Philip P. Dembure

Phenylketonuria provides a human model for the study of the effect of phenylalanine on brain function. Although irreversible mental retardation is preventable through newborn diagnosis and dietary phenylalanine restriction, controversy exists regarding the effects of increased concentrations of phenylalanine in older patients. We have studied ten older, treated, phenylketonuric patients using a triple-blind, multiple trials, crossover design. Each patient was tested at the end of each of three 1-wk periods of high or low phenylalanine intakes. Tests included a repeatable battery of neuropsychological tests, analysis of plasma amino acids, and measurement of urine amino acids, phenyl organic acids, dopamine, and serotonin. In all 10 patients plasma phenylalanine rose (900-4,000 microM). In 9 of 10 patients there was an inverse relationship between plasma phenylalanine and urine dopamine excretion. When blood phenylalanine was elevated, these patients had prolonged performance times on neuropsychological tests of higher but not lower integrative function. Urinary serotonin fell during phenylalanine loading in six patients. The concentration of phenylacids in the urine was not proportional to the plasma phenylalanine at concentrations below 1.5 mM. In one patient, neither performance time nor dopamine excretion varied as blood phenylalanine rose or fell. We interpret these data as follows: blood phenylalanine above 1.3 mM impairs performance on neuropsychological tests of higher integrative function, this effect is reversible, and one mechanism may involve impaired biogenic amine synthesis.

Abstract Since blood spots on filter paper are routinely collected in many areas for phenylketonuria testing of newborn infants, we developed several methods for detecting sickle-cell disease and other hemoglobinopathies using this type of specimen. Simple, inexpensive modifications of cellulose acetate electrophoresis plus citrate-agar and globin-chain electrophoresis are employed, permitting one technician to screen as many as 500 specimens per day at a material cost of $0.03 per specimen. (N Engl J Med 288:1265–1268, 1973)

The human cDNA and gene for galactose-1-phosphate uridyl transferase (GALT) have been cloned and sequenced. A prevalent mutation (Q188R) is known to cause classic galactosemia (G/G). G/G galactosemia has an incidence of 1/38,886 in 1,396,766 Georgia live-born infants, but a more common variant of galactosemia, Duarte, has an unknown incidence. The proposed Duarte biochemical phenotypes of GALT are as follows: D/N, D/D, and D/G, which have approximately 75%, 50%, and 25% of normal GALT activity respectively. In addition, the D allele has isoforms of its enzyme that have more acidic pI than normal. Here we systematically determine (a) the prevalence of an A-to-G transition at base pair 2744 of exon 10 in the GALT gene, transition that produces a codon change converting asparagine to aspartic acid at position 314 (N314D), and (b) the association of this mutation with the Duarte biochemical phenotype. The 2744G nucleotide change adds an AvaII (SinI) cut site, which was identified in PCR-amplified DNA. In 111 biochemically unphenotyped controls with no history of galactosemia, 13 N314D alleles were identified (prevalence 5.9%). In a prospective study, 40 D alleles were biochemically phenotyped, and 40 N314D alleles were found. By contrast, in 36 individuals known not to have the Duarte biochemical phenotype, no N314D alleles were found. We conclude that the N314D mutation is a common allele that probably causes the Duarte GALT biochemical phenotype and occurs in a predominantly Caucasian, nongalactosemic population, with a prevalence of 5.9%.

We describe a stratagem for identifying new mutations in the galactose-1-phosphate uridyl transferase (GALT) gene. GALT enzyme activity and isoforms were defined in erythrocytes from probands and their first-degree relatives. If the biochemical phenotypes segregated in an autosomal recessive pattern, we screened for common mutations by using multiplex PCR and restriction endonuclease digestions. If common mutant alleles were not present, the 11 exons of the GALT gene were amplified by PCR, and variations from the normal nucleotide sequences were identified by SSCP. The suspected region(s) was then analyzed by direct DNA sequencing. We identified 86 mutant GALT alleles that reduced erythrocyte GALT activity. Seventy-five of these GALT genomes had abnormal SSCP patterns, of which 41 were sequenced, yielding 12 new and 21 previously reported, rare mutations. Among the novel group of 12 new mutations, an unusual biochemical phenotype was found in a family whose newborn proband has classical galactosemia. He had inherited two mutations in cis (N314D-E203K) from his father, whose GALT activity was near normal, and an additional GALT mutation in the splice-acceptor site of intron C (IVSC) from his mother. The substitution of a positively charged E203K mutation created a unique isoform-banding pattern. An asymptomatic sister's GALT genes carries three mutations (E203K-N314D/N314D) with eight distinct isoform bands. Surprisingly, her erythrocytes have normal GALT activity. We conclude that the synergism of pedigree, biochemical, SSCP, and direct GALT gene analyses is an efficient protocol for identifying new mutations and speculate that E203K and N314D codon changes produce intraallelic complementation when in cis.