Search

Mouse fibroblasts in which the mthfd2 gene encoding mitochondrial NAD-dependent methylenetetrahydrofolate dehydrogenase-cyclohydrolase (NMDMC) was previously inactivated were infected with retroviral expression constructs of dehydrogenase/cyclohydrolase cDNA. Cellular fractionation confirmed that the expressed proteins were properly targeted to the mitochondria. Expression of the NAD-dependent methylenetetrahydrofolate dehydrogenase-cyclohydrolase enzyme in mitochondria corrected the glycine auxotrophy of the null mutant cells. A construct in which the cyclohydrolase activity of NMDMC was inactivated by point mutation also rescued the glycine auxotrophy, although poorly. This suggests that the cyclohydrolase activity is also required to ensure optimal production of 10-formyltetrahydrofolate. The expression of the NADP-dependent methylenetetrahydrofolate dehydrogenase-cyclohydrolase-synthetase in the mitochondria also reversed the glycine requirement of the null cells demonstrating that the use of the NAD cofactor is not absolutely essential to maintain the flux of one-carbon metabolites. All rescued cells demonstrated a decrease in the ratio of incorporation of exogenous formate to serine in standardized radiolabeling studies. This ratio, which is approximately 2.5 for nmdmc(-/-) cells and 0.3 for the wild type cells under the conditions used, is a qualitative indicator of the ability of the mitochondria of the cells to generate formate.

We have isolated the cDNA and the gene encoding the murine cytoplasmic methylenetetrahydrofolate dehydrogenase-methenyltetrahydrofolate cyclohydrolase-formyltetrahydrofolate synthetase (DCS). Comparison of these sequences with the 3′-untranslated region of the mitochondrial NAD+-dependent methylenetetrahydrofolate dehydrogenase-cyclohydrolase (mt-DC) revealed areas of significant homology. Both exon and intron sequences of the synthetase domain of DCS are homologous to sequences in the untranslated region of mt-DC. A similar comparison between the mt-DC and the DCS sequences of humans as well as Drosophila supports the conclusion that in higher eukcaryotes the bifunctional mt-DC replaced a trifunctional precursor through inactivation of the synthetase domain. The mt-DC should be considered in models of one-carbon folate fluxes in mammals.

Methylenetetrahydrofolate reductase and methyltetrahydrofolate-homocysteine methyltransferase were found to be localized in the soluble fraction of rat brain. They are clearly separated from serine hydroxymethyltransferase and the glycine cleavage complex which are localized in the mitochondria in this tissue. Hence, although the primary, if not the only, site of 5,10-methylenetetrahydrofolate formation in brain appears to be the mitochondrion, the utilization of this compound for 5-methyltetrahydrofolate synthesis and utilization of the latter compound for methylation of homocysteine occurs in the cytosol. Parallel experiments with rat liver confirmed that the reductase and homocysteine methyltransferase of this tissue are also localized in the soluble fraction, while the hydroxymethyltransferase is about evenly divided between mitochondria and cytosol. However, in liver (but not in brain) the reductase activity of the supernatant fraction is only partially expressed unless the fraction is dialyzed. We have found that this phenomenon, which initially suggested the occurrence of an endogenous inhibitor in liver extracts [Ordonez and Wurtman (1973) J. Neurochem., 21, 1447–1455] , is due to loss of the product (5, 10-methyl-enetetrahydrofolate) of the reductase assay via its oxidation by methylenetetrahydrofolate dehydrogenase, in combination with the NADP present in undialyzed extracts. All regions of the rat central nervous system tested contained methylenetetrahydrofolate reductase activity. Regional variations were observed however, with an almost threefold difference between the specific activities of the highest and lowest ranking regions Comparison of the rank order of 12 regions tested with respect to reductase specific activity (this study) and methyltetrahydrofolate-tryptamine N-methyltransferase specific activity [Mandell, Knapp, and Hsu (1974) Life Sci., 14, 1–17] shows a high positive correlation (r = 0.916) between these activities in the selected regions.

A methotrexate-resistant subline of the Ehrlich ascites carcinoma contained approximately 14-fold more dihydrofolate reductase activity than did the drug-sensitive parent neoplasm. The activities of a number of tetrahydrofolate-dependent enzymes (i.e., of N 5 -formyltetrahydrofolate synthetase, N 5 , N 10 -methylenetetrahydrofolate dehydrogenase, N 5 -formyltetrahydrofolate isomerase, and serine hydroxymethylase) in cells of the resistant subline were similar to those of the sensitive neoplasm; thus, the increase in dihydrofolate reductase activity in the methotrexate-resistant cells was specific. Several properties of partially purified dihydrofolate reductase preparations from each cell line were similar, suggesting that no quantitative changes in the enzyme accompanied the transition to the resistant state. The elevated dihydrofolate reductase activity in the drug-resistant variant appeared to be sufficient to account for the refractoriness of this neoplasm, since considerable enzyme activity could be detected in these cells after large amounts of methotrexate. The problems of relating in vitro enzyme assays to events occurring in vivo are discussed.