314 POPESCU ET AL. HUMAN MUTATION 12:314�319 (1998)

© 1998 WILEY-LISS, INC.

RESEARCH ARTICLE

Mutation Spectrum and Phenylalanine HydroxylaseRFLP/VNTR Background in 44 RomanianPhenylketonuric AllelesTeodora Popescu,1* Michaela Blazkova, Libor Kozak,2 Gheorghe Jebeleanu,1 andAntonia Popescu3

1Department of Biochemistry, University of Medicine and Pharmacy, Cluj-Napoca, Romania2Department of Biochemical and Molecular Genetics, Research Institute of Child Health, Brno, Czech Republic3Department of Paediatrics, University of Medicine and Pharmacy, Cluj-Napoca, Romania

Communicated by Randy Eisensmith

The mutation spectrum and polymorphic haplotype background in 22 Romanian families havebeen analysed in this study using the restriction digestion of phenylalanine hydroxylase (PAH)regions specifically amplified or the DGGE/direct sequencing methods. Eleven PAH mutationsspecifically associated with six mutant haplotypes were detected. In spite of the relative heteroge-neity of the molecular defects in the PAH gene, three mutations covered almost 70% of all alleles:R408W, 47.72%, 21/44; K363fsdelG 13.63%, 6/44; and P225T 6.81%, 3/44. Among these, R408W,the most frequent mutation in our population, represented 50% of all the phenylketonuric (PKU)chromosomes. Splice mutation IVS12nt1g®a affected two PAH alleles (4.54%); the remainingseven mutations were rare, each having an effect on just one chromosome (1/44), resulting in arelative frequency of 2.27%. A high frequency was observed in our PKU samples for the relativelyuncommon mutations, K363fsdelG and P225T mutation, suggesting a possible founder effect atorigin. Within the investigated panel, these mutations, both very rare among other Caucasianswere exclusively linked to haplotype 5.8 and 1.7, respectively. These results provide a basis for thedevelopment of a routine molecular analysis of Romanian PKU families. Hum Mutat 12:314–319, 1998. © 1998 Wiley-Liss, Inc.

KEY WORDS: PAH; PKU; HPA; Romanian families

INTRODUCTION

Phenylketonuria (PKU, MIM# 261600), themost common inborn error of amino acid me-tabolism, is caused by a high variety of mutationsin the gene for phenylalanine hydroxylase (PAH)enzyme (E.C. 1.14.16.1). The high clinical andbiochemical heterogeneity observed in thisdisease is a consequence of more than 380 dif-ferent mutations (both expressed and unex-pressed) identified in the PAH gene. The fourknown clinical phenotypes—classical, moderate,mild, and non-PKU hyperphenylalaninemia(HPA)—are clearly defined, even though thereare also overlapping intermediary forms amongthese phenotypes. The mutation spectrum andPAH polymorphic background of the PKU alle-les differ in various populations and geographi-cal regions. In order to define more precisely theorigin, mechanism, and clinical consequences of

different PAH mutations in a certain population,it is necessary to characterise as much of thispopulation as possible.

The purpose of this study was to describe thedisease-causing mutations and the RFLP/VNTRassociations at the PAH locus in a panel of 22families from the Transylvanian part of Romaniaand to compare these results with the results ofother studies. Using this approach, a more effec-tive screening program, improved diagnosis, andearlier implementation of dietary therapy mightbe introduced.

Received 18 February 1998; accepted 18 June 1998.*Correspondence to: Teodora Popescu, Department of Bio-

chemistry, University of Medicine and Pharmacy, Cluj-Napoca,Romania.

Grant sponsor: Grantova agentura Ceske Republiky; Grantnumber: GACR 302/93/2535.

ROMANIAN PKU FAMILIES 315

PATIENTS

Twenty-six Romanian children with HPA (44unrelated mutant alleles) and their parents, whenavailable, were included in the study. None of thepatients belonged to a consanguineous family. Thepatients, including two pairs of univitelline twinsand two sibs, were equally distributed in sexes.They were of various origins: 18 of Romanian an-cestry, 5 of Hungarian, one of Gypsy, one of mixedHungarian-Gypsy, and one of mixed Ukrainian-Romanian ancestry. All patients were diagnosedand followed for diagnosis and treatment of PKUat the Centre of the third Paediatric Clinic, Cluj-Napoca, using the Guthrie bacterial inhibition as-say and high-performance liquid chromatography(HPLC) measurements of the plasma Phe and Tyr.The clinical phenotype of the patients was classi-fied according to the criteria described by Guttleret al. (1993). They were not tested for BH4 defi-ciency; therefore, a potential defect in biopterinmetabolism was not excluded. Eleven of the chil-dren were diagnosed at birth; the other 15 chil-dren, coming from counties in which PKUnewborn screening is not performed, were lateridentified at 9 months to 16 years of age.

Phenotype classification of our probands wasperformed, correlating the biochemical parameters(serum Phe pretreatment value, control Phe valuesduring the dietary treatment, loading test results)with the clinical outcome and the specific geno-type. The effect of mutations was determined whenthe predicted PAH residual activity was identified.

METHODS

Genomic DNA was isolated from peripheralblood leukocytes by standard methods (Lahiri andNumberger, 1990). More common mutationsknown to occur more frequently in our population(Popescu et al., 1994) were screened by means ofPCR and enzyme digestion, using natural or ampli-fication created restriction sites. The presence ofthe following mutations was tested: R408W, R158Q,R261Q, IVS12nt1g→a, L48S, G272X, R252W,I65T, and Y414C. The amplification conditions wereas described elsewhere (Kozak et al., 1995). Remain-ing unidentified PKU alleles underwent denatur-ing gradient gel electrophoresis (DGGE); samplesshowing altered electrophoretic mobility patternwere sequenced by direct sequencing.

DGGE

The first 12 exons of the PAH gene with theirintronic junctions in were amplified under condi-tions described previously (Kozak et al., 1997).

After generation of heteroduplexes, the amplifiedsamples were loaded onto 6% polyacrylamide/TAEgel with 0–9.5 M urea gradient. Electrophoresisran in a ProteanRII XiCell apparatus (Bio-Rad) at170 V, for 4.5 h, in TAE buffer (0.04 M Tris-ac-etate, 1 mM EDTA, pH 8.0) at a constant tem-perature of 57°C. Samples with an unusuallymigrating band pattern identified by DGGE wereanalysed by solid-phase sequencing.

Direct Sequencing

The dideoxy chain termination sequencingmethod of Sanger et al. (1977) was carried outusing the Sequenase version 2.0 DNA Sequenc-ing kit (U.S. Biochemicals) as recommended bythe manufacturer. The sequencing analysis wasperformed as described by Kozak et al. (1997).

RFLP/VNTR Haplotype Analysis

Five of the seven diallelic polymorphismspresent in the PAH gene were determined bydigestion of polymerase chain reaction (PCR)-am-plified DNA with specific restriction endonu-cleases, as previously described: BglII (Dworniczaket al., 1991a); PvuII(a) (Dworniczak et al.,1991b);PvuII(b); MspI (Wedemeyer et al., 1991); XmnI(Goltsov et al.,1992a). The remaining two diallelicsites, EcoRI and EcoRV, were examined by South-ern blot, using a 32P-labelled full-length humanPAH cDNA-phPAH247 (Kwok et al., 1985) as ahybridization probe (Kozak et al., 1993).

The 30-bp multiallelic minisatellite variablenumber of tandem repeats (VNTR) at the 3′ endof the PAH gene was analysed by PCR accordingto the method of Goltsov et al. (1992b).

Restriction fragment-length polymorphism(RFLP)/VNTR haplotypes were numbered accord-ing to Eisensmith and Woo (1992).

RESULTS AND DISCUSSION

An 18-year HPA screening in Transylvania,Romania, demonstrated a PKU incidence of1:8,000 in newborn and 1:110 in retarded children.The study of 26 Romanian patients with HPA be-longing to 22 families identified 11 different PKUmutations. Results are summarized in Table 1. Fivealleles remain unknown, resulting in 88.64% diag-nosis efficiency.

As has already been observed by Tyfield et al.(1997) and Guldberg et al. (1996a), three to sixdifferent mutations are generally responsible foralmost 65% of the mutant alleles. Among these,one to two mutations represent almost 50% of thetotal; the other 25–30% are present, in most cases,

316 POPESCU ET AL.

in just one or two alleles. This is in accordancewith our findings of mutation frequency (Table 1).

Using the methods described above, both mu-tant alleles were identified in 18 patients. Amongthese, 72.72% (16/22) are compound heterozygotesfor two different mutations, and 27.2% (6/22) arehomozygotes for R408W or K363fsdelG. In threepatients, only one mutant allele was identified; inone subject, no abnormal electrophoretic patternwas observed in the first 12 exons of the PAH genewith their splice junctions. There are two possible

explanations for this. The disease-causing muta-tion could be either in exon 13 of PAH gene ordeep in the intronic sequence; or, as might be thecase for the patient with no mutation detected, itcould be caused by BH4 HPA deficiency, especiallywhen the low pretreatment Phe value of 656 µmol/L, the normal DGGE aspect, and the very severeclinical phenotype are taken into account.

The mutations detected were predominantlydistributed in exons 6–12 (38/44, 86.36%). Themost frequently affected (50% of disease-causing

TABLE 1. Genotype, Biochemical, and Clinical Phenotype for 22 Romanian HPA Patients*

Mutation Phe PAH(systematic Mutations Silent pretreatment residualname) (trivial name) polymorphisms Haplotype (µm) Phenotype activity

c.1222C→T/ R408W/ /Q232Q, L385L 2.3/3.8 1920 Classical <1%/<1%c.1315+1g→a IVS12nt1g→a PKUc.1222C→T/ R408W/R408W 2.3/2.3 2520 Classical <1%/<1%c.1222C→T PKUc.1222C→T/ R408W/K363fsdelG 2.3/5.8 2838 Classical <1%/NDc.1089delG PKUc.1222C→T/ R408W/L48S /V245V, Q232Q, 2.3/4.3 1793 Classical <1%/NDc.143T→C IVS3nt-22t→c PKUc.842C→T/ND P281L/ND 1.8/2.3 1388 Classical <1%/ND

PKUc.1222C→T/ R408W/R408W 2.3/2.3 1917 Classical <1%/<1%c.1222C→T PKUc.1222C→T/ R408W/R413P /V245V, Q232Q 2.3/4.3 2014 Classical <1%/<3%c.1238G→C PKU1c.1222C→T/ R408W/K363fsdelG 2.3/5.8 1428 Moderate <1%/NDc.1089delG PKUc.1222C→T/ R408W/R408W 2.3/2.3 1672 Classical <1%/<1%c.1222C→T PKUc.1315+1g→a/ IVS12nt1g→a/ Q232Q, L385L/ 3.8/5.8 1789 Classical <1%/NDc.1089delG K363fsdelG PKUc.1222C→T/ R408W/R261Q 2.3/1.8 1079 Classical <1%/30%c.782G→A PKUc.1222C→T/ R408W/R408W 2.3/2.3 ND Classical <1%/<1%c.1222C→T PKUc.533A→G/ E178G/R252W /L385L 1.7/69.3 1515 Non-PKU ND/<1%c.754C→T HPAc.1222C→T/ R408W/R408W 2.3/2.3 1678 Classical <1%/<1%c.1222C→T PKUc.1222C→T/ R408W/P225T 2.3/1.7 1243 Classical <1%/NDc.673C→A PKUc.1089delG/ND K363fsdelG/ND 5.8/1.8 1600 Classical ND/ND

PKUC.1222C→T/ND R408W/ND /V245V, Q232Q, 2.3/4.3 1538 Classical <1%/ND

IVS3nt-22t→c PKUc.1222C→T/ R408W/P225T 2.3/1.7 ND Classical <1%/NDc.673C→A PKUND/ND ND/ND V245V, Q232Q, 4.3/4.3 656 Classical ND/ND

IVS3nt-22t→c/ PKUV245V, Q232Q,IVS3nt-22t→c

c.1222C→T/ R408W/E390G /V245V, Q232Q, 2.3/4.3 984 Non-PKU <1%/NDc.1169A→G IVS3nt-22t→c HPAc.1089delG/ K363fsdelG/K363fsdelG 5.8/5.8 3319 Classical ND/NDc.1089delG PKUc.1222C→T/ R408W/P225T 2.3/1.7 1515 Classical <1%/NDc.673C→A PKU

PKU, phenylketonuria.*Frequencies of single mutations: R408W 47.72%(21/44); K363fsdelG 13.63%(6/44); P225T 6.81%(3/44); IVS12nt1g→a4.54%(2/44); and L48S, P281L, R413P, R261Q, E178G, R252W, E390G 2.27%(1/44).

ROMANIAN PKU FAMILIES 317

mutations) was exon 12 (R408W, R413P). Exon7, which is known to be the most hypermutableexon of the PAH gene (Kleiman et al., 1992), hadthe most various mutations (R252W, P281L,R261Q). Five of the 11 mutations (R408W,R261Q, P281L, R413P, R252W) involved a CpGhotspot, suggesting recurrent mutation as a mecha-nism of origin. Another mechanism, explaining theoccurrence of mutations P225T and K363fsdelG,is the founder effect; these mutations have highfrequency and are linked with specific haplotypesin studied population.

The most frequent mutation responsible forPKU in the Romanian population is R408W, ac-counting for almost 50% of all mutant alleles.Eleven patients (68.75%) were heterozygous, andfive were homozygous for R408W. The mutationwas always associated with haplotype 2.3 (H2.3)alleles, according to our results. This finding con-firms the association of the mutation with H2.3PKU alleles in Central and Eastern Europe, whereit is considered to be of Balto-Slavic origin (Koneckiet al., 1991; Eisensmith et al., 1995). Even thoughthe Slavic origins are lacking in the Romanian PKUpopulation, frequent Slavic migrations probably foot-printed the genetic apparatus of Romanian ethnicalgroup (Comsa et al., 1987; Heitel et al., 1994; Horedtet al., 1982, 1986).

The second most frequent mutation was foundto be frameshift microdeletion K363fsdelG(Table1). This deletion causes premature termi-nation of the protein chain eliminating exons 13,12 and partly 11. These exons encode the cata-lytic center; thus, inactive enzyme is probablyproduced. One patient homozygous for this mu-tation and four compound heterozygotes, bear-ing null mutation on the other PKU chromosome,sustained the severe classical clinical phenotype,suggesting the severity of this mutation. Furtherin vitro expression studies of the mutant PAHenzyme are necessary to confirm this hypothesis.The deletion is flanked by short repeated se-quences (5′AGAAG, 3′AGAAG); thus, it can beassumed that DNA strands caused slipping ofpolymerase during DNA replication, thus en-abling mutation at this site (Jaruzelska et al.,1992; Kleiman et al., 1992).

The third most frequent mutation in our studywas P225T (Table 1). This mutation is found onlyin the Romanian (H1.7) and Czech (H1.9) PKUpopulation (Hoang et al., 1986; Kozak et al., 1997).Association of one mutation with different num-bers of VNTR is not unusual; it can be explainedby the high mutational rate of such polymorphisms

(Goltsov et al., 1992a,b). Two mutations specificfor very mild non-PKU HPA (E390G and E178G)were identified in our studied samples. In bothcases, null mutation (R408W and R252W, respec-tively), was present on the other PAH allele. Thesetwo non-PKU HPA mutations were also observedin other very mild HPAs combined with severemutations; however, these were different fromthose in our patients (IVS12nt1g→a and IVS10nt-11g→a, R408W, respectively). Our findings con-firm previous observations (Desviat et al., 1987;Economou-Petersen et al., 1992; Guldberg et al.,1994; Hoang et al., 1986) that the combination ofsevere and mild PKU mutation results in mildHPA, rather than classical PKU.

The distribution of PKU mutations in Roma-nian patients included in this study is very similarto that observed in other Caucasian populations(Slavic, Scandinavian, or Mediterranean). How-ever, mutation of Asian origin (R413P) is presentas well. It has been speculated (Eisensmith et al.,1995; Konecki et al., 1991) as to the means ofspreading different mutations in distinct ethnicgroups, e.g., direct or intermediate genetic passageby different population migrations in history. Thus,Mediterranean mutations in Romanian populationare connected with the Daco-Roman origin of thepopulation (Comsa et al., 1987; Heitel et al., 1994,Horedt et al., 1986), while Scandinavian andSlavic mutations arrived probably by the way ofmultiple Germanic and Slavic migrations (Heitelet al., 1994; Horedt et al., 1986, 1982).

The distribution of mutant haplotypes at the PAHlocus in 44 alleles investigated was as follows: H2.3(22/44, 50%), H5.8 (6/44, 13.6%), H4.3 (6/44,13.6%), H1.7 (4/44, 9.09%), H1.8 (3/44, 6.81%),H3.8 (2/44, 4.54%), and H69.3 (1/44, 2.27%).

H2.3 was predominant in our study, in agree-ment with other Central and Eastern Europeanpopulations. H2.3 was found to be associated withsevere R408W mutation (21/22, 95.45%) and withunknown mutations (1/22, 4.54%) in patients withclassical PKU phenotype. The next most frequentwas H1.7 or H1.8 (H1, 5/44, 15.90%), on whichfour different mutations (P225T, E178G, R261Q,P281L) and one unknown mutation were found.H5.8, not frequent in other populations, was inour study in strict linkage disequilibrium with theframe shift deletion K363fsdelG. H4, which is to-gether with H1 relatively common in both normaland mutant PAH alleles (Konecki et al., 1991) wasassociated with various mutations (L48S, R413P,E390G), as well as with three unknown mutations.This finding confirms the previously established

318 POPESCU ET AL.

finding that, unlike H2 and H3, H1 and H4 aregenerally associated with many different mutations.H69.3, which is quite unusual, was present in ourstudy with the R252W Gypsy PAH allele.

Four different polymorphisms (V245V, Q232Q,L385L, IVS3nt-22-t→c) in coding part of the PAHgene, specifically associated with mutated H4.3,H3.8, and H69.3 alleles, have been found in someof the patients studied. The identification of thesesilent mutations on new PAH alleles will predictwith high probability the presence of haplotypes4.3 or 3.8 and of mutations R413P, L48S,IVS12nt1g→a, and E390G.

Taking into account the high mutational het-erogeneity at the PAH locus and the specific mu-tation/haplotype associations in distinct ethnicalgroups, molecular analysis covering most popula-tions aids in elucidating the origin, spread, andmechanism of origin of these mutations.

Nineteen PKU patients had the classical phe-notype; only one was a moderate HPA form, andtwo had HPA non-PKU. The severe classical phe-notype requiring strict dietary regimen was pre-dominant in our cases (86.36%). The combinedstudy of mutations (genotype) and clinical andbiochemical phenotype (PAH residual activity,pretreatment and Phe value during treatment, Pheloading test when performed) demonstrated gen-erally a good genotype–phenotype correlation. Twoexceptions to the rule were found: a case of non-PKU HPA (E178G/R252W) with high pretreat-ment Phe level (1,515 µmol/L), but with 0.1 g/kgby weight loading test eliminated within 24 h, mini-mal neurological changes, and normal IQ, and acase of moderate HPA diagnosed at 16 years ofage because of the patient’s bad school results. Thispatient had a very high pretreatment Phe valueand a genotype (R408W/K363fsdelG) correspond-ing more to the classical phenotype.

Genetic analysis performed in our study of Ro-manian PKU patients is important from the pointof view of diagnosis confirmation, for prognosis andby those correct and earlier optimal dietary therapywill be implemented. Proper genetic counsellingmight then be introduced, with the possibility ofdetecting carriers in families at risk.

REFERENCES

Comsa M (1987) Slawen und Awaren auf rumanischen Boden,ihre Beziehungen zu der bodenstandigen romanischen undspateren fruhrumanischen Bevolkerung in Die VolkerSudosteuropas im 6. bis 8. Jahrhundert, SudosteuropaJahrbuch. Munchen-Berlin 17:219–230.

Desviat LR, Perez B, Ugarte M (1987) Molecular basis of non-PKU hyperphenylalaninemia in Spain: Prevalence of A403V,

a mutation with high residual activity. J Inher Metab Dis2:227–230.

Dworniczak B, Wedemeyer N, Horst J (1991a) PCR detection ofthe BglII RFLP at the human phenylalanine hydroxylase(PAH) locus. Nucleic Acids Res 19:1958.

Dworniczak B, Wedemeyer N, Eigel A, Horst J (1991b) PCR de-tection of the PvuII(Ea) RFLP at the human phenylalaninehydroxylase (PAH) locus. Nucleic Acids Res 19:1958.

Economou-Petersen E, Henriksen KF, Guldberg P, Guttler F(1992) Molecular basis for nonphenylketonuria hyper-phenylalaninemia. Genomics 14:1–5.

Eisensmith R, Woo SLC (1992) Updated listing of haplotypes atthe human phenylalanine hydroxylase (PAH) Locus. Am J.Hum Genet 51:1445–1448.

Eisensmith RC, Goltsov A, O’Neill C, Tyfield L, Schwarz E,Kuzmin A, Baranovskaya S, Tsukerman G, Treacy E, ScriverC, Guttler F, Guldberg P, Eiken H, Apold J, Svensson E,Naaaughten E, Cahalane S, Croke D, Cockburn F, Woo SLC(1995) Recurrence of the R408W mutation in the phenylala-nine hydroxylase locus in Europeans. Am J Hum Genet56:278–286.

Goltsov AA, Eisensmith RC, Woo SLC (1992a) Detection of theXmnI RFLP at the human PAH locus by PCR. Nucleic AcidsRes4:927.

Goltsov AA, Eisensmith RC, Konecki DS, Lichter-Konecki U,Woo SLC (1992b) Associations between mutations and aVNTR in the human phenylalanine hydroxylase gene. Am JHum Genet 51:627–636

Guldberg P, Henriksen KF, Thony B, Blau N, Guttler F (1994) Mo-lecular heterogeneity of nonphenylketonuria hyperphenyl-alaninemia in 25 Danish patients. Genomics 21:453–455.

Guldberg P, Levy H, Hanley WB, Koch R, Matalon R, Rouse B,Trefz F, de la Cruz F, Henriksen KF, Guttler F (1996a) Pheny-lalanine Hydroxylase Gene Mutations in the United States:Report from the Maternal PKU Collaborative Study. Am JHum Genet 59:84–94.

Guldberg P, Levy H, Henriksen KF, Guttler F (1996b) Three preva-lent mutations in a patient with phenylalanine hydroxylasedeficiency: implications for diagnosis and genetic counselling.J Med Genet 33:161–164.

Guttler F, Guldberg P, Henricksen KF, Mikkelsen I, Olsen B, LouH (1993) Molecular basis for the phenotypical diversity ofphenylketonuria and related hyperphenylalaninemias. J In-herit Metab Dis 16:602–604.

Heitel R (1994 )“Zur Frage des Eindringens der Fruhslawen inSiebenburgen” in Siebenburgen zur Zeit der Romer und derVolkerwanderungszeit. Koln-Weimer-Wien, pp 183–207.

Hoang L, Byck S, Prevost L, Scriver CR (1986) PAH mutationanalysis consortium database: A database for disease-produc-ing and other allelic variation at the human PAH locus.Nucleic Acids Res 24:127–131.

Horedt K (1982) “Siebenburgen in spatromischer Zeit.”Bucharest: Kriterion Verlag.

Horedt K (1986) “Siebenburgen im Fruhmittelalter.” Bonn:Antiquitas 4:80–85.

Jaruzelska J, Melle D, Matuszak R, Borski K, Munnich A (1992) Anew 15 bp deletion in exon 11 of the phenylalanine hydroxy-lase gene in phenylketonuria. Hum Mol Genet 9:763–764.

Kleiman S, Li J, Schwartz G, Eisensmith RC, Woo SLC, Shiloh Y(1993) Inactivation of phenylalanine hydroxylase by a mis-sense mutation, R270S, in a Palestinian Kinship with phe-nylketonuria. Hum Mol Genet 5:605–606.

Kleiman S, Schwartz G, Woo SLC, Shiloh Y (1992) A 22-bp de-letion in the phenylalanine hydroxylase gene causing phe-nylketonuria in an Arab Family. Hum Mutat 1:344–346.

ROMANIAN PKU FAMILIES 319

Konecki DS, Lichter-Konecki U (1991) The phenylketonurialocus: Current knowledge about alleles and mutations of thephenylalanine hydroxylase gene in various populations. HumGenet 87:377–388.

Kozak L, Blazkova M, Kuhrova V, Pijackova A, Ruzickova S,Stastna S (1997) Mutation and haplotype analysis of pheny-lalanine hydroxylase alleles in classical PKU patients from theCzech Republic: Identification of four novel mutations. J MedGenet 34:893–898.

Kozak L, Dvorakova D, Pijackova A, Kamaryt J (1993) Haplo-type distribution at the phenylalanine hydroxylase locus inPKU families from the Moravian area of Czechoslovakia. JInher Metab Dis 16:451–456.

Kozak L, Kuhrova V, Blazkova M, Romano V, Fajkusova L,Dvorakova D, Pijackova A (1995) Phenylketonuria mutationsand their relation to RFLP haplotypes at the PAH locus inCzech PKU families. Hum Genet 96:472–476.

Kuzmin A, Eisensmith R, Goltsov A, Sergeeva N, Schwartz E,Woo SLC (1995) Complete spectrum of PAH mutations inTataria: Presence of Slavic, Turkic and Scandinavian muta-tions. Eur J Hum Genet 3:246–255.

Kwok S, Ledley FD, DiLella AG, Robson K, Woo SLC (1985)Nucleotide sequence of a full-length complementary DNAclone and amino acid sequence of human phenylalanine hy-droxylase. Biochemistry 24:556–561.

Lahiri D, Numberger J (1990) A rapid non-enzymatic methodfor the preparation of HMW DNA from blood for RFLP stud-ies. Nucleic Acids Res 19:5444

Popescu A, Andrian T, Guttler F, Guldberg P (1994) Genotype–phenotype correlations in 11 Romanian PKU families. J InherMetab Dis 17:374–375.

Sanger F, Nicklen S, Coulson AR (1977) DNA sequencing withchain-terminating inhibitors. Proc Natl Acad Sci USA74:5463–5467.

Tyfield LA, Stephenson A, Cockburn F, Harvie A, Bidwell JL,Wood NAP, Pilz DT, Harper P, Smith I (1997) Sequence varia-tion at the phenylalanine hydroxylase gene in the British Isles.Am J Hum Genet 60:388–386.

Wedemeyer N, Dworniczak B, Horst J (1991) PCR detection ofthe MspI(Aa) RFLP at the human phenylalanine hydroxy-lase (PAH) locus. Nucleic Acids Res 19:1959.