10-Formyl-7,8-dihydrofolic acid (10-HCO-H2folate) was prepared by controlled air oxidation of 10-formyl-5,6,7,8-tetrahydrofolic acid (10-HCO-H4folate). The UV spectra of the 10-HCO-H2folate preparation has lambda max. 234, 333 nm and lambda min. 301 nm at pH 7.4, and lambda max. 257, 328 nm and lambda min. 229, 307 nm at pH 1. 1H-NMR spectroscopy of 10-HCO-H2folate (in 2H2O; 300 MHz) suggested a pure compound and gave resonances for one formyl group proton, two protons on C-7 and C-9, and no evidence for a C-6 proton, which is consistent with the structure proposed. The spectral properties indicated that the 10-HCO-H2folate preparation is not appreciably contaminated with 10-HCO-H4folate, 5,10-methenyltetrahydrofolic acid (5,10-CH = H4folate) or 10-formylfolic acid (10-HCO-folate). The above data establish that the 10-HCO-H2folate prepared here is authentic. In contrast, a folate with a UV spectrum having lambda max. 272 nm and lambda min. 256 nm at pH 7, which was prepared by 2,6-dichloro-indophenol oxidation of 10-HCO-H4folate and reported to be 97% pure [Baram, Chabner, Drake, Fitzhugh, Sholar and Allegra (1988) J. Biol. Chem. 263, 7105-7111], is apparently not 10-HCO-H2folate. 10-HCO-H2folate is utilized by Jurkat-cell (human T-cell leukaemia) and chicken liver aminoimidazolecarboxamide ribonucleotide transformylase (AICAR T'ase; EC 2.1.2.3) in the presence of excess 5-amino-imidazole-4-carboxamide ribotide (AICAR) resulting in the appearance of approximately 1 mol of H2folate product for each mol of AICAR formylated. The present 10-HCO-H2folate preparation had a kinetic advantage over 10-HCO-H4folate resulting from a difference of approx. 5-fold in K(m) values when both folates were used as cofactors for Jurkat-cell and rat bone marrow AICAR T'ase. No substantial kinetic advantage was observed using chicken liver AICAR T'ase. 10-HCO-H2folate had little or no activity with Jurkat-cell or chicken liver glycinamide ribonucleotide transformylase (GAR T'ase, EC 2.1.2.2). The existence in vivo of 10-HCO-H2folate is suggested in mammals by several reports of detectable amounts of radiolabelled 10-HCO-folate in bile and urine after administration of radiolabelled folic acid.

7-Hydroxymethotrexate (7-OH-MTX) is the major and, frequently the only, pteridine metabolite found in bone-marrow aspirates of patients chronically treated with low-dose oral methotrexate (MTX) [Sonneveld, Schultz, Nooter and Hahlen (1986) Cancer Chemother. Pharmacol. 18, 111-116]. The Ki values for MTX and 7-OH-MTX for avian liver 5-amino-imidazole-4-carboxamide ribonucleotide transformylase differ by 4.5-fold in favour of 7-OH-MTX as the better inhibitor, while Ki values for avian liver glycinamide ribonucleotide transformylase differ by 1.9-fold favouring MTX as the better inhibitor. Thus 7-OH-MTX possesses a different enzyme-inhibiting repertoire from its parent drug and this information may be useful in explaining the mechanism of action of low-dose MTX therapies used to treat autoimmune disease.

Many non-steroidal anti-inflammatory drugs (NSAIDs) (including sulphasalazine, sulindac, indomethacin, naproxen, salicylic acid, ibuprofen, piroxicam and mefenamic acid) were found to be competitive inhibitors (with respect to folate) of avian liver phosphoribosylaminoimidazolecarboxamide formyltransferase (AICAR transformylase, EC 2.1.2.3) and bovine liver dihydrofolate reductase (EC 1.5.1.3). In contrast, aspirin and the antipyretic-analgesic drugs acetaminophen and antipyrine were weak inhibitors of these enzymes. Structure-activity correlation suggests that an aromatic ring with a side chain containing a carboxylic acid is a requirement for competitive inhibition of the transformylase. The above-listed NSAIDs also inhibited the folate-coenzyme-mediated biosynthesis of serine from glycine and formate (i.e., the C1 index) by human blood mononuclear cells (BMCs) in experiments where the drug was added to a culture of BMCs. Acetaminophen had a weak inhibitory effect on the C1 index. Consistent with the results obtained in vitro is the observation that the C1 index of BMCs from rheumatoid-arthritis patients treated with drugs which possess little antifolate activity (e.g. acetaminophen) is higher than the C1 index of BMCs from rheumatoid-arthritis patients treated with NSAIDs possessing more potent antifolate activity (e.g. sulindac, sulphasalazine, naproxen and ibuprofen). The mean activity of the transformylase in BMCs taken from healthy humans was 1.98 nmol of product/h per 10(6) cells and the activity was positively correlated with BMC folate levels. These results are consistent with the hypothesis that (1) the antifolate activity of NSAIDs, and hence cytostatic consequences, are important factors in producing anti-inflammatory activity and (2) aspirin exerts its anti-inflammatory effects after its conversion into salicylic acid, which possesses greater antifolate activity than its parent compound.

The colorimetric assay for 5-aminoimidazole-4-carboxamide ribotide (AICAR) transformylase (phosphoribosylamino-imidazolecarboxamide formyltransferase; EC 2.1.2.3) has been extensively modified. The modified assay is based upon the short-term permanganate oxidation of the folate product, tetrahydrofolate (H4folate) to p-aminobenzoyl glutamate (pABG). The modified assay was used to detect the transformylase activity in crude extracts of peripheral-blood mononuclear cells (PBMCs). Azathioprine and its metabolite, thioinosinic acid (tIMP), are competitive inhibitors (with respect to AICAR) of the chicken liver transformylase and the transformylase from PBMCs of the MRL/lpr mouse, an animal model of systemic autoimmune disease. The Ki values of tIMP and azathioprine for the chicken liver enzyme are 39 +/- 4 microM and 120 +/- 10 microM, whereas the Ki values for the enzyme from PBMCs of the MRL/lpr mouse are 110 +/- 20 microM and 90 +/- 14 microM respectively. The anti-inflammatory drugs ibuprofen and naproxen are also inhibitors of the transformylase.

With the use of a continuous spectrophotometric assay and initial rates determined by the method of Waley [Biochem. J. (1981) 193, 1009-1012] methotrexate was found to be a non-competitive inhibitor, with Ki(intercept) = 72 microM and Ki(slope) = 41 microM, of 5-aminoimidazole-4-carboxamide ribotide transformylase, whereas a polyglutamate of methotrexate containing three gamma-linked glutamate residues was a competitive inhibitor, with Ki = 3.15 microM. Pentaglutamates of folic acid and 10-formylfolic acid were also competitive inhibitors of the transformylase, with Ki values of 0.088 and 1.37 microM respectively. Unexpectedly, the pentaglutamate of 10-formyldihydrofolic acid was a good substrate for the transformylase, with a Km of 0.51 microM and a relative Vmax. of 0.72, which compared favourably with a Km of 0.23 microM and relative Vmax. of 1.0 for the tetrahydro analogue. An analysis of the progress curve of the transformylase-catalysed reaction with the above dihydro coenzyme revealed that the pentaglutamate of dihydrofolic acid was a competitive product inhibitor, with Ki = 0.14 microM. The continuous spectrophotometric assay for adenosine deaminase based on change in the absorbance at 265 nm was shown to be valid with adenosine concentrations above 100 microM, which contradicts a previous report [Murphy, Baker, Behling & Turner (1982) Anal. Biochem. 122, 328-337] that this assay was invalid above this concentration. With the spectrophotometric assay, 5-aminoimidazole-4-carboxamide riboside was found to be a competitive inhibitor of adenosine deaminase, with (Ki = 362 microM), whereas the ribotide was a competitive inhibitor of 5′-adenylate deaminase, with Ki = 1.01 mM. Methotrexate treatment of susceptible cells results in (1) its conversion into polyglutamates, (2) the accumulation of oxidized folate polyglutamates, and (3) the accumulation of 5-aminoimidazole-4-carboxamide riboside and ribotide. The above metabolic events may be integral elements producing the cytotoxic effect of this drug by (1) producing tighter binding of methotrexate to folate-dependent enzymes, (2) producing inhibitors of folate-dependent enzymes from their tetrahydrofolate coenzymes, and (3) trapping toxic amounts of adenine nucleosides and nucleotides as a result of inhibition of adenosine deaminase and 5′-adenylate deaminase respectively.