Phenylalanine hydroxylase (PheH), tyrosine hydroxylase (TyrH), and tryptophan hydroxylase (TrpH) make up the family of tetrahydropterin dependent aromatic amino acid hydroxylases. Scheme 1 illustrates the general reaction these enzymes catalyze, the hydroxylation of the side chain of an aromatic amino acid utilizing a tetrahydropterin as the source of the two electrons required to reduce the other atom of oxygen to the level of water. The eukaryotic forms of these enzymes have been studied the most due to their physiological importance. PheH is a liver enzyme that catalyzes the catabolism of excess phenylalanine in the diet to tyrosine; the vast majority of cases of phenylketonuria are due to deficiencies in this enzyme (1). TyrH is found in the central nervous system and adrenal gland, where its role is to catalyze the first step in the biosynthesis of catecholamines, the formation of dihydroxyphenylalanine (DOPA) from tyrosine. TrpH is present in the brain where it catalyzes the first step in the biosynthesis of serotonin, the hydroxylation of tryptophan to 5-hydroxytryptophan. PheH is also present in bacteria, although not in Escherichia coli; as of this writing, 17 different bacterial genomes include a PheH ortholog. Only the enzyme from Chromabacterium violaceum has been characterized to any extent.

While our knowledge of the structures of these proteins is most advanced in the case of PheH, details of the catalytic mechanism have been worked out with all three. The homology of the catalytic domains of the pterin dependent hydroxylases suggests that they share a common catalytic mechanism. For the present discussion, it will be assumed that this is indeed the case, so that experimental results for all three enzymes can be used to develop a common mechanistic scheme. Scheme 2 summarizes the present understanding of the chemical mechanism of these enzymes. The reaction can be divided into two partial reactions, formation of the hydroxylating intermediate and oxygen transfer to the amino acid substrate. In neither has an intermediate been trapped for direct structural analysis. However, a variety of kinetic and computational approaches is providing insights into the catalytic details.

A 4a-hydroxypterin (Scheme 1) is a product of the reaction catalyzed by all three enzymes (55-58). Molecular oxygen is the source of the oxygen atom incorporated into the pterin product (59) as well as the oxygen atom incorporated into the amino acid (60). This requires that there be a reaction between an oxygen species and the pterin during catalysis, so that the pterin must participate in catalysis to a greater extent than simply supplying electrons. Studies of the autoxidation of tetrahydropterin provide a possible model for the enzymatic reaction between oxygen and tetrahydropterin. The products of the autoxidation reaction are H₂O₂ and quinonoid dihydropterin (61, 62). Scheme 3 summarizes the mechanism proposed by Eberlein et al. (63) for this reaction. The first and rate-limiting step is single electron transfer from the pterin to oxygen to form superoxide and a pterin cation radical. Radical collapse generates a peroxypterin; loss of peroxide from the peroxypterin leaves the dihydropterin. The proposed radical and peroxypterin intermediates have not been directly detected during the autoxidation reaction. The intermediacy of a peroxypterin is supported by the observation that a stable peroxypterin can be formed by reacting a 5-deazatetrahydropterin with singlet oxygen (64), by the production of H₂O₂ as an autoxidation product, and by the direct observation of a similar 4a-peroxyflavin in the reaction of reduced flavoprotein hydroxylases with oxygen (65). Eberlein et al. (63) proposed the intermediacy of a pterin cation radical based upon analyses of the thermodynamics of the autoxidation reaction and the pH insensitivity of the reaction. There is no solvent isotope effect on the autoxidation reaction (66), consistent with rate-limiting single electron transfer to form the pterin cation radical.

One critical difference between the enzyme-catalyzed reaction and the autoxidation reaction is that the enzyme has a metal in the active site; this raises the possibility that the iron is explicitly involved in the reaction between oxygen and tetrahydropterin in the active site. A mechanism consistent with the data is shown in Scheme 4. Binding of oxygen to the Fe(II) atom would generate a complex equivalent to Fe(III)O₂⁻. This could then attack the C(4a) position of the tetrahydropterin to form the peroxypterin. Such a mechanism would avoid the spin-forbidden direct reaction of triplet oxygen with the tetrahydropterin. The oxygen isotope effects for TyrH are too small to be due only to binding of oxygen to Fe(II) (71). They can be rationalized by equilibrium binding of oxygen to the iron followed by an irreversible step involving either electron transfer from the pterin to the oxygen to form Fe(II)O₂⁻ or by nucleophilic attack of the pterin on the iron-bound oxygen to form the Fe(II) μ-peroxypterin species shown in Scheme 2. In the absence of direct detection of the intermediates proposed for the two different mechanisms, one criterion for clarifying the role of the iron in the initial reaction between tetrahydropterin is to determine whether these enzymes can catalyze a reaction between tetrahydropterin and oxygen in the absence of iron. The metal-free PheH from C. violaceum has been reported to catalyze the oxidation of dimethyltetrahydropterin to quinonoid dihydropterin and hydrogen peroxide at about 5% the turnover of the iron-containing enzyme (23), suggesting that the metal is not absolutely required. In contrast, several mutant forms of TyrH in which the metal affinity has been decreased by mutation of the one of the metal ligands still show an absolute requirement for metal to catalyze any tetrahydropterin consumption (35). The metal requirement of the eukaryotic enzyme may reflect a need for bound metal to effect the conformational change that occurs when both amino acid and pterin are bound before any reaction between pterin and oxygen can occur rather than direct involvement of the metal in the reaction. Alternatively, the reaction of the bacterial enzyme may be due to a low level of catalysis of tetrahydropterin oxidation by enzyme containing a metal other than iron. When His336 in TyrH is mutated to glutamine, the enzyme retains substantial hydroxylase activity with an absolute need for iron. However, in the presence of Co(II), this mutant protein will catalyze tyrosine dependent tetrahydropterin oxidation at 14% the rate seen with Fe(II), although it does not catalyze DOPA formation (72). The iron-free C. violaceum PheH was not metal-free but contained substoichiometric amounts of nickel and copper (23).

Studies of the mechanism of oxygen addition to the amino acid substrate are consistent with an electrophilic hydroxylating intermediate such as Fe(IV)O. Probing the mechanism of oxygen addition to the amino acid substrate has been complicated by the fact that other first-order steps in the reaction are slower for both PheH and TyrH (46, 69), so that changes in steady state kinetic parameters with alternate or isotopically labeled substrates do not directly reflect changes in the rate constant for this chemical step. In the case of TyrH, this complication has been circumvented by analyzing product partitioning rather than steady-state kinetics. The k_cat value for this enzyme is relatively constant for a number of ring-substituted phenylalanines (69). However, changes in the electron-donating ability of the substituent at the 4-position of the aromatic ring alter the efficiency of the reaction (85). This is consistent with a kinetic model (Scheme 5) in which the rate of formation of the Fe(IV)O intermediate is insensitive to the reactivity of the amino acid substrate, but its subsequent fate is determined by partitioning between productive hydroxylation and unproductive breakdown. When the relative amount of hydroxylated amino acid formed was determined for a series of 4-substituted phenylalanines, there was a good correlation between the relative rate constants for hydroxylation and the σ values of the substituents, with an average ρ value of −5 for tetrahydrobiopterin and 6-methyltetrahydropterin (85). This result was interpreted as evidence for a cationic transition state for oxygen addition to the aromatic ring. An electrophilic aromatic substitution reaction for hydroxylation involving the cationic species shown in Scheme 2 was proposed based on these data. The proposed cationic intermediate is consistent with the observation that these enzymes exhibit NIH shifts, 1,2 shifts of the substituent at the site of hydroxylation to the adjacent ring carbon (86, 87). The extent of the shift decreases with the electronegativity of the substituent, consistent with a requirement for the substituent to participate in the three-centered transition state for the 1,2 shift (85).

A Fe(IV)O intermediate for the pterin dependent enzymes would be expected to have reactivity comparable to the heme-based cytochrome P450 systems. Other enzymes for which such an intermediate has been invoked (e.g., iso-penicillin synthase and the α-ketoglutarate dependent hydroxylases) are capable of aliphatic hydroxylation (77). Consistent with such an expectation, all three hydroxylases have been shown to catalyze benzylic hydroxylation of methylated aromatic amino acids (54, 85, 88). In addition, PheH has been reported to catalyze aliphatic hydroxylation (21, 46) and epoxidation (23). Only in the case of benzylic hydroxylation by TyrH has the mechanism been investigated in detail. Analysis of the products when 4-methylphenylalanine containing one, two, or three deuterium atoms in the methyl group was used as a substrate allowed determination of the intrinsic primary and secondary deuterium isotope effects for the hydroxylation step (89). The secondary isotope effect of 1.2 demonstrated that there is significant rehybridization of the methyl carbon in the transition state for hydroxylation; such a result is most consistent with the formation of a benzylic radical intermediate by abstraction of a hydrogen atom from the methyl group by the electrophilic Fe(IV)O species, followed by radical rebound of a hydroxyl radical to form the hydroxylated product (Scheme 6). The primary isotope effect for this reaction was 9.6; together with the large secondary effect, this value suggests that there is a significant contribution of quantum mechanical tunneling to the reaction. This mechanism is similar to that proposed for aliphatic hydroxylation by the cytochrome P450 family (90), demonstrating the comparable reactivity of the hydroxylating intermediates of the non-heme pterin dependent enzymes and the heme dependent hydroxylases.

(A) Overlay of the catalytic domains of human PheH (PDB file 1J8U), rat TyrH (PDB file 2TOH), human TrpH (PDB file 1MLW), and C. violaceum PheH (PDB file 1LTV) (B) Combined catalytic and regulatory domains of rat PheH. The structure is based on the PDB file 1PHZ; the protein used for structure determination lacked the C-terminal 24 residues, and the N-terminal 18 residues are not seen in the structure. The orientation is the same as in panel A. (C) Model for the intact rat PheH tetramer. The model was constructed by superimposing the catalytic and regulatory domains of rat PheH shown in panel B with the structure of the tetrameric TyrH catalytic domain (PDB file 2TOH).

Comparison of the iron and tetrahydrobiopterin binding sites of PheH in the absence (A) and presence (B) of β-thienylalanine. The structures are from the PDB files 1J8U and 1KWO.

Amino acid substrate binding site of PheH. The structure is from the PDB file 1KWO.

Movement of the 130–150 loop in PheH upon binding of β-thienylalanine to the tetrahydrobiopterin bound enzyme. The loop with Tyr138 in blue is from the binary complex (PDB file 1J8U); the positions of the active site iron, the substrates, and the metal ligands are from the ternary complex (PDB file 1KWO).