I
HOOC-C>(
CH,CO\
/O
CH$O
-
I11
Spirans. \ \ I . 9-Hj tl ro\? meth?I-3-azaspiro[S.3]undecane~' ' l,k,:l)x \ti11 11.
l ~ i i l !
( t ( J , C'Oiiey6! Of P h U l ' / l l ~ f i ~ ! l ,
II'ctshinyioti, / I . ('. /pectrLt. 4-Carboxycyclohexane-1,l-diaceticAcid Anhydride (11). A. -. A mixture of 25 g of arid I and 100 ml of .IrCl w w he:itr,i1 t o
173
January 1960
SOTES
reflux aird HOAc was added ruitil a clear solution was obtained. The mixture was refluxed for 2 hr and allowed t o cool. -4fter removittg the wlvents the residue was recrystallized from EtOAc, 20 E, m p 190-192°; recrystallization from EtOAc, nip 195-196'; mmp lf3-1%' with acid I. B.--Xcid I 120 P) was refluxed n i t h 50 ml of A%c20for 5 min and alloirtd to' cool. l i t e r cooling petroleum ether (bp 63-75") was added and the product was filtered, 11 g. After two recrysta!lizations iroin 1~:tOAc-prtroleuin ether, the melting point was
Hon,jo6 reported the oxidation of deoxyuridine and its 3-bromo and 5-iodo derivatives and pointed out the difficulty in obtaining a product as the halogen \ubstituent became more electronegative. The synthetic scheme for the oxidation of FUDK to give 5-flu0 ro-2'-deoxyuri d i ne-S'-c a r box y li c R c i d (FUDA) was not as simple as the one-step reaction might indicate. Previous ~ o r k e r s *had , ~ used decidedly alkaline conditions (pH 8-9) ; however, no oxidation product was obtained with analogous treatment of FUDR. Our earlier method' for the preparatioii of phenyl-P-D-glucopyruronoqide uia the stepwise addition of NaHCOB proved to be fruitful. The p H wa. niuiiitained between G and 7 and the reaction was run for 30 hr at 50-60' to give a (30% yield of FUD.4. The structure of FL-Dh wa. eqtablished by its ir and iimr spectra and by elemental analysis. Conipari4oii of the nmr spectra of F L D A 1us. FUDR in DIO coiifirniy the proposed structure of FUDA, both by the integration of nonexchangeable protoris, and by the fact thLit absorption in the methylene region of r U D R at 6 3 . E is completely absent in FUD-1. While FUDR i:, sensitive to acid hydrolyqis, FCDA4 has been found stable to acid. -1Ithough the nierhariiim of acid hydrolysi. of nucleosides is ambiguouq, most workers agree that the sugar ring 0 atom muit be protonated.8 For example, Garrett and co-workers3 had suggested that the three steps involved in the acid hydrolysis of nucleoqides are (1) protonation of the wgar ring 0 atom, (2) formation of the Schiff base, arid (3) decomposition of the Schiff base. Thus, a. plausible explanation for the enhanced stability toward acid hydrolysis of FUDA compared to F U D R can be provided. I n the case of FUDA, protonation of the carbonyl 0 atom can greatly diminish the rate of protoiiation of the sugar ring 0 atom, thereby retarding hydrolysis. FUDA has been shown to have on117 about one-tenth the toxicity of FUDR in civo in mice. Due to the known lower level of esterase in tumor cells compared t o normal cells,g the esters shomi in Table I u-ere prepared in the hope that, if a more cytotoxic ester could be found, such derivatives would become more permeable to tumor cell membrane, u-hereas normal cells, by virtue of their esterase content, would hydrolyze it hack to FUDL4. The method of synthe4s wa3 acid catalysi., both by use of H,SO, and ion-exchange re+, thereby taking advantage of the stability of FUDA toward acid hydrolysis. During these syntheses, no .?-fluorouracil could be detected in the reaction mixture by means of tlc. The use of dicyclohexylcarbodiimide (DCC) as :I condensing agent for the preparation of esters was attempted. KO product could be found with this method and only the anhydride of FUDA was isolated. Its structure was established by its ir spectrum and complete conversion to FUDA upon hydrolysis. On the other hand, DCC was an effective agent in yielding the amide VI1 from P-naphthylamine. The facile formation of this amide has led us to our present attempts
195-1Nj'.
The filtrates from both A arid B o n removal of the sdverits, followed by hj-drolysih ( S a O H ) of the residue and acidifying gave the origiiial acid I . .inal. (C11H,40i) C, H.
4-Carbomethoxycyclohexane-l,l-diaceticAcid Anhydride (IIJ).-.%cid I1 (18g ) was dissolved in 250 nil of T H F and allowed to cool to 15'. CH?Sd (0.15-0.2 mole) in 400 ml of Et20 at 10" was added, and a-: the reactioti proceeded mo& of the product precipitated oiit of .;olritioii. After rtaiidiiig 2 hr the crystalq were filtei,ed (14 g), nip 11'8-120°. Itecrystallization from EtOA-petrolerim ether gave 13 g, mp 1:10-131°. Anal. iC12-
HlaOj) C, IT. N-Dimethylaminopropyl-9-carbomethoxy-3-azaspiro [5.5] undecane-2,4-dione (IV).- The ester anhydride (13 g, 0.054 mole) was mixed n.ith 6 g of 3-dimeth\.lan~inopropylamineand
-
when homogcntxous w a ~hmted a t 200" for 1 hr. After cooling the product was di>tilled, hp l!):3-200° (0.45 mm), yield 7.6 g. Anal. (Ci;H~2Y?O~) C, H, S . The methiodide was prepared i i i the usual maimer, nip 229230" (from EtOH, 1IeOH). Anal. (C,?H,,ISZO1) I.
N-Dimethylaminopropyl-9-hydroxymethyl-3-azaspiro] 5-51undecane (V).-h solution of 7 g of IT' in 200 ml of Et20 was added to 5 g of LiAlH4 dissolved in 500 nil of EtnO. After 4 h r the niixt itre way deeompohed (H20)in the iisual mailtier, filtered, and dried (XarSO4), and the rolvetit was stripped off. The residue W R Y dihtilled, bp 13:3-138" (0.05 mm), yield 5.!5 g. dnal. (Ci6HJJnO) C , FI, 3 . The hydrochloride was prepared with almholic HCI, mp 295-2!16° (EtOH, MeOH). Anal. (Cl6H,,Cl2N,O)C1.
Synthesis of S-Fluoro-2'-deosyuridine-5'-carboxylic Acid and Its Derivatives1 X. C. Tsoc, S. J. SANTOKI, A N D E. E. MILLER Hni I ihon Depaitiizent of SUIgzcal Reseatch, School of Medzcine, Cniz e i t s i / i j of Pennsylvania, Philadelphia, Pennsylmnia 19104
Xlthough .I-fluoro-2'-deoxyuridine (FUDR) is an efiective arititunior :\gent, it is easily hydrolyzed to 5-fluorouracil in z\ivo and in zlitro, both by phosphorylase arid by ~ i c i d . ~Thus , ~ in the clinical use of this drug the toxicit,y t o the gastroint'estirinl system often masks its chemotherapeutic effectiveness. The present work reports the synthesis of ~?-fluoro-:!'-deoxyuridine-5'carboxylic acid which is less toxic and more resistant to hydrolysis than FUDR. 3loss, et c ~ l . first , ~ reported the synthesis of uronic acid derivatives of nucleosides. They successfully oxidized uridine, thymidine, and adenosine to their 5'-carboxylic acids using Pt as a cat,alyst.5 More recently, Imai and (1) This work was supported IIY U. 8. Public Health Service G r a n t CA 07339: presented a t t h e XCS Mid-.ltlantic Regional Meeting of t h e American Ciiernical Society. Philadelphia, P a . , Feh 1968. (2) C. Heidelherger. G. D. Birnie, J. Boohar, a n d D. Wentland. Biochim. B i o p h y s . 4 c l a , 76,315 (1963). (3) E. R. G a r r e t t , J. D. Seyrlel, a n d A. J. Sharpen, J . Org. Chem., 31,2219 ( 1963).
(4) G.P. Moss, C.B . Reese, K. Schofield, R . Shapiro. a n d A . R. T o d d , J . Ciiem. Soc., 1149 (1963). ( 5 ) K . C. Tsou and A . RI. Seligman, J . A m . Chem. Soc., 74, 5605 (1952).
(6) K. I m a i a n d M. Honjo, Chem. Pharm. Bull. (Tokyo), 13, i (1965). ( 7 ) K.C. Tsou a n d A. M. Seligman, J . Am. Chem. Soc., 76, 1042 (1953). ( 8 ) G. W. Kenner, Ciba Foundation Symposium o n t h e Chemistry a n d Biology of Purines, Little, Brown a n d Co., Boston, Mass., 1957,p 312. (9) K. C. Tsou, H. C. Su, C. Segebarth, and U. SIirachi, J . Org. Chem., 26,4987 (1961).
