Functional Domains of Human Tryptophan Hydroxylase 2 (hTPH2)*

Tryptophan hydroxylase (TPH) is the rate-limiting enzyme in serotonin biosynthesis. A novel gene, termed TPH2, has recently been described. This gene is preferentially expressed in the central nervous system, while the original TPH1 is the peripheral gene. We have expressed human tryptophan hydroxylase 2 (hTPH2) and two deletion mutants (NΔ150 and NΔ150/CΔ24) using isopropyl β-d-thiogalactopyranoside-free autoinduction in Escherichia coli. This expression system produced active wild type TPH2 with relatively low solubility. The solubility was increased for mutants lacking the NH2-terminal regulatory domain. The solubility of hTPH2, NΔ150, and NΔ150/CΔ24 are 6.9, 62, and 97.5%, respectively. Removal of the regulatory domain also produced a more than 6-fold increase in enzyme stability (t½ at 37 °C). The wild type hTPH2, like other members of the aromatic amino acid hydroxylase superfamily, exists as a homotetramer (236 kDa on size exclusion chromatography). Similarly, NΔ150 also migrates as a tetramer (168 kDa). In contrast, removal of the NH2-terminal domain and the COOH-terminal, putative leucine zipper tetramerization domain produces monomeric enzyme (39 kDa). Interestingly, removal of the NH2-terminal regulatory domain did not affect the Michaelis constants for either substrate but did increase Vmax values. These data identify the NH2-terminal regulatory domain as the source of hTPH2 instability and reduced solubility.

Tryptophan hydroxylase (TPH 2 ; EC 1.14.16.4), a tetrahydropterin (BH 4 )-dependent amino acid hydroxylase, is the key regulator of serotonin (5-hydroxytryptamine) biosynthesis. Utilizing BH 4 and O 2 as co-substrates and Fe 2ϩ as a cofactor, TPH hydroxylates tryptophan to 5-hydroxytryptophan (Fig. 1). Subsequent decarboxylation of 5-hydroxytryptophan by amino acid decarboxylase generates serotonin (1)(2)(3). This essential monoamine has been found in a variety of tissues and impli-cated in a wide range of physiological functions. In the central nervous system, serotonin is synthesized primarily in the dorsal raphe nucleus and acts as a neurotransmitter; in the pineal gland, it serves as a precursor for melatonin biosynthesis (4). In the periphery, serotonin constricts large blood vessels and regulates platelet adhesion (5)(6)(7). Serotonin, produced by the enterochromaffin cells of the intestinal system, initiates peristaltic and secretory reflexes (8). Altered activity of serotonin is associated with various disorders such as depression, impulsive behaviors, aggression, suicide, drug abuse and alcoholism, sleep disorders, gastrointestinal diseases (such as irritable bowel syndrome), and cardiovascular dysfunction leading to heart failure (9).
Recently, Walther et al. (10) discovered the gene responsible for nervous system TPH (TPH2). Unlike TPH1 now known to be the peripheral enzyme (11), TPH2 is neuron-specific and expressed predominantly in serotonergic neurons of the raphe nuclei (10,12) and in the peripheral myenteric neurons in the gut (13). Human TPH1 and TPH2 display 72% sequence homology and have high sequence identity within the COOHterminal catalytic domain. The chromosomal location of these sister enzymes differ; human TPH1 is located on chromosome 11, while TPH2 is on chromosome 12.
Full-length human TPH1 has been expressed and purified from Escherichia coli and Pichia pastoris (14). However, TPH1 is a notoriously unstable enzyme (15), and formation of insoluble inclusion bodies has limited its purification in large quantities from E. coli (14,16). While there are a number of important and informative crystal structures for the related enzymes phenylalanine hydroxylase and tyrosine hydroxylase, there is single reported structure for TPH1 from Stevens and colleagues (17). This crystal structure, hypothetical models based on other hydroxylase structures (18,19), and NMR studies (20) provide insight into the active site of TPH1 and the respective binding sites for substrates.
The recent discovery and cloning of TPH2 offers new opportunities for understanding serotonin physiology. TPH2 has a divergent NH 2 -terminal regulatory domain, which shows more homology to that of TH (16). Chimerical tyrosine/tryptophan hydroxylase constructs have suggested that the tyrosine hydroxylase regulatory domain serves to stabilize the enzyme (21); therefore, recombinant TPH2 may exhibit more stability compared with TPH1. Additionally, a number of single nucleotide polymorphisms have already been identified in the TPH2 gene that are associated with mental health disorders (22)(23)(24)(25)(26).
Therefore, it is a priority to characterize structure/function relationships within this enzyme to better understand its regulation and role in human health and disease. Indeed, a report from the Haavik laboratory describes expression of human TPH2 (27), while a recent report from Kuhn and colleagues describes the expression of recombinant mouse TPH2 (mTPH2) and the functional characterization of a naturally occurring coding region polymorphism (28).
In this study, we describe expression, purification, and characterization of recombinant hTPH2 and truncation mutants from E. coli using a novel autoinduction procedure (29).

EXPERIMENTAL PROCEDURES
Bacterial Expression and Purification of hTPH2-The bacterial expression construct for hTPH2 was constructed in two steps. First the internal NdeI and BamHI sites were removed through the introduction of silent mutations (using the QuikChange TM site-directed mutagenesis kit; Stratagene, La Jolla, CA). To remove the internal BamHI site the following primers were used: 5Ј-CATGTTCCACTACTTGCGGGTCCTAAG-TTTGCTCA-3Ј (forward) and 5Ј-AACTGAGCAAACTTAGG-ACCCGCAAGTAGTGGAACATG-3Ј (reverse). To remove the internal NdeI site, the following primers were used: 5Ј-GCAAC-TGCGGGCACATGGAGCAGGACTCCTTTCC-3Ј (forward) and 5Ј-AAAGGAGTCCTGCTCCATGTGCCCGCAGTTGC-CCTTC-3Ј (reverse). These changes were constructed through mutagenesis using the QuikChange side-directed mutagenesis kit. Second, an NdeI site was introduced at the start codon and a BamHI site after the TAA stop codon. (The primers were 5Ј-AAGCGACTGGACATATGCAGCCAGC-3Ј and 5Ј-GCA-ACGGATCCTCAAATCCCCAGATATTGG-3Ј). The resultant PCR product was cut with NdeI and BamHI and cloned into the equivalent sites in pET28-tev (a generous gift from Dr. W. Studier) (29). All primers were provided by the DNA Synthesis Core Laboratory of the Pennsylvania State University College of Medicine. PCR conditions were are follows: denaturation at 95°C, annealing at 55°C, extension at 68°C, and 20 cycles (QuikChange) or 35 cycles (amplification).
This recombinant vector contains an NH 2 -terminal hexahistidine tag followed by a cleavage recognition site for tobacco etch virus (TEV) nuclear inclusion la protease. The recombinant construct was then expressed in BL21-CodonPlus(DE3)-RIL (E. coli) cells (Stratagene) in ZYP-5052 autoinduction medium plus 50 M L-tryptophan. Grown at low temperatures (15°C), these cells divide until they exhaust their glucose source and then reliably induce recombinant protein expression at very high titers (see Figs. 2 and 3) (29).
TPH2-positive colonies were grown in ZYP-5052 autoinduction medium at 15°C in an orbital shaker at 300 rpm. The cultures were sampled every 3 h between 33 and 51 h of culture, and their optical densities were measured at 600 nm. In parallel, hTPH2 enzyme activity was monitored from aliquots removed during the time course of enzyme induction. Following induction, E. coli were harvested by low speed centrifugation. The cell pellet was lysed with BugBuster amine-free with benzonase and rLysozyme (EMD Biosciences, Inc., Madison, WI) in the presence of EDTA-free protease inhibitor mixture (Complete EDTA-free; Roche Diagnostics GmbH). The lysate was then centrifuged at high speed (40,000 ϫ g) to obtain soluble proteins. The clarified lysate was subjected to a rapid one-step metal chelate affinity chromatography procedure using either a 5-or 1-ml prepacked HiTrap nickel column (Amersham Biosciences) and an Akta FPLC (Amersham Biosciences) work station at 4°C. Columns were first equilibrated with 25 mM imidazole and His-tagged protein bound. Recombinant protein was then eluted with a phosphate buffer containing 500 mM imidazole and 1 mM dithiothreitol. Eluted proteins were then dialyzed using a Slide-A-Lyzer dialysis cassette with a 10,000 molecular weight cut-off (Pierce) while undergoing buffer exchange into storage buffer containing 10% glycerol at 4°C. Attempts to cleave the hexa-His fusion peptide (26 amino acids) were unsuccessful as (a) the enzyme is highly unstable and lost activity during the digestion, and (b) the recombinant hTPH2 fusion is a poor substrate for TEV protease resulting in incomplete digestion even when digestion was carried out with 1:20 ratio of TEV to hTPH2. In these preliminary experiments, treatment of the soluble protein with TEV protease produced less than 50% cleavage. More importantly, the full-length enzyme was inactive following the incubation. Therefore, the experiments in this report were conducted with the intact fusion protein. The activity of the fusion enzyme is comparable with the wild type TPH1 (30) and TPH2 (28) enzymes (both fusion and native versions) as characterized previously.
Enzyme Activity Assay-TPH2 activity was assayed by using a radioenzymatic 3 H 2 O release assay as previously described by Reinhard et al. (31) and modified by Vrana et al. (32). Activity values derived from each assay were normalized to total protein present in the homogenates, as determined by the BCA protein assay (Pierce), and were expressed as nmol/h/mg. All materials for activity assays were obtained from Sigma with the exception of activated charcoal (Darco G-60 from Fisher Scientific).
Michaelis-Menten kinetic analyses were conducted by varying substrate (tryptophan or BH 4 ) concentrations at a fixed concentration of the other substrate (50 M BH 4 and 50 M tryptophan, respectively) and a fixed and ambient concentration of oxygen. Resulting data were converted to specific enzyme activities (correcting for varying radioactive amino acid concentrations, as nec-FIGURE 1. Tryptophan hydroxylase catalyzes the rate-limiting step in serotonin (5-hydroxytryptamine) biosynthesis using tetrahydroiopterin (BH 4 ) and dioxygen co-substrates. The subsequent and final reaction in serotonin biosynthesis is catalyzed by the aromatic amino acid decarboxylase (AAADC ). essary) and analyzed by Prism (GraphPad Software, San Diego, CA) to determine kinetic constants (K m , V max ).

SDS-PAGE and Western Blot Analysis-Expression of recombinant proteins and confirmation of molecular weights
were evaluated by SDS-PAGE and Western blot analysis. Proteins were resolved on 4 -12% NuPAGE BisTris gels (Invitrogen) run in MOPS buffer at 200 V, 115 A for 50 min. Gels were stained with Coomassie Blue (Pierce) to visualize total protein and to assess purity. In addition, gels were transferred to polyvinylidene fluoride membrane (Immobilon-P, Millipore) using a semidry electroblot apparatus (Owl Scientific, Cambridge, MA). Human TPH2 was detected by probing with a 1:1500 dilution of an affinity-purified sheep anti-TPH monoclonal antibody (Chemicon, Temecula, CA), which recognizes both TPH1 and TPH2. A 1:3000 dilution of donkey anti-sheep IgG secondary horseradish peroxidase conjugate (Upstate Biotechnology) was used to detect recombinant protein following chemiluminescence (ECL; Amersham Biosciences).
Enzyme Mutations by Deletion Analysis-As described in Fig. 3, two different deletion mutations were constructed. For the N⌬150 mutation, an NdeI restriction sequence was inserted into the wild type hTPH2 between amino acids 150 and 151 using the following primers: 5Ј-GAAGAAGAGCTAGAGGATCATATGGTGCCC-TGGTTCCCTCG-3Ј (forward) and 5Ј-CGAGGGAACCAGGG-CACCATATGATCCTCTAGCTCTTCTTC-3Ј (reverse). Then this mutated plasmid was digested with NdeI and religated (elimination the first 150 codons). The N⌬150/C⌬24 deletion construct was created by inserting a stop codon (TGA) into the N⌬150 mutation from the COOH terminus between amino acids 24 and 25 (residues 151-476). This was accomplished with the following primers: 5Ј-AGTATTGAAAATGTGTGATAGGTGCAGGAC-CTTCGC-3Ј (forward) and 5ЈGCGAAGGTCCTGCACCTA-TCACACATTTTCAATACT-3Ј (reverse). Deletion mutagenesis for each construct was carried out, in a termocycler, under the following conditions: denaturation at 95°C, annealing at 55°C, extension at 68°C, and 20 cycles using QuikChange TM site-directed mutagenesis kit. The hTPH2 sequences were determined to confirm the desired deletions and to eliminate unexpected mutations. To delete the regulatory and/or subunit assembly domains, hTPH2 was truncated at the amino terminus (N⌬150) and both the amino and carboxyl termini (N⌬150/C⌬24). Each construct was subjected to the same procedures as described for the wild type construct throughout the study.
Enzyme Solubility-Aliquots of E. coli (5 ml) expressing wild type hTPH2 and deletion mutants (10 OD units/ml) were centrifuged at 5000 ϫ g for 10 min. After discarding the supernatant, each pellet was lysed in 1 ml of lysis solution (described above) and incubated on ice for 30 min. Aliquots (3 l) were taken from each lysate sample and designated as total protein homogenate. The remaining lysate materials were centrifuged at 40,000 ϫ g for 20 min, and aliquots (3 l) of the supernatants were taken from each sample and designated as high speed supernatant. Both total and corresponding high speed supernatants were resolved on 4 -12% NuPAGE BisTris gel. The gel was stained with Coomassie Blue to visualize the protein. The band corresponding to hTPH2 was scanned, the local background subtracted (as determined in a bacterial control), and the amount of soluble hTPH2 expressed as a percentage of that present in the homogenate.
Enzyme Stability-Purified wild type hTPH2 and deletion mutants were incubated at 37°C, and samples were removed and placed on ice every 15 min from initiation to 180 min. Enzyme activity was then determined as described above and analyzed as a function of activity decay (on a logarithmic plot).
Statistical Analysis-The TPH activity determinations were analyzed using GraphPad-4 one-way analysis of variance followed by an unpaired Student's t tests. Statistical significance was associated with values of p Ͻ 0.05.

RESULTS
Active TPH2 Expression following Autoinduction-The expression system in the present study takes advantage of a recent report by Dr. W. Studier (Bookhaven National Laboratories) that permits the induction of expression from pETbased expression plasmids in an IPTG-free medium (29). We compared autoinduction and IPTG expression systems at two different incubation temperatures: 15 and 25°C. In both cases, significantly more soluble protein was obtained at 15°C compared with 25°C. For the autoinduction system, this corresponded to a 15-fold increase in active protein, and for IPTG induction it was a 3.5-fold increase. Comparison of the relative yields of the two induction systems at 15°C showed that autoinduction resulted in 11.8-fold more active hTPH2 protein compared with IPTG induction of mid-log growth cells (data not shown). The onset of the autoinduction of the expression of the wild type hTPH2 construct was determined to be 33 h after the initiation of the culture (Fig. 2). Activity increased dramat- Human TPH2 Domain Structure SEPTEMBER 22, 2006 • VOLUME 281 • NUMBER 38

JOURNAL OF BIOLOGICAL CHEMISTRY 28107
ically for the subsequent 7 h and reached a plateau. Cell growth continued for an additional 10 h reaching a maximum optical density at 600 nm of ϳ14; however, the specific enzyme activity per unit of optical density of the culture declined. This is consistent with the observations reported by McKinney et al. (27). Based upon our growth/activity measurements, E. coli cultures expressing TPH2 and TPH2 variants were harvested 45 h after the initiation of the culture for the following purification and activity studies. The onset of the autoinduction was observed to be similar in other E. coli cultures that express hTPH2 deletion constructs (data not shown).
Although the expression of active enzyme was found to be high, this system was limited in its production of large quantities of soluble hTPH2 as determined by the SDS-gel electrophoresis of the lysate (crude sample) and high speed supernatant (soluble proteins) (Fig. 3B). This is similar to what we have seen with other aromatic amino acid hydroxylases such as TH and TPH1. A number of attempts to increase solubility by harvesting cells at earlier time points or growing the cells at lower temperatures did not further improve solubility (data not shown). Based on our previous studies and published reports by others, it is our belief that the limited solubility of full-length hTPH2 is due the NH 2 -terminal regulatory domain of the enzyme. Therefore, the next series of the studies tested this hypothesis by deleting regions of the enzyme.
Deletion of Both Regulatory and Tetramerization Domains Results in Higher Solubility-In concordance with the notion that the NH 2 -terminal regulatory domain decreases solubility, deletion studies showed an increase in TPH1 solubility. The highest solubility was achieved when both regulatory and tetramerization domains were truncated. Deletion of the regulatory domain resulted in dramatically increased solubility, while deletion of the amino and carboxyl termini generated nearly completely soluble enzyme (Fig.  3B). Solubilities of the three recombinant proteins were estimated to be 6.9, 62, and 97.5% for full-length hTPH2, N⌬150, and N⌬150/C⌬24, respectively. Interestingly, deletion of just the COOH-terminal tetramerization domain did not increase the solubility or stability (data not shown).
A major advantage of the pET28-tev expression system is the ability to conduct a rapid purification following a singlestep Ni 2ϩ affinity column. Both N⌬150 and N⌬150/C⌬24 deletion products were purified with high efficiency as detected by Coomassie staining (Fig. 4, A and B). By contrast, the full-length, wild type hTPH2 purifies as a minor component (Fig. 4, A and B) that can then be enriched by size exclusion chromatography (Fig. 4C). Unfortunately, based on the instability of the wild type hTPH2 (see Fig. 5), even this relatively rapid, two-step approach produces inactive enzyme (data not shown). The enhanced purity of the NH 2terminal truncation mutants stems largely from their greater solubility and consequently they represent a higher fraction of the soluble protein in extracts compared with constructs containing the NH 2 -terminal domain.
Stability Analysis-Inactivation time courses were conducted for the recombinant enzymes (Fig. 5). Half-lives for inactivation at 37°C were found to be 30, 203, and 147 min for FIGURE 3. A, the wild type hTPH2 was truncated at the amino terminus (N⌬150) or both amino and carboxyl termini (N⌬150/C⌬24) to delete the regulatory (R) or the subunit assembly (tetramerization) domains, respectively. C, catalytic domain. B, solubility of wild type hTPH2, its amino (N⌬150) and amino/carboxyl (N⌬150/C⌬24) deletion mutants expressed in ZYP5052-autoinduction medium. Upon lysis of the host bacteria, total homogenate protein (H) and high speed supernatant protein (S) samples were prepared as described under "Experimental Procedures." The samples were then resolved on denaturing polyacrylamide gels and were Coommassie-stained. BenchMark PreStain (Invitrogen) was used as a molecular weight (MW) marker. C, the predicted structure of hTPH2 and its deletion mutation forms. . A, SDS-PAGE gels stained with Coomassie depicting the resulting products from nickel-column purification of hTPH2 and its deletion mutations, N⌬150 and N⌬150/C⌬24. A total of 8 g of purified protein was loaded into each well. The molecular weight marker was See Blue II (Invitrogen). B, Western blot analysis of the resulting products from nickel-column purification of hTPH2, N⌬150, and N⌬150/C⌬24. A total of 0.5 g of purified protein was used per sample. MagicMark XP (Invitrogen) was used as a molecular weight marker. C, wild type hTPH2 was further purified on a Superdex 200 sizing column following nickel chromatography. BenchMark PreStain (Invitrogen) was used as the molecular weight marker.
Although not shown, the C⌬24 truncation mutant was also found to be very unstable with a half-life of less than 30 min.
Analysis of hTPH2 Quaternary Structure-Based on size-exclusion chromatography, both wild type hTPH2 and N⌬150 were determined to be tetramers with a size of 236 and 168 kDa, respectively, reflecting their differential subunit size (hTPH2 ϭ 59 kDa; N⌬150 ϭ 42 kDa). On the other hand, N⌬150/C⌬24 was determined to be monomer with a size of 39 kDa (the size of the monomeric protein; Fig. 6).
Kinetic Analysis-Given the caveat that the three recombinant enzymes have very different stabilities, we performed enzyme kinetic analysis to determine Michaelis-Menten constants (Table 1). To summarize, removal of the NH 2terminal regulatory domain (N⌬150) produced no significant change in either the K m for BH 4 or tryptophan. Ironically, deletion of both the NH 2 -terminal regulatory and COOH-terminal tetramerization domains (N⌬150/C⌬24) produced opposite effects in K m values. That is, the K m,BH4 was increased 2-fold, while the K m,Trp was decreased almost 3-fold. For both deletion constructs, there were apparent increases in V max values that we believe reflect their increased stability and differences in purity. Finally, attempts to extend these findings to the C⌬24 deletion of just the COOH-terminal tetramerization domain produced an insoluble enzyme of such limited stability that it failed to provide reliable kinetic values (data not shown).

DISCUSSION
In 2003, the central nervous system tryptophan hydroxylase isoform, TPH2, was discovered (10). Up until this point, we (and others in the field) had been studying the peripheral form of the enzyme traced back to the original cloning of pineal TPH in 1987 by Woo and colleagues (33). Interestingly, during this period, there were hints that the brain TPH behaved differently and that there were difficulties reconciling apparent discrepancies in the regulation (reviewed in Ref. 11). The discovery of TPH2 opens new vistas of central nervous system research and provides tools to better understand human mental health and disease. However, characterization of the structure, activity, and regulation of hTPH2 (human tryptophan hydroxylase 2; central nervous system isoform) is a prerequisite. The present report describes the expression of human TPH2 in bacteria and its characterization.
As has been observed for TPH1 (16 -18), there are several problems that arise in working with recombinant hTPH2. First, while it is expressed at very high levels in bacteria (Fig.  2), it is not well tolerated and is "packaged" into insoluble inclusion bodies. This is shown in Fig. 3B as the low level of protein in the soluble fraction for the full-length, wild type hTPH2. While this low level of soluble hTPH2 can then be enriched by nickel column chromatography, it then presents a second problem whereby it co-purifies with other proteins that are expressed in bacteria (Fig. 3B). We have already identified the source of this problem as the NH 2 -terminal regulatory domain. It is noteworthy that these problems of solubility are identical to those described by Haavik and colleagues (27) for recombinant fusion construct of human TPH2.
In our first efforts to characterize the recombinant hTPH2, we have conducted deletion mutagenesis to confirm our prediction of a regulatory/catalytic domain structure that is common to all of the aromatic amino acid hydroxylases (reviewed in Refs. 34 -36). The design of these studies is presented in Fig. 3A. Sequence comparisons suggest that, as with the other hydroxylases, hTPH2 is composed of a COOH-terminal catalytic core and an NH 2 -terminal domain that presumably functions as a regulatory domain. Interestingly, all four of the hydroxylases (phenylalanine hydroxylase, tyrosine hydroxylase, TPH1, and TPH2) share a conserved pentapeptide (Val-Pro-Trp-Phe-Pro) that marks the beginning of the catalytic domain, which in the case of hTPH2, begins at residue 151. In addition, McKinney et al. (27) report that hTPH2 is phosphorylated at Ser-19 within this NH 2 -terminal domain. Deletion of the regulatory domain (N⌬150) has a dramatic impact on the enzyme. First, it improves expression of soluble protein (Fig. 3B). This then improves the purification of the protein (Fig. 4). In the latter case, we think this is simply a consequence of the fact that the recombinant truncation mutant represents a high percentage of the total protein and so more effectively competes for binding to the affinity column. This conclusion is reached because a second purification step can produce nearly homogenous enzyme (Fig. 4C), suggesting that there is no strong interaction between wild type hTPH2 and bacterial proteins.
A common problem with TPH1 is a marked instability (reviewed in Ref. 35). In particular, it has been known for some time that the enzyme is difficult to purify, in part, because it loses enzymatic activity during isolation. This is not the result of a proteolytic degradation; rather, the enzyme appears to undergo a loss of function related to denaturation or oxidative damage (the AAAHs all generate reactive oxygen as part of their reaction mechanisms). The same is true of hTPH2 as illustrated in Figs. 4 and 5. First, while we can purify hTPH2 to homogeneity, the enzyme is generally devoid of activity. Previous studies on rabbit TPH1 (21) suggested that the difference between TH (stable) and FIGURE 5. The stability of hTPH2 and deletion mutations was measured during a 3-h incubation at 37°C. The activity for each enzyme was measured every 15 min. Half-lives were found to be 30, 203, and 147 min for hTPH2, N⌬150, and N⌬150/C⌬24, respectively. The wild type hTPH2 activity (solid triangles) rapidly decreased. In contrast, the activity of N⌬150 (open circles) and N⌬150/C⌬24 (solid circles) remained stable. TPH1 (unstable) resides in their removal of the NH 2 -terminal regulatory domains. In the present context, we have established that removal of the regulatory domain stabilizes the protein by a factor of six. However, more detailed fine mapping strategies will be required to identify the specific residues responsible for this destabilization phenomenon.
All of the hydroxylases exist as tetramers (34,35). Significant work has been conducted to map the interaction domains. The existence of a COOH-terminal coiled-coil (putative leucine zipper) was first postulated in 1991 (37). The existence of this interaction motif has subsequently been confirmed by a number of laboratories for all of the hydroxylases (38 -41) including observations in x-ray crystallography studies of a 41 antiparallel coiled coil at the extreme carboxyl terminus of TH (30). The same observations hold true for hTPH2. That is, removal of the carboxyl terminus converts the enzyme from a tetramer to a monomer (Fig. 6). This is in contrast to TPH1 (18) where motifs within the amino-terminal regulatory domain contribute to some subunit interactions. Unfortunately, as noted above, the C⌬24 mutant (that does not contain an amino-terminal deletion) fails to provide reliable, stable, and soluble enzyme for analysis.
Using the modeling programs Accelrys DS Modeling 1.1 and Accelrys DS ViewerPro 5.0 (Accelrys Software Inc.), SwissPdb Viewer (42), and LOOPP (43)(44)(45), we have predicted the structure of hTPH2 and its amino and amino/carboxylterminal deletions (N⌬150 and N⌬150/C⌬24) (Fig. 3C). The term "presumably" is advisable at this point, because deletion of the NH 2terminal domain produces only modest increases in activity and alterations in kinetic constants ( Table 1). The kinetic constants determined for the recombinant hTPH2 enzymes are very comparable for reported values for multiple species of TPH1 (16,46) and TPH2 FIGURE 6. A, purified N⌬150 and N⌬150/C⌬24 chromatography using an analytical Superdex 200 10/300 GL column at room temperature. The N⌬150 peak was detected at 168 kDa and migrated as a tetramer, whereas the N⌬150/C⌬24 peak was detected at 39 kDa as a monomer. Standards (dashed line) were apoferritin (443 kDa), ␤-amylase (200 kDa), alcohol dehydrogenase (150 kDa), bovine serum albumin (66 kDa), and carbonic anhydrase (29 kDa). B, purified wild type hTPH2 size-exclusion chromatography using an analytical Superdex 200 10/300 GL column at 4°C. The main peak was seen at 236 kDa. The identity of this peak as hTPH2 (wild type) was confirmed by Western blot analysis and enzyme activity determination on isolated fractions (data not shown). Wild type hTPH2 migrated as a tetramer. The same standard (dotted line) proteins were used.  (27,28). Most importantly, from a practical standpoint, deletion of the NH 2 -terminal domain dramatically increases the solubility of the enzyme in bacteria (see Fig. 3B) and provides an opportunity to generate sufficient amounts of material for future crystallography studies.