March 5 , 1955
6-METHYL-USy?lZ-TRIAZlNE-3,5 (23)-DIONE RIBOSIDES
1145
ice-water bath and alkaline sodium hypochlorite solution*6 a t 20'. The reaction in the homogeneous mixture was (14.9%, 1550 ml.) added dropwise with stirring so that the followed chromatographically in solvent A , nucle(2tide conmixture refluxed gently. After addition of the sodium hypo- taining spots being eluted with dilute hydrochloric acid chlorite solution (45 minutes) the mixture was stirred a t 20' (0.1 hl) and estimated spectrophotometrically. Reaction For 12 hr. The organic layer was then separated, the aque- was complete in 1-2 hr., PIPz-diadenosine 5'-pyrophosphate ous layer extracted with methylene chloride and the com(62%) being the only new product ( R f , 0.2, AMP, 0.3). bined organic solutions washed with brine and dried over Under otherwise identical conditions the reaction in water Drierite. Removal of the solvent gave a brown oil (170 9.) (0.36 nil.)-pyridine (0.35 ml.) and water (0.6 m1.)-pyridine which was dissolved in light petroleum (b.p. 20-40", 500 (0.1 ml.) gave 25 and IS%, respectively, of the pyrophosml.) and kept at 0' overnight. Unchanged thiourea (23 g.) phate. which separated was removed by filtration and the clear The Reaction of AMP and Orthophosphoric Acid with solution was concentrated to an oil which was distilled in a 1V.-AMP (36.5 mg. of the monohydrate, 0.1 m J 1 1and mm.) to give orthophosphoric acid (116 mg., 1.0 mM) in water (0.42 short path moiecular still a t 45-50" (5 X the carbodiimide a s a pale yellow oil (104.9 g., 67y0) n''D nil.)-pyridine (2.52 ml.) was treated with the diinlide (3.79 1.4861. An analytical sample was redistilled a t 45' (5 X g., 10.0 m h f ) a t 20' (cooling required, initially) for 1 hr. mm.); n22 6~ 1.4862. Anal. Calcd. for C I ~ H Z T S Solvent ~: was removed in vacuo a t 20" and the residual gum C, 70.80; H, 11.47; N, 17.71. Found: C, 70.64; H, 11.46; taken up in acetone (100 ml.) containing barium iodide N, 17.63. (2.136 g. of the dihydrate, 5.0 m M ) . After keeping the The methiodide (IV) was obtained by dissolving the carbo- mixture a t 0" overnight, the precipitate which separated diimide (104 g.) in light petroleum (b.p. 20-40°, 500 ml.) was spun down, washed with acetone and dried. The dry and treating with methyl iodide (100 ml.) a t 0" for 7 days. white powder was suspended in water and stirred with 1.R.The methiodide separated a s a n oil which was redissolved 120 ion-exchange resin (sodium form). The resulting soluin acetone and reprecipitated with light petroleum. This tion was made up to 10 ml. with water and an :aliquot (2 process was repeated until a n aqueous solution of the heavy ml.) chromatographed on an analytical ion-exchan e: column oil was neutral. as described before. The recovery of optical density was Reaction of AMP with l-Cyclohexyl-3-(~-diethylamino- 50%; AMP, 28; ADP, 15; and ATP, 56%. propyl) Carbodiimide Methiodide (IV).-AMP (10 mg., Acknowledgment.-This work was carried out 0.0275 mfif) in a mixture of water (0.1 ml.) and pyridine (0.6 ml.) was treated with the diimide (379 mg., 1.0 mM) under Defence Research Board of Canada grant (25) E. Schmidt, F. Zoller, F. Moosmuller and E. Kammerl, Ann., 585, 230 (1954).
[CONTRIBUTION FROM
THE
number 2020-23 Project D52-20-20-23.
\'ANCOUVER
8, !d.
e.,CANADA
NUTRITION AND PHYSIOLOGY DEPARTMENT, RESEARCH DIVISIOX, AMERICAN CYANAMID Co. ]
Riboside Derivatives of 6-Methyl-asyrn-triazine-3,5(2,4)-dione1 BY Ross H. HALL RECEIVED AUGUST23, 1957 The chemical synthesis of 2-~-ribofuranosyl-6-methyl-asym-tr~az~ne-3,5(2,4)-d~one is described. The synthesis of the corresponding 4-ribosyl and 2,4-diribosyl derivatives and their identification is also reported. A facile ring opening of the 4-substituted derivatives is described.
Various investigators have synthesized analogs of the naturally occurring pyrimidines for use as experimental anti-tumor compounds. The suggestion has been made recently that the nucleosides of such compounds may be more effective than the free bases.2 A publication3 supporting this concept demonstrated the value of the deoxyriboside of 6 - methyl - asym - triazine - 3,5(2,4) - dione (azathymidine) as an inhibitor of deoxyribonucleic acid (DNA) synthesis in contrast to the inertness of the free base. The deoxyriboside is obtainable a t the present time only by enzymic means,3 but in view of the evidence of Roll, et al.,4 and Rose and Schweigert6 that facile conversion of ribosides to deoxyribosides occurs within living cells it was (1) This base is often referred to as 6-azathymine. ( 2 ) Examples of the effectiveness of nucleosides over the corresponding free bases as antimetabolites are: 5-bromodeoxyuridine Versus 5-bromouracil, T. J . Bardos, G . M. Levine, R. R . Herr and H. L. Gordon, TEIS JOURNAL,77, 4279 (1955): 6-azauridine Yeisus 6azauracil, R. E. Handschumacher, Fedemlion Proc., 16, 191 (1957); A. D. Welch and R. Schindler, Scwncc, 126, 548 (1957); azathymidine versus azathymine, W. H. Prusoff and A . D. Welch, J . Bioi. Chcm., 218, 929 (1956), and W. H. Prusoff, L. G. Lajtha and A. D. Welch, Biockim. el Biopkys. Acto, 90, 209 (1956); W. H. Prusoff, J . Biol. Ckcm., 226, 901 (1957). (3) Ross H. Hall and R. Haselkorn, THIS JOURNAL, 80, 1138 (1958). (4) P. Roll. H. Weinfeld, E. Carroll and G. B. Brown, J . Bioi. Chem., 290, 439 (1956). (5) I. A. Rose and B. 9. Schacigert. ibid., 402, 635 (1853).
decided to synthesize 6-methyl-asym-tria,zine-3,5(2,4)-dione riboside via chemical means in order to evaluate its use as an anti-cancer agent. This paper describes the synthesis of a mixture of ribosides, their separation and identification. A mercury salt of 6-methyl-asym-triazine-3,5(2,4)-dione (VIII) was prepared and this was condensed with l-chloro-2,3,5-tri-0-benzoyl-~-ribose,~ according to the procedure of Davoll and Lowy.7 The resulting mixture of products (65-7056 yield) was separated on alumina and silicic acid columns to yield three products, IV, V and VI.s Each of these compounds was catalytically debenzoylated in anhydrous media by sodium methoxide to yield the three riboside derivatives I, I1 and 111. These three derivatives also were obtained by first debenzoylating the condensation mixture, then separating the free nucleosides by partition chromatography on Celite. The analytical data indicated that two of the compounds were monoribosyl derivatives of 6-rnethyl-asym-triazine-3,5(2,4)-dione, while the third was a diribosyl derivative; ( 6 ) R . K . Ness. H. W . Diehl and H . G . Fletcher. THIS JOURNAL, 76, 703 (1954). (7) J . Davoll and B . A . Lowy, i b i d . , 74, 1563 (1952). (8) The glycosidic linkages of all formulas are written arif they were in the 8-configutation, although there is no evidence to suplport this.
1146
ROSS
Vol.
1-r. HALL
’2 0
so
1
CH3NH-C-NH- N OBz 662
+Ip
II HO C-C-CH3 II
t 1
evidence for location of the sugar residues on the group to N-4 was demonstrated by heating the compound with strong alltali. This caused scission base will be given. Structure 111, 4-~-ribofuranosyl-6-methyl-asymof the ring giving the -1-niethylsemicarbazone of triazine-3,5(2,4)-dione,was assigned after study- pyruvic acid, XII, which was identified by coniing the properties of a model compound, 4,G-di- parison with an unambiguously synthesized sammethyl-asym-triazine-3,5 (2,4)-dione (XI). This ple. The ultraviolet absorption spectra of I11 compound was prepared by methylation of VI11 closely resembled that of the model compound XI In particular the alkaline bathowith methyl iodide. Attachment of the methyl (Table T).
March 5 , 1955
6-METHYL-USy??Z-TRIAZINE-3,5(2,4)-DIONE RIBOSIDES
chromic shift was characteristic. Shugar and Foxg reported a similar alkaline bathochromic shift for 3-methyluracil but not for 1-methyluracil. The ultraviolet absorption peak of the other monoriboside, compound I, underwent a shift to shorter wave lengths in alkali; therefore, of the two monoribosides only compound I11 merited consideration as n four substituted asym-triazine. TABLEI ULTRAVIOLET ABSORPTIONOF COMPOUNDS 0.1 N HCI Derivative of 6-methylasym-triazine3,5(2,4)-dione
2-Ribosyl2-Deoxyribosyl2,4-Diribosyl4-Ribosyl4-Methyl-
0.1 N N a O H
--H20--
frna.
€mal
A,,
(X 1000)
(X
Amsx
1000)
A,,,,,
€mar
262 262 267 264 260
4.95 5.80 4.3 5.1 5.2
262 262 267 264 260
5.10 5.80 4.15 5.15 5.0
251 251 ,. 304 298
(X 1000)
5.4 6.2
..
.. 7.9
4-Ribos ylsemicarbazone of pyruvic acid 257
4 . 6 248 6 . 7 248 6 . 0 Semicarhazone of pyruvic acid 257 7.G 245 7 . 5 5 248" 7.55 Semicarbazone of pyruvic acid from 2,4-dirihosyl derivative 257 248 248 Seinicarbazone of pyruvic acid from 4-ribosylderivative 257 7 . 6 248 7 . 5 248 7 . 5 2,4-Dimethyl274 6 . 9 273 6 . 4 273 6 . 3 a In 1 *V alkali A,,, is 279, emax is 7,600. This denotes a new dissociation with p k greater than 13.
Further evidence as to the location of the ribose in compound 111 was forthcoming on studying the infrared spectra of compounds 111 and XI. Both these compounds had the same general absorption pattern. Of particular significance was the pattern in the carbonyl region; here the pattern of the two derivatives was identical. There were two bands at 5.78 and 6.00 K , respectively. This indicated that compound 111,like X I , was in the diketo form. Thus compound I11 was not an 0-glycoside, a conclusion supported by the fact that this compound was relatively stable in boiling 1.0 N hydrochloric acid.'O One other possibility for compound I11 that was evaluated was attachment of the sugar a t position one. This would mean an internally compensated quaternary compound, in order that the analytical data should apply. Two possible structures are
*
,-CHs
I
N-R
A N / 0-
~ - - C I &
JJ-.
+
R
=
O H 2-deoxy-D-ribose
.A positively charged linkage such as this would be expected to decompose easily in acid or alkali analogous to nicotinamide riboside." compound (9) D. Shugar a n d J. J. Fox, Biochim. el Biophys. Acto, 9, 199 (1952). (10) P. A. Levine and H. Sobotka studied the lability of pyrimidine 0-glycosides in acid and base: J . Bid Chem., 66, 469 (!925). (11) N. A. Hughes, G . W Kenner and Sir Alexander Todd, J . Chem. Sac., 3733 (1957); T. P. Singer and E. B. Kearney, Ado. iiz Enrymol., 16, 79 (1954).
1147
111, although relatively acid-stable, did undergo a rapid ring scission between atoms four and five in weak alkali, demonstrating a fundamental instability in this medium. This reaction is described in detail below. Such an occurrence is similar to that reported for the breakdown of 3,5'-cyclo-6dimethylamino- 9 - (3' - deoxy - ,B - D - ribofuranosy1)purine-2',3'-carbonate methanesulfonate which has a charged nitrogen atom in the pyrimidine ring.'? However, the infrared spectrum of compound I11 does indicate the presence of two keto groups, SO that a charged structure like those above does not seem reasonable for this compound. The instability of compound I11 in alk,sli was studied; a t a p H value of 10, in a few minutes, it added one molecule of water undergoing ring opening to a ribof uranosyl semicarbazone of pyruvic acid, IX. Compound I X had an ultraviolet absorption spectra identical with that of the semicarbazone of pyruvic acid (Table I). Thus the sugar residue was in such a position as not to modify the spectral characteristics ; that is to say, i t could be either a t position two or four of the semicarbazone but in accord with structure I11 would have to be a t position four. It should be pointed out that if compound I11 contaiced the sugar residue at N-1 a s discussed above, a rapid alkaline rearrangement of the sugar from atoms one to two would have had to be realized. There is little evidence that such rearrangements occur. That compound I X was indeed a semicarbazone of pyruvic acid, X, was confirmed by removal of the sugar residue by mild acid hydrolysis. The hydrolytic product was identified by comparison with an authentic sample. The alkaline lability of the bond between atoms four and five of compound 111has its analogy in the chemistry of dihydrothymine and dihydrouracil.13 Mild alkaline treatment produces the corresponding ureido compound in each case. As noted above, the 4-methyl substituted asym-triazine required vigorous alkaline treatment in contrast to thl. labile 4-glycosyltriazine for ring scission, but in each case scission did occur in the same location. I11 summation, all the spectral and hydrolytic data are in accord with formulation of compound 111 as 4~-ribofuranosyl-6-methyl-asym-triazine-3~5(2,4) - dione. The structure of compound I was arrived. a t by a process of elimination. Establishment of the position of the sugar residue in compound 111:eliminated N-4 as the point of attachment for I. Much of the same argument for not considering II:[ as an 0-ribosyl derivative or an N-1 substituted t-lerivative also applies to compound I. Mainly, this compound was relatively stable in boiling 1.0 N alkali and acid. Therefore, compound I was considered to be 2-~-ribofuranosyl-6-methyl-as3;mtriazine-3,5(2,4)-dione. The enzymatically prepared deoxyribo~ide~ had an ultraviolet a.bsorption pattern identical with that of I (Table I) and was therefore presumed to be 2-(2'-deoxy-:D-ribofuranosyl)-6-methyl-asynz-triazine-3,5(2,4) -diene. (12) B . R . Baker and J . P. Joseph, THISJ O U R N A L , 77, 1 5 ( 1 9 5 5 ) . (13) R. M . Fink, C . McGaughey, R. E Cline and K . Fink, J . B i d , Chem., P18, 1 (1956); R . D. Batt, J. K . Martin, J . Pluessrr a d .I. Murray, THISJOURNAL, 7 6 , 3663 (1954).
1145
ROSS H. HALI,
From microanalytical data, compound I1 appeared to be a diriboside of VII. The location of the riboside residues at position two and four were determined as follows : Mild acid hydrolysis resulted in a quantitative yield of compound I. Vigorous acid hydrolysis seemed to remove the sugar residue from position one slightly faster than from position three because just prior to complete hydrolysis to the free base VI1 it was possible to demonstrate a small quantity of compound IT1 on paper strips. The identification of this hytlrolytic product was confirmed by Rr value on paper strips, its characteristic ultraviolet absorpticm spectra, and its distinctive behavior to alkali. Like compound 111, compound I1 was extremely labile to alkali and underwent ring opening to give a product which after acid hydrolysis appeared to be identical with the semicarbazone of pyruvic acid, X. The main purpose of this problem was the synof 2-~-ribofuranosy1-6-methyl-asym-trithesis azine-3,5(2,4)-dione (I) and when the chemistry of the three ribosyl derivatives had been established, a more direct pathway to this end became evident. After the initial condensation, the mixed benzoylated isomers were dcbenzoylated ; care was taken not to permit contact with aqueous alkali. The mixture of nucleosides was treated first with dilute acid to hydrolyze I1 to I then with alkali to convert I11 to IX. The problem then became one of separation of I from IX, easily accomplished by ion exchange chromatography. I n this manner compound I was obtained in 41% over-all yield. The compound mas a strong complexing agent so that it was necessary to remove bound metal ions by sulfide precipitation prior to final isolation of the nraduct.
VOl. 80
0.1 N hydrochloric acid and finally with water. Evaporation of the chloroform solution left 21.2 g. of a fluffy solid. This consisted of a mixture of the three isomeric nucleoside5 mid unreacted ribose, al! as the tribcnzoylated derivatives. It is termed mixture .4. Separation of Mixture A on Activated Alumina.-Merck reagent grade alumina (acid-washed) was used directly from the bottle in preparing a column 7.1 cm.2 X 41 cni. (300 g.), with benzene as the initial solvent. Mixture A (10.f3 9.) was dissolved in 40 cc. of benzene and run on the column. The column was washed with the following solvents a t a flow rate of 10 cc./minute: 2.2 1. of benzene, which elutccl 1.B g. of unreacted sugar product (1n.p. 69-72.'); 2 I. of 5 7 0 ethyl acetate in benzene which eluted 3.37 g. of V antl a ribose impurity; 3 1. of 10% ethyl acetate in benzene which eluted 2.8 g. of 117; 2.2 1. of 2SojC ethyl acetate in benzene which eluted 3.1 g. of VI and IV. Purification of VI was achieved by running a second column on alumina of greater activity. The above alumina (130 9.) was heated at 180" for 24 hours and used to prepare A gradient elution between 1 ii column 3.2 cm.2 X 50 cin. 1. of benzene and 1 1. of ethyl acetate was set up with increasing amounts of ethyl acetate. Compound V I (2 g.) uas eluted before IV. Compound VI was crystallized from :J riiixture of methanol antl petroleum ether to give crystals, 111.p. 155.5-1
