HAKSHELMUT BAERAND HERMANN 0. L. FISCHER
5 184
[CONTRIBUTIOS FROM
THE
I-01. S l
DEPARTMENT O F BIOCHEMISTRY, UNIVERSITY O F CALIFORSIA, 'BERKELEY]
Cyclizations of Dialdehydes with Nitromethane. 1I.I Preparation of 3-Amino-3deoxy-D-ribose and 3-Amino-3-deoxy-~-ribose BY
HANSHELMUT BAERAND HERMANN 0. L.
FISCHER
RECEIVEDMARCH24, 1959 The two enantiomeric dialdehydes obtained by periodate cleavage of the methyl pentopyranosides can be cyclized with nitromethane and sodium methoxide in a yield up to 40% to two crystalline methyl 3-nitro-3-deoxy-,9-pentopyranoside sodium salts which belong to the D- and L-series. The configurations of the salts correspond t o the 3-deoxy-riboses (3-deoxyxyloses). Acidification of the sodium salts results in the production of the methyl 3-nitro-3-deoxy-P-ribosides as the main product; the corresponding xylosides are produced in small amounts. Catalytic hydrogenation yields quantitatively the corresponding amino sugar derivatives. This sequence of reactions, followed by acid hydrolysis, presents a new method fur the preparation of the D- and L-forms of 3-amino-3-deoxy-ribose.
The well-known condensation of nitro-alkanes, IT^ c o---drI and in particular nitromethane, with aldehydes I was first applied to aldoses some 15 years ago by Methyl p-D- or 2NaI04 Sowden and Fixher.* In 1948, Grosheintz and a-t-peritopyranoside.; -+ Fischer3 found that 1,2-O-isopropylidene-D-xylotrihydroxyglutaric dialdehyde, a partly blocked Ia CH20dialdehyde of the pentose series, could be condensed with nitromethane and gave, in several steps, a good HC-OCH~ yield of a mixture of nitro-deoxy-inositols. This I ZNaIO, Methyl CY-D- or CHO formation of six-carbon rings made possible the +--6-L-pentopyranosides chemical transformation of D-glucose to myoCHO inosit01.~ It was expected that the direct condensation of a CHzO-2 dialdehyde with nitromethane would lead to the Ib formation of a cyclic structure just as readily as the aqueous-ethanolic solution. The condensation restepwise reaction mentioned above. Furthermore, action started immediately and was indicated by it seemed to be likely that a pyranoside ring could the precipitation of colorless crystals (decompobe formed in a similar way. Therefore, we under- sition point above 160') of the composition took to condense with nitromethane and sodium CaHlaOeNNa. We assigned to this reaction prodmethoxide dialdehydes produced by periodate uct, which was obtained in approximately 3770 cleavage of pentopyranosides. The reaction pro- yield, the formula of a methyl 3-aci-nitro-3ceeded indeed as anticipated and was described, deoxy-P-D-ribo-(xylo)-pyranosidesodium salt IIa without reference to stereochemical consideration, for reasons discussed below. in a preliminary note.' OC H OCM? According to Jackson and Hudson6 periodate oxidation of the four possible methyl P-D-pentopyranosides yields the same dialdehyde, L'-methoxy-diglycolaldehyde (Ia, [ L Y ] ~ O D - 142'). It can also be prepared as well by glycol cleavage of methyl a-L-pentopyranosides as demonstrated by Grosheintz,G who used methyl a-L-arabinopyranoside. The optical antipode of Ia, D'-methoxyThe sodium salt I I a is easily soluble in water, diglycolaldehyde (Ib, 4- 124') can be pre- showing immediately after dissolving a specific pared analogously from the anorners, the methyl rotation of - 160'. The rotation changes quickly a-~-pentopyranosides~ and ~ ~ methyl p-L-pento- and after several hours reaches a constant end pyranosides6 value [ a ] D - 117'. The explanation for this This paper describes first the sequence of unexpected mutarotation appears to be obscure. reactions carried out with the dialdehyde I a pre- However, this may indicate that a profound change pared from methyl p-D-xylopyranoside, and, sub- takes place in the molecule. This problem will be sequently, the analogous results obtained with Ib. investigated a t a later date. Because of the apDialdehyde Ia (levorotatory) was reacted with parent instability of the nitropentoside sodium nitromethane and sodium methoxide a t 0' in salt in water, the next step, namely, the preparation (1) Communication I: H . H. Baer and H . 0. L. Fischer, Proc. Nal. of the free nitropentoside from the sodium salt, A c a d . Sci.. 44, 991 (1958). was carred out under anhydrous conditions. The (2) J. C. Sowden and H . 0. L. Fischer, THISJ O U R N A L , 66, 1312 "dry" acidification of IIa, by grinding it with (1944). For a comprehensive review see J. C. Sowden, Advames in solid potassium bisulfate,*yielded 92y0 of an etherCarbohydiaie Chem., 6 , 291 (1951). (3) J. M ,Grosheintz and H. 0. L. Fischer, THISJOURNAL, 7 0 , 1476, extractable sirup (mixture A ; [ a I z 4 ~ 133'), 1479 (1948). which consisted for the greatest part of methyl (4) Th. Poslernak, Helu. Chim. Acta, 33, 1597 (1950).
1
(5) E. L. Jackson a n d C. S. Hudson, THISJOURNAL, 6 9 , 994 (1937);
63, 1229
(1941).
( G ) J. M. Grosheintz, ibid.. 61, 3379 (1939). (7) W. D. Maclay a n d C. S. Hudson, ibid., 60, 2059 (1938).
(8) Potassium bisulfate has been used previously for acidification of stereochemically labile salts: cf. K . Freudenberg, Ber., 47, 2035 (1914): R. Kuhngand F. Ebel, i b i d . , 68, 929 (1926). We are obliged to Professor K u h n for kindly pointing-out this method.
OCt. 5, 1959
PREPARATION O F 3-AMINO-3-DEOXY-D-RIBOSE AND 3-AMINO-3-DEOXY-L-RIBOSE
3 - nitro - 3 - deoxy - P - D - ribopyranoside (IIIa). This nitroriboside I I I a crystallized in part (43%) out of the sirup in colorless, prismatic crystals, m.p. 92-93' and [ c L ] ~ ~D 117'. By addition of the equivalent of sodium methoxide to the ethanolic solution of IIIa, the sodium salt IIa could be recovered quantitatively. Hydrogenation of I I I a with platinum catalyst in the presence of 1 mole of dilute hydrochloric acid gave quantitatively, with absorption of 3 moles of hydrogen, a new methyl amino-deoxy-pentoside hydrochloride IVa with melting point 171-172' dec. and [ a I z 3 D - 126'. Its acid hydrolysis yielded readily the knowng 3-amino-3-deoxy-Dribose hydrochloride (Va). Comparison of the latter product with an authentic sample of 3arnino-3-deoxy-~-ribosehydrochloride kindly supplied by Dr. R. E. Schaub showed that they were identical. In addition, we have observed that the sugar exhibits, in aqueous solution, a rapid upward mutarotation, [alZ7D- 37.5' (2 min.) -23' (30 min., final), indicating that the crystals represent the @-anomer.
5185
IIO
Fig. 1.
In slightly better yield, one can prepare the 3amino-xyloside VIa by reducing mixture A' in alkaline instead of acid solution. That is to say, the reaction becomes alkaline during the hydrogenation due to the amino sugar formed if one does not add hydrochloric acid. In order to have an alkaline medium from the beginning, a small amount of triethylamine was added. Under these conditions, VIa crystallized directly out of the hydrogenated reaction mixture in a yield of 5.8%. OH The alkalinity apparently causes a partial 3epimerization of the nitroriboside I I I a to the nitroxyloside VIIa via the a ~ i - n i t r o - a n i o n . ~In~ kH2 OH. HCl our case, however, such an epimerization seems to \'a be of secondary importance. In the first place, the yield of aminoxyloside is only slightly higher The fact that our amino sugar Va proves to be upon hydrogenation in alkaline medium; and, 3-amino-3-deoxy-~-ribosehydrochloride also proves secondly, amino riboside is the main product even that the aminopentoside IVa and the nitropento- in that medium. The amino riboside can be isoside I I I a have the proposed formulas of D-ribose lated in the form of its hydrochloride IVa upon derivatives substituted in the 3-position. addition of hydrochloric acid after completion of The sirup (mixture A'), remaining from mixture the hydrogenation. -4 after removing the crystals IIIa, still contained The mode of preparation described above for considerable amounts of I I I a which obviously had derivatives of 3-deoxy-~-riboseand 3-desoxy-Dnot crystallized out completely. By catalytic xylose has been applied to the L-series, starting hydrogenation in the presence of 1 mole of hydro- from the dialdehyde I b which resulted from the chloric acid, an additional 20% of pure IVa could periodate oxidation of methyl P-L-arabinopyranobe isolated. The mother liquor of the hydrogena- side. The same sequence of reactions gave, in the tion, freed of chloride ions by an ion exchanger, same yields, the derivatives IIb, IVb, Vb, VIb. yielded approximately 2.6% of a crystalline methyl The properties of these derivatives agreed very well amino-deoxy-pentoside (VIa) as the free base, with those of the corresponding compounds of the [ c L ] ~ ~D 65'. On account of this low rotation, D-series; only the signs of their optical rotations the product was thought unlikely to be the free were opposite. Consequently, they have to be base of the hydrochloride IIIa. The rotation formulated as enantiomorphs of the compounds corresponded rather to that of methyl 3-amino-3- described above. deoxy-P-D-xylopyranoside,lovll and indeed it proved Mixture B ([(Y]D 132'), the product of acidito be identical, with regard to melting point and fication of IIb, thus far has not shown any tendency infrared spectrum, with an authentic preparation to crystallize. Therefore, the separation and of this compound prepared by Schaub. purification of the free nitro-riboside has not yet This shows that upon acidification the nitro- been achieved in the L-series. Rather, we have pentoside sodium salt yields not only the nitro- hydrogenated the oily mixture B directly both in riboside IIIa but in addition a small amount*2of acid and alkaline medium. Upon acid hydrogenamethyl 3-nitro-3-deoxy-&~-xylopyranoside (VIIa). tion, we could isolate pure methyl 3-amino-3So far VIIa could not be isolated in pure form from deoxy-P-L-ribopyranosidehydrochloride (IVb) showmixture A. ing melting point 173' dec. and [aIz5D 123' in a (9) (a) C. W. Waller, P . W ,Frgth, B. I,. Hutchings and J. H. Wilyield of about 41-53y0 of the theory. liams, THIS JOURNAL, 7 6 , 2025 (19.53); (b) €3. R. Baker and R . E. Upon alkaline hydrogenation, we obtained 11% Schaub, J . Org. Chem., 19, 646 (1984); (c) B. R. Baker, R.E. Schaub of free methyl 3-amino-3-deoxy-~-~-xylopyranoand J. H. Williams, T H I SJOURNAL, 7 7 , 7 (19.55). side (VIb). It proved to be identical with u
