striatal slices were incubated for 60 min with the Cdk5 inhibitor, roscovitine at the indicated concentrations. Homogenates were subjected to SDS-PAGE and immunoblotted for detection of phospho-Ser31 and total TH. Quantitation of phopho-Ser31 from multiple experiments are shown, *p < 0.05, Student's unpaired t-test, n = 4

Fig. 5

Cdk5 both directly phosphorylates Ser31 TH in striatum and regulates its phosphorylation at the same site by ERK1/2. (a) Effect of Cdk5 and MEK1 inhibitors on phospho-Ser31. Striatal slices were incubated for 60 min with roscovitine (Ros), the MEK1 inhibitor U0126 or both. Homogenates were immunoblotted for phospho-Ser31 or total TH. Quantitation of multiple experiments is shown. *p = 0.0069 versus control, **p < 0.001 versus control, †p < 0.02 versus roscovitine, §p < 0.04 versus U0126, Student's unpaired t-test, n = 4–6. (b) Effects of Cdk5 and MEK1 inhibitors on the phosphorylation states of ERK1 and 2. The same membrane used in (a) was also blotted for phospho-Thr202/phospho-Tyr204 levels and total ERK2. ERK1 was not detectable with the total ERK antibody. *p < 0.05 versus control, †p < 0.05 versus P-ERK1, Student's unpaired t-test, n = 4. (c) Phosphorylation of MEK1 by Cdk5 in vitro. Either inactive or active MEK1 was phosphorylated by Cdk5 or incubated in the absence of Cdk5 for 60 min and reaction mixtures were subjected to SDS–PAGE, the gel was Coomassie stained, and the radiographic image was derived using a PhosphorImager.

Fig. 6

The effect of KCl-induced membrane depolarization on TH Ser31 and ERK1/2 phosphorylation. Striatal slices were treated with roscovitine and/or U0126, followed by treatment with 40 mm KCl for 5 min. (a) Striatal slice homogenates were immunoblotted for phospho-Ser31 and total TH. Quantitation of multiple experiments is shown. *p < 0.001 versus control, †p < 0.05, ††p < 0.01 versus KCl, §p < 0.001 versus KCl + U0126, anova and Newman-Keuls test, n = 4. (b) Homogenates were immunoblotted for phospho-ERK1/2 and total ERK2. The ERK2 protein band appears as a doublet in two of the lanes in the total ERK2 blot, where ERK2 was highly activated by KCl. *p < 0.001 versus controls, anova and Newman-Keuls test, n = 4.

Fig. 7

Effect of cocaine self-administration on phospho-Ser31 TH levels following withdrawal. Mean (± SEM) percentage change from naïve rats at each withdrawal period. *p = 0.04, one-way anova (n = 6–8).