Automated Synthesis, Quantum Chemistry and NMR in the Identification of Structurally Related Macrolide Natural Products

合成有机化学具有巨大的工业重要性，多年来，化学家们一直在，并且仍然在不断地寻找更有效和可靠的方法来构建有机分子，以满足高需求。自动化是数字化学未来的一部分，它与机器学习和计算机辅助反合成相结合，有可能极大地改变我们创造分子的方式。在过去的10年里，Aggarwal团队创造了基于硼的新反应和策略，现在已经成熟，可以转移到自动化平台上。我们已经证明，将这种同源化学与最先进的量子化学计算和核磁共振光谱学相结合，提供了一种协同作用，使我们能够设计、合成和阐明三维复杂天然产物的结构，例如鲍拉霉素a。我们现在的目标是解决更复杂分子的结构，抗真菌海洋天然产物caylobolide a，并使用自动化的力量来帮助其合成。carlobolide A的12个立体中心中只有5个在文献中使用Mosher的酯分析被分配，其余的不能用这种方法分辨。该分子的主要结构特征是1,3-和1,5-相关羟基。我们已经开发出了一种方法来创造这种基序的任何异构体，解决了不对称合成中一个长期存在的问题。我们还建立了核磁共振+计算程序，使我们能够探测柔性化合物的立体化学，如caylobolide A，基于量子化学（或机器学习）预测核磁共振参数，如质子和碳化学位移，3JHH， 3JCH和核Overhauser增强。这些方法对彼此接近的-OH取代的立体中心非常有效，但当立体中心距离较远时（如caylobolide a）效果较差，我们可能需要探索对较远距离相互作用敏感的核磁共振技术的应用，例如剩余的各向异性核磁共振参数。

Synthetic organic chemistry has a huge industrial importance and for many years, chemists have been, and are still, constantly in search for more efficient and robust methods for the construction of organic molecules to meet the high demand. Automation is one strand of a digital chemistry future, which, in combination with machine learning and computer assisted retrosynthesis has the potential to dramatically change how we create molecules. Over the last 10 years, the Aggarwal group have created new reactions and strategies based on boron homologations that is now ripe for transferring to an automation platform. We have shown that combining this homologation chemistry with state-of-the-art quantum chemical calculations and NMR spectroscopy provides a synergy that allows us to design, synthesise and elucidate the structure of 3-dimensionally complex natural products, such as baulamycin A. Our goal now is to address the structure of an even more complex molecule, the antifungal marine natural product caylobolide A, and to use the power of automation to aid in its synthesis. Only 5 of the 12 stereocentres of caylobolide A have been assigned in the literature using Mosher's ester analysis, with the remainder not being resolvable by this method. Key structural features of the molecule are 1,3- and 1,5-related hydroxyl groups. We have developed methodology to create any isomer of such motifs, solving a long-standing problem in asymmetric synthesis.We have also established NMR+computation procedures that allow us to probe the stereochemistry of flexible compounds, such as caylobolide A, based on quantum chemical (or machine learning) prediction of NMR parameters such as proton and carbon chemical shifts, 3JHH, 3JCH and nuclear Overhauser enhancements. These work very well for -OH substituted stereocentres that are close to each other, but are less effective when the stereocentres are a long distance apart (as in caylobolide A) and we may need to explore the application of NMR techniques that are sensitive to longer-range interactions, such as residual anisotropic NMR parameters.