In nature, enzymes (like cytochrome P450) are thought to effect site and chemoselective C—H oxidations using reactive metal oxidants (iron oxos) via elaborate protein binding pockets that exploit shape and functional group recognition of the substrate. Previously, it was thought not possible to achieve site-selective C—H methylene oxidation with simple catalysts because these bonds are among the strongest and most ubiquitous, decorating the hydrocarbon skeleton of most molecules. In 2007, our group discovered a simple iron catalyst Fe(PDP) for site-selective aliphatic oxidations. Fe(PDP) catalysis illuminated that C—H bonds in complex molecules exist in different electronic, steric, and stereoelectronic environments that can be preparatively distinguished by simple catalysts that proceed via late, product like transition states. This paradigm shift enabled the blossoming area of late-stage C—H functionalizations where in 2018 such atomistic changes are being done to streamline syntheses and generate diversity.
A key challenge that remained was chemoselectivity, that is the ability to hydroxylate such strong bonds in the presence of oxidatively more labile pi-functionality. Because most drugs have at least one aromatic or heteroaromatic moiety in their hydrocarbon scaffold, this was a critical challenge to overcome. Enzymes restrict access of the aromatic portion of the molecule to the oxidant using the protein binding pocket, however this approach has not been successful for small molecule catalysts. A key insight by us was that the steric and electronic requirements for C—H oxidation may differ from those for aromatic oxidation. Specifically, aromatic oxidation via single electron transfer may be attenuated by replacing the iron metal with manganese, a metal that can still support oxo formation but has a lower oxidation potential. Aromatic oxidation to phenols via epoxidation/hydride shift (“NIH shift’), may be attenuated using a bulkier ligand that should favor C—H oxidation over the more sterically demanding epoxidation pathway. Such thinking ultimately led to the discovery of Mn(CF3-PDP) 1 a highly chemoselective but poorly reactive catalyst. Carboxylic acids are part of the oxidant in Fe(PDP), so we explored their electronic properties to tune the electrophilicity of the manganese oxidant. We found a significant boost in reactivity, without harming chemoselectivity, by the addition of the electron-deficient carboxylic acid additive chloroacetic acid.
Together with a talented, international team [MCW (Greece/USA), Takeshi Nanjo (Japan), Jinpeng Zhao (China, lead author), Emilio de Lucca Jr. (Brazil)], we demonstrated combination of Mn(CF3-PDP) 1/chloroacetic acid has led to a remarkable system that can preparatively oxidize 50 aromatic compounds at remote methylene cites in the presence of medicinally relevant halogen, oxygen, heterocyclic and biaryl moieties. Chemoselectivity in late stage oxidation is showcased in four drug scaffolds that undergo methylene oxidation in the presence of oxidatively labile alkynes, activated benzylic sites, and basic amines. We demonstrate that all other reported oxidants, including TFDO, are not able to effect remote methylene oxidation in the presence of such aromatics. We additionally demonstrate the ability to effect rapid generation of a drug metabolite (piragliatin) from an advanced intermediate. We anticipate that this method will enable late-stage oxidation to gain even more widespread use in drug discovery and development. Additionally, we hope this discovery inspires future catalyst design to strive to solve selectivity challenges by seeking simplicity rather than by adding complexity.