Guided discovery of chemical reaction pathways with imposed activation

Abstract

Computational power and quantum chemical methods have improved immensely since computers were first applied to the study of reactivity, but the de novo prediction of chemical reactions has remained challenging. We show that complex reaction pathways can be efficiently predicted in a guided manner using chemical activation imposed by geometrical constraints of specific reactive modes, which we term imposed activation (IACTA). Our approach is demonstrated on realistic and challenging chemistry, such as a triple cyclization cascade involved in the total synthesis of a natural product, a water-mediated Michael addition, and several oxidative addition reactions of complex drug-like molecules. Notably and in contrast with traditional hand-guided computational chemistry calculations, our method requires minimal human involvement and no prior knowledge of the products or the associated mechanisms. We believe that IACTA will be a transformational tool to screen for chemical reactivity and to study both by-product formation and decomposition pathways in a guided way.

Fig. 1. Diagram of reaction search by conformer exploration with imposed activation. (a) Conformer search methods generate stable three-dimensional molecular structures, such as those shown for complex 1, composed of a molecule of (R)-2-iodobutane and ethoxide anion. (b) Our reaction prediction methodology consists of constraining a specific activation coordinate q^‡ to out-of-equilibrium values (vertical dotted arrows) and performing searches for activated conformers with the other coordinates unconstrained (solid horizontal arrows). This is demonstrated here on 1, using the carbon–iodine bond as q^‡ (green). Further increasing the carbon–iodine distance of the activated conformers yields reaction pathways to multiple products (compounds 2–8). Non-reacting hydrogen atoms are omitted for clarity.

Fig. 2. Epoxide-initiated β-keto ester-terminated triple cyclization reaction and the corresponding products and side-products predicted by IACTA. The polycyclization from 9 to 10 is a key step in a concise total synthesis of berkeleyone A described in the literature. We initiated an IACTA reaction search from protonated enol 9a using the epoxide C–O bond of the tertiary carbon indicated in green as the activating coordinate. Compounds 11–16 are a subset of the predicted side products.

Fig. 3. Selected transition structures found in the IACTA simulation of the epoxide-initiated reactions of berkeleyone A forming triple cyclization products. (a–d) Transition state guesses for the formation of 9b (a), 13 (b), 15 (c) and 16 (d). Bond-forming atoms are connected by thin lines annotated with their interatomic distances. Non-reacting hydrogen atoms are omitted for clarity.

Fig. 4. Water-catalyzed 1,4-addition reaction between acrolein and 4-methylthiophenol and some important side products predicted by IACTA. (a) Reactions were obtained by stretching the carbon–carbon double bond of an acrolein molecule (shown in green) in the presence of 4-methylthiophenol and six explicit water molecules. Predicted products are annotated with estimated activation energy (top) and reaction energy (bottom) at the GFN2-xTB level of theory. (b–e) Reaction pathways forming 17 include a direct proton transfer (b) and proton transfer shuttled by one (c), two (d), and four water molecules (e). Structures (b) and (c) are s-cis and structures (d) and (e) are s-trans, and all of them are compact conformations.

Fig. 5. Formation of oxidative addition complexes of drug-like substrates. (a) The reaction studied here is the formation of Pd(ii) complexes from ^tBuXPhosPd by oxidative addition. The aryl-halide bond is the activating coordinate which is depicted in green. (b) IACTA is performed for the ten drug-like, highly functionalized molecules 19–28.

Fig. 6. Superimposition of approximate transition structures obtained from IACTA at the GFN2-xTB level of theory with optimized transition state structures at the PBE-D3/def2-SVP level of theory for the formation of oxidative addition complexes of drug-like substrates. (a and b) Transition structures are shown for compounds (a) 19 and (b) 20. Notably, the GFN2-xTB transition structures are obtained directly from IACTA, without additional computational refinements.