Long- and Short-Term Regulation of Tyrosine Hydroxylase

Menek Goldstein

Tyrosine hydroxylase (TH) (EC 1.14.16.2) catalyzes the enzymatic conversion of L-tyrosine to L-3,4-dihydroxyphenylalanine (L-dopa), the first step in the biosynthesis of dopamine (DA), norepinephrine (NE), and epinephrine (E). TH is subject to short- and long-term regulation; the latter occurs at transcriptional as well as at translational levels, and the former occurs at post-translational levels. The findings that dibutyryl cyclic AMP (dB-cAMP) stimulates L-dopa biosynthesis from L-tyrosine in striatal slices (4), and that phosphorylation of TH mediated by cAMP-dependent protein kinase A (PKA) phosphorylates and activates TH in vitro, have provided the first evidence that protein kinase-mediated phosphorylation plays a role in regulation of the enzyme activity and catecholamine biosynthesis. Subsequently it was shown that various protein kinases catalyze phosphorylation of TH at multiple serine sites of the enzyme (10), 45). The findings that long- and short-term regulation of TH activity at transcriptional, translational, and post-translational levels, respectively, involve similar second messenger signals raises the possibility that these pathways are interdependent.

Many advances in the mechanisms involving regulation of TH gene and enzyme activity have been made in the last few years, and we will review only some of them. We will focus on areas that have recently made rapid progress and discuss strategies to be pursued in future research.

The long-term regulation of TH occurs at transcriptional and translational levels. Several studies have shown that cAMP and glucocorticoids regulate TH mRNA levels primarily by stimulating the transcription rate of TH gene. Both elevate TH protein as well as mRNA and gene transcription rate in cultured PC12 cells and in adrenal glands (38, 39, 56, 57). The 5¢ flanking region of the TH gene possesses the same response elements which mediate the regulation of several other genes by cAMP and glucocorticoids (39). The absence of the cAMP response element in the TH gene promoter region results in a loss of response to elevated cAMP (39, 56). The transcription rate of the TH gene as estimated from the run-off assays was found to be stimulated 5- to 20-fold by treatment of PC18 cells (a variant of PC12 cells) with cAMP analogues or dexamethasone (14). The pulse-labeling experiments with 4-thiouridine have shown that treatment with a cAMP analogue or dexamethasone increases the rate of TH mRNA synthesis. However, the enhanced stability may also, in part, contribute to the increased steady-state TH mRNA levels elicited by cAMP (14). Because protein kinase C (PKC) seems to mediate both short- and long-term Ca2+-dependent cellular functions, its involvement in the regulation of TH at transcriptional and post-transcriptional levels was investigated (10). Activation of PKC by the phorbol ester, TPA, leads to an increase of TH gene transcription, as well as to the post-translational modulation of TH gene expression in PC12 cells (61). Thus, PKA and PKC are involved in transcriptional regulation of TH, and both pathways may converge to amplify TH expression and activity in some physiological and/or pathological states.

The TH gene regulation in dopaminergic (DA) and noradrenergic (NE) neurons has been the subject of numerous studies. It became apparent that regulation of TH activity and TH enzyme protein in NE neurons differs from that in DA neurons. In the central nervous system (CNS) the reserpine effect on TH activity is restricted to NE neurons, and it does not have an effect on TH activity in the DA neurons of the substantia nigra (51). Nevertheless, one study reports that reserpine slightly increased TH protein levels without affecting the enzymatic activity in the DA neurons of substantia nigra (35). The time course of the induction of TH by reserpine was analyzed in adrenals, locus coeruleus, and substantia nigra (13). Reserpine caused in locus coeruleus and adrenals a significant increase in TH mRNA and in TH activity, while no effect was observed in the substantia nigra (13). The exposure of rats to chronic stress increases the levels of TH mRNA and TH protein in the locus coeruleus (44), 52), but it does not alter TH in the substantia nigra or ventral tegmentum (44). It is noteworthy that a single study reports that isolation stress results in a small transient increase in TH mRNA in both the substantia nigra and ventral tegmental area (VTA); the magnitude of increase is much smaller than in locus coeruleus (5). Repeated administration of morphine increases TH mRNA levels in the locus coeruleus but not in the substantia nigra (22), while chronic treatment with antidepressants decreases the firing rate of NE neurons as well as the steady-state TH mRNA levels and TH protein in locus coeruleus but not in the midbrain DA neurons (44, 52). Because stress activates the hypothalamic–pituitary–adrenal axis, which results in increased levels of adrenocorticotropic hormone (ACTH) and corticotropin-releasing factor (CRF), it was suggested that these neuropeptides might be involved in the regulation of TH. Indeed, CRF seems to mediate the induction of TH in response to chronic footshock and noise stress (43).

Several lines of evidence suggest that the cAMP second messenger system contributes to adaptive responses of NE neurons in locus coeruleus (3). Thus, the cAMP system mediates the long-term effects of stress, catecholamine depletion, and various drug and hormone treatments in NE neurons but not in DA neurons (44). It is of considerable interest that TH phosphorylated at Ser40 position by PKA was localized immunohistochemically in E and NE neurons of the brainstem, and in only a small population of DA neurons in substantia nigra and VTA (36, 37). The findings that E and NE neurons in the steady state are phosphorylated by PKA at TH Ser40, and that most DA neurons are not, imply that a large capacity for phosphorylation of TH above basal levels exists in some midbrain DA neurons.

The question of whether the feedback inhibition of TH activity in vitro by the enzymatic end products L-dopa and DA and by other catecholamines (47, 55) is of physiological significance was extensively studied. The intracellular concentration of free catecholamines might be too low for regulating TH activity, and feedback inhibition of the enzyme by the enzymatic end products might play a role only when intracellular levels are elevated under physiological or pathological conditions. Our findings that site-directed mutagenesis of TH at serine 40 (Ser was substituted with Leu or Tyr) produces an activated form of the enzyme which is less sensitive to feedback inhibition and increases L-dopa synthesis in situ (18, 63) illustrates the effects of reduced inhibition on L-dopa biosynthesis. It was also reported that feedback inhibition involves competition between the catecholamines and the pteridine cofactor for the oxidized form of the enzyme, and catecholamines bind to TH by a direct coordination to Fe3+ at the active enzyme site (24).

The autoreceptor regulation of TH activity was also extensively studied (17, 62), and it was shown that stimulation of autoreceptors by DA or other DA D2/D3 agonists inhibits striatal TH activity and release of DA from nerve terminals, as well as decreases the phosphorylation of TH (17, 62). The question of whether DA autoreceptors which regulate synthesis and release of the transmitter represent the same or different receptor proteins requires further investigation. The findings that DA autoreceptors regulating synthesis of the transmitter are linked to pertussis-toxin-sensitive G proteins, whereas those regulating release are not (6, 8), 25), indicate that these receptors are not coupled to the same effector systems. In light of the available data that DA D2 receptors are coupled to G proteins whereas DA D3 receptors are not, one can postulate that DA autoreceptors regulating DA biosynthesis are DA D2 receptors whereas those regulating release of DA are DA D3 receptors (see also Molecular Biology of the Dopamine Receptor Subtypes and Dopamine Autoreceptor Signal Transduction and Regulation).

In vitro incubation of the enzyme with polyanions, phospholipids, mucopolysaccharides such as heparin, and ribonucleic acid results in increased TH activity (34, 48). The activation of TH by macromolecules seems to be an electrostatic phenomenon rather than a result of a specific chemical interaction, and the physiological relevance is questionable. The activation of TH by protein-kinase-mediated phosphorylation leads to a covalent modification of the enzyme, and phosphorylation/dephosphorylation of TH represents an important mechanism in the short-term regulation of the enzymatic activity. It is now well established that the amino-terminal segment of the enzyme contains several potential phosphorylation sites and that each Ser site is phosphorylated by a distinct protein kinase ().

TH is a substrate for phosphorylation by PKA, and phosphorylation correlates with increased catalytic activity (42, 46). In purified TH preparations from cultured PC12 cells, PKA increased [32]P incorporation from [g-32P]ATP, and the radioactivity is associated with the 62-kD subunit of the enzyme (42). The enzyme phosphorylated by PKA has a higher affinity for the pteridine cofactor and a lower inhibitory affinity for catechols (40, 42, 59). The activity of the nonphosphorylated enzyme has a pH optimum at 6.0, whereas that of the phosphorylated has a broad optimum in the physiological pH range (42). The kinetic parameters and the activity optimum at physiological pH values of the phosphorylated enzyme suggest that phosphorylation of TH by PKA is of importance in regulation of TH activity and catecholamine biosynthesis. The phosphorylated and activated TH is dephosphorylated by phosphatase 2A, suggesting a role of specific phosphatases in the regulation of TH activity by dephosphorylation (23), 58).

TH is phosphorylated by the Ca2+-phospholipid-dependent PKC (see ), and the kinetic properties of the enzyme phosphorylated by this kinase are similar to those phosphorylated by PKA (2). A comparison of phosphopeptides generated by tryptic digestion of TH phosphorylated by PKA and those by PKC revealed that both kinases phosphorylate the enzyme at a similar site (2). However, it was reported that phosphorylation by PKA and not by PKC results in activation of the enzymatic activity (15). As a possible explanation it was suggested that PKC phosphorylates only two out of the four subunits of the enzyme without affecting the enzyme activity, whereas PKA phosphorylates all four subunits, resulting in an increase in the enzymatic activity. The availability of TH antibodies which recognize specifically TH phosphorylated at Ser40 site, anti-TH Ser40p, makes it possible to further investigate the stoichiometry of phosphorylation by PKA and PKC (36, 37). The findings that TH at Ser31 is indirectly phosphorylated by PKC and directly by ERK1 and ERK2 kinases (31) will provide new insights on the regulation of TH by these protein kinases.

Ca2+ CaMpKII phosphorylates TH, but its phosphorylation is not associated with an increase in enzymatic activity (64). Some studies suggest that an "activator protein" is required for increased catalytic activity, and the cloning of cDNA coding for the protein kinase-dependent activator has been reported (32). Other explanations for the lack of enhanced catalytic activity following phosphorylation by Ca2+/CaMpKII—such as the enzyme is phosphorylated at multiple sites, and these sites modulate the activity in a way that the activation is nullified (20)—were also suggested. The idea that Ca2+/CaMpKII, like PKC, phosphorylates only two of the four subunits of the enzyme, which is not sufficient for activation of the enzyme, has also been proposed (20).

A number of serine/threonine kinases which are stimulated in response of cells to growth factors were described. Some of these involve enzymes which are regulated by second messengers, whereas others seem to be independent of them. The second-messenger-independent serine/threonine protein kinases which are activated by growth factors include ribosomal protein S6 kinase (1, 12) and MAP kinase (ERK 1 and ERK 2) (1, 9, 12, 19, 31, 53, 54). Treatment of intact PC12 cells with bradykinin or nerve growth factor (NGF) increased the phosphorylation of TH in situ and the catalytic activity of ERKs (31). TH phosphorylation at the Ser31 site is regulated by multiple signaling pathways which converge at or prior to activation of ERKs (31). The presence in PC12 cells of an NGF-activated protein kinase, designated as N-kinase, was described (53). N-kinase is rapidly activated in PC12 cells by treatment with NGF, epidermal growth factor (EGF), basic fibroblast growth factor (bFGF), phorbol ester or dB-cAMP. Thus, N-kinase can be activated via multiple second messenger pathways, and it might play a role in mediating shared intracellular responses to various extracellular signals (54).

Site-Specific Phosphorylation (see

)

TH is phosphorylated by distinct protein kinases at five phosphorylation sites: Ser8, Ser19, Ser31, Ser40 and Ser153 in the N-terminal region of the enzyme (10, 31). TH is phosphorylated at Ser19 by CaMpKII, at Ser40 by PKA and PKC (and to a small extent by CaMpKII), and at Ser31 by ERK1 and ERK2 kinases and indirectly by PKC. Electrical stimulation of the medial forebrain bundle increases phosphorylation at Ser19, Ser31, and Ser40 of the enzyme (26). Although Ser153 is a substrate for phosphorylation by PKA, no evidence for in situ phosphorylation at this enzymatic site was obtained.

The proline-directed protein kinase (PDPK) isolated from PC12 cells was shown to phosphorylate Ser8 in vitro (60). The levels of this kinase are very low in brain and adrenal glands, but relatively high in PC12 cells. PDPK activity is increased in response to NGF, and this effect seems to be mediated by the high-affinity NGF receptors (59). Inhibition of phosphatase-2A by treatment of synaptosomes with okadaic acid increases by severalfold the [32P] incorporation into Ser8, suggesting a high turnover rate of phosphate on Ser8 (29).

CaM-protein kinase is the only protein kinase known to phosphorylate TH at Ser19, and this kinase is activated in nerve terminals by depolarization (50). The activation of CaMpKII seems to be linked to depolarization-dependent Ca2+ influx, and phosphorylation of TH by this kinase is of physiological significance. The temporal changes following depolarization of striatal terminals on phosphorylation of TH were investigated (29). At relatively short treatment durations the effects of K+ and veratradine were restricted to Ser19 phosphorylation. At longer treatment durations (up to 4 min), an increase in Ser31 phosphorylation and a smaller increase in Ser19 phosphorylation was observed (29). A biphasic increase in Ser19, but not in Ser31, was also observed in chromaffin cells following prolonged exposure to acetylcholine (27, 30). The increased activity and phosphorylation by nicotine and muscarine is primarily associated with phosphorylation of the enzyme at Ser19 site (30).

TH Ser31 was found to be a substrate for ERK1 and ERK2 (31), and a number of receptor systems in PC12 cells stimulate ERKs activity and Ser31 phosphorylation. Thus, muscarinic, bradykinin, ATP, and NGF receptor activation increases Ser31 phosphorylation. Two intracellular signaling pathways, an NGF and a PKC activation pathway, seem to be linked with the ERKs-mediated Ser31 phosphorylation of TH. Muscarine, bradykinin, and ATP activate G-protein-linked receptors, resulting in an increase of phosphatidyloinositol phosphate turnover and activation of PKC. However, the pathways associated with NGF-stimulated increases in Ser31 phosphorylation are not yet established. The high-affinity NGF receptor has been identified as trk-B (the proto-oncogene product), and NGF increases its tyrosine phosphorylation in PC12 cells (33). Although phosphorylation of Ser31 increases TH activity by 20–40% (31), its physiological significance might not be relevant in view of much greater increases associated with phosphorylation of TH at Ser40 by PKA. The interactions between Ser31 phosphorylation and that of Ser40 and/or Ser19 are under investigation.

The enzyme is phosphorylated at Ser40 in vitro by several protein kinases including PKA, PKC, CaMpKII, PKN, and S6 kinase, but the phosphorylation by PKA plays the predominant role in vivo (26). The adenylyl cyclase/cAMP-dependent protein kinase system stimulates TH activity in intact striatal synaptosomes and slices, as well as in vivo by increased phosphorylation of Ser40. Vasoactive intestinal polypeptide and related peptides stimulate TH activity in various catecholaminergic tissues by PKA-mediated increased phosphorylation of Ser40. In intact PC12 cells the substrate specificity of Ser40 seems to be restricted to PKA (27). The involvement of CaMpKII in phosphorylation of Ser40 seems unlikely, because in synaptosomes maximal increases in Ser19 phosphorylation produced by elevated K+ failed to increase Ser40 phosphorylation. On the other hand, the possible involvement of PKC in phosphorylation of Ser40 cannot be excluded because treatment of synaptosomes for a longer time period (15 min) with phorbol dibutyrate increased the phosphorylation of Ser31 and, to a smaller extent, Ser40 (29, 30).

In order to determine the degree to which phosphorylation at serine sites 19, 31, and 40 contributes to an increase in TH activity, we individually substituted the corresponding serine with leucine or tyrosine and transfected AtT-20 cells with TH cDNA constructs (63). The specific enzymatic activity of transiently expressed TH mutant Ser40m was higher as compared with the wild-type enzyme or the mutants Ser19 and Ser31 (63). Kinetic studies with stably expressed recombinant TH revealed that Ser40m has a lower KM for the cofactor 6-methyltetrahydropteridine and a higher Ki for the end product DA than the wild-type enzyme (63). These findings suggest that Ser40 exerts an inhibitory influence on the enzymatic activity; and its replacement with another amino acid (AA) by site-directed mutagenesis, or its modification by phosphorylation, leads to a change in confirmation with an increase in enzymatic activity. The phosphorylation of TH mediated by PKA increases the enzymatic activity of the wild type but has no effect on Ser40m activity, indicating the essential role of Ser40 in the activation of TH by PKA (63).

TH is the product of a single gene, and in most species a single TH mRNA is translated to produce a single form of the protein. However, three additional forms of human TH mRNA which were formed by alternative splicing were detected. The three additional mRNA forms have 12, 81, or 12 + 81 additional nucleotides (21, 49). Antibodies to each of the four forms of TH protein identified all four isoforms of TH in human adrenal glands, in human pheochromocytoma, and in several neuroblastoma cell lines (27). The physiological and pathological significance of the four TH isoforms in the regulation of TH activity is not yet known. It is of interest that the addition of 12 nucleotides transforms the sequence surrounding Ser35 in hTH-2 and -4 (Arg-Gly-Gln-Ser) in such a way that it represents a putative site for phosphorylation by a Ca/CaMpKII.

Considerable progress was made in elucidating the mechanisms involving short- and long-term regulation of TH. Phosphorylation by distinct protein kinases at specific Ser phosphorylation sites of the enzyme were characterized, but their individual contribution to the regulation of TH and catecholamine biosynthesis has to be further investigated. A large body of experimental evidence suggests that phosphorylation at Ser40 by the cAMP-dependent protein kinases increases enzymatic activity. However, Ser40 can be phosphorylated in vitro by a number of other protein kinases such as PKC, Ca/CaMpKII, PKN, and so on, and the role of each of these kinases in the stimulation of enzymatic activity has not yet been established. A Ca2+-dependent activation of TH by depolarization in peripheral and central nervous system has been demonstrated, and this activation is linked to Ca/CaMpKII phosphorylation of TH at the Ser19 site of the enzyme. Studies with mutant enzymes in which single and multiple phosphorylation sites are systematically eliminated will further elucidate the relationships between different sites and their effects on TH activity. With this in mind, we investigated the phosphorylation and activation of TH mutants in which Ser40, Ser31, or Ser19 was substituted with another AA. Our studies indicate that removal of the phosphorylation site at Ser40 affects the magnitude of phosphorylation at other sites (18). The findings that in situ L-dopa biosynthesis in AtT-20 cells catalyzed by TH Ser40m is higher than that by TH WT (18) infer that a single AA mutation might be associated with the overproduction of catecholamines in some specific disorders. It is noteworthy that PKC exhibits both negative and positive cross-talk with multifunctional Ca2+/CaM protein kinase in PC12 cells (41), and multiple signaling pathways might be involved in the regulation of TH activity.

The resistance of TH to regulation by various exogenous and endogenous stimuli in DA, but not in NE/E neurons, suggests that different intracellular transduction pathways converge upon TH in these neuronal populations. cAMP-dependent systems play an essential role in the regulation of neuronal noradrenergic activity, but the intracellular pathways in the DA neurons were not yet fully elucidated. Most studies on phosphorylation of TH mediated by trophic factors [e.g., NGF, acidic fibroblast growth factor (aFGF) or bFGF] were carried out in PC12 cells. However, the importance of these signaling systems in nervous tissues remains to be determined. The presence of neurotrophic factors such as aFGF or bFGF in mesencephalic DA neurons (7, 11) indicates that these factors might modulate TH activity in these neurons. The findings that NGF induction of TH in PC12 cells is regulated by the c-fos gene family (16) raises the possibility that the expression of early genes is associated with regulation and dysregulation of TH neuronal activity. The generation of phospho/dephospho-specific antibodies against a segment of TH (36, 37) will be a useful tool in investigations of the basic regulatory features of TH in catecholamine neurons and their dysregulation in disease state.

The author gratefully acknowledges the support of NIMH, Career Scientist Award MH 14918, and MH 02717.

published 2000