Evidence from chemical degradation studies for a covalent bond from 5-fluoro-2'-deoxyuridylate to N-5 of tetrahydrofolate in the ternary complex of thymidylate synthetase-5-fluoro-2'-deoxyuridylate-5,10-methylenetetrahydrofolate

In the ternary complex of thymidylate synthetase, 5-fluoro-2'-deoxyuridylate (FdUMP), and 5,10-methylenetetrahydrofolate (5,10-CH2H4folate), the 5-fluorouracil moiety is covalently bound to the enzyme by a sulfide linkage from C-6 and to either N-5 or N-10 of H4folate by a methylene bridge from C-5. In an effort to establish the site by which H4folate is attached to FdUMP, the ternary complex was subjected to reagents that cleave the C-9, N-10 bond of folate derivatives. The complex was stable to zinc dust in hydrochloric acid, a reagent that cleaves N-10-substituted but not N-5-substituted folates. The conditions of the Bratton-Marshall reaction, which involve the use of nitrous acid, were found to cleave N-5-substituted folates in yields ranging from 5 to 50%. Exposure of the double-labeled thymidylate synthetase-FdUMP-[2-14C,7,9,3',5'-3H]5,10-CH2H4folate complex to the Bratton-Marshall reaction resulted in 16% cleavage of the C-9, N-10 bond with release solely of p-aminobenzoylglutamate, whereas all of the carbon-14-labeled pterin residue remained covalently bound to the protein. These results demonstrate that in the ternary complex, the 5-fluorouracil residue is connected by a covalent bond to N-5 of H4folate.

Assay and time course of 5-fluorouracil incorporation into RNA of L1210/0 ascites cells in vivo

A method for determination of levels of incorporation of nonradiolabeled 5-fluorouracil (FUra) into RNA (F-RNA) in tissue samples is shown to be applicable to tissues in vivo. BDF1 mice bearing L1210 ascites cells were injected intraperitoneally with [14C]FUra, 100 mg/kg. The time course of F-RNA levels in L1210 cells was determined by following the radiolabeled drug, and by NaB3H4 labeling of isolated and derivatized nucleoside. RNA ribonucleotides were obtained by KOH hydrolysis of perchloric acid precipitates of cell sonicates. FUMP nucleotides were separated from remaining nucleotides by DEAE-cellulose chromatography. FUMP fractions were treated with alkaline phosphatase, and FUrd was separated from non-FUrd nucleoside contaminants by additional DEAE-cellulose chromatography. FUrd was quantitated by periodate oxidation of ribose and NaB3H4 reduction of the resulting nucleoside dialdehydes. Isolation of tritiated FUrd-trialcohol from remaining tissue contaminants and background radioactivity was done by silica gel thin layer chromatography. Comparison of results obtained by isolation of [14C]FUrd with results of NaB3H4 labeling of the same samples showed parallel results with comparable biological standard deviations, although the tritium method consistently gave slightly lower values. The peak level of F-RNA at 3 hr was 1 base substitution per 174 normal nucleotides. The level of F-RNA after 3 hr declined slowly, so that at 96 hr there still remained 1 FUra base per 597 normal nucleotides. Serial determinations of RNA content showed marked decreases, on the basis of either DNA or protein level, that continued up to 96 hr after FUra administration. These biochemical effects are among the most prolonged reported for FUra, suggesting the possibility that F-RNA represents a storage compartment for release of toxic metabolites and emphasize the need for additional study of RNA effects at long time points. Our method for assay of F-RNA appears to be suitable for study of biopsy specimens of tumors and normal tissues following nonradiolabeled-FUra administration.

Deoxyribonucleoside-induced selective modulation of cytotoxicity and mutagenesis

Treatment of Chinese hamster V79 cells with dThd, dCyd, or dThd plus dCyd increased MNNG-induced AGr-, TGr-, and Ouar-mutant frequencies but did not significantly increase background mutant frequencies. All the AGr colonies that were isolated possessed phenotypes characteristic of HGPRT-deficient mutants, and the deoxyribonucleosides did not selectively affect the growth of the mutants, nor the selecting efficiency of AG, and did not significantly enhance background mutagenesis. These data show that both dThd and dCyd facilitated MNNG-induced mutagenesis. This facilitation was maximal when cells were exposed to the deoxyribonucleosides throughout the first doubling time (24 h) after treatment with MNNG and for 4 more doubling times prior to mutant selection with AG. This indicates that one round of DNA replication was sufficient for mispairing of methylated bases in the DNA with the C and T provided by the deoxyribonucleosides, and that 4-6 doublings prior to mutant selection with AG were necessary to deplete pre-existing hypoxanthine: guanine phosphoribosyl transferase in newly mutated cells. The dCyd facilitated mutagenesis by FdUrd, which was not mutagenic without dCyd, indicating that increased dCTP:dTTP ratios were mutagenic. Treatment with FdUrd plus dCyd also induced FdUrdr cells, suggesting that inhibition of dCyd utilization may prevent the development of FdUrd-resistance in cancer chemotherapy. Although dCyd and dThd facilitated mutagenesis in cells treated with monofunctional alkylating agents that methylate DNA oxygens, facilitation of mutagenesis did not occur in cells treated with BCNU, which cross links DNA, nor with benzo(a)pyrene and aflatoxin B1, which are frame shift mutagens, nor with MMS, which produces barely detectable levels of O-methylation in DNA. Virtually non toxic concentrations of dThd potentiated the cytotoxicity of MNNG more than 10-fold but that of MMS was potentiated only about 2-fold showing that O-alkylation of DNA was associated not only with the facilitation of mutagenesis but also with the potentiation of cytotoxicity. The potentiation of MNNG-induced cytotoxicity was maximal in V79 and L1210 cells after only 2 h treatment with dThd, showing that not even one round of DNA replication was necessary for this potentiation. Moreover, dCyd abolished the potentiation, and, at equitoxic concentrations, MNNG induced higher mutant frequencies than did MMS. These data show that the mechanisms by which methylating agents plus dThd induce mutagenesis are fundamentally different from their mechanisms of cytotoxicity.(ABSTRACT TRUNCATED AT 400 WORDS)

Spectroscopic studies of ternary complexes of thymidylate synthetase, deoxyribonucleotides, and folate analogs

Conformational changes accompanying the formation of binary and tightly bound ternary complexes of thymidylate synthetase and all possible combinations of three folate analogs (N-10-ethyl-quinazoline, folic acid triglutamate, and folic acid) and three deoxyribonucleotides (5-fluoro-2'-deoxyuridylic acid (FdUMP), 2'-deoxyuridylic acid (dUMP), and thymidylic acid (dTMP] were studied by means of ultraviolet difference spectroscopy. The amplitudes of the spectral changes upon ternary complex formation were 2-3-fold greater than those generated by formation of binary enzyme-nucleotide and enzyme-folate analog complexes. Difference spectra of the ternary complexes all showed a major increase in absorbance in the region of 320-340 nm, presumably due to perturbations of the folate analog chromophores, whereas decreases in absorbance occurred over a range of 260-310 nm. N-10-ethyl-quinazoline tended to form the complex with the greatest filtration efficiency on nitrocellulose filters, followed by folic acid triglutamate and folic acid, whereas among the nucleotides, the most stable complexes were formed with FdUMP, followed by dUMP and dTMP. A correlation was observed between the apparent stability of the ternary complex and the magnitude of the absorbance change in its difference spectrum. The formation of the various ternary complexes showed three different categories of rate behavior: 1) very rapid formation of the complex; 2) biphasic formation with a rapid phase and a slow phase requiring up to 90 min for completion; and 3) in the case of the ternary complex formed with enzyme, FdUMP, and folic acid, only a slow phase of binding. The slow formation of the latter complex was accompanied by concomitantly slow changes in the difference spectrum. However, in those cases of biphasic formation of the complexes, almost all of the spectral change occurred rapidly, and very little of it corresponded to the slow phase of complex formation. To accommodate these observations, a model is proposed involving a sequential interaction of the two subunits of thymidylate synthetase.

Thymidylate synthetase inhibition in malignant tumors and normal liver of patients given intravenous 5-fluorouracil

Single surgical biopsies of solid tumor were obtained at 20 to 240 min after drug administration in 21 patients given first-dose bolus i.v. 5-fluorouracil (5-FUra), 500 mg/sq m, and assayed for 5-fluorodeoxyuridylate (FdUMP), deoxyuridylate (dUMP), total thymidylate synthetase (TS), and non-FdUMP-bound, free enzyme. Nineteen patients had cancer of gastrointestinal origin, 13 of these colorectal, and 2 patients had breast adenocarcinoma. In 9 patients, synchronous biopsies of surgically normal liver were obtained along with samples of hepatic tumors metastatic from gastrointestinal sites. Total TS averaged 4.18 pmol/g in the malignant tissues and 2.23 pmol/g in liver. FdUMP levels in the gastrointestinal tumors were higher than in normal liver, were highest at the earliest time interval studied, 20 to 30 min, and appeared to decrease exponentially through 120 min. TS inhibition averaged 70 to 80% in gastrointestinal tumor biopsies and less than 50% in normal liver. Levels of dUMP were low and varied little with time. Those gastrointestinal tumors with higher FdUMP:dUMP ratios showed significantly greater TS inhibition. Tumors of 3 patients who benefited from 5-FUra therapy (1 patient with colonic adenocarcinoma and the 2 patients with breast adenocarcinoma) showed greater TS inhibition than did tumors of remaining patients. It is concluded that the apparent time course changes observed in FdUMP, dUMP, and TS in the grouped data are qualitatively similar to findings of murine studies in vivo and that the relationship between FdUMP:dUMP ratios and TS inhibition are consistent with established in vitro enzymic kinetics. Thus, biopsies of tumors at short time periods after 5-FUra administration may be usefully studied for biochemical parameters of TS inhibition, with the objectives of correlation of sensitivity to subsequent 5-FUra therapy and clarification of mechanisms of drug resistance.

Evidence that intracellular synthesis of 5-fluorouridine-5'-phosphate from 5-fluorouracil and 5-fluorouridine is compartmentalized

L1210 cells were exposed to equitoxic concentrations of [14C]5-fluorouracil and [3H]5-fluorouridine for 4 hours. The RNA from these cells was separated into cytosolic and nuclear fractions, and then further fractionated by chromatography on poly-U Sepharose, Sephadex G-200 and DEAE-cellulose. The ratio of tritium to carbon-14 incorporated into various species of RNA differed by as much as 6-fold, indicating that the respective 5-fluorouridine-5'-monophosphates synthesized from the two precursors are localized in separate pools that do not mix rapidly.

Studies on the mechanism of fluoropyrimidine cytotoxicity in l1210 cells: correlation with inhibition of thymidylate synthetase but not with incorporation into RNA

The effects of 5-fluorouracil (FUra), 5-fluorouridine (FUrd), and 5-fluoro-2'-deoxyuridine (FdUrd) on L1210 cells were examined in an effort to determine whether the cytotoxicity of these fluoropyrimidines is more closely associated with incorporation of FUra residues into RNA or inhibition of thymidylate (dTMP) synthetase (5,10-methylenetetrahydrofolate: deoxyuridylate C-methyl-transferase, EC 2.1.2.45) by 5-fluoro-2'-deoxyuridylate (FdUMP). In different batches of cells exposed to equitoxic (LD50) doses of these drugs for 48 hr, the levels of free FdUMP, dUMP, and free dTMP synthetase were found to be very similar. However, the number of FUra residues incorporated into total cellular RNA were in the approximate ratio of 1:10:100 in cells treated with FdUrd, FUrd, and FUra, respectively. Although these results are consistent with a common DNA-directed mechanism of toxicity, thymidine (dThd), which should circumvent dTMP synthetase inhibition, did not rescue the cells from the effects of FUra. However, uridine (Urd), which should compete with FUra for incorporation into RNA, had no effect on the toxicity of FUra either. Urd at 10(-5) M did not decrease the amount of incorporation of 10(-7)M [(3)H]FUra into total RNA, but a limited fractionation of polysomal RNA showed about a 4-fold decrease of incorporation of FUra into mRNA in the presence of Urd. Urd and dThd did effectively decrease the cytotoxicity of FUrd and FdUrd, respectively. These observations suggest that cell rescue experiments may not be reliable indicators of the mechanism of cytotoxicity of antimetabolites with complex mechanisms of action.

Synthesis and biological activity of 5-fluoro-2',3'-dideoxy-3'-fluorouridine and its 5'-phosphate

5-Fluoro-2',3'-dideoxy-3'-fluorouridine (3'-FFdUrd) and 5-fluoro-2',3'-dideoxy-3'-fluorouridine 5'-phosphate (3'-FFdUMP) have been synthesized, and their interactions with thymidine (dThd) phosphorylase and thymidylate (dTMP) synthetase, respectively, have been examined. 3'-FFdUrd is not a substrate for dThd phosphorylase, but is a weak, noncompetitive inhibitor (Ki = 1.7 mM). 3'-FFdUMP inhibits dTMP synthetase competitively with deoxyuridylate (Ki = 0.13 mM) when both the substrate and inhibitor are present simultaneously. However, in the presence of 5,10-methylenetetrahydrofolate, the inhibition increases with time in a first-order manner (konobsd = 0.029 s-1). A complex is formed between [6-3H]3'-FFdUMP and dTMP synthetase, which is isolable on nitrocellulose filters, and has a dissociation rate (koffobsd = 1.4 X 10(-2) min-1) similar to that of the potent inhibitor 5-fluoro-2'-deoxyuridylate (koffobsd = 1.3 X 10(-2) min-1) from its ternary complex with dTMP synthetase. These results are explained in terms of a two-stage model involving the initial formation of a reversible adsorption complex, followed by a slow conversion to a tight-binding catalytic complex.

Human leukemic cells resistant to 5-fluoro-2'-deoxyuridine contain a thymidylate synthetase with lower affinity for nucleotides

A line of human lymphocytic leukemia cells (CCRF-CEM) has been obtained which is 140-fold resistant to the potent cell growth inhibitor 5-fluoro-2'-deoxyuridine (FdUrd). The cells were also 11-fold cross-resistant to 5-fluorouracil. In contrast to several previous studies involving FdUrd-resistant mouse cells, thymidylate synthetase levels were not substantially elevated in these FdUrd-resistant human leukemic cells. Thymidine kinase activity was also unchanged in the resistant cells, although the levels of 5-fluoro-2'-deoxyuridylate (FdUMP), the potent inhibitor of thymidylate synthetase, generated at equimolar doses of FdUrd were about 40% lower than in the sensitive cells. Studies of the kinetics of FdUMP binding to thymidylate synthetase isolated from the FdUrd-resistant cells disclosed a considerably higher dissociation constant (Kd = 1.0 X 10(-9) M) for the ternary covalent enzyme . FdUMP . 5,10-methylene tetrahydrofolate complex compared to the value obtained with enzyme from sensitive cells (Kd = 4.4 X 10(-11) M). The thymidylate synthetase from the FdUrd-resistant cells also showed 17-fold weaker binding of 2'-deoxyuridylate, even though the Km value for 2'-deoxyuridylate was 3-fold lower compared to the enzyme from FdUrd-sensitive cells. The turnover number of the altered enzyme was 1.8-fold higher than that for the normal enzyme but the rate constants for the release of FdUMP from the ternary complex, which is also an enzyme-catalyzed reaction, were identical for both enzymes. Electrophoresis of the radiolabeled ternary complexes on nondenaturing gels showed small but reproducible differences in migration rates. These results demonstrate that the mechanism of resistance to FdUrd in this cell line involves an alteration in the target enzyme, thymidylate synthetase, which causes it have a lower affinity for nucleotides.

Induction of mutations by 5-fluorodeoxyuridine: a mechanism of self-potentiated drug resistance?

In medium containing concentrations of deoxycytidine that occur in vivo, 5-fluorodeoxyuridine induced mutation frequencies 6-90 fold greater than spontaneous mutant frequencies at two genetic loci in Chinese hamster cells. In medium lacking deoxycytidine, 5-fluorodeoxyuridine was more cytotoxic but induced no mutants. Hence, the effectiveness of cancer therapy with 5-fluorodeoxyuridine may be limited by self potentiated development of 5-fluorodeoxyuridine-resistant mutants and enhanced and prolonged by manipulating deoxycytidine metabolism.

Therapeutic response of leukemic mice treated with fluorinated pyrimidines and inhibitors of deoxyuridylate synthesis

The therapeutic efficacy of combinations of fluorinated pyrimidines and inhibitors of either ribonucleotide reductase or deoxycytidylate deaminase was evaluated for the treatment of the L1210 mouse leukemia in DBA/2 mice. Therapeutic synergisms were observed with optimal combinations of 5-fluor-2'-deoxyuridine and either hydroxyurea or guanazole. In addition, mice treated with guanazole combined with 5-fluorouracil survived longer than was observed with any dose of guanazole or with 5-fluorouracil alone. Tetrahydrodeoxyuridine, a potential prodrug of a transition-state analog of deoxycytidylate deaminase, did not have antitumor activity by itself nor did it improve the therapeutic response of leukemic mice to 5-fluoro-2'-deoxyuridine. These results are consistent with the hypothesis that deoxyuridylate accumulation was limited by inhibition of ribonucleotide reductase but not by administration of tetrahydrodeoxyuridine. It is suggested that combination chemotherapy with fluorinated pyrimidines and inhibitors of deoxyuridylate synthesis may improve the therapeutic response to these drugs.

Synthesis and biological activity of 5'-substituted 5-fluoropyrimidine nucleosides

5'-Deoxy-5-fluorouridine (5'-dFUrd, 1) possesses a significantly higher chemotherapeutic index than other fluoropyrimidines as a result of its being selectivity cleaved in tumors to 5-fluorouracil (FUra) by uridine phosphorylase. Because 1 is a relatively poor substrate for this enzyme, we synthesized a series of 5'-deoxy-5'-substituted-5-fluorouridine (FUrd) derivatives in an effort to obtain compounds that might have improved substrate interactions compared to 1 and thus possibly be better prodrugs of FUra. Three derivatives, 5'-O-tosyl-FUrd (13), 5'-O-mesyl-FUrd (14), and 5'-deoxy-5'-bromo-FUrd (15), had cytostatic activity against L1210 and CCRF-CEM leukemic cells in culture superior to that of 1. In preliminary in vivo antitumor studies against L1210 leukemic cells in mice, 5'-deoxy-5'-chloro-FUrd (4), 5'-O-mesyl-FUrd (14), an 5'-deoxy-5'-fluoro-FUrd (18) gave percent increases in life span of 64, 58, and 58, respectively, compared to a value of 20 for compound 1.

Thymidylate synthetase - substrate complex formation

The complexes that thymidylate synthetase (TSase) forms with various potent inhibitors have been intensively studied and thoroughly reviewed. Of particular significance is the covalent ternary complex of TSase-FdUMP-5,10-CH2H4PteGlu. FdUMP is the active metabolite of the widely used anti-cancer drug 5-fluorouracil. This complex is thought to be analogous to a steady-state intermediate of the normal enzyme reaction with the substrate dUMP. In this review, we examine the properties of TSase-dUMP complexes in order to determine if there is an experimental basis for drawing a close analogy between dUMP and FdUMP in their interaction with TSase, and also to evaluate data indicating a potential chemotherapeutic value for TSase-dUMP complexes formed in the presence of folate analogs.

Tight-binding complexes of thymidylate synthetase, folate analogs and deoxyribonucleotides

Tightly-bound ternary complexes are formed when dTMP synthetase is incubated with deoxyribonucleotides such as FdUMP, dUMP, and dTMP, and folate analogs, presumably mediated by conformational changes in the enzyme induced by these ligands. The structure of both the folate analog and the nucleotide determines the tightness of binding in the ternary complex. For example, the first order rate constant for dissociation of FdUMP from the dTMP synthetase-FdUMP PteGlu complex is almost 100-fold larger than from the dTMP synthetase-FdUMP-PteGlu3 complex. The latter had a slower rate of dissociation than did the covalent dTMP synthetase-FdUMP-5,10-CH4H4Pte- Glu complex, which had been shown to have a Kd of 10(-11) M. FdUMP formed tighter complexes than did either dUMP or dTMP when using PteGlu3, PteGlu, or H2PteGlu as the folate components, but with 5,8-deazafolate complexes formed with all three nucleotides appeared to have about the same stability. H2PteGlu also induced the tight binding of dUMP to dTMP synthetase. A Scatchard plot indicated positive cooperativity in the binding of dUMP to both human and bacterial enzyme in the presence of H2PteGlu. These results suggest the possibility that when cells are exposed to MTX the decrease in dTMP synthetase activity may be due largely to formation of inhibitory enzyme-dUMP-H2PteGlun complexes and does not result primarily from depletion of reduced folates because of dihydrofolate reductase inhibition. Because this complex has a slow rate of dissociation, dTMP synthetase activity would still be inhibited for some time after an influx of reduced folates (as in leucovorin rescue) sufficient to support the enzyme reaction under normal conditions.

Large-scale and small-scale methods for the preparation of 5,10-methylenetetrahydrofolate

Two procedures have been developed for the synthesis and isolation of 5,10-methylenetetrahydrofolate, the cofactor for the reaction catalyzed by thymidylate synthetase, one of which can be used for large-scale preparations of the cofactor and the other for small-scale syntheses especially suitable for obtaining the radiolabeled cofactor. The large-scale procedure involves treatment of folic acid with dithionite to give dihydrofolate, which is then converted to tetrahydrofolate by dihydrofolate reductase (L. casei). The small-scale method involves a direct enzymatic reduction of folic acid to tetrahydrofolate by dihydrofolate reductase, and has been used to prepare the double-labeled 5,10-[14C]methylene[3',5',7,9-3H]tetrahydrofolate. In both procedures, after the reduction steps have been performed, the tetrahydrofolate is treated in situ with formaldehyde prior to purification by DEAE-cellulose chromatography, thus allowing the isolation of 5,10-methylenetetrahydrofolate as a dry powder after lyophilization. This product is active in the enzyme reaction without the further addition of excess formaldehyde as in previous procedures. The cofactor prepared in this manner has much improved stability toward oxidation compared to free tetrahydrofolate.

Interaction of 5-ethynyl-2'-deoxyuridylate with thymidylate synthetase

The interaction of 5-ethynyl-2'-deoxyuridylate (5-ethynyl-dUMP; 1) with thymidylate (dTMP) synthetase has been investigated. The compound was an inhibitor of the enzyme, competitive with 2'-deoxyuridylate (dUMP) when the reaction was initiated by addition of enzyme (Ki = 2.7 X 10(-6) M). However, upon preincubation of 1 with dTMP synthetase, the inhibition pattern became noncompetitive. The time course of the enzyme reaction in the presence of 1 was nonlinear, indicating an increase in binding with time. Irreversible inactivation of the enzyme did not occur. The compound did not appear to become altered structurally as a result of interaction with the enzyme. A ternary complex was formed among dTMP synthetase, compound 1, and 5,10-methylenetetrahydrofolate, which was stable enough to survive Sephadex G-25 filtration but dissociated upon denaturation of the enzyme.

Synthesis and biological activity of 5-fluoro-2'-deoxyuridine 5'-phosphorodiamidates

Three 5'-phosphorodiamidate derivatives of 5-fluoro-2'-deoxyuridine (FdUrd), 5-fluoro-2'-deoxyuridine 5'-phosphorodiamidate (4a), 5'-phosphorodiimidazolidate (4b), and 5'-phosphorodimorpholidate (4c), were synthesized by aminolysis of 5-fluoro-2'-deoxyuridine 5'-phosphorodichloridate with the respective amine. In culture, these 5'-phosphorodiamidates inhibited the growth of murine leukemia (L5178Y) cells. 5-Fluoro-2'-deoxyuridine 5'-phosphorodiamidate (4a) was the most active derivative and, on a molar basis, produced a cytostatic effect comparable to that of FdUrd and 5-fluoro-2'-deoxyuridine 5'-monophosphate (FdUrd-5'-P). Compounds 4b and 4c were less active than 4a, with relative rates of activity 4a > 4b > 4c that corresponded to their rates of hydrolysis to FdUrd-5'-P. None of the 5'-phosphorodiamidates inhibited thymidylate synthetase of concentrations up to 1 mM.

Hydrodynamic behavior of human and bacterial thymidylate synthetases and thymidylate synthetase--5-fluoro-2'-deoxyuridylate--5,10-methylenetetrahydrofolate complexes. Evidence for large conformational changes during catalysis

The conformations of thymidylate synthetases from CCRF-CEM human leukemic cells and Lactobacillus casei were studied by hydrodynamic methods. Although the human enzyme has a molecular weight of 70 000--72 500, somewhat smaller than that of the L. casei enzyme, it has a larger Stokes radius and a lower sedimentation coefficient, indicating that the human enzyme is less spherical than the bacterial enzyme. Thymidylate synthetases from the human leukemic cells and the bacterial source both undergo substantial conformational changes upon the formation of a covalent ternary complex with the mechanism-based inhibitor 5-fluoro-2'-deoxyuridylate and 5,10-methylenetetrahydrofolate. The Stokes radius of both proteins decreases by 3.5% when the ternary complex is formed in spite of the 1.8% increase in molecular weight, and the sedimentation coefficient increases by 3.5% after appropriate corrections for the bound ligands. Ternary complex formation results in a more compact structure for both enzymes, with approximately the same reduction in the frictional ratio. Experiments with the bacterial enzyme indicate that approximately 70% of the total conformational change occurs upon binding of 1 mol of ligands/mol of enzyme. These results demonstrate that human and bacterial thymidylate synthetases undergo marked structural changes upon forming a ternary complex which is probably very similar to an activated complex formed with both substrates. These investigations also provide evidence for fundamental similarities in the mechanism of ternary complex formation with 5-fluoro-2'-deoxyuridylate and 5,10-methylenetetrahydrofolate, despite the marked differences in amino acid composition and the dissimilar conformations of these two enzymes obtained from widely divergent sources.

Thymidylate synthetase and 2'-deoxyuridylate form a tight complex in the presence of pteroyltriglutamate

Thymidylate synthetases of human and bacterial origin form a tightly bound complex with the substrate dUMP in the presence of pteroyltriglutamate. This complex and the weaker enzyme . dUMP binary complex can be isolated and conveniently assayed by nitrocellulose disc filtration using [6-3H]dUMP as the radioactive ligand. Intact thymidylate synthetase . dUMP . pteroyltriglutamate complex can be obtained by gel filtration chromatography on Sephadex G-25, but the binary enzyme . dUMP complex dissociates under the same conditions. Scatchard plots show the presence of two nonequivalent dUMP binding sites on the enzyme for the pteroyltriglutamate complex, with dissociation constants of 5 and 95 nM compared to 730 nM for the binary complex. The implications of these findings for folate analog inhibition of thymidylate synthetase are discussed.