THE COMPARISON
A1298C vs C677T: the two common MTHFR variants, compared properly
Most “MTHFR” content on the internet is C677T content wearing a generic label. The two common variants sit in different domains, break different things, and carry different weights of evidence — and much of the older literature confounds them because they are in linkage disequilibrium. This page separates them. Nothing here is medical advice.
Side by side
| PROPERTY | C677T (rs1801133) | A1298C (rs1801131) |
|---|---|---|
| Database name | rs1801133 | rs1801131 |
| HGVS | c.665C>T · p.Ala222Val | c.1286A>C · p.Glu429Ala |
| Exon / domain | Exon 4 — N-terminal catalytic (βα)₈ barrel | Exon 7 — C-terminal regulatory (SAM-binding) |
| Mechanism | Accelerated FAD cofactor loss + monomer dissociation → thermolability; folate and SAM binding slow the loss | Purified enzyme indistinguishable from wild-type; not thermolabile |
| In-vitro activity | 45% ± 10.8 of wild-type | 68% ± 5.0 of wild-type |
| Homocysteine (alone) | Homozygotes: significantly elevated — the strongest common genetic cause of mild hyperhomocysteinemia | Homo- or heterozygotes: not associated with higher homocysteine or lower folate in the discovery data; at most a weak signal in very large cohorts |
Sources: discovery and recombinant-expression work for both variants, structural mechanisms, and the discovery-era homocysteine data.[4][5][7][8][10][9][11]
Why C677T is the “stronger” variant
C677T’s Ala222Val substitution sits in the catalytic core and attacks the enzyme’s physical integrity: it accelerates dissociation of the FAD cofactor and promotes monomer formation, producing a thermolabile enzyme with roughly half of normal activity in TT homozygotes and a clear homocysteine elevation when folate status is low. Folate and SAM binding stabilise the enzyme against that loss — which is why the C677T phenotype is so sensitive to vitamin status.[4][8][10][28]
A1298C’s Glu429Ala substitution sits in the regulatory domain. Its cellular effect is real but smaller — about 60–68% of control activity — its mechanism is less clear, it produces no thermolability, and alone it does not meaningfully raise homocysteine.[7][5]
The purified-enzyme caveat
In 2001, Yamada and colleagues characterised highly purified recombinant human MTHFR carrying the 1298 substitution — and found it “indistinguishable from the wild-type enzyme”. The cellular activity reduction measured in lymphocyte extracts is real; the purified protein behaves normally. Both things are true at once, and holding them both is the discipline: a genetic association, a cellular phenotype and a molecular mechanism are three different claims requiring three different proofs.[8]
Compound heterozygosity: where the two meet
Carry one 677T allele and one 1298C allele and the picture changes. In the Canadian discovery sample about 15% of individuals were compound heterozygotes, with 50–60% of control enzyme activity — lower than single C677T heterozygotes — and no individual homozygous for both variants was observed.[6]
Compound heterozygotes show a C677T-homozygote-like phenotype: reduced MTHFR specific activity (ANOVA P<.0001), higher homocysteine and decreased plasma folate (ANOVA P<.03) in the discovery study, and significantly higher fasting homocysteine than carriers of the 677 variant alone in the Family Heart Study (P<0.05).[5][7]
The clinical boundary, reported honestly: in the discovery study, compound heterozygosity appeared in 28% of neural-tube-defect patients versus 20% of controls — an odds ratio of 2.04 whose 95% CI (0.9–4.7) crosses 1, so it was not statistically significant. And current professional guidance from the ACMG states that MTHFR polymorphism testing, for either variant, has minimal clinical utility and should not be part of routine thrombophilia evaluation.[5][68]
One more asymmetry: riboflavin
The most developed gene–nutrient interaction in this field — riboflavin lowering blood pressure in MTHFR 677TT carriers, tested in repeated randomised trials — rests on the FAD-loss mechanism. That mechanism is established for C677T, not for A1298C: the 1298 variant shows no FAD-loss behaviour and no thermolability, and no equivalent riboflavin-intervention trial for A1298C exists. Even here, though, the 2025 Cochrane review rates the riboflavin–blood-pressure evidence “very uncertain” and calls for larger trials.[29][31][36][37]
Any page implying that riboflavin “answers” A1298C has left the evidence. This one won’t.