The 1298A→C polymorphism in methylenetetrahydrofolate reductase (MTHFR): in vitro expression and association with homocysteine

Atherosclerosis 156 (2001) 409– 415 www.elsevier.com/locate/atherosclerosis

The 1298A“ C polymorphism in methylenetetrahydrofolate reductase (MTHFR): in vitro expression and association with homocysteine Ilan S. Weisberg a, Paul F. Jacques b, Jacob Selhub b, Andrew G. Bostom c, Zhoutao Chen a, R. Curtis Ellison d, John H. Eckfeldt e, Rima Rozen a,* a

Departments of Human Genetics and Pediatrics, McGill Uni6ersity Health Centre, Montreal, Canada Jean Mayer USDA Human Nutrition Research Center on Aging, Tufts Uni6ersity, Boston, MA, USA c Di6ision of General Internal Medicine, Memorial Hospital of Rhode Island, Pro6idence, CT, USA d NHLBI Family Heart Study Field Center, Framingham, MA and Boston Uni6ersity School of Medicine, Boston, MA, USA e NHLBI Family Heart Study Central Laboratory, Department of Laboratory Medicine and Pathology, Uni6ersity of Minnesota, Minnesota, MN, USA b

Received 22 March 2000; received in revised form 4 July 2000; accepted 18 August 2000

Abstract A common mutation in methylenetetrahydrofolate reductase (MTHFR), 677C “ T, is associated with reduced enzyme activity, a thermolabile enzyme and mild hyperhomocysteinemia, a risk factor for vascular disease. Recently, a second common mutation (1298A “C; glutamate to alanine) was reported, but this mutation was suggested to increase homocysteine only in individuals who carried the bp677 variant. To evaluate the functional consequences of this mutation, we performed site-directed mutagenesis and in vitro expression. For in vivo assessment of clinical impact, we examined the 1298A “ C genotypes and plasma homocysteine in 198 individuals from the NHLBI Family Heart Study that had previously been assessed for the 677 substitution. Site-directed mutagenesis of the human cDNA was performed to generate enzymes containing each of the two mutations, as well as an enzyme containing both substitutions. Enzyme activity and thermolability were assessed in bacterial extracts. The activity of the wild-type cDNA was designated as 100%; mutant enzymes containing the 1298 and 677 mutations separately had 68% (9 5.0) and 45% ( 910.8), respectively, of control activity while the enzyme containing both mutations had 41% ( 912.8) of control activity. The 1298 mutation was not associated with a thermolabile enzyme. In the Family Heart Study, fasting homocysteine was significantly higher (PB 0.05) in individuals heterozygous for both substitutions, compared to individuals who carried only the 677C “T variant. This study suggests that two variants in MTHFR should be assessed as genetic risk factors for hyperhomocysteinemia. © 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Homocysteine; Folate; Genetics; Enzymes; Methylenetetrahydrofolate reductase

1. Introduction Numerous studies have demonstrated an elevation of total plasma homocysteine (tHcy) in patients with vascular disease and have suggested that mild hyperhomocysteinemia is an independent risk factor for several vasculopathies including atherosclerosis, acute myocar* Corresponding author. Montreal Children’s Hospital, 4060 St. Catherine West, Montreal, Quebec H3Z 2Z3, Canada. Tel.: +1-5149344400; fax: + 1-514-9344331. E-mail address: [email protected] (R. Rozen).

dial infarction, cerebrovascular disease and carotid artery stenosis [1–4]. Disruptions in homocysteine metabolism, due to genetic or nutritional deficiencies, result in hyperhomocysteinemia. 5-Methyltetrahydrofolate, a carbon donor in homocysteine remethylation to methionine, is synthesized from 5,10- methylenetetrahydrofolate by methylenetetrahydrofolate reductase (MTHFR). Kang et al. described a thermolabile variant of MTHFR that had reduced enzyme activity and was associated with increased tHcy [5]. We isolated the human MTHFR cDNA [6] and described a common

0021-9150/01/$ - see front matter © 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 0 2 1 - 9 1 5 0 ( 0 0 ) 0 0 6 7 1 - 7

410

I.S. Weisberg et al. / Atherosclerosis 156 (2001) 409–415

alanine to valine substitution (677C“ T) which is present in the homozygous state in 10 – 15% of North Americans. This mutation has been expressed in vitro and shown to encode the thermolabile variant [7]. In studies of lymphocyte extracts, individuals homozygous for the V allele displayed decreased enzyme activity, increased enzyme thermolability and increased plasma homocysteine [7,8]. However, the fasting homocysteine concentrations in mutant individuals are dependent upon folate status; individuals homozygous for this mutation were hyperhomocysteinemic only when folate levels were in the lower end of the normal range [9]. Recent data suggest that the protective effect of folate may be mediated by enzyme stabilization [10]. These findings illustrate the interactive effect between folate and MTHFR genotype, and highlight the multifactorial nature of hyperhomocysteinemia. Hyperhomocysteinemia has also been observed in families with neural tube defects. Consequently, the 677C“T variant has been reported to increase the risk for neural tube defects [11,12], as well as for vascular disease [13–15]. To investigate the contribution of other MTHFR mutations to the development of hyperhomocysteinemia, we recently studied a second common variant (1298A“C; glutamate to alanine) in children with spina bifida and in their mothers, as well as in control mother–child pairs [16]. Homozygotes for this variant ( 10% of individuals) had a significant decrease in activity in lymphocyte extracts (60% of control) but this decrease did not appear to be sufficient to affect plasma homocysteine. Heterozygotes had activities that were intermediate between those of controls and homozygous mutants, and also showed normal homocysteine levels. Similar findings were observed in a Dutch study of spina bifida [17]. Homozygotes for the two mutations at bp677 and 1298 have not been reported and only one individual has been described to carry the two variants on the same allele, after investigation of over 1000 individuals [16,17]. However, individuals who are heterozygous for the two variants,  15 – 20% of the general population, have decreased activity in lymphocytes, at 50 – 60% of control values, and could be at risk for hyperhomocysteinemia. The intent of this study was to investigate the effect of the 1298A“C mutation on enzyme function, by site-directed mutagenesis and expression in vitro. We constructed all four possible combinations of the two polymorphisms for evaluation of activity and thermolability. For additional in vivo assessment of this variant, we examined the 1298A“C polymorphism and plasma homocysteine levels in a group of 198 individuals from the NHLBI Family Heart Study. We specifically selected individuals previously confirmed to be heterozygous for the 677C“T variant, since individuals heterozygous or homozygous for the 1298A“ C

variant alone have normal homocysteine levels, and homozygotes for both variants have not been observed. Our results demonstrate good correlation between enzyme activity in vitro and homocysteine levels in vivo, supporting the important role of MTHFR in modifying homocysteine levels.

2. Methods

2.1. Design of constructs Site-directed mutagenesis was used to create cDNAs containing all four possible combinations of the polymorphic variants, designated by the amino acid codons (AE = wild-type, alanine for 677C and glutamate for 1298A; AA= single mutant, alanine for 677C and alanine for 1298A; VE=single mutant, valine for 677T and glutamate for 1298C; VA= double mutant, valine for 677T and alanine for 1298C). Since the original cDNA that we isolated contained the mutant allele for the 1298A“ C polymorphism, we used PCR-based mutagenesis [18] to create the wild type allele. Two sets of primers were designed to generate independent PCR products. Primer 1 is a sense primer complimentary to MTHFR sequences [6] (bp 1082–1101) located 5% of an EagI site. Primer 4 is an antisense primer complimentary to sequences (bp 308 to 327) located 3% to the XbaI site in the pTrc99A plasmid polylinker (Pharmacia). Primers 2 and 3, the sense and antisense mutagenesis primers, were identical to the published human cDNA sequence, except for position 1298 where they contained a single mismatch corresponding to the wild type allele of the 1298A“ C polymorphism. Using Vent Polymerase (NEB) to eliminate incidental mutations, two separate PCR reactions were performed. Amplification with primers 1 and 2, and a separate reaction with primers 3 and 4, generated two distinct PCR fragments containing the mutagenized allele at position 1298. These fragments served as templates for a third PCR reaction, with primers 1 and 4, to amplify one fragment of 1.2 kb. The product was digested with EagI I and XbaI (NEB), and the resultant 1 kb mutagenized cassette was subcloned into the pTrc99A vector containing the cDNA. Mutagenesis was performed for both alanine and valine backgrounds of the 677C“ T polymorphism, which were created in Frosst et al. [7]. Constructs were confirmed by sequencing; the 677C“ T and 1298A“ C polymorphisms were also confirmed by PCR and HinfI or MboII digestion, respectively. The 4 cDNAs were then subcloned into the His-tagged pET23d(+ ) expression vector (Novagen). To generate the MTHFR –His fusion protein, the endogenous MTHFR stop codon had to be removed. Since a convenient restriction site 5% to the stop codon was not available to facilitate direct subcloning of the entire

I.S. Weisberg et al. / Atherosclerosis 156 (2001) 409–415

cDNA into the pET vector, a two-step approach was used to create the final construct. Following digestion with NcoI and BamHI, 1.7 kb of the wild type (AE) cDNA was subcloned into the multiple cloning site of the pET plasmid generating an intermediate vector pETNB. Using PCR to introduce an artificial XhoI site, the remaining 250 bases of sequence coding for the 3% end of the cDNA minus the stop codon were then shuttled into the pETNB vector, thereby creating the full length MTHFR–His fusion protein. The remaining cDNA combinations of the two polymorphisms were generating by digesting with BamHI and EcoRI and subcloning into the wildtype His fusion construct. Sequence analysis of amplified regions and cloning junctions was performed for confirmation of sequence.

2.2. Expression analysis Competent BL21 E. coli bacterial cells (Novagen) were transformed by heat shock and selected on carbenecillin-containing agar plates. Fresh LB broth was inoculated with 10 ml of the overnight culture until the OD600 reached 0.6–0.8. Cells were induced with IPTG (100 ml of 100 mM stock) and grown overnight at 20°C. Bacteria were pelleted, resuspended in protease-containing buffer, and sonicated. Extracts were centrifuged and supernatant protein was used to measure activity, using a modification of Erbe and Rosenblatt [9]. Ninety mg protein were incubated for 60 min, 37°C with 50 ml 1 M phosphate buffer (pH=6.3), 40 ml 0.025 M menadione bisulfate, 4 ml 0.05 M ascorbic acid, 10 ml 0.81 × 10 − 3 M FAD, 4 ml 0.1 M EDTA (pH = 6.3), and 16 ml [14C]CH3THF (1017.8 cpm/nmol) in a total volume of 279 ml. The reaction was terminated with 250 ml NaOAc (0.6 M, pH =4.5), 150 ml 0.4 M dimedone, and 100 ml 0.1 M formaldehyde. The mixture was boiled for 12 min and chilled. After addition of 2.5 ml toluene, samples were vigorously vortexed and centrifuged at 2000 rpm for 10 min. Scintillation counting of supernatant was used to quantitate [14C] formaldehyde– dimedone adducts. Activity was expressed as nmol formaldehyde/h/mg protein. To determine thermolability, reaction mixtures containing protein extracts were incubated at 46°C for 5 min. After chilling on ice, substrate and FAD were added and the assay proceeded as above. Thermolability was designated as percentage of residual activity obtained with unheated samples.

2.3. Western blotting Immunoblotting was performed with His-probe (H15), a rabbit polyclonal antibody raised against the penta-His domains of pET vectors (Santa Cruz Biotechnology). Protein was transferred to a Hybond™ ECL™ (Amersham) membrane and blocked overnight.

411

The membrane was incubated with primary antibody (1 h, room temperature), washed with PBS, and incubated with secondary antibody (arabbit Ig horseradish peroxidase linked antibody) for 20 min. The ECL™ Western Blotting Detection Kit (Amersham) was used for protein visualization.

2.4. Subjects Subjects were participants in the NHLBI Family Heart Study (FHS). The FHS was established to evaluate genetic and non-genetic determinants of coronary heart disease (CHD), preclinical atherosclerosis, and cardiovascular disease risk factors in individuals and families. Probands were recruited from three existing population-based studies: the Framingham Heart Study, the Utah Health Family Tree Study, and two cohorts of the Atherosclerosis Risk in Communities (ARIC) Study. The FHS was divided into two phases. In phase 1, family and personal histories from probands were used to characterize families with respect to history of CHD and determine the extent to which familial aggregation of known risk factors accounted for familial clustering of CHD. In phase 2, 657 families known to be at highest risk for CHD and 588 randomly-selected families were offered clinical examinations and laboratory tests. Probands and their parents, siblings, spouses, and adult offspring were included. Subject examinations began in February 1994 and continued through December 1995. A detailed description of study design is provided elsewhere [19]. Two FHS centers, Framingham and Utah, participated in an ancillary homocysteine study. All persons at these sites undergoing the complete FHS Phase II evaluation were invited to participate in the ancillary project, with the following exclusions: age B 25 or fasting for B 10 h. This homocysteine study protocol was approved by the FHS Steering Committee and Safety Monitoring Board, and the institutional review boards for Boston University School of Medicine, University of Utah, and Tufts-New England Medical Center. The present analyses include 198 subjects from the ancillary homocysteine study who were heterozygotes for the MTHFR 677C“ T polymorphism.

2.5. Fasting blood collection Immediately upon arrival, subjects underwent a fasting (\10 h) phlebotomy. Plasma was promptly separated, and aliquots were stored at − 70°C. Fasting samples were used for plasma concentrations of total homocysteine (tHcy), folate, vitamin B12, and pyridoxal-5%-phosphate (PLP; the active circulating form of vitamin B6).

412

I.S. Weisberg et al. / Atherosclerosis 156 (2001) 409–415

2.6. Determination of homocysteine and 6itamins Plasma homocysteine was determined by HPLC with fluorometric detection [20], plasma folate and vitamin B12 by radioassay (Biorad Quantaphase II, Hercules, CA), and plasma PLP by the tyrosine decarboxylase apoenzyme method [21]. Coefficients of variation were 6% for homocysteine, 9% for folate, 11% for PLP, and 11% for vitamin B12. Methods for serum creatinine determinations, performed as part of the main FHS study, are described elsewhere [22].

2.7. Genotype determination Although the 1298A“ C mutation abolishes an MboII site, it does not provide a suitable diagnostic test for the polymorphism. This is due to the presence of another silent mutation, 1317T“C that results in an identical restriction pattern upon MboII digestion [16]. Consequently, we created an artificial restriction site to distinguish between these two mutations, as reported [16]. Detection of the 1298A“C mutation was accomplished using the sense primer 5%-GGGAGGAGCTGACCAGTGCAG-3% and the antisense primer 5%-GGGGTCAGGCCAGGGGCAG-3% followed by Fnu4HI digestion. This generated a 138 bp PCR fragment, which is cleaved into 119 bp and 19 bp fragments in the presence of the C allele.

family structure was ignored. Only results from the latter analyses are presented. If not otherwise noted, statistical significance refers to PB 0.05.

3. Results

3.1. In 6itro analysis MTHFR cDNAs containing the four possible combinations of the two polymorphisms (AE, AA, VE, VA) were expressed, and specific activity was determined (Fig. 1(A)). Activity was assayed in duplicate for each construct, in five separate experiments. Activity of the wild-type cDNA (AE) was designated as 100%, and all

2.8. Statistical methods Because fasting tHcy concentrations were skewed, we applied logarithmic transformation to these data. Therefore, geometric means were used to summarize homocysteine concentrations. Analyses were performed using SAS [23]. Geometric mean homocysteine concentrations and 95% confidence intervals were estimated in the two 1298A“ C genotype categories (AA, AC) using the PROC GLM LSMEANS procedure. Additional analyses also considered the relation between 1298A“ C genotype and factors known to influence homocysteine concentrations, including age, sex, vitamin supplement use, plasma concentrations of folate, vitamin B12, and PLP, and serum creatinine concentrations. Concentrations of the plasma vitamins and serum creatinine were also skewed and, therefore, logarithmically transformed. We also performed multivariable analyses to adjust the homocysteine differences between genotype groups for age, sex, examination site, and other factors that might also influence homocysteine interactions. In addition, we considered potential interactions (i.e. joint influence) of these vitamins and genotype on homocysteine concentrations. These analyses were also repeated using SAS PROC MIXED to account for the inclusion of genetically related individuals in the FHS. Results were similar to those in which

Fig. 1. Relative activities of four constructs designated by the amino acids of the two polymorphisms: AE, AA, VE and VA. The values, % of wild-type, are mean 9standard error of five experiments, performed in duplicate. (b) Western blot of four constructs. The pET vector without insert was used as a negative control. Protein extracts were incubated with rabbit polyclonal antibody against the penta-His tag that is fused with the cDNA. Western blotting was performed for four different experiments; scanning of the blot was used to determine the amount of MTHFR protein that was used for enzyme assays, as depicted in (a).

I.S. Weisberg et al. / Atherosclerosis 156 (2001) 409–415 Table 1 Characteristics of the FHS participants by MTHFR 1298A “ C genotype. All individuals were heterozygous for the 677C “ T genotype 1298A “C Genotype

N Age (years) Women (%) Fasting total homocysteine (mmol/l)b Plasma folate (nmol/l)b Plasma PLP (nmol/l)b Plasma vitamin B12 (pmol/l)b Serum creatinine (mmol/l)b Regular vitamin supplement use (%)

AA

AC

93 51.3a (48.5–54.0) 50.5 (40.4–60.7) 8.0 (7.5–8.5)

105 49.7 (47.1-52.3) 58.1 (48.6–67.6) 8.9c (8.4–9.4)

11.2 (9.7–13.0) 64.8 (57.1–73.6) 294 (270–320) 83 (80–86) 45.2 (35.1–55.2)

9.5 (8.3–10.9) 53.7c (47.7–60.5) 260c (240–281) 82 (79–85) 37.5 (28.0–47.0)

a

Mean or prevalence (95% CI). Geometric mean. c Means for genotype groups are significantly different (PB0.05). b

other constructs were expressed as a percentage of this value. The relative activities (mean9standard error) for the AA, VE, and VA constructs were 68% (9 5.0), 45% (9 10.8), and 41% (912.8), respectively. These values represent the mean of five independent trials. Western blotting (Fig. 1(B)) was performed to ensure that similar amounts of the His– MTHFR fusion protein were used in the enzyme assays. Therefore it is unlikely that the differences in activity are due to different transfection efficiencies. These in 6itro results are similar to the earlier published observations from lymphocyte extracts [16,17]. MTHFR activity was decreased by 30–40% in individuals homozygous for the 1298A “ C mutation, whereas a 60–70% loss of activity was associated with homozygosity for the 677C“ T polymorphism. These data indicate that the mutant valine allele of the 677 polymorphism is more deleterious than the mutant alanine allele of the 1298 polymorphism. The wild-type enzyme (AE) and the enzyme mutant for the 1298 polymorphism (AA) retained similar activities after heating (40– 50%). However, the two valinecontaining constructs (VE and VA) retained only  10 –12% residual activity after heating. We conclude, therefore, that the enzyme is not rendered thermolabile when mutant for the 1298A“C polymorphism, in contrast to the 677C“ T polymorphism.

3.2. MTHFR genotypes and plasma homocysteine Table 1 outlines the FHS results. Since all individuals had been pre-selected to be carriers of the 677C“ T mutation, there were no homozygotes for the 1298A“ C mutation, as expected if the two mutations do not

413

occur on the same allele. Fasting homocysteine levels (mmol/l) were significantly higher (PB 0.05) in individuals heterozygous for the 1298 A“ C polymorphism, compared to individuals without this mutation (8.9 versus 8.0). AC individuals displayed lower values for plasma vitamin B12 and plasma pyridoxal-5%-phosphate (PB 0.05). Serum creatinine concentrations were similar in the two groups. A slightly larger proportion of those with the AA genotype were vitamin users, but the difference was not significant. However, the differences between genotype groups for plasma vitamin B12 and plasma pyridoxal-5%-phosphate were largely accounted for by this difference in supplement use since these differences were attenuated and no longer statistically significant after adjusting for vitamin supplement use. Adjustment for age, sex, examination site, folate, PLP, vitamin B12 and creatinine did not affect the difference in geometric mean homocysteine concentrations between persons with and without the1298A“C mutation (Table 2). After adjustment, the mean homocysteine concentration was 8.0 mmol/l (95% CI, 7.7– 8.4) for those without the mutation and 8.7 mmol/l (95% CI, 8.4–9.1) for the heterozygotes (PB0.02). Additional adjustment for vitamin supplement use did not affect the observed difference between genotype groups. Table 2 Multivariable adjusteda geometric mean fasting total homocysteine concentration in the Family Heart Study participants by methylenetetrahydrofolate reductase (MTHFR) 1298A “ C genotype and plasma folate status 1298A “ C Genotype

P-valueb

AA

AC

93 8.0 (7.7–8.4)

105 8.7 (8.4–9.1)

0.02

Plasma folate\9 nmol/ld N 51 Geometric mean 7.0 95% CI (6.6–7.5)

48 7.5 (7.2–8.0)

0.12

Plasma folate B9 nmol/l N 42 Geometric mean 9.1 95% CI (8.5–9.8)

57 10.1 (9.5–10.8)

0.04

Full samplec N Geometric mean 95% CI

a

Adjusted for age, sex, examination site, serum creatinine concentrations, and plasma concentrations of pyridoxal phosphate, vitamin B12, and folate. b Test for difference in homocysteine between genotype groups. c All individuals were heterozygous for the MTHFR 677C “ T polymorphism. d The interaction between plasma folate concentrations and the MTHFR 1298A“ C mutation was not statistically significant (P= 0.48).

414

I.S. Weisberg et al. / Atherosclerosis 156 (2001) 409–415

We found no statistical evidence of an interaction between plasma folate and the1298A“ C mutation (P for interaction=0.48) implying that the relation between the polymorphism and homocysteine concentrations was not influenced by folate. However, the stratified analyses suggest that the relation between homocysteine and the1298A“C mutation is stronger in those with folate concentrations below the median folate value than it is among those with folate concentrations greater than the median (Table 2).

4. Discussion Mild hyperhomocysteinemia has been implicated in many studies as a risk factor for cardiovascular disease [1– 4]. Inadequate intake of B vitamins clearly contributes to elevated homocysteine. An important genetic determinant of hyperhomocysteinemia is a common missense mutation (677C“T) in MTHFR [6 – 8]. A second polymorphism (1298A“C) was recently reported in families with spina bifida and shown to be associated with decreased activity in lymphocytes [16,17]. In this study, we directly assessed the effect of this polymorphism on enzyme activity and thermolability by expressing wild-type and mutant enzymes in vitro. We confirmed that both polymorphisms decrease activity, but to different degrees. The activity observed in vitro for the 1298A“C polymorphism ( 65% of control) is quite similar to that previously observed in lymphocyte extracts of homozygous mutant individuals [16,17]. This activity is clearly higher than that observed with the 677C“T mutation ( 40% of control), suggesting that homocysteine metabolism might not be significantly disrupted by the 1298A“ C polymorphism alone. In contrast to the 677C“ T variant, the 1298A“ C mutation does not result in a thermolabile enzyme in vitro. This result is similar to that seen in lymphocyte activities measured in vivo. In the first two studies which examine homocysteine levels in individuals with the 1298A“C change [16,17], homocysteine levels for heterozygotes and homozygotes were not different from those with the wild type genotype (1298AA), supporting the idea that this polymorphism alone might not significantly affect homocysteine metabolism. However, since individuals heterozygous for both the 1298A“ C and 677C“ T mutation had somewhat lower activities in lymphocytes (50– 60% of control) than single 677C“T heterozygotes, the possibility existed that double heterozygotes could have elevated homocysteine. These individuals should represent  15–20% of the general population. The NTD study by van der Put et al. [17] demonstrated increased homocysteine levels for double heterozygotes while our NTD study group, with a smaller sample, did not manifest

this increase. We therefore decided to examine a large series of double heterozygotes to evaluate plasma homocysteine. We had previously reported 677C“T genotypes in a group of 365 individuals from the Family Heart Study [9] and have recently increased our sample size. We specifically selected those that were heterozygous for this variant for the current study. Our data demonstrate that double heterozygotes have higher fasting homocysteine (8.9 mM) than single heterozygotes for the 677C“ T variant (8.0 mM). In our earlier study, individuals who were heterozygous for the 677C“ T mutation alone did not have higher homocysteine values than wild-type individuals. It has been suggested that the two polymorphisms do not appear in cis due to selection against a severe phenotype [17]. Our results with the two polymorphisms expressed in cis in vitro, and the fact that other severe MTHFR mutations have already been reported in cis with the 677C“ T variant [24], seem to refute this assumption. More likely, the two mutations occurred independently and have not had enough time for a recombination event to bring them together on the same chromosome. It is somewhat surprising that the alanine to valine substitution (two non-polar residues) at bp677 could have a profound impact on protein stability and activity, while the replacement of the acidic glutamate by the non-polar alanine, at bp1298, reduces enzyme activity and stability to a much lesser extent. However, according to main chain folding angles, glutamate and alanine are actually the two most structurally similar residues [25]. In addition, the distinct locations of the two polymorphisms could explain the different consequences. The 677 mutation is in the region encoding the N-terminal catalytic domain; the 1298 substitution is found in the C-terminal regulatory domain. The recently-determined structure of bacterial MTHFR suggests that the alanine to valine substitution could indirectly affect FAD binding and destabilize the quaternary structure [10]. Since bacterial MTHFR does not contain the C-terminal regulatory domain, structural predictions of the glutamate to alanine substitution cannot be made from the bacterial model. The 1298A“C polymorphism alone does not affect plasma homocysteine but may have a modest effect when present with the 677C“ T variant. Although we did not identify a significant interaction between folate level and the 1298 genotype in influencing homocysteine levels, the effect of the genotype appeared to be stronger in those with folate levels below the median. An increased risk for heart disease has not yet been reported for this polymorphism. However, since there appears to be a graded rather than a threshold effect of homocysteine on vascular damage, every contribution to homocysteine levels may be significant [2– 4]. This is especially true when folate levels are low or when folate

I.S. Weisberg et al. / Atherosclerosis 156 (2001) 409–415

requirements are increased (during pregnancy, for example). In summary, our data suggest that the overall picture of the causes of hyperhomocysteinemia should include not only nutritional factors, but also two common genetic polymorphisms in MTHFR.

[9]

[10]

Acknowledgements We thank Daniel Suh for technical assistance. This work was supported by the Medical Research Council of Canada and the NIH/NHLBI grant number 533K06-01 and contract number NO1-HC-25106 (The Family Genetics Studies of Cardiovascular Disease) and the US Department of Agriculture, under agreement No. 58-1950-9-001. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the US Department of Agriculture. I. Weisberg is a recipient of the Montreal Children’s Hospital Studentship Award.

[11]

[12]

[13]

[14]

[15] [16]

References [1] Clarke R, Daly L, Robinson, et al. Hyperhomocysteinemia: an independent risk factor for vascular disease. N Engl J Med 1991;324:1149 – 55. [2] Refsum H, Ueland PM, Nygard O, Vollset SE. Homocysteine and vascular disease. Ann Rev Med 1998;48:31 –62. [3] Boushey CJ, Beresford SAA, Omenn GS, Motulsky AG. A quantitative assessment of plasma homocysteine as a risk factor for vascular disease. Probable benefits of increasing folic acid intakes. JAMA 1995;274:1049 –57. [4] Selhub J, Jacques PF, Bostom AG, et al. Association between plasma homocysteine concentrations and extra-cranial carotid artery stenosis. N Engl J Med 1995;332:286 –91. [5] Kang SS, Wong PWK, Susmano A, Sora J, Norusis M, Ruggie N. Thermolabile methylenetetrahydrofolate reductase: an inherited risk factor for coronary artery disease. Am J Hum Genet 1991;48:536 – 45. [6] Goyette P, Sumner JS, Milos R, et al. Human methylenetetrahydrofolate reductase: isolation of cDNA, mapping and mutation identification. Nat Genet 1994;7:195 – 200. [7] Frosst P, Blom HJ, Milos R, et al. A candidate genetic risk factor for vascular disease: a common mutation in methylenetetrahydrofolate reductase. Nat Genet 1995;10:111 –3. [8] Christensen B, Frosst P, Lussier-Cacan S, et al. Correlation of a common mutation in the methylenetetrahydrofolate reductase gene with plasma homocysteine in patients with premature coro-

.

[17]

[18]

[19] [20]

[21]

[22] [23] [24]

[25]

415

nary artery disease. Arterio Thromb Vasc Biol 1997;17:569 – 73. Jacques PF, Bostom AG, Williams RR, et al. Relation between folate status, a common mutation in methylenetetrahydrofolate reductase, and plasma homocysteine concentrations. Circulation 1996;93:7 – 9. Guenther BD, Sheppard CA, Tran P, Rozen R, Matthews RG, Ludwig ML. The structure and properties of methylenetetrahydrofolate reductase from Escherichia coli: a model for the role of folate in ameliorating hyperhomocysteinemia in humans. Nat Struct Biol 1999;6:359 – 65. van der Put N, Steeger-Theunissen R, Frosst P, et al. Mutated methylenetetrahydrofolate reductase as a risk factor for spina bifida. Lancet 1995;346:1070 – 1. Whitehead AS, Gallagher P, Mills JL, et al. A genetic defect in 5,10 methylenetetrahydrofolate reductase in neural tube defects. Q J Med 1988;88:763 – 6. Kluitjmans LAJ, van den Heuvel LP, Boers GHJ, et al. Molecular genetic analysis in mild hyperhomocysteinemia: A common mutation in methylenetetrahydrofolate reductase gene is a risk factor for cardiovascular disease. Am J Hum Genet 1996;58:35 – 41. Morita H, Taguchi J, Kurihara H, et al. Genetic polymorphism of 5,10 methylenetetrahydrofolate reductase (MTHFR) as a risk factor for coronary artery disease. Circulation 1997;95:2032 –6. Fletcher O, Kessling AM. MTHFR association with arteriosclerotic vascular disease? Hum Genet 1998;103:11 – 21. Weisberg I, Tran P, Christensen B, Sibani S, Rozen R. A second genetic polymorphism in methylenetetrahydrofolate reductase (MTHFR) associated with decreased enzyme activity. Mol Gen Metab 1998;64:169 – 72. van der Put NMJ, Fons G, Stevens EMB, et al. A second common mutation in the methylenetetra-hydrofolate reductase gene: An additional risk factor for neural-tube defects? Am J Hum Genet 1998;62:1044 – 51. Cormack B. Directed mutagenesis using the polymerase chain reaction, Current Protocols in Molecular Biology 1997;8.5.1 – 8.5.10. Higgins M, Province M, Heiss G, et al. NHLBI family heart study: objectives and design. Am J Epidemiol 1996;143:1219 –28. Araki A, Sako Y. Determination of free and total homocysteine in human plasma by high-performance liquid chromatography with fluorescence detection. J Chromatogr 1987;422:43 –52. Shin-Buehring Y, Rasshofer R, Endres W. A new enzymatic method for pyridoxal-5%-phosphate determination. J Inherit Metab Disorders 1981;4:123 – 4. NHLBI Family Heart Study: manuals of procedure. St. Louis, MO: Washington University, 1993. SAS User’s Guide. 6th Edition. Cary, NC: SAS Institute, 1989. Goyette P, Christensen B, Rosenblatt DS, Rozen R. Severe and mild mutations in cis for the methylenetetrahydrofolate reductase (MTHFR) gene, and description of five novel mutations in MTHFR. Am J Hum Genet 1996;59:1268 – 75. Niefind K, Schomburg D. Amino acid similarity coefficients for protein modeling and sequence alignment derived from mainchain folding angles. J Mol Biol 1991;219:481 – 97.