HUMAN MUTATION 2:131-137 (1993)

METHODS

Rapid Classification of Phenylketonuria Genotypes

Nigel Wood, Linda Tyfield, and Jeffrey Bidwell* University of Bristol Department of Transplantation Sciences, Bristol Homoeopathic Hospital, Cotham, Bristol BS6 6JU, United Kmgdom (N. W., J. B.) and Department of Clinical Chemistry, Southmead Hospital, Bristol BSlO 5NB, United Kingdom (L. T.); Fax: 0272-506-277

Communicuted by Savio L.C. Woo

We describe a rapid and simple method for phenylketonuria genotyping which identifies five point mutations within exon 12 of the human phenylalanine hydroxylase gene. The method involves PCR amplification of the target exon and hybridization with a PCR-amplifiable synthetic DNA (universal heteroduplex generator, UHG). The UHG contains identifiers consisting of nucleotide substitutions and/or deletions, contiguous with known mutation sites within the target exon. DNA heteroduplexes are resolved by nondenaturing polyacrylamide minigel electrophoresis. Individual mutant genotypes are identified by characteristic banding patterns, in either homozygous or heterozygous states. The method may potentially be applied to rapid genotyping of any mutation or series of mutations within PCR- amplifiable genetic material. o 1993 WiIey-Liss, Inc.

INTRODUCTION

Phenylketonuria (PKU) is probably the most common disorder of amino acid metabolism and results from mutations at the phenylalanine hy- droxylase (PAH) gene. The gene consists of 13 exons and encodes a mRNA of 2.8 kb (Kwok et al., 1985; DiLella et al., 1986). To date, more than 50 mutations have been described at the PAH gene but there are wide variations in the frequency distribution of individual mutations be- tween different populations (Scriver et al., 1992). In the southwest of England, three mutations in exon 12, namely IVS12nt1, R408W, and Y414C, account for just over 40% of PKU chromosomes (Tyfield et al., 1991). In vitro expression analyses reveal 100% loss of PAH activity associated with IVS12ntl and R408W mutations, and 50% reduc- tion in activity for the Y414C mutation (Okano et al., 1991a,b). Within this population, other mu- tations found in the PAH gene are in exon 3 (165T) and intron 10 (IVSIOnt546), resulting in 74 and 100% loss of PAH activity, respectively (John et al., 1992; Dworniczak et al., 1991).

A significant proportion of early treated young adult PKU sufferers shows clinical neurological de- terioration (Thompson et al., 1991; Lou et al., 1992), and thus it is important to research the

possible link between specific PAH gene mutations and long-term prognosis. Because of the large number of disease-associated mutations at this lo- cus it is desirable to have available a mutation detection system which is rapid, safe, and reliable, and is able to screen samples for several mutations at one time. A number of PCR-based techniques for mutation detection have been described in re- cent years. Two widely used techniques which de- tect single, previously defined point mutations are sequence-specific oligonucleotide (SSO) typing (Saiki et al., 1986, 1988), and the amplification refractory mutation system (ARMS) (Newton et al., 1989). Other techniques include those that depend on conformational polymorphisms in am- plified single-stranded DNA (SSCP) (Orita e t al., 1989), chemical cleavage of mismatched DNA se- quences in amplified DNA (Cotton et al., 1988), and those that depend on differences in electro- phoretic melting properties between amplified het- eroduplexes and homoduplexes in denaturing or thermal gradient gels (DGGE and TGGE) (Fisher and Lerman, 1983). However, none of these tech- niques is without some limitation regarding safety

Received December 8, 1992; accepted Februaty 2, 1993. "To whom reprint requestsicomespondence should be ad-

dressed.

0 1993 WILEY-LISS, INC.

132 WOOD ET AL.

FIGURE 1. The nucleotide sequence of a portion of the normal human PAH gene, including exon 12 and part of the 5’ and 3’ flanking introns, is shown. Sequences of the PKU point mu- tations and PAH-UHG are shown aligned to the normal se-

or reliability. More importantly, these techniques are not ideal due to their complexity and the re- quirement for specialist equipment. We describe here the application of a new and technically sim- ple mutation detection system based on DNA het- eroduplex analysis (Bidwell and Hui, 1990), to screening PKU individuals for five mutations in exon 12 of the PAH gene. Originally described as a technique for rapid matching of human HLA- DR-Dw allotypes (Bidwell and Hui, 1990; Clay et al., 1991; Wood et al., 1991), it may be applied with modifications to any genetic locus (Bidwell et al., 1993). We report here its efficacy in PKU genotyping.

MATERIALS AND METHODS Construction of the Universal Heteroduplex Generator (UHG)

A specific UHG (PAH-UHG) was constructed to identify five mutations (R408W, R413P, Y414C, T418P, and IVS12ntl) within exon 12 of the PAH gene. Identifiers for four of the five mu- tations were incorporated within the UHG and consisted of nucleotide substitutions or substitu- tions plus deletions (Fig. 1). A specific identifier for the T418P mutation was not necessary. The PAH-UHG was constructed from two synthetic oligonucleotides (“longmers”) : PAH-LEFT #330

quence: dashes indicate identity. Annealing locations for the PAHl2L and PAHl2R PCR primers within the PAH gene are underlined. The overlapping complementary regions of the two constituent longmers are denoted by = .

(sense strand), a 95mer, and PAH-RIGHT #335 (antisense strand), a 98mer, with a region of 3’- complementarity of 17 bases (Fig. 1). Sequences of the longmers were PAH-LEFT #330, 5‘-TGT GGT TTT GGT CTT AGG AAC TTT GCT GCC ACA ATA CCT CAC CTT CTC AGC CGGCTACATACACCCAAAGGATTG AGG TCT TGG ACA ATA CCC AG-3’; and

C T G A G A A A C C G A G T G G C C T C G T A A GGTGTAAATTAAGCTGTTAATGGA ATC AGC CAA AAT CTT AAG CTG CTG GGT ATT GTC CAA GA-3’. These were syn- thesized on an ABI 391A DNA synthesizer using 40 nmol polystyrene support columns (Applied Biosystems). Following synthesis, the two long- mers were joined to form a double-stranded UHG as follows. Reaction mixtures were set up contain- ing 2.5 pl of a 5 pM aqueous solution of each longmer, 5 p1 dNTP mix (dATP, dCTP, dGTP, and dTTP each at a concentration of 10 mM), 1 p1 of 10 mg/ml bovine serum albumin, 10 pl Vent@ DNA polymerase reaction buffer (New England Biolabs), 58 p1 deionized double distilled water, 2 units Vent@ DNA polymerase (New England Bio- labs), added at 72°C to hot-start the reaction (DAquila et al., 1991), and 0.5 pl E. coli single- stranded DNA binding protein (Bind Aid@,

PAH-RIGHT #335, 5’-AGT CTT CGA TTA

RAPID PHENYLKETONURIA GENOTYPING 133

United States Biochemicals). Mixtures were sub- jected to heating at 94°C for 1 min and cooling at 37°C for 1 min. The temperature was increased to 72°C and 10 p1 each of 5 pM solutions of exon 12 PCR primers was added. The sequences of these primers were left (PAH12L), 5’-TGT GGT TTT G G T CTT AGG AA-3‘ and right (PAH12R),

Thermal cycling conditions were as for genomic DNA amplification (see below). The products of six replicate reactions were pooled and electro- phoresed on two 8% wlv nondenaturing polyacryl- amide gels (Protogel, National Diagnostics) as pre- viously described (Bidwell and Hui, 1990). Gel slices containing the double-stranded UHG band were excised and the DNA extracted by isota- chophoresis (Ofverstedt et al., 1984), precipitated with ethanol, washed in 70% ethanol, and redis- solved in 50 p1 sterile double distilled water. Ten microlitres of a 1:10,000 serial dilution was ampli- fied for use in the heteroduplex analyses.

DNA Heteroduplex Analysis

Genomic DNA was isolated from whole blood according to standard protocols, or alternatively from dried blood spots used in neonatal screening tests, as follows: a 2 mmz area of blood spot was excised from the paper, added to a 0.5 ml micro- centrifuge tube containing 25 pl of a 1 : 10 dilution of AmpliTaqO reaction buffer (Perkin Elmer Ce- tus). Samples were heated at 96°C for 15 min, cooled, diluted with a further 25 pl of 1 x Am- pliTaqO reaction buffer, mixed, and centrifuged at 13,OOOg for 15 min. Supernatants were used di- rectly in the PCR. Separate UHG and genomic DNA PCR amplifications were performed in 100 p1 reaction mixtures containing 10 pl heat-dena- tured diluted UHG or genomic DNA isolated from whole blood (100 ng), respectively, 10 p1 each of 5 KM solutions of PAH12L and PAH12R primers (details above), 2 pl dNTP mix [dATP, dCTP, dGTP, and dTTP (Boehringer Mannheim) each at a concentration of 10 mM], 58 pl deionized double distilled water, 10 pl AmpliTaqO reaction buffer, and 2 units AmpliTaqO DNA polymerase (Perkin Elmer Cetus). The enzyme was added at 72°C to hot-start the reaction (D’Aquila et al., 1991). For PCR amplifications of DNA obtained from dried blood spots, 30 pl of supernatant was used in the reaction mixtures, and the volumes of AmpliTaq@ reaction buffer and water were reduced to 7 and 40.6 pl, respectively. The reaction mixtures were subjected to 35 rounds of thermal cycling at 94°C for 1 min, 48°C for 1 min, and 72°C for 1 min.

5’-AGT CTT CGA TTA CTG AGA AA-3‘.

Fifteen microlitres of PCR-amplified genomic DNA and of PCR-amplified PAH-UHG DNA were mixed and subjected to 3 rounds of thermal cycling at 94°C for 1 min, 65°C for 2.5 min, and 72°C for 2.5 min (Note: 1 amplification of the PAH-UHG was therefore sufficient for 6 reactions with PCR-amplified genomic DNA). The entire mixture was subjected to electrophoresis for 90 min at 200 V on 12% wiv nondenaturing polyacryl- amide minigels (Protogel, National Diagnostics) and heteroduplexes were visualised by staining with ethidium bromide as previously described (Bidwell and Hui, 1990).

RESULTS

The PAH-UHG was able to generate character- istic and discriminatory DNA heteroduplexes for five known exon 12 mutations within DNA from PKU individuals, both in the homozygous and het- erozygous states (Fig. 2a and b).

The reliability and reproducibility of the test were examined by screening two sets of genomic DNA samples collected from PKU individuals in whom exon 12 mutations were established by other methods, either prospectively or retrospectively. In the former set (Fig. 3a) samples were analysed in a blind trial and there was complete concordance between results of the PAH-UHG tests and pro- spectively performed analyses using allele specific oligonucleotide (ASO) probes or DNA sequenc- ing. In the latter set, constituting untreated PKU individuals with reduced intelligence from a Scot- tish population, 8 out of 9 samples were identified as heterozygous for the IVS12ntl mutation (Fig. 3b); these results were confirmed by A S 0 analysis (data not shown), and are not unexpected for this population (Sullivan et al., 1989).

A practical consideration of the potential appli- cation of this technology is in the ability to rapidly genotype infants who are positively diagnosed in the neonatal screen for PKU. We addressed this by analysing genomic DNA prepared from a 2 mm2 area of dried blood spots normally used in the neo- natal screening programme. These samples had been obtained from affected individuals who regu- larly send dried blood spots to the screening labo- ratory for monitoring blood phenylalanine concen- trations. The samples were between 1 to 4 months old and had been left at ambient temperatures dur- ing that time. Samples were analysed from PKU individuals whose genotype had been established previously using AS0 probes or DNA sequencing. In all cases, the visual quality of these results (Fig. 4) compared favourably with those obtained with

134 WOOD ET AL.

$1 I- horn het horn het horn het horn het horn het

b I R408W 1 R413P ~ Y414C T418P !lVSlSNTlI

FIGURE 2. Heteroduplex analysis of PAH gene exon 12 muta- tions. (a) Lanes 1 and 12 show molecular weight markers (M, marker VIII, Boehringer Mannheim). The leading double bands in all other lanes are the homoduplexes of the PAH- UHG (leading band, 176 bp) and PAH gene exon 12 (trailing band, 185 bp) PCR products. Bands of apparent molecular weight 400-500 bp are different conformational forms of DNA heteroduplexes (Fig. 2b). Lanes 2, 11, 13, and 16 show the heteroduplexes observed in normal individuals (controls). Other lanes show the variations in heteroduplexes observed in PAH gene exon 12 mutations, denoted beneath each lane: hom, individual homozygous for mutation; het, individual heterozygous for an exon 12 mutation (nonexon 12 PAH gene

mutation present on other chromosome). Lanes 6, 8, and 14 show individuals with different PAH gene exon 12 mutations on each chromosome. In all cases, characteristic heterodu- plexes (b) are observed for each mutation, and in addition the homozygosity or heterozygosity status is readily deter- mined. (b) Schematic representation of DNA heteroduplex banding patterns obtained for each PAH gene exon 12 mu- tation. For each mutation, banding patterns are shown for homozygous (horn) and heterozygous (het) states. The mo- bilities of the bands (apparent molecular weights) are a, 446 bp; b, 433 bp; c, 416 bp; d, 395 bp; e, 374 bp; f, 359 bp; g, 354 bp. Bands c plus g are observed in normal PAH genes and in genes with nonexon 12 mutations.

DNA isolated from fresh blood samples, and per- mitted unambigous assignment of exon 12 geno- types.

DISCUSS I 0 N We describe here the first specific application of

UHG technology undertaken outside the human

HLA class I1 system (Bidwell et al., 1993). UHG technology is a refinement of “PCR fingerprint- ing,” a DNA heteroduplex-based technique origi- nally applied to matching of human HLA-DRB allotypes: in this test, two or more DNA sequences representing different alleles of a gene are enzymat-

RAPID PHENYLKETONURIA GENOTYPING 135

FIGURE 3. (a) Blind trial of concordance between PAH-UHG analysis and the results of other DNA typing analyses. M, molecular weight markers (for details see Fig. 2 legend); C, normal control; lane 1, R408W heterozygous individual; lanes 2, 8, 9, and 11, IVSl2ntl heterozygous individuals; lane 3, individual with Y414C and lVSl2ntl mutations on opposite alleles; lane 4, Y414C heterozygous individual; lane 5, lVSl2ntl homozygote; lane 7, individual with R408W and

1VS12ntl mutations on opposite alleles. Lanes 6, 10 and 12 are PKU individuals with mutations outside exon 12. (b) Classification of mutations in PKU individuals from a Scot- tish population. M, molecular weight markers (for details see Fig. 2 legend); C, normal control; lanes 1-7 and 9, IVSl2ntl heterozygous individuals; lane 8, PKU individual with muta- tion outside exon 12.

ically coamplified by PCR, and differences be- tween alleles are visualised by electrophoretic sep- aration of the resulting DNA homoduplexes and heteroduplexes (Bidwell and Hui, 1990). In PKU and any other situation where relevant genes are multiallelic but not duplicated within the genome, DNA heteroduplex formation following PCR is possible only in “trans” and therefore only in het- erozygotes for a given gene. In these situations, DNA heteroduplexes can be generated by hybri- dising PCR products with synthetic DNA mole- cules which we have termed universal heterodu- plex generators (UHG) (Bidwell et al., 1992). These are PCR-amplifiable DNA sequences that mimic a genomic DNA sequence but differ from it as a result of controlled nucleotide substitutions, deletions, or insertions at sites contiguous to known mutation sites within the genomic DNA. The UHG and genomic DNA sequences are am- plified with the same PCR primers, and hybridised together post-PCR by heating and slow cooling, resulting in DNA heteroduplexes having different

FIGURE 4. Classification of mutations in PKU individuals using DNA isolated from dried blood spots. C, normal control; M, molecular weight markers (for details see Fig. 2 legend); lane 1, individual with R408W and IVSl2ntl mutations on oppo- site alleles; lanes 2, 3, 5, and6, IVSl2ntl heterozygous indi- viduals; lane 4, IVSlZntl homozygous individual.

conformational and charged forms, and thus dif- ferent electrophoretic mobilities, characteristic for individual alleles of a gene (Bidwell e t al., 1993).

136 WOOD ET AL.

Our technique differs significantly from the het- eroduplex analysis system described by White et al. (1992), where the authors construct a series of lin- earized plasmid heteroduplex generators which, in separate tests, detect individual point mutations by heteroduplex analysis of isotopically labeled PCR product. Our technique uses a single (universal) heteroduplex generator to detect a complete series of mutations in a single PCR test, and does not involve the use of radioisotopes.

For the present study, the technique has proved to be a highly sensitive, rapid, and nonradioactive screening and classification system for PKU in which five mutations in exon 12 of the PAH gene can be identified using a single UHG and one PCR-based test. The tests we have described here are technically simple to perform and would be suitable for the routine screening of large numbers of individuals. Results were obtained within 24 hr of receipt of blood samples, and were of equal vi- sual quality when performed on whole blood (Figs. 2 and 3) or dried blood spots (Fig. 4). Some ques- tions still remain regarding the general reliability of the method in discriminating between individ- ual mutations. In particular, although we have shown that individual alleles generate characteris- tic heteroduplex banding patterns, it is not certain whether these patterns are unique to individual mutations and whether previously undescribed mu- tations within the region covered by the UHG can be detected without a specific identifier within the UHG. In addition, we do not know whether dif- ferent banding patterns would be generated if more than one mutation occurred within the same codon and hence within the same modified region of the UHG. Studies are being conducted to ad- dress these questions. Nevertheless, it is important to highlight the fact that a specific modifier was not introduced into the UHG to identify the T418P mutation but the heteroduplex banding pattern generated by the R413P/T418P combina- tion (lane 14, Fig. 2a) is clearly distinguishable from the others.

The present study illustrates the wide potential of UHG-based DNA heteroduplex technology. The technique appears ideally suited to the geno- typing of multiallelic genes where mutations are clustered within a reasonable distance and are flanked by conserved sequences suitable for PCR priming. Genotyping mutations within separate exons in genomic DNA could be facilitated using multiple exon-specific UHGs, or potentially by a single UHG for genotyping cDNA derived from mRNA by reverse transcription PCR (Myers and

Gelfand, 199 1). We are currently applying these techniques to the study of exon 7 of the PAH gene which is known to carry more than 20 different mutations, and to the genotyping of mutations re- sponsible for a variety of other inherited disorders.

ACKNOWLEDGMENTS

We are indebted to Dr. Randy C. Eisensmith of the Howard Hughes Medical Institute, Houston, Texas, for supplying DNA from R413P/T418P het- erozygous and from Y414C homozygous individu- als.

REFERENCES Bidwell JL, Hui KM (1990) Human HLA-DWDw allotype match-

ing by analysis of HLA-DRB gene PCR product polymorphism (PCR fingerprints). Technique 2:93-100.

Bidwell JL, Clay TM, Wood NAP, Pursall MP, Martin AF, Bra- dley BA, Hui KM (1993) Rapid HLA-DR-Dw and DP match- ing by PCR fingerprinting and related DNA heteroduplex technologies. In Hui KM, Bidwell JL (eds): Handbook of HLA Typing Techniques. Boca Raton: CRC Press (in press).

Clay TM, Bidwell ]L, Howard MR, Bradley BA (1991) PCR- fingerprinting for selection of HLA matched unrelated marrow donors. Lancet 33 7: 1049 -1 052.

Cotton RGH, Rodrigues NR, Campbell RD (1988) Reactivity of cytosine and thymine in single-base-pair mismatches with hy- droxylamine and osmium tetroxide and its application to the study of mutations. Proc Natl Acad Sci USA 85:4397-4401.

DAquila RT, Bechtel RJ, Videler ]A, Eron JJ , Gorczyca P, Kaplan JC (1991) Maximizing sensitivity and specificity of PCR by preamplification heating. Nucl Acids Res 19:3749.

DiLella AG, Kwok SCM, Ledley FD, Marvit J , Woo SLC (1986) Molecular structure and polymorphic map of the human phe- nylalanine hydroxylase gene. Biochemistry 25:743-749.

Dworniczak B, Aulehla-Scholz C, Kalaydjieva L, Bartholome K, Grudda K, Horst J (1991) Aberrant splicing of phenylalanine hydroxylase mRNA: The major cause of phenylketonuria in parts of Southern Europe. Genomics 11:242-246.

Fisher SG, Lerman LS (1983) DNA fragments differing by single- base pair substitutions are separated in denaturing gradient gel: Correspondence with melting theory. Proc Natl Acad Sci USA 80:1579-1583.

John SWM, Scriver CR, Laframboise R, Rozen R (1992) In vitro and in vivo correlations for 165T and M1V mutations at the phenylalanine hydroxylase locus. Hum Mut 1: 147-153.

Kwok SCM, Ledley FD, DiLella AG, Robson KJH, Woo SLC (1985) Nucleotide sequence of a full-length complementary DNA clone and amino acid sequence of human phenylalanine hydroxylase. Biochemistry 24:556-561.

Lou HC, Toft PB, Andresen J , Mikkelsen I, Olsen B, Guttler F, Wieslander S, Henriksen 0 (1992) An occipito-temporal syn- drome in adolescents with optimally controlled hyperphenyl- alaninaemia. J Inher Metab Dis 15:687-695.

Myers TW, Gelfand DH (1991) Reverse transcription and DNA amplification by a Themus thermophilus DNA polymerase. Bio- chemistry 30:7661-7666.

Newton CR, Graham A, Hepinstall LE, Powell SJ, Summers C , Kalsheker N, Smith JC, Markham AF (1989) Analysis of any point mutation in DNA. The amplification refractory muta- tions system (ARMS). Nucl Acids Res 17:2503-2515.

Ofverstedt L-G, Hammarstrom K, Balgobin M, Hjerten S, Pet- tersson U, Chattopadhyaya J (1984) Rapid and quantitative

RAPID PHENYLKETONURIA GENOTYPING 137

recovery of DNA fragments from gels by displacement electro- phoresis (isotachophoresis). Biochim Biophys Acta 782: 120- 126.

Okano Y, Eisenamith RC, Dasovich M, Wang T, Guttler F, Woo SLC (1991a) A prevalent missense mutation in northern Eu- rope associated with hyperphenylalaninaemia. Eur J Pediatr 150:347-352.

Okano Y, Eisensmith RC, Guttler F, Lichter-Konecki U, Konecki D, Trefz FK, Dasovich M, Wang T, Henriksen K, Lou H, Woo SLC (1991b) Molecular basis of phenotypic heterogeneity in phenylketonuria. N Engl J Med 324:1232-1238.

Orita M, Suzuki Y, Sekiya T, Hayashi K (1989) Rapid and sensi- tive detection of point mutations and DNA polymorphisms using the polymerase chain reaction. Genomics 5:874-879.

Saiki RK, Bugawan TL, Horn GT, Mullis KB, Erlich HA (1986) Analysis of enzymatically amplified P globin and HLA-DQa DNA with allele-specific oligonucleotide probes. Nature (Lon- don) 324:163-166.

Saiki RK, Chang C-A, Levenson CH, Warren TC, Boehm HH, Kazazian HH, Erlich HA (1988) Diagnosis of sickle cell anemia and P-thalassemia with enzymatically amplified DNA and non-

radioactive allele-specific oligonucleotide probes. N Engl J Med 319:537-541.

Scriver CR, John SMW, Rozen R, Eisensmith RC, Woo SLC (1992) Associations between population, PKU mutations and RFLP haplotypes at the PAH locus. Brain Dysfunct (in press).

Sullivan SE, Moore SD, Connor JM, King M, Cockburn F, Stein- mann B, Gitzelmann R, Daiger SP, Woo SLC (1989) Haplo- type distribution of the human phenylalanine hydroxylase locus in Scotland and Switzerland. Am J Hum Genet 44:652- 659.

Thompson AJ, Smith 1, Kendall BE, You1 BD, Brenton D (1991) MRI changes in early treated patients with phenylketonuria. Lancet 337:1224.

Tyfield LA, Osborn MJ, Holton JB (1991) Molecular heterogene- ity at the phenylalanine hydroxylase locus in the population of the South-west of England. J Med Genet 28:244-247.

White MB, Carvalho M, Derse D, OBrien SJ, Dean M (1992) Detecting single base substitutions as heteroduplex polymor- phisms. Genomics 12301-306.

Wood NAP, Clay TM, Bidwell JL (1991) HLA-DWDw matching by PCR fingerprinting: The origin of PCR fingerprints and further applications. Eur J Immunogenet 18:147-153.