TPH (tryptophan hydroxylase) catalyses the rate-limiting step in the synthesis of serotonin, and exists in two isoforms: TPH1, mainly found in peripheral tissues and the pineal body, and TPH2, a neuronal form. In the present study human TPH2 was expressed in Escherichia coli and in HEK (human embryonic kidney)-293 cells and phosphorylated using several different mammalian protein kinases. TPH2 was rapidly phosphorylated to a stoichiometry of 2 mol of phosphate/mol of subunit by PKA (protein kinase A), but only to a stoichiometry of 0.2 by Ca2+/calmodulin dependent protein kinase II. Both kinases phosphorylated Ser19, but PKA also phosphorylated Ser104, as determined by MS, phosphospecific antibodies and site-directed mutagenesis of several possible phosphorylation sites, i.e. Ser19, Ser99, Ser104 and Ser306. On average, purified TPH2 WT (wild-type) was activated by 30% after PKA phosphorylation and studies of the mutant enzymes showed that enzyme activation was mainly due to phosphorylation at Ser19. This site was phosphorylated to a stoichiometry of up to 50% in HEK-293 cells expressing TPH2, and the enzyme activity and phosphorylation stoichiometry was further increased upon treatment with forskolin. Purified PKA-phosphorylated TPH2 bound to the 14-3-3 proteins γ, ϵ and BMH1 with high affinity, causing a further increase in enzyme stability and activity. This indicates that 14-3-3 proteins could play a role in consolidating and strengthening the effects of phosphorylation on TPH2 and that they may be important for the regulation of serotonin function in the nervous system.
Serotonin (5-hydroxytryptamine) is a key hormone and neurotransmitter involved in a range of physiological functions. TPH (tryptophan hydroxylase) [1,2] converts tryptophan into 5-hydroxytryptophan in the rate-limiting step for synthesis of 5-hydroxytryptamine. TPH belongs to the AAAH (aromatic amino acid hydroxylase) family of enzymes, which have strong structural and functional similarities and are dependent on BH4 (tetrahydrobiopterin) and iron(II) for activity .
Vertebrates have two distinct TPH genes (tph1 and tph2) which code for two different TPH enzymes . TPH1 is selectively expressed in the foetal brain , in non-neuronal serotonergic tissues, e.g. enterochromaffin cells of the gut, the pineal body and the skin . TPH2, on the other hand, is expressed mainly in serotonergic neurons of the raphe nuclei of the brain stem and mysenteric plexus . The two human TPH enzymes have an overall sequence identity of 71% . They are almost identical in the catalytic domain, whereas they differ in their N-terminal regulatory domains and kinetic properties .
Regulation of the enzyme by phosphorylation has been addressed in several species. Several reports suggest that phosphorylation by PKA (protein kinase A) [7–9] and by CaMKII (Ca2+/calmodulin-dependent protein kinase II) [10–13] activates TPH in brain tissues from various mammalian species, whereas other investigators have reported unaltered enzyme activities upon phosphorylation . However, the significance of these results remains unclear due to differences in experimental conditions and the possibility of heterogeneity within these enzyme samples because they were published prior to the identification of TPH2.
Rabbit and rat TPH1 have been reported to be phosphorylated on Ser58 [14,15] and Ser260  and to be substrates for both PKA [1,6,7,14] and CaMKII [11,16]. Similarly to the closely related AAAH TH (tyrosine hydroxylase), TPH1 appears to be activated upon phosphorylation and the phosphorylated enzyme binds to 14-3-3 proteins. It has been shown that phosphorylation targets TPH for degradation  and we have hypothesized that it may affect the solubility and stability of the enzyme . Little is known about the regulation of TPH1, and even less about TPH2. Previously, we have shown that human TPH2 is phosphorylated on Ser19 by PKA , and that additional unidentified phosphorylation sites are present on the enzyme.
The 14-3-3 proteins constitute a family of several closely related, acidic, highly conserved proteins, present in all eukaryotic organisms . These proteins bind to many phosphorylated proteins and have multiple effects in cell-signalling pathways, but they were first identified as activators of TPH and TH [20,21].
In the present study we have investigated the phosphorylation of human TPH2 and its effects on enzyme activity, stability and its binding specificity for 14-3-3 proteins. A series of TPH2 serine/alanine and serine/glutamate mutants were generated, to determine the contribution of phosphorylation at individual sites to enzyme activation, stabilization and interaction between phosphorylation sites.
MATERIALS AND METHODS
Chromatography materials for enzyme purification were purchased from Amersham Biosciences (GE Healthcare), unless otherwise indicated, and all other reagents were from Sigma–Aldrich. BH4 was purchased from Dr B. Schircks Laboratories. Plasmids encoding mammalian 14-3-3 γ, η, ζ, ϵ and His–BMH1 were gifts from Dr Alastair Aitken (School of Biomedical and Clinical Laboratory Sciences, University of Edinburgh, Edinburgh, Scotland, U.K.) and Dr Greg Moorhead (Department of Biological Sciences, University of Calgary, Calgary, AB, Canada).
To analyse TPH2 phosphorylation in human cells by immunoblotting, antibodies against Ser19-phosphorylated TPH2 were generated in sheep (Division of Signal Transduction Therapy, University of Dundee, Dundee, Scotland, U.K.). Phosphopeptides containing the sequence surrounding Ser19 [ARRGF(p)SLDSAV] were used for immunization of sheep and the antibodies were purified on affinity columns with the bound phosphopeptide. Anti-phospho-TPH1 Ser58 and Ser260 antibodies (PA1-4644 and PA1-4645 respectively) were purchased from Affinity Bioreagents.
Construction of plasmids
Missense mutations in Ser19, Ser99, Ser104 and Ser306 were introduced into TPH2 cDNA using the QuikChange® site-directed mutagenesis kit from Stratagene.
Mutagenesis was carried out using the following primers from MWG Biotech: S19A forward, 5′-CGGAGAGGGTTTGCCCTGGATTCAGCAGTGC and reverse, 5′-GCACTGCTGAATCCAGGGCAAACCCTCTCCG; S19E forward, 5′-GCACGGAGAGGGTTTGAACTGGATTCAGCAGTGC and reverse, 5′-GCACTGCTGAATCCAGTTCAAACCCTCTCCG; S99A forward, 3′-GAATCCAGGAAAGCTCGGCGAAGAAGT and reverse, 3′-ACTTCTTCGCCGAGCTTTCCTGGATTC; S99E forward, 5′-GAATCCAGGAAAGAACGGCGA and reverse, 3′-ACTTCTTCGCCGTTCTTTCCTGGATTC; S104A forward, 5′-CGGCGAAGAAGTGCTGAGGTTGAAATCTTTGTGG and reverse, 5′-CCACAAAGATTTCAAGCTCAGCACTTCTTCGCCG; S104E forward, 5′-CGGCGAAGAAGTGAAGAGGTTGAAATCTTTGTGG and reverse, 5′-CCACAAAGATTTCAAGCTCTTCACTTCTTCGCCG; S306A forward, 5′-CGGCATGGCGCAGATCCCCTCTACACCCC and reverse, 5′-GGGGTGTAGAGGGGATCTGCGCCATGCCG; S306E forward, 5′-CCCAGTACATCCGGCATGGCGAAGATCCCCTCTACACC and reverse, 3′-GGTGTAGAGGGGATCTTCGCCATGCCGGATGTACTGGG. The bold type indicates the codons corresponding to the changed amino acid. The sequences of all TPH2 expression clones were verified by DNA sequencing.
Expression and purification of proteins expressed in Escherichia coli
WT (wild-type) and mutant TPH2 proteins were expressed as N-terminal 6×His–MBP (maltose-binding protein) fusion proteins and cleaved with TEV (tobacco etch virus) protease . After cleavage, the proteins were further purified on a Superdex HR 200 column from Amersham equilibrated with 20 mM NaHepes (sodium salt of Hepes) (pH 7.0), 200 mM NaCl and 10% glycerol. The yeast 14-3-3 protein BMH1 was purified by Ni-NTA (Ni2+-nitrilotriacetate) as described previously . The mammalian 14-3-3γ, η and ζ were expressed as GST (glutathione transferase) fusion proteins and 14-3-3ϵ as a fusion with MBP and the proteins were purified as described in .
Phosphorylation of TPH2
For all assays, the stoichiometry (mol of phosphate/mol of TPH subunit) was determined by radioactive labelling using [γ-32P]ATP. TPH2 was phosphorylated by the catalytic subunit of PKA for 60 min at 30 °C, under the following assay conditions: 5 mM MgCl2 and 0.2 mM ATP in 10 mM NaHepes (pH 7.0). In this assay 100 μg/ml PKA C-subunit was used. CaMKII phosphorylation of TPH2 was performed for 30 min at 30 °C, under the following assay conditions: 5 mM MgCl2, 0.2 mM CaCl2, 15 μg/ml calmodulin (Calbiochem), 0.2 mM ATP, 1.5–3.5 μg/reaction of CaMKII and 200 μg/ml TPH2 in 10 mM NaHepes (pH 7.4). Phosphorylation with purified MAPK (mitogen-activated protein kinase; New England Biolabs), MAPKAPK1 and 2 (MAPK-activated protein kinase 1 and 2), MSK1 (mitogen- and stress-activated kinase), PRAK (p38- regulated/activated protein kinase) and PKCζ (protein kinase Cζ; Calbiochem) were performed under the same conditions as with PKA.
Cell extracts and purified proteins were analysed by SDS/PAGE (12% gels), transferred on to a nitrocellulose membrane and immunostained. All three antibodies [the primary monoclonal antibody PH8 (MAB5278; Chemicon), the primary polyclonal antibody anti-TPH2 (ab40846; AbCam) and the secondary goat anti-mouse antibody (172-1011; Bio-Rad)] were applied at a 1:2000 dilution. The phosphospecific antibodies described above were applied at a 1:1000 dilution [anti-(phospho-Ser58) and anti-(phospho-Ser260)] or 0.1 μg/ml [anti-(phospho-Ser19)] and detection was by enhanced chemiluminescence (GE Healthcare). Gel images were acquired and band intensities were measured using Quantity One software (Bio-Rad). In all immunostaining experiments phosphorylated TPH2 was used as an internal standard.
Protein G beads (GE Healthcare) were washed with 50 mM NaHepes (pH 7.0) before affinity-purified anti-(phospho-Ser19) antibody was added to a 50% slurry of beads (5 μg per 30 μl of Protein G) and incubated for 12 h at 4 °C with constant shaking. The beads were washed and resuspended in 0.2 M sodium borate (pH 9.3), exposed to the cross-linking reagent (25 mM dimethyl pimelimidate) for 30 min at room temperature (23 °C), washed for 2 h with 0.2 M ethanolamine (pH 8.0) and finally resuspended in 50 mM NaHepes (pH 7.0) containing 0.1% sodium azide and stored at 4 °C. For immunoprecipitation, chicken brain extracts (1 mg of protein) were incubated with 30 μl of Protein G slurry for 12 h at 4 °C, washed and proteins were eluted with 20 μl of SDS/PAGE sample buffer.
TPH activity was assayed at 30 °C in a standard reaction mixture (100 μl final volume) containing 40 mM NaHepes (pH 7.0), 0.05 mg/ml catalase, 10 μM ferrous ammonium sulfate, 2.5 mM dithiothreitol, 100 μM L-tryptophan and 250 μM BH4, essentially as described previously [18,25,26]. In order to measure kinetic effects of phosphorylation, all activity reactions were performed after an incubation period of 1 h at 30 °C in the presence of PKA or buffer only. Kinetic parameters were calculated by non-linear regression curve fitting using the equation
which allows for determination of a Ksi (substrate inhibition constant).
Measurement of thermoinactivation rates
Enzyme samples (3 μM) in 400 mM NaHepes (pH 7.0) were incubated at 37 °C for different time periods (0, 2, 5, 10, 15, 20 and 30 min) and then assayed for TPH activity as described previously . The data were fitted with exponential decay curves and rate constants were calculated using the equation
where E0 is the initial enzyme activity, Et is the activity at time t, k is the decay rate constant and a represents an apparent plateau value. In the experiments testing the effect of 14-3-3 proteins on TPH2 stability, 120 μg/ml 14-3-3 proteins were incubated with 50 μg/ml TPH2, corresponding to a 4:1 subunit molar ratio; 120 μg/ml purified His–MBP was used as a negative control.
MS, Nano LC-ESI (liquid chromatography-electrospray ionization) Q-TOF (quadruple–time-of-flight) MS/MS (tandem MS)
For two of the kinases used in the present study (PKA and CaMKII), the phosphorylation sites were determined by MS. Analysis was performed using a Q-TOF (Ultima Global) MS equipped with an LC Packings Ultimate nano-HPLC system. The gradient used was 5% solvent B to 80% solvent B over 35 min followed by 10 min at 100% B. The composition of solvents A and B was 0.1% (v/v) formic acid in 2% (v/v) acetonitrile/water and 90% (v/v) acetonitrile/water respectively. The biphasic acetonitrile gradient was run at 200 nl/min on to the analytical column (Reprosil-Pur 5 μm C18 resin, Dr Maisch Gmbh; packed in a 15 cm×75 μm ID fused silica capillary). The Q-TOF was operated in DDA mode using a 1 s MS survey scan. CID spectrum acquisition was allowed for up to a total of 2 s on each precursor ion before a new MS to MS/MS cycle was started. Precursors were excluded from MS/MS experiments for 1 min and singly charged ions were excluded as precursors for MS/MS. Phosphopeptides were enriched using TiO2 affinity resins as described previously [27,28].
Expression of human TPH2 in HEK-293 cells and chicken brain tissue
HEK-293 cells were transfected with the pcDNA5-TPH2 expression vector (WT and mutant forms) using Lipofectamine™ (Gibco) as described by the manufacturer, and the proteins were expressed as described previously [29,30]. The cells were harvested after 48 h, lysed in a buffer containing 50 mM NaHepes (pH 7.0), 1 mM orthovanadate, 50 mM NaF, 0.27 M sucrose, 1% Triton X-100, 1 μM okadaic acid, 10 μM leupeptin and 5 μM pepstatin A before the cell lysates were centrifuged (16000 g at 4 °C for 10 min in an Eppendorf microcentrifuge) and the crude cell extracts were kept in liquid nitrogen until used for SDS/PAGE and immunoblotting and assayed for enzyme activity. Chicken brains were provided by a local chicken slaughterhouse and collected on dry ice until they were dissected and homogenized in the same buffer as used for the HEK-293 cells.
SPR analyses were made using the Biacore 3000 biosensor system (Biacore AB). The BMH1, 14-3-3γ, 14-3-3η, 14-3-3ζ and 14-3-3ϵ proteins  were diluted in 10 mM sodium acetate (pH 5.17, 4.73, 4.73, 5.17 and 4.56 respectively), to a concentration of 0.1 mg/ml and immobilized covalently to the hydrophilic carboxymethylated dextran matrix (Sensor Chip CM5, exposure time ∼10 min) by the standard primary amine-coupling reaction as described by the manufacturer. A reference surface was subjected to the same procedure, but with no protein. A stable baseline was obtained in the cell with immobilized protein by a continuous flow (20 μl/min) of running buffer [10 mM NaHepes (pH 7.4), 150 mM NaCl and 0.005% (v/v) surfactant P20] for approx. 1 h. All binding studies were performed at 25 °C. The phosphorylated and unphosphorylated human TPH2 enzyme in both WT and mutant forms (0.1 mg/ml) were injected in a volume of 60 μl for 3 min at a flow rate of 20 μl/min. The sensograms were analysed in the BIAevaluation version 3.2 program (Biacore AB).
All results presented in graphs are shown as means±S.E.M. Comparisons were analysed using ANOVA. Statistical significance was considered as P<0.05.
RESULTS AND DISCUSSION
Phosphorylation of TPH2 with PKA and CaMKII
Purified human TPH2 was typically phosphorylated to a stoichiometry of 2 mol of phosphate/mol of enzyme subunit after 3 h incubation with the catalytic subunit of bovine PKA at 30 °C (Figure 1). The initial rate of the phosphorylation reaction was rapid, but decreased after a phosphate incorporation of 1 mol of phosphate/mol of subunit had been reached. Further addition of PKA after 1 h phosphorylation did not increase the phosphorylation stoichiometry. Compared with PKA, phosphorylation of TPH2 by rat brain CaMKII occurred slowly (only 15% of the initial rate with PKA). At CaMKII concentrations that gave a stoichiometry of 0.6 mol of phosphate/mol of subunit for TH, TPH2 was phosphorylated to a maximal stoichiometry of 0.2 mol of phosphate/mol of TPH2 monomer after 10 min, but the stoichiometry subsequently slowly decreased (Figure 1). The stoichiometry of CaMKII phosphorylation obtained in the present experiments was similar to that obtained using TPH from rat brain , but lower than that recently observed for recombinant mouse TPH2 .
Phosphorylation of TPH2 by PKA and CaMKII
In addition to PKA and CaMKII, other protein kinases that have previously been shown to phosphorylate TH, which is functionally and structurally related , were assayed for their ability to phosphorylate TPH2. The functional integrity of these protein kinases was first verified using purified human TH1, a well-characterized substrate for all of the protein kinases [33,34]. MAPK, MAPKAPK 1 and 2, MSK1, PRAK or PKCζ all phosphorylated TPH2 to a stoichiometry of <0.1 mol of phosphate/mol of subunit (results not shown). Thus of all protein kinases tested, PKA and CaMKII were the only enzymes demonstrating comparable activities against human TPH2 and TH, suggesting that these kinases could play a part in phosphorylation of TPH2 in cells.
Identification of phosphorylation sites
MS was used to determine PKA- and CaMKII-dependent phosphorylation sites on TPH2. Purified TPH2 that was phosphorylated in vitro with either PKA or CaMKII was subjected to SDS/PAGE. The proteins were then digested either with trypsin or endoproteinase LysC and subjected to analysis by nano-LC MS/MS. Tryptic digestion resulted in identification of 83% of the sequence of TPH2 (GenBank® accession number Q8IWU9, EC 188.8.131.52). As we have shown previously, Ser19 was phosphorylated by PKA  and this site was also phosphorylated by CaMKII. Furthermore, it has recently been shown that this site also is phosphorylated in the mouse enzyme .
PKA-phosphorylated preparations also resulted in the identification of a peptide ion with m/z 947.91, corresponding to the doubly charged, singly phosphorylated TPH2 peptide, phospho-RSSEVEIFVDCECGK (residues 102–116), which fragmented to produce the spectrum shown in Figure 2. The peptide sequence revealed two possible sites of phosphorylation at either Ser103 or Ser104. The y-ions and b-ions together with the neutral loss fragment of the precursor ion (m/z 899.41) strongly suggest that the peptide is phosphorylated, and that the phosphate is positioned on serine 3 of the peptide (neutral loss of b3 ion at m/z 313.16 minus the b2 ion at m/z 244.15) corresponding to Ser104 in the sequence of human TPH2. The phosphopeptide was absent in control experiments where the kinase was excluded, or when the enzyme was phosphorylated with CaMKII.
Analysis of phosphorylation sites of TPH2 by MS/MS
In contrast, we could not find evidence for the phosphorylation of Ser306, corresponding to the previously reported phosphorylation site Ser260 in rabbit TPH1 . However, a peptide with a mass of 3107.49 (expected 3107.50), equivalent to the non-phosphorylated peptide Arg303–Phe332, was isolated in high yields after trypsin digestion.
Ser19 and Ser104 were identified as putative PKA and CaMKII consensus sites using the program NetPhos 2.0 , where Ser99 and Ser104 had the highest score of all serine residues in TPH2. Ser19 is missing in TPH1, but Ser104 corresponds to Ser58 in human TPH1 (Figure 3). In addition to being in agreement with the consensus sequences of PKA and CaMKII , these sites also resemble the mode I binding sites for 14-3-3 binding to target proteins .
Sequence alignment of TPH enzymes from various vertebrates
A sequence comparison of TPH1 and TPH2 sequences from 12 different species that are relatively well-characterized at the protein level revealed that Ser19 and the amino acids in its immediate vicinity are conserved, with small variations in TPH2, from all species sequenced so far. Similarly, the phosphorylation sites corresponding to Ser58 in TPH1 and Ser104 in TPH2 are conserved in most species examined. The phosphorylation site Ser260 in TPH1 (Ser306 in TPH2) is also conserved in most species. However, according to sequence alignment, the most C-terminal of all phosphorylation sites that have been proposed for TPH1, i.e. Ser443 , is replaced by a glycine in mammalian TPH2 using the sequence alignments described below. Thus our results for recombinant human TPH2 have important relevance for the interpretation of earlier studies on TPH from other species [8–13].
Phosphorylation of TPH2 mutants
In the closely related enzyme TH, it is well-established that the N-terminal phosphorylation sites are not phosphorylated independently, but have a strong kinetic interaction , which results in a hierarchical TH phosphorylation. This has been demonstrated using various techniques, including site-specific protein kinases and Ser→Glu mutants that partially mimic phosphorylated serine residues [34,39,40]. Ser→Ala (commonly used to specifically eliminate phosphorylation sites) and Ser→Glu point mutations in TPH2, corresponding to the phosphorylation sites identified in TPH1 or TPH2 were constructed in order to explore the effects of a neutral or negative charge at these phosphorylation sites. In addition, double and triple glutamate mutant enzymes were prepared, i.e. S19E and S104E or S19E, S104E and S306E.
The S99A, S99E, S306A and S306E mutants of TPH2, either alone or in combination with the other mutations, were enzymatically active, but difficult to purify, as their MBP fusion proteins precipitated during purification. Thus only a partial characterization of these proteins was possible and the destabilizing effects produced on modification of these sites could indicate that they are not likely to be responsible for activation of TPH2 by phosphorylation. In contrast, the other serine mutants of TPH2 appeared to be in a native folded state, as shown by their gel-filtration profiles which were identical with the WT. They were soluble, stable and were purified in yields that were either similar or higher than those obtained for the WT. Furthermore, the kinetic values of the mutant proteins were comparable with those of the WT (Table 1).
|Enzyme form||Vmax (nmol/min/mg)||Km (μM)||Vmax (nmol/min/mg)||Km (μM)|
|Enzyme form||Vmax (nmol/min/mg)||Km (μM)||Vmax (nmol/min/mg)||Km (μM)|
The TPH2 mutants were incubated with PKA under conditions where the incorporation of 1.5–1.8 mol of phosphate into the WT enzyme was achieved. As shown in Figure 4, all of the proteins were phosphorylated at lower initial rates than the WT. TPH2 S19A, TPH2 S19E and TPH2 S104A were phosphorylated to a final stoichiometry of approx. 30% of the WT TPH2, whereas TPH2 S104E was phosphorylated with a stoichiometry of approx. 60% of the WT. Interestingly, for Ser19 mutants the initial rates of TPH2 phosphorylation were significantly lower for the Ser→Ala than for the Ser→Glu mutant (P=0.02; one-way ANOVA). Although the S19E substitution does not lead to a higher final incorporation of phosphate into other site(s), it gives an almost similar initial phosphorylation rate as for the S104E mutant protein. Surprisingly, the double mutant TPH2 S19E/S104E was phosphorylated by PKA to a similar stoichiometry as the single Ser→Glu mutants (results not shown). It remains to be established whether the putative third phosphorylation site is present in native TPH2, or represents a site only accessible due to effects associated with the phosphomimicking mutant enzyme.
Phosphorylation of TPH2 mutants
Similar experiments were also performed using CaMKII phosphorylation. As for the phosphorylation with PKA, the Ser→Glu mutants of Ser19 and Ser104 were phosphorylated 1.5–2.5-fold faster than the Ser→Ala counterparts. This indicates that this modification of TPH2 increases its rate of phosphorylation, independently of phosphorylation conditions. However, CaMKII appears to have a higher selectivity for Ser19 than PKA, as the S19A and S19E mutants incorporated only 14–19% of the amount of phosphate relative to the WT TPH2, whereas TPH2 S104E incorporated 82% of the amount of phosphate, compared with the WT (results not shown).
Phosphospecific antibodies against Ser58 and Ser260 (TPH1) are commercially available, and we generated a sheep antibody towards phospho-Ser19 TPH2. Ser104 and Ser306 in TPH2, correspond to Ser58 and Ser260 in human TPH1 respectively. Thus the identification of Ser19 and Ser104 as PKA phosphorylation targets in purified TPH2 was verified using Western blotting (Figures 5A and 5B). The antibody directed against phospho-Ser260 (Ser306 in TPH2) did not have immunoreactivity for PKA-phosphorylated enzyme (results not shown), providing further evidence that Ser306 is not a PKA phosphorylation site in TPH2. These experiments also confirmed that Ser19 was the major phosphorylation site recognized in TPH2 by both PKA and CaMKII. For CaMKII, >85% of the total immunoreactivity was assigned to the Ser19 site, and the antibody reactivities against phosphorylated Ser104 and Ser306 were on the level of the non-phosphorylated enzyme (results not shown). In addition, we analysed the Ser→Ala and Ser→Glu mutants of Ser19 and Ser104 and could demonstrate that these mutations blocked the immunoreactivity for its corresponding phosphospecific antibodies (Figures 5A and 5B).
Phosphorylation of purified TPH2 in vitro (A and B), in chicken lysate (C) and in intact HEK-293 cells (D)
Together, these results suggest that the presence of a negative charge at either Ser19 or Ser104 strongly facilitates the phosphorylation at the remaining site(s). This could imply that TPH2 can respond synergistically to combinations of cellular signalling inputs, i.e. that cAMP- and Ca2+-mediated inputs on the protein are mutually stimulatory. For TH, mutual stimulation is not found, rather it has been found that a Ca2+-mediated phosphorylation of Ser19 would potentiate a cAMP-mediated response (Ser40 phosphorylation), but not in the reverse order .
Effects of phosphorylation on TPH2 enzyme kinetics
Using substrate concentrations at or slightly above the physiological range (20 and 100 μM L-tryptophan and BH4), the enzyme activity of PKA-phosphorylated preparations of TPH2 was on average 30% higher than the non-phosphorylated enzyme. Based on this observation, a systematic study of the kinetic properties of phosphorylated and non-phosphorylated enzyme was undertaken, with the aim of disentangling the contributions from the different phosphorylation sites. Phosphorylation of purified WT TPH2 by PKA produced opposite but modest effects on the kinetic constants measured for L-tryptophan and BH4. The Vmax was relatively unaffected, the Km value for tryptophan was decreased by 20% and the Km for BH4 was increased by the same amount. A statistically significant effect was only observed for BH4, due to higher precision for that measurement (Table 1).
The site-specific effects of phosphorylation were studied using the phosphomimicking Ser→Glu mutants. The greatest effect was observed for the S104E mutant, which had reduced catalytic efficiency at standard assay conditions and gave higher yields upon purification than the other mutants (results not shown). The reduced affinity constant (Vmax/Km) of this mutant is mainly due to a 4-fold increase in the Km for BH4 compared with the WT. Smaller effects were observed for the kinetic properties of the S19E mutant, and they were different for tryptophan and BH4. Thus an apparent activation (increased Vmax, reduced Km) was found for the substrate (L-tryptophan) whereas the opposite was found for the cofactor (BH4).
We also studied Ser→Ala mutants to reveal any effects that might be related to mutation of serine residues but not specific to glutamate. We observed reduced catalytic activity for all the Ser→Ala mutants. We suspect that this reflects a decreased stability of the Ser→Ala enzymes, since the yields of purified protein upon purification were lower than for the Ser→Glu mutants (results not shown). Surprisingly, the S104A and S104E mutants showed similar changes in their kinetic properties, mainly an increase in the Km for the cofactor. Phosphorylation of the S19A mutant resulted in a significant decrease in the Vmax for BH4 (28%) whereas phosphorylation of S104A resulted in a significant increase in the Vmax for L-tryptophan (56%) and BH4 (15%).
Assuming that glutamate is partially mimicking a phosphorylated serine, we conclude that phosphorylation of Ser19 is mainly responsible for the increased Vmax (Trp) of TPH2, whereas phosphorylation of Ser104 has a neutral or slightly negative effect on the Vmax (Trp). Such opposite effects of phosphorylation at Ser19 and Ser104 could explain the relative insensitivity of the Vmax of the WT to enzyme phosphorylation. Our results for the alanine/glutamate mutants suggest that phosphorylation of Ser104 may have an important effect on the Km for BH4. Thus Ser104 may also be important for the stability and kinetic properties of the enzyme.
As the S19E and S104E mutants had higher solubility than the S19A, S104A or WT enzymes (see below), the effects of mutations on enzyme kinetics could also be mediated by an increase in the amount of native enzyme being present during the phosphorylation reaction, rather than being due to a direct effect on the active site. Thus compared with the moderate enzyme activation produced by phosphorylation reported in previous studies [8–13] and in the present study (Table 1 and Figure 8), the impact of phosphorylation on TPH2 stability is much more important.
Phosphorylation of TPH2 in HEK-293 cells and brain tissue
In order to determine whether the enzyme is phosphorylated in vivo in brain tissue, we performed immunoprecipitation of brain lysates with antibodies directed against Ser19-phosphorylated TPH2. Chicken brain was selected for these studies, as the region surrounding Ser19 is identical in human and chicken TPH2, whereas Phe18 is replaced by a leucine residue in the rat and mouse sequences (Figure 3). As shown in Figure 5(C), a weak band corresponding to phosphorylated TPH2 appeared in lysates of chicken brainstem, a region known to contain large amounts of TPH2 mRNA in humans  whereas this band was not apparent in lysates of cerebellum (results not shown), a brain region that contains lower amounts of TPH2 .
Since only a weak band was detected when immunoprecipitating phosphorylated TPH2 in brain tissue, further study was needed in order to determine whether the enzyme is phosphorylated in intact human cells. TPH2 was transiently expressed in HEK-293 cells and as recently shown, these cells do not express TPH2 when transfected with an empty vector, but have a high expression level when transfected with vectors containing either WT or mutant forms of TPH2 [22,42]. As shown in Figure 5(D), a strong band corresponding to TPH2 was observed in Western blots of cell lysates using the anti-(phospho-Ser19) antibody, showing that this site was phosphorylated in HEK-293 cells. The band intensity was further increased 2-fold after incubation with an excess of PKA, indicating that the Ser19 phosphorylation stoichiometry was maximally 0.5 mol of phosphate/mol of enzyme subunit in resting cells. A modest (10–20%), but significant (P=0.044; one-way ANOVA), increase in Ser19 phosphorylation was observed after treatment with okadaic acid (protein phosphatase inhibitor) and forskolin (adenylate cyclase activator), but not by okadaic acid alone. The physiological TPH2 protein phosphatases have not been identified, but it has been shown that PP2A (protein phosphatase 2A) is the major TH phosphatase and its inhibition by okadaic acid strongly increases the level of TH phosphorylation. HEK-293 cells express both CaMKII, PKA, PP2A and 14-3-3 proteins including 14-3-3γ , and these proteins could be involved in both the phosphorylation and the activation processes of TPH2.
In order to determine whether the increase in phosphorylation stoichiometry was accompanied by an increased TPH activity, enzyme assays were also performed. As the Km value for tryptophan could be altered upon phosphorylation, TPH activity in cell lysates was determined using two different concentrations of L-tryptophan (20 and 100 μM). Under both conditions, the enzyme activity increased by 10–17% in the presence of okadaic acid and 11–12% in the presence of okadaic acid and forskolin. In both instances, the changes were only marginally significant, as determined by one-way ANOVA, but interestingly these moderate changes corresponded well with the small increase in TPH Ser19 phosphorylation stoichiometry. Using Western blotting with the anti-TPH2 antibody, we also showed that the total amount of TPH2 was unchanged after treatment with either okadaic acid or forskolin during 20 min, as was also the case with the TPH2 mRNA levels (<4.5% difference from the control cells). The high level of endogenous Ser19 phosphorylation of TPH2 transfected in HEK-293 cells (Figure 5D) could explain why additional stimulation of cAMP phosphorylation by forskolin only gave a weak additional phosphorylation and enzyme activation. As the commercial antibodies directed against phospho-Ser104 (Ser58 in TPH1) and phospho-Ser306 gave multiple bands on blots of crude cell lysates, it was not possible to conclude whether these sites were also phosphorylated in HEK-293 cells or brain tissue.
Binding of 14-3-3 proteins to phosphorylated TPH2
Binding of 14-3-3 to TPH and TH has been reported to increase their hydroxylase activity , the first biological activity assigned to 14-3-3 proteins [20,21], but the binding specificity of phosphorylated TPH2 is not known. When the interactions of recombinant TPH2 and different 14-3-3 isoforms were examined using SPR, we found very little binding to non-phosphorylated TPH2 (Figure 6), as previously reported . However, after phosphorylation of TPH2 by PKA, the binding of BMH1, 14-3-3γ and MBP–14-3-3ϵ to TPH2 increased dramatically (Figure 6A), in particular for BMH1. As shown in Figure 6(A), the binding affinity appeared in the following order: 14-3-3ϵ>14-3-3γ≥BMH1, with apparent dissociation constants (KD) ranging from 1 to 170 nM. Only weak binding to bovine 14-3-3ζ and η was observed for all of the TPH2 constructs examined (results not shown).
Association between different 14-3-3 isoforms and human TPH2 (WT and mutants)
In contrast with the WT enzyme, neither phosphorylated TPH2 S19E nor TPH2 S19A bound more strongly to 14-3-3 proteins than non-phosphorylated TPH2 (Figure 6B). However, when analysing the phosphorylated S104E mutants, the binding intensity to the 14-3-3γ increased 50-fold relative to the non-phosphorylated S104E mutant enzyme. Thus phosphorylation-dependent binding to 14-3-3s was observed for all forms of TPH2 having an intact Ser19 phosphorylation site. Although phosphorylation of Ser104 did not appear to be directly involved in 14-3-3 binding to TPH2, this phosphorylation site could contribute indirectly by increasing enzyme stability and stoichiometry of Ser19 phosphorylation.
The negative charge introduced in the Ser→Glu mutants of Ser19 and Ser104 had a clear effect on TPH2 activity and stability and influenced its rate of phosphorylation. However, none of the glutamate mutants were able to bind to 14-3-3 proteins with affinities comparable with that of the phosphorylated enzyme, as observed for other aspartate or glutamate substitutions of phosphoserine residues [44,45] that also regulate 14-3-3 binding.
Effect of binding of 14-3-3 on the enzyme kinetics of phosphorylated TPH2
Addition of the 14-3-3 proteins BMH1, or MBP–14-3-3ϵ increased the catalytic activity of phosphorylated TPH2, but had no effect on the non-phosphorylated enzyme (Figure 7). Furthermore, His-tagged MBP had no effect on TPH activity on either phosphorylated or non-phosphorylated TPH2 at identical protein concentrations. The relative ability of 14-3-3s to stimulate the activity of TPH2 correlated with their binding affinities to phosphorylated TPH2. Enzyme assays performed either at saturating concentrations of substrates (500 μM) or below the Km value for either tryptophan or BH4 gave similar results, indicating that the enzyme activation was due to an increase in the Vmax value of TPH2.
The effect of 14-3-3 proteins on the activity of TPH2 phosphorylated by PKA
In the experiment shown in Figure 7, a maximal stimulatory effect of the 14-3-3 proteins was observed at approx. 50 μg/ml, a concentration that was also able to significantly stabilize the protein (see below). The identity of the phosphorylation site responsible for the enzyme activation was determined using the Ser→Glu and Ser→Ala mutants of TPH2. The S19E mutant was not activated by phosphorylation nor addition of 14-3-3 (results not shown), whereas enzymes having an intact Ser19 phosphorylation site (TPH2 WT and TPH2 S104E) were activated (Figures 7A and 7B). Similarly, the mutants TPH2 S19A and TPH2 S19E/S104E were not activated by 14-3-3 proteins, whereas TPH2 S104A was activated to the same extent as the WT (results not shown). Thus we conclude that phosphorylation of Ser19 is mainly responsible for the activation of TPH2 by phosphorylation and 14-3-3 binding. The ability of 14-3-3 proteins to further activate the enzyme appears to vary between the isoforms (Figure 7). Thus the extent of TPH activation found in previous studies is dependent not only on 14-3-3 concentrations, but also on the specific isoforms used.
Effect of 14-3-3 binding on TPH2 stability
Binding of 14-3-3 has also been reported to either increase or decrease the stability of different target proteins [46,47]. To determine whether this is the case for TPH2, thermostability assays were performed. TPH2 proteins were first phosphorylated, and then both the phosphorylated and the non-phosphorylated TPH2 were incubated at 37 °C for various periods of time in the presence or absence of 14-3-3 proteins and further assayed for activity. After 60 min incubation non-phosphorylated TPH2 lost 65% of its activity, compared with 50% for the phosphorylated enzyme. In comparison, in the presence of BMH1, phosphorylated TPH2 had an activity loss of 35%, but in the presence of 14-3-3ϵ, the activity loss was only 15% (Figure 8). Control experiments showed that non-phosphorylated TPH2 was also stabilized by the 14-3-3 proteins, but to a much lower extent than the phosphorylated enzyme (P=0.01; one-way ANOVA).
The effect on enzyme stability of 14-3-3 binding to phosphorylated TPH2
Stabilization of phosphorylated TPH2 by 14-3-3 proteins may be more important for its molecular regulation and for its function in serotonin biosynthesis. Considering the many early reports on TPH regulation by phosphorylation, where most studies were performed on impure TPH in cell extracts, the observed enzyme activation could have been due to the presence of 14-3-3 proteins in the extracts. Interestingly, 14-3-3 binding also has a stimulatory effect on the further conversion of serotonin into melatonin . Phosphorylation of TPH1 has been reported to trigger the proteasomal degradation of the enzyme , but it is not known whether this effect is mediated by 14-3-3 proteins and whether phosphorylation also affects TPH2 turnover in intact cells.
The results of the present study show that human TPH2 is phosphorylated in vitro to a stoichiometry of at least 2 mol of phosphate/mol of subunit by PKA  and to a lower stoichiometry by several other serine/threonine-specific protein kinases. Both Ser19 and Ser104 have been identified as phosphorylation sites by MS, phosphospecific antibodies and mutagenesis. A negative charge at Ser104 appears to facilitate phosphorylation on the other site(s) whereas Ser19 seems to be the most important phosphorylation site for activation and 14-3-3 binding. Binding of 14-3-3 to phosphorylated TPH2 increases enzyme stability and thereby activates TPH2. Furthermore, we found a strong positive correlation between the amount of bound 14-3-3 to phosphorylated TPH2, the affinity of binding and the stimulation of enzyme activity.
We greatly appreciate the technical expertise of Sidsel E. Riise and Ali S. Muñoz. The TPH2 plasmid was prepared from expression vectors provided by Gunter Stier (EMBL Heidelberg, Heidelberg, Germany) and the 14-3-3 plasmids were provided by Dr Alistair Aitken, University of Edinburgh. This work was supported by grants from the Research Council of Norway, Locus of Neuroscience and Helse-Vest.
aromatic amino acid hydroxylase
Ca2+/calmodulin-dependent protein kinase II
human embryonic kidney
mitogen-activated protein kinase
- MAPKAPK1 and 2
MAPK-activated protein kinase 1 and 2
mitogen- and stress-activated kinase 1
sodium salt of Hepes
protein kinase A
protein kinase C
protein phosphatase 2A
p38-regulated/activated protein kinase