6202
COMMUNICATIONS TO THE EDITOR
Vol. s 2
Acknowledgment.-The author is most grateful to the National Science Foundation for its support of this work under Grant NSF-G11274. He is also indebted to Dr. Wallace S. Brey, Jr., for obtaining and aid in interpreting the n.m.r. spectra.
285-286'). The indentification was made by analysis (Calcd. for C5H6C1N6-0.5H20: C, 33.25; H, 3.91; N, 38.79. Found: C, 33.38; H, 4.01; N, 38.62) and ultraviolet spectrum.9 A picrate (m.p. 190-192') formed from the hydrochloride failed to depress the melting point of authentic DEPARTMENT OF CHEXISTRY UNIVERSITY OF FLORIDA adenine picrate.8 W. M. JOKES GAINESVILLE, FLA. Reaction of triacanthine with benzylamine and RECEIVED AUGUST30, 1960 benzylamine hydrochloride in a sealed tube a t 180' for 12 hours yielded the "exchange amination"lO product N-benzyltriacanthine (111), m.p. 150" THE STRUCTURE OF TRIACANTHINE' (Calcd. for C17HlSN6:C, 69.60; H, 6.53; N, 23.88; Sir: mol. wt., 293. Found: C, 69.34; H, 6.52; N , Early reports suggested the presence of alka- 24.06; electrometric titration mol. wt.,6 310 loidal material in species of G l e d i t s i ~ ,of~ .the ~ Legu- lo), " : : :A (neutral) 292 mp ( E 17,100) and A:", minosae family, and recently the alkaloid "tri- (pH 1) 287 (24,300). The formation of this acanthine" has been isolated from Gleditsia tria- product demonstrated that the amino group of canthos L. and the formula C8HloN4 a s ~ i g n e d . ~adenine is unsubstituted in triacanthine. An We have repeated the described extraction pro- n.m.r. spectrum (CDCl,) of N-benzyltriacanthine cedure on the young leaves of G. triacanthos col- showed peaks a t = 2.05 and 2.23," which must lected by one of us (J.A.D.) in April a t Dixon derive from the ring hydrogens of adenine.12 Springs, Illinois,6 and have established that the Thus, triacanthine must be represented by adenine correct molecular formula for triacanthine is substituted only a t position 7 or 9. The choice C10H13N6 and that the structure is 6-amino-7- between these two alternatives is made readily in (y,y-dimethylallyl)-purine (I). favor of position 7 by studying representative 7and 9-substituted 6-aminopurines. For exRNH R' ample, adenosine (9-p-D-ribofuranosidoadenine) I, R = H, R' CHzCH=C(CH,)z 11, R = H, R' = H shows X,:"."," (pH 13) 260 mp ( E 14,300) and A%."," 111, R = CHsCeHs, R' = CHzCH=C(CHa)z (pH 1) 260 (14,200) while 7-~-ribofuranosidoadenIV, R = H, R' = CHzCHzCH(CH3)z ine shows (pH 13) 271 (9,800) and A:","," (OH 1) 273 (13,600).13 The n.m.r. spectrum of Analytical data for triacanthine, m.p. 228-229' N-benzyltriacanthine allowed a n accurate determi(reported4 227-228') (Calcd. for C10H13NB:C, nation of the number of methyl groups in triacan59.09; H, 6.45; N , 34.46. Found: C, 59.29; thine. By using the area of the 2.77 T peak repreH, 6.32; N, 34.59), and triacanthine hydrochloride, senting the five phenyl hydrogens as an internal m.p. 218-220" dec. (reported4 218-219') (Calcd. standard, in comparison to the area of the singlet for Cl0Hl4ClN5: C, 50.09; H, 5.89; N, 29.22. CH3-C peak a t 8.14 7 (after extrapolating to zero Found: C, 50.24; H, 6.04; N , 28.96), as well as rf. power),14it was established that there were two the molecular weight determined by electrometric C-CHg groups present in triacanthine (a Kuhntitration6(Calcd. for C10H13N5:203. Found: 211 f Roth C-methyl determination on triacanthine gave 10; pK', 5.4) supported the formula C10H13N5. Tri- a value of 5.73 (calcd. for two C-CH, groups, acanthine showed 1%;" (neutral) 273 mp ( E 12,500) 14.80)). I n addition, the r value for the C-CH3 and (pH 1) 277 (18,300). The mass spectrum peak indicated that the methyl groups are atof triacanthine? showed peaks a t mass 203 (confirm- tached to an olefinic carbon atom.15 ing the molecular weight), 188 (establishing the presCatalytic reduction of triacanthine with hydroence of a t least one methyl group), and 135. Chemi- gen and platinum in glacial acetic acid yielded dical realization of the cleavage suggested by the 135 hydrotriacanthine (IV), m.p. 230-231' (strongly mass peak was achieved by treatment of triacan- depressed when mixed with triacanthine) (Calcd. thine with concentrated hydrochloric acid a t 80" for CI0Hl5N5: C, 58.51; H, 7.37. Found: C, for 8 hours, which produced crystalline adenine 58.55; H, 7.17.), (neutral) 273 (13,020) (11) hydrochloride, m.p. 283-284' (reporteds and X:"j (pH 1) 277 (18,950). I n addition to dihydrotriacanthine, a considerable amount of (1) This investigation was supported by a research grant (USPHSRG5829) from the National Institutes of Health, U. S. Public Health adenine was produced by hydrogenolysis. l6 The g813
Service. (2) C. Wehmer, "Die Pflanzenstoffe," Verlag von Gustav Fischer, Jena, 1929, p. 508. (3) M. Greshoff, "Mededeelingen uit ' s Lands Plantentuin," 29, 67 (1900), G. Kolff and Co., Batavia, 1900. (4) A. S. Belikov, A. I. Bankowsky and 31. V. Tsarev, Z k u r . Obskchei K k i m . , 24, 919 (1954). (5) We wish t o record our special thanks t o Professors W.R. Boggess and A. R. Gilmore of the Department of Forestry, University of Illinois, for helping us t o obtain the leaf material. (6) We are grateful t o Dr. Harold Boaz of Eli Lilly and Company, Indianapolis, Indiana, for this information. (7) Determined by Dr. Klaus Biemann of the hfassachusetts Institute of Technology, t o whom we are pleased t o acknowledge our thanks. (8) H. R. Bentley, K . G. Cunningham and P. S. Spring, J. Chem.
SOC.,
a a u (1~51).
(9) J. M . Gulland and E. R. Holiday, ibid., 765 (1936). (10) C. W. Whitehead and J. J. Traverso, THISJOURNAL, 82, 3971 (1960). We are grateful t o these authors for providing an advance copy of their article. (11) G . V. D. Tiers, J . P k y s . Chem., 62, 1151 (1958). (12) C. D. Jardetzky and 0. Jardetzky, THISJ O U R N A L , 82, 222 (1960). (13) W. Friedrich and K. Bernhauer, Ckem. Bey., 89, 2507 (1956). (14) Cf.,J. Pople, W. Schneider and H . Bernstein, "High-resolution Nuclear Magnetic Resonance," McGraw-Hill Book C o . , New York, N . Y . , 1959, p. 77. (15) G. V. D. Tiers "Table of 7 Values for a Variety of Organic Compounds," Minnesota Mining and Manufacturing Company, St. Paul, Minn., 1958. (16) Triacanthine is not cleaved b y acetie acid under the room temperature conditions ob the hydroyeuation.
Dec. 5, 1960
COMMUNICATIONS TO THE EDITOR
6203
ease with which triacanthine is cleaved by acid in such a way that i t fulfilled the dual roles of reand by hydrogenolysis suggests that the double actant and container, with results identical with bond is allylic to the 7-nitrogen atom.” Dihy- those obtained by the alternative procedure. All drotriacanthine is stable to the conditions under cutting and transfer operations involving the lithwhich triacanthine is cleaved by hydrochloric acid. ium were made in a protective bath of pentane The similarity between the ultraviolet spectra of which ultimately was pumped off, along with the triacanthine and dihydrotriacanthine shows that air in the system. Just prior t o the ignition the double bond in triacanthine is not part of the step, several flushings with argon also were made. In every case the surface of the lithium remained chromophore. Finally, n.m.r. spectra of triacanthine and di- free of appreciable contamination throughout the hydrotriacanthine (30% in liquid sulfur dioxide), entire cycle of operations. The reactions normally were carried out on a in addition to the previously mentioned data, permit unambiguous assignment of structure I to 50-100 mg. scale with the lithium present in about triacanthine. Triacanthine shows a single unre- a tenfold mole-excess. During the few minutes solved peak in the CH3-C region (7 = 8.13), required to encompass the 20&500° temperature from (CH,)zC=, and, in addition, a triplet (7 = range, some physical attack on the platinum strip 4.49) and a doublet (7 = 5.01)-a splitting pattern occurred. This particular phase of the process, which is indicative of the =CH-CHzmoiety. which we do not believe involves a chemical reDihydrotriacanthine exhibits a doublet (7 = 9.05) action, is described more fully below. Somewhat which came from the methyl groups in (CH&CH-. above 500’ a more spectacular phenomenon obFuture publication will deal with the synthesis, tains. At some point, usually on the perimeter of the further chemistry of triacanthine, and the re- the metal-liquid metal interface, we observe consults of tests for biological activity. siderable agitation accompanied by a moderately (17) Cf. F. M. Dean in “Fortschritte der Chemie Organischer intense, reddish glow, These effects spread rapidly Naturstoffe,” ed. by L.Zechmeister, 9 , 269 (1952),for the behavior of over the entire contact area, and in less than a a I,?-dimethylallyl ether toward acid and catalytic hydrogenation. second the reaction is completed. During this THENOYESCHEMICAL LABORATORY short time some lithium distills (or perhaps spatUNIVERSITY OF ILLINOIS NELSONJ. LEONARDters) rather uniformly onto the walls of the glass URBANA, ILLINOIS JAMES A. DEYRUP envelope, a traverse distance of about a centiRECEIVED OCTOBER 14, 1960 meter. It is possible to remove the burner a t the instant of ignition, and the effects we describe INTERMETALLIC COMPOUNDS OF ALKALI METALS will proceed in an entirely self-sustaining fashion. WITH PLATINUM, A NOVEL PREPARATION OF A Although we have not, as yet, examined this COLLOIDAL PLATINUM HYDROGENATION system calorimetrically, some relatively crude measCATALYST urements indicate that the surface temperature of sir: the crucible immediately below the reacting maWe have found that in vacuo or under an at- terials increases by several hundred degrees. mosphere of argon, molten lithium and metallic It is unlikely that contaminant amounts of platinum undergo a violently exothermic reaction lithium oxide or lithium hydroxide are responsible having an ignition temperature of about 540’. for the phenomenon, for we have used a burner X-Ray diffraction powder patterns of the resulting to heat both of these compounds to bright-red brittle, grayish product, taken after the excess incandescence in the presence of platinum, with lithium was removed by treatment with water or no effects which in any way resemble those ob3 M hydrochloric acid, showed broad lines whose served with the pure metals. intensities and positions were in good agreement By heating the raw materials to about 650’ in an with calculations based on an MgCuz cubic ‘‘Faves argon atmosphere, we have also obtained, in poor phase” lattice having a0 = 7.60 i 0.05 A. A yield, a sodium-platinum compound which is isoflame photometric determination of the water- structural with LiPtz, and hence probably is NaPtz. inactive lithium content of the reaction product This latter compound has a lattice parameter of (dissolved in aqua regia) showed over 85% of the 7.48 0.02 A. Our results complement those of theoretical amount required for LiPtz proportions. Compton and collaborator^,^^^ who very recently The combined lithium and platinum analyses ac- identified a number of cubic Laves phase comcount for over 99% of the product by weight. pounds containing noble metals in combination Because of the great disparity in the atomic weights, with both alkaline earth and rare earth elements. however, the lithium analysis is clearly the one We also have found that molten lithium, a t which is most important as regards the identifi- temperatures well below that a t which appreciable cation of the species formed. reaction occurs, has a remarkable ability to peneA few experimental details are given. Most of trate the platinum lattice. A similar type of atthe reactions were carried out in a glass system tack, occurring near the melting point of lithium, with the reactants contained in molybdenum cruci- has been noted previously by Grosse.4 We have bles fitted with thermocouples. Molybdenum has observed that, overnight, lithium will climb several been shown previously’ to be inert to molten alkali centimeters up a platinum strip whose bottom is in metals. Nevertheless, several experiments were contact with a lithium pool contained in a molybperformed in which the platinum strip was bent denum crucible a t 400”. Apparently a diffusive
*
(1) E.E.Hoffman and W. D. Manly in “Handling and Uses of the Alkali Metals,” Advances in Chemistry Series No. 19, American Chemical Society, Washington ,D. C., 1957. p. 81.
(2) E.A. Wood and V. B. Compton, Act5 Crysf., 11, 429 (1958). (3) V. B.Compton and B. T. Matthias, i b i d . , 12, 651 (1950). (4) A. V. Grosse, 2. Nolurfwsch., Sb, 533 (1953).
