The chemistry associated with the anomeric center is cationic by nature. One interesting approach to C-glycoside synthesis is to reverse the electronic character of this center from electrophilic to nucleophilic. Hence anions at the anomeric center (gly-cosyl anions) have now become fairly common in the synthesis of C-glycosides. Of course, one limitation of this method was the presence of heteroatoms at C2, since any discrete anion at C1 would surely cause an elimination reaction to occur and in turn produce the corresponding glycal (55 ^ 58, Scheme 13). This is in fact a very common way of preparing sugar glycals from sugars. As a consequence of this limitation, the chemistry of 2-deoxy C1 anions (56 ^ 59) has been extensively explored. An interesting development in this area came with the observation that if the hydroxyl group at O2 is unprotected, deprotonation followed by anion formation at C1 can occur to give species 57, which can go on to react with a suitable elec-trophile to deliver 60, after workup. It should be noted that a publication dealing with the preparation of anomeric stannanes from 2,3-dihydro-4H-pyran-4-ones has appeared . A related, and important observation was that anomeric organosa-marium species possessing a protected oxygen atom at C2 are stable enough to react via the anomeric carbon atom with suitable electrophiles to give access to C-gly-cosides (61 ^ 62, Scheme 13) .
B. Dianion Approaches
Work by the groups of Vasella and Kessler (Scheme 14) focused on the preparation of dianions from the corresponding 1,2-anhydro sugars. Thus treatment of 63 with Ph3SnLi gave 64 as the major product in 65% yield. Exposure to an excess of n-BuLi resulted in formation of dianion 65, which was then quenched with a variety
of electrophiles to give the ¡3 isomers (66-69 ) as the major compounds in all cases, attesting to the configurational stability of these anions at low temperature .
This strategy was then applied to the preparation of the anomeric ¡3 and a acids 72 and 76. The 3 isomer 72 was prepared by application of the epoxide-opening strategy above, while the a acid 76 was synthesized by using the procedure originally developed by Kessler and coworkers. Once prepared, the acids were coupled with the protected asparagine amino acid 78 to give the unnatural C-glycosylated amino acids 73 and 74 (Scheme 15) .
Glycosyl phenyl sulfoxides also serve as a good source of dianions. The a-sulfoxide was metallated to the dianionic species, treated with a suitable aldehyde to give alcohol 81 and reduced product 82 (Scheme 16) .
The Kessler group then investigated the question of whether similar dianion chemistry could be carried out with the corresponding 2-deoxy-2-amino sugars. Accordingly, compound 83 was treated with thionyl chloride to give the intermediate 2-deoxy-2-aminoglucosyl chloride, which was then treated sequentially with ra-bu-tyllithium and lithium naphthalenide to give intermediate 84. Quenching with a variety of electrophiles gave modest to good yields of the 2-deoxy-2-amino-a-C-gly-
cosides 85 stereoselectively (Scheme 17) . Similar chemistry was also carried out in the galacto series .
The corresponding /3 dianion 84 could be accessed by first converting 83 to the a-pyranosyl chloride followed by SN2 displacement with tributylstannyllithium to give 84. Deprotonation and subsequent halogen metal exchange was followed by quenching with various electrophiles to provide the ^-isomeric C-glycosides 87 . The chemistry was then applied to the preparation of a C-glycoside-based analog of V-glucoaspargine 88 (Scheme 18) .
C. Anomeric Organosamarium Species: O2 Sugars
Work in the area of O2 sugars was spurred by the following observations: upon treatment of 89 with samarium iodide (to obtain an anomeric free radical that would
then cyclize onto the pendant acetylene), three products were obtained. The first, 91, was the expected product of free radical cyclization in 25% yield; the product of elimination 58 (11%) and the product of simple deoxygenation 90 (60%) also were isolated (Scheme 19). This was very surprising, since the elimination product is produced from the organometallic anomeric samarium species that then undergoes a 1,2-elimination, but 90 is derived from protonation of the same intermediate organ-ometallic species! This implies that the organometallic species is stable enough to be deprotonated to give 90 .
The workers then took advantage of this reactivity to generate anions at the anomeric center of 2-oxygenated sugars and have succeeded in showing that pyridyl sulfones make excellent precursors for such reactions. When O2 is protected at a t-butyldimethylsilyl ether, acceptable yields of 0-C-glycosides are obtained from the corresponding 0-pyridyl sulfone (92 ^ 93, Scheme 20). When the a-manno derivative 94 was used, a good yield of the corresponding a-C-glycoside 95 was obtained . The authors attribute the discrepancy in yields to the ability of the manno intermediate to attain a reactive boat conformation that is more stable than the gluco derivative. With anomeric phosphates such as 96 as precursors, Wong has used a similar reaction to prepare a-C-manno pyranosides. Good yields of a-C-glycosides were obtained (Scheme 20). In the absence of the electrophile, the 1-deoxy sugars were the main products formed. The use of furanoside 98 gave good yields of C-glycoside products . The results were obtained in the gluco series paralleled those of Beau and coworkers.
Scheme 21 shows the reaction of the 2-deoxy-2-amino sugar 100 reacting with samarium iodide and cyclohexanone to give the desired C-glycoside 101 as the major compound . When the azide was used in lieu of the protected amine function, only the product of elimination was formed. This indicates a possible complexation between the acetimido oxygen atom (Scheme 21, second equation) and the samarium atom, resulting in stabilization of the a-organosamarium species, since normally,
anomerization to the more stable ^-organosamarium species would be expected to take place. In this case the complexation holds the organosamarium intermediate in the a configuration. It may also be possible that a chair flip of 102 occurs to give 103, which can then react via the equatorial direction to ultimately give the a-C-glycoside .
The same workers applied their methodology to the preparation of a C-gly-coside analog of the Tn antigen. The Tn antigen epitope is expressed in over 70%
of human epithelial cancers. This epitope is a major tumor-associated O-linked glycoprotein. The same antigen has also been identified as a partial structure of the HIV envelope glycoprotein gp120. The workers designed an analog in which the inter-glycosidic serine oxygen atom was replaced with a methylene group. The anomeric pyridyl sulfone 100 was coupled with aldehyde 104 to give 105 as a mixture of isomers. Radical deoxygenation was followed by standard manipulations to deliver the analog building block 106 (Scheme 22) .
Lindhart and coworkers also relied on a samarium-mediated condensation reaction to prepare a-C-glycosides related to V-acetylneuraminic acid (Neu5Ac). Neu5Ac is frequently found as the terminating residue of cell surface glycoproteins and glycolipids and is involved in a number of important biological events such as intracellular interactions, adhesion, and aggregation. A stable analog that would be hydrolytically inert was desired, and thus sulfone 109 was exposed to standard samarium iodide conditions (Scheme 23). However, it gave only the net product of reduction 110 .
When the same reaction was carried out with pyridyl sulfone 111 in the presence of the shown cyclic ketone, a 90% yield of 112 was obtained. Several other carbonyl compounds were also used as electrophiles with the corresponding pyridyl sulfone and gave good yields of coupled products. However, the reaction did not proceed with hindered carbonyl compounds . Matsuda and coworkers have found that if such 2-keto-thioglycosides are treated with samarium iodide in THF followed by quenching with a carbonyl compound, good yields of C-glycosides are obtained, with the a isomer (114) predominating in a ratio of 9:1 (Scheme 24) .
Kocienski et al. showed that the a-glycosyl copper(I) compound 115 reacts with the molybdenum complex 116 to give an excellent yield of the allylic a-C-glycoside 117 (Scheme 25). The corresponding ^-glycosyl copper(I) reagent gave a
49% yield of the corresponding ^-C-glycoside. When the mirror image of the cati-onic complex was used, the stereochemistry of the newly formed stereogenic center was found to be reversed .
Was this article helpful?