configured products predominantly, the α-D-mannose-derived bromides
(3r , 3x ) favored α-products. Unfortunately,
2-deoxyglucose failed to yield products due to the decomposition of
glycosyl bromide and chloride. When furanose derivatives were
investigated, D -ribofuranose-derived chloride produced the
desired product in low yield with high β-selectivity (3u ),
although D -arabinofuranose-derived bromide resulted in
significant hydrolysis products. The authors found that ligand could
play a significant role in the stereochemical outcome of the reaction.
Take the synthesis of phenyl C-mannoside as an example. They found that
by using tBuTerpy, a 2.9:1 ratio of α/β isomers was produced
(3q ). However, good-to-excellent α-selectivities were achieved
using Terpy and PyBox in good efficiency (3q ). These results
suggested that different ligand catalysts may function by different
mechanisms, with the observed selectivity indicative of a stereochemical
mismatch between the ligated nickel catalysis and carbohydrate
substrates. Catalyst such as PyBox/Ni(COD)2, which is
smaller, provided good α-arylation selectivities likely viasubstrate control. Mechanistically, the authors proposed two variants of
the nickel-catalyzed Negishi cross-coupling pathway. As shown in Scheme
1-C, the reaction could occur either by halogen abstraction (inner
sphere, top) or by a single electron transfer (outer sphere, bottom)
mechanism.[9] Lastly, the authors demonstrated the
utility of this protocol with the successful total synthesis of
Salmochelin SX (8 , Scheme 1-B).
Scheme 2 α-Vinyl/Aryl C-glycosides via nickel-catalyzed
reductive couplings