Differential Regulation of the Human Tyrosine Hydroxylase Isoforms via Hierarchical Phosphorylation*

Tyrosine hydroxylase (TH) is the rate-limiting enzyme in the biosynthesis of the catecholamines dopamine, noradrenaline, and adrenaline. In response to short term stimuli TH activity is primarily controlled by phosphorylation of serine 40. We have previously shown that phosphorylation of serine 19 in TH can indirectly activate TH via a hierarchical mechanism by increasing the rate of phosphorylation of serine 40. Here we show that phosphorylation of serine 31 in rat TH increases the rate of serine 40 phosphorylation 9-fold in vitro. Phosphorylation of serine 31 in intact bovine chromaffin cells potentiated the forskolin-induced increase in serine 40 phosphorylation and TH activity more than 2-fold. Humans are unique in that they contain four TH isoforms but to date no significant differences have been shown in the regulation of these isoforms. Phosphorylation of the human TH isoform 1 at serine 31 by extracellular signal-regulated protein kinase (ERK) also produced a 9-fold increase in the rate of phosphorylation of serine 40, whereas little effect was seen in the TH isoforms 3 and 4. ERK did not phosphorylate human TH isoform 2. The effect of serine 19 phosphorylation on serine 40 (44 in TH2) phosphorylation is stronger in TH2 than in TH1. Thus hierarchical phosphorylation provides a mechanism whereby the two major human TH isoforms (1 and 2) can be differentially regulated with only isoform 1 responding to the ERK pathway, whereas isoform 2 is more sensitive to calcium-mediated events.

Tyrosine hydroxylase (TH) is the rate-limiting enzyme in the biosynthesis of the catecholamines dopamine, noradrenaline, and adrenaline. In response to short term stimuli TH activity is primarily controlled by phosphorylation of serine 40. We have previously shown that phosphorylation of serine 19 in TH can indirectly activate TH via a hierarchical mechanism by increasing the rate of phosphorylation of serine 40. Here we show that phosphorylation of serine 31 in rat TH increases the rate of serine 40 phosphorylation 9-fold in vitro. Phosphorylation of serine 31 in intact bovine chromaffin cells potentiated the forskolin-induced increase in serine 40 phosphorylation and TH activity more than 2-fold. Humans are unique in that they contain four TH isoforms but to date no significant differences have been shown in the regulation of these isoforms. Phosphorylation of the human TH isoform 1 at serine 31 by extracellular signal-regulated protein kinase (ERK) also produced a 9-fold increase in the rate of phosphorylation of serine 40, whereas little effect was seen in the TH isoforms 3 and 4. ERK did not phosphorylate human TH isoform 2. The effect of serine 19 phosphorylation on serine 40 (44 in TH2) phosphorylation is stronger in TH2 than in TH1. Thus hierarchical phosphorylation provides a mechanism whereby the two major human TH isoforms (1 and 2) can be differentially regulated with only isoform 1 responding to the ERK pathway, whereas isoform 2 is more sensitive to calcium-mediated events.
Tyrosine hydroxylase (TH) 2 [EC1.14. 16.2] is the rate-limiting enzyme in the biosynthesis of the catecholamines dopamine, noradrenaline, and adrenaline (1). Short term regulation of TH is accomplished by dynamic changes in the phosphorylation state of the enzyme (2,3). Although four serine residues have been shown to be phosphorylated in TH, only three of these serine residues (Ser 19 , Ser 31 , and Ser 40 ) are regulated in vivo (4). The most important mechanism of TH activation is phosphorylation of Ser 40 , which decreases the feedback inhibition by the catecholamines (5)(6)(7). Phosphorylation of dopamine bound TH at Ser 40 by protein kinase A (PKA) can activate TH by up to 20-fold (5). The direct effect of Ser 19 and Ser 31 phosphorylation on TH activation is much more modest. Phosphorylation of Ser 19 by calcium calmodulin-dependent protein kinase (CaMKII) will only increase TH activity in the presence of the 14-3-3 protein (8 -10), and this results only in a 2-fold increase in the activity. The phosphorylation of Ser 31 by extracellular signal-regulated protein kinase (ERK) produces less than a 2-fold increase in TH activity, primarily by decreasing the affinity of the cofactor tetrahydrobiopterin (BH 4 ) (11)(12)(13). Phosphorylation of Ser 31 in TH has also been shown to increase the stability of TH (14).
Humans have four TH protein isoforms, whereas anthropoids have two, and other mammalian species only have one (15). TH is encoded by a single gene, and the multiple isoforms are because of multiple mRNAs generated by alternative splicing of the single gene (16 -18). The human TH1 (hTH1) variant is like the subunits in all other species. The other human TH isoforms hTH2, hTH3, and hTH4 have inserts that lead to the expression of proteins containing 4, 27, and 31 (4 ϩ 27) amino acids inserted immediately N-terminal to Ser 31 in the hTH1 isoform. The hTH1 and hTH2 isoforms are the most prominent forms in human tissue samples and human cell lines (15,19,20). Analysis of recombinant forms of the four human isoforms indicate that their steady state kinetic parameters are comparable among the non-phosphorylated forms and among the phosphorylated forms (12,(21)(22)(23). All four human TH isoforms showed the same dopamine binding characteristics as rat TH (rTH) (23).
As there is little direct effect of Ser 19 and Ser 31 phosphorylation on TH activity we have been exploring the possibility that phosphorylation of these sites may indirectly effect TH activation by hierarchical phosphorylation (3). That is, phosphorylation of Ser 19 and Ser 31 could alter the rate of Ser 40 phosphorylation and therefore TH activation. We have shown that phosphorylation of Ser 19 increases the rate of phosphorylation of Ser 40 in TH 3-fold, whereas Ser 40 phosphorylation has no effect on the rate of Ser 19 phosphorylation (24). Therefore the phosphorylation of Ser 19 increases the rate of phosphorylation of Ser 40 in a hierarchical manner. This result has been confirmed both qualitatively and quantitatively by others using a different methodology (10). We have further shown that Ser 19 phosphorylation can potentiate Ser 40 phosphorylation and TH activation in intact cells (25).
In this report we have investigated whether Ser 31 phosphorylation could also alter the rate of Ser 40 phosphorylation. The results in this report show that Ser 31 phosphorylation has a major effect on the rate of Ser 40 phosphorylation and TH activation in intact cells and in vitro. These results also provide, for the first time, a mechanism by which the four human TH isoforms can be differentially regulated by hierarchical phosphorylation.

EXPERIMENTAL PROCEDURES
Materials-Protease Inhibitor Mixture was from Roche Applied Science. Angiotensin II (AngII), HEPES, phenol red, EGTA, EDTA, and PKA were obtained from Sigma. BH 4 was obtained from Dr. B. Schirck's laboratories, Jona, Switzerland. Forskolin was obtained from Biomol. SDS-PAGE reagents were from Bio-Rad Laboratories. Molecular weight PAGE standards, nitrocellulose membrane (Hybond ECL), ECL plus kit, anti-rabbit immunoglobulin (horseradish peroxidase-linked whole antibody from donkey), and anti-mouse immunoglobulin (horseradish peroxidase-linked whole antibody from sheep) were obtained from Amersham Biosciences. The hTH2 Ser 35 phosphopeptide (GQS P PR) was synthesized by AUSPEP (Australia). Tissue culture reagents were from Sigma and were of analytical or tissue culture grade. Forskolin was dissolved in dimethyl sulfoxide before use. Appropriate solvent controls were performed for every condition.
Expression and Purification of TH and ERK and Purification of CaMKII-rTH was expressed in Escherichia coli and purified according to previously described procedures (5) with modifications as described (24). The recombinant rTH is not phosphorylated and does not contain bound catecholamine. The TH mutants were generated using the QuikChange site directed mutagenesis kit (Stratagene). The hTH2 and hTH4 isoforms were generated from the hTH1 and hTH3 isoforms, respectively, using the conditions described (23). The introduction of the mutation and the absence of other mutations were confirmed by DNA sequencing. The S19A mutations were further confirmed by the absence of Ser 19 phosphorylation using the Ser(P) 19 -specific antibody after phosphorylation with CaMKII. ERK2 was expressed from the plasmid NPT7-5His-ERK2-R4F and purified essentially as described (26) except that elution from the nickel-charged chelating-Superose column was with an imidazole gradient (0 -400 mM). CaMKII was purified from whole brain by calmodulin-Sepharose chromatography (27). All of the recombinant DNA experiments were carried out in accordance with the guidelines of the Office of the Gene Technology Regulator, Australia.
Determination of TH Activity and the Incorporation of Radioactivity into TH-The phosphorylation reactions for TH were performed in 50 mM Tris-HCl, pH 7.5, 100 M ATP, 12.5 mM MgCl 2 except for CaMKII, which were performed at 1 mM ATP. To obtain dopamine-bound TH, the TH was initially incubated in a buffer containing 75 mM MOPS, pH 7.2, 75 mM GSH, 15 M dopamine for 20 min at 20°C. Phosphorylation reactions were performed in the same buffer but with the addition of 100 M ATP and 12.5 mM MgCl 2 . TH activity was determined using the tritiated water release assay of Reinhard et al. (28). After dopamine binding and phosphorylation of TH, the TH was purified over a heparin-Sepharose column. The TH activity was then measured in a buffer containing 25 M tyrosine, 60 mM potassium phosphate, pH 7.2, 0.006% ␤-mercaptoethanol, 36 g/ml catalase, 500 M BH 4 at 25°C for 8 min. For the analysis of incorporation of radioactivity into TH, [␥-32 P]ATP was included in the phosphorylation reactions. The phosphorylation reactions were stopped by the addition of an equal volume of 4% SDS, 2 mM EDTA, 1% dithiothreitol, 50 mM Tris, pH 6.8. Proteins were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis. The TH bands were detected by staining with Coomassie Blue-R250, the gel section containing TH was cut out, and the incorporated radioactivity was measured in a scintillation counter (Wallac 1410, Pharmacia).
TH Site Analysis-After TH was phosphorylated in the presence of [␥-32 P]ATP as described above the samples were applied to sodium dodecyl sulfate polyacrylamide gels and then transferred to nitrocellulose as described by Jarvie and Dunkley (29). The nitrocellulose filter was then subject to Ponceau staining. The TH band was cut out of the nitrocellulose, and the TH was eluted from the nitrocellulose by trypsin digestion. The HPLC analysis of the resultant phosphopeptides was performed as described previously (30). In the CaMKII rate experiments TH was precipitated by 10% trichloroacetic acid, washed twice with 10% trichloroacetic acid, and then once with acetone; the TH was then digested with trypsin and then directly injected onto the HPLC column.
Measurement of Site-specific TH Phosphorylation, TH Activity, Ser 40 Kinase Activity and the Stoichiometry of Ser 31 Phosphorylation in Bovine Adrenal Chromaffin Cell Extracts-Bovine adrenal chromaffin cells (BACCs) were prepared as described (31) with modifications (32). BACCs were plated and treated as described (25) except that cells were initially treated with 100 nM AngII for 30 min rather than anisomycin. Site-specific TH phosphorylation, TH activity, and Ser 40 kinase activity in BACC extracts were measured as described (25). For measurement of the stoichiometry of Ser 31 phosphorylation in BACCs, recombinant rat TH was maximally phosphorylated at Ser 31 by incubation with ERK and was used as a calibration standard. A stoichiometry of 0.45 was assumed for the ERK-phosphorylated TH, as this was the average stoichiometry determined in a number of experiments. The stoichiometry of Ser 31 phosphorylation in untreated BACCs was determined as described for Ser 19 and Ser 40 phosphorylation (25).
Statistical Analysis-For multiple comparisons statistical significance was assessed by the Tukey's test for multiple comparisons, protected by a one-way analysis of variance. Student's t test was used for pairwise comparisons.

The Effect of Ser 31 Phosphorylation on the Forskolin-induced Increase in Ser 40 Phosphorylation and TH Activity in Intact
Cells-To examine the effect of Ser 31 phosphorylation on Ser 40 phosphorylation in intact cells we needed to establish conditions in which BACCs could be stimulated to independently increase the level of Ser 31 or Ser 40 phosphorylation without increasing Ser 19 phosphorylation. The data in Fig. 1A show that when BACCs were incubated with AngII for 30 min, the level of Ser 31 phosphorylation increased 3-4-fold with respect to control, whereas Ser 40 phosphorylation and Ser 19 phosphorylation did not change compared with the control ( p Ͼ 0.05). An incubation of BACCs with forskolin for 4 min increased only Ser 40 phosphorylation. The stoichiometry of phosphorylation of Ser 31 was measured as described under "Experimental Procedures." A value for Ser 31 stoichiometry in untreated chromaffin cells of 0.04 Ϯ 0.003 (mean Ϯ S.E., n ϭ 8) was obtained. This indicates that the stoichiometry of Ser 31 phosphorylation after AngII treatment would be between 0.12 and 0.16.
When the BACCs were treated first with AngII and then forskolin, the change in the levels of phosphorylation of Ser 19 and Ser 31 were the same as when the cells were individually treated with AngII or forskolin. This suggests that the effects of AngII and forskolin on BACCs are independent of each other. In contrast to this result, the increase in the level of Ser 40 phosphorylation in cells treated with AngII plus forskolin was almost 2-fold greater than the increase in Ser 40 phosphorylation in cells treated with forskolin alone. This difference was significant ( p Ͻ 0.001). The effect of AngII and forskolin on TH activity is also shown in Fig. 1A. The treatment of cells with AngII produced no significant increase in TH activity when compared with the activity in untreated cells ( p Ͼ 0.05). The addition of forskolin increased TH activity in BACCs ϳ2-fold. In cells treated with forskolin plus AngII, the increase of TH activity over control was 2.2-fold greater than that in cells treated with forskolin alone. This difference was statistically significant ( p Ͻ 0.001).
The data shown above suggest that phosphorylation of Ser 31 mediated by AngII can potentiate the forskolin-induced increase in Ser 40 phosphorylation. If this were true then AngII should not increase the Ser 40 phosphorylating activity in forskolin-treated cells. The level of TH Ser 40 kinase activity in extracts of cells was determined as described under "Experimental Procedures" using exogenous TH as a substrate. The cells were treated with AngII, forskolin, or both forskolin and AngII. The data in Fig. 1B show the results from a phosphopeptide analysis of TH under the conditions used for the Ser 40 kinase assay after treatment with AngII and forskolin. The results show a single radioactive peak with an elution position similar to that obtained for the Ser(P) 40 tryptic peptide in our system (33). This indicates that essentially only the Ser 40 is phosphorylated. The very low stoichiometry of Ser 19 and Ser 31 under these conditions would mean that we would not expect any potentiation of exogenous TH Ser 40 phosphorylation because of Ser 19 or Ser 31 phosphorylation in the Ser 40 kinase assay and so would be measuring the true Ser 40 kinase activity. The results in Fig. 1B show that treatment of cells with forskolin produced ϳ17-fold increase in Ser 40 kinase activity in the cell extracts. When extracts of cells treated with AngII and forskolin were analyzed for Ser 40 kinase activity, the activity was not significantly different from that in extracts of cells treated with forskolin alone ( p Ͼ 0.05). Therefore the phosphorylation of Ser 31 mediated by AngII must be potentiating the forskolin-induced increase in TH Ser 40 phosphorylation.
Stoichiometry and Site Analysis of the Phosphorylation of rTH by ERK and PKA-The data shown above indicated that phosphorylation of Ser 31 could increase the rate of phosphorylation of Ser 40 and activation of TH in intact cells. To understand the mechanism of this we examined the effect of Ser 31 phosphorylation on Ser 40 phosphorylation and TH activation in vitro. In these experiments ERK was used to phosphorylate Ser 31 . To phosphorylate Ser 31 in rTH, recombinant ERK was expressed in E. coli and purified as described under "Experimental Procedures." The specificity of ERK phosphorylation of rTH was determined. rTH was maximally phosphorylated with ERK. The ERK-phosphorylated rTH was then subjected to tryptic digestion, and the resulting labeled phosphopeptides were analyzed using HPLC as described under "Experimental Procedures." The results in Fig. 2, panel E show a major peak eluting at ϳ20 min and three minor peaks. Taking into account the elution position of the Ser(P) 40 peptide (PKA-phosphorylated rTH; see Fig. 2, panel P) the elution position of the major peak was consistent with that previously determined for the Ser(P) 31 peptide from rat TH (11). Essentially identical results were obtained with the ERK phosphorylation of dopamine-bound rTH ( Fig. 2 panel DE). Previous results have shown that ERK could also phosphorylate Ser 8 in rTH but only at a low level (11). More recent results suggested that ERK could phosphorylate Ser 8 at a rate ϳ9-fold less than Ser 31 (34), so it was possible that under the conditions used to maximally phosphorylate Ser 31 there may have been significant phosphorylation of Ser 8 . The elution position of the smallest of the three minor peaks was consistent with that previously defined to be the elution position of the Ser(P) 8 peptide in rTH (11). To confirm this, ERK was used to phosphorylate a S8A mutant of rTH. As can be seen in Fig. 2, panel E (S8A), the middle peak of the three minor peaks disappears in the S8A mutant, confirming that this peak is in fact the Ser(P) 8 peptide. The very low level of Ser 8 phosphorylation suggested that it was unlikely to have any significant impact on Ser 40 phosphorylation and so wild-type rTH was used in all subsequent experiments. The nature of the two other minor peaks is unclear. Once rTH was maximally phosphorylated by ERK, the subsequent addition of PKA only resulted in phosphorylation of Ser 40 in dopamine-free ( Fig. 2 panel EP) or dopamine-bound (Fig. 2, panel DEP) TH.
Analysis of the Effect of Ser 31 Phosphorylation on Ser 40 Phosphorylation and TH Activity in rTH-We first determined whether the effect of Ser 31 phosphorylation on TH in intact cells was because of Ser 31 phosphorylation increasing the ability of PKA to phosphorylate Ser 40 and activate dopamine-bound rTH. The results in Fig. 3A show that addition of dopamine inhibited rTH activity ϳ40-fold (D). Phosphorylation of Ser 31 by ERK did not significantly activate dopamine-bound rTH (Fig.  3A, DE). The addition of very high levels of PKA was able to fully reactivate the dopamine-bound rTH (DP150). The addition of a lower con-FIGURE 1. Effect of Ser 31 phosphorylation on the forskolin-induced increase in Ser 40 phosphorylation and TH activation in BACCs. A, BACCs were incubated with or without 100 nM AngII at 37°C for 30 min. The cells were then incubated with or without 1 M forskolin for a further 4 min at 37°C. The cells were then processed, and the level of phosphorylation in TH of Ser 19 (pSer 19 ), Ser 31 (pSer 31 ), Ser 40 (pSer 40 ) or TH activity was determined as described under "Experimental Procedures." Con, control basal; Fors, forskolin. Results are presented as the percent of the mean of the control Ϯ S.E. For each condition the number of individual experiments was between 6 and 10. Statistical analysis: Ser 19 , all conditions not significantly different from control ( p Ͼ 0.05); Ser 31 , Con versus Fors and AngII versus ForsϩAngII were not significant ( p Ͼ 0.05), all other comparisons were significant ( p Ͻ 0.001); Ser 40 , Con versus AngII was not significant ( p Ͼ 0.05), all other comparisons were significant ( p Ͻ 0.001); TH activity, Con versus AngII was not significant ( p Ͼ 0.05), all other comparisons were significant ( p Ͻ 0.001). B, TH site. Cells treated with AngII and forskolin as described above were processed, and the Ser 40 kinase assay was run as described under "Experimental Procedures," except that [␥-32 P]ATP was also included in the assay, and HPLC analysis of the radiolabeled tryptic peptides was performed as described under "Experimental Procedures." The elution profile is shown. In Ser 40 kinase, cells were treated as described above and then processed, and the level of TH Ser 40 kinase activity was determined as described under "Experimental Procedures." The results are presented as the percent of the mean of the control Ϯ S.E. For each condition the number of individual experiments was four. Statistical analysis, Con versus AngII and Fors versus ForsϩAngII were not significant ( p Ͼ 0.05), all other comparisons were significant ( p Ͻ 0.001).
centration of PKA was able to partially reactivate rTH (Fig. 3A, DP). Prior phosphorylation of Ser 31 by ERK had no effect on the ability of this lower PKA concentration to reactivate rTH (DEP is not significantly different from DP). The effect of Ser 31 phosphorylation on Ser 40 phosphorylation in dopamine-bound TH was examined directly (Fig. 3B). Again, prior phosphorylation of Ser 31 did not alter the PKA-mediated increase in the phosphorylation of Ser 40 in dopamine-bound TH (DEP is not significantly different from DP). The results show that phosphorylation of Ser 31 could not increase the rate of phosphorylation of Ser 40 nor increase the activation of dopamine-bound rTH by PKA.
We therefore determined whether the potentiation of TH activation by Ser 31 phosphorylation in intact cells could be because of Ser 31 phosphorylation increasing the ability of PKA to phosphorylate Ser 40 in dopamine-free TH. As there is only a very small effect on rTH activity by Ser 40 phosphorylation in dopamine-free TH and as Ser 31 phosphorylation produces only a small change in rTH activity, we examined Ser 40 phosphorylation directly. The effect of Ser 31 phosphorylation on the initial rate of Ser 40 phosphorylation by PKA in dopamine-free rTH was determined, and the results are presented in Fig. 3C. The data show that the rate of Ser 40 phosphorylation by PKA for Ser 31 phosphorylated rTH (Fig. 3C, EP) is 9-fold greater than that for rTH not phosphorylated at Ser 31 (P). In contrast, prior phosphorylation of Ser 40 did not alter the rate of phosphorylation of Ser 31 by ERK (not shown), indicating that the effect of the phosphorylation of Ser 31 is hierarchical.
Stoichiometry and Site Analysis of the Phosphorylation of the Human TH Isoforms by ERK and CaMKII-The sequence comparison of the human TH isoforms is shown in Fig. 4A. It can be seen that the human TH isoforms differ in sequence only around the Ser 31 site in hTH1. This suggested that the human TH isoforms may differ with respect to the effect of hierarchical phosphorylation via Ser 31 (or its equivalent) phosphorylation. The four human isoforms were expressed and purified as described under "Experimental Procedures." The phosphorylation of the isoforms by ERK was examined. hTH1 was phosphorylated to a maximum stoichiometry of 0.5, whereas hTH3 and hTH4 could be phosphorylated to a maximum stoichiometry of 1.0, but there was very little phosphorylation of hTH2 by ERK (not shown). These results are consistent with that determined previously (12). The site analysis of hTH1 phosphorylated by ERK and CaMKII is shown in Fig. 4B. The relative positions of the peaks found for the Ser 31 , Ser 19 , and Ser 40 tryptic phosphopeptides are consistent with those determined previously (12). When hTH1 was phosphorylated by ERK, there was also a small peak that eluted at 20 min. This peak was also found when hTH2 was phosphorylated by ERK, which would suggest that this peak is because of the phosphorylation of a protein in the ERK preparation rather than hTH1. hTH2 contains an additional four amino acids inserted N-terminal to Ser 31 in hTH1. There is no information in the literature on the elution position of the hTH2 tryptic Ser 35 phosphopeptide (GQS P PR). The hTH2 tryptic Ser 35 phosphopeptide was therefore synthesized, and as can be seen in Fig. 4B this peptide elutes very early under the conditions used. As expected, when hTH2 was phosphorylated by ERK the major   Ͼ 0.05). B, Ser 40 phosphorylation of rTH was determined using Western blotting as described under "Experimental Procedures" after various treatments. P, dopamine-free rTH phosphorylated with 20 units of PKA for 20 min; DP, dopamine-bound rTH phosphorylated with 20 units of PKA for 20 min; DEP, dopamine-bound rTH maximally phosphorylated with ERK and then phosphorylated with 20 units of PKA for 20 min. Ser 40 phosphorylation is shown the relative to dopamine-free rTH phosphorylated with PKA, which was defined as 100%. For each condition the number of individual experiments was six. Statistical analysis, DEP was not significantly different from DP ( p Ͼ 0.05). C, dopamine-free rTH was incubated with ERK (EP) or without ERK (P) in the presence of unlabeled ATP for 25 min, [␥-32 P]ATP was then added, and the incorporation of radiolabel into rTH was measured as described under "Experimental Procedures" for 3.5 min after addition of PKA. Initial rates were determined by linear regression analysis. For each condition three rate experiments were performed. The results are presented relative to the mean rate of phosphorylation of rTH by PKA (P), which was assigned a value of one. Statistical analysis, EP was significantly different from P ( p Ͻ 0.001).

Hierarchical Phosphorylation in Tyrosine Hydroxylase
peak found in hTH1 was no longer found (Fig. 4B, hTH2 ERK). Rather, in addition to the same small peak found in hTH1 that eluted at around 20 min, there was an additional minor peak that eluted very early. This early eluting peak was not found when hTH1 was phosphorylated by ERK, which suggests that it is derived from the phosphorylation of hTH2. This "peak" appeared to be two to three unresolved peaks. These peaks consistently eluted at least 2 min later than the synthesized hTH2 tryptic Ser 35 phosphopeptide and so this suggests that ERK cannot phosphorylate hTH2 at Ser 35 but may be phosphorylating another site(s) at very low levels. When hTH2 was phosphorylated by CaMKII (Fig. 4B, hTH2 CaMKII), the result was essentially the same as that for hTH1 phosphorylated by CaMKII (Fig. 4B, hTH1 CaMKII). This was surprising as it has been claimed that the Ser 35 site in hTH2 can be phosphorylated by CaMKII (22). We could find no evidence for the phosphorylation of an hTH2 tryptic Ser 35 phosphopeptide by CaMKII with mobility similar to that of the synthesized Ser 35 phosphopeptide. This result is consistent with previous work where attempts to demonstrate the phosphorylation of hTH2 Ser 35 in intact human neuroblastoma cells (prelabeled with 32 P i , treated with veratridine, forskolin, or phorbol ester, immunoprecipitated with isoform-specific antibodies, trypsinized, and analyzed by HPLC with radiochemical detection) indicated, rather, that hTH2 Ser 35 was essentially not phosphorylated under any of the conditions. 3 Effect of ERK Phosphorylation on the Human TH Isoforms-We examined the effect of ERK phosphorylation on the rate of activation of the dopamine-bound human TH isoforms by PKA. The data is shown in Fig. 5A. All four isoforms showed similar levels of inhibition by dopamine. Addition of low concentrations of PKA produced partial reactivation of each isoform (Fig. 5, DP). In all four isoforms, prior phosphorylation by ERK (Fig. 5, DEP) had no effect on the PKA-mediated activation as, in each case, DEP was not significantly different from DP.
We next investigated whether Ser 31 phosphorylation could potentiate Ser 40 phosphorylation in dopamine-free human TH isoforms. The results in Fig. 5B show that the effect seen in hTH1 was essentially the same as that for rTH in that Ser 31 phosphorylation by ERK increased the rate of Ser 40 phosphorylation by PKA ϳ9-fold. In contrast, phosphorylation of Ser 58 in hTH3 by ERK increased the rate of Ser 67 phosphorylation by PKA only ϳ1.7-fold. In the hTH4 isoform, Ser 62 phosphorylation by ERK did not alter the rate of Ser 71 phosphorylation by PKA. The Ser 35 residue in hTH2 could not be significantly phosphorylated by ERK or CaMKII and so for this isoform of TH the equivalent potentiation could not occur.
Determination of the Role of Ser 19 Phosphorylation in the Phosphorylation of Ser 40 in hTH1 and Ser 44 in hTH2-When the phosphorylation of hTH1 and hTH2 was examined by site analysis we con-

FIGURE 5. Effect of ERK phosphorylation on the PKA-mediated increase in TH activity and Ser 40 (or its equivalent) phosphorylation in the human TH isoforms.
A, TH activity of the four human isoforms in the presence of dopamine was determined in vitro after various treatments as described under "Experimental Procedures." The conditions and nomenclature are the same as in Fig. 3. TH activity is shown relative to untreated TH, which was defined as 100%. For each condition the number of individual experiments was three. Statistical analysis, for each isoform the value for DP was not significantly different from DEP ( p Ͼ 0.05). B, the effect of ERK phosphorylation on the rate of PKA phosphorylation of the dopamine-free hTH1, hTH3, and hTH4 isoforms was determined as described under "Experimental Procedures." The conditions and nomenclature are as described in Fig. 3. The results are presented relative to P, which was assigned a value of one. For each condition the number of individual experiments was four. Statistical analysis, for hTH1 and hTH3 the value for P was significantly different from EP ( p Ͻ 0.01), whereas for hTH4 the value of P was not significantly different from EP ( p Ͼ 0.05).
sistently found that when hTH1 and hTH2 were phosphorylated by CaMKII, the ratio of Ser 44 to Ser 19 phosphorylation in hTH2 was higher than the ratio of Ser 40 to Ser 19 phosphorylation in hTH1 (see Fig. 4B). This raised the possibility that the 4-amino-acid insert in hTH2 was either directly altering the rate of phosphorylation of Ser 44 or altering the effect of Ser 19 phosphorylation on the rate of Ser 44 phosphorylation. We therefore examined the rate of phosphorylation of Ser 19 , Ser 40 , and Ser 44 in the two isoforms by CaMKII. This was done by quantitating the radioactivity incorporated into Ser 19 , Ser 40 , and Ser 44 by site-analysis as described in methods. The results in Fig. 6A show that when hTH1 and hTH2 are maximally phosphorylated by addition of high levels of CaMKII and incubation for 300 s, the incorporation of radioactivity into Ser 19 , Ser 40 , and Ser 44 was, as expected, the same. As the amount of radioactivity incorporated into hTH1 and hTH2 was the same, the data were directly comparable. The results show that the rate of phosphorylation of Ser 19 was the same in hTH1 and hTH2. In contrast the rate of phosphorylation of Ser 44 in hTH2 was increased by 150% when compared with the rate of phosphorylation of Ser 40 in hTH1. To examine whether this was because of the effect of Ser 19 phosphorylation or because of the 4-amino-acid insert in hTH2 directly effecting CaMKII phosphorylation of Ser 44 , we prepared mutants of hTH1 and hTH2 where Ser 19 was converted to alanine so Ser 19 could not be phosphorylated. We compared the rates of phosphorylation by CaMKII of Ser 40 in S19A hTH1 with that of Ser 44 in S19A hTH2. The results in Fig. 6B show that CaMKII could phosphorylate Ser 44 in hTH2 faster than Ser 40 in hTH1 in the absence of Ser 19 phosphorylation, but the effect was small, with there being only a 30% increase. This direct effect (30% increase) can only contribute in a minor way to the effect seen in the presence of Ser 19 phosphorylation (150% increase). We therefore believe that it is the phosphorylation of Ser 19 that is primarily responsible for the increased rate of Ser 44 phosphorylation in hTH2. The rate of phosphorylation of Ser 40 in hTH1 by PKA (which does not phosphorylate Ser 19 ) was the same as the rate of phosphorylation of Ser 44 in hTH2 by PKA (Fig. 6C). Thus the presence of the 4-aminoacid insert in hTH2 does not affect the rate at which PKA phosphorylates Ser 44 compared with the rate at which PKA phosphorylates Ser 40 in hTH1.

DISCUSSION
Previously we have shown that phosphorylation of Ser 19 in TH can potentiate the phosphorylation of Ser 40 in vitro (24) and potentiate the phosphorylation of Ser 40 and lead to TH activation in situ (25). The data reported here extends this work by showing that the phosphorylation of Ser 31 can also increase the rate of Ser 40 phosphorylation and TH activation both in vitro and in situ. The phosphorylation of Ser 40 had no effect on the rate of phosphorylation of either Ser 19 or Ser 31 . Therefore the effect of phosphorylation of Ser 19 and Ser 31 is hierarchical in nature. In light of the very modest effect that phosphorylation of Ser 19 or Ser 31 have on TH activity we would suggest that the main role of the phosphorylation of these sites is to potentiate the rate of phosphorylation of Ser 40 .
Phosphorylation of Ser 31 increased the rate of phosphorylation of Ser 40 in dopamine-free TH but did not have any effect on the PKA activation of dopamine-bound TH. This is the same as the situation with Ser 19 where Ser 19 phosphorylation increased the rate of Ser 40 phosphorylation in dopamine-free TH (24) but had no effect on the PKA activation of dopamine-bound TH. 4 The phosphorylation of Ser 40 in dopamine-free TH has very little effect on TH activity (5) and so the question arises as to how Ser 31 (or Ser 19 ) phosphorylation produces the potentiation of TH activation that we have found in intact cells. In response to a stimulus Ser 40 phosphorylation in TH increases in the first 2-3 min. Clearly, phosphorylation of Ser 19 or Ser 31 does not effect the initial activation of catecholamine-bound TH. If there is continued stimulation the level of Ser 40 , phosphorylation remains relatively stable (35). This means that after the initial activation phase there is an equilibrium phase where the rate of phosphorylation of Ser 40 must equal the rate of dephosphorylation of Ser 40 in order that the level of phosphorylation of Ser 40 in the cell remains constant. This equilibrium phase is shown diagrammatically in Fig. 7. In the absence of Ser 31 phosphorylation, the level of Ser 40 phosphorylation will reach a stable point when the rate of phosphorylation equals the rate of dephosphorylation (Fig. 7, dotted line). Phosphorylation of Ser 31 will increase the rate of phosphorylation of 4 J. Daniel, P. Dunkley, and P. Dickson, unpublished observations. FIGURE 6. Effect of Ser 19 phosphorylation on the phosphorylation of Ser 40 in hTH1 and the phosphorylation of Ser 44 in hTH2. A, hTH1 and hTH2 were phosphorylated by CaMKII in the presence of [␥-32 P]ATP for the times shown up to 80 s. 6-Fold more CaMKII was then added, and the sample was incubated for a further 3.5 min to obtain maximal phosphorylation of each TH isoform. The incorporation of radioactivity into each site was measured as described under "Experimental Procedures." The data is combined from three independent experiments. Open circles, Ser 19 ; closed circles, Ser 40 (hTH1) or Ser 44 (hTH2). B, S19A hTH1 and S19A hTH2 were phosphorylated by CaMKII in the presence of [␥- 32  Ser 40 in catecholamine-free TH. This will have the effect of altering the equilibrium to favor the phosphorylated form and so the level of Ser 40 phosphorylation will increase. Ultimately a new equilibrium will be established but it will be at higher level of Ser 40 phosphorylation than in the absence of Ser 31 phosphorylation (Fig. 7, solid line). Therefore we believe that the role of Ser 31 (and Ser 19 ) phosphorylation in catecholamine-free TH is to increase the rate of rephosphorylation of Ser 40 in TH after it has been dephosphorylated (but before dopamine is rebound). This will have the effect of increasing Ser 40 phosphorylation in intact cells and, in particular, reducing the possibility of the catecholamine binding to the dephosphorylated TH and reinhibiting TH.
This work also provides a major insight into the way the human TH isoforms are regulated. Previous work has failed to identify any significant differences in the activity or regulation of the four human TH isoforms (12,(21)(22)(23). The fact that the only sequence difference among the four isoforms is around Ser 31 in hTH1 suggests that the effect of Ser 31 (or its equivalent) phosphorylation in increasing Ser 40 (or its equivalent) phosphorylation may be different in different isoforms. This was in fact the case with the strong potentiation via Ser 31 phosphorylation being only found in the hTH1 isoform. The other major human TH isoform hTH2 showed stronger potentiation via Ser 19 phosphorylation than hTH1. How would these differences between the two major TH isoforms impact on the regulation of TH in cells? When cells are stimulated either via depolarization or receptor-mediated mechanisms the level of phosphorylation of Ser 19 increases rapidly and thereafter starts to decrease (3). In contrast, the level of Ser 31 phosphorylation only increases after extended stimulation of the cell (3). This means that in response to short term stimuli, hTH2 will be activated to a greater extent than hTH1 because of the stronger potentiation via Ser 19 phosphorylation. When cells are stimulated for a longer period of time, the ERK pathway is activated. Activation of the ERK pathway will strongly potentiate the activation of hTH1 but will have no effect on hTH2. Therefore hTH1 and hTH2 can be differentially activated depending on the length of the stimulus involved.
Hierarchical phosphorylation via Ser 31 may also provide a mechanism by which hTH1, or the TH in other mammalian species that are homologous to hTH1, may be differentially activated in different tissues. We have shown here that in unstimulated adrenal chromaffin cells the stoichiometry of Ser 31 phosphorylation is very low with only 4% of TH phosphorylated at Ser 31 . This is similar to the results from unstimulated adrenal-derived PC12 cells where only 7% of TH molecules are phosphorylated at Ser 31 (36). In contrast, the basal level of Ser 31 phosphorylation was much higher in the brain with as much as 32% of TH phosphorylated at Ser 31 in the striatum (36). This would mean that in response to short term stimuli the high stoichiometry of Ser 31 phosphorylation the striatum would lead to strong potentiation of hTH1 activation, whereas this would not occur in adrenal cells where the stoichiometry of Ser 31 phosphorylation is very low. For many years it was thought that ERK was the only kinase that could phosphorylate Ser 31 in TH. The recent discovery that Ser 31 can also be phosphorylated by CDK5 (14,37) means that this potentiation of activation of hTH1 can be modulated by at least two different pathways and so could provide the basis of quite complex regulation of hTH1.
In summary we have shown the Ser 31 phosphorylation can strongly potentiate Ser 40 phosphorylation and TH activation. This has also provided for the first time a mechanism by which the different human TH isoforms are differentially regulated.