Search for the optimal reaction conditions
Based on our mechanistic assumption depicted in Fig. 2C, we initially evaluated an array of bifunctional (NH, N)-ligands on the palladium-catalyzed double isocyanide insertion reaction by subjecting bromobenzene (1) and tert-butyl isocyanide (2) to PdCl2 (10 mol%), Cs2CO3 (1.0 equiv), H2O (100 μL) in DMSO (1.0 mL) under an Ar atmosphere at 90 ºC for 12 hours (Table 1, entries 1−7).116 The desired α-ketoamide product 3 was smoothly generated in 36% yield in the presence of 1,1,3,3-tetramethylguanidine (L1, TMG) (Table 1, entry 1). Disappointingly, benzimidamide (L2), propionimidamide (L3), 1,3-diphenylguanidine (L4), 1,3,5-triazine-2,4,6-triamine (L5) and 2-methylpyrimidine-4,6(1H,5H)-dione (L6) were ineffective for this transformation (Table 1, entries 2−6). To our surprise, 1,5,7-triazabicyclo[4.4.0]dec-5-ene (L7, TBD) was proven to be the best cocatalyst, offering 3 in 49% yield together with trace amount of the single isocyanide insertion product amide 4 (Table 1, entry 7). Inspired by these fascinating results, other reaction parameters, including palladium catalysts, bases, solvents, et al, were further investigated in the presence of TBD (Table 1, entries 8−18). Replacement of PdCl2 with Pd2(dba)3 resulted in raised 63% yield of 3 (Table 1, entry 12), while Pd(OAc)2, Pd(TFA)2, Pd(OPiv)2 and Pd(PPh3)4 deliver the product in slightly lower yields (Table 1, entries 8−11). However, screening of inorganic bases failed to improve conversion (Table 1, entries 13−15). A survey of solvents revealed that DMSO was always the best choice (Table 1, entry 12 vs entries 16−18), even trace amount of product was yielded by harnessing MeCN or PhMe as the reaction medium (Table 1, entries 17 and 18). To our delight, the yield of 3 promptly increased to 81% from 63% when the reaction was performed with 2.5 equiv of 2 and 2.0 equiv of Cs2CO3 (Table 1, entry 19). As expected, the yield of 3 reduced as the loading of Pd2(dba)3 decreases (Table 1, entry 20). Of particularly noted was that the selectivity of products was quickly switched when the reaction was performed under classic palladium-catalytic conditions, giving amide 4 in 74% yield, while α-ketoamide 3 was not observed (Table 1, entry 21).96
Scope of (hetero)aryl bromides
Having established the optimal reaction conditions for the formation of α-ketoamides, the generality of current protocol with regard to (hetero)aryl bromides was initially tested as outlined in Fig. 3. Whether the substituent group is an electron-donating or -withdrawing moieties at the ortho-, meta-, and para-position of bromoarenes, the reactions proceeded smoothly to deliver the desired α-ketoamides (5−20) in good yields (50−88%).117 In addition, this approach could be extended to disubstituted bromoarenes bearing electron-rich or electron-poor groups, affording the corresponding products 21−26 in good results. Given the importance of heteroarenes in pharmaceutical industry, the scope of this double insertion of isocyanides event toward heteroaryl bromides were further evaluated with the established methodology. Delightfully, α-ketoamides incorporating xanthone (27), quinoxaline (28), pyridine (29−32), quinoline (33) and isoquinoline (34) were smoothly accessed in up to 90% yield. Bromoarenes featuring a π-extended system also proved compatible to provide the corresponding α-ketoamides 35−37 in good results. Moreover, the sterically hindered 2-bromo-9,9'-spirobi[fluorene] and 4-bromo[2.2]paracyclophane were successfully applied to this TBD-switchable assembly platform, furnishing the desired products 38 in 80% yield and 39 in 83% yield, respectively.
Scope of (hetero)aryl sulfonates
In order to further explore the generality of current methodology, pseudohalides were investigated as depicted in Fig. 4.116 Notably, the double isocyanide insertion process appears to be laborious for (hetero)aryl sulfonates, in which a higher temperature (120 ºC) was required for the installation of desired α-ketoamides with low yields.
Scope of (hetero)aryl iodides
Subsequently, we explored the scope with respect to highly active iodoarenes for this TBD enabled Pd-catalyzed double insertion of isocyanides (Fig. 5).116 Incontrovertibly, aryl iodides with diversified functional groups reacted efficiently with isocyanides and H2O, giving the desired products (3, 5−7, 11−14, 16, 17, 19−22, 24, 40 and 43−49) in good to excellent yields within 1 h. It is worth noting that some sensitive functional groups, including -NHAc, -CN and -CO2Me, could also tolerate the current conditions, no observing further hydrolysates. The substrate scope was successfully extended to heteroaryl iodides, allowing for the corresponding α-ketoamides incorporated pyridine (29, 30 and 32), quinoline (33), indole (50), pyrazole (51), thiophene (52 and 53), benzo[b]thiophene (54), dibenzo[b,d]furan (55) and carbazole (56) in a highly efficient manner (in up to 98% yield). Although alkyl isocyanides, such as n-butyl isocyanide and 1-fluoro-2-(2-isocyanoethyl)benzene, were well compatible with the transformation to give 57 and 58 in good results, aryl isocyanides were found to be poor substrates under current catalytic conditions, and the corresponding products were not observed, supposedly due to the facile dimerization of aryl isocyanides.
Gram scale reactions
To exhibit the scalability of this protocol, three gram-scale reactions were performed as depicted in Table 2. The transformations of bromobenzene and iodobenzene under their individual reaction conditions were amenable to scale-up production of 3, albeit with loss of efficiency (Table 2, entries 1 and 2). Pleasingly, good result (1.50 g of 3 in 73% yield) was also gained with only 5 mol% of palladium catalyst (Table 2, entry 3).
Transformations of 3
After developing the novel strategy, we attempted to explore its utility for the synthesis of some valuable intermediates as depicted in Fig. 6. Compound 3 could be smoothly transformed by condensation with N2H4·H2O, hydrogenation reduction with NaBH4, reductive deoxygenation with N2H4·H2O and nucleophilic addition with PhMgBr, leading to 59 in 98% yield, 60 in 92% yield, 61 in 83% yield, and 62 in 94% yield, respectively. In addition, deprotection of the N-tertbutyl group using trifluoroacetic acid gave 63 in 60% yield, which could be further converted into pyrrol-2-one 65 through Cu-catalyzed C-N cross-coupling and Mo-catalyzed deoxygenative cross-coupling reactions. Moreover, oximation of 3 and subsequent hydrogenation reduction performed well to furnish 67 in good result. Surprisingly, unnatural amino acid 69 could be smoothly attained with 78% overall yield in two steps from 3 via successive hydrolysis and reductive amination.
Late-stage modification of pharmaceuticals and their derivatives
Its application in the late-stage modification of pharmaceuticals and their derivatives further demonstrated the inherent value of this approach. As outlined in Fig. 7, drug-like haloarenes derived from sulfadimethoxine, clofibrate, celecoxib, loratadine, δ-vitamin E and trametinib underwent the current reaction smoothly and furnished corresponding α-ketoamides 70−75 in good to excellent yields, with only one exception of 74, in this case, a slightly lower yield was obtained.
Synthesis of pharmaceutically active molecules
Considering the frequency of α-ketoamide moieties in pharmaceuticals, this protocol thus provides an alternative method for the applications in drug discovery and development. The ligand-switchable assembly platform was successfully applied to three de novo synthesis of pharmaceutically active molecules. As depicted in Fig. 8, oxerin receptor agonist (76), epoxide hydrolase inhibitor (77) and RARγ agonist (78) were smoothly forged in 3−4 steps form the safe, cheap, and commercially available feedstock chemicals.
Mechanistic investigation on the TBD enabled Pd-catalyzed double insertion of isocyanides
Mechanistic experiments were next carried out to probe the proposed reaction pathway. First, we found that cross-product 58 was not detected when 2-(3-fluorophenyl)ethan-1-amine was added to the model reaction (Fig. 9A). This finding indicates that acylpalladium species was not involved in current catalytic system. Compound 3 was not detected in the absence of TBD (Fig. 9B), and the reaction without Cs2CO3 led to only 9% yield of desired product (Fig. 9C), indicating that TBD was not acted as a base, but it is likely to participate in the reaction as a ligand. In order to verify this conjecture, control experiment with MTBD by methylation of the N−H group of TBD led to decreases in activity (Fig. 9D), probably because the steric hindrance effect or the methyl group eliminates the hydrogen bond with the imine moiety of imidoylpalladium(II) intermediate. Surprisingly, α-ketoimine amide 79 was attained in 58% yield by recrystallization (Fig. 9E).117 In addition, 18O-labeled α-ketoimine amide 80 was generated in 64% yield, which rationalized that the oxygen atom in amide moiety derived from H2O of the reaction system (Fig. 9F). To our delight, a linearly dicoordinated palladium complex 81 was isolated and its structure was determined by single crystal X-ray diffraction (Fig. 9G).117 Similarly, α-ketoimine amide 79 was obtained in 45% yield when imidoylpalladium complex 81 was subjected to the reaction (Fig. 9H). Treatment of α-ketoimine amide 79 by silica gel column chromatography utilizing a mixed solvent of ethyl acetate and petroleum ether (volume ratio 20:1) as eluant led to hydrolysis product 11 in 93% yield (Fig. 9I), suggesting that compounds 79 and 81 are possible intermediates for the formation of target product 11, and that the oxygen atom in ketone carbonyl group comes from the silica gel utilized in column chromatography. To probe the electronic effects on the TBD enabled Pd-catalyzed double insertion of isocyanides, seven reactions with bromoarenes bearing para-substituted groups (-OMe, -Me, -H, -F, -Cl, -CF3, -CN) under the optimal conditions were conducted. As shown in Fig. 9J, a linear Hammett plot (R2 = 0.961) with a positive ρ value of 1.235 was obtained, which indicated that the double insertion of isocyanides is clearly facilitated by electron withdrawing groups. Finally, the orders from the initial rates for five individual components [p-bromofluorobenzene, tert-butyl isocyanide, H2O, Pd2(dba)3 and TBD] in the reaction was surveyed (Fig. 9K). The reaction is zero-order relationship with respect to p-bromofluorobenzene, tert-butyl isocyanide and TBD, suggesting that the oxidative addition of haloarenes to Pd(0) species, migratory insertion of isocyanides and coordination of imidoylpalladium(II) intermediate with TBD are not the rate-limiting step. Remarkably, the kinetic experiment of H2O did not implicate a simple zero- or first-order reaction, but a saturation kinetics—a hydrolysis pre-equilibrium of Cs2CO3 was involved, which implied that the halogen exchange with OH- might be the rate-limiting step. Similarly, a pre-equilibrium process between the ligand exchange of Pd2(dba)3 and [Pd(CNtBu)2] determined by a saturation dependence on Pd2(dba)3 was observed, which is consistent with the mechanistic study on palladium-catalyzed imidoylative couplings using isocyanides as both substrates and ligands.118
Proposed mechanism
Based on the above experiments and related literatures,97−107,118 a revised mechanism toward the TBD enabled Pd-catalyzed double insertion of isocyanides was proposed (Fig. 10). Initially, [Pd(CNR)2] complex formed by the ligand exchange of Pd2(dba)3 with isocyanides (CNR). Next, imidoylpalladium(II) B was smoothly yielded through oxidative addition of [Pd(CNR)2] to haloarenes and migratory insertion of isocyanides. Subsequently, the coordination of intermediate B with TBD to form a chelating seven-membered ring transition state I (TS I), in which the N-H moiety is associated to the imine of B and the N(sp2) atom in TBD is coordinated with the Pd(II) center. And then, the second migratory insertion of isocyanides and followed by dissociation of TBD occurred to give α-ketoimine imidoylpalladium(II) D, which was trapped by OH- in situ generated from Cs2CO3 and H2O to deliver palladium(II) complex F. Lastly, α-ketoamides was formed through successive reductive elimination and amide-iminol tautomerization, followed by silica gel-assisted hydrolysis process, while Pd(0) species is regenerated into the next catalytic cycle.