EDWARD C. TAYLOR AND KLAUS S. HXRTKE
2456
FRICKCHEMICAL LARORATORP, PRINCETON
[CONTRIBUTION FROM THE
VOl. 81 USIVERSITY]
The Reaction of Malononitrile with Substituted Hydrazines: New Routes to 4Arninopyrazolo [3,4-d]pyrirnidines112 B Y EDWARD c. TAYLOR AKD
KL.%US
s. H A R T K E ~
RECEIVED SOVEMBER 22, 19,58 The reaction of malononitrile or ma1ono:iitrile dimer (1,1,3-tricyano-2-aminopropene-l) with phenylhydrazine and inethylhgdrazine is showii t o give 1-phenyl- and l-rnethyl-3-cyanomethy1-4-cyan0-5-atninopyrazole ( I and XVIII), respec(XII’), and the structure of tively. The structure of I is established by degradation to l-phenyl-3-methyl-5-aminopyrazole XVIII is established by drgradation t o 1,3-dirnetii!;l-4-nitroso-li-pyrazolone (XXIX). These readily available pyrazoles are versatile intermediates for the synthesis of condensed pyrazole heterocycles, including a number of derivatives of 1phenyl- and l-methyl-4-aminopyrazolo[3,4-d] pyrimidine.
In the preceding paper,4 the reaction of malononitrile with hydrazine was shown to yield %cyanomethyl-4-cyano-5-aminopyrazoleand not 3,S-diaminopyrazole as previously c l a i ~ n e d . ~The present paper discusses the structures of the products formed by reaction of malononitrile with two representative substituted hydrazines, phenylhydrazine and methylhydrazine, and illustrates the usefulness of the resulting products as intermediates for the synthesis of condensed pyrazole heterocycles. The reaction of malononitrile or malononitrile dimer4with phenylhydrazine led to a colorless, crystalline solid, m p. 166’, with the molecular formula C12HJi6.6 Of the six possible formulations ( I VI) for this product, only I and I1 are consistent with all of the following observations: (1) the compound is recovered unchanged after prolonged heating in ethanolic sodium ethoxide solution (excluding structures V and VI) ; ( 2 ) the compound shows infrared absorption bands a t 2.93, 3.05 and K C n C H , C N H2N4
.a
?
Ph
I
NC-yCHzCN HJ”‘K ,iX-Ph I1
IV
Ph I11
NCCH2CLC(CKj2
XCCH,C=CiCN)Z
I
I
SH LT
I h”Ph
?-Ph
\‘I
“2
3.13 p (N-H bands) and two nitrile bands, one a t 4.46 (unconjugated) and one a t 4.54 p (conjugated) (excluding structures I11 and IV); (3) the ultraviolet absorption spectrum given by the compound shows only decreasing absorption with increasing wave length, with no maximum above 220 rnp (ex(1) This work was supported in part by a grant (C-25B1) t o Princeton University from t h e National Cancer Institute of the Kational Institutes of Health, Public Health Service. (2) Presented before t h e Divisioo of Organic Chemistry a t the 2nd Delaware Valley Meeting ef the A.C.S., February 5 , 19511, in Pniladelphia, Pa., and before the Division of Organic Chemistry a t the 133rd Annual ACS Meeting, April 13-iE5, 1958, in San Francisco, Calif, (3) V i s i t i n g Scholrrr from the University of Liarburp. sponsored by Der Stifterverband fur die Deutsche Wissenscbaft. ( 4 ) E. C. Taylor and K. S. Hartke, TIIX? JOUIINAL, 81, 2452 (1930). (5) R. “on Rothenburg. B17..27, 685 (1894). (6) von Rothenburg (ref. 5 ) reported t h a t t h e reaction of malononitrile with phenylhydrazine yielded a dark brown oil for which no structure was suggested.
cluding structures V and VI. since malononitrile dimer shows a strong absorption maximum a t 273 mp) ; (4) both the infrared and ultravolet absorption spectra given by the compound are similar to the spectra given by the product of the reaction of malononitrile with hydrazine itself which has been shown? to be 3-cyanomethyl-4-cyano-5-aminopyrazole. The choice is thus narrowed to structures I and 11. In the subsequent discussion of the proof of structure of the compound ClrHeNs, structure I , which will be shown to be correct, will be employed throughout for the purposes of both brevity and clarity. Initial attempts to degrade I to a simple pyrazole were unsuccessful. Hydrolysis with 5% sodium hydroxide yielded 1-phenyl-3-carboxymethyl-4-carboxamido-3-aminopyrazole(VII), but attempted decarboxylation of this compound led only to decomposition. Conversion of I to the imino ether IX, followed by mild aqueous hydrolysis to the ester X, and finally saponification with sodium carbonate, gave 1-phenyl-3-carboxymethyl-4-cyano-5aminopyrazole (XI), but attempted decarboxylation of X I resulted in riiig closure to give 2-phenyl3 - amino - 4,6 - dihydroxypyrazolo(4,3 - c)pyridine (VIII). Compound VI11 could also be formed by treatment of I or VI1 with strong acids. Hydrolysis of X with dilute sodium hydroxide gave VII. This observation confirms the assigned positions of the carboxamido and carboxyl groups of VII, since the infrared spectrum of X definitely places the nitrile group in the 4-position of the pyrazole ring (conjugated nitrile a t 4.51 p). A projected Curtius degradation sequence starting with the ester X likewise proved to be fruitless, for it was not possible to convert it either to the amide or to the hydrazide. For example, X was recovered unchanged after heating with alcoholic ammonia a t 150” for 8 hours. I t seems probable that removal of a proton from X yields a resonance stabilized anion (XII) which resists further nucleophilic attack by ammonia or hydrazine. Complete hydrolysis of both nitrile groups of I therefore appeared to be a necessary prelude to further degradation of the molecule. Treatment of I with boiling 30’% sodium hydroxide for 8-10 hours resulted in vigorous evolution of ammonia and the formation of a monocarboxylic acid, CI1HIIN102. Proof that this compound was l-phenyl-3-carboxymethyl-5-aminopyrazole (XIII) and not the isomeric l-phenyl-3-rnethyl-4-carboxy5aminopyrazole(XV) was obtained as follows : ~
MALONONITRILE WITH SUBSTITUTED HYDRAZINES
May 20, 1959
2457
to be l-phenyl-3-cyanomethyl-4-cyano-5-aminopyrazole (I).9 The reaction of malononitrile or malononitrile dimer with methylhydrazine yielded a white, crystalline solid, m.p. 178', with the molecular formula C7H,N6. As a result of arguments similar to those employed previously in the discussion of the structure of I , it was apparent that this product was either 1-methyl- or 2-methyl-3-cyanomethyl-4-cyano-5-aminopyrazole (XVIII or XIX, respectively). I t was unfortunately not possible to assign struc-
NCI!CHz HZN NCr-+H "N-CHB
HzN
$J
AH3
VH
IOH-
I VI1
Ph
XI11
XVII
COOH
On the basis of the above degradation sequence, the structure of the compound ClzHgNs, formed by the reaction of malononitrile or malononitrile dimer with phenylhydrazine, is firmly established (7) C. C. Cheng and R. K. Robins, J . Or& Chcm., 21,1240 (1956). (8) E. Mohr, J. p m k l . Chcm., 79, 1 (IQOQ).
XVIII
XIX
ture XVIII to this product by analogy with the established structure I for the phenylhydrazine product because of the different basicities of the nitrogen atoms in methyl- and phenylhydrazine. Thus, aromatic hydrazines are known to acylate exclusively on the 6-nitrogen, while aliphatic hydrazines react predominantly a t the more basic a-nitrogen.'l-13 This difference is also manifested in pyrazole syntheses. For example, reaction of ethyl 0-amino-6-ethoxyacrylate with phenylhydrazine yields l-phenyl-3-amino-5-pyrazolone, while reaction with methylhydrazine leads predominantly to the other isomer, l-methyl-3-hydroxy-5-iminopyrazole I 4 b I 6 An independent structure proof for the methylhydrazine product C7H7Nb was therefore mandatory. However, in order to simplify the following discussion, the methyl group will be assigned its correct position (structure XVIII). A number of the reactions of XVIII proved to be analogous to the reactions previously discussed of I. For example, the action of dilute alkali on XVIII led to l-methyl-3-carboxymethyl-4-carboxamido-3-aminopyrazole (XX), while treatment with strong acids yielded 2-methyl-3-amino-4,6dihydroxypyrazolo(4,3-c)pyridine (XXI) ; X X was converted with a mixture of ethyl orthoformate and acetic anhydride by cyclization across the active methylene group and the carboxamido group to 2methyl- 3 - amino-4-hydroxy-7-carboxypyrazolo (4,3c)pyridine (XXII). A similar reaction carried out with VI1 led to the corresponding phenyl derivative XVII. Hydrolysis of XVIII with 30% sodium hydroxide yielded a monocarboxylic acid which was shown to be l-methyl-3-carboxymethyl-5-aminopyrazole (XXIII) rather than the isomeric 1,3-dimethyl-4-carboxy-5-aminopyrazole(XXVI) by arguments similar to those employed in the establishment of the structure of XIII. Thus, 1J-dimethyl(Q)Since the initial presentation of this material (see ref. 2), the reaction of malononitrile dimer with Phenylhydrazine has been described (R. A. Carboni, D. D. Coffman and E. G. Howard, THISJOURNAL, 80, 2838 (1958)) as leading to a compound believed to be 2phenyl-3-cyanomethyl-4-cyano-5-aminopyrazole(II), although it was stated t h a t the isomeric structure I could not be eliminated as a possibility. I t is now shown that the product of this reaction is indeed I. (10) E.Fischer, A m . , 190, 67 (1878). (11) bl. T. Folpmers, Rec. Iran. chim.. SC, 34 (1915). (12) A. iMichaelis and E. Hadanck, Bey., 41, 3285 (1909). (13) R.L. Hinman and D. Fulton, THISJOURNAL, 80, 1885 (1958). (14) A. Weissberger, H. D. Porter and W. A. Gregory, ;bid., 66, 1851 (1944). (15) B. Graham, H. D. Porter and A. Weissberger, ibid., 71, 983 (1949).
243
EDWARD C . TAYLOR .\ED KLAUS S. HARTHE
VOl. 81
4-cyano-5-aminopyrazole (XXV)7 was prepared from methylethoxymethylenemalononitrile and methylhydrazine and hydrolyzed with 30y0 sodium hydroxide. T h e only product which could be isolated was 1,3-dimethyl-5-aminopyrazole (XXIV), for the intermediate &-carboxylic acid XXVI proved to be extremely unstable and decarboxylated spontaneously a t room temperature. Conversion of X X I I I to XXIV was achieved by xx ioH CH? heating a t 200' in vacuo. Finally, a third independent synthesis of XXIV was carried out by treating the dimer of acetonitrile (1-cyano-2-aminopropene-1) (XXVIII) with methylhydrazine. None of the above syntheses, however, establishes with certainty the position of the N-methyl group in XXIV. The structure of 1,3-dimethyl-4cyano-5-aminopyrazole (XXV) had been assigned by analogy with the phenyl isomer, whose structure had been rigorously established,' but it has already been pointed out that analogies based on the assumption that phenylhydrazine and methylhydrazine react similarly are not compelling. The synthesis of XXIV from the dimer of acetonitrile and methylhydrazine confirms the pyrazole nature of the product, but sheds no light on the position of the N-methyl group. It remained, therefore, to establish with certainty the structure of 1,3-dimethyl5-aminopyrazole (XXIV) . Both 1,3-dimethyl-5-pyrazolone (XXX) 16-18 and 1,5-dimethyl-3-pyrazolone (XXXI) 19 are known, and i t was thought the establishment of a relationship between XXIV and one or the other of these XXVII XXVIII CH3 XXIX CH3 compounds would serve t o establish its structHO.0 ture with certainty. Therefore, XXIV was nitroCH ,CCH,COOEt C H , N H N H ~ 7 CH3 sated to 1,3-dimethyl-4-nitroso-5-aminopyrazole I1 (XXVIII), which was smoothly converted by heat0 O=yXH 1 ing with alkali to 1,3-dimethyl-4-nitroso-5-pyrazoCH? lone (XXIX). 1,3-Dimethyl-5-pyrazolone (XXX) CH3CCHzCOOEt * xxx was then prepared by the known methodlS from II i ethyl acetoacetate and methylhydrazine, and nitroA, sation yielded a product identical in all respects H:J2kCOSH2 with X X I X . A careful reading of the literature revealed, howThe product of the reaction of malononitrile or ever, that even the long accepted structures of X X X malononitrile dimer with methylhydrazine is thus and X X X I had never been firmly established. conclusively established as 1-methyl-3-cyanoKnorrlGassigned the structure 1,3-dimethyl-5-py- methyl-4-cyano-5-aminopyrazole (XVIII). razolone (XXX) to the product of the reaction of Considerable recent attention has been given to ethyl acetoacetate and methylhydrazine appar- the 4-aminopyrazolo(3,4-d)pyrimidinesystem beently on the assumption that the first step in the cause of the significant antimitotic activity exhibreaction must involve the formation of the methyl- ited by a number of its derivatives.z1 The reachydrazone of ethyl acetoacetate. The structure of tion of malononitrile or malononitrile dimer with the isomeric 1,5-dimethyl-3-pyrazolone(XXXI) substituted hydrazines has been shown to lead was then apparently assigned by default.lg Fortu- directly and in good yield to pyrazole intermediates nately, an unequivocal confirmation of the correct- ( I and XVIII) potentially capable of ready converness of Knorr's structural assignment was possible. sion to this class of condensed pyrazole heterocycle. It has been reportedz0 that 3-methyl-5-pyrazolone We therefore investigated methods for their possimay be prepared by gentle pyrolysis of the seniicar- ble utilization for such syntheses. bazone of ethyl acetoacetate. We therefore prel-Phenyl-3-cyanomethyl-4-cyano - 5 - aminopyrapared the N-methylsemicarbazone of ethyl aceto- zole (I), upon treatment with diethyl oxalate in acetate, which upon heating was converted into the presence of potassium ethoxide, yielded an 1,3-dimethyl-5-pyrazolone (XXX), identical in all ethoxalyl derivative to which structure X X X I I respects with Knorr's product. was assigned on the basis of its subsequent reacf l l i ) L. Knorr, A w ; / . ,279, 232 (1x94). tions. Treatment of X X X I I in tetrahydrofuran ( 1 7 ) I-. Wolff and W. Schreiner, B e y . , 41, 6.50 {1!lO8) solution with diazomethane yielded a monomethyl (18) K. v. Auwers and F. I\-iemeyer. J . p i i i k l . Ckeii! , 110, 13J
JCH,NHNH~
-
(1512,5).
C A . Rojahn, B e y , , 5 5 , 2939 ( 1 9 2 2 ) . ( 2 0 ) J Thiele and 0. Stange. Aiin.. 283, I (1894).
(1'1)
(21) €1. E . Skipper, R. K . Robins, J. R. Thomson, C. C. Cheng, R .
\V.Brockman and F. M Schabel, J r , Cancer Research, 17, 579 (1957), and references cited therein.
May20, 1959
derivative which can only be 1-phenyl-3-cyanomethyl- 4-cyano - 5 - carbethoxymethoxymethylene aminopyrazole (XXXIII), since its infrared spectrum shows only one carbonyl peak (5.73 p ) and no N-H band. By contrast, X X X I I shows two carbonyl bands (5.64 and 5.76 p ) and one N-H band (3.12 p), thus conclusively placing the position of methylation on oxygen rather than on nitrogen. Finally, treatment of XXXIII with alcoholic ammonia resulted in a facile ring closure to give 1-phenyl-3-cyanomethyl 4-amino-6-carbethoxypyrazolo(3,4-d)pyrimidine(XXXIV) in almost quantitative yield. The infrared spectrum of XXXIV indicated the loss of the conjugated nitrile group and the retention of the unconjugated ni-
present in the spectrum of I were missing in the ethoxymethylene derivative, and the positions of the nitrile bands were unchanged, thus conclusively eliminating X L from consideration. The action of ethanolic ammonia on XLI a t room temperature then led to the separation in almost quantitative yield of l-phenyl-3-cyanomethyl-4-aminopyrazolo(3,4-d)pyrimidine (XLII). As predicted, the ultraviolet absorption spectrum of XLII was almost superimposable with the spectrum of XXXV. Compound XLII was also formed by treatment of XLI with guanidine, apparently by elimination of cyanamide from the intermediate XLIII. Treatment of XLI with hydrazine and n-butylamine yielded l-phenyl-3-cyanomethyl-4-imino-5-amino4,5-dihydropyrazolo(3,4-d)pyrimidine(XLIV) and 1-phenyl - 3 - cyanomethyl- 4- imino-5-(n-butyl)-4,5dihydropyrazolo (3,4-d)pyrimidine (XLV), respectively.
Nkh
XXXII. R = - P h XXXVII, R = - C H j
I, R = - P h XVIII, R =-CHS
H-XH
HzN-C
CH,CN
N EtOCL$kN,N 0
EtOC-C=NA
R
HCO”,
R XXXIII, R = -Ph XXXI-111, R = - C H j
>
By a similar sequence of reactions, l-methyl-3cyanomethyl-4-cyano-5-aminopyrazole (XVIII) was converted into the pyrazolo(3,4-d)pyrimidines XLVII, XLVIII and XLIX. The conversion of aromatic o-aminonitriles into systems containing a fused 4-amino- or 4-iminopyrimidine ring via an intermediate ethoxymethyleneamino derivative (such as XLI and XLVI) is somewhat similar to the method recently exploited so successfully by ShawZ2for the synthesis of pyrimidine nucleosides, and provides an attractive potential route from the appropriate o-aminonitriles to glycosides of systems containing a fused pyrimi-
