Independent Mutations at Codon 3500 of the Apolipoprotein B Gene Are Associated With Hyperlipidemia
Abstract The apoB arginine-to-glutamine change at codon 3500 has become established as a cause of failure of binding of the LDL particle to its receptor and the consequent hypercholesterolemia of familial defective apoB 100. A search for further similar mutations was undertaken by systematic screening of a candidate region of the apoB gene from individuals with hypercholesterolemia. Polymerase chain reaction and denaturing gradient gel electrophoresis were used. We describe two families in which a different mutation in the codon 3500 causes an arginine-to-tryptophan substitution. Most adults in these families who have this mutation have hypercholesterolemia. LDL derived from all who have inherited the mutation is dysfunctional in that it allows only poor growth of an LDL cholesterol–dependent cell line. We conclude that this arginine 3500 is essential to the function of apoB and that its loss and replacement by glutamine or tryptophan is responsible for the hypercholesterolemia of familial defective apoB 100.
Reprint requests to Dr Dairena Gaffney, Institute of Pathological Biochemistry, Glasgow Royal Infirmary, 4th Floor Queen Elizabeth Bldg, Alexandra Parade, Glasgow G31 2ER, Scotland.
- Received February 27, 1995.
- Accepted May 3, 1995.
The west of Scotland has one of the highest incidences of heart disease in the world. There is a high prevalence of classic environmental risk factors, but underlying genetic causes are also being sought. ApoB is the major protein associated with the circulating atherogenic LDL particle. Normally LDL is cleared when it binds to the LDL receptor, but mutation in either the ligand apoB gene or the LDL receptor gene can disrupt binding. The apoB arginine-to-glutamine change at codon 3500 has become established as a cause of failure of binding of LDL to its receptor and consequent hypercholesterolemia.1 2 This condition is known as familial defective apoB 100 (FDB). (We suggest that the term FDB be recognized as the generic acronym for familial defective apoB. Because there now exists more than one genotype, we suggest that the amino acid position continue to be used to classify subjects [eg, FDB3500]. However, for a situation in which there are two different functional mutations in one codon, we suggest a single-letter code be used to indicate the unusual amino acid [eg, FDB3500Q for the original FDB described1 ]. This is in line with the suggestion by Myant.2 ) Despite the fact that codon 3500 is outside the original putative receptor binding region,3 ample evidence confirms that LDL containing apoB with the arginine-to-glutamine change at codon 3500 is defective in binding to the LDL receptor.2 4 5 Thus, the 3500 residue of apoB is clearly in an important site, and it is predictable that other similar functional defects may result when base pair changes occur nearby. Confirmatory evidence for the importance of the 3500 residue and its flanking regions comes from the use of monoclonal antibodies to specific regions of the apoB protein, as well as investigation of available epitopes when LDL is bound to its receptor.6 In an attempt to identify new functional mutations, we screened all patients attending the lipid clinic of Glasgow Royal Infirmary for changes in the coding sequences of apoB in a region of 206 bp surrounding the 3500 codon. DNA from 907 hyperlipidemic individuals was examined by amplification of this region by use of PCR and DGGE to detect any changes.
Blood samples were obtained from 907 unselected hyperlipidemic patients attending the lipid clinic of the Glasgow Royal Infirmary. These patients represented all referrals to the clinic for hyperlipidemia and were not prescreened because the phenotype of the mutations sought was as yet unknown. There were 48 patients with FH in the clinic (identified by total cholesterol greater than 7.5 mmol/L on two occasions and tendon xanthoma in the patient or first-degree relative). They were not excluded from the screening because it was considered possible that patients could have mutations in both apoB and the LDL receptor gene.
Blood was collected in tubes containing EDTA as anticoagulant. For PCR the DNA was prepared from 0.5 mL of fresh or frozen samples by use of salt precipitation of proteins after lysis of cell nuclei.7 For RFLPs, the DNA was prepared with 10 mL of frozen blood by lysis of the cell nuclei and phenol and chloroform extraction.8
The region of apoB including codon 3500 was amplified by use of the primers 5′-CGCCCGCCGCGCCCCGCGCCCGT-CCCGCCGCCCCCGCGAATATATGCGTTGGAGTGTGGC and 5′-GGAGCAGTTGACCACAAGCTTAGCTTGGAAA, which flank 206 bp of the sequence from nucleotide 10582 to nucleotide 10787.3 The first primer included a synthetic GC-rich tail of 38 nucleotides that acted as a GC clamp when the products were run on the gel. PCR conditions for the buffer, primer, and enzyme concentrations were as recommended by the manufacturer (Boehringer Mannheim). Thirty-five cycles, consisting of denaturation (95°C for 1 minute), annealing (67°C for 1 minute), and extension (72°C for 1.5 minutes), were performed. The initial denaturation step was carried out for 5 minutes. After the final extension step the products were heated to 98°C for 12 minutes to totally denature all products formed and then cooled slowly to room temperature. During this cooling, if the PCR products from the two apoB chromosomes differed, heteroduplex and homoduplex molecules were formed in approximately equimolar amounts.
The products were separated on an 8% polyacryamide gel at 60°C for 16 hours at 62.5 V. Gel dimensions were 16 cm×18 cm×1 mm. A Hoeffer apparatus was used with a custom-made lower buffer chamber designed in such a way that the buffer recycled and also acted as a thermostatic bath. At the top of the gel the concentration of urea was 15.7% (wt/vol) and that of formamide was 14% (vol/vol), and this denaturing gradient increased so that at the bottom of the gel the concentration of urea was 27% (wt/vol) and that of formamide was 24% (vol/vol).9 The double-stranded PCR products were electrophoresed into the urea and formamide gradient. While the GC-rich tail remained hybridized at all times, the remainder of the duplex at a given urea and formamide concentration became partially denatured so the movement of the molecule through the gel was virtually halted. The less stable heteroduplex molecules, if present, migrated less far than homoduplexes.
Genotypes of individuals and their family members were determined by RFLP analysis with the Southern blot technique.10 Three polymorphisms within the apoB gene were determined. The Xba I and Msp I RFLPs were detected with the PAB 3.5 C probe (a 3.5-kb EcoRI genomic fragment cloned in pUC 8) and the EcoRI RFLP was detected by use of a 2.0-kb HindIII genomic fragment cloned in pUC 8. In all cases, + indicates presence of the restriction site and − indicates absence of the site. For example, an individual with genotype Xba I+− is heterozygous for the Xba I site, and an individual with genotype Msp I++ is homozygous for the cutting allele of the Msp I site.
Cloning and Sequencing Strategies
The PCR product from the proband in the R family (I.3, Table 1⇓) was the first to be identified with an altered banding pattern on DGGE, distinct from that of classic FDB (Fig 1⇓). The PCR product was cloned by use of the T vector system supplied by Novagen (Northumbria Biologicals Ltd). A clone containing the mutation was identified by PCR with the same primers followed by DGGE. The clone was sequenced by use of the Sequenase Version 2.0 kit (United States Biochemical, supplied by Amersham Life Science) and the base pair change was identified. Molecular Biology Laboratory, Strathclyde University, sequenced the whole of the cloned insert with an automated sequencer (ABI 373A) and showed that there were no other base pair changes in the amplified segment. A separately isolated PCR product from the same subject (the proband in the R family) was sequenced directly, as were PCR products from other key individuals.
Lipid measurements were performed according to the standard Lipid Research Clinics Program protocol.11
U937 Cell Growth Assay for LDL Function
LDL cholesterol was prepared from normal individuals (n=9), individuals with FH (n=18), individuals with classic FDB (hereafter termed FDB3500Q) (n=9), and individuals with the new mutation FDB3500W (n=11). LDL from individuals with FH has normal binding properties, and because the subjects are hypercholesterolemic they are a good comparative group.12 A standard normal LDL (always from the same person) was included in each set of growth measurements.
Before the assay, actively growing U937 cells were washed in PBS and resuspended in lipid-deficient serum overnight. The next day they were counted, washed in PBS, and diluted to 1×105 cells/mL in serum-free medium. The cell suspension (150 μL/well) was used to inoculate a 96-well plate. Quantities of LDL cholesterol from 0 to 20 μg/mL were added to the wells.13 After 4 days of incubation, cell numbers were evaluated colorimetically with MTT.14 The greatest distinction in the ability to promote growth was observed when 5 μg/mL LDL cholesterol was provided. Thus, growth at 5 μg/mL was divided by the maximal growth, which occurred at 20 μg/mL for each individual, and to eliminate any variation due to the state of the cells each value was expressed relative to the standard normal in the same batch.
Screening With DGGE
Of the 907 hyperlipidemic patients screened, five with the classic FDB3500Q mutation were found. Two other patients also displayed a four-band pattern on DGGE, but the positions of the homoduplex and heteroduplex mutant bands were slightly different from those of the FDB3500Q mutation (Fig 1⇑). We initially termed this previously undescribed mutation FDBGlasgow. None of the individuals found with unusual band patterns on DGGE suffered from FH (see subject characteristics in Tables 1 through 3⇑⇓⇓). All patients with classic FDB3500Q and the members of the R family were of Scottish descent. Members of the S family were Asian.
The available relatives of the probands with the new mutation were screened, and the pedigrees are presented in Fig 2⇓.
Sequencing of the cloned PCR product from the proband, I.3, in family R revealed that the mutation in FDBGlasgow was in the first base pair in codon 3500. In normal individuals codon 3500, CGG, coded for arginine; in subjects with FDB3500Q CAG encoded glutamine; and in subjects with FDBGlasgow a new codon, TGG, coded for tryptophan (Fig 3⇓), resulting in the suggested designation FDB3500W for this new mutation. Direct sequencing of a fresh batch of the PCR product from the proband’s DNA confirmed this C-to-T change, and bands were visible in both C and T tracks. Confirmatory sequence data were also obtained from the PCR product from the affected daughter, II.5 (Table 1⇑).
The same C-to-T mutation was confirmed in the proband of the S family by sequencing of his PCR product.
Lipoprotein Concentrations in Family Members
The characteristics of R and S family members with and without the mutation are shown in Tables 1⇑ and 2⇑. Not all individuals with FDB3500W were hypercholesterolemic, and this was particularly true for the younger subjects. Lipid characteristics for the individuals identified during screening as carrying FDB3500Q are shown in Table 3⇑ for comparision with the two FDB3500W probands.
DNA analysis from individuals in the FDB3500W families indicated that the C-to-T mutation arose independently from the G-to-A mutation causing FDB3500Q. FDB3500W was associated with the haplotype Xba I+/Msp I−/EcoRI+ in the R family. In the S family, FDB3500W was associated with Xba I−/Msp I+/EcoRI+. Of the five FDB3500Q individuals who were also detected in the screening, four had genotypes compatible with the common Xba I−/Msp I+/EcoRI− FDB3500Q haplotype.15 The other individual, IS (Table 3⇑), had the genotype Xba I+−/Msp I++/EcoRI++. Analysis of DNA from two of his children showed that he carried the FDB3500Q mutation on a chromosome with the haplotype Xba I−/Msp I+/EcoRI+. This is the same as the haplotype of the apoB allele associated with FDB in a Chinese man described by Bersot et al16 and may have the same origin.
LDL Functional Assay
It has been established that the replacement of an arginine with glutamine at position 3500 in apoB adversely affects binding of LDL to its receptor. The cell line U937 has an absolute requirement for extracellular LDL cholesterol for growth, and FDB3500Q-derived LDL has been shown to be less efficient at promoting growth of these cells.12 13 LDL from individuals with either form of FDB 3500 and from their affected relatives was examined for its ability to promote growth of U937 cells (Fig 4⇓). LDL from normal individuals and from individuals with FH was used for comparison. The relative growth rates of LDL from subjects with FDB3500Q and FDB3500W were about half those of LDL from normal subjects or subjects with FH. No difference was detected between the results from 9 subjects with “classic” FDB (FDB3500Q) and 11 subjects with “new” FDB (FDB3500W). Although some of the subjects with FDB3500W did not express hyperlipidemia, their LDL cholesterol was invariably abnormal, as shown by the reduced growth of the U937 cells.
Many laboratories have sought mutations in the apoB gene that will help to confirm the location of the region important for binding to the LDL receptor. When the sequence of apoB was originally published, a putative receptor-binding region was defined,3 but it is clear that other regions of apoB are also important.6 17 The portion of apoB flanking the codon for amino acid 3500 is highly conserved18 ; however, by systematic screening, we have identified one new and important mutation. Other workers have reported apoB coding changes; for example a C-to-T mutation in exon 29, called apoB-100 Hopkins, has been found in one family, but the mutation was not linked to a hyperlipidemic phenotype.19 An FDB3531 arginine-to-cysteine change has been reported that appears to cause defective LDL binding (69.8±4% of normal activity) and moderate hypercholesterolemia.20 This is the next arginine downstream from amino acid 3500. Our primers would not have detected this mutation even if it were present in the Glasgow population.
This situation of alternative mutations in the same CGG codon is exactly analogous to that found in the sterol 27-hydroxylase gene. Codon 441 is normally CGG for arginine. However, two mutations that disrupt the function of the protein have been found: one is CAG for glutamine and the other is TGG for tryptophan.21 The different haplotype associated with FDB3500W in our two families indicated independent mutations of the CG dinucleotide to a TG, which is a relatively common human mutation.22 Alternative mutations in the same codon, both of which are clinically significant, are also found for Alzheimer’s disease, in which two different susceptibility mutations were found to map to the same codon 717 of the β-amyloid gene.23
FDB3500W is the result of a mutation in a base pair adjacent to the previously described FDB3500Q. The discovery of this independent mutation with the same phenotype as FDB3500Q suggests that it could be the loss of the arginine residue rather than the appearance of another amino acid that is causing the impaired ability to bind the LDL receptor. An alternative explanation is that the residue in position 3500 influences the conformation of the peptide chain, thereby affecting the proposed structure of the putative receptor-binding domain of apoB.6 The discovery of a new functional mutation provides a further opportunity to understand the molecular biology of apoB, its interaction with the LDL receptor, and its relevance to heart disease.
Note added in proof. A more detailed account of FDB3531 has been published. Pullinger CR, Hennessy LK, Chatterton JE, Liu W, Love JA, Mendel CM, Frost PH, Malloy MJ, Schumaker VN, Kane JP. Familial ligand-defective apolipoprotein B: identification of a new mutation that decreases LDL receptor binding affinity. J Clin Invest 1995;95:1225-1234.
Selected Abbreviations and Acronyms
|DGGE||=||denaturing gradient gel electrophoresis|
|PCR||=||polymerase chain reaction|
|RFLP(s)||=||restriction fragment length polymorphism(s)|
This study was supported by Greater Glasgow Health Board grant RSG/END/9293/z to Dr Gaffney. Thanks to Dr D. St. J. O’Reilly, Dr L.F. Squires, Dr M.J. Murphy, Dr J.H. McIlroy, and Dr P.J. Galloway from the Glasgow Royal Infirmary Lipid Clinic for patient samples and to Dr P. Wenham of the Western General Hospital, Edinburgh, for confirming all the FDB 3500Q individuals by use of mutagenic primers.
Soria LF, Ludwig EH, Clarke HRG, Vega GL, Grundy SM, McCarthy BJ. Association between a specific apolipoprotein B mutation and familial defective apolipoprotein B-100. Proc Natl Acad Sci U S A. 1989;86:587-591.
Knott TJ, Wallis SC, Powell LM, Pease RJ, Lusis AJ, Blackhart B, McCarthy BJ, Mahley RW, Levy-Wilson B, Scott J. Complete cDNA and derived protein sequence of human apolipoprotein B-100. Nucleic Acids Res. 1986;14:7501-7503.
Innerarity TL, Mahley RW, Weisgraber KH, Bersot TP, Krauss RM, Vega GR, Grundy SM, Friedl W, Davignon J, McCarthy BJ. Familial defective apolipoprotein B-100: a mutation of apolipoprotein B that causes hypercholesterolaemia. J Lipid Res. 1990;31:1337-1349.
Marz W, Baumstark MW, Scharnagl H, Ruzicka V, Buxbaum S, Herwig J, Pohl T, Russ A, Schaaf L, Berg A, Bohles H-J, Usadel KH, Gross W. Accumulation of “small dense” low density lipoproteins (LDL) in a homozygous patient with familial defective apolipoprotein B-100 results from heterogeneous interaction of LDL subfractions with the LDL receptor. J Clin Invest. 1993;92:2922-2933.
Milne R, Theolis R Jr, Maurice R, Pease RJ, Weech PK, Rassart E, Fruchart J-C, Scott J, Marcel YL. The use of monoclonal antibodies to localize the low density lipoprotein receptor-binding domain of apolipoprotein B. J Biol Chem. 1989;33:19754-19760.
Gaffney D, Campbell RA. A PCR based method to determine the Kalow allele of the cholinesterase gene: the E1k allele frequency and its significance in the normal population. J Med Genet. 1994;31:248-250.
Kunkel LM, Smith KD, Boyer SH, Borgaonkar DS, Wachtel SS, Miller OJ, Breg WR, Jones HW Jr, Rary JM. Analysis of human Y-chromosome-specific reiterated DNA in chromosome variants. Proc Natl Acad Sci U S A. 1977;74:1245-1249.
Barni N, Talmud PJ, Carlsson P, Azoulay M, Darnfors C, Harding D, Weil D, Grzeschik KH, Bjursell G, Junien C, Williamson R, Humphries SE. The isolation of genomic recombinants for the human apolipoprotein B gene and the mapping of three common DNA polymorphisms of the gene: a useful marker for human chromosome 2. Hum Genet. 1986;73:313-319.
Lipid Research Clinics Program. Lipid and Lipoprotein Analysis: Manual of Laboratory Operations, Volume 1. Bethesda, Md: 1974. US Dept of Health, Education, and Welfare publication NIH 75-628.
Frostegard J, Hamsten A, Gidlund M, Nilsson J. Low density lipoprotein-induced growth of U937 cells: a novel method to determine the receptor binding of low density lipoprotein. J Lipid Res. 1990;31:37-44.
Bersot TP, Russell SJ, Thatcher SR, Pomernacki NK, Mahley RW, Weisgraber KH, Innerarity TL, Fox CS. A unique haplotype of the apolipoprotein B-100 allele associated with familial defective apolipoprotein B-100 in a Chinese man discovered during a study of the prevalence of this disorder. J Lipid Res. 1993;34:1149-1154.
Dunning AM, Houlston R, Frostegard J, Revill J, Nilsson J, Hamsten A, Talmud P, Humphries S. Genetic evidence that the putative receptor binding domain of apolipoprotein B (residues 3130 to 3630) is not the only region of the protein involved in interaction with the low density lipoprotein receptor. Biochim Biophys Acta. 1991;1096:231-237.
Avoustin P, Mostachi H, Perret B, Cambou JP, Cambien F, de Preval C. A very conservative region of apoB-100 in the putative binding region to the LDL receptor in the Toulouse population. Hum Genet. 1993;90:460-463.
Ladias JAA, Kwiterovich PO Jr, Smith HH, Miller M, Bachorik PS, Forte T, Lusis AJ, Antonarakis SE. Apolipoprotein B-100 Hopkins (arginine 4019→tryptophan): a new apolipoprotein B-100 variant in a family with premature atherosclerosis and hyperapobetalipoproteinemia. JAMA. 1989;262:1980-1988.
Pullinger CR, Hennessy LK, Love JA, Frost PH, Mendel CM, Weiqun L, Malloy MJ, Kane JP. Familial ligand-defective apolipoprotein B: identification of a new mutation that decreases LDL receptor binding affinity. Circulation. 1993;88(suppl I):I-322. Abstract.
Kim KS, Kubota S, Kuriyama M, Fujiyama J, Bjorkhem I, Eggertsen G, Seyama Y. Identification of new mutations in sterol 27-hydroxylase gene in Japanese patients with cerebrotendinous xanthomatosis (CTX). J Lipid Res. 1994;35:1031-1039.