Evaluation of reaction conditions. We began this investigation with PA-protected homoallylic amine 1a as the model substrate, phenylboronic acid and phenyl disulfide as coupling partners (Table 1). After systematic manipulation of the reaction parameters, the reaction in the presence of 20 mol% of NiBr2·DME and 3 equiv. K3PO4 at 100 °C afforded the desired three-component conjunctive cross-coupling product 2a with excellent regioselectivity in 87% isolated yield (entry 1). Using other nickel sources (Ni(COD)2 or NiBr2) led to diminished yield (entries 2, 3). Slightly lower conversion was observed when the catalyst loading was reduced to 15 mol% (entry 4). Phenylboronic esters were regularly used in alkene difunctionalization (entry 5), but PhB(OH)2 was more effective in our reaction system. Screening of the bases revealed that K3PO4 was optimal for this reaction (entries 6, 7). Replacement of the solvent, a mixture of DMF and methanol with a single solvent led to a substantial decrease in the yield (entries 8, 9). Use of tert-butanol as the solvent also resulted in decreased yield and with 15% hydroarylation byproduct (entry 10). Inferior results were obtained when conducting the reaction at 80 °C (entry 11).
Substrate scope. With the optimal conditions in hand, we proceeded to examine the scope of the nucleophilic coupling partners (Fig 2a). Arylboronic acids bearing electron-donating substituents in the ortho, meta, and para positions are found to perform well, furnishing the corresponding products in high yields (2b-2g). On the other hand, arylboronic acids bearing electron-withdrawing substituents showed relatively lower reactivity, but still resulted in moderate yields (2h-2m). The reaction showed excellent functional group tolerance and a range of reactive groups, such as bromo (2k), aldehyde (2n), ketone (2o), alkenyl (2p), free hydroxyl (2q), and phenolic hydroxyl (2r), were well tolerated under optimized conditions. It merits mention that heterocycles frequently found in medicinally relevant molecules, such as furan (2s) and thiophene (2t), were also compatible with the current protocol. In addition, alkenylboronic acid worked well to deliver the alkenylthiolation product in 61% yield without isomerization of the alkene geometry.
Subsequently, we studied the electrophile scope with respect to disulfides under the standard conditions (Fig 2b). To our delight, both electron-rich and electron-deficient diaryl disulfides reacted well to give the corresponding thioethers in good to excellent yields and regioselectivities. Notably, sterically hindered ortho-substituted substrates are also excellent reaction partners (2x and 2ab). A series of potential coupling motifs, including bromo (2af), ester (2ah), secondary amide (2ai), cyano (2aj) and allyloxy (2ak) remain intact, showing both the excellent chemoselectivity of this protocol and the opportunity for further derivatization. Additionally, heteroaryl disulfides showed attenuated reactivity (2al-2am), giving products in a lower yield with large amounts of unreacted starting materials. Unfortunately, alkyl disulfides were ineffective under the reaction conditions. On the other hand, as analogues of disulfides, diphenyl diselenide was successfully converted into the desired arylselenolation product (2an) in 65% yield.
Next, we explored the scope of alkene substrates (Fig 2c). The terminal homoallylic amines containing α-branching were first tested under the standard conditions. The alkenes bearing alkyl- or aryl-substitution proceeded readily to afford the α,γ-dibranched amines in moderate yields with high diastereoselectivities (2ao-2ar). The trans-stereochemistry of substrates bearing two skipped stereocenters was confirmed by X-ray crystallography of the sulfone derived from the oxidation of 2ap. This stereochemical outcome arises from formation of the more trans-nickelacycle upon C-S bond reductive elimination and is similar to results from our previously reported analogous alkene difunctionalizations.37,43 Notably, arylboronic acid derived from estrone effectively underwent arylthiolation to give corresponding product (2ar) with diastereocontrol, which demonstrates the robustness of this protocol. We then explored challenging internal alkene substrates, Z or E isomers furnished the corresponding products in nearly 1:1 diastereomeric ratios. We hypothesize that reversible sulfur-Ni(III) homolysis/recombination process or observed partial stereoisomerization of the internal alkenes ablated the diastereoseletivity. In addition, isobutyl-, phenyl-, benzyl-, and thioether-substituted internal alkenes were competent substrates in this reaction.
Synthetic potential. This methodology is amenable to gram-scale operation, affording product 2a in 85% yield on a 5 mmol scale (Fig. 3). The PA directing group could be readily removed under basic conditions, giving the functionalized primary amine 3a in nearly quantitative yield after extraction without extensive purification by column chromatography. The PA auxiliary could also be easily modified from the products through C-H functionalization. For example, a Rh-catalyzed C-H olefination/cyclization of 2a yielded the pyrido pyrrolone derivative 3b with E-stereoisomer.53 Furthermore, the synthetic utility of the sulfide functionality was explored. Controlled oxidation of 2a with H2O2 and m-CPBA furnished corresponding sulfoxide 3c and sulfone 3d in 55% and 64% yields, respectively. Treatment of 2a with (diacetoxyiodo)benzene and ammonium carbonate led to the synthesis of the corresponding sulfoximine 3e in 76% yield.54
Mechanistic consideration. To elucidate the mechanism, we first conducted a radical scavenger experiment (Fig. 5a). As it turned out, the reaction efficiency was not affected by the addition of TEMPO or BHT, indicating that the carbosulfenylation likely did not involve a radical pathway or, alternatively, radical formation followed by a fast recombination with Ni within the solvent cage. Furthermore, a control experiment with in-situ prepared Ni(I) intermediate A43 was carried out. The intermediate A reacted with diphenyl disulfide at 80 °C for 6 h afforded 2a in 69% yield along with 30% hydroarylation product 5 (Fig. 5b). This result demonstrated that the Ni(I) species could be intercepted by diaryl disulfide to form the desired product.
A proposed NiI/NiIII catalytic cycle for the carbosulfenylation of unactivated alkenes was depicted in Fig. 6. Initially, the NiI species (I) was formed from a NiII precursor with a catalytic amount of arylboronic acid. It then underwent transmetalation with arylboronic acid to generate a nickel-aryl intermediate (II), followed by olefin migratory insertion to form the nickel-alkyl species (III). This 5,5-membered nickeladicycle stabilized by bidentate PA directing group preferentially underwent oxidative addition with the diaryl disulfide rather than undesired β-hydride elimination or protonation with the methanol, leading to the formation of NiIII species (IV). Finally, reductive elimination from the reactive high-valent nickel adduct and subsequent ligand exchange delivered the carbosulfenylation product and regenerated the active catalyst I.