the desired thermodynamic products at N-9 for purines and N-1 for pyrimidines, the condensation reactions
are often mechanistically much more complex.^9 Thus, there is considerable evidence for pyrimidines that
reaction initially gives an O-glycoside or even an O^2 ,O^4 -diglycoside that is then transformed into the
desired N-glycoside. For purines, condensation initially takes place on N-3 for adenine and its derivatives or
alternatively at N-7, particularly for bases with a 6-keto substituent (e.g. N^2 -acetylguanine and hypoxanthine).
The general mechanism for N-3→N-9 glycosylation is shown in Figure 3.5 and proceeds stereoselectively
owing to the formation of an acyloxonium ion intermediate (see Section 3.1.1.7).
3.1.1.2 Fusion Synthesis of Nucleosides. Two disadvantages of the above methods are the poor
solubility of the mercury derivatives and the instability of the halogeno-sugar derivative. Furthermore, the
biological activity of a number of nucleosides synthesised in this way has often been wrongly assigned owing
to the presence of trace amounts of mercury in samples. One early improvement was the combination of
1-acetoxy sugars with Lewis acids such as TiCl 4 or SnCl 4 as a means of generating the reactive halogeno-sugar
in situ. That led to the fusion process, in which a melt of the 1-acetoxy sugar and a suitable base in vacuo,
often with a trace of an acid catalyst, can give acceptable yields of nucleosides.^10 Thus, 1,2,3,5-tetra-O-acetyl-
D-ribofuranose fused with 2,6-dichloropurine^11 or 3-bromo-5-nitro-1,2,4-triazole^12 gives useful yields
of the corresponding acylated nucleosides (Figure 3.6). This method works best for purines that contain
electron-withdrawing groups and have low melting points. Recent examples include the syntheses of
2 -deoxyribonucleosides of purines, but such methods result in anomeric mixtures of nucleosides.
3.1.1.3 The Quaternization Procedure: Hilbert Johnson Reaction. Hilbert and Johnson noticed
that substituted pyrimidines are sufficiently nucleophilic to react directly with halogeno-sugars without
any need for electrophilic catalysis. The method, which bears their name, involves the alkylation of a
2-alkoxypyrimidine with a halogeno-sugar13,14and has been reviewed.^15 The initial product is a quaternary
salt, which at higher temperatures eliminates an alkyl halide to give an intermediate condensation product.
Further chemical modification of substituents on the pyrimidine ring can lead to a range of natural and
artificial bases (Figure 3.7). Such condensations frequently give mixtures of - and -anomers although
the use of HgBr 2 increases the proportion of the -anomer.
3.1.1.4 Silyl Base Procedure. A major improvement came from the utilisation of silylated bases
(silyl-Hilbert–Johnson method), developed independently by Nishimura,^16 Birkofer^17 and Wittenberg.^18
Silylated bases have three advantages: (1) they are easily prepared, (2) they react smoothly with sugars in
homogeneous solution due to their increased solubilities and greater nucleophilicities, and (3) they give
intermediate products that can be easily converted into modified bases. The early use of mercuric oxide as
80 Chapter 3
NNOEtHgClO+RBr
NNOEtRONNNH 2R'ONNHOR'O(i)(ii)(iii)Figure 3.4 Chloromercuri route for synthesis of pyrimidine nucleosides. Reagents: (i) xylene, 120°C; (ii) NH 3 ,
MeOH; and (iii) NaOH aq. Rprotected ribofuranosyl; R1--D-ribofuranosyl