stereochemical preferences ... fluorouracil and showed that this compound had potent an- ... pro-S hydrogen at carbon 5 and the hydrogen at the chiral carbon 6 ...
THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1990 by The American Society for Biochemistry
Vol. 265, No. 24, Issue of August 25, pp. 14242-14249,199O Printed in U.S.A.
and Molecular Biology, Inc.
Synthesis and Antiproliferative three-5-Fluoro-L-dihydroorotate”
Activity
of (Receivedfor publication, March 2, 1990)
Pradipsinh
K. RathodS
and Ashok
From the Institute for Biomolecular Washington, D. C. >0064
Studies,
Khatri of Biology,
Department
Fluorinated compounds play an important role in enzymology as well as clinical medicine. Based on the stereochemical preferences of dihydroorotate oxidase and enzymes that use fluoroaspartate, it was anticipated that threo-5-fluoro-L-dihydroorotate (t-FDHO) would have the properties of an antimetabolite. Thus, t-FDHO was synthesized via the reduction of B-fluoroorotate using NADH and dihydroorotate dehydrogenase that was free of dihydroorotase. When the product was purified and studied by high field proton and carbon 13 NMR, the fluorine, the five carbons, and all the nonexchangeable protons were readily observed. Confirmation of threo configuration was obtained by examining the vicinal coupling constants between the substituents on carbon 5 and carbon 6 of the newly synthesized compound. Additionally, t-FDHO could be reoxidized to 5-fluoroorotate in the presence of dihydroorotate dehydrogenase and NAD+. Treatment of tFDHO with dihydroorotase generated N-carbamylthreo-3-fluoro-L-aspartate (CTF-ASP) which was also purified and characterized by NMR. The antiproliferative activity of t-FDHO was determined against a diploid human fibrosarcoma cell line (HT-1080). Fifty PM t-FDHO caused 50% inhibition of HT-1080 cell proliferation. During the 46-h toxicity study, extracellular t-FDHO underwent significant hydrolysis to CTF-ASP. Further extracellular degradation to fluoroaspartate was not seen. The antiproliferative activity of t-FDHO was not due to extracellular degradation since CTF-ASP itself was essentially nontoxic.
Substitution of a fluorine for a proton has played an important role in the development of antimetabolites (Walsh, 1983). Due to its small Van der Waal’s radius, substitution of a fluorine for a proton represents an insignificant stearic problem, but the strong electronegative effect of fluorine can seriously alter the biochemical properties of a compound. This, in turn, can effect the fate of the cell that actively metabolizes fluorinated compounds. Perhaps the best studied and clinically relevant fluorinated antimetabolites are the family of 5-fluorinated pyrimidines. In 1957, Heidelberger and co-workers (Heidelberger et al., 1957) synthesized 5fluorouracil and showed that this compound had potent antiproliferative activity against mammalian cells. Since then, 5-fluoropyrimidine bases and nucleosides have been tested for cancer chemotherapy and against select microbial cells (Hei*This work was supported in part by Grant AI-26912 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Z#To whom correspondence should be addressed.
The
Catholic
University
of America,
delberger et al., 1983; Bennett, 1977; Rathod, et al., 1989; Gomez and Rathod, 1990). 5-Fluoropyrimidines have also served as important biochemical tools. Enzymologists have used these compounds to study methylation reactions, and cell biologists have followed genetic traits using fluoropyrimidines (Santi, 1981; Lewis and Dunlap, 1981; Boeke et aZ., 1984). The toxicity of 5-fluoropyrimidine bases and nucleosides arises through their activation to 5-fluoropyrimidine nucleotides. The nucleoside triphosphate derivatives can be incorporated into newly synthesized RNA and DNA molecules (Heidelberger et al., 1983). In many mammalian cells the cytotoxicity of fluorinated pyrimidines is correlated with their incorporation into RNA molecules (Kufe and Major, 1981). In addition, 5-fluoro 2’-deoxyuridylate is a potent inactivator of thymidylate synthetase (Santi, 1981; Lewis and Dunlap, 1981). While many fluorinated derivatives of pyrimidines have been synthesized and shown to be clinically important, fluorinated analogs of nonaromatic precursors of pyrimidine biosynthesis have not received adequate attention; perhaps with the exception of some aspartate analogs (Hageman et al., 1977; Wanner et al., 1980; Stern et al., 1982). To fill this void, the synthesis and biological activity of fluorinated analogs of dihydroorotate and of N-carbamyl-L-aspartate were considered. (GS)-Dihydroorotate is an intermediate of de nouo pyrimidine biosynthesis (Lieberman and Kornberg, 1954; Lieberman et al., 1955). Introduction of a fluorine on carbon 5 of dihydroorotate would generate a second chiral center on the molecule. The published literature does not describe the synthesis or chemical characterization of any of the four diastereomers of 5-fluorodihydroorotate. The pharmacological properties of the three isomer of 5-fluoro+dihydroorotate were of particular interest due to the unique stereospecificity of cellular enzymes. One of the de nova pyrimidine biosynthesis enzymes, dihydroorotate oxidase (EC 1.3.3.1), was known to catalyze the oxidation of dihydroorotate by removal of the pro-S hydrogen at carbon 5 and the hydrogen at the chiral carbon 6 (Blattman and Retey, 1972; Pascal and Walsh, 1984; Keys and Johnston, 1985). This stereoselectivity suggested that an enzyme catalyzed conversion of threo-5-fluoro-L-dihydroorotate (t-FDHO;‘, 2 structure A in Fig. 1) to 5-fluoroor’ The abbreviations used are: t-FDHO, threo-5-fluoro-L-dihydroorotate; CTF-ASP, N-carbamyl-threo-3-fluoro-L-aspartate; Tris-DTT buffer, 20 mM Tris-HCl, pH 7.4 with 2 mM dithiothreitol; MEM, minimum essential medium; PBS, phosphate-buffered saline; FPLC, fast protein liquid chromatography. 2 three-5-Fluoro-L-dihydroorotate describes (5R,GR)-[5-fluoroldihydroorotate. Note that the priority assignments for determining the stereochemistry at carbon 6 are influenced by the fluorine substitution at carbon 5; the non-fluorinated compound with an identical stereochemistry at carbon 6 is called (6S)-dihydroorotata. N-carba-
14242
threo-5-Fluorodihydroorotate
HN
-00
” +NH3
FIG. 1. Potential metabolic fate of three-5-fluoro-L-dihydroorotate. Structures of the following compounds are shown in this figure. A, three-5-fluorodihydroorotate; B, N-carbamyl-three-3-fluoro-L-aspartate; C, three-3-fluoro-L-aspartate; D, 5-fluoroorotate; E, 5-fluorouridine 5’-monophosphate. The enzymes that would participate in such metabolism are: i, dihydroorotase; ii, aspartate transcarbamvlase: iii. dihydroorotate oxidase; iv, orotate phosphoribosyltransferase; u, ‘orotidyiate decarboxylase. The asterisk (*)-on struckes A, B, D, and E shows the fate of the ‘“C label during metabolism of tFDHO.
otic acid (structure D in Fig. 1) would be possible without impedance from the fluorine substituent. The resulting 5fluoroorotate was then expected to proceed to the synthesis of 5-fluorouridine 5’-monophosphate (structure E in Fig. 1) and other 5-fluoropyrimidine nucleotides (Dahl et al., 1959). In the other direction, hydrolysis of t-FDHO was expected to generate the three analog of &fluoroaspartic acid (structure C in Fig. 1). The conversion of t-FDHO to threo-/3-fluoro-Laspartate was expected to begin with the hydrolysis of the dihydroorotate analog to N-carbamyl-threo-3-fluoro-L-aspartate (CTF-ASP, structure B in Fig. 1)) by the enzyme dihydroorotase (EC 3.5.2.3). In principle, fluorinated carbamylaspartate would generate threo+fluoroaspartate with the aid of aspartate transcarbamylase. Previously, stereospecific fluorinated analogs of aspartic acid had been shown to have cytotoxic activity (Stern et al., 1982,1984; Rathod et al., 1986; Casey et al,, 1986). Threo-@-fluoroaspartate and threo-/3-fluoroasparagine were far more toxic to mammalian cells than the erythro isomers (Stern et al., 1982, 1984). The present investigation was undertaken to synthesize tFDHO and to test the antiproliferative activity of this compound. In the course of this work CTF-ASP was also synthesized and tested for its ability to inhibit the growth of mammalian cells in culture.
14243
Characterization
enzyme would reduce 5-fluoroorotic acid to generate t-FDHO. Commercial dihydroorotate dehydrogenase contained significant amounts of dihydroorotase as a contaminant (Kensler et al., 1981). With dihydroorotase in the reaction mixture, the yield of t-FDHO in intact form would be seriously compromised. To avoid this problem, the two activities, dihydroorotase and dihydroorotate dehydrogenase, were separated by anion-exchange chromatography on a FPLC system (see Miniprint; Fig. MP-1). Partially purified dihydroorotate dehydrogenase was then incubated with the starting reagents 5-fluoroorotate and NADH. A large excess of NADH was used in order to favor the reduction reaction. Unlike 5-fluoroorotate, t-FDHO was not expected to absorb near UV light. To facilitate its purification and characterization, low specific activity radiolabeled 5-fluoroorotate was used as a substrate. Analysis of the reduction reaction by anion-exchange chromatography revealed the presence of a single new radioactive peak that was wellseparated from 5-fluoroorotate (see Fig. MP-2). As expected, the radioactive product lacked UV absorbance, and it was less anionic than 5-fluoroorotate. From the amount of radioactivity under the new peak, it was calculated that 55% of the starting 5-fluoroorotate had been reduced. The new radioactive material was initially separated from the bulk of the reactants by preparative anion-exchange chromatography. Further purification was achieved by anion-exchange and ionpair reversed-phase chromatography on a FPLC system. The final product was pure as judged by reverse-phase chromatography and anion-exchange chromatography (see Fig. 2.4). The preparation showed no adventitious UV absorbing or radioactive materials. Enzymatic modifications of the isolated product suggested that we had successfully synthesized t-FDHO. When t-FDHO was treated with a combination of dihydroorotate dehydrogenase and NAD+, the radioactive material was converted back to 5-fluoroorotate in near quan2ot----
1
EXPERIMENTALPROCEDURES RESULTS Synthesis
of threo-5-Fluoro-L-dihydroorotate-Dihydroor-
otate dehydrogenase from Zymobacterium oroticum was known to catalyze the trans-hydrogenation of erotic acid (Blattman and Retey, 1972). It was anticipated that this myl-three-3-fluoro-L-aspartate describes N-carbamyl-@R,3R)-[3-fluorolaspartate; the biologically active, non-fluorinated metabolite, Ncarbamyl-L-aspartate is called N-carbamyl-(3S)-aspartate. The corresponding fluoroaspartate analog is called threo+fluoroaspartate (Wanner et al., 1980; Stern et al., 1982). 3 Portions of this paper (including part of “Experimental Procedures” and Figs. MP-I-MP-3) are presented in mininrint at the end of this papery Miniprint is easily -read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.
FRACTION NUMBER FIG. 2.
Characterization of threo-5-fluoro-L-dihydroorotate by chromatography. A, elution of pure radioactive t-FDHO off an anion-exchange column. B, elution of radioactivity after tFDHO was treated with dihydroorotate dehydrogenase and NAD’. Standard 5-fluoroorotate elutes in fractions 19-21. C, elution of radioactivity after t-FDHO was treated with dihydroorotase. Details of chromatography conditions and incubation conditions are presented in the Miniprint.
14244
threo-5-Fluorodihydroorotate
titative yield (Fig. ZB). In addition, t-FDHO was a competent substrate for dihydroorotase (Fig. 2C and below). In addition to the chromatographic evidence, high field nuclear magnetic resonance spectra of the preparation were consistent with a pure product having the desired structure (Fig. 3). A 500 MHz proton FT-NMR spectrum in D,O revealed a doublet-doublet proton signal at 5.33 ppm that was assigned to the 5-H proton. A second doublet-doublet proton signal at 4.46 ppm was assigned to the 6-H proton. The presence of an aliphatic fluorine on carbon 5 was evident from the 47 Hz fluorine coupling with the C-5 geminal proton and a 13 Hz fluorine coupling with the C-6 vicinal proton. Coupling of 7 Hz was also observed between the C-5 proton and the C-6 proton. Fig. 3 also shows the proton-decoupled carbon 13 FT-NMR spectrum of the newly synthesized compound. This spectrum revealed the five signals at 58.5 ppm (C-6), 86.5 ppm (C-5), 156.5 ppm (C-2), 164.1 ppm (C-4) and 174.5 ppm (carboxyl group). The signal at 58.5 ppm showed 23 Hz coupling to the fluorine and the 86.5 ppm signal showed a 179 Hz coupling to the fluorine. These spin-spin coupling constants are characteristic of two-bond and one-bond 13ClgF interactions, respectively (Breitmaier and Voelter, 1987; Jameson, 1987). The spectrum was not strong enough to resolve any coupling between any carbonyl carbons and the fluorine. Stability of three-SFluoro-L-dihydroorotate-The purification of t-FDHO was performed under cold conditions, and the purified compound was always stored at -20 “C. Since the biological characterization of t-FDHO required prolonged exposures at 37 “C, it was necessary to determine the stability of t-FDHO at this temperature. t-FDHO was incubated at
:: 5: -:::q
A
J
F.-+-J
4.6
4.4
PPM
Characterization
37 “C either in phosphate-buffered saline, serum-supplemented MEM (complete medium), or in complete medium with HT-1080 cells attached to the substratum. Using anionexchange chromatography the media were analyzed for the breakdown of t-FDHO. In all cases, anion-exchange chromatography of the media showed a time-dependent decrease in the amount of radioactivity associated with the t-FDHO peak. This decrease in tFDHO was accompanied by the time-dependent appearance of a single new radioactive peak in fractions 17 and 18 of the anion-exchange analytical system (see Miniprint). In order to determine if there were other products of t-FDHO breakdown, the media were also analyzed for the presence of radiolabeled CO,. When the media were acidified, no radiolabeled CO, release was detectable. Table I shows the factors that influenced the degradation of t-FDHO. After 24 h, only 4.5% of tFDHO broke down in PBS. The degradation in complete MEM was 19%, whereas the breakdown of t-FDHO in the well containing HT-1080 cells was 27.5%. After 48 h, the breakdown of t-FDHO was 13% in PBS, 34% in complete MEM, and 45.6% with HT-1080 cells in complete MEM. It was suspected that the radioactive breakdown product observed by anion-exchange chromatography was CTF-ASP arising from the hydrolysis of t-FDHO. A systematic synthesis of CTF-ASP was undertaken so that its properties could be compared with those of the breakdown product of t-FDHO. Synthesis of N-Carbamyl-three-3-fluoro-L-pa&ate-In order to synthesize CTF-ASP, radioactive t-FDHO was incubated with partially purified 2. oroticum dihydroorotase, an enzyme which normally hydrolyzes L-5,6dihydroorotate to N-carbamyl-L-aspartate (Fig. MP-1). Under these conditions, t-FDHO was readily converted to a more acidic radioactive compound that eluted differently from t-FDHO and 5-fluoroorotate (see Fig. 2C). In addition to the manner in which it was prepared, two lines of evidence suggested that the newly synthesized compound was CTF-ASP. First, since the compound was synthesized from 5-fluoro [2-‘4C]dihydroorotate, the retention of radioactivity indicated that the carbamyl group was intact (see Fig. 1). Indeed, brief treatment of the compound with acid did not release radioactive carbon dioxide. Second, a proton nuclear magnetic resonance spectrum of the isolated material in deuterated water clearly showed the expected two proton signals. The chemical shifts and the spin-spin coupling constants were consistent with the expected structure for CTF-ASP (see Miniprint and Table II). The chromatographic properties of the newly synthesized TABLE
Conditions that effect PBS, phosphate-buffered
’
’
65
I” 70
75
60
8 8 II 65
8
‘II 60
I
the degradation of threo-5-fluarodihydroorotate saline. Fraction of material eluting as t-FDHO
Incubation conditions
g PPM
%
Fraction of material eluting as CTF-ASP %
Oh Starting
I’ 175
r
I n 170
’
e
’
I”’ 165
8
I n 160
r
n
‘I 155
PPM
FIG. 3. Characterization of three-5-fluoro-L-dihydroorotate by nuclear magnetic resonance. A, 500 MHz proton NMR of 0.7 mg of t-FDHO in 0.5 ml of DzO. The abnormal shape at 4.8 ppm is due to incomplete suppression of the HOD signal. B, 125 MHz proton-decoupled carbon 13 NMR spectrum in D20. Details of the data acquisition parameters are presented in the Miniprint.
sample
100
24 h PBS Complete culture Culture medium
medium on HT-1080
48 h PBS Complete culture Culture medium
medium on HT-1080
0
cells
95.5 80.6 72.5
4.5 19.4 27.5
cells
87.0 66.0 54.3
13.0 34.0 45.7
threo-5Fluorodihydroorotute
14245
Characterization
TABLE II Vicinal and geminal coupling constants from proton NMR of aspartate and dihydroorotate analogs Compared with the noncyclic compounds, stearic restrictions in the cyclic compound, t-FDHO, influence the vicinal H-H and vicinal H-F coupling constants as described under “Discussion.” Note, H, is used to refer to the hydrogens on carbon 2 of aspartic acid analogs and on carbon 6 of the dihydroorotate derivatives. Ho is used to refer to the hvdrogens on carbon 3 of aspartic acid derivatives and on carbon 5 of dihydroorotate and t-FDHO. Compound
H--H, ~I
I
H-.-F
Refs.
Ha-F
HZ Aspartate Erythro-@-fluoroaspartate Threo-P-fluoroaspartate
5-6 2.5 2.4
27 29
50 45
Roberts and Jardetzky (1970) Wanner et al. (1980) Stern et al. (1982)
Asparagine Erythro-P-fluoroasparagine Threo-fl-fluoroasparagine
5-7 2.5 2
26 33
50 45
Roberts and Jardetzky (1970) Wanner al. (1980) Stern et al. (1982)
N-Carbamylaspartate N-carbamyl-threo-3-fluoroaspartate
6-8 2
26
48
This study This study
L-Dihydroorotate Threo-5-fluoro-L-dihvdroorotate
5-7 7
13
47
Keys and Johnson (1985) This study
et
0
Em
t-FDHO
CTF-ASP
1 0 \
0 ‘0.
i
Iii’
12
16”
t '\ 0 IO’
CONC. OF FLUORINATED
FIG. 5. Comparison t-FDHO
FIG. 4. Inhibition coma
(HT-1080)
of the cells by
Cont.,
(M)
proliferation of human three-5-fluoro+dihydroorotate.
fibrosar-
Cells were seeded at a density of 50,000/9.6 square cm well. Each well contained 2.5 ml of growth medium (see Miniprint). After 24 h of incubation at 37 “C in a CO2 incubator, the used medium was replaced with fresh medium containing varying concentrations of t-FDHO (note, the concentration values shown in the figure represent starting t-FDHO concentrations; the values were not adjusted for uptake or breakdown of the antimetabolite). Two control wells had fresh MEM without an inhibitor. The cells were incubated for an additional 48 h. Proliferation of cells was determined directly by measuring the number of viable cells/well (see Miniprint). Each data point represents an average of two determinations. Control wells showed a net increase of 350,000 cells during the 48-h incubation. Cells in control wells were in logarithmic growth phase throughout the course of the toxicity study. CTF-ASP were identical to those of the breakdown product of t-FDHO during the incubation experiment described above (data not shown). Antiproliferative Activity of threo-5-Fluoro-L-dihydroorotate and IV-Carbamyl-threo-3-fluoro-L-aspartate-When t-FDHO was incubated with HT-1080 cells, it interfered with the proliferation of these cells. From the growth inhibition curve of t-FDHO (Fig. 4), the I& for HT-1080 cells was estimated to be 20 PM. Since it was known that as much as 45% of tFDHO hydrolyzed to CTF-ASP during the cell culture experiment, it was important to determine if CTF-ASP had significant cytotoxic effects of its own. The toxicity of CTF-ASP was found to be negligible (Fig. 5). No toxicity was observed at 1 or 10 PM concentration of the N-carbamylaspartate
between
three-5-fluoro-L-dihydroorotate fluoroaspartate. Toxicity
the
COMPD.,
(M)
antiproliferative effects and N-carbamyl-three-3-
studies were performed
of
as described in
the legend to Fig. 4. derivative; only 7.5% of the cells died at 100 pM concentration. It was concluded that most of the toxicity observed from tFDHO was due to the inherent antiproliferative activity of this compound and not due to the breakdown of this compound to CTF-ASP. Often it is possible to determine the mechanism of an antimetabolite by observing whether an intermediate or an end product of a metabolic pathway can dilute the effects of the toxic agent. Incubation of HT-1080 cells with t-FDHO in the presence of aspartate and/or uridine decreased the cytotoxic effects of t-FDHO. However, the experiments failed to provide clear insights on the dominant mechanism behind the antimetabolic activity of this compound. Incubation of HT-1080 cells with 1 pM t-FDHO alone caused 30% inhibition of growth. When cells were treated with 1 PM t-FDHO along with 1 mM aspartate or 1 mM uridine, the cells were inhibited by only 14 and 15%, respectively. The rescuing effects of uridine and aspartate were not additive. When the cells were incubated with 1 pM t-FDHO in the presence of a combination of 1 mM aspartate and 1 mM uridine, they were inhibited by 13%. When HT-1080 cells were treated with 100 pM t-FDHO alone, the cells were inhibited by 68%. At this higher concentration of t-FDHO, 1 mM uridine was ineffective as a rescuing agent (67% inhibition of cell growth). In contrast, in the presence of 100 I.IM t-FDHO, 1 mM aspartate reduced inhibition of cell growth from 68 to 57%. When the cells were
14246
threo-5-Fluorodihydroorotate
Characterization
incubated with 100 PM t-FDHO in the presence of a combination of 1 mM aspartate and 1 mM uridine, they were inhibited by 56%. Thus, at more toxic levels of t-FDHO, uridine was virtually ineffective at decreasing toxicity whether used alone or in combination with aspartate. DISCUSSION
A stereospecifically fluorinated analog of dihydroorotate, tFDHO, was synthesized, chemically characterized, and shown to inhibit the proliferation of human fibrosarcoma cells. The hydrolyzed derivative, CTF-ASP, had insignificant antiproliferative activity of its own. The synthesis of t-FDHO was achieved by reducing 5fluoroorotate with NADH in the presence of dihydroorotate dehydrogenase. A proton NMR and a proton-decoupled carbon NMR clearly showed the deshielding effects and the spinspin coupling effects of an aliphatic fluorine on carbon 5 of the dihydroorotate analog. Additionally the NMR data could account for all the hydrogens and the carbons of t-FDHO. Earlier studies had shown that 6-substituted dihydrouracils in deuterated buffer underwent nonenzymatic exchange of enolic C-5 hydrogens (Keys and Johnson, 1985). This acidity of C-5 hydrogen in dihydroorotate was IO-fold higher for the pro-R hydrogen compared with the pro-S hydrogen. The hydrogen on carbon-5 of t-FDHO was expected to be in the same position as the less acidic pro-S hydrogen of dihydroorotate. Nonetheless, it was important to consider whether tFDHO synthesized and purified in the present study was optically pure or whether acidity of carbon 5 had caused racemization of this molecule. The NMR data was consistent with the synthesis and purification of the correct stereoisomer of fluorodihydroorotate. Since the ring structure of t-FDHO was expected to be somewhat rigid, the NMR data could be analyzed in the context of the predicted stereochemistry of t-FDHO (Katritzky et al., 1969; Keys and Johnson, 1985). The spatial relationships amongst atoms of t-FDHO and the alternate possible diastereomer, erythro-5-fluorodihydroorotate, are shown in Fig. 6. For each diastereomer, the two limiting conformers are shown as Newman projections. The NMR information under discussion is the magnitude of the H-5 and H-6 coupling constant (7 Hz) and also the magnitude of the F-5 and H-6 coupIing constant (13 Hz). It is known that on an absolute scale, vicinal H-H coupling constants can be as high as 15 Hz, and vicinal H-F coupling constants can be as high as 58 Hz. This information is of little use in the analysis of t-FDHO structure because, in addition to the configuration of the substituents, a number of factors contribute to the magnitude of vicinal coupling constants. These factors include the electronegativity of substituents on the carbon-carbon bond, the nature of the equilibrium between limiting conformations, temperature, pH, and solvent. Vicinal proton-proton coupling constants in quasiplaner fluoropyrimidine rings are known to vary from 2 to 20 Hz (Haas and Kortmann, 1981; Palling et al., 1980). The intermediate values of 7 and 13 Hz seen with putative tFDHO could be assigned to certain conformations of threo as well as erythro isomers of fluorinated dihydroorotate. A more fruitful approach to analyzing the effects of conformational restrictions on the vicinal coupling constants of tFDHO was to compare the NMR spectrum of the isolated compound with that of closely related molecules. Table II compares the NMR properties of aspartate, asparagine, Ncarbamylaspartate, and dihydroorotate to the properties of their fluorinated analogs. Information compiled in this way allowed two important observations. First, for noncyclic com-
THREO
ISOMER
H.6
coo-
’ ‘6
H,
ERYTHRO
>
< F
ISOMER
coo-
FIG. 6. Limiting conformations assumed by two diastereomers, three-5-flioro-L-dihydroorotate, and erythro-Ei-fluoro-L-dihydroorotate. The Newman projections in the figure illustrate the possible dihedral angles between Hh-H, and Fe-He. The consequences of these configurational restrictions on the proton NMR of r-FDHO are addressed under “Discussion.” pounds which have unrestricted rotation between the Ly-carbon and the p-carbon (i.e. aspartate, asparagine, and iVcarbamylaspartate), addition of a @-fluorine causes a decrease in the proton-proton coupling constant from about 7 Hz to about 2 Hz. This decrease in vicinal proton-proton coupling constant is independent of the stereochemistry of the fluorine substitution. In contrast, conversion of conformationally restricted dihydroorotate to t-FDHO was not associated with a decrease in the H,-HB coupling constant. The 7 Hz protonproton coupling on this fluorinated compound could be explained on the basis of stereochemistry. In a cyclic system such as t-FDHO, the effect of an electronegative fluorine on the H-5 to H-6 vicinal constant was expected to depend on the stearic disposition of the fluorine substituent with respect to the H-6 proton. Maximum effect of the fluorine in reducing the H-5 and H-6 vicinal coupling constant would be observed when the fluorine and the H-6 proton were trans-coplanar (Hall and Jones, 1973; Atta-ur-Rahman, 1986). Examination of Fig. 6 shows that restricted bond angles in the threo isomer of fluorodihydroorotate preclude a conformation where F-5 and H-6 are coplanar. In contrast, the fluorine on the erythro isomer of fluorodihydroorotate should be able to assume a tram-coplanar conformation with respect to H-6, and thereby an erythro isomer would have shown a substantial decrease
threo-5-Fluorodihydroorotate in the H-5 to H-6 vicinal coupling constant.4 The vicinal H-F coupling constants in Table II also indicates that the isolated compound is the threo isomer of fluorodihydroorotate. For all fluorinated analogs of aspartate, asparagine, and N-carbamylaspartate, the Ha-F0 coupling constant is between 26-33 Hz. In contrast, the fluorinated analog of dihydroorotate isolated in this study shows a much lower coupling constant between F-5 and H-6 (13 Hz). Again, while dihedral angles between F-5 and H-6 of the erythro isomer would allow for maximum H-F coupling constants, such possibilities are not available to the threo isomer. The NMR, then, is consistent with the synthesis of the correct stereoisomer of fluorinated dihydroorotate. Treatment of the newly synthesized t-FDHO with dihydroorotate dehydrogenase and NAD’ resulted in quantitative conversion back to 5-fluoroorotate. Again, knowing the stereoselectivity of dihydroorotate dehydrogenase (Blattmann and Retey, 19’72), this suggested that the compound was indeed in the form of the threo isomer and that, under the careful purification conditions, the compound had not racemized to an appreciable level. Overall, it was concluded that the newly synthesized and purified compound was threo-5fluoro-L-dihydroorotate. To test its antiproliferative activity, t-FDHO was incubated with human fibrosarcoma (HT-1080) cells in culture. Previously, this cell line had served as a good model for determining the toxicity of fluorinated pyrimidines toward mammalian cells (Rathod et al., 1989). t-FDHO showed modest toxicity against HT-1080 cells (apparent I& = 20 PM). This value resembled the toxicity of 5-fluoroorotate against HT-1080 cells (ICsO = 10 wM, Rathod et al., 1989). HT-1080 cells are much more vulnerable to 5-fluorouracil (I& = 2 PM), 5fluorouridine (I& = 30 nM), and 5-fluoro 2’-deoxyuridine (I& = 10 nM). Threo+fluoroaspartate has not been tested against HT-1080 cells, but with other mammalian cells it showed I&o values between lo-150 pM (Stern et al., 1982). The lower toxicity of 5-fluoroorotate and t-FDHO could be due to poor transport of these anionic compounds into mammalian cells or due to inefficient activation of the intracellular compounds to fluoropyrimidine nucleotides. During the determination of the cytotoxic effects of tFDHO, a significant amount of extracellular t-FDHO was found hydrolyzed to CTF-ASP. Direct incubation of HT-1080 cells with CTF-ASP resulted in little or no toxicity, suggesting that this compound was not contributing significantly to the cytotoxicity of t-FDHO. The biochemical basis for the low toxicity of extracellular CTF-ASP is not known. It could be due to poor uptake of this compound by HT-1080 cells or due to severely limited ability of cells to convert CTF-ASP to threo-@-fluoroaspartate. It was conceivable that t-FDHO caused the observed toxicity by first undergoing hydrolysis to threo-fi-fluoroaspartate in the extracellular medium. The possibility, however, could not be supported. Such degradation would release radioactive carbamylphosphate, carbamate, and, eventually, bicarbonate in the medium (Allen and Jones, 1964). Since acidification of media that had been exposed to t-FDHO did not result in the release of radioactive carbon dioxide, it was concluded that tFDHO did not degrade beyond CTF-ASP. As noted above, incubation of the cells with CTF-ASP resulted in no cytotoxicity. This also argues against extracellular degradation of 4 This prediction will be tested directly upon the successful synthesis of the evthro isomer of 5-fluorodihydroorotate. Several attempts at synthesizing this compound via the chemical reduction of 5fluoroorotate resulted in elimination of the fluorine. Other synthetic approaches are being tested in our laboratory.
14247
Characterization
this intermediate to generate cytotoxic threo+fluoroaspartate. Finally, aspartate caused minimal rescue of HT-1080 cells from t-FDHO cytotoxicity. If cytotoxicity from t-FDHO was primarily due to extracellular threo-@-fluoroaspartate, one would expect near complete rescue with 1 mM aspartate (Stern et al., 1982). While t-FDHO was shown to have antiproliferative activity, it was not very clear how this compound exerted its cytotoxic actions. Addition of uridine or aspartate allowed HT-1080 cells to recover slightly suggesting that at least some of the toxicity resulted from interference with the pyrimidine pathway and aspartate metabolism. The fact that uridine did not rescue the cells completely was not surprising. Uridine offers incomplete rescue to HT-1080 cells treated with 5-fluoroorotate or 5-fluorouracil (Rathod et al., 1989). Present efforts are directed at producing larger amounts of t-FDHO and also at producing t-FDHO with a radiolabel on the aspartate backbone. The availability of such compounds will allow us to determine the complete metabolic fate of radioactive t-FDHO. Such a study would also reveal the metabolic target(s) of tFDHO. Finally, the cytotoxic effects of this compound against various mammalian cell lines and against pathogenic microorganisms needs to be examined in detail, particularly since 5-fluoroorotate has recently been shown to have selective antimalarial activity (Rathod et al., 1989; Gomez and Rathod, 1990). Acknowledgment-We wish to thank Colonel Milhous (Walter Reed Armv Institute of Research) for providing us with 5-fluoroorotate, Dr. Michael Summers (U&e&ity of &Iaryland, Baltimore County) for performing high resolution nuclear magnetic resonance analysis on our samples, Dr. James Greene (The Catholic University of America) for providing us with HT-1080, and Jon Mills for his assistance in the purification of dihydroorotate dehydrogenase. REFERENCES Allen, C. M., Jr., and Jones, M. E. (1964) Biochemistry 3,1238-1247 Atta-ur-Rahman (1986) Nuclear Magnetic Resonance: Basic Principies, pp. 72-74, Springer-Verlag, New York Bennett. J. E. (1977) Ann. Intern. Med. 86.319-322 Blattmann, P.,‘and ketey, J. (1972) Eur. J.‘Biochem. 30, 130-137 Boeke, J. D., LaCroute, F., and Fink, G. R. (1984) Mol. Gen. Genet.
197,345-346 Breitmaier, E., and Voelter, W. (1987) Carbon-13 NMR Spectroscopy: High Resolution Methods and Applications in Organic Chemistry and Biochemistry, 3rd Ed., pp. 205-206, VCH Publishers, New York Casey, P. J., Abeles, R. H., and Lowenstein, J. M. (1986) J. Biol. Chem. 261,13634-13642 Dahl, J. L., Way, J. L., and Parks, R. E., Jr. (1959) J. Biol. Chem. 234,2998-3002 Friedmann, H. C., and Vennesland, B. (1958) J. Biol. C&m. 233, 1398-1406 G6mez, Z., and Rathod, P. K. (1990) Antimicrob. Agents Chemother. 34,1371-1375 Haas, A., and Kortmann, D. (1981) Chem. Ber. 114,1176-1179 Hall, L. D., and Jones, D. L. (1973) Can. J. Chem. 51,2925-2929 Hageman, J. J. M., Wanner, M. J., Koomen, G. J., and Pandit, U. K. (1977) J. Med. Chem. 20, 1677-1679 Heidelberger, C., Chaudhuri, N. K., Danneberg, P., Mooren, D., Griesbach, L., Duschinsky, R., Schnitzer, R. J., Pleven, E., and Scheiner, J. (1957) Nature 179, 663-666 Heidelberger, C., Danenberg, P. V., and Moran, R. G. (1983) Adu. Enzymol. 54,57-119 Jameson, C. J. (1987) in Multinuclear NMR (Mason, J., ed) pp. 437446, Plenum Publishing Co., New York Katritzky, A. R., Nesbit, M. R., Kurtev, B. J., Lyapova, M., and Pojarlieff, I. G. (1969) Tetrahedron 25, 3807-3824 Keys, L. D., IV, and Johnston, M. (1985) J. Am. Chem. Sot. 107,
486-492 Kensler, T. W., Han, N., and Cooney, D. A. (1981) 111,49-53 Kufe,
D., and
Major,
P. (1981)
J. Biol.
Chem.
266,
Anal. 9802-9805
Biochem.
14248 Lieberman, Lieberman,
threo-5-Fluorodihydroorotate I., and Kornberg, A. (1954) J. Biol. Chem. 207,911-924 I., Kornberg, A., and Simms, E. (1955) J. Biol. Chem.
215,403-415 Lewis, C. A., and Dunlap, R. B. (1981) in Topics in Molecular Biology (Burger, A. S. V., and Roberts, G. C. K., eds) pp. 169-219, Elsevier/ North-Holland Biomedical Press, New York Palling, D. J., Atkins, P. J., and Hall, C. D. (1980) J. Chem. Sot. Perkin II 1460-1466 Pascal, R. A., Jr., and Walsh, C. T. (1984) Biochemistry 23, 27452752 Rathod, P. K., Khatri, A., Hubbert, T., and Milhous, W. K. (1989) Antimicrob. Agents Chemother. 33, 1090-1094 Rathod, P. K., Tashjian, A. H., Jr., and Abeles, R. H. (1986) J. Biol. Chem. 261,6461-6469
Characterization Roberts, G. V. K., and Jardetzky, 0. (1970) Adu. Protein Chem. 24, 447-545 Santi, D. V. (1981) in Molecular Actions and Targets for Cancer Chemotherapeutic Agents (Sartorelli, A. C., Lazo, J. S., and Bertino, J. R., eds) pp. 285-300, Academic Press, New York Stern, A. M., Foxman, B. M., Tashjian, A. H., Jr., and Abeles, R. H. (1982) J. Med. Chem. 25,544-550 Stern, A. M., Abeles, R. H., and Tashjian, A. H., Jr. (1984) Cancer Res. 44,5614-5618 Walsh, C. (1983) Ado. Enzymol. Rel. Areas Mol. Biol. 65, 197-289 Wanner, M. J., Hageman, J. J. M., Koomen, G.-J., and Pandit, U. K. (1980) J. Med. Chem. 23,85-87 Washabaugh, M. W., and Collins, K. D. (1984) J. Biol. Chem. 259, 3293-3298
thread-Fluorodihydroorotate
Characterization
