The Reaction Mechanism of Phenylalanine Hydroxylase. - a Question of Coordination

Teigen Κ. et al.: The Reaction Mechanism of Phenylalanine Hydroxylase 27

Pteridines Vol. 16, 2005, pp. 27 - 34

The Reaction Mechanism of Phenylalanine Hydroxylase. - A Question of Coordination.

Knut Teigen1, Vidar R. Jensen2, Aurora Martinez1

'Department of Biomedicine, University of Bergen, Jonas Lies vei 91, 5009-Bergen, Norway, ^Department of Chemistry, University of Bergen, Allégaten 41, 5007-Bergen, Norway

Abstract

Phenylalanine hydroxylase (PAH) is a non-heme iron and tetrahydrobiopterin-dependent enzyme that catalyzes the hydroxylation of L-phenylalanine to L-tyrosine using dioxygen as additional substrate. The cofactor tetrahydrobiopterin accepts one of the oxygen atoms of dioxygen during catalysis and also seems to be involved in pre- reduction of the active site iron from the ferric to the activated ferrous form. Structures of the truncated form of PAH in complex with substrate and cofactor are available, but the oxygen binding site and the actual mechanism of electron transfer are uncertain. It is believed that dioxygen binds directly to the metal, where it is activated, and several reaction mechanisms involving end-on binding of 02 have been proposed based on both experimental studies and quantum mechanical calculations. However, in this work we aimed to investigate the possibility of side-on binding of dioxygen to the iron. Furthermore, NMR and recent high-resolution crystallographic studies also place the cofactor in closer proximity to the iron, challenging the mechanistic conclusions from earlier crystallographic and computational studies. In this paper we report preliminary results from a density functional theory (DFT) study of the coordination of dioxygen to a structural model of PAH based on a recent crystallographic structure. These results are compared with existing computational and experimental data and their implications for the mechanism of the PAH-reaction are discussed. Particular attention is paid to the binding-mode of dioxygen and the iron-cofactor distance.

Key words: Phenylalanine hydroxylase. Reaction mechanism, 5,6,7,8-tetrahydrobiopterin, Density functional theory

Abbreviations: AAH, aromatic amino acid hydroxylases; BH2. L-erythro-7,8-dihydrobiopterin; BH4, (6R)-L-erythro-5,6,7.8-tetrahydrobiopterin; DFT, density functional theory; DZV, double zeta valence; EFP, effective core potentials: L-Phe, L-phenylalanine; MD, molecular dynamics; MM, molecular mechanics; NLE, norleucine; PAH, phenylalanine hydroxylase; QM, quantum mechanics; RMS, root mean square: SCF, self consistent field; SCI-PCM, self-consistent isodensity polarized continuum model; SP, single-point; TH, tyrosine hydroxylase; THA, thienylalanine; TS, transition state; TZDP, triple zeta double polarization

Introduction prior to the disruption of the dioxygen bond. Initial crystallographic studies showed that in the resting The reaction catalyzed by phenylalanine hydroxy- enzyme the oxidized iron is coordinated by two his- lase (PAH; phenylalanine 4-monooxygenase) (EC tidines (His285 and His290) and one glutamic acid 1.14.16.1) involves the incorporation of one of the (Glu330) in addition to three water molecules, forming oxygen atoms from dioxygen into the substrate L-Phe, a six-coordinate octahedral iron environment (3). converting it to L-Tyr, and the other into the cofactor Subsequently, the binding site for both the cofactor and BH4 (1, 2). The active site iron in the ferrous state is the amino acid substrate in PAH was first solved by a believed to be responsible for the activation of oxygen combination of NMR and molecular modelling (4) and

Correspondence to: Knut Teigen. Department of Biomedicine, University of Bergen, Jonas Lies vei 91, 5009-Bergen, Norway, Tel.: 1(47)55586328: Fax: +(47)55586360; e-mail: knul.teigen(S'biomed.uib.no

Pteridines/Vol. 16/No. 1 28 Teigen Κ. et al.: The Reaction Mechanism of Phenylalanine Hydroxylase

later by X-ray crystallography (5-8). cules. L-Phe was found to bind in the second coordi- Many experiments have been carried out in order to nation sphere of the iron (4). The crystal structure of delineate the catalytic mechanism of the aromatic the ferrous binary complex of the enzyme with BH4 amino acid hydroxylases (AAH), but little experimen- showed, however, that in this state the iron was six- tal evidence has been presented about the exact nature coordinate with three bound water molecules (6). of the reaction. Despite this, several mechanisms of Later, the complex with BH4 and a substrate analogue hydroxylation have been proposed for the AAH (1,2, was reported to have only one water molecule bound to 7, 9-13) although the actual catalytic mechanism, the ferrous iron and Glu330 coordinated in a bidentate including the chemical nature of the hydroxylating manner, giving rise to a fivc-coordinatc metal atom intermediate, is still not clear. It has been proven for (7). However, in a repeated study with higher resolu- the catalytic cycle in mammalian PAH that the iron tion, the ferrous iron appeared to have no water mole- atom, which is in the ferric form in the isolated cules coordinated, rendering the iron in a four-coordi- enzyme, is prereduced to its ferrous form by the pterin nate configuration with the partially negatively cofactor prior to catalysis (1, 14). For tyrosine hydrox- charged keto-oxygen atom of the BH4 cofactor at a dis- ylase (EC 1.14.16.2, TH) it has also been shown that tance of 3.1 A from the iron (8). Thus, these recent the steady-state kinetic mechanism is sequential with crystallographic investigations indicate fewer, if any, the tetrahydropterin cofactor binding first, followed by coordinated water molecules, as well as a tighter asso- molecular oxygen and then the substrate (15). The ciation between iron and cofactor. Nevertheless, the steady-state kinetic mechanism of PAH seems to be coordination number of the iron prior to catalysis is not sequential as well, with some degree of randomness in completely clarified and in the present study we have the order of substrate addition ( 1, 14). For PAH from performed quantum chemical investigations based on the bacterium C. violaceum it has recently been report- DFT to investigate this issue. Our results are discussed ed that the mechanism is fully ordered, with the cofac- and compared to existing computational and experi- tor binding the active site first, L-Phe second and mental data. The implications of a different dioxygen dioxygen last ( 16). It seems clear for both enzymes binding mode and a tighter association of the cofactor that no product or intermediate is released prior to the on the reaction mechanism are discussed. binding of all the three substrates ( 1 ). Formation of a quaternaiy PAH-Fe(ll)BH4L-Phe02 complex also trims the iron atom for catalysis ( 1 7). The first observ- Methods able product of the pterin cofactor in the TH and PAH catalyzed reactions is a 4a-hydroxy-tetrahydropterin in The structure of the complexes of PAH with the which the oxygen atom in position 4a is derived from substrate and pterin cofactor was analyzed using the molecular oxygen (1, 10, 18). The other half of the Insightll software package and WebLabViewer (both oxygen molecule is found in the hydroxylated product from Accelrys Inc.). The quantum chemical geometry (L-Tyr). It has been postulated that a Felx=0 complex optimizations w ere performed using the three-parame- is an intermediate in the reaction mechanism (12). The ter hybrid density functional method of Becke (termed existence of FeIN-0 has been verified in model com- "B3LYP") (22) as implemented in the Gaussian 03 set pounds by a combination of Mössbauer and X-ray of programs (23). The unrestricted formulation was absorption spectroscopy ( 19, 20). Recent density func- used, i.e., "UB3LYP". Stationary points were opti- tional theory (DFT) studies confirmed that a reaction mized and characterized using algorithms involving mechanism for PAH based on an Fclv=0 intermediate analytic calculation of the first and second derivatives could be possible (11,21). In this mechanism, binding of the energy. Numerical integrations were performed of both substrate and cofactor triggers removal of one using the default "line" grid of Gaussian 03, consisting of the water molecules coordinated to the iron atom, of 75 radial shells and 302 angular points per shell, and and the water molecule is then replaced by dioxygen, the Gaussian 03 default values were chosen for the self thus forming an iron-peroxo-pterin intermediate. Upon consistent field (SCF) and geometry optimization con- heterolytic O-O bond cleavage, 4a-OH-BH4 is generat- vergence criteria. Thermochemical values were com- ed together with an 0H-Fe,v=0 intermediate, which in puted within the harmonic-oscillator, rigid-rotor, and turn is believed to be responsible for the hydroxylation ideal-gas approximations. The B3LYP-optimized of the aromatic amino acid. geometries were subjected to single-point (SP) energy and properties calculations. The SP calculations were Our group has investigated the binding of BH4 and performed using the "fine" grid described above and L-Phe to PAH by NMR (4). It was postulated that the the SCF procedure was converged to a RMS change of cofactor directly coordinates the active site iron, most the density matrix below 1.0-10 \ The basis sets that probably replacing one of the coordinated water mole- were used for the geometry optimizations (termed

Pteridines/Vol. 16/Νυ. 1 Teigen Κ. et αι.: The Reaction Mechanism of Phenylalanine Hydroxylase 29

Table 1. Cofactor and substrate distances to the active site iron observed in different PAH complexes. Cofactor" and substrate'1' distances to the active site iron. The number of iron-coordinated water molecules"' are listed. Results obtained from NMR are commensurate with a model where one water molecule'1' is coordinated to the iron when both substrate and cofactor are present. Observed distance ranges0 from the MD simulations in (34) and unpublished results" from a 500 ps MD simulation of L-Phe in complex with hPAH are given. The distances observed in the average structure from the simulations are also tabulated51. MD simulations were performed without solvent present, but also with two water molecules restricted to be coordinated to the iron1" (34). cvPAH, C.vio/aceum PAH; hPAH, human PAH; rPAH, rat PAH.

Distance from iron [Ä] Coord. pd b- Complex C4aa) 04a) C«bl waterc' Method code Resolution Reference 3 cvPAH-Fe *-BH2 6.38 4.18 - 2 X-ray 1 LTZ 1.4 A (33) 3+ hPAH-Fe -BH2 5.96 3.67 - 3 X-ray 1LRM 2.1 Λ (7) 3 hPAH-Fe "-BH2 6.06 3.82 - 3 X-ray 1DMW 2.0 A (5)

2 hPAH-Fe '-BH4 5.92 3.81 -, 3 X-ray 1J8U 1.5 A (6) 2+ hPAH-Fe -BH4-THA 4.45 3.37 7.43 1 X-ray 1KWO 2.5 A (7) 2+ hPAH-Fe -BH4-THA 4.42 3.12 7.30 0 X-ray 1MMK 2.0 A (8) 2+ hPAH-Fe -BH4-NLE 4.60 3.53 7.48 1 X-ray 1MMT 2.0 A (8)

3+ dl hPAH-Fe -BH2-Phe 4.29 2.46 7.00 1 NMR _ (4) 4.16-6.34 e> 2.04-4.45e) 6.35-8.23fl 2t hl hPAH-Fe -BH4 0-2 MD * - (34) ~5.119) -3.029' ~7.17'91

3+ rPAH-Fe - 3 X-ray 2PHM 2.6 A (35)

3 rPAH-Fe *-Seri6-P04 - - * 1 X-ray 1 PHZ 2.2 A (35)

DZV) are described as follows: Hay and Wadt effec- Results and Discussion tive core potentials (ECP) (24) were used for iron, replacing the Is, 2s, and 2p electrons. The 3s, 3p. 3d. The BH4 Binding Site 4s, and 4p orbitals of Fe were described by the Hay and The cofactor binding pocket in PAH is located at the

Wadt primitive basis set (24) contracted to [3s.3p,2d]. bottom of the hydrophobic active site opening. BH4 Oxygen, nitrogen, carbon and hydrogen atoms were (Fig. 1 ) makes stacking interactions with an invariant described by standard Dunning and Hay valence double-zeta basis sets (25). The SP energy and properties calculations involved basis sets (termed TZDP) that were improved compared to those used during the geometry optimizations: For iron, the Hay and Wadt primitive basis was contracted to [4s,4p,3d]. For carbon, nitrogen and oxygen atoms, the standard valence double-zeta basis set was extended with a diffuse ρ primitive and a single d polarization primitive (26) and ocl Ν Ν contracted to [3s.3p,ld], whereas the hydrogen basis set was extended with a diffuse s primitive and a single ρ polarization primitive and contracted to [3s. lp] Figure 1. Structure of BH4 with atom labelling (26). Dielectric effects from the surrounding protein (hydrogens not shown). were included in these SP calculations using the Self- Consistent Isodensity Polarized Continuum Model phenylalanine (Phe254) and the N3 and the amino (SCI-PCM) (27). The solute cavity was defined by the group at C2 hydrogen bonds with the carboxylic group 0.00001 au isodensity surface. of a conserved glutamate residue (Glu286). In the crys-

Pteridines/Vol. 16/No. 1 30 Teigen Κ. eí al.: The Reaction Mechanism of Phenylalanine Hydroxylase

tal structure of the binary PAHBH4 complex these Density Functional Theory Calculations hydrogen bonds are water-mediated (6). The subse- In order to investigate possible mechanistic impli- quent crystal structure of the ternary complex of PAH cations from the iron coordination environment with with the substrate analogue thienylalanine (THA) and few or no water ligands present as deduced from NMR BH4 still shows one water molecule bridging the inter- studies (4) as well as from the most recent model from action between the cofactor and the iron (7). However, crystallography (8), wc have embarked on quantum in a repeated study with higher resolution, the coordi- chemical investigations. We have followed a model- nating water molcculc was omitted in the final model ling-strategy close to that recommended by Siegbahn building (8) and the cofactor is actually placed in the for quantum chemical gas-phase calculations on the first, not the second, coordination sphere of the iron catalytic centers of metalloenzymes in general (3 1 ) and atom (Table 1.). This latest crystal structure and the also as applied in particular in the study of the reaction corresponding distance from the cofactor to the iron is catalyzed by PAH performed by the same research more in agreement with our previous model of the ter- group (11,21). Thus, imidazole is used as a model for nary complex obtained by NMR and molecular dock- histidines, whereas glutamates are modeled by for- ing (4) (Table 1.). It has been shown in other iron con- mates. Also, 5,6,7,8-tetrahydropterin is used instead of taining enzymes that exclusion of water molecules the (6R)-erythro-5,6,7,8-tetrahydrobiopterin (BH,) from the active site is essential for efficient enzymatic (Fig. 1), i.e., the (ClI(OII))2CH3-substituent in posi- function. For example, upon substrate binding, water tion C6 is replaced by a hydrogen atom for reasons of has to be displaced from the active site of cytochrome computational efficiency. Details of the DFT calcula- P450 to prevent reaction uncoupling that could result tions can be found in the Methods section. in production of hydrogen peroxide (28). We have started by investigating the interaction betw een the iron site and dioxygen and the pterin The L-Phe Binding Site cofactor. In the crystal structures, iron is coordinated The crystal structures of PAH in complex with a by His285 and His290, and one bidentate carboxylate substrate analogue, i.e. THA (7. 8) and norleucine (Glu330), here modeled by two imidazole ligands and (NLE) (8), show binding at the same distance and posi- one formate ion (Fig. 2). The more distant Glu286 is tion from iron as reported by NMR (29) (see Table 1 also included as a formate ion, resulting in an overall for Ca-iron distances). His285 and Trp326 are neutral model. No water molecules were included (see involved in stacking interactions with the aromatic ring of the L-Phe substrate, while Arg270, Ser349 and Thr278 interact with the charged α-groups of the substrate. Because of the slightly different orientation of Ν •· the α-amino group in L-Phe and THA, it has been argued that they interact somewhat differently with the protein (7). However, these differences are within the ' e"' - Κ1 experimental error of both techniques, and the binding O ' mode found by the two techniques is identical for all - •ν :. ι o i practical purposes. 1 92 Λ / 2 .05 A Ο , h" The Oxygen Rinding Sire r Attempts to elucidate the oxygen binding mode have so far been unsuccessful. However, crystallography studies of PAH in complex w ith substrate analogues (THA and NLE) and BH4 together with CO have been reported (8). Some undefined electron density that could be derived from CO was observed close to the iron atom and, according to the authors of this work, this density could be in agreement with traces of side-on bound CO (8). However, this electron density was omitted in the final model building. Side-on binding of molecular oxygen to iron has been observed in Figure 2. Top view of the septet ground state of the naphthalene dioxygenase (30) and it seems plausible iron-dioxygen complex as optimized using DFT. The that dioxygen could interact with the iron atom in a dioxygen molecule is seen in front of the iron atom. similar manner in PAH. The O-O distance of dioxygen is 1.35 A in the complex.

Pteridines/Vol. 16/Νο. I Teigen Κ. et al.: The Reaction Mechanism of Phenylalanine Hydroxylase 31

above). The geometry optimizations of the dioxygen tions involved a pterin-peroxo intermediate where the complexes immediately reveal that the iron atom has dioxygen bound end-on to the active site iron atom, considerable affinity for the pterin cofactor. In the bridging the interaction with the pterin cofactor. A het- septet ground state (S = 3. M = 7), the geometry opti- erolytic cleavage of the dioxygen bond then gave rise mization reduces the Fe-04 distance from 3.12 A (the to a hydroxylated cofactor and a Fe(IV)-oxo interme- distance in the crystal structure) to 1.92 A (see Fig. 2). diate that subsequently drove the incorporation of oxy- The resulting complex contains a relatively strong gen in the amino acid substrate. On the other hand, a Fe(II)-pterin dative bond, whereas the formate iigand side-on binding of dioxygen to the iron as assumed in now bonds in a monodentate fashion. The original this work would imply that the cofactor would have to four-coordinate iron complex is thus still four-coordi- be closer to the iron to be an acceptor for one of the nate, not counting dioxygen. The complexation of the oxygen atoms. A closer association between the metal latter molecule should be termed side-on rather than atom and the cofactor is already suggested by NMR end-on, although the Fe-O distances (2.28 and 2.51 A) and molecular modelling and by the most recent crys- reveal a slightly asymmetric bonding mode. tallographic studies (Table 1). and the pterin might in fact be directly coordinated during catalysis, as depict- Implications for the Reaction Mechanism ed in Fig. 3. Cleavage of the dioxygen molecule could The structural model previously employed to inves- then give rise to simultaneous hydroxylated L-Phe and tigate the reaction mechanism by Bassan et al. ( 11, 2 1 ) BH4, in agreement with the finding that both substrates was based on the crystal structure solved for the bina- are being released at the same time ( 1 ). There are some ry complex of PAH with L-erythro-7,8-dihydro- very important differences observed in our DFT results biopterin (BH2) (5). The quantum chemical calcula- as compared to the study by Bassan et al. (21).

Figure 3. Schematic representation of a hypothetical hydroxylating transition state (TS) based in a binding of the cofactor in the first coordination sphere of the iron. Dioxygen interacts side-on w ith the iron, generating an iron- peroxide-cofactor-substrate TS where cleavage of the dioxygen bond could give rise to simultaneously hydroxylation of cofactor and substrate.

Pteridines/Vol. 16/No. 1 32 Teigen Κ. et cil.: The Reaction Mechanism of Phenylalanine Hydroxylase

Although only one Fe-O distance is reported in the approach and be a plausible acceptor of the oxygen study by these authors, their septet dioxygen complex atom, thus explain the decoupling of the reaction. It appears to be significantly more asymmetric (thus end- should be emphasized that even if the reaction mecha- on) as judged from the figures given in the paper. On nism presented here differs from that postulated by the other hand, the presently obtained electronic struc- Bassan et al. and others, both mechanism do not need ture of the iron complex agrees well with that reported to be mutually exclusive. It has been shown for other by Bassan et al. The formal electronic configuration is systems that a single reaction might follow several dif- best described as a hybrid between the forms Fe- 02 ferent pathways. For instance the proton transfer in and Fe5 O2. The spin on iron is 4.04 and the total spin bacteriorhodopsin has been shown to be able to on dioxygen is 1.53. Our results in this context, in par- progress through several different pathways with very ticular regarding the initial iron-pterin bond formation, similar energy barriers (32). It could be possible that appear to be in stark contrast to those of Bassan et al., PAH is able to hydro.xylate L-Phe by more than one and it must be stressed that this discrepancy is mainly mechanistic pathway. a result of the different starting points, i.e. different X- It is important to note that the current calculations ray crystal structures have been used when building the do not take into account the protein environment models. The present calculations are based on a struc- (except for polarization effects, see the Methods sec- ture without the presence of water ligands and the tion) and hydrogen bonds connecting our iron site coordination of the pterin at the expense of the second model to the macrotnolecule. The presently obtained oxygen of the formate ion suggests that the iron has a tight iron-pterin complex indicates that hydrogen surprisingly high affinity for the pterin keto-function. bonds between the pterin and nearby residues as well Both the tight iron-cofactor association and the side-on as significant rigidity in the latter would be needed in binding of dioxygen suggest that oxygenation ot the order to avoid coordination of pterin to the iron atom. substrate and the pterin might, in fact, take place in Formation of an iron-pterin complex such as found in concert. The transition state (TS) for such a reaction the present study could be relevant to the mechanism would involve the simultaneous binding of L-Phe. of the AAH and more detailed calculations of the iron- molecular oxygen and BH4 (a sketch of such a hypo- pterin interaction should therefore be undertaken. thetical TS is given in Fig. 3). Such a TS does not only Nevertheless, accurate calculation of the delicate bal- indicate why all the reagents need to be bound for ance between bidentatc complexation of carboxylate catalysis to occur, but might also offer a molecular and formation of an iron-pterin bond would involve explanation for the fact that some cofactor analogues \ery elaborate QM/MM calculations and must be left give rise to uncoupling of the reaction. It is well known for future investigations. that some cofactor analogues result in an unproductive catalytic reaction, where the relative amount of oxidized cofactor exceeds the amount of hydroxylated substrate ( 1 ). The acceptor for the second oxygen atom Aknowledgements is believed to be water, and hydrogen peroxide is the product of this hydroxylation. It is difficult to explain Part of this work was presented at the 23 d Winter how an altered configuration of the cofactor gives rise Workshop on Clinical, Chemical, and Biochemical to an uncoupled reaction in the mechanism by Bassan Aspects of Pteridines held on February 28"' to March et al. in which the generation of the hydroxylating 6th. 2004 in St. Christoph. Austria. We gratefully intermediate (the Fc-lV-oxo species) is postulated to acknowledge CPU resources granted by the occur prior to hydroxylation of the substrate. Thus, the Norwegian Research Council through the NOTUR hydroxylation of the cofactor and the substrate are supercomputing program. K. Teigen gratefully imagined as separate events. Once generated, the sub- acknowledges support from the Sigurd K. Thoresen sequent fate of the Fe-IV-oxo intermediate should not Foundation and the Norwegian Research Council. be dependent on the cofactor involved in generating it. unless the proper binding of the substrate is dependent on the structure of the cofactor. How e\ er. the hydrox- References ylating intermediate as depicted in Fig. 3 is compatible with the finding that an altered configuration of the 1 Kappock TJ. Caradonna J P. Pterin-Dependent cofactor results in a slightly altered conformation (and Amino Acid Hydroxylases. Chem. Rev. thus to a higher energy) of the pterin-peroxo-iron TS 1996:96:2659-2756. and an altered orientation of the substrate. With such a 2 Fitzpatrick PF. Mechanism of aromatic amino acid concerted TS, it is probable that water could then hydroxylation. Biochemistry 2003:42(48): 14083- 91.

Pteridines/Vol. 16/No. 1 Teigen Κ. et ai.: The Reaction Mechanism of Phenylalanine Hydroxylase 33

3 Erlandsen H, Fusetti F, Martinez Λ, Hough E, tem. Adv. Enzymol. Rclat. Arcas Mol. Biol. Fiatmark T, Stevens RC. Crystal structure of the 1993;67:77-264. catalytic domain of human phenylalanine hydroxy- 15 Fitzpatrick PF. Steady-state kinetic mechanism of lase reveals the structural basis for phenylke- rat tyrosine hydroxylase. Biochemistry tonuria. Nat. Struct. Biol. 1997;4:995-1000. 1991 ;30( 1 5):3658-62. 4 Teigen Κ. Froystein NA. Martínez A. The structur- 16 Volner A, Zoidakis .1, Abu-Omar MM. Order of al basis of the recognition of phenylalanine and substrate binding in bacterial phenylalanine pterin cofactors by phenylalanine hydroxylase: hydroxylase and its mechanistic implication for implications for the catalytic mechanism. J. Mol. pterin-dependent oxygenases. J Biol Inorg Chem Biol. 1999;294(3):807-823. 20()3;8( l-2):121-8. 5 Erlandsen H, Bjorgo E, Fiatmark T, Stevens RC. 1 7 Hagedoorn PL, Schmidt PP, Andersson KK, Hägen Crystal Structure and Site-Specific Mutagenesis of WR. Fiatmark T, Martinez A. The effect of sub- Pterin-Bound Human Phenylalanine Hydroxylase. strate, dihydrobiopterin, and dopamine on the EPR Biochemistry 2000;39(9):2208-221 7. spectroscopic properties and the midpoint potential 6 Andersen OA, Fiatmark T, Hough E. High of the catalytic iron in recombinant human pheny- Resolution Crystal Structures of the Catalytic lalanine hydroxylase. J Biol Chem Domain of Human Phenylalanine Hydroxylase in 2001 ;276(25):22850-6. its Catalytically Active Fe(II) Form and Binary 18 Haavik J, Fiatmark T. Isolation and characteriza- Complex with Tetrahydrobiopterin. J. Mol. Biol. tion of tetrahydropterin oxidation products generat- 2001 ;314(2):266-278. ed in the tyrosine 3-monooxygenase (tyrosine 7 Andersen OA, Fiatmark T. Hough E. Crystal hydroxylase) reaction. Eur. J. Biochcm. Structure of the Ternary Complex of the Catalytic- 1987; 168( 1 ):21 -26. Domain of Human Phenylalanine Hydroxylase 19 Lim ΜΗ. Rohde JU, Stubna A, Bukowski MR, with Tetrahydrobiopterin and 3-(2-Thienyl)-L-ala- Costas M, Ho RY, et al. An FelV=0 complex of a nine, and its Implications for the Mechanism of tetradentate tripodal nonheme ligand. Proc Natl Catalysis and Substrate Activation. J. Mol. Biol. Acad Sci USA 2003; 100(7):3665-70. 2002;320(5): 1095-1108. 20 Rohde JU, In JH. Lim ΜΗ, Brennessel WW, 8 Andersen OA, Stokka A.I, Fiatmark T, Hough E. Bukowski MR, Stubna A, et al. Crystallographic 2.OA resolution crystal structures of the ternary and spectroscopic characterization of a nonheme complexes of human phenylalanine hydroxylase Fe(IV)-0 complex. Science 2003:299(5609): 1037- catalytic domain with tetrahydrobiopterin and 3-(2- 9. thienyO-L-alanine or L-norleucine: substrate speci- 21 Bassan A, Blomberg MR, Siegbahn PE. ficity and molecular motions related to substrate Mechanism of aromatic hydroxylation by an acti- binding. J Mol Biol 2003;333(4):747-57. vated FeIV=0 core in tetrahydrobiopterin-depend- 9 Benkovic S, Wallick D, Bloom L, Gaffney BJ. ent hydroxylases. Chemistry 2003;9( 1 7):4055-67. Domanico P, Dix T, et al. On the mechanism of 22 Becke AD. Density functional thermochemistry. action of phenylalanine hydroxylase. Bioehem Soe III. The role of exact exchange. Journal of Trans 1985; 13(2):436-8. " Chemical Physics 1993;98:5648-5652. 10 Dix TA, Benkovic S.I. Mechanism of oxygen acti- 23 Frisch MJ, Trucks GW, Schlegel HB. Scuseria GE. vation by pteridine-dependent moooxygenases. Robb MA. Cheeseman JR, et al. Gaussian 03. In. Acc. Chem. Res. 1 988;2 1:101 -1 07. Revision B.04 ed. Pittsburg PA: Gaussian, Inc.; 11 Bassan A, Blomberg MR. Siegbahn PF. 2003. Mechanism of dioxygen cleavage in tetrahvdro- 24 Hay PJ. Wadt WR. Ab initio effective core poten- biopterin-dependent amino acid hydroxylases. tials for molecular calculations. Potentials for Κ to Chemistry 2003;9(1 ): 106-15. Au including the outermost core orbitals. Journal 12 Francisco WA, Tian GC. Fitzpatrick PF, Klinman of Chemicaf Physics 1985:82:299-3 10. JP. Oxygen-1 8 kinetic isotope ellect studies ol the 25 Dunning TH. Hay PJ. In: Schaefer HF, editor. tyrosine hydroxylase reaction: Evidence of rate Methods of Electronic Structure Theory. New limiting oxygen activation. .1. Am. Chem. Soc. York: Plenum Press; 1977. p. 1-27. 1998; 120( 1 7):4057-4062. 26 Check CE, Faust TO, Bailey JM, Wright BJ. 13 Solomon El, Decker A, Lehnert N. Non-heme iron Gilbert TM, Sunderlin LS. Addition of Polarization enzymes: contrasts to heme catalysis. Proc. Natl. and Diffuse Functions to the LANL2DZ Basis Set Acad. Sci. 2003; 100(7):3589-94. for P-Block Elements. Journal of Physical 14 Kaufman S. The phenylalanine hydroxylating sys- Chemistry A 2001 ; 1 05:81 1 1.

Pteridincs/Vol. 16/No. 1 34 Teigen Κ. et a!.: The Reaction Mechanism of Phenylalanine Hydroxylase

27 Foresman JB, Keith TA, Wibcrg KB, Snoonian J, 32 Bondar AN, Elstner M, Suhai S, Smith JC, Fischer Frisch MJ. Solvent Effects. 5. Influence of Cavity S. Mechanism of primary proton transfer in bacte- Shape, Truncation of Electrostatics, and Electron riorhodopsin. Structure (Camb) 2004; 12(7): 1281- Correlation on ab Initio Reaction Field 8. Calculations. Journal of Physical Chemistry 33 Erlandsen H, Kim JY, Patch MG, Han A, Volner A, 1996; 100:16098. Abu-Omar MM, et al. Structural Comparison of 28 OpreaTI, Hummer G, Garcia AE. Identification of Bacterial and Human Iron-dependent a functional water channel in cytochrome P450 Phenylalanine Hydroxylases: Similar Fold, enzymes. Pioc. Natl. Acad. Sci. USA Different Stability and Reaction Rates. J. Mol. 1997;94(6):2133-8. Biol. 2002:320(31:645-661. 29 Teigen Κ, Dao KK, Froystein NA. Corren AC. 34 Teigen Κ. Martínez A. Probing cofactor specificity Mayer Β, McKinney J, et al. The conformation of in pheny lalanine hy droxylase by molecular dynam- tetrahydrobiopterin free and bound to aromatic ics simulations. J. Biomol. Struct. Dyn. amino acid hydroxylases. Chemistry and Biology 2003:20( 6): 73 3-740. of Pteridines and Folates 2001:66-72. 35 Kobe B. Jennings IG, House CM, Feil SC, Micheli 30 Karlsson A, Parales JV, Parales RE. Gibson DT. BJ. Tiganis T. et al. Regulation and crystallization Eklund H, Ramaswamy S. Crystal structure of of phosphorylated and dephosphorylated forms of naphthalene dioxygenase: side-on binding of truncated dimeric phenylalanine hydroxylase. dioxygen to iron. Science 2003:299(5609): 1039- Protein Sci. 1997;6:1352-1357. 42. 31 Siegbahn PEM. Modeling Aspects of Mechanisms for Reactions Catalyzed by Metalloenzymes. Journal of Computational Chemistry 2001; 22:1634-1645.

Ptcridines/Vol. 16/No. 1