3'-AMINO-3 '-DEOXYADENOSINE
Nov. 20, 1955 [CONTRIBUTION FROM
THE
5911
MEDICINALCHEMICAL RESEARCHSECTION, LEDERLELABORATORIES, RESEARCHDIVISION, AMERICAN CYANAMID COMPANY]
Puromycin. Synthetic Studies. XV.
3 ' - h i n 0 - 3 '-deoxyadenosine
BY B. R.BAKER,ROBERTE. SCHAUB AND HENRYM.
KISSMAN
RECEIVED J U N E 9, 1955 The title conipouiid 1-1,the S-demethylated analog of the biologically active aminonucleojide I from puromycin, has been synthesized by condensation of chloromercuri-6-benzamidopurine with 1-0-acetyl-2,5-di-0-benzoy1-3-acetamido-3deoxy-D-ribofuranose (111), in the presence of titanium tetrachloride, followed by deacylation. Both a - and B-nucleosides were obtained, the @nucleoside IV being anomerized t o the or-nucleoside V by the titanium tetrachloride. Condensation of chloromercuri-6-benzamidopurine with 2,5-di-0-benzoyl-3-phthalimido-3-deoxy-~-~-ribofuranosyl chloride followed by deacylation gave a much higher over-all yield of the title compound VI since no anomerization to a-nucleoside took place. I t was shown that "basic" nucleosides could be separated from "neutral" nucleosides by absorption on a carboxylic acid ionexchange resin.
The active moiety of the puromycin molecule against Trypanosoma equiperdum and the transplanted mammary adenocarcinoma of the C3H mouse (J-tumor) has been shown t o be the aminonucleoside, 6-dimethylamino-9-(3'-amino-3'-deoxyp - D - ribofuranosyl) - purine (I).l The activity
culture, VI is about twenty times as active as the aminonucleoside I, but neither I nor VI show a selective in vitro action against tumor cells over normal celk.6 I n the CsH mouse, 3'-amino-3'-deoxyadenosine (VI) is much more toxic than the aminonucleoside I agreeing with the tissue culture data. When the activity of V I against the J-tumor YMe, in the C3Hmouse was tested a t the maximum tolerI ated dose (1/50 of I), the growth of the tumors was reduced to 487, of the tumor controls, compared to a 6% growth of tumors when the mice were treat5d with aminonucleoside I.' This combination of zn vivo and in vitro data shows t h a t the selective in vivo action of I against the J-tumor3 cannot be rationalized by enzymatic demethylation of I to 3'-amino3'-deoxyadenosine (VI), otherwise I should have shown a selective action against tumor cells in NH, b H vitro. I t is probable that the aminonucleoside I is I metabolized to some unknown active compouncl against Trypanosoma equiperdum can be reversed in which has a selective action on tumor cells. Furvivo by a number of purines, including adenine and, ther synthesis and testing of suitable additional strikingly, even the purine moiety of puromycin, analogs of I may lead to this unknown active metanamely, 6-dimethylaminopurine. The activity bolic product. Synthesis.-Condensation of l-O-acetyl-2,Z-diagainst the J-tumor in vivo is not reversed by these (111 p ~ r i n e sindicating ,~ that in these two test systems O-benzoyl-3-acetamido-3-deoxy-~-ribofuranose the mechanisms of action are not necessarily re- anomeric mixture)8 with chloromercuri-6-benlated. Microbiologically, i t has been observed zamidopurineg in ethylene dichloride in the presence t h a t large amounts of 6-dimethylaminopurine can of titanium tetrachloride'O gave a 57% yield of replace the adenine necessary for growth of E. crude nucleoside IV. During debenzoylation of IV in methanolic soc o l i 4 T h e fact t h a t 6-diethylaminopurine can replace adenine in these two biological systems dium methoxide a pure, highly insoluble, crystalline can be rationalized by enzymatic demethylation of acetamido nucleoside, m.p. 271" dec., [ a ] D +64", the 6-dimethylaminopurine to adenine. If such is separated from the boiling methanol solution in the case, then it is equally possible that the amino- 10% yield (from 111). Concentration of the mother nucleoside I from puromycin is also enzymatically liquor gave 33% of another crystalline acetamino demethylated to 3'-amino-3'-deoxyadenosine (VI). nucleoside which, when purified, had a 1n.p. 247' The synthesis of the latter compound is the subject dec. and [ a ] +lo". ~ Both compounds had the of this paper. combustion analyses, ultraviolet and infrared specBiological Activity.-3 '-Amino-3 '-deoxyadenosine tra expected for 3'-ace tamino-3 '-deoxyadenosine. (VI) has only '/z the activity of the aminonucleoside Thus the higher melting isomer, by reason of its I against Trypanosoma e p ~ i p e r d u mthus , ~ indicating t h a t enzymatic demethylation is not a factor in higher positive rotation, appeared to be the a-nu(6) Private communication from Dr. P. A . Eichorn of these labornthe activity. Against the J-tumor grown in tissue tories.
(1) B. R . Baker, J. P. Joseph and J. H . Williams, THIS JOURNAL, (7) Private communication from Dr. J. J . Oleson of these labora77, 1 (19551,paper VI1 of this series. tories. (2) R. I. Hewitt, A. R. Gumble, W. S. Wallace and J. H. Williams, ( 8 ) B . R . Baker, R. E. Schaub, J. P . Joseph and J. H . Williams, TRIS Anfibiolics and Chcmotheyapy, 4, 1222 (1954). JOURNAL, 77, 12 (1955), paper IX of this series. (3) J. J. Oleson, P . L. Bennett, S. L. Halliday and J. H . Williams, 6th (9) J. Davoll and B . A. Lowy, i b i d . , 73, 1650 (1951). International Cancer Congress, Sao Paulo, Brazil, July, 1954. (lo) This procedure was found superior to the methodS of pre(4) G. B. Elion, G . H . Hitchings and H. Vanderwerff, J. Biol. Chcm., formation of the chloro sugar-titanium chloride complex for the syn198, 505 (1951). ( 5 ) Private communication from Dr. R . I. Hewitt of these labora-
tories.
thesis of 6-dimethylamino-9-(3'-ncetamino-3'-deoxy-cr-~-arabinofuranosy1)-purine; d.B R . Baker and R . E . Schaub, THIS JOURNAL, 77, 2396 (1955), paper XI1 of this series.
5912
E. K.BAKER,K. E. SCHAUB AKD H. 11.KIssi\r,is
Vol. '77
absorbed from neutral solution on a carboxylic acid type ion-exchange resin. However, any nucleoside bearing the much stronger aliphatic amine in the sugar moiety mould be expected to be absorbed.'? Experimentally, it has now been shown that the carboxylic acid resin, -4mberlite IRC-50 (H-form) was not strong enough, as anticipated, to hold a "neutral" nucleoside such as 6-dimethylamino-9P-D-ribofuranosylpurine. However, the basic nucleoside, 6-dimethylaniino-9-(3'-a1nino-3'-deoxyp-D-ribofuranosy1)-purine (I), was completely absorbed from an aqueous solution and washing with large quantities of water failed to remove it. The basic nucleoside I then could readily be eluted from the resin with 2 LV ammonium hy~1roxide.l~In these experiments the concentration of nucleosides in the effluent fractions was conveniently follow~ed by ultraviolet inspection. These data showed that it should be possible to separate quantitatively a ' 13, 30 (0 ti)2 "neutral" nucleoside from a "basic" nucleoside. , L Y T h e aqueous solution of VI obtained from alkaline hydrolysis of VI1 was freed from barium hydroxide with carbon dioxide, then passed through a column of Amberlite IRC-50 ( H i , The colunin was then washed with 509; methanol until the effluent contained no appreciable purine observable a t 260 mp in the ultraviolet. The column was then washed with 2 LV ammonium hydroxide in 50% methanol until no more of the basic nucleoside mas Ix XI eluted. The desired 3'-arnino-3'-deoxyadenosinc cleoside VI11 and the lower melting isomer the ex(VI) crystallized directly from the richest fraction. pected p-nucleoside VII. T h a t VI1 was the de- A total of 2 f 1 7 of ~ crystalline VI could be isolated by sired 0-D-ribofuranosyl nucleoside VI1 was une- this method. Although VI is not appreciably water quivocally demonstrated by (a) rebenzoylation soluble, direct seeding of the hydrolysis solution back to IV and hydrolysis with the refluxing two still failed to give any crystals of VI. Spparently phase system 2 N hydrochloric acid-ethylene di- this compound readily supersaturated in water in chlorides to 2,5-di-O-benzoy1-3-acetamido-3-deoxythe presence of certain impurities, a property also D-ribose ( X ) ; (b) removal of the N-acetyl group characteristic of adenosine. T h a t VI had the P to V I and periodate oxidation to IX, discussed in furanose configuration was checked by periodate detail later. Xlthough VI11 could be rebenzoyl- oxidation in pH 4.5 buffer. The resultant solution ated back to I-,hydrolysis with the two-phase hydrolysis system did not give any X (or any other of dialdehyde IX had MD -66400°, in rcasonable agreement with the same dialdehyde obtained crystalline product). However, hydrolysis of T'III from adenosine14(MD - 5600" to - T7OO" dependwith 27, hydrochloric acid gave authentic 3-aminoing on the p H j and in marked contrast to the di3-amino-3-deoxy-~-ribose," showing that in \711 aldehyde of oxonfiguration, f1FD +4000°,obtained no rearrangement of the sugar configuration had by periodate oxidation of 9-a-D-arabinofuranosyltaken place. It is reasonable to assume that the adenine.15 ring size of the sugar in the nucleoside 1'111 was I n order to prove that no change in the basic still furanose since ring expansion in a nucleoside structure of VI1 had taken place during hydrolysis, condensation in the presence of titanium tetra(VI) was fully pure 3'-aniino-3'-deoxyadenosi11e chloride has not been observed previously.8s'0 acetylated with acetic anhydride and pyridine, Thus the higher melting acetamino nucleoside most then 0-deacetylated to give pure VII, 111.p. 247" probably, though not unequivocally, has structure dec., [ff]D +i 2". \TII. The formation of the a-nucleoside V as a byHydrolysis of the N-acetyl &nucleoside VI1 to VI with 0.5 ili barium hydroxide a t 100" appeared to product during nucleoside synthesis contradicts the (12) J . C. Winters and R. Kunin, I d . E n g . Chcm., 4 1 , 461 (19491, be complete in 1 hour since the amorphous product have used Amberlite IRC-50 for separation of basic amino acids from no longer had amide carbonyl absorption a t 5.90 p neutral amino acids. They have also recorded that t h e weekly basic (characteristic of T'II). However, the product purine, adenine, is not held by this resin. (13) Basic amino acids have been eluted from Amberlite IRC-50 could not be crystallized by conventional methods, with hydrochloric acid.12 Acid elution was not considered compatible but was crystallized after development of a new with the acid lability of t h e sugar-purine linkage. This difficulty technique for the purification of nucleosides bearing was avoided b y elution with ammonium hydroxide, a procedure also an amino group in the sugar moiety. used b y H . Tauber, Pvoc. SOC.Enp. B i d . M e d . , 86, 838 (1954). for Since the purine moiety of a nucleoside is a very elution of amino acids from the sulfonic acid type ion-exchange resin weak organic base, it would not be expected to be Amberlite Ili-105.
*
(14) J . Tlavoll, 11. 1,ythpir
( 1 1 ) I< I< li.ihtr 2nd I< I p i1)c1 I I I u f t h i i i l r , ,
i
\cliiuh
I Oii:
rhPiii
19 1141, (I'lil)
(I !l46),
:tuil
A. R
'l'tt~If., 833
Nov. 20, 1955
3 '-AMINo-~ '-DEOXYADENOSINE
5913
previously postulated ruleI6 that when a halo acyl- nucleoside gave V I in 60% yield. Thus it can be ated sugar is coupled with a heavy metal salt of a calculated that the crude synthetic 3'-phthalimidopurine, the resultant nucleoside will have a CI-CZ 3'-deoxyadenosine had a purity of about 68%. trans-configuration. The formation of the a- Slightly higher over-all yields of V I can be obtained nucleoside V with a C1-C2 &-configuration can be if the intermediate N-phthalyl derivative is not reasonably explained by assuming t h a t the &nu- purified. The over-all yield of V I from l-O-acetyl-2,5-dicleoside IV with a C1-C2 trans-configuration is is formed initially, then IV is anomerized by the tita- O-benzoyl-3-phthalimido-3-deoxy-~-ribofuranose nium chloride present to the a-nucleoside V in the almost five times that obtained from 1-0-acetyl-ribofuransame fashion that acylated methyl glycosides are 2,5-di-O-benzoyl-3-acetamido-3-deoxy-~ anomerically equilibrated by titanium tetrachlo- ose (111). With the N-phthalyl group, no formaride." Attempts to prove t h a t this mechanism was tion of a-nucleoside was noted, although a relatively possible by rebenzoylation of the pure a- or @-nu- small yield could have escaped isolation. cleoside back to the tribenzoate I V (or V), treatAcknowledgment.-The authors wish to thank ment with titanium tetrachloride in ethylene di- W. Fulmor and staff for rotations and spectrochloride, then debenzoylation back to IV and/or V photometric data and L. Brancone and staff for did not give conclusive results. microanalyses. It is interesting to note t h a t this anomerization Experimental has so far only been observed with nucleosides of and p h o m e r s of 3'-Acetamido-J'-deoxyadenosine 6-benzamidopurine; an example of the same (VI1 and VIII). (A).-A mixture of 8.5 g. of chloromercurianomerization with a different sugar as 6-benzami- 6-benzamidopurine (11),9 10 g. of Celite (diatomaceous dopurine has been found and will be published in a earth), 6.36 g. of I11 (anomeric mixture)8 and 570 cc. of future paper. Nucleosides from 2-methyl-mer- ethylene chloride was freed from traces of moisture by discapto-6-dimethylaminopurine with either the a- tillation of 50 cc. of solvent. Then a solution of 1.95 cc. titanium tetrachloride in 15 cc. of ethylene dichloride was or @-configurationdo not anomerize in the presence of added portionwise with stirring. After being refluxed and of titanium chloride.*~'O A number of other purines stirred for 19 hours, the mixture was cooled to room temhave been used in this method and also do not perature, treated with 275 cc. of 1 N hydrochloric acid and anomerize. I n spite of the fact that 6-benzamido- stirred for 1 hour. The mixture was filtered and the Celite washed with 100 cc. of hot chloroform. The Celite purine and others may be discovered is an excep- cake cake was stirred vigorously with 125 cc. of chloroform and tion to the rule of formation of c 1 - C ~trans-nucleo- 125 cc. of 1 N hydrochloric acid to break down the remainder sides, this rule has been valuable for predicting the of the titanium complex. The mixture was filtered and the route to stereochemically controlled syntheses of solids washed with two 100-cc. portions of boiling chloroform. The combined organic solutions and washings, nucleosides.10s18 This difficulty can be avoided by separated from water, were evaporated to dryness in vacuo. the use of other substituted purines which do not The residue was dissolved in 100 cc. of chloroform and anomerize and which by further transformation of yashed with 100 cc. of 30YGaqueous potassium iodide, then the nucleosides lead to adenine nucleosides (to be water. Dried with magnesium sulfate, the chloroform solution was evaporated to dryness in vacuo leaving 5.49 g. published). 230 mp (e (57%) of a crude mixture of IV and V, A&. Since the anomerization of the @-nucleoside IV 37,700), 278 mp ( E 14,100). to the &-nucleoside V by titanium chloride causes A solution of 5.48 g. of this benzoate in 75 cc. of hot methconsiderable loss of the desired @-nucleoside, a anol was clarified with Norit and filtered through Celite. synthesis was sought which would avoid the anom- After the addition of 2.7 cc. of 1 N methanolic sodium methoxide the solution was refluxed for 30 minutes, the pH reerization. One of the answers found was the use maining < 10 when spotted on moist indicator paper. After of the N-phthalyl blocking grouplg for the amino 15 minutes white crystals had begun to separate. The sugar moiety. Condensation of crystalline 2,5- solution was filtered hot and the crystals of a-anomer VI11 di-0-benzoyl-3-phthalimido-3-deoxy-@-~-ribofur- were washed with hot methanol; yield 440 mg. (9.9% from anosyl chloridelg with chloromercuri-6-benzamido- 1111, m.p. 271" dec., [ c Y ] ~ ~fD6 4 =k 1' (2% in 0.1 N HCI), 2.98 p (broad) (OH, NH), 259 m p ( e 14,800); purine followed by debenzoylation gave a 58% 5.88 p (C=O), 6.02, 6.13 p (C=N). yield of crude non-crystalline 3'-phthalimido-3'Anal. Calcd. for Ci2H1~NsO4:C, 46.8; H, 5.23; N, deoxyadenosine. Removal of the N-phthalyl group 27.2. Found: C, 46.8; H , 5.63; N, 26.8. with hydrazinelg and isolation of the product by This compound was analytically, optically and spectrothe ion-exchange column technique gave crystal- graphically (ultraviolet) pure. When twice recrystallized line 3'-amino-3'-deoxyadenosine (VI) in 41y0yield by extraction with methanol in a Soxhlet apparatus the or an over-all yield of 24% from the original chloro above properties did not change from the original in each recrystallization. The a-anomer then had [ a I z 5$63 ~ f sugar. Reaction of pure VI with phthalic anhy- 1' (2% in 0.1 N HCl). dride in boiling dimethylformamide gave crystalThe filtrate from the 440 mg. was concentrated to about line 3'-phthalimido-3'-deoxyadenosine which could '/3 in vacuo, chilled in a n ice-bath, the gelatinous crystals VI1 were collected and mashed with cold be recrystallized from water. Recrystallization of of crude p-anomer then with benzene to remove methyl benzoate; the crude synthetic 3'-phthalimido-3'-deoxyadeno- methanol, yield 1.49 g. (33%), m.p. 241' dec.,: : :A 259 m p ( e 12,200). sine from water afforded a 61% recovery of pure Thus, ultraviolet analysis showed a purity of 82% based on material. Dephthalation of this pure phthalimido the mol. wt. of VI1 (or VIII). It can also be estimated from (Y-
(16) B. R. Baker, J. P. Joseph, R. E. Schauh and J. H. Williams, J. OUK. Chcm., 19, 1786 (1954), paper V of this series. (17) E. Pacsu, THISJOURNAL, 62, 2563, 2568, 2571 (1930): E. Pie1 and C. B. Purves, ibid., 61, 2978 (1939). (18) B. R. Baker and R. E. Schauh, ibid., 77, 5900 (1955), paper XI11 of this series. (19) B. R. Baker, J. P. Joseph and R. E. Schaub, i b i d . , 77, 5903 (l!].?.?), paper X I V of this series.
recrystallization data that this crude \'I1 contains about 10% VIII. Recrystallization of 500 mg. from 6 cc. of water with the aid of Norit gave 310 mg. (62% recovery) of white needles of VIII, m.p. 247" dec., [ c Y ] ~ ~+DI 0 =k 2" (1.5% in 0.1 N HCI),: : :A 259 mp (E 14,700); 2.97 fi (broad) (OH, XH), 5.90 p (C=O), 6.03, 6.14 p (C=S), which were analytically, spectrographically (ultraviolet) and optically pure. The compound forms a hydrate which is obtained
anhydrous after drying a t 110' for 18 hours, but not 3 was removed (about 75 cc. distilled). After the addition hours, in high vacuum. of a warm solution of 2,5-di-O-benzoyl-3-phthalimido-3A n d . Calcd. for C12HleK's04:C, 46.8; H, 5.23; S, deoxy-p-D-ribofuranosyl chloride's in 50 cc. of xylene, the mixture was r d u x e d with stirring for 3 hours. The hot 27.2. Found: C, 46.8; H,5.04; S,27.3. Further recrystallizations from water did not change the solution was filtered and the filter cake was washed with 50 optical rotation. The P-anomer VI1 is readily separated cc. of hot toluene. The combined filtrate and washing was from a-anomer VI11 since the latter, when dissolved in hot evaporated t o dryness in vucuo. Meanwhile the filter cake water, does not crystallize out from solution after standing mas washed with four 25-cc. portions of boiling chloroform. for several days. The rotation of neither VI1 nor VI11 The residue from the xylene solution was dissolved in the changed when the solutions in 0.1 N hydrochloric acid were chloroform solution. Washed with 100 cc. of 30y0 aqueous potassium iodide, then water, the chloroform solution was allowed t o stand for 48 hours. dried with magnesium sulfate. Evaporation t o dryness in I n a pilot run the yield of combined VI1 and VI11 was 50:; aacuo left 3.3 g. (99%) of crude 6-benzamido-9-(2',5'-di-O(173 mg.). Replacing I11 with 2,5-0-benzoyl-3-acetamido3-deoxy-~-ribofuranose(X) gave a 29% yield of combined benzoyl-3'-phthalimido-3 '-deoxy+ -D -ribofuranosyl) - purine as a cream colored glass. VI1 and VIII. When the 2,5-di-O-benzoyl-3-acetamido-3To the 3.3 g. of blocked nucleoside was added 66 cc. of deoxy-D-ribofuranosyl chloride titanium chloride complex was preformed,8 then condensed with chloromercuri-6-ben- reagent methanol and 5.15 cc. of 1 N methanolic sodium zamidopurine the yield of combined VI1 and VI11 mas 47y0. methoxide. The mixture was refluxed for 40 minutes, solution being complete after 2 minutes a t the b.p.; thepH All three runs were the same size. remained
