Configurational Inversion within a Furanoside Ring by Anchimerically

Configurational Inversion within a Furanoside Ring by Anchimerically Assisted Displacement: 5-Deoxy-D-ribose from 5-Deoxy-D-xylose. Kenneth J. Ryan, H...
0 downloads 0 Views 972KB Size
June 20, 1964 [CONTRIBUTION FROM

DEO OXY-D-RIBOSE

FROM

2497

~-DEOXY-D-XYLOSE

LIFE SCIESCES RESEARCH, STANFORD RESEARCH INSTITUTE, MENLOPARK.

CALIF.]

Configurational Inversion within a Furanoside Ring by Anchimerically Assisted Displacement : 5-Deoxy-D-ribose from 5-Deoxy-D-xylose‘ BY KENNETH J. RYAN,HENRIARZOUMANIAN, EDWARD M. ACT ON,^

LEONGOODMAN

AND

RECEIVED JANUARY 10, 1964

Treatment with sodium benzoate in boiling dimethylformamide easily displaced t h e mesyl group in methyl 2-O-benzoyl-5-deoxy-3-O-mesyl-~,~-~-xylofuranos~de ( V I I I ) with inversion a t C-3, forming a derivative of 5deoxy-D-ribose as the only sugar detected in the product. T h e presence of 5-deoxy-D-ribose and absence of 5-deoxy-~-xyloseand 5-deoxy-D-arabinosewere determined from a gas chromatographic comparison of bis-0(trimethylsilyl) derivatives of the methyl furanosides Since methyl 2-0-benzyl-5-deoxy-3-O-mesyl-a-~xylofuranoside (a-XXIV), an analog of VI11 lacking a participating group, was found more resistant t o this type of displacement, i t is therefore concluded t h a t t h e inversion observed with VI11 occurs via participation of the neighboring benzoyl group. Furthermore, the more drastic conditions necessary for the conversion of a-XXIV t o a derivative of B-deoxy-~-riboseled to the formation of by-products, notably the 3,4-olefin XXVII.

ful with hydrazine,* however, makes it clear that t h e successful transformation of I1 to I via sulfonate displacement could not be predicted from the work of the Birmingham group.* Furthermore, it seemed desirable to investigate facilitation of the displacement by a substituent capable of neighboring group participation, since recent studies in open-chain s u g a r ~ pointed ~ ~ ~ out ~ the ease of displacement of secondary sulfonate esters with anchimeric assistance using the sodium benzoateFHzOH DMF reagent. It was also important t o clarify any doubts posed by the fact that attemptsx1 (with a less favorable solvent) to utilize such neighboring group participation in solvolyses of ring sulfonates failed to I \ convert a D-glucopyranoside or a D-altropyranoside to a OH OH OH D-allopyranoside. I II In the initial work on the conversion of I1 t o I, a simple model system was sought which would minimize ary sugar sulfonate esters, but most of this has inthe blocking and deblocking problems necessary in such volved sulfonates in open-chain sugar derivatives or the an investigation. This paper reports the facile displaceexocyclic secondary sulfonates in ring sugars. In ment of a secondary ring sulfonate ester, assisted by a particular, in the case where the over-all result of the neighboring benzoate, in the model compound methyl displacement is simple epimerization of a secondary 2-0-benzoyl-5-deoxy-3 - 0 - mesyl- a,P-D - xylofuranoside sugar hydroxyl ( L e . , the sulfonate ester is displaced by (VIII) ; and it provides a procedure for the preparation an oxygen of the nucleophile), only one example is reof 5-deoxy-D-allose from 11, which will be described in a corded where a secondary sulfonate ester in a sugar ring separate paper. was involved. In this e ~ a m p l e ,the ~ conversion of methyl 3-0-mesyl-2,5-di-0-methyl-a-~-rhamnofurano- In the model compound VIII, it was seen that inverside t o methyl 3-O-benzoyl-6-deoxy-2,5-di-O-methyl-sion could occur either by neighboring group participation or by direct displacement at C-3, if that was easy. a-L-altrofuranoside with sodium benzoate in dimethylformamide ( D M F ) occurred with relative ease, in a reaction that would appear to be s N 2 in character. The sodium benzoate-DMF reagent5 has been widely exploited in such epimerizations. I t is known that secondary sulfonates are much less readily displaced than primary, and that the presence of the sulfonate in a sugar ring increases the difficulty of displacement.6 The fact that the sulfonate ester of 1 , 2 : 5,G-di-0-isopropylidene-3-O-tosyl-D-glucose was not displaced by7 sodium benzoate-DMF, a displacement that is success-

In a continuing series of studies3 on the synthesis of 5-deoxy-~-allose( I ) , desired as a precursor for the preparation of “homoadenosine,” a scheme was considered which involved the conversion of 5-deoxy-D-glucose (11) to I by inversion of the C-3 hydroxyl in an appropriate derivative of 11. A considerable amount of work has been done on the nucleophilic displacements of second-

Y

(1) This work was carried o u t under t h e auspices of t h e Cancer Chemotherapy National Service C e n t e r , National Cancer I n s t i t u t e , National Institutes of H e a l t h , Public Health Service, Contract No. PH43-64-.500. T h e opinions expressed in this paper a l e those of t h e authors and not necessarily those of t h e Cancer Chemotherapy S a t i o n a l Service Center. (2) T o whom reprint requests should be addressed. (3) For t h e first paper in this series, see H . Arzoumanian, E. M. Acton, a n d L. Goodman, J . A m . Chem. Soc., 8 6 , 74 (1964). (4) A. B. Foster, J. L e h m a n n , and M . Stacey, J , Chem S o c . , 4649 (1961). ( 5 ) E. J. Reist, I> Goodman, a n d B R . Baker, J. A m . Chem. Soc., 80, 577.5 (19.58) (6) R . S. Tipson, Advan. Carbohydvafe Chem., 8 , 180, 212 (19.53); J. M. Sugihara, rbid., 8 , 26 (1953). (7) E. J. Reist a n d R . Spencer, unpublished results from these laboratories.

o,

o,

Ph+OH IV

I



o=c P I

Ph

Direct s N 2 displacement could afford a derivative only of 5-deoxy-~-ribose. Participation, through the inter(8) M. I,. Wolfram, J. Bernsmann, and I). H o r t o n , J . Org, C h r m . , 27, 4,505 (1962). (9) B. R . Baker and A. H . Haines. i b i d . , 28, 438 (1963). (10) M . A. Bukhari, A. B. Foster, J. Lehmann, M. H . Randall, a n d J . M. Webber, J. Chem. Soc , 4 1 6 7 (1963). (11) R . W . Jeanloz and D A. Jeanloz, J . A m . Chem Soc., 80, 5692 (19.58).

K. J. RYAN,H. ARZOUMANIAN, E. M. ACTON,AND L. GOODMAN

2498

mediate acylonium ion 111, could afford this and two additional possibilities, depending on how I11 subsequently reacts Attachment of water, from moisture present during the reaction or from the work-up, to the benzoyl carbon (path a ) and opening of the resultant ortho ester IV results in the ribo configuration. Attack on 111 a t C-3 (path b) or C-2 (path c) by benzoate anion or other nucleophile present could give the xylo ( L e , starting material) or the arabino configuration, respectively. These alternatives needed to be studied and, hopefully, paths b and c excluded before the inversion method could be applied to the glucose-allose transformation. In the preparation of VIII, the known 5-deoxy-l,2-0isopropylidene-~-xylofuranose~~~~~ (V) was first converted to the 3 - O - m e ~ y l a t e VI. '~ This material, analytically pure but noncrystalline in our hands, was con-

group, as shown by n.m.r. and infrared spectra. Evidence from these spectra that the product existed largely as a monohydroxybenzoate suggested that i t was formed by path a . Direct displacement or either of paths b and c should form, a t least initially, a dibenzoate, and it was quite unlikely that very extensive saponification of a benzoate to a hydroxyl had taken place subsequently during the reaction. The product was investigated further after removal of protecting groups.' Debenzoylation afforded a water-soluble methyl 5deoxypentofuranoside which was potentially a mixture of the ~ - r i b o( X I ) , D-xylo ( X I I ) , and D-arabino ( X I I I ) compounds. The most definitive evidence for the nature of the material came from a gas chromatographic study of the 2,3-bis-O-(trimethylsilyl) derivative. The trimethyl-

OR XI, R XIY, R

V, R = H VI, R = M s

OR

VII, R = H VIII, R = BZ

OR

IX, R = H X, R = B z

verted to the a,@-methylfuranoside VI1 upon removal of the isopropylidene group in boiling 2y0 methanolic hydrogen chloride. The resultant sirup was benzoylated to form VIII, which was purified by chromatography on alumina. Benzene elution afforded analytically pure a,p-VIIl in several fractions. The n.m.r. s1)ectruiti indicated the proportion of anomers was ca. ? C Y / $ ;this was verified by comparison with a sample of pure ~ ~ - 1 ~ 1 later 1 1 , obtained by another route (see Experimental). Further elution of the original chromatogram with methanol unexpectedly afforded a fraction of nearly double the molecular weight of VIII, and in which the ratio of sulfonate to carbonyl absorption in the infrared was qualitatively higher than that in VIII. These facts, don: with the elemental analyses, suggested structurc. X for this material. Lye suspect dimerization occurred in the methanolysis step by a process akin to that called 'ireversion,"15~'6 and that the resultant I X was benzoylated to form X along with VIII. Treatment of a,P-VIII with sodium benzoate in refluxing dimethylformamide for 0 hr., the general procecl~ire,~~ caused complete ejection of the mesyl A . Levene a n d J . C o m p t o n , J . Bioi. C h e m . , 111, 32.5 (1935) P A . Levene a n d A . L. R a y m o n d , ibid., 102, 317 (1933). (14) H. K u z u h a r a and S . E m o t o , A g i . Biol Chem. ( T o k y o ) ,27, 689 (1963).

(12) E' 11:1!

( l a ) W . Pigman, " T h e Carbohydrates," Academic Press, I n c . , N e w Y o r k , N . Y . , 1957, pp. 59-60, 486. (16) This was, a p p a r e n t l y , largely o r completely avoided in ref. 14 b y preparing V I 1 f r o m \'I b y a n acetolysis-mild methanolysis sequence. (17) When a,p-T-III was heated in I I 3 I F for 6 hi- in t h e absence of sodium I,cr,zoatc. t h e mixtut-e darkened badly and there was spectral evidence f o r civzr:i