基于高通量实验与机器学习的Buchwald-Hartwig反应区域选择性的研究

C-N键生成反应的研究在有机合成中是一个重要的领域。此类反应可以制备胺及其衍生物、含氮杂环等，在药物研发中具有很高的应用价值。Buchwald-Hartwig偶联反应是形成C-N键的经典反应之一，也是合成芳胺化合物的有力工具。但是，在Buchwald-Hartwig偶联反应中，当存在竞争性的胺组合与卤代芳烃同时发生偶联反应时，一般会产生两种C-N偶联产物，仅依靠化学经验通常难以实现对该反应区域选择性的精准预测。因此，通过设计和筛选反应条件，降低胺的竞争性，提高区域选择性，让偶联反应更加干净和高效在合成化学领域具有重要意义。针对这一目标，本文提出了一种解决方案，即通过高通量实验结合机器学习建立分类模型，实现对Buchwald-Hartwig竞争偶联反应区域选择性的预测，从而快速得到高区域选择性的条件。

通过两轮高通量实验筛选，3-氨基-4-甲基吡啶和4-氨基-3-氟吡啶，3-氨基-4-甲基吡啶和苯胺这两组胺组合被确定存在激烈竞争。随后，这两组胺组合中的每一组和5种不同的溴化物在94个条件下分别反应。我们收集了其中282个反应数据，通过机器学习建立分类模型，实现了对3-氨基-4-甲基吡啶和4-氨基-3-氟吡啶的N-芳基化产物生成倾向的预测，并得到高区域选择性的条件。本次研究表明通过高通量实验技术结合机器学习方法可以预测此类反应的区域选择性，指导实验的合成方向，降低实验成本。同时，本文为研究此类反应提供新的研究思路，为未来建立更加精准的预测模型，以预测竞争胺组合的选择性比值打下了良好基础。

[1] Beletskaya I P,Cheprakov A V. The Complementary Competitors: Palladium and Copper in C–N Cross-Coupling Reactions[J]. Organometallics, 2012, 31(22):7753-7808.
[2] Bhunia S, Pawar G G, Kumar S V, et al. Selected Copper-Based Reactions for C−N, C−O, C−S, and C−C Bond Formation[J]. Angewandte Chemie International Edition, 2017, 56(51):16136-16179.
[3] Desnoyer A N,Love J A. Recent advances in well-defined, late transition metal complexes that make and/or break C–N, C–O and C–S bonds[J]. Chemical Society Reviews, 2017, 46(1):197-238.
[4] Rotella D P. The Critical Role of Organic Chemistry in Drug Discovery[J]. ACS Chemical Neuroscience, 2016, 7(10):1315-1316.
[5] Grygorenko O O, Volochnyuk D M, Ryabukhin S V, et al. The Symbiotic Relationship Between Drug Discovery and Organic Chemistry[J]. Chemistry – A European Journal, 2020, 26(6):1196-1237.
[6] Heravi M M, Kheilkordi Z, Zadsirjan V, et al. Buchwald-Hartwig reaction: An overview[J]. Journal of Organometallic Chemistry, 2018, 861:17-104.
[7] Forero-Cortés P A,Haydl A M. The 25th Anniversary of the Buchwald–Hartwig Amination: Development, Applications, and Outlook[J]. Organic Process Research & Development, 2019, 23(8):1478-1483.
[8] Biscoe M R, Barder T E, Buchwald S L. Electronic Effects on the Selectivity of Pd-Catalyzed CN Bond-Forming Reactions Using Biarylphosphine Ligands: The Competitive Roles of Amine Binding and Acidity[J]. Angewandte Chemie International Edition, 2007, 46(38):7232-7235.
[9] Huang X, Anderson K W, Zim D, et al. Expanding Pd-Catalyzed C−N Bond-Forming Processes:  The First Amidation of Aryl Sulfonates, Aqueous Amination, and Complementarity with Cu-Catalyzed Reactions[J]. Journal of the American Chemical Society, 2003, 125(22):6653-6655.
[10] Buitrago Santanilla A, Regalado Erik L, Pereira T, et al. Nanomole-scale high-throughput chemistry for the synthesis of complex molecules[J]. Science, 2015, 347(6217):49-53.
[11] Tu N P, Dombrowski A W, Goshu G M, et al. High-Throughput Reaction Screening with Nanomoles of Solid Reagents Coated on Glass Beads[J]. Angewandte Chemie International Edition, 2019, 58(24):7987-7991.
[12] Meuwly M. Machine Learning for Chemical Reactions[J]. Chemical Reviews, 2021, 121(16):10218-10239.
[13] Reid J P,Sigman M S. Holistic prediction of enantioselectivity in asymmetric catalysis[J]. Nature, 2019, 571(7765):343-348.
[14] Zahrt Andrew F, Henle Jeremy J, Rose Brennan T, et al. Prediction of higher-selectivity catalysts by computer-driven workflow and machine learning[J]. Science, 2019, 363(6424):eaau5631.
[15] Hartwig J F,Paul F. Oxidative Addition of Aryl Bromide after Dissociation of Phosphine from a Two-Coordinate Palladium(0) Complex, Bis(tri-o-tolylphosphine)Palladium(0)[J]. Journal of the American Chemical Society, 1995, 117(19):5373-5374.
[16] Guram A S, Rennels R A, Buchwald S L. A Simple Catalytic Method for the Conversion of Aryl Bromides to Arylamines[J]. Angewandte Chemie International Edition in English, 1995, 34(12):1348-1350.
[17] Louie J,Hartwig J F. Palladium-catalyzed synthesis of arylamines from aryl halides. Mechanistic studies lead to coupling in the absence of tin reagents[J]. Tetrahedron Letters, 1995, 36(21):3609-3612.
[18] Dorel R, Grugel C P, Haydl A M. The Buchwald–Hartwig Amination After 25 Years[J]. Angewandte Chemie International Edition, 2019, 58(48):17118-17129.
[19] Surry D S,Buchwald S L. Dialkylbiaryl phosphines in Pd-catalyzed amination: a user's guide[J]. Chemical Science, 2011, 2(1):27-50.
[20] Job G E,Buchwald S L. Copper-Catalyzed Arylation of β-Amino Alcohols[J]. Organic Letters, 2002, 4(21):3703-3706.
[21] Altman R A, Hyde A M, Huang X, et al. Orthogonal Pd- and Cu-Based Catalyst Systems for C- and N-Arylation of Oxindoles[J]. Journal of the American Chemical Society, 2008, 130(29):9613-9620.
[22] Strieth-Kalthoff F, Sandfort F, Segler M H S, et al. Machine learning the ropes: principles, applications and directions in synthetic chemistry[J]. Chemical Society Reviews, 2020, 49(17):6154-6168.
[23] Machine Learning[J]. The Journal of Physical Chemistry C, 2018, 122(4):1889-1889.
[24] Weininger D. SMILES, a chemical language and information system. 1. Introduction to methodology and encoding rules[J]. J. Chem. Inf. Comput. Sci., 1988, 28:31-36.
[25] Heller S, McNaught A, Stein S, et al. InChI - the worldwide chemical structure identifier standard[J]. Journal of Cheminformatics, 2013, 5(1):7.
[26] Sandfort F, Strieth-Kalthoff F, Kühnemund M, et al. A Structure-Based Platform for Predicting Chemical Reactivity[J]. Chem, 2020, 6(6):1379-1390.
[27] Rogers D,Hahn M. Extended-Connectivity Fingerprints[J]. Journal of Chemical Information and Modeling, 2010, 50(5):742-754.
[28] Reid J P, Proctor R S J, Sigman M S, et al. Predictive Multivariate Linear Regression Analysis Guides Successful Catalytic Enantioselective Minisci Reactions of Diazines[J]. Journal of the American Chemical Society, 2019, 141(48):19178-19185.
[29] Ahneman D T, Estrada J G, Lin S, et al. Predicting reaction performance in C–N cross-coupling using machine learning[J]. Science, 2018, 360(6385):186.
[30] Rupp M, Tkatchenko A, Müller K-R, et al. Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning[J]. Physical Review Letters, 2012, 108(5):058301.
[31] Brethomé A V, Fletcher S P, Paton R S. Conformational Effects on Physical-Organic Descriptors: The Case of Sterimol Steric Parameters[J]. ACS Catalysis, 2019, 9(3):2313-2323.
[32] Harper K C, Bess E N, Sigman M S. Multidimensional steric parameters in the analysis of asymmetric catalytic reactions[J]. Nature Chemistry, 2012, 4(5):366-374.
[33] Graziano G. Fingerprints of molecular reactivity[J]. Nature Reviews Chemistry, 2020, 4(5):227-227.
[34] Wender P A,Miller B L. Synthesis at the molecular frontier[J]. Nature, 2009, 460(7252):197-201.
[35] Muratov E N, Bajorath J, Sheridan R P, et al. QSAR without borders[J]. Chemical Society Reviews, 2020, 49(11):3525-3564.
[36] Chan B. Use of Low-Cost Quantum Chemistry Procedures for Geometry Optimization and Vibrational Frequency Calculations: Determination of Frequency Scale Factors and Application to Reactions of Large Systems[J]. Journal of Chemical Theory and Computation, 2017, 13(12):6052-6060.
[37] Selekman J, Qiu J, Tran K, et al. High-Throughput Automation in Chemical Process Development[J]. Annual Review of Chemical and Biomolecular Engineering, 2017, 8.
[38] Macarron R, Banks M N, Bojanic D, et al. Impact of high-throughput screening in biomedical research[J]. Nat Rev Drug Discov, 2011, 10(3):188-95.
[39] McMullen J P,Jensen K F. Integrated microreactors for reaction automation: new approaches to reaction development[J]. Annual review of analytical chemistry, 2010, 3:19-42.
[40] Gaunt M J, Janey J M, Schultz D M, et al. Myths of high-throughput experimentation and automation in chemistry[J]. Chem, 2021, 7(9):2259-2260.
[41] Krska S W, DiRocco D A, Dreher S D, et al. The Evolution of Chemical High-Throughput Experimentation To Address Challenging Problems in Pharmaceutical Synthesis[J]. Accounts of Chemical Research, 2017, 50(12):2976-2985.
[42] Takáts Z, Wiseman Justin M, Gologan B, et al. Mass Spectrometry Sampling Under Ambient Conditions with Desorption Electrospray Ionization[J]. Science, 2004, 306(5695):471-473.
[43] Ingoglia B T, Wagen C C, Buchwald S L. Biaryl monophosphine ligands in palladium-catalyzed C–N coupling: An updated User's guide[J]. Tetrahedron, 2019, 75(32):4199-4211.
[44] Huang F-D, Xu C, Lu D-D, et al. Pd-PEPPSI-IPentAn Promoted Deactivated Amination of Aryl Chlorides with Amines under Aerobic Conditions[J]. The Journal of Organic Chemistry, 2018, 83(16):9144-9155.
[45] Fors B P,Buchwald S L. A Multiligand Based Pd Catalyst for C−N Cross-Coupling Reactions[J]. Journal of the American Chemical Society, 2010, 132(45):15914-15917.
[46] Vo G D,Hartwig J F. Palladium-Catalyzed Coupling of Ammonia with Aryl Chlorides, Bromides, Iodides, and Sulfonates: A General Method for the Preparation of Primary Arylamines[J]. Journal of the American Chemical Society, 2009, 131(31):11049-11061.
[47] Marion N, Ecarnot E C, Navarro O, et al. (IPr)Pd(acac)Cl:  An Easily Synthesized, Efficient, and Versatile Precatalyst for C−N and C−C Bond Formation[J]. The Journal of Organic Chemistry, 2006, 71(10):3816-3821.
[48] Xie X, Zhang T Y, Zhang Z. Synthesis of Bulky and Electron-Rich MOP-type Ligands and Their Applications in Palladium-Catalyzed C−N Bond Formation[J]. The Journal of Organic Chemistry, 2006, 71(17):6522-6529.
[49] Fors B P, Krattiger P, Strieter E, et al. Water-Mediated Catalyst Preactivation: An Efficient Protocol for C−N Cross-Coupling Reactions[J]. Organic Letters, 2008, 10(16):3505-3508.
[50] Shen Q,Hartwig J F. [(CyPF-tBu)PdCl2]: An Air-Stable, One-Component, Highly Efficient Catalyst for Amination of Heteroaryl and Aryl Halides[J]. Organic Letters, 2008, 10(18):4109-4112.