Nucleosides. XI. Synthesis of 1-β-D-Arabinofuranosul-5-Fluorouracil

fluorouracil and subsequently to products derived therefrom).6 The discovery7 that l-/S-D-arabino- furanosyluracil is phosphorylated enzymically to...
0 downloads 0 Views 842KB Size
4060

N. C. YUSG.J. E-I. BURCHENAL, RONALD FECHER, ROBERT DUSCHINSKY AND JACK J. Fox

Vol. 83

[CONTRIBUTION FROM THE DIVISIONS OF NUCLEOPROTEIN CHEMISTRY AND EXPERIMENTAL CHEMOTHERAPY, SLOAN-KETTERINSTITUTE FOR CANCER RESEARCH;SLOAN-KETTERING DIVISIONOF CORNELL UNIVERSITY MEDICAL COLLEGE, NEW YORK 21, NEW Y O R K AND HOFFMANN LA ROCHE, INC.,NUTLEY, N E W JERSEY]

ING

Nucleosides.

XI.

Synthesis of I-~-D-Arabinofur~osyl-5-~Fluorourac~l and Related Nucleosides1

BY NAISHUNC. Y m c , JOSEPH H. BURCHENAL, RONALD FECHER,ROBERTDUSCHINSKY AND JACK J. F o x RECEIVED APRIL 24, 1961 Methodsaredescribed for the synthesis in good yields of 1-8-D-arabinofuranosyl-andl-~-~-lyxofuranosyl-5-fluorouracils from Bfluorouridine Via anhydro nucleoside intermediates. Synthesis of Biluorouridine in improved yield from Siluorouracil Via the mercuri procedure is described. Preliminary studies on the effect of these nucleosides against transplanted mouse leukemia B82 are reported.

The usefulness of 5-fluoro-2‘-deoxyuridine2as uridine and l-fLD-ribofuranosylthymine12 via 2,2‘an anti-tumor agent in experimental tumors3 and anhydronucleoside intermediates to their corin clinical trials4is believed to be due to the inhibi- responding arabino nucleosides. The preparation tion 0 the metabolic conversion of 2’-deoxyuridylic of 5-fluorouridineh by the mercuri processI2 was in the acid t o thymidylic acid by the nucleotide of 5- improved by use of 5-fluorouracilmer~ury~~ fluoro-2’-deoxyuridine.5 The efficacy of 5-flUOrO- condensation reaction with tri-O-benzoyl-D-ribo2’-deoxyuridirie is weakened, however, by catabolic furanosyl chloride. l a Removal of the benzoyl processes (e.g., cleavage by nucleosidases t o 5- blocking groups with dilute alkali afforded 5fluorouracil and subsequently to products derived fluorouridine (I) in 65% yield (based upon 5therefrom) .6 The discovery? that l-@-D-arabino- fluorouracil). Tritylation of I afforded an almost furanosyluracil is phosphorylated enzymically to quantitative yield of the 5’-O-trityl derivative the 5’-nucleotide which, further, can partake in (11). Tosylation of I1 with p-toluenesulfonyl the enzymic methylation step (albeit poorly) is chloride in pyridine afforded a mixture from which of significance. Since l-P-D-aIdopentofuranosyl- the 2’-O-tosyl derivative (111) crystallized in 52% pyrimidines other than ribosyl are generally poor yield.I4 Detritylation of I11 with ethanolic hydrosubstrates for nucleosidase activity (glycosyl cleav- gen chloride a t 60’ for ten minutes yielded the 2’age) the 1-8-D-arabinofuranosyl-5-fluorouracil 0-tosyl derivative (IV) which was converted to the and related nucleosides were synthesized to de- 2,2’-anhydronucleoside (V, R = F) by treatment termine whether they could exert anti-tumor with one equivalent of alkali. Treatment of V activity but with decreased toxicity relative to with dilute alkali a t room temperature afforded 5-fluoro-2’-deoxyuridine. A preliminary report VI. A more practical approach to VI was effected by dealing with the syntheses of these compounds has treatment of I11 with two equivalents of alkali in appeared.9 The synthesis of l-B-D-arabinofuranosy1-5-ffuoro- 50% ethanol. By this procedure, the anhydrouracil (VI) was accomplished by modifications of nucleoside (VII) is formed and cleaved in situ the procedureslot” used for the epimerization of by the excess alkali to VI11 which precipitates. VI11 was detritylated with dilute acid in 50% (1) This investigation was supported in part (to the Sloan-Kettering Institute) b y funds from the National Cancer Institute of the Naethanol to form VI. Though intermediates were tional Institutes of Health, Public Health Service (Grant No. C Y not purified, the over-all yield of VI (based upon 3190), the Cancer Chemotherapy National Service Center, Research 111) was SS%. Contract SA-43-ph-2445 and the American Cancer Society. Proof that VI is l-~-~-arabinofuranosyl-5-fluoro(2) (a) R. Duschinsky, E.Pleven, E. Malbica and C . Heidelherger, Abstr. 132nd Meeting, Am. Chem. Sac., 1957, p. 19-C; (b) M. Hoffer, uracil rests on the following data: The ultraviolet R . Duschinsky, J. 1. Fox and N. Yung, J. A m . Chem. SOL, 81, 4112 absorption spectrum of VI was similar t o that given (1959). by I. When treated with metaperiodate, VI con(3) (a) C. Heideiberger, L. Griesbach, 0. Cruz, R. J. Schnitzer and sumed one mole of reagent per mole of nucleoside E. Grunberg, PYOC.SOC.Exp. Bioi. Mcd., 97, 470 (1958): (b) J. H. Burchenal, E. A. D. Holmberg, J. J. Fox, S. C. Hemphill and J. A. slowly12 without the liberation of formic acid in Reppert, Cancer Research, 19, 494 (1959). accord with a furanosyl structure containing a (4) A. R. Curreri and F. Ansfield, Cancer Ckemofherapy Reports trans a-glycol system. Finally, reduction of VI (Cancer Chemotherapy National Service Center), 2, 8 (1959); F. Ansfield and A. R. Curreri, ibid., 6, 21 (19GO). with palladium-charcoal afforded 1-b-D-arabino(5) S. S. Cohen, J. C. Flaks, H. D. Barner, M. 11. Loeb and J. furanosyluracil. Lichensteiu, Proc. Null. Acad. Sci., 44, 1004 (1958); L. Bosch, E. The lyxo nucleoside (XII) also was prepared Harbers and C. Heidelberger, Cancer Research. 18, 33.5 (1958). from 5’-O-trityl-5-fluorouridine (11). Mesylation (6) P;. K. Chaudhuri, R. L. Miikherjee and C. Heidelherger. Riochemica; Pharmacology, 1, 328 (1959). of I1 afforded a di-0-mesyl derivative (IX) as an ,‘e8

(7) L. I. Pizer and S. S. Cohen, J. Bid. Chem., 296, 2387 (1960); Abstr. 136th Meeting, Am. Chem. SOC., 1959, p. 9C. (8) J. J. Fox. J. F. Codington, N. C. Yung, L. Eaplan and J. 0. Lampen, J . A m . Chem. Soc., 80, 5155 (1958). (9) (a! J. J. F o x , N. Yung, I. Wempen, R. Duschinsky and L. Rnplan, Abstr. Intl. Union Pure and Applied Chemistry (Symposium on Natural Products) .4nstralin, 1960, p. 6 6 ; (b! the synthesis of I-8-Darabinofuranosyl-5-fiuoroura~lby au alternate route has been reported as a communication recently by E. J. Reist, J. E.Osiecki. L. Goodman and B. R. Baker, J . .4m. ChPm. Soc., 83, 2208 (1961). (10) D. M. Brown, A. R. Todd and S. Varadarajan, J . Chem. Soc.. 2388 (1956).

(11) J. J. Fox, N. C. Yung and A. Bendich. J. A m . Chem. SOC..79. 2775 (1957). (12) J. J. F o x , N. Yung, J. Davoll and G. B. Brown, i b i d . , 78, 2117 (1.956).

(13) l-Q-Acety1-2,3,5-tri-O-benzoyI-~-ribose (R. K. Ness, H. W . Diehl and €1. G. Fletcher, Jr., ibid., 76, 783 (1954); H. M. Kissman. C . Pidacks and B. R. Baker, ibid., ‘77, 18 (1955), waB used for the preparation of this halogenose. (14) The mother liquor from this reaction is rich in nucleoside derivatives. The composition of this mother liquor and its use in further syntheses are described later in the text.

Oct. 5, 1961

SYNTHESIS OF 1-@-D-ARABJNOFURANOSYL-5-FLUOROURAClL

amorphous solid. (Attempts to prepare a mono-0-mesyl derivative of 11 using one molecular equivalent of methanesulfonyl chloride were unsuccessful. The product obtained was invariably the di-0-mesylated derivative (fX) along with starting material.) Detritylatjon of I X afforded the 2’,3 ’-di-0-mesyl derivative (X) of 5-fluorouridine which, upon boiling in water for 4 hr., was converted to ~-P-D-

406 1

lyxofuranosyl -5-fluorouracil (XII). Treatment of X with one equivalent of alkali formed the 2,2’-anhydronucleoside (XI, R = F, R’ = OH) which, upon refluxing with dilute acid, also gave XII. When treated with metaperiodate, 1-P-D-lyxofuranosyl-5-fluorouracil consumed one mole of oxidant per mole without the liberation of formic acid in accordance with a pentofuranosyl structure containing an a-cis glycol system. A comparison of the molecTr = triphenylmethyl Ms methylsulfonyl Ts = F - toiuenssulfonyl ular rotations of these fluorinated nucleosides verFlow Chart sus their non-fluorinated analogs (see Table I) is in accord with configura- hydrolytic cleavage of the anhydro linkage is tional assignments made for I, VI and XII. shown by the fact that V (R = F) is also inert in The conversion of X I (R = I?, R’ = OH) t o refluxing water for 4 hr., and, again, acid “priming” the lyxo nucleoside (XII) was not unexpected in of the refluxing solution is essential for its converlight of previous studies in this Laboratory’6 sion t o VI. On the other hand, the non-fluorinated which demonstrated conclusively that 5’-substi- analog of V (R = H) undergoes appreciable contuted derivatives of X I (R = H, R’ = 0 mesyl, version to l-@-D-arabinofuranosyhracd without 0-benzoyl or hydrogen) are converted to 1-/3-~-acid “priming” under identical reflux conditions. lpxofuranosyluracils by merely boiling in water. I n the conversion of X t o XII, however, acid The mechanism elaborated16 involves first the hy- priming is not necessary since methylsulfonic acid drolytic cleavage of the 2,2’-anhydro linkage t o is liberated by the formation of the 2,2‘-anhydrothe arabino nucleoside followed by formation of a nucleoside intermediate (XI, R = F, R’ = OH). 2,3’-anhydro-lyxosyl nucleoside with the liberation The liberated acid catalyzes the cleavage of the of one mole of methylsulfonic acid. Under the anhydro linkage. acidic conditions engendered, the 2,3’-anhydro The acid requirement needed for the cleavage bond is hydrolyzed with the formation of lyxo of the anhydro linkages of the 5-fluorinated derivanucleosides. tives (V and XI) as versus their un-fluorinated Whereas the non-fluorinated derivatives (XI, analogs is probably a reflection of the electron withR = H, R’ = 0-benzoyl, 0-mesyl or hydrogen) drawing properties of the 5-fluor0 atom exerted are converted by refluxing in water for 4-5 hr. to through the conjugated system in the aglycon lyxosyluracils, l6 the fluorinated derivative (XI, It is probable that the electronegative 5-fluor0 R = F, R’ = OH) remained unaltered under these atom renders the formation of the conjugate acid reaction conditions. However, if the refluxing of the anhydro nucleoside more difficult by water. aqueous solution of X I (R = F) is “primed” with This obstacle would be overcome as the hydrogen dilute acid (one molecular equivalent of methyl- ion concentration in the refluxing medium is sulfonic acid), the formation of the 5-fluorolyxosyl increased. nucleoside is complete in 4 hr. That the 3‘As mentioned above, a 52% yield of I11 was obmesyloxy function plays no significant role in the tained upon tosylation of 11. The mother liquor from this reaction (designated as XIII-mother ( 1 5 ) R. Fecher, J. F. Codington and J. J. Fox, J . Am. Chem. Soc., liquor on the flow chart) was also utilized for the 88, 1889 (1961).

4062 N.

c. YUNG, J. H. B U R C H E N A L , RONALD FECIIER, ROBERTDUSCHINSKV A N D JACK J. F O X

VOl. 83

synthesis of l-fl-~-lyxofuranosyl-5-fluorouracil ment of XIII-mother liquor signified that this (XII). XIII-Mother liquor could contain four (or mother liquor contains the di-0-tosylated derivative less) possible nucleosides : starting material (II), (XIII, R,R’ = p-tosyl). Since only one equivalent additional 2’-O-tosyl nucleoside (111), the isomeric of p-toluenesulfonyl chloride was used in the re3’-0-tosyl-5‘-0-trityl derivative and, finally, a 5’- action of I1 to 111, it may be concluded that an O-trityl-2’,3’-di-O-tosylated 5-fluorouridine (XIII, appreciable amount of unreacted I1 is also conR,R’ = p-tosyl). tained in XIII-mother liquor. Treatment of XIII-mother liquor with an excess After the isolation of XV, the remaining solution of methylsulfonyl chloride in pyridine afforded a (hereinafter designated as “XV-mother liquor”) sirup (XIV, R,R’ = mesyloxy or tosylosy) which was examined by paper electrophoresis (borate did not crystallize. This sirup should contain buffers, $H 6 and 9.2).lS Only two spots of ap(regardless of the relative composition of nucleo- proximately equal intensities were revealed, neither sides in XIII-mother liquor) only 5’-0-trityl-5- of which corresponded to XV (in p H 9.2 borate). fluorouridines bearing sulfonyloxy groups on both The spot migrating the farther (anodically) corthe 2’- and 3’-positions. The sirup was detrity- responded to 5-fluorouridine (I) while the second lated with acid and the detritylated nucleosides spot migrated similarly to VI. (The lyxo isomer refluxed in water for 20 hr. A 73% yield of (XII) is absent and the xylo isomer may be excrystalline l-~-D-lXyOfUranOSyl-5-fluorouraCil (based cluded on the basis of previous studies by Gordon, upon XIII-mother liquor) was obtained. Intrieri and BrownIs who demonstrated that a A good indication of the composition of nucleo- good electrophoretic separation of 1-p-D-aldopentosides contained in XIII-mother liquor was obtained furanosylthymine isomers (borate p H 6) is achieved by treatment of this mother liquor with excess alkali by this method.) The ultraviolet absorption a t room temperature. Under these conditions, spectra of eluates of the two spots resembled those the following nucleosides would be expected to given by I and VI. Neither of these spectra reform : From the remaining 2’-O-tosyl derivative semble that expected for 3‘ - 0 - tosyl - 5 - fluo(111), anhydro formation followed by cleavage of rouridine. l9 Thus, in the over-all tosylation of I1 with one the anhydro bond with the excess alkali would lead ( v b VI1 and VIII) to l-fl-~-arabinofuranosyl-5- mole of p-toluenesulfonyl chloride, the major fluorouracil (VI). If any 3’-O-tosyl isomer of product (111) is formed in approximately 60% I11 were present in XIII-mother liquor, only 3’- yield (of which 54% was isolated in crystalline 0 - tosyl - 5’ - 0 - trityl - 5 - fluorouridine should form). A small amount of 2’,3’-di-O-p-tosylate result since i t had been demonstrated previously16 is formed leaving some unreacted 11. The 3’that under these conditions 3’-O-sulfonyloxyuridines 0-tosyl isomer of I11 is not formed in detectable (such as 3’-O-mesyluridine and its 2’,5’-di-0- amounts. These data differ from the results obtained by trityl derivative) are unaffected by aqueous alkali. Should any 2’3‘-di-O-tosylate of I1 be present in Brown, et a1.20 They tosylated 5’-O-acetyluridine XIII-mother liquor, alkaline treatment should TABLE I form a 2,2‘-anhydro nucleoside which, upon cleavDifferage of the anhydro linkage, would provide a trans ence in system a t 2‘,3‘ amenable to formation of a 2‘,3‘- 1-8-D-Aldopentofuranosyluracils [n]D [MID [MID epoxylyxosyl derivative. Epoxy-nucleosides of this Uridine 10“ 2,440 -2020 type have been synthesized previously in this Ribosyl-5-fluorouracil ( I ) 17 4,460 Laboratory’’ by alkaline treatment of XI (R = Arabinosyluracil +12621 $30,770 -2790 H, R’ = OBz, OMS, or H). Finally, if starting Arabinosyl-5-fluorouracil (VI) 128 +33,560 material I1 were present in the mother liquor, Lyxosyluracil 9515 +23,200 -2230 alkaline treatment followed by detritylation with Lyxosyl-5-fluorouracil ( X I I ) 97 +25,430 acid should yield I. 34h + 7,690 2’,3’-Epoxy-lyxosyluraci117 XIII-Mother liquor was treated with three 2‘,3’-Epoxy-lyxosyl-5-fluoro-2570 equivalents of dilute alkali a t room temperature uracil ( X V ) + 42 +10,260 for one day and the reaction mixture detritylated a D. T. Elmore, J . Chem. Soc., 2084 (1950). * The with dilute hydrochloric acid. A precipitate authors are indebted to Dr. J. F. Codington for this value. (XV) was obtained which exhibited a nucleoside spectrum but which did not consume metaperiodate. under similar conditions and isolated 2’-O-tosylThe melting point, optical rotation and absorption 5‘-O-acetyluridine as a major product as well as spectrum differed from that exhibited by the anhy- an approximately 2070 yield of the 3’-O-tosyl dronucleoside V(R = F) while the elemental isomer. (The presence of a 2’,3’-di-O-tosylate analysis was consonant with that for a 5-fluor0 was not reported by them.) nucleoside minus one molecule of water. A corn(18) M. P. Gordon, 0. M. Intrieri and G. B. Brown, J. A m C h e i i i . parison of the molecular rotation of XV (see Table Soc., 80, 5161 (1958). I) with that for the 2‘,3’-epoxide of l-p-D-lyxo(19) 3’-0-Tosyl-5-fluorouridine is unfortunately not available for f u r a n ~ s y l u r a c i lpermits ~~ the designation of XV comparison. However, i t is reasonable to assume t h a t the spectrum as the 2’,3’-epoxide of l-p-~-lyxofuranosyl-5- of this compound would resemble that for its isomer, 2’-0-tosyl-5fluorouridine (IV). The spectrum of the 2‘-O-tosyl isomer is quite fluorouracil. from that for I or VI (see experimental). The formation of this epoxide by alkaline treat- different (20) D. M. Brown, D. B. Parihar, A. R. Todd and S. Varadarajan,

++

+ +

+ + + +

(16) N. C. Yung and J. J. Fox, J. A m . Chcm. SOC., 88, 3060 (1961). (17) J. F.Codington, R. Fecher and J. J. Fox, Abstr. 139th Meeting, Am. Chem. SOC., St. Louis, April 1961, p. 13D.

J. Chem. SOC.,3028 (1958). (21) W. Bergmann and D. C. Burke, J. Org. Chem., 10, 1501 (1955).

Oct. 5 , 1961

SYNTHESIS OF l - P - D - A R . 4 1 3 1 S o F v R . ~ ~ ~ S ~ ~ L - ~ - F L U O R O U R A C I L

4063

Chemotherapy Studies.-Studies on these comE~perirnental~~ pounds were initiated against transplanted mouse 5-Fluorouracilmercury.2b~2LMercuricacetate (0.1 mole) leukemia B82.Sb This line of leukemia has in the was dissolved in 600 ml. of methanol under stirring and refluxing. A hot solution of 13.0 g. (0.1 mole) of 5-fluorourapast demonstrated sensitivity t o various B-fluori- cil in 250 ml. of water was added causing an immediate prenated pyrimidine derivative^,^^.^^ and the results cipitation of 5-fluorouracilmercury. The mixture was allowed of these studies are shown in Table 11. As can t o cool while stirring. The fine precipitate was filtered and be seen, l-~-~-arabinofuranosy~-5-fluorouraci~ (VI) dried in a desiccator. The yield (33 9.) was quantitative. had approximately as good a chemotherapeutic The compound does not melt below 360'. Anal. Calcd. for C4HFN202Hg: N, 8.52. Found: N, index as 5-fluoro-2'-deoxyuridine but required four 8.02. times as much. VI showed a considerably better 1-( 2 ',3',5'-Tri-O-benzoyl-,8-~-ribosyl)-S-fluorouracil.chemotherapeutic index than 5-fluorouridine (I). A mixture of l-0-acetyl-2,3,5-tri-O-benzoyl-~-ribose~~ (0.02 The lyxosyl analog (XII) was totally inactive a t mole) in 250 ml. of anhydrous ether was saturated with all dose levels tried.24 It is to be noted that there hydrogen chloride a t 0". After six days a t 5", the solvent removed i n vacuo and the light sirup treated with 10 ml. was no weight loss with XI1 a t any of the doses that was of benzene. The benzene was removed i n vacuo (bath were used so that i t appears likely that much temperature not exceeding 40') and the sirup treated again higher doses could be given. with benzene. After removal of the solvent, the sirup was treated with toluene and the solution added to a previously azeotroped, refluxing mixture of 5-fluorouracilmercury (0.01 mole) in 350 ml. of dry toluene. The well-stirred DOSE RESPONSEO F LEUKEMIAB82 TO 5-FLUOROURACIL mixture became homogeneous with slight yellowing. After 0.5 hr. of refluxing, the hot toluene was filtered and the NUCLEOSIDES filtrate concentrated i n vacuo to a glass. The glass was dissolved in 100 ml. of ethyl acetate and washed twice with 50InhibiDose, tion, mg./kg./ A wt. Tumor wt. ml. portions of 3oY0 potassium iodide and finally with water. d X 10 Rx/c Rr/c Nucleoside % The organic layer was dried over sodium sulfate, filtered and concentrated under vacuum t o a glass. The glass was 98 100 -4.4/+1.0 24/1229 5-Fluoro-2'treated with 25 ml. of warm chloroform whereupon crystal91 deoxy-uridine 50 -0.5/+1.3 77/852 lization occurred. After cooling, the solid was removed and washed with a small amount of cold chloroform and 82 25 -0.4/+1.8 159/902 ether. From two crops, a yield of 7470 was obtained, m.p. 57 12.5 +0.2/+2.0 419/981 207-209n0, [CY]** -33" ( c 2.0, CHC13). Recrystallization from ethyl acetate-ether gave an analytical sample, m. p. 92 5-Fluorouridine 3.3 -3.6/+1.1 35/453 207-209" 1 . 6 +0.1/+1.1 199/453 56 (1) Anal. Calcd. for C3oH23FN2O9: C, 62.72; H, 4.04; N, 24 0.8 +1.7/+1.1 341/453 4.88. Found: C,62.54; H,4.01; N,4.86. 5-Fluorouridhe (I).-The above tri-0-benzoate (4.5 g.) 100 471 -5.8/+2.0 0/1032 1-p-D-kabinoin 100 ml. of 50% ethanol was treated with stirring with 20 393 +O .5/+4.0 46/723 94 furanosyl-5ml. of 1 N sodium hydroxide. After 2 hr. the PH of the 98 225 +1.7/ ;3.9 10/658 fluorouracil light-yellow solution was between 10-12. ( I t is important that the consumption of alkali with time ceases.) The solu150 1.4/+3.9 38/658 94 (VI) tion was neutralized with glacial acetic acid and most of the 74 100 +2.0/+3.9 170/658 ethanol was removed. The aqueous solution was treated batchwise with Dowex 50 ( H + ) . (During the successive -5 ~-@-D-LYxo210 +0.9/+1.0 975/931 treatments with Dowex, benzoic acid precipitated.) The 0 furanosyl-5105 +0.6/+1.0 935/931 precipitated benzoic acid was removed along with the Dowex 4 fluorouracil 52 +1.4/+1.0 895/931 resin by filtration. (The resin treatment is complete when benzoic acid no longer precipitates and the filtrate exhibits (XI11 a negative flame test for sodium ion.) The aqueous filtrate was washed three times with chloroform and then taken t o It would appear from these preliminary data that dryness in vacuo. The semi-solid obtained was azeotroped the arabinosyl analog (VI) is behaving more like with toluene to remove water and dissolved in a minimum 5-fluoro-2'-deoxyuridine than 5-fluorouridine, both amount of hot absolute ethanol. Upon cooling, 5-fluorouridine crystallized, m.p. 180-182" (88y0). Recrystallizain its over-all toxicity and in the relationship of tion from absolute ethanol did not alter the melting point. toxicity to chemotherapeutic effect. ~-P-D-LYxo- This material exhibited only one spot (paper electrofuranosyl-5-fluorouracil, on the other hand, ap- phoresis, borate, pH 6.2) and was sufficiently pure for use pears to be totally inactive a t the levels tested and in the further syntheses described below. For chemostudies and elemental analyses, a sample of 5to be inactive at a dose level twice that a t which VI therapeutic fluorouridine was chromatographed on Dowex 1 (formate). is effective. These data would appear to demon- The material obtained melted a t 184-185', [ c x ] ~ ~ Df 1 7 " strate the importance of alterations in the sugar (c 2.0, water). Light absorption data: in 1 N hydrochloric moiety on the toxicity and therapeutic activity of acid, maximum a t 269 mp, emax 8950; minimum a t 234 mp, em,* 1680. Spectrophotometrically determined pK, = these 5-fluorinated nucleoside analogs. 7.57. 27 Anal. Calcd. for CgHllFN206: C, 41.22; H, 4.23; F, Acknowledgments.-The authors wish to thank N, 10.68. Found: C, 41.04; H, 4.67; F , 7.12; hT, the Cancer Chemotherapy National Service Center 7.25; 10.73.

TABLE I1

+

for some of the 1-0-acetyl-2,3,5-tri-o-benzoyl-~- 1-(5'-0-Trityl-p-D-ribofuranosyl)-S-fluorouracil(II).ribose used in this investigation. The authors are 5-Fluorouridine (8.3 g.) was treated with 9.8 g. of trideeply indebted t o Dr. George Bosworth Brown of phenylmethyl chloride in 80 ml. of dry pyridine. The amber this Institute for helpful suggestions and con(25) Melting points are corrected unless stated otherwise. Analytitinued interest. (22) J. H. Burchenal and H. F. Oettgen, Cancer Chemofhernpy Rep o r f s (Cancer Chemotherapy National Service Center), 1. 16 (1959). (23) J. H. Burchenal, H. F. Oettgen, J. A. Reppert and V. Coley, ibid., 6 , 1 (1960). (24) Higher doses of XI1 will be tried when more ia availabled

cal data by Spang Microanalytical Laboratories, Ann Arbor, Mich., and by Dr. AI Steyermark of Hoffmann LaRoche, Inc. (26) The authors are indebted to Dr. Max Hoffer of Hoffmann LaRoche, Inc., for his unpublished data for this preparation. (27) The authors are indebted to Miss Iris Wempen for spectrophotometric determination of pKa values.

4064 N.

e. YUNG,J. H. BURCHENAL,RON.!.LD

FECHER, ROBERTDUSCHINSKY A S D JACK J. F O X

vd. 83

solution was kept a t 5' for 16 hr. after which it was heated ml. of solvent was removed in vacuo and the precipitated on a steam-bath for 2 hr. The cooled solution was poured triphenylcdrbinol was fdtered and discarded. The aqueous in a thin stream into one liter of stirred ice-water whereupon filtrate was extracted three times with 15-ml. portions of a reddish gum separated. After decantation, the gum was chloroform and finally concentrated to a light sirup (bath treated again with ice-water. After decantation, the temperature was kept below 50"). Ethanol was added and gummy residue was dissolved in ether and dried over sodium evaporated in vacuo repeatedly. The neutral sirup was sulfate, filtered and the solution was evaporated to a glass dissolved in about 10 ml. of absolute ethanol in which zn z'acuo. The glass was dissolved in a minimum amount crystallization of colorless prisms occurred. A crude yield of ether (-100 ml.). Benzene was added t o faint opales- (8870) was obtained in three crops. Recrystallization from cence and with cooling and scratching long, colorless needles ethanol gave an analytical sample, m.p. 187-188", [aIa4D (This product contains separated (94yO), m.p. 115-119'. $128' (c 0.21, H 2 0 ) . (Reist, et a l . , report +108°9b). solvent of crysta!lization.) Kecrystal1iz;dtion from ethanol Ultraviolet absorption data: at p H 5-6, maximum a t 269.5 Ultraviolet ab- mp, em. 9170; minimum at 234 mp, rmin 1780. Spectrogave colorless rosettes, m.p. 202-203 sorption data: in ethanol, maximum a t 268.5 mp, emer photometrically determined pK. = 7.63." Treatment of 9350; minimum a t 245 mp, emln 4780; [ a ] 2 4 +42" ~ ( c 0.4, VI with sodium metaperiodate and titration of aliquots ethanol). with arsenite showed a slow uptake of one mole of oxidant A n d . Calcd. for C28H2LFN206:C, 66.67; €1, 5.00; N, per mole which was essentially complete after approximately 24 hr. Formic acid was not liberated. 5.55. Found: C,66.77; H, 5.12; N, 5.76. I-( 2'-~-~-Tosyl-5'-O-trityl-~-~-ribosyl)-5-fluorouracil Anal. Calcd. for CleH~FN206: C, 41.23; H, 4.23; F. (m).-A solution of I1 (4.5 g.) in 20 ml. of dry pyridine was 7.25; N, 10.68. Found: C, 41.21; H, 4.48; F, 7.09; N, Treated with 1.87 g. of p-toluenesulfonyl chloride. The 10.68. mixture was agitated until a clear solution was obtained and l-~-~-i%rabinofuranosyl-5-fluorouracil (VI) from (R = allowed to remain a t room temperature for 16 hr. Ethanol (20 ml.) was added and the solvents removed. The reddish F).-A sample of V ( K = F, 4 mg.) in 2 ml. of water was sirup was dissolved in ethanol (60 ml.) with warming (40") refluxed for 4 hr. The reaction was followed spectrophotowhereby precipitation of a granular solid occurred. Com- metrically and showed no alteration of the spectrum during plete crystallization was eflected by cooling. The solid this time period. Hydrochloric acid (2 N, 2 drops) was (52%) was collected on a filter and washed thoroughly with added and refluxing continued for 4 hr. Paper electrocold ethanol. Recrystallization from ethanol gave colorless phoresis ( 3 hr., 700 volts, pH 9.2, borate buffer18) of the refluxed solution showed only one spot a t an identical rosettes, m.p. 178-179"; [ C Y ] ~ ~ D 24' ( c 0.24, EtOH). Ultraviolet absorption data: in ethanol, maximum a t 263 distance from the origin as an authentic sample of VI. l-~-D-Arabhofuranosyluracil.~~Palladium-charcoa1 6800; minimum at 248 mp, em,n 4650. mp, JOO mg.) was added to 100 mg. of VI in 10 ml. ethanol Anal. Calcd. for Ct&IslFPt'yOS: C, 63.81; H, 4.74; (5%, containing one drop of triethylamine. The mixture N, 4.25; S, 4 . S . Found: C, 63.98; H, 5.20; N, 4.49; was shaken under hydrogen at room temperature and S,4.85. atmospheric pressure for 20 hr. The mixture was passed T;he ethanolic mother liquors (designated as XIII-mother through a filter and the filtrate concentrated t o a light sirup liquor on thejlow chart) were combined and used i n subsequent zn vacuo. Dilute ammonium hydroxide was added t o the reactions. sirup and the solution refluxed for 20 min. Solvents again 1-(2'-~-p-~oe;y~-~-~-r~bofuranosyl)-5-~uorourac~ (N).- were removed, then absolute ethanol was added and the 111 (1.5 8.) was added to 50 ml. of ethanol (previously satu- solution cooled. Colorless prisms were collected, m.p. rated with hydrogen chloride) and warmed for 10 min. at 217-219'. A mixed melting point with a sample of 1-650-60". The solution was concentrated i n vacuo t o a sirup D-arabinofuranosyluracil prepared by another routem (bath temperature below 40'). The sirup was triturated showed no depression. with ether repeatedly in order to remove triphenylcarbinol. l-,S-D-Arabinofuranosyluracil from V (R = H).-A sample A granular solid remained (0.84 9.) which was crystallized of VIO,po( R = H, 2 mg.) was refluxed for 4 hr. in 2 ml. of from water-ethanol (3: 1) to give pure IV, m.p. 162-163", water and the reaction followed spectrophotometrically. [ a ] ~-34' * ~ (c 0.55, EtOH). Ultraviolet absorption data: Spectral shifts were noted when reflux temperature was in ethanol, maximum at 266 LUM, emax 7120; minimum a t reached. The solution was chromatographed electro245 mp, emln4000. phoretically (3 hr., 700 volts, p H 9.2 borate buffer) after Ana?. Calcd. for ClSH17FN208S: N, 6.73; S, 7.71. 4 hr. of reflux. Two spots were found, one at the same Found: N, 6.67; S, 7.43. distance as a known sample of 1-8-D-arabinofuranosylura2,2'-Anhydro-l-(~-~-arabinofuranosyl)-5-fluorouracilci1.28 The other spot (like V) did not leave the origin. 1-(2 ',3'-Di-O-mesyl-S '-O-trityl-@-~-ribosyl)-5-fluorouracil (V, R = F).-Sodium hydroxide (0.5 N , 0.9 ml.) was added t o 0.2 g. of IV in 10 ml. of ethanol. The solution was al- (IX).-A solution of I1 ( 3 9.) in 30 ml. of dry pyridine at lowed to stand for 10 min., then adjusted to p H 5-6 with 0-5" was treated while stirring with 0.9 ml. of methylsuldilute acetic acid and the solvent evaporated in vacuo. fonyl chloride. The mixture was allowed to stand a t 5' The residue was dissolved in 5 mi. of water, and the solu- for 16 hr., then poured into 500 ml. of ice-water and stirred for 1 hr. An ochre-colored solid separated and was coltion was passed through a small column of Dowex 50 (H' form) (12 X 1.5 cm). The column was washed with water lected on a filter and washed thoroughly with water. After until no ultraviolet absorbing material was found in the drying in vacuo the amorphous solid weighed 3.15 g. Crystalline IX was not obtained, and the amorphous solid was eluate. The acidic eluates were cumbixied and evaporated used in the following reaction without further purification. irt vacuo at room temperature. Repeated ether extractions 1-(2',3'-Di-O-mesyl-~-~-~bofuranosyl)-5-fluorouraci1 (x). of the s h p gave a residue free of acid (ether extractions discarded). Product crystallizecl readily from absolute -Concentrated hydrochloric acid (10 drops) was added to crude I X (3.15 g.) in 150 ml. of ethanol and was refluxed for ethanol as colorless platelets, 100 mg., m.p. 196-197', 10 min. The ethanolic solution was evaporated to dryness [ a ] % 4 ~ -61" ( c 0.28, ECOH). Ultraviolet absorption data: 7210 and 9050, and the residue triturated repeatedly with hot ether until it a t pK 5-6, maxima at 222.5 and 254 mM, hmnx became crystalline. ,4 yellow solid, 2.0 g., m.p. 160-163", respectively; minimum a t 234 mp, bln 5620. Anal. Calcd. for C ~ H O F N ~ O C,~ :44.28: H, 3.73; N, was obtained. Recrystallization, using preheated 50% ethanol, gave a pure sample, m.p. 172-173", [of] 24D +26' 11.47, Found: C. 44.46; II,3.95; N, 11.42. ( c 0.2, EtOH). Ultraviolet absorption data: In ethanol, I-~-~-i%rabinofuranosy~-5-fluorouracil (VI) from Ill.-maximum a t 263 mjc, ems= 9190; minimum a t 234 mp, Sodium hydroxide (1 N , 12 ml.) was added to a stirring emin3400. suspension of 4 g. of 111 in 100 ml. of 50% ethanol. The Anal. Calcd. for CIIHlrN201&: C, 31.57; H, 3.61. mixture became clear after the addition of alkali and w2s stirred at room temperature fur 3 hr. The solution was Found: C, 31.39; H, 3.77. adjusted to p H 5-6 with acetic acid and most of the ethanol was removed i n vacuo. The precipitated VIII was col(28) J. F. Codingtoa, R. Fecher and J. J. Fox, J . A m . Chem. S O L , lected on a filter and washed thoroughly with water (to 89, 2794 (1960). remove sodium tosylate). The crude nucleoside was placed (29) The authors are indebted to Dr. D. M. Brown of the University Chemicnl Laboratory, Cambridge, England, for a sample of thir in 100 ml. of 50y0 ethanol, containing 5 drops of concentrated hydrochloric acid and refluxed for 1 hr. Approximately 76 compound.

.

v

+

Oct. 5, 1961

SYNTHESIS OF 1-~-D-ARABINOFURANOSYL-5-hUQROURACIL

4065

2 ,Z’-Anhydro-1-(3’-0-mesyl-p-~-arabhofuranosyl)-5-ethanol was concentrated to a light sirup and treated refluorouracil (XI, R = F, R’ = OH).-Sodium hydroxide peatedly with water to remove salts. The aqueous tritu(0.1 N,2 ml.) was added to a solution of 80 mg. of X in rates were discarded and the amorphous solids treated mi th 400 mi. of 50y0 ethanol. Sodium hydroxide (50 ml. of 1 0 ml. of water. The resulting solution was neutral. Crystallization occurred on standing and the colorless prisms (60 N)was added t o the stirred mixture a t room temperature. After stirring overnight, the clear solution was neutralized mg.) were collected, m.p. 192-197’ dec. Recrystallization from water gave a pure sample, m.p. 196-198’ dec., [ a ] 2 4 D with acetic acid (to pH 6) and evaporated in vacuo to a gum. -75” (c 0.26, HsO). Ultraviolet absorption data: a t The gum was dissolved in 500 ml. of chloroform and the PH 5-6, maxima a t 222.5 and 254 mp, emSx 6730 and 8850, chloroform solution washed several times with water. The aqueous washings were discarded and the organic layer was respectively; minimum a t 233 mp, e,in 5540. A n a l . Calcd. for CloH1IFNaO&: C, 37.26; H, 3.44; taken to a sirup (14 g.). The sirup was treated with 200 ml. of 5070 ethanol containing 1 ml. of concentrated hydroN, 8.69; S, 9.96. Found: C, 37.20; H , 3.79; N, 8.91; chloric acid and the mixture refluxed for 1 hr. The clear s, 9.93. solution was reduced in volume in vacuo to approximately l-p-~-Lyxofuranosyl-5-fluorouracil ( X I ) . Method A 10 ml., 50 ml. of water added and the solution extracted from X.-Water (50 ml.) was added t o 1 g. of X and re- several with chloroform. (The chloroform extract fluxed for 4-5 hr. Titration of an aliquot showed that two contains times 6.5 g. of triphenylcarbinol, theor = 7.3 9.) The equivalents of methylsulfonic acid were liberated. The aqueous layer was concentrated to a light sirup and treated solvent was concentrated t o a light sirup and triturated with ca. 15 ml. of ethanol. Colorless short needles precipirepeatedly with ether t o extract the acid. Crystallization tated, 1.0 g., (14% yield based on XIII-mother liquor or occurred on addition of ethanol, 0.45 g., m a p . 195-196’. 6.3% based on the over-all reaction from II), m.p. 193Recrystallization from 9570 ethanol afforded a pure sample, 197’. One recrystallization from ethanol gave an analytical m.p. 203.5-204”, [ U ] ~ ~ D 97’ ( c 0.3, H2O). Ultraviolet ~ W(O c 0.75, water). sample of XV, m.p. 198-200’, [ a ] 2 3absorption data: at pH 5-6, maximum at 269 mp, emax Ultraviolet absorption data: a t pH 5-6, maximum a t 266 9190; minimum a t 234 mp, emin 1760. When treated with mp, emnx 8740; minimum a t 232.5 mp, emin 1910. The epoxmetaperiodate, XI1 consumed one mole of oxidant per mole ide (XV) did not consume metaperiodate over a period of of nucleoside rapidly (within 3 minutes) without the libera- one day. tion of formic acid. Titration of aliquots after this 3 minute -4naZ. Calcd. for ChIIgFNaOa: C, 41.26; H, 3.71; S, period showed no further uptake of metaperiodate. 11.47. Found: C, 44.50; H,3.48; N, 11.50, 11.69. Anal. Calcd. for CQHIIFNZO~: C, 41.22; H, 4.23; F, The mother liquors after epoxide remoual (hereinafter desig7.25; N, 10.69. Found: C, 40.98; H, 4.62; F, 7.11; nated as “X V-mother liquor”) were used in the next erperiN, 10.70. nwnt Method B from XI (R = F, R’ = OH).-hl:ethylsulfonic The isolation of this epoxide perrilits the conclusion (see acid (1 N , 0.06 ml.) was added to 20 mg. of XI (R = F, text) that XI11 (R,R’ = p-tosyl) is a component of XIIIR’ = OH) in 5 ml. of water and refluxed for 4 hr. Only mother liquor in amounts equivalent to the percentage of one ultraviolet absorbing spot was found, migrating identic- XV ally as l-p-~-lyxofuranosyl-5-fluorouracil ( X I I ) , in paper B. Presence of I1 and III in XV-Mother Liquor Proved electrophoresis (4 hr., 700 volts, pH 6, borate bufferla). by Determination of I and VI.-The ethanolic mother liquors 1-p-D-Lyxofuranosyl-5-fluorouracil(XII) from XIII- remaining after isolation of epoxide (XV) above was exMother Liquor.-The ethanolic mother liquor (containing amined electrophoretically in two systems (borate bufrers, theoretically 0,009 mole of products) from the reaction Only two ultra700 volts, pH 6, 4 hr.; pH 9.2, 3 hr.’”). of I1 to I11 (vide supra) was concentrated in vacuo to a sirup violet absorbing spots were revealed. The farther spot and dried azeotropically with small portions of toluene. The (anodically) corresponded to 5-fiuorouridine ( I ) while the sirup was dissolved in 50 ml. of anhydrous pyridine, cooled other corresponded to VI. (Epoxide XV appears to be t o 5” and treated dropwise with an excess (3 ml.) of methyl- absent in this mother liquor. Its electrophoretic migration sulfonyl chloride. After storage a t 6-10’ overnight, the differs from I and VI in the pH 9.2 borate system.) Each contents were poured in a thin stream into 1 1. of stirred of these spots was eluted and its spectrum determined. ice-water. The amorphous solid was collected and washcd The spectra were similar to those for I and VI and did not with water. The solids were added to 300 mi. of 30% resemble the spectrum of IV.lD For the spot corresponding ethanol and treated with 1 ml. of concentrated hpdro- to I, maximum a t 268 mw, minimum a t 234 mp, ratio chloric acid. This mixture was refluxed for 20 hr., after max./min. = 5.7. For pure I (see above), ratio max./min. which the clear, yellow solution was concentrated to a small = 5.3. For the spot corresponding to VI, maximum a t volume (10 ml.) in v a c w (bath temperature not exceeding 268 mp, minimum a t 234 mp, ratio max./mill. = 4.4. 40’). An amorphous solid formed, was collected and For pure VI (see above), ratio max./min. = 5.15. For pure triturated with warm diethyl ether to remove sulfonic acids IV (see above) ratio max./min. = 1.78. and triphenylcarbinol. The solid was dissolved in -250 Spectrophotometric Studies .-The p K , valuesn were deml. of warm ethanol, treated with charcoal, filtered and the termined spectrally by methods previously e m p l ~ y e d ~ , ~ ~ filtrate cooled. Colorless needles were obtained, 73% and are accurate to within 0.05 PH unit. (based upon XIII-mother liquor). Recrystallization from Metaperiodate Studies.--Periodate oxidatioiw were per95% ethanol gave a pure sample of X I I , m.p. 203.5-204’. formed according to procedures employed in previous papers The product was identical with XI1 obtained previously. in this series.8J2 Determination of Components in XIII-Mother Liquor. A. Presence of 5’-0-Trityl-2’,3’-di-O-p-tosyl-S-fluorouridine (39) D. Shugilr nnd J. J. Fox, Riochirn. et B i o p h y s . A d a , 9, 199 Proved by Synthesis of 1-(2’,3’-Epoxy-p-~-lyxofuranosyl)- (1952). 5-fluorouracil (XV).-XIII-Mother liquor (0.028 mole) in (31) J. J. Fox and D. Shugar, Bull. snc. chim. Beiges. 61, 44 (1952)

+

.

.