Novel strategies in asymmetric synthesis

Asymmetric catalysis is an evergreen research subject in chemistry1 since molecular chirality plays a key role in science and technology. Therefore, the discovery of efficient methods to achieve the preparation of enantiopure compounds has become a challenge for chemists. The synthesis of chiral molecules can be achieved by asymmetric metal catalyzed, organocatalyzed or enzymatic reactions. The field of metal catalysis has been established by the 2001 Nobel Prize awarded to K. B. Sharpless, R. Noyori and W. Knowles. Organocatalysis, on the other side, experienced a significant expansion and received increasing attention and recognition as witnessed by the exponential number of publications that appeared within the last 14 years2. It is now recognised as a versatile method for the synthesis of chiral compounds both in academia and in industry, as it features simple experimental procedures, mild and non-anhydrous reaction conditions and the usage of low-toxic catalysts, most of all commercially available3. Despite impressive advancement, researchers are still eagerly looking for new asymmetric catalyzed reactions, since only a fraction of the known chemical transformations present an asymmetric version which shows wide substrate generality. The organocatalytic activation of the substrate can be realized through the generation of covalent intermediates (catalysis via enamine, iminium ion or dienamine and SOMO catalysis) or through the formation of hydrogen-bonds between the substrate and the catalyst4. The development of chiral Brønsted acids such as ureas, thioureas, diols, phosphoric acids and chiral binaphtols to be employed as catalysts can be, in many cases, complex since the stereochemical control arises from weak interactions between the substrate and the catalyst. Nevertheless, these compounds have been successfully employed delivering the desired products in high yield and stereoselectivity. An issue met during reaction optimisation is the lack of a rational strategy usage. Generally, trial-and-error approaches and chemical intuition are employed. When the catalyst forms a covalent bond with the substrate the resulting adduct can be characterized by spectroscopic techniques and their geometry studied/calculated by computational chemistry. Predictive models can therefore be built. On the contrary, the construction of a reliable model employing non-covalent catalysis is much more difficult, considering that enantioselectivity arises from a minimal energy difference (just 2 kcal/mol for 90% ee) between two diastereomorphic transition states. If the reactions were to be studied by computational chemistry, the error introduced could be comparable or larger than the absolute value arising from the calculations. Several variables are to be taken into consideration, such as solvent interaction, the catalyst‘s molecular weight, usually in the range of 300-600 Dalton, and a transition state involving weak interactions (hydrogen bonding) between the catalyst and the substrate(s). In order to trackle this issue a rational approach is highly desirable. Currently, the preferred strategy is the One-Variable-at-a-Time (OVAT) approach. However for a non-covalent asymmetric reaction, variables (concentration, !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 1 a) “Comprehensive Asymmetric Catalysis.” (Eds. Jacobsen, Pfaltz, Yamamoto), Springer, 1999; b) “Catalytic Asymmetric Synthesis.” (Ed. Ojima), Wiley-VHC, 2000; c) Heitbaum, M.; Glorius, F.; Escher, I. Angew. Chem. In. Ed. 2006, 29, 4732. “Asymmetric Catalysis on Industrial Scale,” (Eds. Blaser, Schmidt), Wiley-VCH, 2004. 2 From a research on Scopus database (September 2014) more than 3200 articles deal with organocatalysis. 3 Melchiorre P.; Marigo M.; Carlone, A.; Bartoli, G. Angew. Chem. Int. Ed. 2008, 47, 6138. 4 a) Bertelsen, S.; Jørgensen K. A. Chem. Soc. Rev. 2009, 38, 217; b) Doyle, A. G.; Jacobsen E. N. Chem. Rev. 2007, 107, 5713.! 7 ! choice of the solvent, temperature, catalyst loading) are likely to strongly interact with one another. Therefore, the rational exploration of chemical space requires the investigation of a multidimensional surface, where chemical intuition can be of limited help. The exploration of chemical space by means of Design of Experiment (DoE) methodologies allows to have an overall picture of the transformation leading to optimisation in a rational way. Considering this aspects, the aim of my Ph-D project has been the development and application of strategies which were not exploited before in the field of non-covalent asymmetric catalysis. Two transformations, the aza-Michael addition of imides and the kinetic resolution of 4-substituted oxazinones, have been choosen as model reactions on which test the viability of the approaches we decided to apply. The search for the best performing catalyst, which is one of the most time and resource consuming tasks to overcome, can be accelerated by the synergistic use of multifunctional and cooperative (multiple) catalysis5. It will be shown in the asymmetric aza-Michael addition of imides that the combination of these two catalysis types is even a more powerful way to gain the reaction optimisation than the usage of a single approach. An other important aspect is the way how the optimisation process is carried on. Chemical intuition might indeed lead to discover unique reactivity and phenomenon, but a rational strategy is surely more reliable in order to deliver results. Choosing a rational approach through application of Design of Experiments techniques, we successfully optimised the aza-Michael addition of imides. In the second year of me Ph-D, I had the possibility to deepen my DoE knowledge being the principal actor of an international collaboration among the group of Prof. A. Berkessel from Köln University (Germany), and Dr. Reddy’s Laboratories, a pharmaceutical company based in Cambridge (UK), where I worked under the supervison of Dr. A. Carlone. We chose the kinetic resolution of 4-substituted oxazinones, developed by Berkessel group in 2005, as model reaction to show that this strategy could serve as standard protocol for kinetic resolution allowing to improve on results to a level which would not have been possible to achieve with any other approaches. This strategy has never been applied within academia to such transformation. The yield and the stereoselectivity of the chosen reaction have been greatly improved with respect to previous results. It should be stressed that this approach also allows a significant saving on time and resources, because only a small number of target experiments are actually run, and the maximum information is extracted from them. The main advantage is, however, the possibility to fully optimise a reaction in a rational way, in contrast to trial-and-error approaches.