Modular, Stereocontrolled Cβ–H/Cα–C Activation of Alkyl Carboxylic Acids Ming Shang,⊥,† Karla S. Feu,⊥,† Julien C. Vantourout,† Lisa M. Barton,† Heather L. Osswald,† Nobutaka Kato,§ Kerstin Gagaring,‡ Case W. McNamara,‡ Gang Chen,† Liang Hu,† Shengyang Ni,† Paula Fernández-Canelas,† Miao Chen,† Rohan R. Merchant,† Tian Qin,† Stuart L. Schreiber,ǁ,§ Bruno Melillo,*,†,§ Jin-Quan Yu,*,† Phil S. Baran*,† †Department of Chemistry, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037 ‡Calibr at The Scripps Research Institute, La Jolla, California 92037 §Chemical Biology and Therapeutics Science Program, Broad Institute, Cambridge, Massachusetts 02142 ǁDepartment of Chemistry and Chemical Biology, Harvard University, Cambridge, Massachusetts 02138 ABSTRACT: The union of two powerful transformations, directed C–H activation and decarboxylative cross-coupling, for the enantioselective synthesis of vicinally functionalized alkyl, carbocyclic, and heterocyclic compounds is described. Starting from simple carboxylic acid building blocks, this modular sequence exploits the residual directing group to access more than 50 scaffolds that would be otherwise extremely difficult to prepare. The tactical use of these two transformations accomplishes a formal vicinal difunctionalization of carbon centers in a way that is modular and thus amenable to rapid diversity incorporation. A simplification of routes to known preclinical drug candidates is presented along with the rapid diversification of an antimalarial compound series. Introduction block like 3-(3-bromophenyl) propanoic acid (4) could be It is becoming increasingly clear that practitioners are no longer employed as its precursor (Figure 1b). Within the privileged bound by notion that the pervasive C–H bond is unresponsive realm of saturated cyclic heterocycles, such logic could be to manipulation. In fact, the past two decades have seen a employed to rapidly access libraries of enantiopure scaffolds 9 dramatic increase in the use of C–H functionalization logic1 to that would be rather difficult to otherwise prepare (Figure 1c). assemble molecules.2 At this juncture it can be considered part For example, chiral pyrrolidines such as 5 have previously been of the mainstream in terms of the way students learn prepared through labor-intensive routes that require chiral retrosynthetic analysis.3 One of today’s workhorse C–H resolution and are not amenable to late-stage diversity 10 activation strategies involves the use of native functional groups incorporation. In stark contrast, a combination of C–H and 8 to direct and guide the site of functionalization.4 As the most radical cross coupling strategies could access the same ubiquitous functional group in organic chemistry, carboxylic architectures in fewer steps with exquisite control of acids and their derivatives have naturally risen to the top in stereochemistry and allow for diverse arenes to be installed at terms of directed C–H functionalization reactions available to the end of the route starting from simple commercial carboxylic the practitioner (Figure 1a).5 With over one thousand reports acids. The difficulty in preparing such seemingly simple now present for the use of such guided C–H activations1 in molecules is directly related to the challenge of “escaping the 11 synthesis, it is fair to say that this is staple reaction manifold for flatland” as articulated by many in the field. Herein we present modern organic synthesis. In this context, an exploration of a strategy for the net vicinal difunctionalization of cyclic and serial reactivity in which the lingering carboxylate group is acyclic systems via sequential functionalization initiated by employed in successive reaction has been limited in scope. Of stereoselective C–H activation followed by decarboxylative the few notable examples, nearly all are restricted to restoration cross-coupling (dCC) to form a variety C–C and C–X bonds, 12 13 14 15,16 17 of the parent carboxylic acid followed by classic reactions such including aryl, alkenyl, alkynyl, alkyl, and boryl. The as amidation and esterification (Figure 1a).6,7 The recent inherent modularity of this strategic advance allows access to a development of robust methods to decarboxylate such systems wealth of acyclic and cyclic systems, some of which have been and programmably replace them with new C–C and C–B bonds prepared before in more laborious ways. Application to a in a stereochemically predictable way, a formal type of C–C promising series of heretofore inaccessible azetidine-based activation, opens new opportunities to leverage the power of antimalarial agents is also disclosed. carboxylate-directed C–H activation chemistry. This Results and discussion 8 combination of one- and two-electron disconnections would Proof of concept. In order to obtain a first proof of concept for enable pathways to potentially valuable chiral acyclic building the underlying strategy, a set of enantiopure carboxylic acids, blocks such as 3 that could be considered “retrosynthetically prepared using Pd-catalyzed ligand enabled asymmetric sp3 C– opaque” as it is not immediately apparent how a simple building H activation, were employed (Figure 2).18 Recalling the suite of 1 19 DG dCC reactions developed over the last several years, the 1 R -I, L* Removal R2–M 1 succeeding reaction was found to be tolerant of a variety H NHArF R R2 pendant functional groups, including esters (19 and 20) and R O Ligand Controlled dCC R boronic esters (14 and 16), as well as both free (18) and silyl- ArF = Asymmetric B = Borylation N = Negishi [Chiral] capped alkynes (15), all of which can be used in yet another 4-(CF3)C6H4 C–H Activation G = Giese S = Suzuki reaction sequence on liberation. Importantly, pyridine and boron can be incorporated in the challenging context of strained Me O iPr carbocycles to access trans-disubstituted cyclobutane (23) and Et N N N cyclopropane (24) rings with high enantiomeric purity, the NHAc L3 latter of which could be conducted without a directing group. Et L* = AcHN AcHN The ligand-controlled nature of the C–H activation step allows Ph for easily tunable access to either desired enantiomer (i.e. 14 vs. tBu tBu 18 NMe2 21). AcHN Scope of saturated heterocycles. In acknowledgement of the tBu L1 tBu L2 L4 increasing demand for saturated heterocycles in drug TMS development,20 the efficacy of this reaction sequence was Br Br Me Me demonstrated using an array of commercial heterocyclic acids a Combination of C-H Activation with Decarboxylative Cross-Coupling O Bpin Me Bpin Me H DG Me Me Previous R1 10 10 Work OA* 14 [L1, B]: 15 [L1, N]: 16 [L1, B]: 17 [L1, N]: R O R2 R3 A* = NHPI, TCNHPI 46%, 90% ee 48%, 90% ee 85%, 88% ee 72%, 88% ee DG = OH, NHR etc. Me Me Me [ -C-H [Decarboxylative β >1000 papers Ref. 12-17 Activation] ca 60 reviews Cross-Coupling] CO2Bn R4 DG R4 = Aryl, R1 R4 Alkyl, Alkyne, R2 Me Me Me R O R3 BPin 10 10 10 18 [L1, N]: 19 [L1, G]: 20 [L1, N]: CO Et This Work 2 53%, 88% ee 52%, 88% ee 58%, 88% ee Me Me Ph [β-C-H 1 (1 e–) R H DG Activation] 23 [L3, S]: R2 45%, 94% ee R R O (2 e–) [Decarboxylative 2 F 1 Cross-Coupling] N b Asymmetric Synthesis of Hypothetical Med. Chem. Building Block Br Br Bpin med. chem. Me [enantioselective] F building block Me 1. Asymmetric H H O C-H Activation 21 [L2, N]: 22 [L2, N]: 24 [L4, B]: Br 39%, 84% ee 51%, 84% ee 33%, 92% ee 2e– OH 2. Decarboxylative Figure 2. Proof of concept for the Cβ–H/Cα–C activation strategy. Br alkylation 4 ee were measured after the C–H activation step. 3-(3-bromophenyl) [modular] propanoic acid 3 1e– [programmable] ($3/g) (both enantiomers of each are available) as shown in Figure 3. Beyond their importance as a framework for the celebrated β- c Synthesis of Biologically Active Saturated Heterocycles 21 [not modular] [modular] lactam therapeutic class, azetidines have garnered recent OMe [racemic] CO2Et [diastereoselective] interest as scaffolds in diversity-oriented synthesis, through C-H which a wide variety of constrained (i.e., bridged or fused) or 22 7 1,4-addition; Activation; densely-substituted nitrogenous ring systems can be accessed. 5 DCC resolution CO H + N 2 To this point, arylated intermediates derived from N-protected Cbz 6 azetidine-2-carboxylic acid (Figure 3a) can be rapidly MeO [Previous] [Current] 8 Steps 6 Steps commercial diversified under the Suzuki (29-32), Negishi (28), and Giese N N H2N N H (26 and 27) protocols. In a similar vein, elaboration of Cbz- 8 protected proline via directed C–H arylation at C3 followed by C-H Activation/DCC of Commercially Available Chiral Acids dCC furnished a range of enantiopure trans-1,2- CO2H O CO2H difunctionalized pyrrolidines bearing diverse substituents such N O CO2H as aryl (35 and 39), heteroaryl (36-38 and 40), cycloalkyl (33), H 10 N CO2H N R N CO2H and alkenyl (34) functionalities (Figure 3b). Substrate 45 is of 9 N CO2H Boc Cbz R = Boc (11) 12 Cbz ent-6 13 particular note as prior routes to access such scaffolds involved [>50 examples] [modular] [regioselective] [diastereoselective] early-stage incorporation of methyl and phenyl substituents and a tedious separation of diastereomers.23 Figure 3c features the Figure 1. Introduction to the modular, stereocontrolled Cβ–H/Cα– C activation of alkyl carboxylic acids. same titular sequence as applied to a tetrahydrofuran core to furnish a variety of enantiopure THF-based building blocks (46- 52). Among the privileged N-heterocyclic fragments, piperidine 2 R1 R1 R1 H [>40 examples] N H 1 DG 2 R -I R –M BocN 2