Toward manufacturing drug substances under continuous flow conditions, homogeneous reaction systems are generally preferred to enable the process control. However, this approach toward the homogeneous reaction systems limits the implementation scope of continuous flow technology and may deviate from the ideal process because various reaction systems including heterogeneous reaction systems have to be handled in the recent pharmaceutical development. The author tried to establish the technological platform to enable the control of the various reaction systems under continuous flow conditions, which will truly contribute to the dissemination of continuous flow technology in the pharmaceutical development.

AI-based research has progressed rapidly with the widespread use of AI modelling tools and the increase in electronic data in chemistry. In the field of medicinal chemistry, AI has been used for several decades to improve the success rate of drug discovery, with a long history in in-silico screening, prediction of mechanism of action and physical properties such as solubility. The authors have developed AI models with the aim of building our own platform for AI-based drug discovery. In the field of synthetic organic chemistry, databases of chemical reactions collected from patents and literature are used to train AI models for proposing new reactions, predicting reaction yields, etc. In this review, I will introduce our AI models for drug discovery, which include molecular design, activity prediction and ligand binding pose prediction. As the latest trend of AI development in synthetic organic chemistry, recent studies on discovering new chemical reactions and predicting reaction yields are presented. In addition, the challenges of AI in drug discovery and synthetic organic chemistry are discussed for the further development of AI studies in chemistry.

2-Silylaryl triflates such as 2-(trimethylsilyl)phenyl triflate, known as Kobayashi reagent, have been widely used as efficient aryne precursors in organic synthesis, and various synthetic transformations have been developed based on this approach to date. On the other hand, synthesis of new organosilicon compounds by retaining the silicon substituents of 2-silylaryl triflates without forming arynes could be regarded as another attractive way of utilizing these compounds. In particular, silicon-containing cyclic compounds are expected to be useful for various applications such as biologically active substances and optoelectronic materials. In this regard, we have been actively engaged in developing new synthetic methods of silacycles by palladium-catalyzed reactions of 2-silylaryl triflates without the formation of aryne intermediates. As a result, we have succeeded in the selective synthesis of various silacycles through the activation of C-H and/or C-Si bonds as well as their intramolecular bond exchange process with C-Pd bonds that are generated during the catalytic reactions. The target compounds of our reactions include silicon-stereogenic dibenzosiloles, highly substituted benzosilacyclobutenes, silicon-stereogenic 5,10-dihydrophenazasilines, benzophenanthrosilines, benzofluorenosilepins, silabenzoxazines, and tetrahydrophenanthrosiloles. Herein an overview of these transformations is described along with their mechanistic aspects.

Boron catalysis enabled the chemoselective α-functionalization of carboxylic acids in the presence of other carbonyl groups such as ketones, esters, and amides. The reversible formation of a B-O covalent bond between the carboxy group and the boron catalyst increases the acidity of α-proton in carboxylic acids, thereby facilitating deprotonation to form enolates, diboron enediolates, using mild organic bases. As a result, Mannich reaction, aldol reaction, α-amination, and α-trifluoromethylthiolation were successfully developed including enantioselective reactions. Moreover, boron-palladium and boron-iridium hybrid catalysis have been developed to promote the asymmetric O to C migratory allylation of allyl esters. Photoirradiation of the in-situ generated diboron enediolates induces single-electron transfer from photoexcited diboron enediolates to electron-accepting allylsulfones, promoting radical reaction to provide α-allylation products.

Amine compounds are widely used in various fields and many synthetic methods have been developed. Reductive amination is an excellent one-step reaction to synthesize the corresponding amine compound from a carbonyl compound, but there are some problems such as the use of toxic and stoichiometric reductant or low catalytic activity. We challenged to develop new catalysts to solve these problems of reductive amination reactions. We examined various ligands and developed an Ir-PA complex having a picolinamidato ligand and an Ir-QN complex having a quinolinolato ligand. These catalysts exhibit high catalytic activity and functional group tolerance under mild transfer hydrogenation conditions using formic acid or formate as a hydrogen source and do not require a high-pressure vessel. We have also developed an asymmetric catalyst based on the knowledge of reductive amination catalysts. As a result of various tuning of the ligand structure, an iridium complex with a N-(2-picolyl)sulfonamidato ligand (Ir-PSA) was developed for the efficient synthesis of optically active amines from ketones by using an inexpensive amino alcohol as an auxiliary chiral group and nitrogen source. The asymmetric auxiliary groups used can be removed under mild oxidative conditions, and the catalytic reaction and deprotection can be performed in a one-pot process. Furthermore, this reaction can be adapted to α-keto carboxylic acids, enabling the synthesis of α-amino acids, including various unnatural α-amino acids. This article describes the properties of the developed catalysts and their reaction applications.

Superbasic carbanions such as phenyl anion and t-butyl anion have been commonly utilized in organic synthesis, typically prepared using stoichiometric organometallic reagents (e.g., alkyllithium, Grignard reagents). The activation of organosilicon compounds with Lewis base represented by Hosomi-Sakurai allylation is another attractive way to generate reactive carbanions in relatively mild conditions. While this reaction system has been extensively investigated so far, its extension to superbasic carbanions remains unexplored. This short review describes the recent advancements in catalytic methodologies for the production of superbasic carbanions from organosilicon compounds.