Structural Studies on Phenylalanine Hydroxylase and Implications Toward Understanding and Treating Phenylketonuria

Transcription

1 Structural Studies on Phenylalanine Hydroxylase and Implications Toward Understanding and Treating Phenylketonuria Heidi Erlandsen, DrSci; Marianne G. Patch, PhD; Alejandra Gamez, PhD; Mary Straub; and Raymond C. Stevens, PhD ABSTRACT. Mutations in the gene encoding for phenylalanine hydroxylase (PAH) result in phenylketonuria (PKU) or hyperphenylalaninemia (HPA). Several 3-dimensional structures of truncated forms of PAH have been determined in our laboratory and by others, using x-ray crystallographic techniques. These structures have allowed for a detailed mapping of the >250 missense mutations known to cause PKU or HPA found throughout the 3 domains of PAH. This structural information has helped formulate rules that might aid in predicting the likely effects of unclassified or newly discovered PAH mutations. Also, with the aid of recent crystal structure determinations of co-factor and substrate analogs bound at the PAH active site, the recently discovered tetrahydrobiopterin-responsive PKU/HPA genotypes can be mapped onto the PAH structure, providing a molecular basis for this tetrahydrobiopterin response. Pediatrics 2003;112: ; phenylalanine hydroxylase, 3-dimensional structure, structural basis, phenylketonuria, BH 4 - responsive hyperphenylalaninemia, co-factor therapy. ABBREVIATIONS. PAH, phenylalanine hydroxylase; l-phe, l- phenylalanine; BH 4, tetrahydrobiopterin; PKU, phenylketonuria; HPA, hyperphenylalaninemia; ARS, autoregulatory sequence; CBR, co-factor binding region. Human phenylalanine hydroxylase (PAH) converts the essential amino acid l-phenylalanine (l-phe) into l-tyrosine using the co-factor (6R)-l-erythro-5,6,7,8-tetrahydrobiopterin (BH 4 ) and molecular oxygen. Catalysis by this irondependent enzyme is the major pathway for catabolic degradation of dietary l-phe and accounts for approximately 75% of the l-phe disposal. The autosomal recessive disorder phenylketonuria (PKU) is the result of a deficiency of PAH enzymatic activity or loss of enzyme expression as a result of mutations in the PAH gene. Because of extensive newborn screening for PKU and genotyping of the PAH alleles, 400 mutations in the PAH gene are known to cause PKU or the milder form referred to as hyperphenylalaninemia (HPA). The recently solved crystal structures of PAH 1 6 provide a structural scaffold to explain the effects of PAH mutations. This review covers the structural work done so far on PAH and From the Department of Molecular Biology, Scripps Research Institute, La Jolla, California. Reprint requests (R.C.S.) to Department of Molecular Biology, Scripps Research Institute, North Torrey Pines Rd, La Jolla, CA stevens@scripps.edu PEDIATRICS (ISSN ). Copyright 2003 by the American Academy of Pediatrics. discusses some of the structural effects of the currently known mutations in the PAH gene, including some of the BH 4 -responsive PKU/HPA mutations (PKU database at and PAH Mutation Analysis Consortium Newsletter, December 2001). Human (liver) PAH (EC ) exists in a phdependent equilibrium of homotetramers and homodimers, 7 and, like the 2 other aromatic amino acid hydroxylases tyrosine hydroxylase (EC ) and tryptophan hydroxylase (EC ), consists of 3 domains: an N-terminal regulatory domain (residues 1 142), a catalytic domain (residues ), and a C-terminal tetramerization domain (residues ; Fig 1). Because of the difficulty of crystallizing full-length PAH, no full-length tetrameric structures exist for PAH. However, several truncated forms of PAH have been structurally characterized, including a dimeric form containing the regulatory and catalytic domains 3 and a tetrameric form containing the catalytic and tetramerization domains. 2 On the basis of these structures and a higher-resolution dimeric double-truncated form of PAH, 1 a composite full-length structural model was constructed by superimposing the respective catalytic domain regions (Figs 2 4). 8 The regulatory domain of PAH contains an - sandwich with an interlocking double motif ( topology) 3 (Fig 4). The N-terminal autoregulatory sequence (ARS; residues 19 33) extends over the active site in the catalytic domain. An N-terminal truncated form of PAH that includes the catalytic and tetramerization domains (residues ) crystallized as a tetramer (dimer of dimers). 2 The tetramerization domain contains 2 -strands, forming a -ribbon, and a 40-Å-long -helix. 2,9 The 4 -helices (1 from each monomer) pack into a tight antiparallel coiled-coil motif in the center of the tetramer structure (Fig 4). Recombinant double-truncated human PAH ( NH102- COOH428, hpah Cat ) represents a fully activated form of the enzyme, without any cooperativity of l-phe binding. 10 Thus, this structure represents the activated, or R state, of PAH. The catalytic domain region of the PAH structure (hpah Cat ) has a basket-like arrangement, with a total of 13 -helices and 8 -strands. Because part of the C-terminal tetramerization domain is missing, this double-truncated PAH construct forms a dimer in solution, as well as in the crystal. PEDIATRICS Vol. 112 No. 6 December

2 Fig 1. Scheme showing the secondary structure assignment of human PAH sequence (SWISS-PROT P00439). Secondary structure was assigned using the program DSSP 56 on the composite PAH model. 8 Residues that have PKU mutations associated with them are marked with grey dots. Secondary structural elements of the 3 regulatory domain are indicated as arrows for -strands and coils for -helical regions. Reprinted with permission from Erlandsen and Stevens. 49 Fig 2. Bar graph showing the truncations made in PAH to form crystals. The figure contains information about the oligomeric state of the protein along with the resolution of the crystal structures. The active site, located in the center of the catalytic domain, consists of a 13-Å-deep and 10-Å-wide pocket (Fig 3). 1 Adjacent to the active site is a 16-Ålong and 8-Å-wide channel that may provide substrate access to the active site. The majority of the 34 amino acids lining the active site are hydrophobic, but 3 charged glutamates, 2 histidines, and 1 tyrosine are also located in this region. Covering the entrance to the active site is a short loop (residues ) that contains some of the highest B-factors in the structure (60 80 Å2), indicative of a flexible, or disordered, loop region. As isolated, the PAH protein contains an active site Fe(III) atom. 1,11 In the crystal structures, the iron atom is located 10 Å below the surface of the protein on the floor of the active site, at the intersection of the channel and the active site pocket. The Fe(III) atom is coordinated to His285, His290, and 1 oxygen atom in Glu330. Both His285 and His290 have been shown by site-directed mutagenesis to be required for iron binding. 12 Well-defined electron density is also observed for 3 water molecules coordinated to the iron. The iron ligands are arranged in an octahedral geometry, making the iron 6-coordinate as previously suggested from spectroscopic studies. 13 The PAH hydroxylation reaction requires the binding of BH 4, dioxygen, and Fe(II) before hydroxylation of l-phe can occur. A stretch of 27 PAH amino acids (His263 to His289), which is highly conserved in all 3 aromatic amino acid hydroxylases, was 1558 STUDIES ON PAH AND UNDERSTANDING AND TREATING PKU

3 Fig 3. Structure of a monomer of human PAH full-length composite model. The regulatory domain (residues ), the catalytic domain (residues ), and the tetramerization domain comprise the full-length monomer. The active site iron is shown as a sphere. Reprinted with permission from Erlandsen and Stevens. 8 thought to be responsible for tetrahydrobiopterin binding 14,15 before any structures were known. In the PAH structures, 10 residues (Phe263, Cys265, Thr266, Thr278, Pro279, Glu280, Pro281, His285, Glu286, and Gly289) in this co-factor binding motif are present in the active site. Based on co-crystal structures of double-truncated Fe(III)-containing human PAH with oxidized cofactor (hpah Cat 7,8-BH 2 ) or Fe(II)-containing human PAH with reduced (natural) co-factor (hpah Cat BH 4 ), 5 structural details of co-factor binding have been determined. The pterin binds close to the catalytic iron and forms hydrogen bonding interactions with 2 of the water molecules coordinated to the iron, as well as to the main chain carbonyl oxygens of Ala322, Gly247, and Leu249; the main chain amide of Leu249; and the O atom of Ser251. 4,5 The pterin ring forms an aromatic -stacking interaction with Phe254, and the pterin ring and dihydroxypropyl side-chain are positioned by a Tyr325 hydrophobic interaction. The BH 2 dihydroxypropyl side chain O2 atom hydrogen bonds to the carbonyl oxygen of Ala322, and Glu286 hydrogen bonds to 1 of the water molecules coordinated to the iron and also hydrogen bonds to a water molecule that is hydrogen bonded to the N3 position of the pterin ring. In addition, a major conformational change occurs in the active site upon pterin binding; residues 245 to 250 move in the direction of the iron, allowing several protein hydrogen bonds to the pterin ring to be formed. Bound pterin co-factor is in an ideal location for dioxygen binding in a bridging position between the iron and the pterin, as is presumed to happen during the l-phe hydroxylation reaction. Although several alternative co-factor analogs are capable of being used in the PAH hydroxylation reaction, only the natural co-factor (BH 4 ) inhibits the l-phe induced activation of the enzyme 13 ; the molecular mechanism of this inhibition is not yet understood. 3,13,16 BH 2 inhibits the activation by l-phe less than the inhibition observed with BH Hydrogen bonding between the pterin dihydroxypropyl side chain and the carbonyl oxygen of Ala322, in combination with the PAH loop 245 to 250 conformational changes observed on co-factor binding, could provide a specific regulatory function of the pterin upon binding at the active site. It is interesting that when the binary complex hpah Cat 7,8-BH 2 structure is superimposed onto the structure of the ligand-free C-terminal truncated rat PAH (rpah Reg containing the regulatory and catalytic domains 3 ), BH 2 interacts with the N-terminal ARS region. This interaction of pterin with the ARS gives a plausible structural explanation for the inhibitory effect of BH 4 on the rate of phosphorylation of Ser16 in rpah by the camp-dependent protein kinase, which is specific for the R-isomer of the natural Fig 4. Three views of the tetrameric form of the composite model of PAH. The active site iron is shown as a sphere. A, Front view. B, Side view, seen in the plane of the paper along the x axis as compared with A. C, Side view, seen in the plane of the paper along the y axis as compared with A. Reprinted with permission from Erlandsen and Stevens. 49 SUPPLEMENT 1559

4 co-factor (BH 4 ). 18 Furthermore, phosphorylation of Ser16 in the mobile ARS region may facilitate access of l-phe to the active site. These conclusions are consistent with the finding that the phosphorylated form of rpah requires less l-phe to be activated (S M) than the nonphosphorylated form (S M). 18 Thus, the regulatory properties of BH 4 in full-length PAH does not require an additional binding site of the co-factor in the regulatory domain as previously postulated. 3 STRUCTURAL BASIS FOR HPA AND PKU Currently, 269 missense point mutations are known for the PAH gene (PAH Mutation Analysis Consortium Newsletter, December 2001 issue; PKU database at in addition to 23 nonsense mutations and 10 silent mutations. Most of the point mutations map onto the exon 5 (PAH residue 148) to exon 12 (PAH residue 438) region of the sequence. As shown in Figs 1 and 5, 57 PAH mutations are located in the regulatory domain sequence (residues 1 142), whereas 231 PAH mutations are located in the catalytic domain (residues ) and 14 are located in the tetramerization domain sequence (residues ). Previously, a summary of genotype/phenotype/structural interpretations for these mutations was published. 8 In vitro expression analyses have been used to characterize PAH gene mutations. In general, the PAH mutations that lead to PKU or HPA genotypes result in reduced enzyme activity and stability to varying extents, and some mutations also have been found to alter the oligomeric state of the protein. 19,20 At least 3 groups of HPA/PKU mutations have been categorized, dependent on the kinetic behavior and/or stability that has been observed. The first category involves mutations that affect both PAH kinetics and stability. A second class describes structurally stable mutations with altered kinetic properties, whereas a third class encompasses PAH mutations displaying normal kinetics but with reduced stability in vitro and in vivo. 21 In addition, on the basis of the 3-dimensional structure of PAH, mutations that cause PKU/HPA affect residues in 5 different categories: 1) residues located at the active site, 2) structural residues, 3) residues involving interdomain interactions in a monomer, 4) residues that interact with the N-terminal ARS, and 5) residues at the dimer or tetramer interface regions of the structure. The most prevalent single point mutations in the PAH gene are discussed according to these 5 different structural groups. PAH Active Site Mutations In total, 31 point mutations can be found in the residues lining the active site. Seven of these can be found in the putative pterin-binding motif, consisting of residues 264 to (PKU database at www. pahdb.mcgill.ca). The D143G mutation is associated with a severe PKU phenotype. 22 Asp143 is located at the entrance to the active site and may be involved in controlling substrate/co-factor access to the active site. Another PKU/HPA mutation is Thr278, which is also located at the entrance to the active site, forming a hydrogen bond with Glu ,24 The mutations T278I and T278A substitute a polar amino acid into different hydrophobic amino acids, losing an important hydrogen bond that would perturb the structure around the entrance to the active site. The more Fig 5. C -trace of 1 monomer of the full-length model of PAH. The trace is shaded for residues associated with PKU mutations, and the active site iron is shown for reference as a sphere. Note that sites of PKU mutations are located throughout the 3 domains of PAH. Reprinted with permission from Erlandsen and Stevens STUDIES ON PAH AND UNDERSTANDING AND TREATING PKU

5 electrostatically conserved mutation T278N would most likely result in a changed hydrogen bonding pattern that would have the same effect as the 2 other mutations, although presumably slightly smaller in magnitude. A frequent mutation among black Americans is the L255S mutation, which causes severe PKU. 25 Leu255 is most likely involved in controlling the separation of helices C 6 and C 9, and in this active site mutation, a nonpolar group is substituted with a smaller polar side chain that would be expected to result in significant structural perturbation. Another mutation in the active site is the E280K mutation, found within the pterin-binding motif 15 and associated with mild to severe PKU phenotypes. 26 Glu280 hydrogen bonds to His146 and also forms a salt bridge to Arg158. Also, in the active site of PAH, there are only 2 free, charged groups (both glutamic acids), and substitution of Glu280 into a lysine represents a dramatic change in the electrostatic potential of the active site. As expected, expression of E280K human PAH in Escherichia coli results in an enzyme with only approximately 1% of the specific activity of wild-type PAH. 27 The R158Q mutation is a frequent mutation in patients with PKU. As mentioned above, Arg158 forms a salt bridge to Glu280, but also forms a hydrogen bond to Tyr268. Both of these interactions are important for conserving the shape of the active site, and substitution into a glutamine or the larger aromatic residue tryptophan (R158W) will alter the active site architecture and lower enzymatic activity. 28 Another active site residue that is involved in BH 4 co-factor binding is Phe254, which -stacks onto the pterin ring in hpah Cat 7,8-BH 2. 4 Phe254 is located 5.9 Å away from the active site iron, and substitution into an isoleucine, as in PKU mutant F254I, would interfere with the proper binding of the pterin and possibly the Phe substrate. One of the most frequent PAH mutations found in southeastern Europe is the P281L mutation that is associated with HPA to severe PKU phenotypes. 29 Pro281 helps to define the shape of the active site very close to the iron. Therefore, substitution to a less rigid leucine will change the conformation of the active site by removing the conformational constraints imposed by the proline. As expected, in vitro expression of P281L PAH in E coli or African green monkey Cercopithecus aethiops kidney (COS) cells results in complete loss of PAH activity. 30,31 Three residues in the active site that are located near the (putative) location of substrate binding or in the region near the catalytic iron are Phe331, Ala345, and Gly346. Phe331 -stacks onto Trp326, a residue that may be involved in determining substrate specificity in the aromatic amino acid hydroxylases. 32 Two PKU mutations have been reported for Phe331, F331C and F331L, 33,34 and both of these mutations will remove important -stacking interactions of the active site. Both Ala345 and Gly346 are located close to the triad of residues that bind the iron (His285, His290, and Glu330). Ala345 has 2 reported PKU mutations, A345S and A345T, and Gly346 has 1 mutation, G346R. All 3 of these mutations into larger and more polar residues would interfere with the location and electrostatics of the residues that act as iron ligands, thus destabilizing the active site structure. One additional residue that is important in holding the catalytic iron in place is Ser349, which hydrogen bonds to His285, one of the iron ligands. Two PKU mutations have been reported for this residue, one that results in classical PKU (S349P 35 ). This substitution into a proline so close to the catalytic iron will change the shape of the active site. As expected, expression of S349P hpah in E coli or COS cells results in only 0.2 to 1% residual activity. The second mutation of Ser349, S349L, 36 also has minimal activity in E coli or COS cells. Both mutations will alter the hydrogen bond to His285, having severe detriment on the PAH catalytic activity. PAH Mutations in Structural Residues The G46S mutation in the regulatory domain is associated with classical PKU and is one of the most frequent mutations found clinically. Gly46 is located on the surface of the regulatory domain in a loop preceding helix R 1. Substitution into a serine would generate potential side-chain hydrogen bonding to 1 or more residues in close proximity to Gly46, resulting in distortions of the regulatory domain secondary structure. This G46S mutation thus potentially results in the formation of inactive PAH aggregates, as previously reported. 37 Another PAH mutation found in the regulatory domain is I65T. This mutation is associated with non-pku HPA to variant PKU. 38 Ile65 is located in the hydrophobic core of the regulatory domain, and substitution into a polar threonine (or asparagine as in the I65N mutant) would result in a significant structural perturbation of the regulatory domain core. A frequent PAH mutation found in patients with PKU is A104D. This mutation is associated with variant PKU, and Ala104 is located in a loop between R 2 and R 4 inthe regulatory domain. Substitution into a larger and charged residue may destabilize this loop structure. Consistent with this hypothesis, in vitro expression led to increased PAH aggregation and marginal activity. 39 Cys203 is involved in a disulfide bond with Cys334 in the rat PAH structure, 3 but this disulfide link is not observed in any of the human PAH structures. 1,2 Both cysteine residues are associated with PKU mutations. Arg243 is located at the end of C 1, forming a salt bridge to Asp129 in the C 1 helix. In vitro expression of either of 2 PKU mutants (R243Q or R243X) in COS cells resulted in 10% residual activity 40 or 1% residual activity. 26 Arg252 forms a salt bridge to Asp315 and hydrogen-bonds to the carbonyl oxygen of Ala313 as well as the side chain of Asp27 in the ARS. There are 3 PKU mutations associated with this residue: R252Q, R252G, and R252W. The R252W mutation results in classical PKU, and in vitro expression of the R252G and R252W mutations results in 1% residual activity. 25,27,28 Recombinant expression of the R252Q mutation, however, results in somewhat larger residual activity (16% and 3%). 25 Thus, any substitution at Arg252 results in disruption of stabilizing interactions in the catalytic domain. Ala259 is buried in a hydrophobic pocket approx- SUPPLEMENT 1561

6 imately 4 Å away from Leu311 and Leu308, which both hydrogen-bond to Arg408. These residues hold the tetramerization arm close to the catalytic domain, and thus substitution of Ala 259 into a larger or polar residue, as in A259T and A259V, cannot be accommodated. Both A259T and A259V hpah when expressed in vitro result in 0.2% to 0.3% residual activity. 27 One of the most frequent mutations in the PKU database is R261Q. This mutation results in phenotypes varying from variant PKU to classical PKU. Arg261 is located in a loop between helix C 6 and C 2 in the catalytic domain, forming hydrogen bonds to Gln304 and Thr238, and substituting an arginine for proline (R261P) would destabilize the catalytic domain structure. PAH Mutations Involved in Interdomain Interactions in a Monomer The most important residues at the interface between the catalytic and tetramerization domains are Leu311, Leu308, and Arg408. Arg408, located in the loop between C 12 and T 1, forms hydrogen bonds to the carbonyl oxygens of Leu308 (at the end of C 8) and Leu311 (in the loop between C 8 and C 9). R408W is the single most frequent mutation in the PKU database, resulting in a severe PKU phenotype, with a low ( 1% to 2.7%) recombinant PAH activity. 41,42 Another mutation, R408Q, associated with an HPA phenotype, reportedly shows 55% residual PAH activity. 42 Substitution of Arg408 into a larger and bulkier tryptophan would alter the hydrogen bonding network at the interface of the tetramerization and catalytic domains, interfering with the correct positioning of the -ribbon (T 1 and T 2). PAH Mutations in Residues That Interact With the N-Terminal ARS Tyr377 is 1 of the residues in the catalytic domain that is in contact with the N-terminal ARS 3 and hydrogen-bonds to Ser23 in the regulatory domain, along with stacking on top of Trp326 (a residue involved in determining substrate specificity for the aromatic amino acid hydroxylases). The proximity of Tyr377 to the ARS and the putative substrate binding site may regulate access to the active site. In the PAH PKU mutant Y377C, the hydrogen bond to Ser23 would be lost, altering access to the active site. PAH Mutation in Residues at the Dimer or Tetramer Interfaces Residues Ser67, Glu76, Arg297, Gln304, and Glu422 can be found at the dimer interface and have PKU mutations associated with them. 1 3 Ser67 is located in -strand R 2, forming a hydrogen bond to Tyr216 in another monomer; the PKU mutation S67P would alter the PAH structure at this interface. Glu76 is located in -strand R 3 on the surface of the regulatory domain, hydrogen-bonding to His208 in another monomer, as well as Asn73 in the same monomer. The PKU mutations E76A and E76G are substitutions of a charged residue into surface-exposed hydrophobic residues, resulting in destabilization of the secondary structure in the regulatory domain. Arg297 is located in -helix C 8 and hydrogen-bonds to Arg71 as well as Glu422 in a second monomer. Arg297 has 2 PKU mutations, R297H and R297C, 37 and both substitutions will result in disrupting the dimer-stabilizing hydrogen bonds, resulting in unstable PAH. Gln304 in -helix C 8 is also involved in dimer/tetramer interactions, forming hydrogen bonds to Tyr417 in a second monomer, and Arg261. Gln304 has 2 PKU mutations, a silent mutation (Q304Q) 43 and Q304R; substitution of a polar glutamine to a charged arginine would alter both hydrogen bonds and destabilize PAH. Last, both Pro362 and Pro366 position Glu368, which mediate dimer contacts. Thus, the PKU mutations P362T 44 and P366H 45 will most likely prevent proper dimer formation. POTENTIAL PKU/HPA THERAPEUTIC APPROACHES In light of the difficulty of PKU/HPA patients in adhering to a restrictive diet for life, 46,47 new and less restrictive therapies for PKU/HPA are needed. Because gene therapy solutions are many years off, alternative solutions such as enzyme replacement therapy using recombinantly produced PAH (in a truncated, monomeric, and constitutively active form) or the plant enzyme phenylalinine ammonia lyase may provide some relief to PKU/HPA Phe load. For example, phenylalanine has been shown to lower l-phe levels, as well as reduce the symptoms of HPA, in mice engineered with a model for HPA. 48 BH 4 -RESPONSIVE HPA/PKU MUTATIONS Recently, a number of patients harboring a subset of PAH mutations showed a marked lowering (normalization) of their blood l-phe levels upon oral administration of the PAH co-factor BH 4 (10 20 mg BH 4 /kg body weight). 49,51 It is believed that this subset of mutations results in expressed mutant enzymes that are K m variants of PAH with an altered binding affinity for BH 4. More and more PAH mutations are continuing to be reported as being responsive to BH 4, 50,51 and characterization of the biochemical, molecular, and physiologic bases for their BH 4 responsiveness is required for a detailed understanding of this phenomenon. Most of the BH 4 -responsive PKU mutations are in the PAH catalytic domain, either located in co-factor binding regions (CBR 1 [residues ], CBR 2 [residues ], CBR 3 [residues ], and CBR 4 [residues ]) or in locations that directly interact with the CBR regions involved with co-factor binding. The following PAH genotypes were found to be responsive to BH 4 : P407S/R252W, IVS4 1g- a/a373t, R413P/R241C, 52 IVS10nt-11g- a/ E390G, 53 A313T/1099insC, V190A/R243X, A300S/ A403V, and R241C/A403V. 54 These mutations are found in distinct regions of the PAH primary sequence (Fig 1) and tertiary structure of the composite model of hpah (Fig 6). Neither residue in the R252W/P407S genotype interacts directly with the co-factor, but Arg252 follows Ser251 and is located in CBR 1 (Figs 1 and 6). Ser251 is presumed to position BH 4 by hydrogen bonding to the dihydroxypropyl side chain. As previously men STUDIES ON PAH AND UNDERSTANDING AND TREATING PKU

7 Fig 6. Backbone structure of a monomer of the composite model of PAH. Side chains of the residues involved in BH 4 responsiveness are shown. 49 The active site iron is shown as a sphere. The co-factor BH 4 is positioned into the structure based on the electron density for the co-factor analog 7,8-dihydro-l-biopterin co-complex crystal structure with human double-truncated PAH. 4 The CBRs are shown as darkened loops (CBR 1 [residues ], CBR 2 [residues ], CBR 3 [residues ], and CBR 4 [residues ]). Reprinted with permission from Erlandsen and Stevens. 49 tioned, Arg252 hydrogen bonds to residues Ala313, Asp315, and Asp27 in the ARS. The second allele residue in the R252W/P407S genotype, Pro407, precedes the start of the tetramerization domain. Mutation into a less rigid serine residue might inhibit PAH tetramerization; however, dimeric or monomeric P407S mutant PAH might still retain co-factor binding affinity and activity. Kure et al 52 also reported a non BH 4 -responsive patient, having a P407S/R111X genotype. The R111X mutation results in a truncated PAH lacking enzymatic activity. Therefore, enzymatic activity for this second genotype must be attributable to the P407S PAH. On the basis of these observations, BH 4 responsiveness is dependent on the specific PAH genotype. The R413P/R241C genotype was also responsive to co-factor loading. 52 Arg241 is located on the surface of the catalytic domain preceding 1 of the CBRs, at the start of C 1. In the tetrameric structure of PAH, Arg241 is hydrogen-bonded to Gln419 of the tetramerization domain, participating in important intradomain interactions. 2 Arg413 is located in the tetramerization domain and hydrogen-bonds to Glu422, forming 1 of a few hydrogen bonds that hold the -ribbon formed by T 1 and T 2 together. This interaction ensures proper tetramer formation by positioning the 40-Å-long tetramerization domain -helix for intermonomer coiled-coil interactions. Three PKU mutations have been found for Arg413: R413S, R413C, and R413P. The R241C mutation, as a result of the mild effect of this mutation on the PAH structure, as compared with the R413P mutation, must be responsible for the BH 4 responsiveness of the R413P/R241C genotype. Trefz et al 53 reported a patient with PKU who was responsive to BH 4 oral loading, having an IVS10nt- 11g- a/e390g genotype. The latter mutation, E390G, must be responsible for the BH 4 responsiveness, because the other mutation leads to a splicing defect in intron 10 of the PAH gene, with no functional protein expressed. The E390G PAH mutation shows 70% residual activity when expressed in COS cells (PKU database). In the structure of tetrameric PAH, Glu390 is located on the surface in the catalytic domain, following 1 of the CBRs (CBR 4) 2 (Figs 1 and 6). The Glu390 side chain points toward the catalytic domain of a second monomer and also toward the tetramerization interface. Because Glu390 is located on the surface, mutation into a flexible glycine may destabilize the PAH tetramer and reduce enzymatic activity somewhat. The E390G mutant PAH most likely retains co-factor binding ability but with an increased K m ; however, this hypothesis must still be experimentally confirmed. Another BH 4 -responsive genotype reported by Kure et al 52 is IVS4 1g- a/a373t, with a mild HPA phenotype. The first mutation affects intron 4, resulting in nonfunctional protein expression. Ala373 forms hydrophobic interactions with helix C 9 containing Tyr325, Trp326, and Ala322 (located in CBR 3), which hydrogen-bonds to the dihydroxypropyl side chain of the co-factor. An A373T substitution might be relatively easily accommodated in this region of the structure and might result in minimal structural perturbation of the C 9 helix and the active site. Thus, this A373T PAH retains some binding affinity for BH 4, providing for the observed BH 4 responsiveness. The variant HPA genotype A313T/1099insC was also found to be responsive to BH 4 loading. 54 The SUPPLEMENT 1563

8 A313T allele is most likely BH 4 responsive, because the 1099insC mutation results in a frameshift after Leu367, but this frameshift mutation has all of the residues needed for catalysis and so might retain some enzymatic activity. Unfortunately, no expression information for either of these mutants exists. Ala313 hydrogen-bonds to Arg252 (in CBR 1) and is located in a loop between helices C 8 and C 9, at the interface with the regulatory domain (close to Pro119). The effect of the A313T substitution on cofactor binding might not be too large because the residue preceding Ala313 is a glycine, which might compensate for a distortion in the backbone imposed by a threonine mutation at Ala313. Genotype V190A/R243X also displays variant HPA, and the BH 4 -responsive mutation must be V190A, because the R243X mutation results in a truncated enzyme missing all of the catalytically required residues. Val190 is located in helix C 3, behind His285, and close to Cys284 and Arg270. Substitution into an alanine at position 190 would create space where the valine side chain was originally located. The mutant enzyme might have Cys284 (located at the end of CBR 2 [residues ]) and His285 rearranged into the empty space created by the V190A substitution. These alterations to the co-factor binding site might require higher concentrations of co-factor for productive catalysis, consistent with the BH 4 responsiveness observed. Variant HPA genotypes A300S/A403V and R241C/A403V both were found responsive to BH 4 loading. Both genotypes contain the A403V allele. Ala403 is located at the end of helix C 12, close to Ala309 in helix C 8. Alanine or another smaller residue might be necessary for close packing of helices C 8 and C 12. Thus, substitution into a larger valine as in A403V PAH might result in a less stable protein. The R241C/A403V genotype also shares 1 allele with R413P/R241C, and in this genotype, the R241C mutation was presumed to be responsible for the BH 4 responsiveness, making the R241C allele the responsive mutation for the R241C/A403V genotype as well. In A300S/A403V, Ala300 is located in the middle of helix C 8, lining the bottom of the active site. Ala300 does not form any hydrogen bonds, so substitution into a serine might be allowed, with minor structural perturbations. The closest residue to Ala300 is Arg261, located in CBR 1 (residues ). Substitution of Ala300 into a slightly larger serine could alter the position of Arg261, altering the co-factor binding site. Thus, the BH 4 binding site in A403V PAH might be only slightly different as compared with the wild-type PAH structure, explaining the BH 4 responsiveness of this genotype. BH 4 oral therapy has the potential to become more widely used, provided the cost of purchasing the necessary large amounts of the co-factor does not continue to be an issue. It remains to be seen exactly what percentage of patients with PKU/HPA will show a response to BH 4 dietary supplementation, but similar to other high-dose vitamin therapies, 55 it is hoped that BH 4 therapy will become an important addition to the strict PKU/HPA diet, making these conditions easier to maintain for life. CONCLUSIONS Crystal structures of different truncated forms of human and rat PAH have allowed for the construction of a composite structural model for full-length, tetrameric PAH and provided a structural basis for the numerous mutations resulting in deficient PAH activity. In addition, PAH crystal structures with substrate analog, inhibitor, or co-factor bound at the active site have provided details of ligand binding. This structural information has helped formulate rules that may aid in predicting the likely effects of unclassified or newly discovered PAH mutations. Also, with the aid of recent crystal structure determinations, recently discovered BH 4 -responsive PKU/HPA genotypes can be mapped onto PAH structure, providing a molecular basis for this BH 4 - dependent response. However, more structural work is needed, such as determination of the structure of full-length PAH, as well as more site-directed mutagenesis studies of PAH, to understand completely the catalytic mechanism and substrate specificity of PAH and tyrosine hydroxylase/tryptophan hydroxylase, the other members of the aromatic amino acid hydroxylases. ACKNOWLEDGMENTS This study was funded by National Institutes of Health grant no. R01-HD38718 from the National Institute of Child Health and Human Development (Bethesda, MD). REFERENCES 1. Erlandsen H, Fusetti F, Martinez A, Hough E, Flatmark T, Stevens RC. Crystal structure of the catalytic domain of human phenylalanine hydroxylase reveals the structural basis for phenylketonuria. Nat Struct Biol. 1997;4: Fusetti F, Erlandsen H, Flatmark T, Stevens RC. Structure of tetrameric human phenylalanine hydroxylase and its implications for phenylketonuria. J Biol Chem. 1998;273: Kobe B, Jennings IG, House CM, et al. Structural basis of autoregulation of phenylalanine hydroxylase. Nat Struct Biol. 1999;6: Erlandsen H, Bjørgo E, Flatmark T, Stevens RC. Crystal structure and site-specific mutagenesis of pterin-bound human phenylalanine hydroxylase. Biochemistry. 2000;39: Andersen OA, Flatmark T, Hough E. High resolution crystal structures of the catalytic domain of human phenylalanine hydroxylase in its catalytically active Fe(II) form and binary complex with tetrahydrobiopterin. J Mol Biol. 2001;314: Andersen OA, Flatmark T, Hough E. Crystal structure of the ternary complex of the catalytic domain of human phenylalanine hydroxylase with tetrahydrobiopterin and 3-(2-thienyl)-L-alanine, and its implications for the mechanism of catalysis and substrate activation. J Mol Biol. 2002;320: Martinez A, Knappskog PM, Olafdottir S, et al. Expression of recombinant human phenylalanine hydroxylase as fusion protein in Escherichia coli circumvents proteolytic degradation by host cell proteases. Biochem J. 1995;306: Erlandsen H, Stevens RC. The structural basis of phenylketonuria. Mol Genet Metab. 1999;68: Goodwill KE, Sabatier C, Marks C, Raag R, Fitzpatrick PF, Stevens RC. Crystal structure of tyrosine hydroxylase at 2.3 Å and its implications for inherited neurodegenerative diseases. Nat Struct Biol. 1997;4: Knappskog PM, Flatmark T, Aarden JM, Haavik J, Martinez A. Structure/function relationships in human phenylalanine hydroxylase. Effect of terminal deletions on the oligomerization, activation and cooperativity of substrate binding to the enzyme. Eur J Biochem. 1996;242: Martinez A, Andersson KK, Haavik J, Flatmark T. EPR and 1H-NMR spectroscopic studies on the paramagnetic iron at the active site of phenylalanine hydroxylase and its interaction with substrates and inhibitors. Eur J Biochem. 1991;198: STUDIES ON PAH AND UNDERSTANDING AND TREATING PKU

9 12. Gibbs BS, Wojchowski D, Benkovic SJ. Expression of rat liver phenylalanine hydroxylase in insect cells and site-directed mutagenesis of putative non-heme iron-binding sites. J Biol Chem. 1993;268: Kappock TJ, Caradonna JP. Pterin-dependent amino acid hydroxylases. Chem Rev. 1996;96: Hufton SE, Jennings IG, Cotton RGH. Structure and function of the aromatic amino acid hydroxylases. Biochem J. 1995;311: Jennings IG, Kemp BE, Cotton RGH. Localization of cofactor binding sites with monoclonal anti-idiotype antibodies: phenylalanine hydroxylase. Proc Natl Acad Sci U S A. 1991;88: Flatmark T, Stevens RC. Structural insights into the aromatic amino acid hydroxylases and their disease-related mutant forms. Chem Rev. 1999; 99: Xia, T, Gray DW, Shiman R. Regulation of rat liver phenylalanine hydroxylase. III. Control of catalysis by (6R)-tetrahydrobiopterin and phenylalanine. J Biol Chem. 1994;269: Døskeland AP, Døskeland SO, Øgreid D, Flatmark T. The effect of ligands of phenylalanine 4-monooxygenase on the camp-dependent phosphorylation of the enzyme. J Biol Chem. 1984;259: Gamez A, Perez B, Ugarte M, Desviat LR. Expression analysis of phenylketonuria mutations. Effect on folding and stability of the phenylalanine hydroxylase protein. J Biol Chem. 2000;275: Kayaalp E, Treacy E, Waters PJ, Byck S, Nowacki P, Scriver CR. Human phenylalanine hydroxylase mutations and hyperphenylalaninemia phenotypes: a metanalysis of genotype-phenotype correlations. Am J Hum Genet. 1997;61: Flatmark T, Knappskog PM, Bjørgo E, Martinez A. Molecular characterization of disease related mutant forms of human phenylalanine hydroxylase and tyrosine hydroxylase. In: Pfleiderer W, Rokos H, eds. Chemistry and Biology of Pteridines and Folates. London, UK: Blackwell Science; 1997: Knappskog PM, Eiken HG, Martinez A, Bruland O, Apold J, Flatmark T. PKU mutation (D143G) associated with an apparent high residual enzyme activity: expression of a kinetic variant form of phenylalanine hydroxylase in three different systems. Hum Mutat. 1996;8: Guldberg P, Romano V, Ceratto N, et al. Mutational spectrum of phenylalanine hydroxylase deficiency in Sicily: implications for diagnosis of hyperphenylalaninemia in southern Europe. Hum Mol Genet. 1993;2: Takahashi K, Masamune A, Kure S, Matsubara Y, Narisawa K. Ectopic transcription: an application to the analysis of splicing errors in phenylalanine hydroxylase mrna. Acta Paediatr Suppl. 1994;407: Bjørgo E, Knappskog PM, Martinez A, Stevens RC, Flatmark T. Partial characterization and three-dimensional-structural localization of eight mutations in exon 7 of the human phenylalanine hydroxylase gene associated with phenylketonuria. Eur J Biochem. 1998;257: Okano Y, Eisensmith RC, Guttler F, et al. Molecular basis of phenotypic heterogeneity in phenylketonuria. N Engl J Med. 1991;324: Knappskog PM, Olafsdottir S, Haavik J, et al. Expression of wildtype and mutant forms of human phenylalanine hydroxylase in E. coli. In: Ayling JE, Gopal Nair M, Baugh CM, eds. Chemistry and Biology of Pteridines and Folates. New York, NY: Plenum Press; 1993: Okano Y, Wang T, Eisensmith RC, Steinmann B, Gitzelmann R, Woo SL. Missense mutations associated with RFLP haplotypes 1 and 4 of the human phenylalanine hydroxylase gene. Am J Hum Genet. 1990;46: Baric I, Mardesic D, Sarnavoka V, Lichter-Konecki U, Konecki DS, Trefz FK. Geographical distribution of the P281L mutation at the phenylalanine hydroxylase locus: possible origin in southeastern Europe. J Inherit Metab Dis. 1994;7: Dworniczak B, Grudda K, Stumper J, Bartholome K, Aulehla-Scholz C, Horst J. Phenylalanine hydroxylase gene: novel missense mutation in exon 7 causing severe phenylketonuria. Genomics. 1991;9: Okano Y, Wang T, Eisensmith RC, et al. Phenylketonuria missense mutations in the Mediterranean. Genomics. 1991;9: McKinney J, Teigen K, Frøystein NA, et al. Conformation of the substrate and pterin cofactor bound to human tryptophan hydroxylase. Important role of Phe313 in substrate specificity. Biochemistry. 2001;40: Benit P, Rey F, Melle D, Munnich A, Rey J. Five novel missense mutations of the phenylalanine hydroxylase gene in phenylketonuria. Hum Mutat. 1994;4: Guldberg P, Levy HL, Hanley WB, et al. Phenylalanine hydroxylase gene mutations in the United States: report from the Maternal PKU Collaborative Study. Am J Hum Genet. 1996;59: Knappskog PM, Eiken HG, Martinez A, Flatmark T, Apold J. The PKU mutation S349P causes complete loss of catalytic activity in the recombinant phenylalanine hydroxylase enzyme. Hum Genet. 1995;95: De Lucca M, Perez B, Desviat LR, Ugarte M. Molecular basis of phenylketonuria in Venezuela: presence of two novel null mutations. Hum Mutat. 1998;11: Eiken HG, Knappskog PM, Boman H, et al. Relative frequency, heterogeneity and geographic clustering of PKU mutations in Norway. Eur J Hum Genet. 1996;4: John SW, Scriver CR, Laframboise R, Rozen R. In vitro and in vivo correlations for I65T and M1V mutations at the phenylalanine hydroxylase locus. Hum Mutat. 1992;1: Waters PJ, Parniak MA, Hewson AS, Scriver CR. Alterations in protein aggregation and degradation due to mild and severe missense mutations (A104D, R157N) in the human phenylalanine hydroxylase gene (PAH). Hum Mutat. 1998;12: Wang T, Okano Y, Eisensmith RC, et al. Missense mutations prevalent in Orientals with phenylketonuria: molecular characterization and clinical implications. Genomics. 1991;10: DiLella AG, Marvit J, Brayton K, Woo SL. An amino-acid substitution involved in phenylketonuria is in linkage disequilibrium with DNA haplotype 2. Nature. 1987;327: Svensson E, Eisensmith RC, Dworniczak B, et al. Two missense mutations causing mild hyperphenylalaninemia associated with DNA haplotype 12. Hum Mutat. 1992;1: Perez B, Desviat LR, De Lucca M, Ugarte M. Spectrum and origin of phenylketonuria mutations in Spain. Acta Paediatr Suppl. 1994;407: Mallolas J, Campistol J, Lambruschini N, et al. Two novel mutations in exon 11 of the PAH gene (V1163del TG and P362T) associated with classic phenylketonuria and mild phenylketonuria. Hum Mutat. 1998; 11: Guldberg P, Henriksen KF, Guttler F. Molecular analysis of phenylketonuria in Denmark: 99% of the mutations detected by denaturing gradient gel electrophoresis. Genomics. 1993;17: Walter JH, White FJ, Hall SK, et al. How practical are recommendations for dietary control in phenylketonuria? Lancet. 2002;360: Levy HL. Phenylketonuria: old disease, new approach to treatment. Proc Natl Acad Sci U S A. 1999;96: Sarkissian CN, Shao Z, Blain F, et al. A different approach to treatment of phenylketonuria: phenylalanine degradation with recombinant phenylalanine ammonia lyase. Proc Natl Acad Sci U S A.1999;96: Erlandsen H, Stevens RC. A structural hypothesis for BH 4 responsiveness in patients with mild forms of hyperphenylalaninaemia and phenylketonuria. J Inherit Metab Dis. 2001;24: Lassker U, Zschocke J, Blau N, Santer R. Tetrahydrobiopterin responsiveness in phenylketonuria. Two new cases and a review of molecular genetic findings. J Inherit Metab Dis. 2002;25: Blau N, Trefz FK. Tetrahydrobiopterin-responsive phenylalanine hydroxylase deficiency: possible regulation of gene expression in a patient with the homozygous L48S mutation. Mol Genet Metab. 2002;75: Kure S, Hou DC, Ohura T, et al. Tetrahydrobiopterin-responsive phenylalanine hydroxylase deficiency. J Pediatr. 1999;135: Trefz F, Blau N, Aulehla-Scholz C, Korall H. Frauendienst-Egger G. Treatment of mild phenylketonuria (PKU) by tetrahydrobiopterin (BH 4 ). J Inherit Metab Dis. 2000;23(suppl 1): Spaapen LJM, Bakker JA, Velter C, et al. Tetrahydrobiopterinresponsive hyperphenylalaninemia (HPA) in Dutch neonates. J Inherit Metab Dis. 2000;23(suppl 1): Ames BN, Elson-Schwab I, Silver EA. High-dose vitamin therapy stimulates variant enzymes with decreased coenzyme binding affinity (increased K(m)): relevance to genetic disease and polymorphisms. Am J Clin Nutr. 2002;75: Kabsch W, Sander C. Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features. Biopolymers. 1983;22: SUPPLEMENT 1565

10 Structural Studies on Phenylalanine Hydroxylase and Implications Toward Understanding and Treating Phenylketonuria Heidi Erlandsen, Marianne G. Patch, Alejandra Gamez, Mary Straub and Raymond C. Stevens Pediatrics 2003;112;1557 Updated Information & Services References Subspecialty Collections Permissions & Licensing Reprints including high resolution figures, can be found at: This article cites 54 articles, 11 of which you can access for free at: #BIBL This article, along with others on similar topics, appears in the following collection(s): Genetics Information about reproducing this article in parts (figures, tables) or in its entirety can be found online at: Information about ordering reprints can be found online:

11 Structural Studies on Phenylalanine Hydroxylase and Implications Toward Understanding and Treating Phenylketonuria Heidi Erlandsen, Marianne G. Patch, Alejandra Gamez, Mary Straub and Raymond C. Stevens Pediatrics 2003;112;1557 The online version of this article, along with updated information and services, is located on the World Wide Web at: Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it has been published continuously since Pediatrics is owned, published, and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois, Copyright 2003 by the American Academy of Pediatrics. All rights reserved. Print ISSN: