Historically, the investigation of nature's mistakes in the form of so-called inborn errors of metabolism has attracted wide and justifiable interest (1). Some of these errors, such as phenylketonuria, are important in their own right; although rare, they are treatable and thus are of considerable public health interest (2). Others, although less common and less treatable, such as the inborn errors that affect cobalamin (or Vitamin B12) metabolism are important because they give an insight into the role of an enzyme or biochemical pathway in the maintenance of normal function and health (3). It is obvious, however, that these historical examples do not include a whole side of genetic variation and disease that remain largely uninvestigated. Epidemiological studies indicate that there is a very significant familial component to virtually all chronic diseases such as cardiovascular disease, colorectal cancer, or common birth defects, such as spina bifida and other neural tube defects (NTDs; ref. 4). As the human genome is sequenced scientists find that certain sequences are highly conserved. Yet even within the conserved sequences there are some genetic variations. Infrequently there are insertions or deletions of sections of the gene (5). Frequently there are single nucleotide changes (polymorphisms). The alteration of a base in the sequence could result in a variety of outcomes. If the nucleotide is in the 85% of the gene that makes up the intronic regions, it would have no effect unless it is at a splice site with an exon. If it is in the regulatory region, it might affect the binding of a regulator of gene expression. If it is within one of the exons, the base change might lead to a new triplet code that encodes for the same amino acid, resulting in no change in the protein. Finally, a change might lead to a different amino acid being encoded. To what degree, if any, this amino acid change would affect function could be highly variable and very difficult to predict. Because proteins are folded into a three-dimensional structure, it is impossible to say with certainty which amino acids are involved in essential functions. Even if one knows, through x-ray crystallography or site-directed mutagenesis, which amino acids are involved in the catalytic site or in regulatory sites, it could not be assumed that a change in the amino acid sequence distant for these regions would be unimportant. Such a change might alter the stability of the protein or how a protein interacts with other proteins or membranes. In fact, the changes in genotype that are now attracting the most interest are the ones in which small changes in sequence result in an outcome (phenotype) that is subtly different from the normal and difficult to detect. If the altered genotype resulted in a dramatically altered phenotype, it probably would not have survived during evolution. If it did survive, it would be very rare, and if recognized it would fall into the category of an inborn error in metabolism. The mapping of the human genome has now presented the opportunity to identify, within a very short period, all the common genetic polymorphisms. The challenge will be to determine whether they result in an altered phenotype and whether this alteration in phenotype results in the increased risk of a particular disease. As discussed, we know that these genotype changes are common, and we can anticipate that the resultant phenotype change will range from none to something quite subtle (they will not be similar to inborn errors of metabolism). We can also predict that multiple genotypes will be studied for associations with multiple diseases. Such multiple testing will result inevitably in a number of purely chance associations between an altered genotype and the risk of a particular disease. Knowing whether the genotype results in an altered phenotype will then be crucial in sorting out real from chance associations. The anticipated phenotype alteration will be small at best, and our current ability to detect small changes in function is poor in most areas. In this issue of PNAS new research explores an important novel approach to determining the likelihood of how an altered genotype might result in an altered phenotype (6).
The approach used by Yamada et al. was to employ recombinant DNA technology to express high levels of the wild-type and variant forms of the enzyme methylenetetrahydrofolate reductase. Thus, the authors have available to them the wild-type enzyme and its two common variants, each purified to homogeneity and available in sufficient quantities to carry out a full biochemical characterization. One of the variants (677 C → T) brings about an amino acid change from Ala-222 to Val, and the other (1298 A → C) changes Glu-429 to Ala. This paper compares the biochemical properties of the wild-type enzyme with the 677 C → T polymorphism and the 1298 A → C polymorphism and makes a further comparison between these preparations and that from the enzyme containing the double mutation of 677 C → T and 1298 A → C.
New research explores an important novel approach to determining the likelihood of how an altered genotype might result in an altered phenotype.
Having sufficiently purified enzyme in each instance, Yamada et al. made a thorough examination of all the kinetic and biochemical properties of the different enzymes. This examination includes comparing kinetic functions such as Vmax and Km, which are found to be the same, as are the Ki for the inhibitor S-adenosylmethionine. Thus, these polymorphisms either individually or combined do not affect catalytic function or regulation. The authors already had established that a similar polymorphism in the expressed form of the enzyme from an Escherichia coli readily loses its flavin cofactor (FAD), unlike the wild type (7). They now also find that the human 677 C → T variant releases its FAD cofactor three times faster than the wild-type enzyme. The 1298 A → C variant shows a pattern identical to that of the wild type. The double mutant behaves similarly to the 677 C → T variant. They conclude that it is the loss of the FAD cofactor at the in vivo relevant concentration that results in the 30% decrease in activity seen in the cell extracts of the human variant compared with the wild type (8). They also find that the loss of activity of the enzyme can be diminished by increasing the concentration of the enzyme's naturally occurring substrate 5CH3H4 folate, its inhibitor S-adenosylmethionine, and its cofactor FAD.
Comparisons of what has emerged in the literature with respect to the expressed phenotype of these two variants gives reassurance but produces challenges. The 677 C → T variant is shown to be associated with decreased enzyme activity both in extracts from tissue culture (8) and from lymphocytes taken from the homozygous, heterozygous, and wild-type genotypes (9). Yamada et al. (6) find that the purified (677 C → T) variant form of the enzyme has the same stability as the wild type in vitro but it depends on the 5CH3H4 folate, S-adenosylmethionine, and FAD concentrations. Under conditions in which folate status and thus intracellular folate concentrations are low, the 677 C → T variant may be particularly prone to losing its FAD and thus to inactivation (10). Of equal interest is the apparent activity of the different forms of the enzyme with respect to riboflavin status. Riboflavin is the vitamin precursor of FAD. In a recent study it was shown that reductions in enzyme activity, as measured by elevations of the plasma homocysteine, contrast markedly between those with the variant TT or CT form of the enzyme compared with the wild-type CC genotype.† The T variant seems to have equal stability to the wild type in subjects with replete riboflavin status. It is only in subjects deficient in riboflavin that the T variant shows reduction in activity in vivo, as measured by elevation of plasma homocysteine. These findings are consistent with the suggestions by the authors that it is through loss of the FAD cofactor that the variant causes reduced activity and thus a detectable phenotype. All this suggests that this novel approach of studying the biochemical properties of the particular enzyme will have to be done in parallel with, and not as a replacement for, other studies to determine the presence of a phenotype.
With respect to the 1298 A → C polymorphism, the literature suggests that the polymorphism also results in decreased enzymatic activity in individuals homozygous for the variant even though there is no increase in plasma homocysteine (12–14). The biochemical approach of Yamada et al. (6) gives a plausible explanation as to why the 677 C → T variant would be unstable because of the loss of its FAD. However, no such altered properties were found for the 1298 A → C polymorphism. In fact, one of the earlier studies also concluded that the 1298 A → C variant, unlike the 677 C → T variant, was not thermolabile. This result reinforces the argument that many different approaches may be needed, from the biochemical studies described by Yamada et al. (6) to measurement of enzyme extracts (8, 9) to determination of the altered level of metabolites (10, 15), to establish the presence or absence of a phenotype.
Likewise, with respect to risk of disease, the biochemical study (6) would infer that, particularly in those with low folate status, the 677 C→T variant would be associated with increased clinical risk. This certainly has been found to be the case with respect to NTDs (16) and may be true in cardiovascular disease (11). Using the biochemical approach, the 1289 A → C polymorphism shows no difference from wild type either in its kinetic or stabilizing properties (6). Two studies have found that, unlike 677 C → T, the 1298 A → C polymorphism does not increase the risk of an NTD-affected birth. The study by van der Put et al. (12) suggests that the combination of being heterozygous for both the 677 C → T and the 1298 A → C variant causes an increase in NTO risk. In light of the fact that the biochemical studies (6) find no additive effect when both amino acids are altered, the claim that the double variant increases risk needs to be reevaluated.
The most important role that the biochemical approach may play in the future is where a variant of some enzyme or protein has been associated with an increased risk of disease in a convincing way and in a number of studies, but there may be a total absence of evidence of any phenotype change. This clearly is not the case for the 677 C → T variant, for which there is ample evidence of a change in phenotype [altered enzyme levels (8, 9) raised plasma homocysteine and lowered red cell and plasma folate levels (15)]. Also associated with the 677 C → T is convincing evidence that it is a risk factor for NTDs (16) and more controversial evidence that it is a risk for cardiovascular disease (11). If one looks at the metabolic role of methylenetetrahydrofolate reductase, it would be hard to imagine that any significant reduction in its activity would not alter plasma homocysteine. However, if we took another enzyme in folate metabolism (e.g. dihydrofolate reductase), an altered phenotype might be harder to detect (Fig. 1). This enzyme is important only in cells that are actively synthesizing pyrimidines and thus DNA. A variant in dihydrofolate reductase therefore might have clinical consequences but little or no biochemical evidence of them. In such instances the demonstration of a phenotype from an altered genotype might be achieved only by the biochemical approach used in the paper by Yamada et al. (6).
Figure 1.
The principle folate-dependent enzymes and pathways.
It also must be borne in mind that the absence of a biochemical phenotype by measuring an analyte such as homocysteine does not necessarily rule out the possible importance of a polymorphism in disease. The reason for this is because homocysteine metabolism, and its resultant plasma level, is principally caused by the role of the enzymes involved in methionine and homocysteine metabolism in the liver and kidney. A particular polymorphism might cause increased risk during a critical stage in the closure of the neural tube but not have a parallel in altered metabolite levels in plasma or in levels of the enzyme in lymphocytes or cell culture extract. It thus would become crucial to know whether such a polymorphism has any evidence of biochemical alterations in phenotype. This type of scenario might only be detected by using the approach in this paper (6).
Footnotes
See companion article on page 14853.
McNulty, H., McKinley, M. C., Wilson, B., Strain, J. J., Weir, D. G. & Scott, J. M. (2001) Ann. Nutr. Metab. 45, Suppl. 1, 60 (abstr.).
References
- 1.Mudd S H, Levy H L, Skovby F. In: The Metabolic Basis of Inherited Diseases. 7th Ed. Scriver C R, Beaudet A L, Sly W S, Valle D, editors. New York: McGraw–Hill; 1995. pp. 1279–1327. [Google Scholar]
- 2.Murray R K. In: Harper's Biochemistry. 24th Ed. Murray R K, Granner D K, Mayes P A, Rodwell V W, editors. London, U.K.: Prentice–Hall International; 1996. pp. 779–793. [Google Scholar]
- 3.Rosenblatt D S. In: Homocysteine in Health and Disease. Carmel R, Jacobsen D W, editors. Cambridge, U.K.: Cambridge Univ. Press; 2001. pp. 244–258. [Google Scholar]
- 4.Scott J M, Weir D G, Kirke P. In: Folate in Health and Disease. Baily L, editor. New York: Dekker; 1994. pp. 329–360. [Google Scholar]
- 5.Strachan T, Read A P. Human Molecular Genetics. 2nd Ed. Oxford, U.K.: BIOS Scientific; 1999. [Google Scholar]
- 6.Yamada K, Chen Z, Rozen R, Matthews R G. Proc Natl Acad Sci USA. 2001;98:14853–14858. doi: 10.1073/pnas.261469998. . (First Published December 11, 2001; 10.1073/pnas.261469998) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Guenther B D, Sheppard C A, Tran P, Rozen R, Matthews R G, Ludwig M L. Nat Struct Biol. 1999;6:359–365. doi: 10.1038/7594. [DOI] [PubMed] [Google Scholar]
- 8.Frosst P, Blom H J, Milos R, Goyette P, Sheppard C A, Matthews R G, Boers G J, den Heijer M, Kluijtmans L A, van den Huevel L P, et al. Nat Genet. 1995;10:111–113. doi: 10.1038/ng0595-111. [DOI] [PubMed] [Google Scholar]
- 9.Van der Put N M J, Steegers-Theunissen R P M, Frosst P, Trijbels F J M, Eskes T K A B, van den Heuvel L P, Mariman E C M, den Heyer M, Rozen R, Blom H J. Lancet. 1995;346:1070–1071. doi: 10.1016/s0140-6736(95)91743-8. [DOI] [PubMed] [Google Scholar]
- 10.Jacques P F, Bostom A G, Williams R R, Ellison R C, Eckfeldt J H, Rosenberg I H, Selhub J, Rozen R. Circulation. 1996;93:7–9. doi: 10.1161/01.cir.93.1.7. [DOI] [PubMed] [Google Scholar]
- 11.Ueland P M, Refsum H, Beresford S A, Vollset S M. Am J Clin Nutr. 2000;72:324–332. doi: 10.1093/ajcn/72.2.324. [DOI] [PubMed] [Google Scholar]
- 12.Van der Put N M J, Gabreels F, Stevens E M B, Smeitink J A M, Trijbels F J M, Eskes T K A B, van den Heuvel L P, Blom H J. Am J Hum Genet. 1998;62:1044–1051. doi: 10.1086/301825. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Weisberg I, Tran P, Christensen B, Sibani S, Rozen R. Mol Genet Metab. 1998;64:169–172. doi: 10.1006/mgme.1998.2714. [DOI] [PubMed] [Google Scholar]
- 14.Weisberg I S, Jaques P F, Selhub J, Bostom A G, Chen Z, Ellison R C, Eckfeldt J H, Rozen R. Atherosclerosis. 2001;156:409–415. doi: 10.1016/s0021-9150(00)00671-7. [DOI] [PubMed] [Google Scholar]
- 15.Molloy A M, Daly S, Mills J L, Whitehead A S, Ramsbottom D, Conley M, Weir D G, Scott J M. Lancet. 1997;349:1591–1593. doi: 10.1016/S0140-6736(96)12049-3. [DOI] [PubMed] [Google Scholar]
- 16.Shields D C, Kirke P N, Mills J L, Ramsbottom D, Molloy A M, Weir D G, Scott J M, Whitehead A S. Am J Med Genet. 1999;64:1045–1055. doi: 10.1086/302310. [DOI] [PMC free article] [PubMed] [Google Scholar]