D. B. Whitehouse

Congenital disorders of glycosylation type I leads to altered processing of N-linked glycans, as well as underglycosylation

The N-linked glycans on transferrin and alpha(1)-antitrypsin from patients with congenital disorders of glycosylation type I have increased fucosylation and branching relative to normal controls. The elevated levels of monofucosylated biantennary glycans are probably due to increased alpha-(1-->6) fucosylation. The presence of bi- and trifucosylated triantennary and tetra-antennary glycans indicated that peripheral alpha-(1-->3), as well as core alpha-(1-->6), fucosylation is increased. Altered processing was observed on both the fully and underglycosylated glycoforms.

Variation in alpha-1-antitrypsin phenotypes associated with penicillamine therapy

During a detailed study of alpha-1-antitrypsin (AAT) by isoelectric focusing, of 130 individuals of the PiZ phenotype and their families, an unusual alpha-1-antitrypsin protein pattern was identified which related directly to the effect of penicillamine therapy. The effect, believed to be mediated by reaction with the free cysteine of the alpha-1-antitrypsin molecule, was demonstrable by in vitro experiments. It was not apparently associated with loss of the anti-proteinase properties of alpha-1-antitrypsin and was not specific to the Z gene products. It is however, a source of possible confusion in the accurate assessment of Pi phenotypes.

Isozyme and DNA analysis of human S-adenosyl-L-homocysteine hydrolase (AHCY).

Erythrocyte and tissue isozymes of human AHCY have been studied by starch gel electrophoresis, cellulose acetate electrophoresis, isoelectric focusing and Na dodecyl sulphate electrophoresis. The same isozyme was observed in all the tissues studied, suggesting that human AHCY is encoded by a single structural locus. Two variant alleles were identified in erythrocyte AHCY using starch gel electrophoresis in a sample of 166 unrelated individuals from the British population. The gene frequencies were 0–024 for AHCY*2 and 0.006 for AHCY *3. The variant isozyme patterns could not be distinguished by isoelectric focusing.

A new alpha 1-antitrypsin mutation, Thr-Met 85, (PI Zbristol) associated with novel electrophoretic properties.

A new AAT allele (PI Zbristol) has been discovered in a woman with an obstetric history of three perinatal deaths from fulminant liver disease and no living offspring. She and her father were both PI M1Zbristol heterozygotes. The Zbristol protein is active as a proteinase inhibitor but appeared to be deficient in the plasma to about the same degree as the S protein in MS heterozygotes. It focuses on the basic side of Z and lacks the normal pattern of secondary isoforms associated with the commonly occurring AAT variants and migrates faster than normal on an SDS electrophoresis gel. The Zbristol mutation was found to be a C to T transition at codon 85 changing ACG (Thr) to ATG (Met). This disrupts the N‐glycosylation site starting at Asn 83 preventing glycosylation at residue 83 in the PI Zbristol protein and explains the protein isoelectric focusing and SDS gel electrophoresis results. An analysis of haplotypes in the propositus and her father indicated that the Zbristol mutation occurred on the common M1(Val 213) genetic background. The new mutation also led to the generation of an NlaIII restriction endonuclease recognition site. Cell lines from two offspring tested for the presence of this NlaIII site revealed that one had the variant and the other did not. Thus, the relationship between Zbristol and fulminant liver disease in the offspring is unclear.

Genetic studies on a new deficiency gene (PI*Ztun) at the PI locus.

During a study of the alpha 1 antitrypsin (AAT) protein and its locus (PI) by high resolution isoelectric focusing and direct molecular analysis of 106 PIZ probands and their families, a new allele (Ztun) was identified that resembles Z in many of its properties. Two sibs, both compound heterozygotes for Ztun and Z, showed similar evidence of mild liver involvement that was indistinguishable from that associated with classical ZZ homozygotes. The Ztun protein appeared to be deficient in the plasma to about the same degree as the Z protein. Allele specific oligonucleotide analysis of amplified genomic DNA indicated that the new allele is the result of a mutation in exon V that is identical to the classical G----A transition at codon 342 that results in the Glu----Lys substitution characteristic of the Z allele. An analysis of DNA haplotypes constructed from polymorphic restriction enzyme recognition sites in and around the PI locus confirmed that Ztun probably represents a new mutation at codon 342 that has occurred on an M2-like genetic background.

Analysis of genetic variation in two human thyroxine-binding plasma proteins by immunodetection after isoelectric focusing.

An immunological method for the detection of two thyroxine‐binding plasma proteins, thyroxine‐binding globulin (TBG) and thyroxine‐binding prealbumin (TBPA), following polyacrylamide gel isoelectric focusing is described. Both proteins show complex electrophoretic patterns, attributable to post‐translational glycosylation, which are affected by unfavourable conditions of sample storage. No genetic variants of TBG or TBPA were encountered in a study of N. European individuals (n= 50). The X‐linked TBG polymorphism, previously detected using a radiolabelled ligand binding technique in Pacific Islanders, was, however, demonstrable by our new immunological method in a previously untested population (n= 139) from the island of Maewo in Vanuatu.

Rhodanese isozymes in human tissues

An investigation of a range of tissue homogenates by various electrophoretic methods, followed by staining for specific enzyme activity, has revealed a series of isozymes of human rhodanese. Polyacrylamide gel isoelectric focusing provided the most data and rhodanese activity was found in all of the tissues examined. The simplest isozyme pattern was found in red cell lysates; liver homogenates generated the most complex pattern which included the ‘red cell’ forms together with a set of more basic ‘tissue’ isozymes. Variation in isozyme patterns thought to be attributable to storage changes affecting reactive sulphydryl residues was observed in ‘red cell’ rhodanese but no genetic variants of either ‘red cell’ or ‘tissue’ rhodanese were encountered in a study of material from the European population. We conclude that ‘red cell’ and ‘tissue’ rhodanese are determined by separate genes but more than one locus may be concerned with the synthesis of the heterogeneous ‘tissue’ isozymes.

‘SZ like’α1‐antitrypsin phenotypes in PI ZZ children with liver disease

Using high resolution isoelectric focusing, α1‐antitrypsin phenotypes were studied in 106 individuals of the PI ZZ genotype including 71 with liver disease, 22 with chest disease and 13 healthy subjects. The resulting Z patterns were found to be highly variable. In the majority of cases (89/106) the maximum staining intensity was either in the most basic isoform or shared equally between two basic isoforms of the Z phenotype. However, in 17 cases there was a marked intensification of the more acidic isoforms resulting in a pattern which closely resembled the SZ phenotype. This ‘SZ like’ pattern occurred more frequently in the liver group (16/71) than the chest group (0/22) or healthy (1/13) controls. One possible consequence of the ‘SZ like’ pattern is confusion with the genuine SZ phenotype leading to misclassification. If this were so, there could be an erroneous exaggeration of the actual incidence of childhood liver disease associated with PI SZ.

Identification and characterisation of polymorphisms in human phosphoglucomutase (PGM1)

This study is part of our effort to map recombination hotspots in two regions (site A, 18 kb; site B, 40 kb) of the human phosphoglucomutase PGM1 gene. Twenty‐two PCR amplified fragments comprising six groups, covering about 5.2 kb, were screened for single nucleotide polymorphisms (SNPs) using non‐isotopic single stranded conformation polymorphism (SSCP) analysis. Fourteen fragments were variable and seven of these showed common polymorphism. Our strategy for screening for polymorphic sites in the PGM1 gene was based on the results of allelic association analysis between each new marker and the sites of the classical isozyme polymorphism (2/1 in exon 4 and +/− in exon 8). Samples from four populations (Caucasian, Chinese, Vietnamese and New Guinean) were typed for each of the seven polymorphic markers. Between two and four common alleles were found in each case, together with a few rare alleles. Co‐dominant inheritance patterns were demonstrated by family studies. The molecular basis of each new marker was determined by direct sequencing of the PCR products: most were SNPs except two that were small insertions/deletions. Direct sequence analysis of a 2.1 kb segment in sixteen individuals revealed no additional nucleotide variation indicating a very high level of efficiency of the SSCP screening method used in this study. The overall nucleotide diversity (θ) for PGM1 was estimated as 0.9×10−3based on 33 segregating sites in a sequence of 5187 nt and a sample size of 614 individuals.

Genetic polymorphism in the 3‘ ntranslated region of human phosphoglucomutase-1

A 317‐bp segment of DNA from the 3′ region of the human phosphoglucomutase‐1 (PGMl) gene has been examined by a non‐radioactive technique for the occurrence of single‐strand conformation polymorphism (SSCP), Eight phenotypes were detected and attributed to the presence of four alleles. Genetic analysis of 75 unrelated individuals and six CEPH families whose PGMl protein phenotypes were known revealed strong association between the PGMl ‘+’ and ‘−’ isozyme phenotypes and the variation detected in this region, but no association with the PGMl 1 and PGMl 2 isozyme phenotypes. DNA sequence analysis demonstrated the presence of three nucleotide substitutions underlying the alleles, which were located in the untranslated region of the PGMl gene. There was complete correlation between the nucleotide sequence and the phenotype detected by SSCP analysis. This study provides support for the model that the PGMl isozyme polymorphism is determined at two distinct sites in the coding sequence, one coding for the ‘1’ and ‘2’ alleles and the other coding for the ‘+’ and ‘−’ alleles, separated by a region where intragenic recombination occurs.