(tcleapaopotpgegtrrhttbs

nt

n

7

a

MA

MINIREVIEW

The Structural Basis of Phenylketonuria

Heidi Erlandsen and Raymond C. Stevens1

The Scripps Research Institute, Department of Molecular Biology and Institute for Childhood

olecular Genetics and Metabolism 68, 103–125 (1999)rticle ID mgme.1999.2922, available online at http://www.idealibrary.com on

and Neglected Diseases, La Jolla, California 92037

ugust

teL

rspp

eUdcfs2amabzcaH(pg

nhds

Received A

The human phenylalanine hydroxylase genePAH) (locus on human chromosome 12q24.1) con-ains the expressed nucleotide sequence which en-odes the hepatic enzyme phenylalanine hydroxy-ase (PheOH). The PheOH enzyme hydroxylates thessential amino acid L-phenylalanine resulting innother amino acid, tyrosine. This is the majorathway for catabolizing dietary L-phenylalaninend accounts for approximately 75% of the disposalf this amino acid. The autosomal recessive diseasehenylketonuria (PKU) is the result of a deficiencyf PheOH enzymatic activity due to mutations inhe PAH gene. Of the mutant alleles that cause hy-erphenylalaninemia or PKU 99% map to the PAHene. The remaining 1% maps to several genes thatncode enzymes involved in the biosynthesis or re-eneration of the cofactor ((6R)-L-erythro-5,6,7,8-etrahydrobiopterin) regenerating the cofactor (tet-ahydrobiopterin) necessary for the hydroxylationeaction. The recently solved crystal structures ofuman phenylalanine hydroxylase provide a struc-ural scaffold for explaining the effects of some ofhe mutations in the PAH gene and suggest futureiochemical studies that may increase our under-tanding of the PKU mutations. © 1999 Academic Press

Key Words: phenylalanine hydroxylase; hyperphe-ylalaninemia; phenylketonuria; structure-func-ion relationship.

Phenylalanine hydroxylase (PheOH,2 phenylala-ine 4-monooxygenase, EC 1.14.16.1) is an iron-con-

1 To whom correspondence should be addressed. Fax: (858)

84-9483. E-mail: stevens@scripps.edu.

2 Abbreviations used: PKU; phenylketonuria, PheOH; phenyl-lanine hydroxylase, HPA; hyperphenylalaninemia; PAH: phe-

sm

103

10, 1999

aining enzyme that catalyzes the conversion of thessential amino acid L-phenylalanine (L-Phe) into-tyrosine, utilizing the cofactor 6(R)-L-erythro-tet-ahydrobiopterin (BH4) and dixoygen (for reviews,ee (1–3)). This catabolic pathway accounts for ap-roximately 75% or more of the disposal of dietaryhenylalanine (4).Mutations in the human gene encoding the PheOH

nzyme (PAH; GenBank cDNA Reference Sequence49897) result in the autosomal recessively inheritedisease hyperphenylalaninemia (HPA). The resultinglinical phenotypes can range in severity from mildorms of non-PKU HPA, with benign outcome, to theevere form, phenylketonuria (PKU3; OMIM Entry61600). Impaired function of PheOH results in theccumulation of high levels of blood plasma L-Phe andental impairment if not treated with diet at an early

ge. Most of the mutant alleles map to the PAH gene,ut approximately 1% map to several genes for en-ymes involved in the biosynthesis or regeneration ofofactor (BH4) (5). Due to an extensive and widelypplied population screening worldwide for neonatalPA, more than 380 mutations have been identified

including missense, nonsense, deletion, insertion,olymorphisms, and splice mutations) in the PAHene (PAH Mutation Analysis Consortium database

ylalanine hydroxylase-encoding gene, BH4; 6(R)-L-erythro-tetra-ydrobiopterin (natural cofactor), 6-MPH4; 6-methyltetrahy-ropterin, TnT; T7-coupled transcription-translation expressionystem.

3 The PKU database is accessible at the PAH Mutation Analy-is Consortium database and web site, located at http://www.cgill.ca/pahdb (6–8).

1096-7192/99 $30.00Copyright © 1999 by Academic Press

All rights of reproduction in any form reserved.

a6htfiattcsy1btairPttpqmt

S

le(maudapau(maostlom1f

sldpurple, the catalytic domain in gold, and the tetramerizationdomain in green. The iron is shown as a red sphere. (B, C) Twoperpendicular views of the full-length PheOH model structure,

cim

1 AND

nd web site at http://www.mcgill.ca/pahdb, Refs.–8). Work on in vitro expression analysis of inheriteduman mutations is being carried out in order to linkhe genotype to the phenotype (4). This work can con-rm that individual mutations impair PheOH enzymectivity and to a certain degree predict the severity ofhe corresponding metabolic phenotype. In this con-ext, the recently solved structures of several trun-ated forms of PheOH (9–11) as well as the recenttructure of the homologous enzyme tyrosine hydrox-lase (TyrOH, tyrosine 3-monooxygenase, EC.14.16.2) with bound cofactor analogue (7,8-dihydro-iopterin) (12) will prove to be invaluable in the iden-ification of amino acid residues involved in catalysisnd regulation of enzymatic activity, as well as mak-ng it possible to give a complete analysis of the cur-ent (Spring 1998, PKU Newsletter) mutations in theAH gene. Thus, in this review, a structural interpre-ation of the known mutations will be compared withhe known data on in vitro expression analyses, PKUhenotypes, and both predicted and observed fre-uency of mutation in the PAH gene in order to have aore complete picture of the disease-causing PKU mu-

ations.

tructure of Human Phenylalanine Hydroxylase

Recombinant human liver phenylalanine hydroxy-ase is reported to exist in solution as a pH-dependentquilibrium between homotetramers and homodimers13). The molecular mass of one subunit is approxi-ately 50 kDa (between 50 and 53 kDa). It has an

bsolute requirement for ferrous iron to be active andtilizes the cofactors BH4 and O2 to perform the hy-roxylation reaction. PheOH, like the two otherromatic amino acid hydroxylases (TyrOH and trypto-han hydroxylase (TrpOH, tryptophan 5-monooxygen-se, EC 1.14.16.4)), consists of three domains: the reg-latory domain (residues 1–142), the catalytic domainresidues 143–410), and a short tetramerization do-ain (residues 411–452). The regulatory domain hascommon a-b sandwich motif commonly observed for

ther regulatory domains, consisting of a four-tranded antiparallel b-sheet flanked on one side bywo short a-helices and on the other side by the cata-ytic domain (11). The active-site iron resides in anpen and solvent accessible region in the catalytic do-ain, the bottom of a novel basket-like arrangement of

4 a-helices and 8 b-strands. The short tetrameriza-

STEVENS

FIG. 1. (A) Structure of human phenylalanine hydroxylaseull-length composite model. The model was generated using thetructures of catalytic/tetramerization domains (9,10) and regu-atory/catalytic domains (11) and superimposing the catalyticomains of the two models. The regulatory domain is shown in

04 ERLANDSEN

olored from red (N-terminus in monomer A) to blue (C-terminusn monomer D). The iron is shown as a gray sphere in all four

onomers.

tabtorrbls

dHwcwswpritl

ec

P

PRrsigtlaftttll

ron. Tl that haP her wi

ion domain consists of two antiparallel b-strands andsingle 40 Å long C-terminal a-helix that is responsi-

le for forming the domain-swapped coiled-coil core ofhe PheOH tetramer. Based on the crystal structuresf several truncated forms of PheOH, including theegulatory/catalytic domains (11) and catalytic/tet-amerization domains (9,10), a full-length model cane constructed by superimposing the respective cata-ytic domains (Fig. 1). The details of the secondarytructure of the full-length model are shown in Fig. 2.

The active-site catalytic iron is ligated in an octahe-ral manner by two histidine residues (His285 andis290), one glutamic acid residue (Glu330), and threeater molecules (Fig. 2). There also is a narrow tunnel

onnected to the active-site pocket, and this may behere the substrate is directed into the active site. A

hort loop made up of residues 378 to 381 forms oneall of the tunnel that runs out into the active-siteocket. This loop shows high mobility in the high-esolution structure of dimeric PheOH (9) and could be

FIG. 2. Close-up of the active site surrounding the catalytic iiganded to the iron are shown as blue spheres. Some residueshe331) and are located close to the iron in the active site, toget

THE STRUCTU

nvolved in restricting accessibility of substrates intohe active site. The majority of the 34 amino acidsining the active site “bowl” are hydrophobic residues

etb

xcept for three charged glutamic acids, two putativelyharged histidines, and an uncharged tyrosine.

KU Mutations

Currently, most of the mutations that cause HPA orKU have been characterized only at the DNA level.ecent genotype-phenotype correlation studies haveevealed that the human PheOH genotype does notufficiently explain the mutant phenotype in all cases,mplying that other factors are involved in the emer-ent property of plasma homostasis (8). Some evidenceoward understanding the molecular mechanismseading to the phenotypic heterogeneity have beenchieved through in vitro expression of recombinantorms of human PheOH (14). This method has provedo be important when it comes to characterization ofhe different mutations’ individual catalytic, regula-ory, and stability properties, mainly due to the veryimited availability of endogeneous (wild-type, humaniver) enzyme. These expression analyses, some in sev-

he iron is shown as a red sphere and the three water moleculesve reported PKU mutations (Thr278, Glu280, Pro281, Trp326,

th their interacting residues, are also shown.

105ASIS OF PKU

RAL B

ral in vitro expression systems, have revealed at leasthree groups of mutations that differ in their kineticehavior and/or stability in vivo: (i) mutations affect-

1

TABLE 1Single-Point Mutations Identified in the PAH Gene (not Including Silent

and Missense Mutations)

06 ERLANDSEN AND STEVENS

TABLE 1—Continued

107THE STRUCTURAL BASIS OF PKU

1

TABLE 1—Continued

08 ERLANDSEN AND STEVENS

TABLE 1—Continued

109THE STRUCTURAL BASIS OF PKU

1

TABLE 1—Continued

10 ERLANDSEN AND STEVENS

TABLE 1—Continued

111THE STRUCTURAL BASIS OF PKU

i(pbctiam

S

kC2

Mr(tdltsm(im

z {¥ (phave m

1

ng both kinetic and stability properties of the enzyme;ii) structurally stable mutations with altered kineticroperties; and (iii) mutations with normal kinetics,ut reduced stability in vitro and in vivo (14). Futureontinuation of these expression analyses, in parallelo the known three-dimensional structure of PheOH,s important for a detailed characterization of the cat-lytic and regulatory properties of the disease-causingutants in the PAH gene.

ingle-Point Mutations

There are currently 234 missense point mutations

TABLE

Note. E; exon, I; intron; X, nonsense mutation, insekidney cells; NS, not stated; RD, regulatory domain;eukaryotic cells (COS cells); Non-PKU HPA definitiClassical PKU, blood plasma L-phe levels above 1000classical PKU nor non-PKU HPA; TC, reference not pHydroxylase Mutation Analysis Consortium Databas

% Predicted frequency of mutation (FOM) based on& Calculated frequency of mutation (FOM) based o

496 patients currently phenotyped. Data taken from PDatabase (1998) information on PKU genotype versu

Calculated FOM: a 5 {No. of times a mutation occu100%.

Calculated FOM (normalized to predicted FOM) 5 aaaver 5 averaged calculated FOM for residues that

12 ERLANDSEN

nown for the PAH gene (PAH Mutation Analysisonsortium Newsletter, Spring 1998), in addition to2 nonsense mutations and 11 silent mutations.

he

ost of the single-point mutations map onto theegion consisting of exon 5 (residue 148) to exon 12residue 438). Forty-nine of the total number of mu-ations are located in the regulatory domain (resi-ues 1–142), 209 mutations are located in the cata-ytic domain (residues 143–410), and 10 are locatedo the tetramerization domain (residues 411–452). Aummary of the known genotype/phenotype infor-ation about these mutations is shown in Table 1

illustrated in Fig. 3), together with a structuralnterpretation of the predicted effect of the selected

issense mutations on the enzyme function.

ntinued

mature stop codon; ND, not detectable; Hkc, humanatalytic domain; TD, tetramerization domain; Euk.,od plasma L-phe levels between 120 and 1000 mM;Variant PKU, blood plasma L-phe levels fits neithered in any journals. Submitted to the Phenylalanine8).rogram MUTPRED (15)—weighted data.l amount of known mutation and their frequency inalanine Hydroxylase Mutation Analysis Consortiumotype.ne allele/No. patients for which data exist (5 495)} z

redicted FOM (weighted))/¥ (aaver (for each residue)}.ore than one mutations known.

STEVENS

1—Co

rts preCD, c

on, blomM;ublishe (199the p

n totahenyls phenrs on o

AND

In an attempt to relate the variable prevalence ofuman genetic disease (PKU) to the mutability inher-nt in the nucleotide sequence of the PAH gene, a

mprPt(LtpYWRQstwmti(Ccfomgam

cadtqimi

A

icctsrtp

mTtavstaattwtdtdBoptfilpqLtsmstCweeoasteatmk

mtGsa

RAL B

utability profile of the PAH cDNA sequence waserformed using the program MUTPRED (15). Theesults are presented in Table 1 for the residues inheOH that are found to result in single-point muta-ions in the PKU database. The silent mutationsG10G, S137S, C203C, Q232Q, V245V, Q304Q, T323T,385L, K398K, V399V and Y414Y) are not shown inhe table since these do not introduce changes in therotein structure. The nonsense mutations (Q20X,77X, R111X, W120X, Y166X, Q172X, R176X,187X, Y204X, Y206X, Y216X, Q232X, R243X,261X, G272X, S295X, W326X, Q336X, K341X,355X, Y356X, S359X) are not shown in the table

ince these mutations lead to a premature step inranslation. The MUTPRED program generates un-eighted or weighted profiles for which the relativeutabilities of CpG dinucleotides are multiplied by

he likelihood of clinical detection depending on chem-cal difference in the translated amino acid residuesFig. 3B). It is based on a higher mutation frequency inpG dinucleotides than in other dinucleotides, thus

ausing human disease. To compare this theoreticalrequency of mutation to the actual detected frequencyf mutation in the PAH gene, an analysis of the knownutations in the PKU database was done for the sin-

le-point mutations known (Fig. 3B). These data arelso presented in Table 1 as calculated frequency ofutations in the PAH gene.

STRUCTURAL INTERPRETATION OFPREVALENT MUTATIONS

Structurally, the mutations causing PKU/HPAan be shown to affect residues in five categories: (a)ctive-site residues, (b) structural residues, (c) inter-omain interactions in a monomer, (d) residues in-eracting with the N-terminal autoregulatory se-uence, and (e) residues in the dimer or tetramernterfaces. Thus, the most prevalent single-point

utations in the PAH gene will be discussed accord-ng to these five groups.

ctive-Site Mutations

Thirty-one missense point mutations can be foundn the residues lining the active site. Seven of thesean be found in the putative pterin-binding motif,onsisting of residues 264–290 (16). Five of the ac-ive-site mutations have been expressed in vitro,ome even in multiple in vitro systems. The first

THE STRUCTU

esidue in the catalytic domain (Asp143) is reportedo be a PKU mutant associated with the severe PKUhenotype (17) in patients heterozygous for the ter-

gtc

ination mutant G272X and the D143G mutation.his residue is located to the surface at the entranceo the active site and may be involved in controllingccess to the active site for substrate/cofactor. Initro expression of the mutant in three differentystems (Escherichia coli, human kidney cells, and aranscription-translation system (TnT)) resulted inrelatively high residual enzymatic activity (52, 33,nd 102%, respectively) as compared to the wild-ype PheOH. Also, when assayed at lower concen-rations of L-Phe and BH4, the residual activitiesere compatible with severely reduced hydroxyla-

ion of L-Phe, thus suggesting that this residue in-eed plays a role in controlling substrate access tohe active site. Another one of the active-site resi-ues which may be involved in the binding of theH4 cofactor in PheOH; Phe254 is found to p-stackn the pterin ring in the structure of TyrOH com-lexed with 7,8-dihydrobiopterin (12). In the struc-ure of PheOH, this residue is located 5.9 Å awayrom the active-site iron, and a substitution into asoleucine as in the PKU mutant F254I would mostikely interfere with proper binding of pterin andossibly also the phenylalanine substrate. A fre-uent mutation among black Americans is the255S mutation, which causes severe PKU (18). Inhis mutation a nonpolar group is substituted with amaller polar side chain in a hydrophobic environ-ent and this is expected to result in significant

tructural perturbation. Leu255 seems to be impor-ant in controlling the separation of helix Ca6 anda9. Expression of the two mutations associatedith this residue, L255S and L255V, in three differ-nt in vitro expression systems, showed that, asxpected, the L255S mutation has a far larger effectn the residual enzyme activity (only 1% residualctivity detected in the TnT expression system ver-us wild-type) than the more conserved L255V mu-ation, which revealed 11% residual activity (TnTxpression system), 44% residual activity (E. coli),nd 13% residual activity (COS cells), versus wild-ype PheOH (18). As can be expected, based on theore conserved mutation, the L255V mutation is

nown to cause mild PKU.The active-site residue Thr278 has three differentutations associated with it. Thr278 is located at

he entrance to the active site and hydrogen bonds tolu280 (Fig. 2). The mutations T278I and T278A are

ubstitutions of a polar amino acid into hydrophobicmino acids and will result in this important hydro-

113ASIS OF PKU

en bond being lost and also perturb the structure athe active-site entrance. The more electrostaticallyonserved mutation T278N would putatively also

1

14 ERLANDSEN AND STEVENS

rtt

tfa(

hstPo

swAiSmmmM green( thanw being

esult in a change of hydrogen-bonding pattern, andhus have the same perturbing effect as that of thewo other mutations, if only slightly smaller.

Another well-studied mutation located to the ac-ive site is the E280K mutation. This residue is also

FIG. 3. (A) Scheme showing the secondary structure assigecondary structure was assigned using DSSP (43) on the compoith them are marked with red dots. (B) Predicted mutability profilso shown for comparison (in red circles) is the calculated freque

n 496 alleles causing HPA and PKU (genotype/phenotype data copring 1998, and Ref. 44)). The data are listed in Table 1 for allonomer of the full-length model of phenylalanine hydroxylase.utations, and the active-site iron is shown for reference as a grutations associated with it. The regions colored blue have resUTPRED) (Table 1) higher than 10, and the regions colored

calculated from data compiled by E. Kayaalp and Ref. 44) higherith both the calculated and the predicted frequency of mutation

THE STRUCTU

ound within the so-called pterin-binding motif (16),nd the mutation itself causes mild to severe PKU19). In the PheOH structures, Glu280 is seen to

mam

ydrogen-bond to His146 and to form an importantalt bridge to Arg158 (Fig. 2). Also, there are onlywo free, charged groups in the active-site crevice ofheOH (both glutamic acids) and the substitution ofne of them (Glu280) by a lysine represents a dra-

of the human PheOH sequence (SWISS-PROT P00439). TheeOH model. The residues that have PKU mutations associated

the human PAH gene calculated by the program MUTPRED (15).mutation based on the known frequency of mutations occurringby E. Kayaalp (PAH Mutation Analysis Consortium Newsletter,single-point mutations causing HPA/PKU. (C) Ca-trace of one

ace is colored yellow for residues not associated with any PKUhere. The region colored red consists of residues that have PKUthat show a predicted frequency of mutation (as calculated by

have residues that show a calculated frequency of mutation10 (Table 1). The four residues colored magenta are the residues

high (Arg158, Arg252, Arg261, and Arg408).

115ASIS OF PKU

nmentsite Phle forncy of

mpiledknownThe treen spidues

RAL B

atic change in the electrostatic potential in thective site. The expression of this mutation in hu-an PheOH in E. coli resulted in an enzyme with

o(a(

enlcilasoPrt

PolrabFtomootaa

hbmtswmC,aP(edBgo

S

afbpMrTspPtTgecqtnpaldasswtftPbsdvgrfe

asamftia

1 AND

nly ;1% of the specific activity of wild-type PheOH20) as assayed with both the natural cofactor BH4

nd the synthetic cofactor 6-methyltetrahydropterin6-MPH4).

One of the most frequent mutations in southeast-rn Europe is the P281L mutation (21), causing phe-otypes ranging from HPA to severe PKU. This pro-

ine helps to define the shape of the active site verylose to the iron in the PheOH structure (Fig. 2). Its therefore almost certain that a substitution to aess rigid leucine will change the conformation of thective site by removing the conformational con-traints imposed by the proline. In vitro expressionf this mutant in E. coli resulted in complete loss ofheOH activity (22), and expression in COS cellsesulted in ,1% enzyme activity (23), as comparedo wild-type PheOH.

Three residues that are in the active site arehe331, Ala345, and Gly346. Phe331 p-stacksnto Trp326, and both Ala345 and Gly346 areocated right behind the triad of residues that areesponsible for binding the iron (His285, His290,nd Glu330) (Fig. 2). Two PKU mutations haveeen reported in the Phe331 residue: F331C and331L (24,25). Both mutations will abolish impor-

ant p-stacking interactions that stabilize the wallf the active site. Ala345 has two reported PKUutations, A345S and A345T, and Gly346 has

ne, G346R. All three mutations increase the sizef the side chain which would putatively result inhe iron ligands being shifted further out into thective-site crevice, and thus destabilize the entirective-site structure.One more residue which seems very important in

olding the iron in place is Ser349, which hydrogen-onds to His285, one of the iron ligands. Two PKUutations have been reported for this residue, one of

hem resulting in classical PKU (S349P) (26). Thisubstitution of a proline so close to the catalytic ironill change the entire shape of the active site. Theutation has been expressed in both E. coli andOS cells and has been found to have only ,0.2 to1% residual activity and no detectable levels ofctivity, respectively, as compared to the wild-typeheOH. The second mutation in this residue

S349L) (27) is also reported to give similar in vitroxpression in E. coli and COS cells of ,0.2% or noetectable levels of PheOH activity, respectively.

16 ERLANDSEN

oth mutations disrupt the very important hydro-en bond to His285, and thus have severe influencesn the catalytic activity of PheOH.

awv

tructural Residues

The G46S mutation in the regulatory domain isssociated with classical PKU and is one of the mostrequent mutations found in the PKU database (Ta-le 1). The cDNA site for residue 46 has also beenredicted to be a site of mutation by the programUTPRED. Gly46 is located on the surface of the

egulatory domain in a loop just before helix Ra1.he substitution with a serine seems to involve theerine side-chain hydrogen bonding to one or moreutative residues in close proximity to Gly46 in theheOH structure, and thus results in distortions ofhe secondary structure in the regulatory domain.his could result in the formation of inactive aggre-ates as reportedly found by Eiken et al. (28) uponxpressing the mutant in three different systems (E.oli, human kidney cells, and TnT). Another fre-uent mutation found in the regulatory domain ishe I65T mutation. This mutation is associated withon-PKU HPA to variant PKU and has been ex-ressed in COS cells to find a residual activity ofpproximately 26% of wild-type PheOH (29). Ile65 isocated in the hydrophobic core of the regulatoryomain. A substitution with a polar threonine (orsparagine as in the I65N mutant) would result inignificant structural perturbation. A more con-erved substitution, like in the I65A substitution,ould probably not cause as large a perturbation as

he two others. Another mutation of relatively highrequency in patients with PKU is the A104D mu-ation. This mutation is associated with variantKU, and the Ala104 residue is located in the loopetween Ra2 and Rb4 in the regulatory domain. Aubstitution by a larger and charged residue mayestabilize the loop structure. When expressed initro in three different systems, the increased aggre-ation of the resulting enzyme was observed, andesidual activity as compared to the wild type variedrom 26% (human kidney cells) to no detectable lev-ls (E. coli) (30).Arg158 is the site of predicted frequent mutation

s determined by MUTPRED based on the cDNAequence of PheOH (15). The R158Q mutation isssociated with classical PKU and is also a frequentutation in patients with PKU (Table 1). Arg158

orms a salt bridge to Glu280 and one hydrogen bondo Tyr268 (Fig. 4A). Both of these interactions aremportant for conserving the shape of the active site,nd substitution with a glutamine, or the larger

STEVENS

romatic residue, tryptophan (R158W mutation),ill necessarily distort the active-site structure. Initro expression of the R158Q mutant was found to

rP

wtat(itidTasrC((rcAoaPRmitrte(sso

pwdasiaemrt

dPtf

PabsteeC3de

pnmecaactl

I

biACbeahMsbste

Rsppcabmt

RAL B

esult in approximately 10% of normal levels ofheOH activity (31).Cys203 is found to be involved in a disulfide bondith Cys334 in the structure of rat PheOH including

he regulatory domain (11). This is not observed inny of the structures of human PheOH (9,10), buthat may be a product of the expression systeminsect cells versus E. coli). Both cysteines involvedn the disulfide bond are associated with PKU mu-ations and are predicted to be sites of disease-caus-ng mutations (by MUTPRED) (15). Another pre-icted site of disease-causing mutation is in Arg243.his residue is located at the end of Cb1 and formssalt bridge to Asp129 in Ca1 helix. In vitro expres-

ion of the two PKU mutants associated with thisesidue (R243Q and a nonsense mutant R243X) inOS cells resulted in ,10% residual activity

R243Q) (32) and ,1% residual activity (R243X)19). Expression of the R243X mutant in E. coliesulted in activity levels similar to those in COSells (,0.2%) (20). Arg252 forms a salt bridge tosp315 and hydrogen-bonds to the carbonyl oxygenf Ala313 as well as the side chain of Asp27 in theutoregulatory sequence (Fig. 4B). There are threeKU mutations associated with this residue:252Q, R252G, and R252W. The latter of the threeutations is reported to result in classical PKU, and

n vitro expression of the R252G and R252W muta-ions in E. coli, TnT, and COS cells results in ,1%esidual activity as compared to that in the wild-ype PheOH (18,20,31). The R252Q mutation, how-ver, results in somewhat larger residual activity16 and 3%) when expressed in E. coli and TnTystems (18). Thus it is quite apparent that anyubstitution in this residue will result in disruptionf stabilizing interactions in the catalytic domain.The residue Ala259 is buried in a hydrophobic

ocket about 4 Å away from Leu311 and Leu308,hich are hydrogen-bonded to Arg408. These resi-ues are important for holding the tetramerizationrm close to the catalytic domain, and thus a sub-titution of Ala259 into a larger or polar residue, asn the PKU mutations A259T and A259V, cannot beccommodated without disturbing the surroundingnvironment (including Leu311 and Leu308). Bothutations have been expressed in vitro in E. coli and

esult in 0.1–0.8% residual activity as compared tohat in wild-type PheOH (18,20).

One of the most frequent mutations based on theata in the PKU database (also predicted by MUT-

THE STRUCTU

RED) (15) is the R261Q PKU mutation. This mu-ation is reported to result in phenotypes varyingrom variant PKU to the more severe form, classical

p

t

KU. Arg261 is located in a loop between helix Ca6nd Cb2 in the catalytic domain, where it hydrogen-onds to Gln304 and Thr238 (Fig. 4C). This helps totabilize the structure of the active site, and substi-uting the arginine to a proline (R261P) would nec-ssarily destabilize the active-site structure. In vitroxpression of the R261Q mutant in E. coli (20) andOS cells (19) results in comparable levels (32 and0%, respectively) of enzymatic activity. Thus, theestabilization of the active-site structure is notnough to completely abolish PheOH activity.One interesting mutant within the putative

terin-binding region in the catalytic domain is theonsense mutant G272X. Expression analysis of thisutant results in no detectable PheOH activity in

ither E. coli or COS cells (17,18,33). Glu272 is lo-ated in a loop between Cb2 and Ca7 just before thective-site histidines that bind the catalytic iron,nd the mutation would putatively result in a trun-ated form of PheOH that has none of the residueshat are responsible for binding iron, thus no cata-ytic activity should be observed.

nterdomain Interactions in a Monomer

The most important residues at the interfaceetween the catalytic domain and the tetramer-zation domain are residues Leu311, Leu308, andrg408. Arg408, which is in the loop betweena12 and Tb1, is buried and forms hydrogenonds to the carbonyl oxygens of Leu308 (at thend of Ca8) and Leu311 (in the loop between Ca8nd Ca9) (Fig. 5A). This residue is predicted toave a high frequency of mutation as predicted byUTPRED (15) (Table 1), but is also the site of the

ingle most frequent mutation in the PKU data-ase (R408W mutation). The mutation results in aevere metabolic PKU phenotype and is reportedo give low (,1 to ,2.7%) residual activity whenxpressed in COS cells (34,35).A second frequent mutation in this residue is the408Q mutation, which is associated with a lessevere metabolic non-PKU HPA phenotype and re-ortedly gives 55% residual PheOH activity, as com-ared to that of the wild type, if expressed in COSells (35). The substitution of Arg408 into the largernd bulkier tryptophan would alter the hydrogen-onding network that holds the tetramerization do-ain together with the catalytic domain, and

hrough steric hindrance interfere with the correct

117ASIS OF PKU

ositioning of the b-ribbon (Tb1 and Tb2).The mutation Y414C has been found to be a rela-

ively frequent mutation and result in non-PKU

Hi2tlto

R

taidsaambsshmawf

R

ciTaidmnpntmmsGbttatad

tions associated with them, including their hydrogen-bondingpartners. The active-site iron and its ligands are also shown forreference (A) Arg158, (B) Arg252, and (C) Arg261.

1 AND

PA to variant PKU phenotypes. In vitro expressionn COS cells results in an enzyme which retains8–50% of normal activity (23). The mild effect ofhis mutation can be observed also at a structuralevel; a substitution from a tyrosine to a cysteine inhis position should not have drastic consequencesn the interaction with the neighboring residues.

esidues Interacting with the N-TerminalAutoregulatory Sequence

One of the few residues in the catalytic domainhat seems to be in contact with the N-terminalutoregulatory sequence is Tyr377 (11). This residues found to hydrogen-bond to Ser23 in the regulatoryomain and also stacks on top of Trp326 (Fig. 5B). Ithould be important for regulating access to thective site and therefore also regulates the substratend cofactor access to the catalytic iron. In the PKUutant Y377C, the important hydrogen bond would

e disrupted, thus exposing the active site and de-troying the effects of the regulatory domain uponubstrate binding. As mentioned previously, Asp27ydrogen-bonds to Ne of Arg252, and in the PKUutations associated with Arg252 (R252Q, R252G,

nd R252W) the disruption of this hydrogen bondill most likely also change access to the active site

or substrate/cofactor.

esidues at the Dimer or Tetramer Interfaces

Several residues that have PKU mutations asso-iated with them can be found at the dimer interfacen all structures of PheOH solved so far (9–11).hese are residues Ser67, Glu76, Arg297, Gln304,nd Glu422. Ser67 is located in b-strand Rb2, andts main chain amide hydrogen-bonds to the hy-roxyl group of Tyr216 in the other monomers thatake up the dimer. The PKU mutation S67P would

ecessarily distort the structure in this region, im-osing more rigid torsion angles than those of theative serine (36). Glu76 is located on the surface ofhe regulatory domain, in b-strand Rb3, where theain-chain carbonyl oxygen of Glu76 in one mono-er is hydrogen-bonded to Ne2 of His208 in the

econd monomer, as well as Asn73, just prior tolu76 in sequence (Fig. 6A). This contact seems toe important for forming the dimer that makes uphe “dimer of dimers” seen in the tetrameric struc-ure of PheOH (10). Glu76 has two PKU mutationsssociated with it, E76A and E76G (Table 1). Both of

STEVENS

FIG. 4. Various active-site residues that have PKU muta-

18 ERLANDSEN

hese mutations are substitutions of a chargedmino acid into surface-exposed hydrophobic resi-ues, and this could result in perturbations of the

sAMTm

sm

P the catP l regulT own ins

econdary structure in the regulatory domain.rg297 is located in a-helix Ca8 and is predicted (by

FIG. 5. (A) Arg408 is the site of the most frequent PKU mutatheOH monomer. This residue seems to be important for holdingKU-mutated residue Tyr377 is associated with the N-terminarp326. The N-terminal regulatory domain Ca-atom trace is shhown in gray ribbon.

THE STRUCTU

UTPRED) (15) to be a site of frequent mutation.he side chain of Arg297 is hydrogen-bonded to theain-chain carbonyl oxygen of Arg71 as well as the

lRd

ide-chain oxygen of Glu422 in the other dimericonomer (Fig. 6B). There are two PKU mutations

408W). It is hydrogen-bonded to Leu311 and Leu308 in the samealytic domain together with the tetramerization domain. (B) Theatory sequence. It hydrogen-bonds to Ser23 and p-stacks ontoyellow ribbon, whereas the catalytic domain Ca-atom trace is

119ASIS OF PKU

ion (R

RAL B

isted for Arg297 in the PKU database, R297H and297C (28). Both substitutions will result in theisruption of the above-mentioned dimer-stabilizing

hfidit(m(pbe

tdtPtd

D

aTiPctopap

ewcrfimiap

dAsn73 and His208 (of monomer B). The Ca-trace of the catalyticdomain and the regulatory domain in one monomer are shown ingray and yellow, respectively, whereas the other monomer in the

dhTpCgythT

1 AND

ydrogen bonds and would possibly result in theormation of unstable PheOH enzyme. The last res-due associated with PKU mutations and involved inimer/tetramer interactions is Gln304. This residues located in a-helix Ca8 and forms hydrogen bondso the hydroxyl group of Tyr414 (in monomer B)Fig. 6C), as well as to Arg261. Gln304 has two PKU

utations associated with it, one silent mutationQ304Q) (37) and Q304R. The substitution of theolar glutamine to a charged arginine could disruptoth hydrogen bonds and destabilize the PheOHnzyme formed.Lastly, both Pro362 and Pro366 seem to be impor-

ant for positioning Glu368 correctly in order to me-iate dimer contacts between the same Glu368 inwo different subunits. Thus the PKU mutations362T (38) and P366H (39) will both destabilizehese parts of the structure and prevent properimers from being formed.

eletion and Insertion Mutants

Most of the non-single-point mutations that mayffect the structure of PheOH are listed in Table 2.hese include deletion, insertion, and one prevalent

ntron mutation which affect the structure ofheOH. The frameshift mutations have been ex-luded from this review paper since the effect ofhese mutations is very difficult to interpret basedn the structures of PheOH. Most of these mutationsutatively lead to truncated versions of hPheOHnd no detectable level of PheOH enzyme is ex-ressed.One of the two deletion mutants that have been

xpressed in vitro, the 3-bp in-frame deletion of I94,as expressed using human colon adenocarcinoma

ells (SW613-12A1) (40). This mutant showed 27%esidual activity and markedly reduced binding af-nity for L-Phe substrate. Ile94 is located in theiddle of a-helix Ra2. A deletion of this hydrophobic

soleucine would disturb the helix packing by givingone residue shift in the hydrophobic versus hydro-hilic residues of helix Ra2.

imer is shown in orange to ease comparisons. (B) Arg297 with itsydrogen-bonding partners Glu422 and Arg71 (of monomer B).he presence of a twofold symmetry axis perpendicular to thelane of the figure is shown in the middle of the picture. Botha-traces of the catalytic domains in each monomer are shown inray, and the Ca-trace of the regulatory domain is shown inellow. The catalytic 2-His-1-Glu triad of coordinating residues to

STEVENS

FIG. 6. Various PKU-associated residues at the tetramer/imer interfaces. (A) Glu76 with its hydrogen-bonding partners

20 ERLANDSEN

he iron is shown for structural reference. (C) Gln304 with itsydrogen-bonding partners Arg261 and Tyr414 (of monomer B).he same color scheme as in (A) is employed.

L(rwpmem

oapvtttit

aiPmPtiItsdcta

d

(

G(

d

G

d(

((

(

(

(

(

(and C

d tion inncated

The second PKU deletion mutation is delL364.eu364 is positioned between two proline residues

Pro362 and Pro366) and may be a necessary spaceresidue for the prolines. Deletion of this residueould result in improper positioning of the rigidrolines and thus perturb the entire catalytic do-ain fold, corresponding to no detectable levels of

nzymatic activity as found upon expression of theutant in COS cells (33).The splicing mutation IVS12 1 1g3 a in intron 12

f the PAH gene is the most prevalent PKU allelemong Caucasians (41). The clinical and metabolichenotype of homozygotes with the mutation is a se-ere (“classical”) form of PKU. The mutation leads tohe synthesis of a truncated form of PheOH lackinghe last 52 amino acids in the C-terminal tetrameriza-

TADeletion and Insertion Mutati

cDNA mutation PheOH mutation

elE1/E2 delE1/E2 Deletpro

115–117)del3bp delF39 In cenDelcore

AAdel E43/E44del Delet131–133)del3bp K42–V45delAAG In loo

pacel3bpTTT delF55 At cen

at Rg/Aa delE2/IVS2nt1 Delet

proelE3 delE3 Delet208–210)del3bp delS70 At en

pac280–282)del3bp delI94 In mi442–706)del 442–706delE5/E6 Delet

ver895–897)del3bp delF299 In a h

Del967–969)delACA delT323 In Ca

Del1066 3 1071)ins6bp Y356/C357ins6bp At su

smareg

1090 3 1092)del3bp delL364 No sepropos

1090 3 1104) delLLPLEdel15bp DeletTD

el15bp 1315nt1g 3 a IVS12 1 1g 3 a Mutatru

THE STRUCTU

ion domain. Expression studies in eukaryotic systemsndicate that the deletion abolishes PheOH activity inhe cell as a result of protein instability (42).

(rd

SUMMARY

The recently solved crystal structures of humannd rat PheOH have proven to be valuable in thenterpretation of many of the mutations that causeKU. However, the effects of some of the reportedutations cannot be predicted based on either theheOH structure or the PAH gene sequence andhus in vitro and in vivo expression analyses aremportant tools in determining the details of PKU.n order to develop the complete picture of the mu-ations that lead to PKU, a great deal of informationtill needs to be collected on both the frequency ofifferent PKU mutations around the world and aontinuing effort on the expression analyses of mu-ations in the PAH gene. The frequency of mutationss calculated by the occurrence of CpG dinucleotides

2dentified in the PheOH Gene

Structural contacts/comments Reference

residues 1–57 in RD. Some active PheOHay be expressed.

(87)

RD in cluster of aromatic residues.f Phe disturbs packing of RD hydrophobic

(39)

both Glu disrupts hydrogen bonds in RD. TCeen Rb1 and Ra1. Deletion disturbs RD (36)

RD. Deletion disturbs hydrophobic packinge.

(77)

residues 21 to 56 in RD. Some active PheOHay be expressed.

(88)

residues 57 to 118 of RD. (89)2. H bonds keep loop in place for Rb2 to

erly.TC

Ra2. Deletion disturbs helix formation. (40)residues 147 to 235 in CD. May result inable or little PheOH protein expressed.

TC

hobic area of CD towards end of Ca8.ill disturb CD fold.

(52)

onds to Glu319 carbonyl oxygen.isturbs CD fold.

(52)

f CD. Room for two extra residues with onlynges to secondary structure in this loop

(36)

y structure in region. Located between two62 and 366). Residue may be necessary to

rolines for correct fold of CD.

(33)

large portion of CD. Wrong CD fold affectsD.

TC

intron 12 of PheOH. Leads to expression ofPheOH missing last 52 residues.

(42)

121ASIS OF PKU

BLEons I

ion oftein mter of

etion o.

ion ofp betwking.ter ofD cor

ion oftein mion ofd of Rbk propddle ofion ofy unstydropetion w9. H betion drface oll cha

ion.condarlines (3ition pion of

RAL B

by the program MUTPRED) (15) and actual occur-ence of mutations in the cases reported at the PKUatabase do seem to match in most cases (Fig. 3B),

lAttfIPs

DdP1pnhaB2oNd

1

1

1

1

1

1

1

1

1

1

2

2

1 AND

ike for Arg158, Arg252, Arg261, Arg408, andsp415. These are all sites of CpG dinucleotides in

he PAH gene sequence and are thus prone to mu-ability (15). In other cases, residues that originaterom non-CpG dinucleotides, like Gly46, Leu48,le65, and Val388, still have a high frequency ofKU mutation and some are also associated with aevere PKU phenotype.

ACKNOWLEDGMENTS

The authors thank Dr. Charles R. Scriver and co-workers ateBelle Laboratory for Biochemical Genetics, Montreal Chil-ren’s Hospital, Montreal, Canada, for supplying us with theAH Mutation Analysis Consortium Newsletter for the Spring of998, which was used extensively in the preparation of this pa-er. Supplying information was also taken from the Phenylala-ine Hydroxylase Mutation Analysis Consortium Database atttp://www.mcgill.ca/pahdb. Atomic coordinates of the phenylal-nine hydroxylase structures are available at the Protein Dataank (PDB) at http://www.rcsb.org/pdb, using the id codes 1pah,pah, 1phz, and 2phm. Professor Torgeir Flatmark, Departmentf Biochemistry and Molecular Biology, University of Bergen,orway, and Albert E. Beuscher are thanked for valuable help oniscussions and preparation of this manuscript.

REFERENCES

1. Kaufman S. The phenylalanine hydroxylating system. AdvEnzymol 70:77–264, 1993.

2. Kappock TJ, Caradonna JP. Pterin-dependent amino acidhydroxylases. Chem Rev 96:2659–2756, 1996.

3. Hufton SE, Jennings IG, Cotton RGH. Structure and func-tion of the aromatic amino acid hydroxylases. Biochem J311:353–366, 1995.

4. Waters PJ, Parniak MA, Nowacki P, Scriver CR. In vitroexpression analysis of mutations in phenylalanine hydroxy-lase: Linking genotype to phenotype and structure to func-tion. Hum Mutat 11:4–17, 1998.

5. Scriver CR, Kaufman S, Eisensmith RC, Woo SLC. Thehyperphenylalaninemias. In The Metabolic and MolecularBases of Inherited Disease, 7th ed. (Scriver CR, Beaudet AL,Sly WS, Valle D, Eds.). New York: McGraw-Hill, pp 1015–1075, 1995.

6. Hoang L, Byck S, Prevost L, Scriver CR. PAH MutationAnalysis Consortium Database: A database for disease-pro-ducing and other allelic variations at the human PAH locus.Nucleic Acids Res 24:127–131, 1996.

7. Nowacki P, Byck S, Prevost L, Scriver CS. The PAH muta-tion analysis consortium database: Update 1996. NucleicAcids Res 25:139–142, 1997.

8. Nowacki P, Byck S, Prevost L, Scriver CR. PAH MutationAnalysis Consortium Database: 1997. Prototype for rela-tional locus-specific mutation databases. Nucleic Acids Res26:220–225, 1998.

22 ERLANDSEN

9. Erlandsen H, Fusetti F, Martinez A, Hough E, Flatmark T,Stevens RC. Crystal structure of the catalytic domain ofhuman phenylalanine hydroxylase reveals the structural

2

basis for phenylketonuria. Nature Struct Biol 4:995–1000,1997.

0. Fusetti F, Erlandsen H, Flatmark T, Stevens RC. Structureof tetrameric human phenylalanine hydroxylase and its im-plications for phenylketonuria. J Biol Chem 273:16962–16967, 1998.

1. Kobe B, Jennings IG, House CM, Michell BJ, Goodwill KE,Santarsiero BD, Stevens RC, Cotton RGH, Kemp BE. Struc-tural basis of autoregulation of phenylalanine hydroxylase.Nature Struct Biol 6:442–448, 1999.

2. Goodwill KE, Sabatier C, Stevens RC. Crystal structure oftyrosine hydroxylase with bound cofactor analogue and ironat 2.3A resolution: Self-hydroxylation of Phe300 and thepterin-binding site. Biochemistry 37:13437–13445, 1998.

3. Martinez A, Knappskog PM, Olafdottir S, Døskeland AP,Eiken HG, Svebak RM, Bozzini M, Apold J, Flatmark T.Expression of recombinant human phenylalanine hydroxy-lase as fusion protein in Escherichia coli circumvents pro-teolytic degradation by host cell proteases. Biochem J 306:589–597, 1995.

4. Flatmark T, Knappskog PM, E, Martinez A. Molecular char-acterization of disease related mutant forms of human phe-nylalanine hydroxylase and tyrosine hydroxylase. In Chem-istry and Biology of Pteridines and Folates (Pfleiderer W,Rokos H, Eds.). London: Blackwell Science, pp 503–508,1997.

5. Cooper DN, Krawczak M. The mutational spectrum of singlebase-pair substitutions causing human genetic disease; pat-terns and predictions. Hum Genet 85:55–74, 1990.

6. Jennings IG, Kemp BE, Cotton RGH. Localization of cofac-tor binding sites with monoclonal anti-idiotype antibodies:phenylalanine hydroxylase. Proc Natl Acad Sci USA 88:5734–5738, 1991.

7. Knappskog PM, Eiken HG, Martinez A, Bruland O, Apold J,Flatmark T. PKU mutation (D143G) associated with anapparent high residual enzyme activity: Expression of akinetic variant form of phenylalanine hydroxylase in threedifferent systems. Hum Mutat 8:236–246, 1996.

8. Bjørgo E, Knappskog PM, Martinez A, Stevens RC, Flat-mark T. Partial characterization and three-dimensional-structural localization of eight mutations in exon 7 of thehuman phenylalanine hydroxylase gene associated withphenylketonuria. Eur J Biochem 257:1–10, 1998.

9. Okano Y, Eisensmith RC, Guttler F, Lichter-Konecki U,Konecki DS, Trefz FK, Dasovich M, Wang T, Henriksen K,Lou H, et al. Molecular basis of phenotypic heterogeneity inphenylketonuria. N Engl J Med 324:1232–1238, 1991.

0. Knappskog PM, Eiken HG, Martinez A, Olafsdottir S, Haa-vik J, Flatmark T, Apold J. Expression of wild-type andmutant forms of human phenylalanine hydroxylase in E.coli. In Chemistry and Biology of Pteridines and Folates(Ayling JE, et al., Eds.). New York: Plenum Press, pp 59–62,1993.

1. Baric I, Mardesic D, Sarnavoka V, Lichter-Konecki U, Ko-necki DS, Trefz FK. Geographical distribution of the P281Lmutation at the phenylalanine hydroxylase locus: Possibleorigin in southeastern Europe. J Inherit Metab Dis 7:376–

STEVENS

377, 1994.

2. Dworniczak B, Grudda K, Stumper J, Bartholome K,Aulehla-Scholz C, Horst J. Phenylalanine hydroxylase gene:

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

4

4

4

4

4

4

4

4

5

RAL B

Novel missense mutation in exon 7 causing severe phenyl-ketonuria. Genomics 9:193–199, 1991.

3. Okano Y, Wang T, Eisensmith RC, Longhi R, Riva E, Gio-vannini M, Cerone R, Romano C, Woo SL. Phenylketonuriamissense mutations in the Mediterranean. Genomics 9:96–103, 1991.

4. Benit P, Rey F, Melle D, Munnich A, Rey J. Five novelmissense mutations of the phenylalanine hydroxylase genein phenylketonuria. Hum Mutat 4:229–231, 1994.

5. Guldberg P, Levy HL, Hanley WB, Koch R, Matalon R,Rouse BM, Trefz F, de la Cruz F, Henriksen KF, Guttler F.Phenylalanine hydroxylase gene mutations in the UnitedStates: Report from the Maternal PKU Collaborative Study.Am J Hum Genet 59:84–94, 1996.

6. Knappskog PM, Eiken HG, Martinez A, Flatmark T, ApoldJ. The PKU mutation S349P causes complete loss of cata-lytic activity in the recombinant phenylalanine hydroxylaseenzyme. Hum Genet 95:171–173, 1995.

7. De Lucca M, Perez B, Desviat LR, Ugarte M. Molecularbasis of phenylketonuria in Venezuela: Presence of twonovel null mutations. Hum Mutat 11:354–359, 1998.

8. Eiken HG, Knappskog PM, Boman H, Thune KS, Kaada G,Motzfeldt K, Apold J. Relative frequency, heterogeneity, andgeographic clustering of PKU mutations in Norway. Eur JHum Genet 4:205–213, 1996.

9. John SWM, Scriver CR, Laframboise R, Rozen R. In vitroand in vivo correlations for I65T and M1V mutations at thephenylalanine hydroxylase locus. Hum Mutat 1:147–153,1992.

0. Waters PJ, Parniak MA, Hewson AS, Scriver CR. Alter-ations in protein aggregation and degradation due to mildand severe missense mutations (A104D, R157N) in the hu-man phenylalanine hydroxylase gene (PAH). Hum Mutat12:344–354, 1998.

1. Okano Y, Wang T, Eisensmith RC, Steinmann B, Gitzel-mann R, Woo SL. Missense mutations associated with RFLPhaplotypes 1 and 4 of the human phenylalanine hydroxylasegene. Am J Hum Genet 46:18–25, 1990.

2. Wang T, Okano Y, Eisensmith RC, Lo WH, Huang SZ, ZengYT, Yuan LF, Liu SR, Woo SL. Missense mutations preva-lent in Orientals with phenylketonuria: Molecular charac-terization and clinical implications. Genomics 10:449–456,1991.

3. Svensson E, von Dobeln U, Eisensmith RC, Hagenfeldt L,Woo SL. Relation between genotype and phenotype in Swed-ish phenylketonuria and hyperphenylalaninemia patients.Eur J Pediatr 152:132–139, 1993.

4. DiLella AG, Marvit J, Brayton K, Woo SL. An amino-acidsubstitution involved in phenylketonuria is in linkage dis-equilibrium with DNA haplotype 2. Nature 327:333–336,1987.

5. Svensson E, Eisensmith RC, Dworniczak B, von Dobeln U,Hagenfeldt L, Horst J, Woo SL. Two missense mutationscausing mild hyperphenylalaninemia associated with DNAhaplotype 12. Hum Mutat 1:129–137, 1992.

6. Tyfield LA, Stephenson A, Cockburn F, Harvie A, Bidwell

THE STRUCTU

JL, Wood NA, Pilz DT, Harper P, Smith I. Sequence varia-tion at the phenylalanine hydroxylase gene in the BritishIsles. Am J Hum Genet 60:388–396, 1997. 5

7. Perez B, Desviat LR, De Lucca M, Ugarte M. Spectrum andorigin of phenylketonuria mutations in Spain. Acta PaediatrSuppl 407:34–36, 1994.

8. Mallolas J, Campistol J, Lambruschini N, Vilaseca MA,Cambra FJ, Estivill X, Mila M. Two novel mutations in exon11 of the PAH gene (V1163del TG and P362T) associatedwith classic phenylketonuria and mild phenylketonuria.Hum Mutat 11:482, 1998.

9. Guldberg P, Henriksen KF, Guttler F. Molecular analysis ofphenylketonuria in Denmark: 99% of the mutations de-tected by denaturing gradient gel electrophoresis. Genomics17:141–146, 1993.

0. Caillaud C, Lyonnet S, Rey F, Melle D, Frebourg T, Berthe-lon M, Vilarinho L, Vaz Osorio R, Rey J, Munnich A. A3-base pair in-frame deletion of the phenylalanine hydrox-ylase gene results in a kinetic variant of phenylketonuria.J Biol Chem 266:9351–9354, 1991.

1. DiLella AG, Marvit J, Lidsky AS, Guttler F, Woo SLC. Tightlinkage between a splicing mutation and a specific DNAhaplotype in phenylketonuria. Nature 322:799–803, 1986.

2. Marvit J, DiLella AG, Brayton K, Ledley FD, Robson KJ,Woo SL. GT to AT transition at a splice donor site causesskipping of the preceding exon in phenylketonuria. NucleicAcids Res 15:5613–5628, 1987.

3. Kabsch W, Sander C. Dictionary of protein secondary struc-ture:pattern recognition of hydrogen bonded and geometri-cal features. Biopolymers 22:2577–2637, 1983.

4. Kayaalp E, Treacy E, Waters PJ, Byck S, Nowacki P, ScriverCR. Human phenylalanine hydroxylase mutations and hy-perphenylalaninemia phenotypes: A metanalysis of geno-type-phenotype correlations. Am J Hum Genet 61:1309–1317, 1997.

5. Forrest SM, Dahl HH, Howells DW, Dianzani I, Cotton RG.Mutation detection in phenylketonuria by using chemicalcleavage of mismatch: Importance of using probes from bothnormal and patient samples. Am J Hum Genet 49:175–183,1991.

6. Carter KC, Rozen R, Byck S, Scriver CR. Novel mutationsand heterogeneity of the phenylalanine hydroxylase (PAH)gene on Quebec PKU chromosomes. Am J Hum Genet57(Suppl):A161 (Abstract), 915, 1995.

7. Guldberg P, Henriksen KF, Thony B, Blau N, Guttler F.Molecular heterogeneity of nonphenylketonuria hyperphe-nylalaninemia in 25 Danish patients. Genomics 21:453–455,1994.

8. Lichter-Konecki U, Schlotter M, Trefz FK, Konecki DS.Identification of the new mutations at the phenylalaninehydroxylase gene locus. Vth International Congress InbornErrors of Metabolism W4.4, 1990 (abstract).

9. Li J, Eisensmith RC, Wang T, Lo WH, Huang SZ, Zeng YT,Yuan LF, Liu SR, Woo SL. Phenylketonuria in China: Iden-tification and characterization of three novel nucleotide sub-stitutions in the human phenylalanine hydroxylase gene.Hum Mutat 3:312–314, 1994.

0. Bosco P, Cali F, Meli C, Mollica F, Zammarchi E, Cerone R,Vanni C, Palillo L, Greco D, Romano V. Eight new muta-tions of the phenylalanine hydroxylase gene in Italian pa-

123ASIS OF PKU

tients with hyperphenylalaninemia. Hum Mutat 11:240–243, 1998.

1. Horst J, Eigel A, Auleha-Scholz C, Kalaydjieva L, Zygulska

5

5

5

5

5

5

5

5

6

6

6

6

6

6

6

6

6

6

7

7

7

7

7

7

7

7

7

7

8

8

1 AND

M, Kunert E, Dworniczak B. Molecular basis of phenylketo-nuria: Report of an extensive study of various Caucasianpopulations. Am J Hum Genet 49(Suppl): A2302, 1991.

2. Eisensmith RC, Martinez D, Kuzmin A, Goltsov A, WooSLC. Molecular basis of phenylketonuria in a heterogeneousUS population (Abstract). Am J Hum Genet 57:A163 (926),1995.

3. Perez B, Desviat LR, Ugarte M. Analysis of the phenylala-nine hydroxylase gene in the Spanish population: Mutationprofile and association with intragenic polymorphic mark-ers. Am J Hum Genet 60:95–102, 1997.

4. Ramus SJ, Treacy EP, Cotton RG. Characterization of phe-nylalanine hydroxylase alleles in untreated phenylketonu-ria patients from Victoria, Australia: Origin of alleles andhaplotypes. Am J Hum Genet 56:1034–1041, 1995.

5. Li J, Eisensmith RC, Wang T, Lo WH, Huang SZ, Zeng YT,Yuan LF, Liu SR, Woo SL. Identification of three novelmissense PKU mutations among Chinese. Genomics13:894–895, 1992.

6. Takarada Y, Kalanin J, Yamashita K, Ohtsuka N, KagawaS, Matsuoka A. Phenylketonuria mutant alleles in differentpopulations: Missense mutation in exon 7 of phenylalaninehydroxylase gene. Clin Chem 39:2354–2355, 1993.

7. Shirahase W, Shimada AM. Genetic study on Japanese clas-sical phenylketonuria (family analysis by PCR-SSCP anal-ysis). Acta Pediatr Japan 96:939–945, 1992.

8. Goltsov A, Kouzmine A, Eisensmith RC, Effat L, TempamyS, Rushdi S, Abdel-Meguid N, Woo SLC. Molecular basis ofphenylketonuria in Egypt (abstract). Am J Hum Genet 55:1288, 1994.

9. Kuzmin AI, Eisensmith RC, Goltsov AA, Sergeeva NA,Schwartz EI, Woo SL. Complete spectrum of PAH mutationsin Tataria: Presence of Slavic, Turkic and Scandinavianmutations. Eur J Hum Genet 3:246–255, 1995.

0. Zschocke J, Graham CA, Stewart FJ, Carson DJ, Nevin NC.Non-phenylketonuria hyperphenylalaninaemia in NorthernIreland: Frequent mutation allows screening and early di-agnosis. Hum Mutat 4:114–118, 1994.

1. van der Sijs-Bos CJ, Diepstraten CM, Juyn JA, Plaisier M,Giltay JC, van Spronsen FJ, Smit GP, Berger R, SmeitinkJA, Poll-The BT, Ploos van Amstel JK. Phenylketonuria inThe Netherlands: 93% of the mutations are detected bysingle-strand conformation analysis. Hum Hered 46:185–190, 1996.

2. Hashem N, Bosco P, Chiavetta V, Cali F, Ceratto N, RomanoV. Preliminary studies on the molecular basis of hyperphe-nylalaninemia in Egypt. Hum Genet 98:3–6, 1996.

3. Popescu A, Andrian T, Guttler F, Guldberg P. Genotype-phenotype correlation in 11 Romanian PKU families. J In-herit Metab Dis 17:374–375, 1994.

4. Dianzani I, Giannattasio S, de Sanctis L, Alliaudi C, Lat-tanzio P, Vici CD, Burlina A, Burroni M, Sebastio G,Carnevale F, et al. Characterization of phenylketonuria al-leles in the Italian population. Eur J Hum Genet 3:294–302,1995.

5. Dworniczak B, Kalaydjieva L, Pankoke S, Aulehla-Scholz C,Allen G, Horst J. Analysis of exon 7 of the human phenyl-

24 ERLANDSEN

alanine hydroxylase gene: A mutation hot spot? Hum Mutat1:138–146, 1992.

6. Zschocke J, Graham CA, Carson DJ, Nevin NC. Phenylke-8

tonuria mutation analysis in Northern Ireland: A rapidstepwise approach. Am J Hum Genet 57:1311–1317, 1995.

7. Okano Y, Hase Y, Lee DH, Takada G, Shigematsu Y, OuraT, Isshiki G. Molecular and population genetics of phenyl-ketonuria in Orientals: Correlation between phenotype andgenotype. J Inherit Metab Dis 17:156–159, 1994.

8. Perez B, Desviat LR, Ugarte M. Expression analysis of mu-tation P244L, which causes mild hyperphenylalaninemia.Hum Mutat 5:188–190, 1995.

9. Caillaud C, Vilarinho L, Vilarinho A, Rey F, Berthelon M,Santos R, Lyonnet S, Briard ML, Osorio RV, Rey J, et al.Linkage disequilibrium between phenylketonuria andRFLP haplotype 1 at the phenylalanine hydroxylase locus inPortugal. Hum Genet 89:69–72, 1992.

0. Shirahase W, Oya N, Shimada M. A new single base substi-tution in a Japanese phenylketonuria (PKU) patient. BrainDev 13:283–284, 1991.

1. Leandro P, Rivera I, Ribeiro V, Tavares de Almeida I, Lech-ner MC. Sequencing analysis of PAH genomic DNA reveals4 novel mutations affecting exons 7 and 11 in a PortugesePKU population. Abstr. SSIEM, Manchester, 1993.

2. Kleiman S, Li J, Schwartz G, Eisensmith RC, Woo SL,Shiloh Y. Inactivation of phenylalanine hydroxylase by amissense mutation, R270S, in a Palestinian kinship withphenylketonuria. Hum Mol Genet 2:605–606, 1993.

3. Melle D, Verelst P, Rey F, Berthelon M, Francois B, Mun-nich A, Lyonnet S. Two distinct mutations at a singleBamHI site in phenylketonuria. J Med Genet 28:38–40,1991.

4. Goebel-Schreiner B, Schreiner R. Identification of a newmissense mutation in Japanese phenylketonuric patients.J Inherit Metab Dis 16:950–956, 1993.

5. Labrune P, Melle D, Rey F, Berthelon M, Caillaud C, Rey J,Munnich A, Lyonnet S. Single-strand conformation poly-morphism for detection of mutations and base substitutionsin phenylketonuria. Am J Hum Genet 48:1115–1120, 1991.

6. Takahashi K, Masamune A, Kure S, Matsubara Y, Nari-sawa K. Ectopic transcription: An application to the analysisof splicing errors in phenylalanine hydroxylase mRNA. ActaPaediatr Suppl 407:45–46, 1994.

7. Zygulska M, Eigel A, Pietrzyk JJ, Horst J. Phenylketonuriain southern Poland: A new splice mutation in intron 9 at thePAH locus. Hum Mutat 4:297–299, 1994.

8. Eisensmith RC, Woo SL. Molecular basis of phenylketonuriaand related hyperphenylalaninemeias: Mutations and poly-morphisms in the human phenylalanine gene. Hum Mutat1:13–23, 1992.

9. Kalaydjieva L, Dworniczak B, Kremensky I, Koprivarova K,Radeva B, Milusheva R, Aulehla-Scholz C, Horst J. Heter-ogeneity of mutations in Bulgarian phenylketonuria haplo-type 1 and 4 alleles. Clin Genet 41:123–128, 1992.

0. Economou-Petersen E, Henriksen KF, Guldberg P, GuttlerF. Molecular basis for nonphenylketonuria hyperphenylala-ninemia. Genomics 14:1–5, 1992.

1. Rozen R, Mascisch A, Lambert M, Laframboise R, ScriverCR. Mutation profiles of phenylketonuria in Quebec popu-lations: Evidence of stratification and novel mutations. Am J

STEVENS

Hum Genet 55:321–326, 1994.

2. Lichter-Konecki U, Konecki DS, DiLella AG, Brayton K,Marvit J, Hahn TM, Trefz FK, Woo SL. Phenylalanine hy-

8

8

8

8

8

8

RAL B

droxylase deficiency caused by a single base substitution inan exon of the human phenylalanine hydroxylase gene. Bio-chemistry 27:2881–2885, 1988.

3. Weinstein M, Eisensmith RC, Abadie V, Avigad S, LyonnetS, Schwartz G, Munnich A, Woo SL, Shiloh Y. A missensemutation, S349P, completely inactivates phenylalanine hy-droxylase in north African Jews with phenylketonuria. HumGenet 90:645–649, 1993.

4. Desviat LR, Perez B, De Lucca M, Cornejo V, Schmidt B,Ugarte M. Evidence in Latin America of recurrence ofV388M, a phenylketonuria mutation with high in vitro re-sidual activity. Am J Hum Genet 57:337–342, 1995.

5. Desviat LR, Perez B, Ugarte M. Molecular basis of non-PKUhyperphenylalaninaemia in Spain: Prevalence of A403V, a

THE STRUCTU

mutation with high residual activity. J Inherit Metab Dis19:227–230, 1996.

6. Wang T, Okano Y, Eisensmith RC, Harvey ML, Wilson HYL,

8

Huang S-Z, Zeng Y-T, Yuan L-F, Furuyama J-I, Oura T,Sommer SS, Woo SLC. Founder effect of a prevalent phe-nylketonuria mutation in the Oriental population. Proc NatlAcad Sci USA 88:2146–2150, 1991.

7. Sullivan SE, Lidsky AS, Brayton K, DiLella AG, King M,Connor M, Cockburn F, Woo SLC. Phenylalanine hydroxy-lase deletion mutant from a patient with classical PKU.Am J Hum Genet 37(Suppl):A177, 1985.

8. Kleiman S, Bernstein J, Schwartz G, Eisensmith RC, WooSL, Shiloh Y. A defective splice site at the phenylalaninehydroxylase gene in phenylketonuria and benign hyperphe-nylalaninemia among Palestinian Arabs. Hum Mutat1:340–343, 1992.

125ASIS OF PKU

9. Avigad S, Cohen BE, Bauer S, Schwartz G, Frydman M,Woo SL, Niny Y, Shiloh Y. A single origin of phenylketonu-ria in Yemenite Jews. Nature 344:168–170, 1990.