An enzyme that catalyzes the hydrolysis of folic acid and the antifolate methotrexate nearly 20 times more rapidly than the hydrolysis of 5-methyltetrahydrofolate was extracted from a gram-negative bacterium tentatively identified as a Flavobacterium sp. The enzyme was purified 500-fold and found to have a molecular weight of about 53,000. Apparently a metallo-enzyme, it is inhibited by citrate and ethylenediaminetetraacetic acid (EDTA). Ca2+, Co2+, Mg2+, and Zn2+ reverse inhibition by EDTA, whereas Ca2+ and Zn2+ are weak activators in the absence of EDTA. The enzymatic reaction releases the carboxy-terminal glutamyl moiety of derivatives of pteroyl-mono-L-glutamic acid. Substituents on N5 of the pteridine ring decrease the velocity of hydrolysis. Some non-specificity for the terminal amino acid is expressed. The strikingly different rates of hydrolysis of methotrexate and 5-methyltetrahydrofolate have stimulated interest in this enzyme for its potential clinical value in improving the therapeutic index of methotrexate. The folic acid analog methotrexate (4-amino-N10-methylpteroylglutamic acid [MTX]) is a potent antineoplastic agent (1). Its clinical application has been advanced with the use of leucovorn (N6-formyltetrahydrofolic acid) rescue (8, 11, 13, 18, 26), which has been relatively successful in relieving the toxicity of antifolate therapy. The report of a bacterial carboxypeptidase G1 (CPD G1) with hydrolytic activity for MTX introduced the possibility of the enzymatic approach to ameliorating drug toxicity (12, 23). In this report, we describe the partial purification and characterization of another enzyme that catalyzes the hydrolysis of MTX, folic acid (FA), and other derivatives of FA by the reaction shown in Fig. 1. This carboxypeptidase is produced by a bacterium tentatively designated Flavobacterium sp. and differs from CPD G and CPD G1 (15, 19, 23). MATERIALS AND METHODS Chemicals. FA (pteroyl-L-glutamic acid) was purchased from Grand Island Biological Co. dl-N5-methyltetrahydrofolic acid was obtained from Sigma Chemical Co. as the barium salt. It was converted to the sodium salt by the addition of an equal volume of 0.1 M sodium phosphate, pH 6.5, to a 1% solution of the folate in 0.1 M sodium bicarbonate-0.05% potassium ascorbate. Leucovorin (dl-N-5-formyltetrahydrofolic acid, calcium salt) and MTX were provided by Bernard C. Clark of Lederle Laboratories, American Cy-anamid Co., and by Harry B. Wood, Jr., Drug Development Branch, Drug Research and Development, Silver Spring, Md. Chromatographicaily purified MTX (4) was used for kinetic studies. Purified aminopterin was a gift from Francis M. Sirotnak, Sloan-Kettering Institute. Harry B. Wood, Jr., also provided bioautographically pure samples of methasquin (17),

The activity of transketolase in mouse brain was 5.7 nmol/min/mg protein measured by an enzyme-coupled spectrophotometric assay. The apparent Km for ribose-5-phosphate was 330 microM, for D-xylulose-5-phosphate was 120 microM, and for thiamine pyrophosphate was 7 microM. However, thiamine pyrophosphate remained tightly bound to transketolase in homogenates in which it dissociated completely from another thiamine pyrophosphate-dependent enzyme, the pyruvate dehydrogenase complex. These data suggest that loss of transketolase activity is likely to be a later consequence of thiamine deficiency in mammalian brain than is decreased activity of pyruvate dehydrogenase complex.