Eur J Pediatr (1991) 150 : 347-352 034061999100055L
European Journal of
Pediatrics
9 Springer-Verlag1991
A prevalent missense mutation in Northern Europe associated with hyperphenylalaninaemia Y. Okano 1, R. C. Eisensmith 1, M. Dasovich 1, T. Wang 1, F. Giittler 2, and S. L. C. Woo 1 ~Howard Hughes Medical Institute, Department of Cell Biology and Institute of Molecular Genetics, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030, USA 2john F. Kennedy Institutiet, Gammel Landevej 7, DK-26000 Glostrup, Denmark Received June 8, 1990 / Accepted July 30, 1990
Abstract. A missense mutation has been identified in the phenylalanine hydroxylase (PAH) gene of a Danish patient with hyperphenylalaninaemia (HPA). An A-to-G transition at the second base of codon 414 results in the substitution of Cys for Tyr in the mutant P A H protein. In in vitro expression studies, the Tyr414-to-Cys 414 mutant construct produced a protein which exhibited a significant amount of normal P A H enzyme activity, which is consistent with both in vitro and in vivo measurements of P A H activity in H P A patients. Population genetic studies reveal that this mutation is present on 50% of mutant haplotype 4 chromosomes in the Danish population. Together with the previously reported codon 158 mutation, these two mutant alleles comprise over 90% of all mutant haplotype 4 chromosomes in the Northern European population. Thus, two allele-specific oligonucleotide probes can detect most mutant haplotype 4 chromosomes in Northern Europe,
Key words: Phenylketonuria - Hyperphenylalaninemia - Phenylalanine hydroxylase - Gene mutations - D N A amplification
Introduction Classical phenylketonuria (PKU) is an autosomal recessive genetic disorder caused by a deficiency of hepatic phenylalanine hydroxylase (PAH) and is among the most common inborn errors of amino acid metabolism. Clinically, this metabolic disorder is very heterogeneous. Offprint requests to: S. L. C. Woo Abbreviations: ASO = allele-specificoligonucleotide; HPA =
hyperphenylalaninemia; PAH = phenylalanine hydroxylase; PKU = phenylketonuria; RFLP = restriction fragment length polymorphism
Since the introduction of mass neontal screening programs, physicians have indentified patients with apparent P A H defects ranging from severe PKU, which requires rigid dietary restriction to prevent severe postnatal mental retardation, to benign hyperphenylalaninaemia (HPA), which does not need dietary therapy [11, 28]. The study of PKU at the molecular level began with the detection of eight restriction fragment length polymorphisms (RFLPs) in or near the P A H locus [21, 33]. These RFLPs are sufficiently prevalent among Caucasians that about 90% of all individuals are heterozygous at this locus [5]. RFLP analysis of normal and PKU families has demonstrated a strict concordance between the presence of PKU alleles and the disease state [31]. Thus, RFLP haplotype analysis can be used to perform prenatal diagnosis in most PKU families [20]. Approximately 50 different RFLP haplotypes have now been observed in Caucasians [32]. However, about 80% of all PKU chromosomes in Caucasian populations are represented by RFLP haplotypes 1-4. Haplotypes 2 and 3 are especially prevalent among PKU chromosomes, despite the fact that these haplotypes are relatively rare among normal chromosomes [5]. The relative prevalence of haplotypes 2 and 3 among PKU chromosomes suggested that there might be a close association between specific haplotypes and mutations in the P A H locus, as had been previously observed at the ~3-globin locus [26]. This prediction was confirmed by the subsequent identification of the Arg4~ 4~ mutation associated with haplotype 2 chromosomes and a splicing mutation at the exon 12/intron 12 boundary associated with haplotype 3 chromosomes. In both of these eases, the linkage between a given mutation and its respective haplotype was nearly absolute [7, 8]. In other words, nearly all haplotype 2 PKU chromosomes bear the Arg4~ Trp 4~ mutation, while nearly all haplotype 3 PKU chromosomes bear the splicing mutation. In contrast to haplotypes 2 and 3, haplotypes 1 and 4 are well represented among normal chromosomes in Caucasian populations
348 and should therefore have a higher probability of sustaining multiple mutational events. In fact, distinct mutations have b e e n detected on m u t a n t h a p l o t y p e 1 and 4 c h r o m o s o m e s in Caucasians: Arg261-to-Gln 26x and Glu 28~ to-Lys 28~ mutations are on a haplotype 1 background [1, 24, 25], while Arg158-to-Gln 158 and Arg243-to-Ter mutations are on haplotype 4 background [10, 24, 30]. More recently, several other P K U mutations have been identified in different populations, each linked to different haplotypes [2, 13, 19, 22]. O f the mutations associated with haplotype 4, the c o d o n 158 m u t a t i o n is prevalent, comprising 33% of all m u t a n t haplotype 4 c h r o m o s o m e s in Switzerland [24], while the c o d o n 243 m u t a t i o n is rare a m o n g Caucasians [30]. T o identify the remaining prevalent mutations associated with m u t a n t haplotype 4 chrom o s o m e s , detailed sequence analysis was p e r f o r m e d on a Danish P K U patient bearing m u t a n t haplotype 3 and 4 chromosomes.
Materials and methods
antibody [27]. PAH RNA and enzymatic activity were determined in cellular extracts as described by Ledley et al. [17, 18]. 13-galactosidase enzyme activity was measured as described by Nielsen et al. [23].
Results
Identification of a missense mutation in the human P A H gene Direct D N A sequence analysis of exon 12 plus flanking intronic regions of the P A H gene amplified f r o m the genomic D N A of Danish P K U patient revealed two bands (G and A ) , each present with r e d u c e d intensity, at a position where only a single b a n d (A) is present in the n o r m a l D N A sequence (Fig. 1). This result demonstrates the presence of two alleles within the patient's D N A , one of which is n o r m a l and one of which bears an A - t o G transition at the second base of c o d o n 414. This mutation results in the substitution of Cys for Tyr in the P A H protein.
Mutation analysis Genomic DNA was isolated from leukocytes of a Danish PKU patient bearing haplotype 3 and 4 mutant chromosomes. Amplification primers complementary to intronic regions of the PAH gene were synthesized by Genetic Designs Inc., Houston. The 13 exoncontaining regions of the 90 kilobase (kb) PAH gene were amplified via the polymerase chain reaction for 35 cycles as previously described [30]. Amplified exonic DNAs were purified by elution from agarose gels and recovered using Gene-Clean (Bio 101) according to the manufacturer's instructions. The purified doublestrand DNA was sequenced directly using the method of Wang et al. [30].
Dot-blot hybridization analysis Amplified exonic samples were denatured and applied to Zetaprobe membrane (Bio Rad, Richmond, CA, USA) using a dotblot manifold (S and S) [9]. Allele-specific oligonucleotide (ASO) probes 17-mers) were endlabeled with gamma-[32p[-ATP (6000 Ci/ mmol, NEN), using polynucleotide kinase (Pharmacia, Piscataway, NJ, USA). Hybridization and stringent washing was performed as described by DiLella and Woo [6].
Mutation verification by expression analysis in mammalian cells T h e p C D N A - 1 expression vector containing either the 414 m u t a n t or the n o r m a l full-length h u m a n P A H c D N A was co-transfected into C O S cells along with p C D N A - 1 containing the [3-galactosidase c D N A . J3-galactosidase activity was not significantly different in cells transfected with either the m u t a n t construct or the n o r m a l construct, demonstrating that the transfection efficiency of both constructs was c o m p a r a b l e (data not shown). N o r t h e r n and dot-blot hybridization studies of transfected cells indicated that the n o r m a l and m u t a n t c D N A constructs expressed similar m R N A levels of the correct sized (Fig. 2). These results suggest that n o r m a l and m u t a n t P A H m R N A s are present at similar steady-state levels. T h e a m o u n t of P A H enzymatic activity in 0-200 gg of protein derived from crude extracts of C O S (CV-1 origin,
Expression analysis The site-directed mutagenesis method of Kunkel et al. [16] was used to create a mutant PAH cDNA for expression analysis. Briefly, a full-length human PAH cDNA fragment was first inserted into the EcoR1 site of M13mp18 in the sense orientation, then transformed into Escherichia coli strain CJ236 to obtain a uracil-containing template for mutagenesis. A 17-mer mutagenizing primer was designed to introduce a specific nucloetide substitution in codon 414 of the PAH cDNA and double-strand DNA containing the mutant sequence was synthesized by T4 DNA polymerase and T4 DNA ligase (Bio Rad) at 37~ for 90 min. The mutant clone was then transformed into TG-1 cells and its authenticity was confirmed by DNA sequence analysis. The mutant and normal PAH cDNAs were then subcloned into the eukaryotic expression vector pCDNA-1 (Invitrogen, San Diego) following digestion with EcoR1 and extraction from low-melting agarose gel. Each of the resultant expression vectors (70gg) and pCDNA-1 containing a 13-galactosidase cDNA (30 ~tg were co-transfected into monkey kidney COS ceils by electroporation [4]. Immunoreactive PAH was identified by Western blotting with goat anti-rat PAH
Normal
Mutant
'C T A G"C T A G~ T A C
, Tyr 414
T A+G C
, Cys 414
Fig.1. Identification of a missense mutation in exon 12 of the human PAH gene, The exon 12-containing regions from a normal individual and a PKU patient were amplified by PCR using sense strand primer (5'-ATGCCACTGAGAACTCTCTT-3') complementary to intron 11 and an antisense strand primer (5'-AGTCTT CGATTACTGAGAAA-3') complementary to intron 12 and sequenced with the primer (5'-CCGAGTGGCCTCGTAAG-3'). Two bands (G and A) are present at the same position in the mutant sequence, while only a single band (G) band is present in the normal sequence. This A-to-G transition in exon 12 results in the substitution of a Cys for Tyr at amino acid 414 of the PAH protein
349
1
3
2
1
2
3
4
PAH
PAH a
1
2
Fig. 4. Western blot analysis of PAH in COS cells transfected with normal and mutant human PAH cDNA. Lane 1-3, 200, 100 and 50 pg of protein extracts from the normal PAH cDNA construct, respectively. Lane 4, 200 gg of protein extract from the mutant 414 PAH cDNA construct
3 (pg)
10
5 2.5 f~
3~4
1.2 5
b Fig. 2 a, b. Analysis of PAH mRNA in COS cells transfected with the normal or mutant PAH cDNA constructs, a Northern blot hybridization for qualitative RNA analysis using the PAH cDNA as a probe. 10 gg of total RNA extract from transfected COS cells was applied to each lane. b Dot-blot hybridization for quantitative RNA analysis using the PAH cDNA as a probe. Serially-diluted RNA samples containing 10, 5, 2.5 and 1.25 gg of total RNA were applied to each lane. Lane 1, mutant 414 PAH cDNA construct; Lane 2, normal PAH cDNA construct; Lane 3, mock-transfected cells
gl00
>~ 80 o
60
o 20 n
0
l
3"/4" Normal
probe
Mutant
probe
4/4"
Fig. 5. Transmission of the missense mutant alleles in a Danish PKU family. Exon 12-containing region (245 bp) of PAH was amplified from genomic DNA by PCR, dot-blotted onto z-probe membrane and hybridized with ASO probes. The following probes were used to detect the substitution at the codon 414 in exon 12: the normal probe (5'-AGTTCGCTACGACCCAT-3') is the sense DNA strand and the mutant probe (5'-ATGGGTCGCAGCGA ACT-3') is the antisense DNA strand. II, Mutant haplotype 3 allele; [Z, mutant haplotype 4 allele; [~, normal haplotype alleles
40
"r"
European Journal of
Pediatrics
9 Springer-Verlag1991
A prevalent missense mutation in Northern Europe associated with hyperphenylalaninaemia Y. Okano 1, R. C. Eisensmith 1, M. Dasovich 1, T. Wang 1, F. Giittler 2, and S. L. C. Woo 1 ~Howard Hughes Medical Institute, Department of Cell Biology and Institute of Molecular Genetics, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030, USA 2john F. Kennedy Institutiet, Gammel Landevej 7, DK-26000 Glostrup, Denmark Received June 8, 1990 / Accepted July 30, 1990
Abstract. A missense mutation has been identified in the phenylalanine hydroxylase (PAH) gene of a Danish patient with hyperphenylalaninaemia (HPA). An A-to-G transition at the second base of codon 414 results in the substitution of Cys for Tyr in the mutant P A H protein. In in vitro expression studies, the Tyr414-to-Cys 414 mutant construct produced a protein which exhibited a significant amount of normal P A H enzyme activity, which is consistent with both in vitro and in vivo measurements of P A H activity in H P A patients. Population genetic studies reveal that this mutation is present on 50% of mutant haplotype 4 chromosomes in the Danish population. Together with the previously reported codon 158 mutation, these two mutant alleles comprise over 90% of all mutant haplotype 4 chromosomes in the Northern European population. Thus, two allele-specific oligonucleotide probes can detect most mutant haplotype 4 chromosomes in Northern Europe,
Key words: Phenylketonuria - Hyperphenylalaninemia - Phenylalanine hydroxylase - Gene mutations - D N A amplification
Introduction Classical phenylketonuria (PKU) is an autosomal recessive genetic disorder caused by a deficiency of hepatic phenylalanine hydroxylase (PAH) and is among the most common inborn errors of amino acid metabolism. Clinically, this metabolic disorder is very heterogeneous. Offprint requests to: S. L. C. Woo Abbreviations: ASO = allele-specificoligonucleotide; HPA =
hyperphenylalaninemia; PAH = phenylalanine hydroxylase; PKU = phenylketonuria; RFLP = restriction fragment length polymorphism
Since the introduction of mass neontal screening programs, physicians have indentified patients with apparent P A H defects ranging from severe PKU, which requires rigid dietary restriction to prevent severe postnatal mental retardation, to benign hyperphenylalaninaemia (HPA), which does not need dietary therapy [11, 28]. The study of PKU at the molecular level began with the detection of eight restriction fragment length polymorphisms (RFLPs) in or near the P A H locus [21, 33]. These RFLPs are sufficiently prevalent among Caucasians that about 90% of all individuals are heterozygous at this locus [5]. RFLP analysis of normal and PKU families has demonstrated a strict concordance between the presence of PKU alleles and the disease state [31]. Thus, RFLP haplotype analysis can be used to perform prenatal diagnosis in most PKU families [20]. Approximately 50 different RFLP haplotypes have now been observed in Caucasians [32]. However, about 80% of all PKU chromosomes in Caucasian populations are represented by RFLP haplotypes 1-4. Haplotypes 2 and 3 are especially prevalent among PKU chromosomes, despite the fact that these haplotypes are relatively rare among normal chromosomes [5]. The relative prevalence of haplotypes 2 and 3 among PKU chromosomes suggested that there might be a close association between specific haplotypes and mutations in the P A H locus, as had been previously observed at the ~3-globin locus [26]. This prediction was confirmed by the subsequent identification of the Arg4~ 4~ mutation associated with haplotype 2 chromosomes and a splicing mutation at the exon 12/intron 12 boundary associated with haplotype 3 chromosomes. In both of these eases, the linkage between a given mutation and its respective haplotype was nearly absolute [7, 8]. In other words, nearly all haplotype 2 PKU chromosomes bear the Arg4~ Trp 4~ mutation, while nearly all haplotype 3 PKU chromosomes bear the splicing mutation. In contrast to haplotypes 2 and 3, haplotypes 1 and 4 are well represented among normal chromosomes in Caucasian populations
348 and should therefore have a higher probability of sustaining multiple mutational events. In fact, distinct mutations have b e e n detected on m u t a n t h a p l o t y p e 1 and 4 c h r o m o s o m e s in Caucasians: Arg261-to-Gln 26x and Glu 28~ to-Lys 28~ mutations are on a haplotype 1 background [1, 24, 25], while Arg158-to-Gln 158 and Arg243-to-Ter mutations are on haplotype 4 background [10, 24, 30]. More recently, several other P K U mutations have been identified in different populations, each linked to different haplotypes [2, 13, 19, 22]. O f the mutations associated with haplotype 4, the c o d o n 158 m u t a t i o n is prevalent, comprising 33% of all m u t a n t haplotype 4 c h r o m o s o m e s in Switzerland [24], while the c o d o n 243 m u t a t i o n is rare a m o n g Caucasians [30]. T o identify the remaining prevalent mutations associated with m u t a n t haplotype 4 chrom o s o m e s , detailed sequence analysis was p e r f o r m e d on a Danish P K U patient bearing m u t a n t haplotype 3 and 4 chromosomes.
Materials and methods
antibody [27]. PAH RNA and enzymatic activity were determined in cellular extracts as described by Ledley et al. [17, 18]. 13-galactosidase enzyme activity was measured as described by Nielsen et al. [23].
Results
Identification of a missense mutation in the human P A H gene Direct D N A sequence analysis of exon 12 plus flanking intronic regions of the P A H gene amplified f r o m the genomic D N A of Danish P K U patient revealed two bands (G and A ) , each present with r e d u c e d intensity, at a position where only a single b a n d (A) is present in the n o r m a l D N A sequence (Fig. 1). This result demonstrates the presence of two alleles within the patient's D N A , one of which is n o r m a l and one of which bears an A - t o G transition at the second base of c o d o n 414. This mutation results in the substitution of Cys for Tyr in the P A H protein.
Mutation analysis Genomic DNA was isolated from leukocytes of a Danish PKU patient bearing haplotype 3 and 4 mutant chromosomes. Amplification primers complementary to intronic regions of the PAH gene were synthesized by Genetic Designs Inc., Houston. The 13 exoncontaining regions of the 90 kilobase (kb) PAH gene were amplified via the polymerase chain reaction for 35 cycles as previously described [30]. Amplified exonic DNAs were purified by elution from agarose gels and recovered using Gene-Clean (Bio 101) according to the manufacturer's instructions. The purified doublestrand DNA was sequenced directly using the method of Wang et al. [30].
Dot-blot hybridization analysis Amplified exonic samples were denatured and applied to Zetaprobe membrane (Bio Rad, Richmond, CA, USA) using a dotblot manifold (S and S) [9]. Allele-specific oligonucleotide (ASO) probes 17-mers) were endlabeled with gamma-[32p[-ATP (6000 Ci/ mmol, NEN), using polynucleotide kinase (Pharmacia, Piscataway, NJ, USA). Hybridization and stringent washing was performed as described by DiLella and Woo [6].
Mutation verification by expression analysis in mammalian cells T h e p C D N A - 1 expression vector containing either the 414 m u t a n t or the n o r m a l full-length h u m a n P A H c D N A was co-transfected into C O S cells along with p C D N A - 1 containing the [3-galactosidase c D N A . J3-galactosidase activity was not significantly different in cells transfected with either the m u t a n t construct or the n o r m a l construct, demonstrating that the transfection efficiency of both constructs was c o m p a r a b l e (data not shown). N o r t h e r n and dot-blot hybridization studies of transfected cells indicated that the n o r m a l and m u t a n t c D N A constructs expressed similar m R N A levels of the correct sized (Fig. 2). These results suggest that n o r m a l and m u t a n t P A H m R N A s are present at similar steady-state levels. T h e a m o u n t of P A H enzymatic activity in 0-200 gg of protein derived from crude extracts of C O S (CV-1 origin,
Expression analysis The site-directed mutagenesis method of Kunkel et al. [16] was used to create a mutant PAH cDNA for expression analysis. Briefly, a full-length human PAH cDNA fragment was first inserted into the EcoR1 site of M13mp18 in the sense orientation, then transformed into Escherichia coli strain CJ236 to obtain a uracil-containing template for mutagenesis. A 17-mer mutagenizing primer was designed to introduce a specific nucloetide substitution in codon 414 of the PAH cDNA and double-strand DNA containing the mutant sequence was synthesized by T4 DNA polymerase and T4 DNA ligase (Bio Rad) at 37~ for 90 min. The mutant clone was then transformed into TG-1 cells and its authenticity was confirmed by DNA sequence analysis. The mutant and normal PAH cDNAs were then subcloned into the eukaryotic expression vector pCDNA-1 (Invitrogen, San Diego) following digestion with EcoR1 and extraction from low-melting agarose gel. Each of the resultant expression vectors (70gg) and pCDNA-1 containing a 13-galactosidase cDNA (30 ~tg were co-transfected into monkey kidney COS ceils by electroporation [4]. Immunoreactive PAH was identified by Western blotting with goat anti-rat PAH
Normal
Mutant
'C T A G"C T A G~ T A C
, Tyr 414
T A+G C
, Cys 414
Fig.1. Identification of a missense mutation in exon 12 of the human PAH gene, The exon 12-containing regions from a normal individual and a PKU patient were amplified by PCR using sense strand primer (5'-ATGCCACTGAGAACTCTCTT-3') complementary to intron 11 and an antisense strand primer (5'-AGTCTT CGATTACTGAGAAA-3') complementary to intron 12 and sequenced with the primer (5'-CCGAGTGGCCTCGTAAG-3'). Two bands (G and A) are present at the same position in the mutant sequence, while only a single band (G) band is present in the normal sequence. This A-to-G transition in exon 12 results in the substitution of a Cys for Tyr at amino acid 414 of the PAH protein
349
1
3
2
1
2
3
4
PAH
PAH a
1
2
Fig. 4. Western blot analysis of PAH in COS cells transfected with normal and mutant human PAH cDNA. Lane 1-3, 200, 100 and 50 pg of protein extracts from the normal PAH cDNA construct, respectively. Lane 4, 200 gg of protein extract from the mutant 414 PAH cDNA construct
3 (pg)
10
5 2.5 f~
3~4
1.2 5
b Fig. 2 a, b. Analysis of PAH mRNA in COS cells transfected with the normal or mutant PAH cDNA constructs, a Northern blot hybridization for qualitative RNA analysis using the PAH cDNA as a probe. 10 gg of total RNA extract from transfected COS cells was applied to each lane. b Dot-blot hybridization for quantitative RNA analysis using the PAH cDNA as a probe. Serially-diluted RNA samples containing 10, 5, 2.5 and 1.25 gg of total RNA were applied to each lane. Lane 1, mutant 414 PAH cDNA construct; Lane 2, normal PAH cDNA construct; Lane 3, mock-transfected cells
gl00
>~ 80 o
60
o 20 n
0
l
3"/4" Normal
probe
Mutant
probe
4/4"
Fig. 5. Transmission of the missense mutant alleles in a Danish PKU family. Exon 12-containing region (245 bp) of PAH was amplified from genomic DNA by PCR, dot-blotted onto z-probe membrane and hybridized with ASO probes. The following probes were used to detect the substitution at the codon 414 in exon 12: the normal probe (5'-AGTTCGCTACGACCCAT-3') is the sense DNA strand and the mutant probe (5'-ATGGGTCGCAGCGA ACT-3') is the antisense DNA strand. II, Mutant haplotype 3 allele; [Z, mutant haplotype 4 allele; [~, normal haplotype alleles
40
"r"
