Pyrimidine nucleobase ligands of orotate phosphoribosyltransferase from Toxoplasma gondii.

Sixty-seven pyrimidine nucleobase analogues were evaluated as ligands of Toxoplasma gondii orotate phosphoribosyltransferase (OPRTase, EC 2.4.2.10) by measuring their ability to inhibit this enzyme in vitro. Apparent Ki values were determined for compounds that inhibited T. gondii OPRTase by greater than 20% at a concentration of 400 microM. 1-Deazaorotic acid (0.47 microM) and 5-azaorotic acid (2.1 microM) were found to bind better (8.3- and 1.9-fold, respectively) to T. gondii OPRTase than orotic acid, the natural substrate of the enzyme. Based on these results, a structure-activity relationship of ligand binding to OPRTase was formulated using uracil, barbituric acid, and orotic acid as reference compounds. It was concluded that the following structural features of pyrimidine nucleobase analogues were required or strongly preferred for binding: (i) an endocyclic pyridine-type nitrogen or methine at the 1-position; (ii) exocyclic oxo groups at the 2- and 4-positions; (iii) a protonated endocyclic pyridine-type nitrogen at the 3-position; (iv) an endocyclic pyridine-type nitrogen or methine at the 5-position; (v) an exocyclic hydrogen or fluorine at the 5-position; (vi) an endocyclic pyridine-type nitrogen or methine at the 6-position; and (vii) an exocyclic negatively charged or electron-withdrawing group at the 6-position. A comparison of the results from the present study with those from a previous study on mammalian OPRTase [Niedzwicki et al., Biochem Pharmacol 33: 2383-2395, 1984] identified four compounds (6-chlorouracil, 5-azaorotic acid, 1-deazaorotic acid, and 6-iodouracil) that may bind selectively to T. gondii OPRTase.

Effects of modifications in the pentose moiety and conformational changes on the binding of nucleoside ligands to uridine phosphorylase from Toxoplasma gondii.

One hundred and fifty analogues of uridine, with various modifications to the uracil and pentose moieties, have been tested and compared with uridine with respect to their potency to bind to uridine phosphorylase (UrdPase, EC 2.4.2.3) from Toxoplasma gondii. The effects of the alpha- and beta-anomers, the L- and D-enantiomers, as well as restricted syn and anti rotamers, on binding were examined. Pseudo-, lyxo-, 2,3'-anhydro-2'-deoxy-, 6,5'-cyclo-, 6,3'-methano-, O5',6-methano- and carbocyclic uridines did not bind to the enzyme. Ribosides bound better than the corresponding xylosides, which were better than the deoxyribosides. The binding of deoxyribosides was in the following manner: 2',3'-dideoxynucleosides > 2',5'-dideoxynucleosides > 2'-deoxyribosides > 3'- and 5'-deoxyribosides. alpha-2'-Deoxyribosides bound to the enzyme, albeit less tightly than the corresponding beta-anomers. The acyclo- and 2,2'-anhydrouridines bound strongly, with the 2,2'-anhydro-derivatives being the better ligands. 2,5'-Anhydrouridine bound to UrdPase less effectively than 2,2'-anhydrouridine and acyclouridine. Arabinosyluracil was at best a very poor ligand, but bound better if a benzyl group was present at the 5-position of the pyrimidine ring. This binding was enhanced further by adding a 5-benzyloxybenzyl group. A similar enhancement of the binding by increased hydrophobicity at the 5-position of the pyrimidine ring was observed with ribosides, alpha- and beta-anomers of the 2'-deoxyribosides, acyclonucleosides, and 2,2'-anhydronucleosides. Among all the compounds tested, 5-(benzyloxybenzyl)-2,2'-anhydrouridine was identified as the best ligand of T. gondii UrdPase with an apparent Ki value of 60 +/- 3 nM. It is concluded that the presence of an N-glycosyl bond is a prerequisite for a nucleoside ligand to bind to T. gondii UrdPase. On the other hand, the presence of a 2'-, 3'-, or 5'-hydroxyl group, or an N-glycosyl bond in the beta-configuration, enhanced but was not essential for binding. Furthermore, the potency of the binding of 2,2'-anhydrouridines (fixed high syn isomers) in contrast to the weaker binding of the 6,1'-anhydro- or 2,5'-anhydrouridines (fixed syn isomers), and the complete lack of binding of the 6,5'-cyclo, O5',6-methano- and 6,3'-methanouridines (fixed anti isomers) to T. gondii UrdPase indicate that the binding of ligands to this enzyme is in the syn/high syn conformation around the N-glycosyl bond. The results also indicate that the parasite but not the mammalian host UrdPase can participate in hydrogen bonding with N3 of the pyrimidine ring of nucleoside ligands. T. gondii UrdPase also has a larger hydrophobic pocket adjacent to the C5 of the pyrimidine moiety than the host enzyme, and can accommodate modifications in the pentose moiety which cannot be tolerated by the host enzyme. Most prominent among these modifications is the absence and/or lack of the ribo orientation of the 3'-hydroxyl group, which is a requirement for a ligand to bind to mammalian UrdPase. These differences between the parasite and host, enzymes can be useful in designing specific inhibitors or "subversive" substrates for T. gondii UrdPase.

Structure-activity relationships for the binding of ligands to xanthine or guanine phosphoribosyl-transferase from Toxoplasma gondii.

Preliminary characterization of Toxoplasma gondii phosphoribosyltransferase activity towards purine nucleobases indicates that there are at least two enzymes present in these parasites. One enzyme uses hypoxanthine, guanine, and xanthine as substrates, while a second enzyme uses only adenine. Furthermore, competition experiments using the four possible substrates suggest that there may be a third enzyme that uses xanthine. Therefore, sixty-eight purine analogues and thirteen related derivatives were evaluated as ligands of T. gondii phosphoribosyltransferase, using xanthine or guanine as substrates, by examining their ability to inhibit these reactions in vitro. Inhibition was quantified by determining apparent Ki values for compounds that inhibited these activities by greater than 10% at a concentration of 0.9 mM. On the basis of these data, a structure-activity relationship for the binding of ligands to these enzymes was formulated using hypoxanthine (6-oxopurine) as a reference compound. It was concluded that the following structural features of purine analogues are required or strongly preferred for binding to both enzymes: (1) a pyrrole-type nitrogen (lactam form) at the 1-position; (2) a methine (= CH-), a pyridine type nitrogen (= N-), or an exocyclic amino or oxo group at the 2-position; (3) no exocyclic substituents at the 3-position; (4) an exocyclic oxo or thio group in the one or thione tautomeric form at the 6-position; (5) a pyridine-type nitrogen (= N-) or a methine group at the 7-position; (6) a methine group at the 8-position; (7) a pyrrole-type nitrogen or a carbon at the 9-position; and (8) no exocyclic substituents at the 9-position. These findings provide the basis for the rational design of additional ligands of hypoxanthine, guanine, and xanthine phosphoribosyltransferase activities in T. gondii.

Structure-activity relationship for the binding of nucleoside ligands to adenosine kinase from Toxoplasma gondii.

One hundred and twenty-eight purine nucleoside analogs were evaluated as ligands of Toxoplasma gondii adenosine kinase (EC 2.7.1.20) by examining their ability to inhibit this enzyme in vitro. Inhibition was quantified by determining apparent Ki (appKi) values for those compounds that inhibited this enzyme by greater than 10% at a concentration of 1 mM. Two compounds, N6-(p-methoxybenzoyl)adenosine and 7-iodo-7-deazaadenosine (iodotubercidin), were found to bind to the enzyme (appKi = 3.9 and 1.6 microM, respectively) better than adenosine. On the basis of these data, a structure-activity relationship for the binding of ligands to T. gondii adenosine kinase was formulated using adenosine as a reference compound. It was concluded that the following structural features of purine nucleoside analogs are required or strongly preferred for binding: (1) "pyridine-type" endocyclic nitrogens at the 1- and 3-positions; (2) an exocyclic hydrogen at the 2-position; (3) 6-position exocyclic substituents in the lactim tautomeric form; (4) a "pyridine-type" endocyclic nitrogen at the 7-position or hydrophobic exocyclic substituents attached to an endocyclic carbon at the 7-position; (5) an endocyclic methine or "pyridine-type" nitrogen at the 8-position; (6) an endocyclic nitrogen at the 9-position; (7) a pentose or "pentose-like" (e.g. hydroxylated cyclopentene) moiety attached to the 9-position nitrogen; (8) hydroxyl groups at the 2'- and 3'-positions in a ribose configuration; (9) a hydroxymethyl or methyl (i.e. 5'-deoxy) group at the 5'-position; (10) a beta-D-nucleoside configuration; and (11) an anti conformation around the N-glycosidic bond. In addition, there appears to be a "pocket" in the catalytic site of T. gondii adenosine kinase, adjacent to the 6-position of adenosine, that can accommodate large (preferably unsaturated or aromatic) substituents (e.g. phenyl). These findings provide the basis for the rational design of additional ligands of T. gondii adenosine kinase.

Structure-activity relationship of ligands of uracil phosphoribosyltransferase from Toxoplasma gondii.

One hundred compounds were evaluated as ligands of Toxoplasma gondii uracil phosphoribosyltransferase (UPRTase, EC 2.4.2.9) by examining their ability to inhibit this enzyme in vitro. Inhibition was quantified by determining apparent Ki values for those compounds that inhibited T. gondii UPRTase by greater than 10% at a concentration of 2 mM. Five compounds (4-thiopyridine, 2-thiopyrimidine, trithiocyanuric acid, 1-deazauracil and 2,4-dithiouracil) bound to the enzyme better than two known substrates for T. gondii UPRTase, 5-fluorouracil and emimycin, which have antitoxoplasmal activity (Pfefferkorn ER, Exp Parasitol 44: 26-35, 1978; Pfefferkorn et al., Exp Parasitol 69: 129-139, 1989). In addition, several selected compounds were evaluated as substrates for T. gondii UPRTase, and it was found that 2,4-dithiouracil is also a substrate for this enzyme. On the basis of these data, a structure-activity relationship for the binding of ligands to T. gondii UPRTase was determined using uracil as a reference compound.

Structure-activity relationship of nucleobase ligands of uridine phosphorylase from Toxoplasma gondii.

Seventy-nine nucleobase analogs were evaluated as potential inhibitors of Toxoplasma gondii uridine phosphorylase (UrdPase), and the apparent Ki (appKi) values for these compounds were determined. Based on the inhibition data, a structure-activity relationship for the binding of nucleobase analogs to the enzyme was formulated, using uracil as a reference compound. Two compounds were identified as very potent inhibitors of T. gondii UrdPase, 5-benzyloxybenzylbarbituric acid and 5-benzyloxybenzyluracil, which had appKi values of 0.32 and 2.5 microM, respectively. A comparison of the results from the present study, with similar studies on mammalian UrdPase and thymidine phosphorylase (dThdPase) (Niedzwicki et al., Biochem Pharmacol 32: 399-415, 1993) revealed that there are both similarities and differences between the catalytic site of T. gondii UrdPase and the catalytic sites of the mammalian enzymes with respect to binding of uracil analogs. One compound, 6-benzyl-2-thiouracil, was identified as a potent, specific inhibitor (appKi = 14 microM) of T. gondii UrdPase, relative to mammalian UrdPase and dThdPase.

Pyrimidine salvage pathways in Toxoplasma gondii.

Pyrimidine salvage enzyme activities in cell-free extracts of Toxoplasma gondii were assayed in order to determine which of these enzyme activities are present in these parasites. Enzyme activities that were detected included phosphoribosyltransferase activity towards uracil (but not cytosine or thymine), nucleoside phosphorylase activity towards uridine, deoxyuridine and thymidine (but not cytidine or deoxycytidine), deaminase activity towards cytidine and deoxycytidine (but not cytosine, cytidine 5'-monophosphate or deoxycytidine 5'-monophosphate), and nucleoside 5'-monophosphate phosphohydrolase activity towards all nucleotides tested. No nucleoside kinase or phosphotransferase activity was detected, indicating that T. gondii lack the ability to directly phosphorylate nucleosides. Toxoplasma gondii appear to have a single non-specific uridine phosphorylase enzyme which can catalyze the reversible phosphorolysis of uridine, deoxyuridine and thymidine, and a single cytidine deaminase activity which can deaminate both cytidine and deoxycytidine. These results indicate that pyrimidine salvage in T. gondii probably occurs via the following reactions: cytidine and deoxycytidine are deaminated by cytidine deaminase to uridine and deoxyuridine, respectively; uridine and deoxyuridine are cleaved to uracil by uridine phosphorylase; and uracil is metabolized to uridine 5'-monophosphate by uracil phosphoribosyltransferase. Thus, uridine 5'-monophosphate is the end-product of both de novo pyrimidine biosynthesis and pyrimidine salvage in T. gondii.

Cloning and characterization of a cDNA coding for the alpha-subunit of a stimulatory G protein from Schistosoma mansoni.

Guanine nucleotide-binding proteins (G proteins) mediate signals between serotonin receptors and adenylate cyclase in Schistosoma mansoni. A bovine Gs alpha cDNA probe was used to isolate a cDNA clone, SG12, encoding the entire alpha-subunit of a G protein of S. mansoni. The cDNA is 1897 base pairs long, contains an open reading frame of 1137 base pairs, and codes for a deduced protein of 379 amino acids. The putative protein encoded by the clone has an exact amino acid match with bovine Gs alpha of 65% and a 78% match when conserved amino acid substitutions are considered. In contrast, the exact and conserved matches of the schistosome alpha-subunit with bovine Gi are 41 and 61%, respectively. A comparison of the deduced amino acid sequence of SG12 with a variety of different G alpha proteins indicates that all the major structural features characteristic of a Gs alpha protein are present in the S. mansoni gene. The schistosome clone contains the putative site for ADP-ribosylation by cholera toxin found in Gs alpha but does not contain the ADP-ribosylation site for pertussis toxin present in Gi alpha. The amino acids are completely conserved at the GTP-binding sites. On a Northern blot, the cDNA hybridizes to a major band of 3.1 kilobases in RNA from adult schistosomes. The message appears to be absent in miracidia and cercariae, but a faint 3.1-kilobase band is visible in the early schistosomule stage preceding adulthood. This evidence, when added to previous biochemical data, indicates that the expression of this gene is developmentally controlled.

Uridine phosphorylase from Schistosoma mansoni.

Uridine phosphorylase is the only pyrimidine nucleoside cleaving activity that can be detected in extracts of Schistosoma mansoni. The enzyme is distinct from the two purine nucleoside phosphorylases contained in this parasite. Although Urd is the preferred substrate, uridine phosphorylase can also catalyze the reversible phosphorolysis of dUrd and dThd, but not Cyd, dCyd, or orotidine. The enzyme was purified 170-fold to a specific activity of 2.76 nmol/min/mg of protein with a 16% yield. It has a Mr of 56,000 as determined by molecular sieving on Sephadex G-100. The mechanism of uridine phosphorylase is sequential. When Urd was the substrate, the KUrd = 13 microM and the KPi = 533 +/- 78 microM. When dThd was used as a substrate, the KdThd = 54 microM and the KPi = 762 +/- 297 microM. The Vmax with dThd was 53 +/- 9.8% that of Urd. dThd was a competitive inhibitor when Urd was used as a substrate. The enzyme showed substrate inhibition by Urd, dThd (greater than 0.125 mM) and phosphate (greater than 10 mM). 5-(Benzyloxybenzyloxybenzyl)acyclouridine was identified as a potent and specific inhibitor of parasite (Ki = 0.98 microM) but not host uridine phosphorylase. Structure-activity relationship studies suggest that uridine phosphorylase from S. mansoni has a hydrophobic pocket adjacent to the 5-position of the pyrimidine ring and indicate differences between the binding sites of the mammalian and parasite enzymes. These differences may be useful in designing specific inhibitors for schistosomal uridine phosphorylase which will interfere selectively with nucleic acids synthesis in this parasite.

Phosphofructokinase from Fasciola hepatica: activation by phosphorylation and other regulatory properties distinct from the mammalian enzyme.

Phosphofructokinase from the liver fluke, Fasciola hepatica, was phosphorylated by the catalytic subunit of cyclic AMP-dependent protein kinase isolated from this organism. Phosphorylated fluke phosphofructokinase had a sevenfold lower apparent Km for its substrate, Fru-6-P, and an eightfold higher 0.5 Vopt for ATP, the enzyme's primary inhibitor, than native phosphofructokinase. Activation of fluke phosphofructokinase following phorphorylation by a mammalian protein kinase catalytic subunit was previously reported (E. S. Kamemoto and T. E. Mansour (1986) J. Biol. Chem. 261, 4346-4351). The catalytic subunit of protein kinase isolated from the liver fluke phosphorylated sites on fluke phosphofructokinase similar to those phosphorylated by the mammalian enzyme. Maximal phosphate incorporation was 0.3 mol P/mol of protomer. The native enzyme was found to contain 1.3 mol P/mol of protomer. In contrast to fluke phosphofructokinase, activity of the mammalian heart enzyme was slightly decreased following phosphorylation. The dependence of allosteric interaction on an acidic pH observed with the mammalian phosphofructokinase was not observed with the fluke enzyme. Unlike mammalian phosphofructokinase, allosteric kinetics of the fluke enzyme was observed at alkaline pH (8.0). Fluke phosphofructokinase was found to be relatively insensitive to inhibition by citrate, a known potent inhibitor of the mammalian enzyme. Fru-2,6-P2, a potent modifier of phosphofructokinase from a variety of sources, was found to activate both native and phosphorylated fluke phosphofructokinase. The most potent activators of fluke phosphofructokinase were found to be Fru-2,6-P2, AMP, and phosphorylation. The endogenous level of Fru-2,6-P2 in the flukes was determined to be 29 +/- 1.3 nmol/g wet wt, a level that may well modulate enzyme activity. Fru-6-P,2-kinase, the enzyme responsible for synthesis of Fru-2,6-P2, was found to be present in the flukes. Our results suggest physiological roles for phosphorylation and Fru-2,6-P2 in regulation of fluke phosphofructokinase.

Partial purification and characterization of cAMP-dependent protein kinase from Fasciola hepatica.

Three cyclic AMP (cAMP)-dependent protein kinases, designated A1, A2, and B, were isolated from the liver fluke Fasciola hepatica using Phenyl-Sepharose and DEAE-cellulose chromatography. These enzymes differed with respect to activation by cAMP and their molecular weights. The half-maximal activation constant for cAMP-dependent protein kinases A1 and B was 20 nM, while that of A2 was about five-fold higher (110 nM). The estimated molecular weights for cAMP-dependent protein kinases A1 and A2 (both 98,000) suggest a dimeric form for these enzymes; whereas, the higher molecular weight for cAMP-dependent protein kinase B (187,000) indicates that this enzyme is a tetramer. The physical and kinetic properties of the catalytic subunit of fluke cAMP-dependent protein kinase were similar to those reported for the mammalian enzyme. The molecular weight of the catalytic subunit was estimated to be 41,000. The pH optimum for the enzyme was 6.0, 6.5, or 7.0 when casein, histone, or protamine were used as substrates. The protein substrate specificity was in the order histone greater than arginine-rich histone greater than casein greater than protamine greater than lysine-rich histone. Free Mg2+ 'stimulated' enzyme activity at low concentrations (0.5 to 5 mM), whereas at higher concentrations (greater than 5 mM) it became inhibitory. Of the divalent cations tested, only Co2+ and Mn2+ could substitute for Mg2+. Kinetic studies indicated that the reaction mechanism of this enzyme is sequential and that MgATP and MgADP are competitive ligands. Reconstitution experiments using the subunits of fluke and bovine heart cAMP-dependent protein kinase showed that there is sufficient structural homology between these enzymes such that the catalytic subunit from one species can combine with the regulatory subunit of the other species to form inactive holoenzyme. Thus, the present results indicate that cAMP-dependent protein kinase from F. hepatica is similar but not identical to the mammalian enzyme.

Effects of N,N-dimethylformamide and sodium butyrate on enzymes of pyrimidine metabolism in cultured human tumor cells.

Effects of a 7-day treatment with the maturational agents DMF and sodium butyrate on enzymes of pyrimidine metabolism, growth rate and cell maturation were assessed in 5 human tumor cell lines, ARH-77 (myeloma), K-562 (chronic myeloid leukemia), KG-1 (myeloid leukemia), HL-60 (promyelocytic leukemia) and RWLy-1 (non-Hodgkin's lymphoma). DMF lengthened the doubling times of all five cell lines while sodium butyrate lengthened only those of K-562, HL-60 and RWLy-1. Full maturation was induced only in HL-60 by either agent and in K-562 by butyrate. Exposure resulted in a decreased activity of the anabolic enzyme orotate phosphoribosyltransferase (EC 2.4.2.10) and increased activities of the catabolic enzymes thymidine phosphorylase (EC 2.4.2.4) and dihydrouracil dehydrogenase (EC 1.3.1.2). Changes in the amphibolic enzyme, uridine phosphorylase (EC 2.4.2.3) did not follow any apparent pattern. This study indicates that the pattern of pyrimidine metabolism differs between the differentiated and slowly growing, and undifferentiated rapidly growing counterpart of several human tumors, suggesting that enzymes of pyrimidine metabolism can be used as markers for cellular growth and/or maturity.

Kinetic studies of thymidine phosphorylase from mouse liver.

Initial velocity and product inhibition studies of thymidine phosphorylase from mouse liver revealed that the basic reaction mechanism of this enzyme is a rapid equilibrium random bi-bi mechanism with an enzyme-phosphate-thymine dead-end complex. Thymine displayed both substrate inhibition and nonlinear product inhibition, i.e., slope and intercept replots vs. 1/[thymine] were nonlinear, indicating that there is more than one binding site on the enzyme for thymine and that when thymine is bound to one of these sites, the enzyme is inhibited. Furthermore, both thymidine and phosphate showed "cooperative effects" in the presence of thymine at concentrations above 60 microM, suggesting that the enzyme may have multiple interacting allosteric and/or catalytic sites. The deoxyribosyl transferase reaction catalyzed by this enzyme is phosphate-dependent, requires nonstoichiometric amounts of phosphate, and can proceed by an "enzyme-bound" 2-deoxyribose 1-phosphate intermediate. These findings are in accord with the rapid equilibrium random bi-bi mechanism and demonstrate that deoxyribosyl transfer by this enzyme involves an indirect-transfer mechanism. These results strongly suggest that phosphorolysis and deoxyribosyl transfer are catalyzed by the same site on thymidine phosphorylase.

Structure-activity relationship of pyrimidine base analogs as ligands of orotate phosphoribosyltransferase.

Eighty pyrimidine base analogs were evaluated as inhibitors of mouse liver orotate phosphoribosyltransferase (OPRTase, EC 2.4.2.10). Based on these findings and an extensive literature review, a structure-activity relationship has been formulated for the binding of pyrimidine base analogs to OPRTase. This study provides a basis for the rational design of new inhibitors of this enzyme, and several such compounds are proposed. Additionally, 4,6-dihydroxypyrimidine has been found to be a potent OPRTase inhibitor. Eleven OPRTase inhibitors were also evaluated as inhibitors of orotidine 5'-monophosphate decarboxylase (ODCase, EC 4.1.2.23). 5-Azauracil, 5-azaorotate, and barbituric acid inhibited ODCase significantly only after preincubation with PRPP and MgCl2 in the presence of cytosol.

Enzymes of uridine 5'-monophosphate biosynthesis in Schistosoma mansoni.

In Schistosoma mansoni, the major product of in vitro orotate metabolism was orotidine 5'-monophosphate (OMP), whereas in mouse liver it was UMP. In contrast to mammalian cells, OMP appeared not to be 'channeled' from orotate phosphoribosyltransferase to OMP decarboxylase in S. mansoni, resulting in substantial degradation of OMP to orotidine. Significant differences were observed in the inhibitor specificity of phosphoribosyltransferase between S. mansoni and mouse liver, indicating that this enzyme may be a potential chemotherapeutic target in S. mansoni. Two distinct phosphoribosyltransferases were found in S. mansoni. One enzyme, having the higher molecular weight, utilized orotate, 5-fluorouracil and uracil as substrates, while the other only orotate. Both enzymes were inhibited by 5-azaorotic acid (oxonic acid) but only the 'orotate-specific' enzyme was inhibited by 4,6-dihydroxypyrimidine. OMP decarboxylase activity co-eluted with both phosphoribosyltransferases from Sephadex G-100 gel chromatography. We suggest that phosphoribosyltransferase in S. mansoni plays a role in both de novo UMP biosynthesis as well as in the salvage of uracil and uridine.

Potentiation of 5-fluoro-2'-deoxyuridine antineoplastic activity by the uridine phosphorylase inhibitors benzylacyclouridine and benzyloxybenzylacyclouridine.

At a nontoxic dose (50 microM), the two potent uridine phosphorylase inhibitors, benzylacyclouridine and benzyloxybenzylacyclouridine (BBAU), potentiated 5-fluoro-2'-deoxyuridine (FdUrd) growth inhibition of human pancreatic carcinoma (DAN) and, to a lesser extent, human lung carcinoma (LX-1) cells in culture. BBAU was more effective than benzylacyclouridine. BBAU (50 microM) enhanced the cytocidal effect of FdUrd (1 microM, 3 hr) on DAN grown on soft agar from 75 to 88%. In antithymocyte serum-immunosuppressed mice bearing DAN, the mean tumor weight in animals treated with FdUrd (50 mg/kg/day for 2 days) was 11% less than that of untreated controls. When BBAU (10 mg/kg/day for 2 days) was coadministered, the mean tumor weight at Day 10 was 78% less than untreated controls, with no apparent host toxicity, clearly demonstrating the potentiation of the antitumor effects of FdUrd by BBAU. The fact that DAN responded better than LX-1 to benzylacyclouridine and BBAU could be due, in part, to the lower relative activity of thymidine phosphorylase to uridine phosphorylase in DAN compared to LX-1. The activities of other enzymes involved in FdUrd metabolism, thymidine kinase, uridine kinase, orotate phosphoribosyltransferase, 5'-nucleotidase, and dihydrouracil dehydrogenase, did not differ between the two cell lines.