March 20, 1964
COMMUNICATIONS TO T H E EDITOR
1253
tained by mild acid hydrolysis of I which gave theo2,3-dihydropyran and 2 , 3 - d i h y d r o f ~ r a n . ~However, ,~ phylline and a carbohydrate residue identified as 4,6earlier attempts to employ the use of various glycals di-0-acetyl-2,3-didehydro-2,3-dideoxy-~ - erythro - hexose under similar reaction conditions as for 2,3-dihydro(di-0-acetylpseudo-D-glucal) by comparison with an pyran2 involving the use of ethyl acetate as a solvent authentic sample’ as judged by paper chromatography did not result in nucleoside formation. in two different solvent systems. N o other products T h e recent fusion procedures5 successfully applied were detected in the hydrolysate. Deacetylation of I to nucleoside syntheses suggested the possibility of the (1.0 g.) with methanolic ammonia gave 0.52 g. of 7direct reaction of the requisite purine and glycal in the (2’,3’-didehydro-2’,3’- dideoxy- D - glucopyranosy1)theoabsence of a solvent, since it appeared quite possible ,-1nul. Calcd. for C13phylline (11), m.p. 197-198’. t h a t a similar resonance-stabilized carbonium ion a t C-1 might be involved in the alkylation process. This reaction now has been shown to occur with G-chloroX suspension purine and 3,4-di-O-acetyl-~-arabinal. of 6.10 g. of G-chloropurine6 &nd 8.80 g. of 3,4-di-0a c e t y l - ~ - a r a b i n a lwas ~ heated to 120’ in the presence of 50 mg. of p-toluenesulfonic acid in a manner similar to t h a t described6 for the fusion of 6-chloropurine and 1,3,3-t r i - 0 - acetyl-2-deoxy-6-D-ribopyranose. After a similar isolation procedure,5 1.02 g. of crystalline product was isolated, m.p. 175-l&’,. Examination revealed t h a t this material was an anomeric mixture Ro@ RO H H and -/3-~of G-chloro-9-(3’,4’-di-O-acety1-2’-deoxy-aribopyranosy1)purine. Fractional crystallization from 0 absolute ethanol gave 0.9 g. of 6-chloro-9-(3’,4’-di-OI, R=C-CHs II acetyl-2‘-deoxy-a-D-ribopyranosyl)purine, m.p. 2052 0 i 0 , [ a j Z 6+22.5’ ~ (c 0.75, acetone). .-1naZ. Calcd. 11, R = H for CI4Hl5ClN4O5:C, 47.4; H , 4.23; N , 15.8. Found: C, 47.3; H, 4.26; S , 15.9. The ultraviolet, infrared, HleNj06: C , 50.6; H , 5 . 2 ; N , 18.2. Fouzd: C , 50.3; and proton magnetic resonance spectra and paper H , 5.6; N, 17.9. Such unsaturated nucleosides should chromatography established the fact that this comprove most interesting synthetic intermediates for pound was identical with that anomer prepared5 by further work. Additional current interest in 2’,3’fusion of G-chloropurine and 1,3,4-tri-O-acety1-2-deoxy- unsaturated nucleosides stems from the fact that such p-D-ribopyranose. The presence of A-chloro-9-(3’,4’compounds have been postulated as possible biodi-0-acetyl-2’-deoxy-~-~-ribopyranosyl)pur~ne~ in the chemical intermediates in the enzymatic conversion of ethanolic filtrates was also confirmed by paper chrovarious purine and pyrimidine ribonucleotides to the matography in two systems. corresponding deoxyribonucleotides. The appliIn an effort to study the scope of this reaction, 3.1,6- cation of this procedure for the preparation of unusual tri-0-acetyl-D-gluca18(7.5 g.) and theophylline ( 5 . 0 g.) nucleosides aza the use of additional glycals and the were similarly fused in vucuo in the presence of 50 mg. detailed study of the structure of resulting nucleoside of p-toluenesulfonic acid. During this process, howderivatives are problems under active investigation in ever, a copious evolution of acetic acid was noted. our laboratory. From the reaction mixture was isolated, after recrystal(11) E Fischer, Chrm B i v . 47, 196 11914) lization from ethanol, 5.0g. of a crystalline nucleoside (12) P Reichard, J B i d C h e m , 2 3 7 , :3>1:3 (1562) (presumably an anomeric mixture, I ) , m.p. 102-104’, (13) A. Larsson, ibid , 238, 3 1 1 1 (1963) [a]25+ ~ l e g o (c 1.0, ethanol). The ultraviolet abDEPARTMENT OF CHEMISTRY WILLIAMA . BOWLES sorption spectra, 273 mp ( E 7800) and’:!A: 231. ARIZONASTATE L~XIVERSITY ROLAND E;. ROBINS 273 mk ( 6 4300, 9000), indicated T - s u b s t i t u t i ~ n . ~ TEMPE, ARIZONA =1nul. Calcd. for C17H20X407: C, 52.0; H, 5.10; N , RECEIVED DECEMBER 13, 1963 14.3. Found: C , 51.7; H , 5.24; N , 14.4. Proton magnetic resonance spectra in deuteriochloroform (TMS internal standard) showed the presence of The Specific Chemical Synthesis of y-P3* Labeled only two acetylmethyl groups a t 6 2.02 and 2.15, respectively. The two vinyl protons occur in the Adenosine 5‘-Triphosphate region 6 6.1-6.3. Sir : On this basis and on the analogy of a similar acidcatalyzed reaction of 3,4,6-tri-O-acetyl-u-glucal and Nucleoside j’-triphosphates labeled specifically with p-nitrophenol as the aglycone,I0 the structure of I P3zin the p- or y-position constitute a valuable tool is tentatively assigned as 7- (4’, 6 ’-di-O-acetyl-2I,3’for the elucidation of the mechanism of many biological didehydro-2’,3’ - dideoxy - D - glucopyranosy1)theophylreactions. Successful syntheses of these compounds line. Additional evidence for this structure was obhave invariably been enzymatic in nature’ and frequently are restricted to adenosine polyphosphates (3) I, R I.ewis, F H Schneider, and R K Robins, J Ovg Chem., 26, by enzyme specificities. As yet, chemical attempts 3837 (1961) (4) W. A Bowler, F. H Schneider, I,. R Lewis, and R K . Robins, J have led only to products with nonspecific labeling.2 . 3 M e d . Chem , 6 , 471 (1563) We now describe an entirely specific chemical synthesis (.5j M J . Robins, W . A Bowles, and R K Robins, J A m . Chem. S o c , of ATP-y-P32which may be extended to any nucleo86, 12.51 (1564). and references cited therein side j’-triphosphate and to 6-P32nucleoside 5’-tetra(6) A Bendich, P . J . Russel, J r , and J . J Fox, i b i d . , 76, 6073 (19.54); see also A. G . Beaman and R K . Robins, J . A p p l Chem., 12, 432 phosphates.
’
(1962) (7) I, Vargha and J . Kuszmann, Chem Bel, , 96, 411 (1963) (8) B Helferich, E h‘ Mulcahy, and H . Ziegler, ibid., 87, 233 (1951) (9) J M. Gulland, E. R. Holiday, and T F. Macrae, J . Chem Soc., 1635 (1934) (10) R . J. Ferrier, W. G. Overend, and A E . R y a n , i b i d . , 3667 (1962)
(1) See, e.g.. A. Kornberg. S. G . Kornberg, and E S. Simms, Riochim. B i o p h y s . A r l n , 2 0 , 21.5 (15,56j, G. PfleidPrer, ibid., 47, 389 (1561) (2) J. M. Lowenstein. Btochem. J , 66, 157 ( 1 5 j 7 ) . (3) A mixed chemical and enzymatic method has been used by R . T a n a k a [ J . Biochem. (Tokyo), 47, 207 (1560)l f o r t h e synthesis of ATP-B-PS2.
COMMUNICATIONS TO THE EDITOR
1254
A solution of dicyclohexylcarbodiimide (5 mmoles) in t-butyl alcohol was added dropwise over 9 hr. to a refluxing solution of the morpholine salt of A D P (1.0 mmole) and morpholine (2.4 mmoles) in 50% aqueous t-butyl alcohol. After a further 12 hr. under reflux, the mixture was evaporated and partitioned between water and ether. T h e water phase was separated by ion-exchange chromatography on Dowex-2 (HCO,-) using a gradient of triethylammonium bicarbonate giving adenosine 5’-phosph~romorpholidate~ ( 1 9%) and PI-aden >sine-5’ P2-(4-morpho1ino)pyrophosphate (I, 81%). )hromatographically pure I was isolated as its calcium salt. Anal. Calcd. for CI4H~oN6Ol0PzCa~2H20: c, 29.48; H , 4.24; N , 14.73; P, 10.86 Found: C, 30.36; H, 3.62; N , 14.50; P, 10.68. As expected, the P-N bond formed involved exclusively the 6-phosphorus of the ADP. This follows from the known reactivity of phosphoric acid monoesters, and resistance of phosphoric acid diester^,^ toward formation of phosphoramidates and could be confirmed by enzymatic means. In a similar reaction in which morpholine was replaced by ammonium hydroxide, the interesting analog PI-adenosine-5’ P2aminopyrophosphate (11) was prepared from A D P in 45% yield. Direct extension of the reaction to the synthesis of a terminally activated A T P was also achieved. Thus the reaction of ATP, morpholine, and dicyclohexylcarbodiimide followed by ion-exchange chromatography gave P1-adenosine-5’ P3(4-morpho1ino)triphosphate (111) in 72% yield. Anal. Calcd. for C14H20N6013P3Cal.j.4H20: 23.84; H , 4.00; N, 11.94; total P : labile P’:adenosine, 3.00: 2.CIO: 1.00. Found: C, 24.14; H, 3.97; N , 11.06; total P:labile P:adenosine, 2.96: 1.94: 1.00,
c,
n O
?
1
I
0 -
VOl. 86
which were separated on a micro-DEAE cellulose (HC03-) column. While the starting material and the unreacted A T P had specific activities of 75,400 c.p.m. per optical density unit a t 259 mp, the A L I P contained no isotope and the A D P fractions only 740 c.p.m. per optical density unit. This minor unidentified activity was chromatographically shown not to be due to A D P itself. ,4t any rate, greater than 9Oy0of the P32was located in the y-position. The apparent generality, specificity, and good yield of these reactions makes attractive the synthesis of other labeled and unlabeled compounds which will be described shortly. (6) Financial support fur this work fiom t h e “Stiftung fur Stipendien auf dem Gebiete der Chemie” (Switzerland) is gratefully acknowledged.
COSTRIBUTION No. 15 D. L. M . VERHEYDEN SYNTEX INSTITUTE FOR MOLECULAR BIOLOGY W. E. W E H R L I ~ PALOALTO, CALIFORNA J . G . MOFFATT RECEIVEDI ~ E C E M B E 19, R 1963
The Dismutation of Nucleoside 5’-Polyphosphates
Sir: Several years ago i t was reported that the 4-morpholine-N,N ‘-dicyclohexylcarboxamidine salt of adenosine 5‘-triphosphate (ATP) was unstable in anhydrous pyridine at room temperature, and a mixture of the 5’-mono-, di-, tri-, and tetraphosphates was present within a few days.’ A related degradation of the initially formed nucleoside 5’-triphosphates to the diand monophosphates during syntheses in pyridine also has been reported by several workers2 We have now examined this reaction in detail and are prepared to offer a tentative mechanism. Ion-exchange chromatography on DEAE-cellulose (HC03-) of the products resulting from storage of a rigorously anhydrous solution of tetrakis(tributy1ammonium) adenosine 5’-triphosphate (1 mmole) in pyridine (10 ml.) for 3.5 days gave a complex pattern of eight well-resolved peaks. The main products were a homologous series of nucleoside 5’-polyphosphates (I, n = 0, 1, 2, 3, 4, 5, 6) containing from one to seven
R
N-fl-O-P-O-POCH2
‘I1
m
OH
OH
The synthesis of ATP-y-P3* from I utilized the known reactivity of phosphoromorpholidates in pyrophosphate s y n t h e ~ e s . ~Thus, .~ the reaction of the 4-morpholine N,N’-dicyclohexylcarboxamidine salt of I ( 0 . 1 mmole) with tributylammonium orthophosphate (0.3 mmole containing 2 mc. of P3*)in rigorously anhydrous dimethyl sulfoxide (2 ml.) for 45 hr. at 35’ gave a mixture of products which were separated by ion-exchange chromatography on D E A E cellulose (HCOs-). Minor amounts of unreacted I, AMP, ADP, and API were cleanly separated from ATPy-P3*(M%) which was isolated as the chromatographically homogeneous sodium salt with a specific activity of 3 pc. per pmole. Controlled partial degradation of this product (1 pmole) with E . coli alkaline phosphatase (40 min. a t 35’ with 20 pl. of dialyzed enzyme) gave a mixture of adenosine, AMP, ADP, and unreacted ATP, ( 4 ) J. G hloffatt and H. G . Khorana, J . A m C h e m . S n c . , 83, 649 (1961) ( 5 ) S . Roseman, J . J 1)intler. J . G. Moffatt. and H G Khorana. i b r d . . 8 8 , 05;l ( 1 R f i l ) : J. G . Moffatt and H. G. Khorana. i b i d . , 8 8 . 663 (1961).
I
I1
phosphate groups? Further chromatography of many of these peaks showed the presence of a second series of compounds, in much smaller amounts (5-10% of the total peak), which has been characterized by analytical and enzymatic means as a,w-di(adenosine5 ’ ) polyphosphates (11, n = 1, 2, 3, 4, 5 , 6 ) . These latter compounds are completely resistant to the action of E . coli alkaline phosphatase but are rapidly split by purified venom phosphodiesterase to A M P and an adenosine 5 ’-polyphosphate which then is cleaved more slowly to A M P and an inorganic polyphosphate. A homologous series of inorganic polyphosphates also is (1) J. G. Moffatt and H. G. Khorana, J . A m . Chem. Soc., 88, 649 (1961). ( 2 ) (a) M . Ikehara and E Ohtsuka, C h e m . P h a r m Bull (Tokyo), 11, 43.5 (1963), (b) J . Zemlicka, J. S m r t , and F Sorm, Colleclion Czech. C h e m . C o m m u n , 2 8 , 241 (1963) (3) T h e following abbreviations are used: A M P , A D P , A T P , API, APa. AP6. a n d APT for t h e adenosine .5’-mono-, d i - , t r i - , t e t r a - , penta-, hexa-, and heptaphosphates; AP& APIA, A P I A , APsA, APsA, and APTA for the appropriate a,w~diadenosine5’-polyphosphates; I p , APsA is PL,Ps-di(adenosine~5‘)pentaphosphate.
