Optimization of the catalyst structure to simultaneously improve multiple reaction objectives (e.g., yield, enantioselectivity, and regioselectivity) remains a formidable challenge. Herein, we describe a machine learning workflow for the multi-objective optimization of catalytic reactions that employ chiral bisphosphine ligands. This was demonstrated through the optimization of two sequential reactions required in the asymmetric synthesis of an active pharmaceutical ingredient. To accomplish this, a density functional theory-derived database of >550 bisphosphine ligands was constructed, and a designer chemical space mapping technique was established. The protocol used classification methods to identify active catalysts, followed by linear regression to model reaction selectivity. This led to the prediction and validation of significantly improved ligands for all reaction outputs, suggesting a general strategy that can be readily implemented for reaction optimizations where performance is controlled by bisphosphine ligands.
An efficient asymmetric synthesis of a potent KRAS G12C
covalent
inhibitor, GDC-6036 (1), is reported. The synthesis features
a highly atroposelective Negishi coupling to construct the key C–C
bond between two highly functionalized pyridine and quinazoline moieties
by employing a Pd/Walphos catalytic system. Statistical modeling by
comparing computational descriptors of a range of Walphos chiral bisphosphine
ligands to a training set of experimental results was used to inform
the selection of the best ligand, W057-2, which afforded
the desired Negishi coupling product (
R
a
)-3 in excellent selectivity.
A subsequent telescoped reaction sequence of alkoxylation, global
deprotection, and acrylamide formation, followed by a final adipate
salt formation, furnished GDC-6036 (1) in 40% overall
yield from starting materials pyridine 5 and quinazoline 6.
High-throughput experimentation
and multivariate modeling allow identification of noncovalent interactions
(NCIs) in monoaryloxy-pyrrolide Mo imido alkylidene metathesis catalysts
prepared in situ as a key driver for high activity
in a representative metathesis reaction (homodimerization of 1-nonene).
Statistical univariate and multivariate modeling categorizes catalytic
data from 35 phenolic ligands into two groups, depending on the substitution
in the ortho position of the phenol ligand. The catalytic
activity descriptor TON1h correlates predominantly with
attractive NCIs when phenols bear ortho aryl substituents
and, conversely, with repulsive NCIs when the phenol has no aryl ortho substituents. Energetic span analysis is deployed
to relate the observed NCI and the cycloreversion metathesis step
such that aryloxide ligands with no ortho aryls mainly
impact the energy of metallacyclobutane intermediates (SP/TBP isomers),
whereas aryloxides with pendant ortho aryls influence
the transition state energy for the cycloreversion step. While the
electronic effects from the aryloxide ligands also play a role, our
work outlines how NCIs may be exploited for the design of improved
d0 metathesis catalysts.
A synthetic method for the palladium-catalyzed cyanation of aryl boronic acids using bench stable and nontoxic N-cyanosuccinimide has been developed. High-throughput experimentation facilitated the screen of 90 different ligands and the resultant statistical data analysis identified that ligand σ-donation, π-acidity and sterics are key drivers that govern yield. Categorization into three ligand groups -monophosphines, bisphosphines and miscellaneous -was performed before the analysis. For the monophosphines, the yield of the reaction increases for strong σ-donating, weak π-accepting ligands, with flexible pendant substituents. For the bisphosphines, the yield predominantly correlates with ligand lability. The applicability of the designed reaction to a wider substrate scope was investigated, showing good functional group tolerance in particular with boronic acids bearing electron-withdrawing substituents. This work outlines the development of a novel reaction, coupled with a fast and efficient workflow to gain understanding of the optimal ligand properties for the design of improved palladium cross-coupling catalysts.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.