a Reaction conditions: Unless otherwise mentioned, 0.2 mmol of 1a , 0.3 mmol of 2 , 0.04 mmol of catalyst, 1.5 mL of solvent, 100 °C, air, 15 h.b Isolated yields. c0.01 mmol of catalyst was used. d 80 °C.e 120 °C. f 24 h.g Under O2.
With the optimal reaction conditions in hand, we first studied the substrate scope of the [4 + 1] spirocyclization reaction by reacting various aromatic amides (1 ) with 2 (Scheme 2). The results demonstrated that substrate 1 with either an electron-donating group (EDG) such as methoxy, tert -butyl, methyl and phenyl or electron-withdrawing group (EWG) such as halides, trifluoromethyl as the R1 unit attached on the 4-pisition of the phenyl ring worked smoothly to give the desired products (3b -3i ) in good to excellent yields. In following study, it was found that substrates 1 bearing 3- methoxy, 3-methyl, 3-chloro and 3-bromo substituted phenyl scaffold coupled readily with 2 , delivering products3j -3m in good yields, respectively, at the less hindered position in a highly regioselective manner. The structure of3k was confirmed by single-crystal X-ray diffraction analysis (see the SI). Moreover, the reaction of substrates 1 having 2-methyl, 2-chloro or 2-bromo substituted phenyl ring with 2afforded the corresponding [4 + 1] spiroannulation products in diminished yields probably due to the steric perturbation of these substrates (3n -3p ). Additionally, the reaction ofN -(quinolin-8-yl)-2-naph- thamide andN -(quinolin-8-yl)-1-naphthamide with 2 also proceeded smoothly to generate 3q and 3r in 89% and 54% yields, respectively. Furthermore, we were pleased to find that 5-methoxyquinolin-8-amine and 5-chloroquinolin- 8-amine as directing groups were also compatible with this transformation to afford3s and 3t in good yields. Unfortunately, 2-methoxy-N -(quinolin-8-yl)benzamide andN -(quinolin-8-yl)furan -2-carboxamide failed to provide the desired product, presumably because of the coordination of the oxygen with the Co-catalyst. In addition, 2-diazocyclopentane-1,3-dione and 2-diazocyclohexane- 1,3-dione could not participate in this reaction under the standard conditions.
Scheme 2 Substrate scope for the synthesis of3 a ,b
a Reaction conditions: Unless otherwise mentioned, 0.2 mmol of 1a , 0.3 mmol of 2 , 0.04 mmol of Co(acac)2, 0.6 mmol of Ag2CO3, 1.5 mL of EtOAc, 100 °C, air, 15 h. b Isolated yields.c under Ar.
Next, we investigated the substrate scope of the [4 + 1 + 1] spirocyclization reaction for the preparation of spiro isochroman-3,1’-isoindolinone 4 (Scheme 3). Various aromatic amides 1 bearing diverse functional groups such as methoxy,tert -butyl, methyl, phenyl fluoro, chloro, or bromo, trifluoromethyl on the para -, meta -, or ortho -site of phenyl moiety smoothly underwent the [4 + 1 + 1] spirocyclization reactions, delivering products 4b -4p in moderate to good yields. Of note, 2-methyl, 2-fluoro, 2-chloro, or 2-bromo substituted N -(quinolin-8-yl)benzamides reacted with 2under standard conditions for the formation of 4a , furnishing the corresponding products 4m -4p in 58-70% yields. We assumed that under the conditions of generating 4 , the solvent CH3CN could replace quinolin-8-amine serve as a ligand for the Co-catalyst, reducing the influence of steric hindrance effect. To our delight, the developed methodology is also applicable forN -(quinolin-8-yl)-1-naphthamide,N -(5-methoxyquinolin-8-yl) benzamide andN -(5-chloroquinolin-8-yl)benzamide giving corresponding annulated product 4q -4s in good yields.
In order to get more insights into the reaction mechanism, a series of control experiments were conducted (Scheme 4). First, deuterium incorporation experiments were performed by treating 1a with CD3OD or D2O under the standard conditions for 3a , and the results showed that no H/D exchange was detected. Furthermore, the same results were observed when treating1a and 2 with CD3OD or D2O under the standard conditions. These results indicated that the C−H bond cleavage is irreversible (Scheme 4a). Second, the intermolecular competition reaction of the 1:1
Scheme 3 Substrate scope for the synthesis of4 a ,b
a Reaction conditions: 0.2 mmol of 1a , 0.3 mmol of 2a , 0.04 mmol of Co(acac)2, 0.6 mmol of Ag2CO3, 1.5 mL of CH3CN, 100 °C, O2, 24 h.b Isolated yields.
mixture of methoxy substituted benzamide (1b ) and trifluoromethyl substituted benzamide (1i ) with 2 gave a mixture of 3b and 3i in 18% and 17% yields, respectively. This outcome suggests that electron-rich substrate is slightly favorable for this reaction than an electron-deficient one (Scheme 4b). Third, an intermolecular competitive reaction was performed by treating an equimolar mixture of 1a and1a-d5 with 2 , from which3a and 3a-d4 were gained in a ratio of 0.53:0.47 upon analyzing 1H NMR spectra, and competitive kinetic isotope effect (KIE) value of 1.13 was calculated, revealing that the C(sp2)−H bond activation might not be involved in the turnover-limiting step (Scheme 4c).
Furthermore, to testing the proposal that 4a might be formed through transformation of 3a , we performed a series of control experiments (Scheme 5). Firstly, 3a was subjected to the standard reaction conditions for the formation of 4a , affording4a in 87% yield. In contrast, under Ar atmosphere, 4awas not observed (Scheme 5a and 5b). Secondly, 4a was generated in 84% yield in the presence of 0.5 eq. Ag2CO3 and O2. However, no desired product was yielded in the absence silver salt (Scheme 5c and 5d). Finally, 3a could not be converted 4a under standard conditions in the presence of radical scavengers such as TEMPO and BHT (2,6-di-tert-butyl-4-methylphenol) (Scheme 5e). These results showed that 4a was formed by oxygen participation in Baeyer-Villiger oxidation of 3a promoted by Ag+.
On the basis of control experiments and known reports,[9-13] a plausible mechanism for the cobalt-catalyzed [4 + 1] and [4 + 1 + 1] spirocyclization reactions have been depicted in Scheme 6. At first, the coordination of the two nitrogen atom of 1a to a Co(III) complex generated via oxidant by Ag2CO3 forms the five-membered cobaltacycle intermediate I , which undergoes the concerted metalation-deprotonation pathway to give intermediateII . The coordination of 2 to the metal center inII followed by release of N2 affords the cobalt-carbene species intermediate III . Subsequently,
Scheme 4 Mechanistic studies (I)
Scheme 5 Mechanistic studies (II)
the migratory insertion yields intermediate IV , which undergoes reductive elimination to form the [4 + 1] spirocyclization product3a and delivers Co(I) complex, which is then reoxidized to regenerate the active Co(III) species. In another aspect, under the promotion of Ag+ and O2, the in situ formed 3a undergoes a Baeyer-Villiger oxidation to generate the product 4a .[13]
To demonstrate the synthetic utility of the current protocol, the large scale experiment and directing group removal experiment were performed. As shown in Scheme 7a and 7b, a mixture of 1a (3 mmol) and2a (4.5 mmol) was subjected to the aforementioned both standard reaction conditions, yielding the [4 + 1] and [4 + 1 + 1] spirocyclization products 3a and 4a in 80% and 62% yields, respectively. On the other hand, 3s and 4rcould be easily transformed into spiro[indene-2,1’-isoindoline]-1,3,3’-trione 5 and spiro[isochroman-3,1’-isoindoline]-1,3’,4-trione 6 through removing the 5-methoxyquinolin-8-amine directing group in moderate overall yields, respectively (Scheme 8).
Scheme 6 Proposed mechanism
Scheme 7 Enlarged scale synthesis of 3a and