Making carbon–carbon bonds is hard. Linking racemic mixtures of two alkyls while simultaneously controlling the stereochemistry of both ends of the product is nigh on impossible. Now Gregory Fu, together with his colleagues at the California Institute of Technology, has found a way to make it easy. Using a chiral nickel catalyst, Fu coupled a racemic mixture of electrophiles and nucleophiles with up to 82% yield and 95% stereoselectivity (Science, 2020, DOI: 10.1126/science.aaz3855). What’s more, this unlikely reaction is compatible with 19 functional groups. Most substitution reactions for forming the C–C bonds critical to pharmaceutical manufacturing have major limitations. They only work for some alkyl compounds, and often lead to side reactions, especially with bulky substituents. Plus, these reactions tend to form a racemic mixture of products. But usually only one isomer of a pharmaceutical compound is biologically active.
In order for the catalyst to be stereoselective, its ligand has to be able to distinguish between the different alkyl substituents. Fu’s hunch about how to make this happen hit pay dirt: the group used a ligand, an isoquinoline–oxazoline, that is bidentate. That means the open spots on the Ni can bind with the oxygen on the amide nucleophiles (shown in blue) so the electrophiles (shown in red) can couple to give one stereoisomeric compound out of the four potential products. The ligand is relatively easy to make, with cheap Ni and a three-step synthesis.
Fu hypothesizes that the new process goes through a radical mechanism. The radical intermediates lose their stereochemistry as they are formed, allowing stereo-convergence to occur. This is why it can accommodate so many functional groups, including a thiophene, aldehyde, and aryl bromide. “Radicals aren’t these horribly promiscuous species,” he says.
This work was proof of concept, but the method has potential to become a major tool in synthesis, says Jin-Quan Yu, an organic chemist from the Scripps Research Institute.