Folylpoly-γ-glutamate synthetase: Generation of isozymes and the role in one carbon metabolism and antifolate cytotoxicity

A single human gene encodes both mitochondrial and cytosolic isoforms of the enzyme. The major mRNA species in human cells encodes the mitochondrial isoform but alternate translation initiation at a downstream in-frame ATG also generates the cytosolic isoform. Cytosolic FPGS may also be generated by use of alternate transcription initiation start sites 3' to the start ATG of the mitochondrial FPGS. Three additional human FPGS mRNAs differing in exon 1 have been identified. One of these is a major species in HEP-G2 cells and other tissue culture cells, and can encode a protein lacking the first 8 amino acids of cytosolic FPGS. A protein of the predicted size is observed in coupled transcription/translation systems. However, expression of this protein in E. coli does not generate an active enzyme. Mutagenesis studies indicate that Tyr-3 of the missing N terminal residues is required for enzyme activity. The major cellular folate pools are in the cytosol and mitochondria and FPGS activity is normally distributed in both compartments. Mitochondrial FPGS activity is required for mitochondrial folate accumulation, and cells lacking this isozyme are auxotrophic for glycine. Overexpression of cytosolic FPGS does not complement the lack of mitochondrial activity. Cells expressing FPGS activity solely in the mitochondria are glycine prototrophs, but also possess cytosolic folylpolyglutamates and are prototrophic for thymidine and purines, products of cytosolic one carbon metabolism. Although cytosolic folylpolyglutamates cannot enter the mitochondrion, mitochondrial folylpolyglutamates are released intact into the cytosolic compartment. Cellular accumulation of some antifolates and their cytotoxic efficacy is highly responsive to the level of FPGS activity. Polyglutamylation of methotrexate (MTX) has little affect on its affinity for dihydrofolate reductase, its target enzyme, but does affect the cellular accumulation of the drug. The sensitivity of model cells, expressing a range of FPGS activities similar to that observed in leukemia blasts, to MTX varied over four orders of magnitude. MTX toxicity was dependent on cytosolic FPGS activity as this drug does not enter the mitochondria, and cells expressing very high levels of FPGS solely in the mitochondria were resistant to MTX. The cytotoxic efficacy of other folate antagonists that are transported into the mitochondria was enhanced by mitochondrial FPGS activity, even when their loci of inhibition was a cytosolic enzyme. Mitochondrial metabolism of these drugs increased cytosolic drug levels. Compartmentalization of antifolate metabolism has to be considered in evaluating mechanisms for increased drug cytotoxicity and for the development of acquired resistance to these agents.