[2+2+2] Annulation of N-(1-Naphthyl)acetamide with Two Alkynoates via Cleavage of Adjacent C–H and C–N Bonds Catalyzed by an Electron-Deficient Rhodium(III) Complex

It has been established that an electron-deficient cationic CpE-rhodium(III) complex catalyzes the non-oxidative [2+2+2] annulation of N-(1-naphthyl)acetamide with two alkynoates via cleavage of the adjacent C–H and C–N bonds to give densely substituted phenanthrenes under mild conditions (at 40 °C under air). In this reaction, a dearomatized spiro compound was isolated, which may support the formation of a cationic spiro rhodacycle intermediate in the catalytic cycle. The use of N-(1-naphthyl)acetamide in place of acetanilide switched the reaction pathway from the oxidative [2+2+2] annulation-lactamization via C–H/C–H cleavage to the non-oxidative [2+2+2] annulation via C–H/C–N cleavage. This chemoselectivity switch may arise from stabilization of the carbocation in the above cationic spiro rhodacycle by the neighboring phenyl and acetylamino groups, resulting in the nucleophilic C–C bond formation followed by β-nitrogen elimination.


Results
In the course of our study of the cationic Cp E -rhodium(III) complex-catalyzed oxidative tandem [2+2+2] annulation-lactamization of acetanilides with two alkynoates, leading to banzo[cd]indolones, the reaction of 2-methyl acetanilide 2b and ethyl 2-butynoates (3a) was examined. As already shown in Scheme 1, the expected banzo[cd]indolones 4b were generated after treatment with KOH in ethanol [23]. Surprisingly, the use of N-(1-naphthyl)acetamide (2c) in place of 2b failed to afford the expected naphtho[cd]indolone 4ca. Instead, densely substituted phenanthrene 5ca was generated as a major product along with a mixture of the corresponding regioisomeric [3+2] annulation products 6ca/6ca (Scheme 2). In addition to the above products, unidentified oligomerization products derived from 2c and 3a were generated as by-products. Then the screening of the reaction conditions and the acyl groups on the nitrogen of 1-aminonaphthalene was conducted as shown in Table 1. Elevating the reaction temperature (40 • C) slightly increased the yields of 5ca and 6ca/6ca (entry 2). Increasing the amount of 3a increased the yield of 5ca and decreased that of 6ca/6ca (entry 3). However, this increase was very small, therefore, the conditions of entry 2 were selected as the best conditions. With respect to the acyl groups on the nitrogen of the 1-aminonaphthalene moiety, electron-poor N-(1-naphthyl)amide 2d, possessing the highly acidic amide proton, was tested in place of 2c, while no reaction was observed even at 80 • C (entry 4). Sterically demanding N-(1-naphthyl)amide 2e was also tested with expectation for acceleration of reductive elimination. Unfortunately, the use of 2e significantly increased the yield of not the [2+2+2] annulation product 5ea but the [3+2] annulation products 6ea/6ea (entry 5). Finally, the reaction was conducted using the Cp*-rhodium(III) complex instead of the Cp E -rhodium(III) complex 1, which resulted in significant decrease of the yields of 5ca and 6ca/6ca (entry 6). The scope of the present cationic Cp E -rhodium(III) complex-catalyzed [2+2+2] annulation is shown in Table 2. As with our previously reported cationic Cp E -rhodium(III) complex-catalyzed oxidative tandem [2+2+2] annulation-lactamization of acetanilides with two alkynoates, leading to banzo[cd]indolones, a wide variety of primary alkyl-substituted alkynoates 3a-e reacted with N-(1-naphthyl)acetamide (2c) to give the corresponding [2+2+2] annulation products 5ca-ce (entries 1-5). However, phenyl-substituted alkynoate 3f failed to afford the corresponding [2+2+2] annulation product 5cf and gave the corresponding [3+2] annulation product 6cf in good yield with perfect regioselectivity (entry 6). This product switch is also the same as our previously reported tandem [2+2+2] annulation-lactamization of acetanilide (2a) with 3f [23]. Unfortunately, the use electron-deficient alkynes did not afford the corresponding annulation products at all. For example, the use of ethyl propiolate afforded a complex mixture of products and no reaction was observed when using ethyl 2-butynoate. The structure and regiochemistry of the [2+2+2] annulation product 5ce were unambiguously determined by the X-ray crystallographic analysis as shown in Figure 1.  The scope of the present cationic Cp E -rhodium(III) complex-catalyzed [2+2+2] annulation is shown in Table 2. As with our previously reported cationic Cp E -rhodium(III) complex-catalyzed oxidative tandem [2+2+2] annulation-lactamization of acetanilides with two alkynoates, leading to banzo[cd]indolones, a wide variety of primary alkyl-substituted alkynoates 3a-e reacted with N-(1-naphthyl)acetamide (2c) to give the corresponding [2+2+2] annulation products 5ca-ce (entries 1-5). However, phenyl-substituted alkynoate 3f failed to afford the corresponding [2+2+2] annulation product 5cf and gave the corresponding [3+2] annulation product 6cf′ in good yield with perfect regioselectivity (entry 6). This product switch is also the same as our previously reported tandem [2+2+2] annulation-lactamization of acetanilide (2a) with 3f [23]. Unfortunately, the use electron-deficient alkynes did not afford the corresponding annulation products at all. For example, the use of ethyl propiolate afforded a complex mixture of products and no reaction was observed when using ethyl 2-butynoate. The structure and regiochemistry of the [2+2+2] annulation product 5ce were unambiguously determined by the X-ray crystallographic analysis as shown in Figure 1.  Importantly, an interesting by-product was generated in the reaction of 2c and 3a as shown in Scheme 3. When using 3 equiv of 3a, a very small amount of dearomatized spiro compound 7ca was detected in the crude reaction mixture (3% yield). An isolable amount (13% yield) of 7ca was generated when using an excess amount of 3a. Scheme 3. Formation of dearomatized spiro compound 7ca. The yields were determined by 1 H NMR yield using C6Me6 as an internal standard.
Plausible mechanisms for the formation of 5ca, 6ca and 7ca from 2c and 3a are shown in Scheme 4. First, C-H bond cleavage of 2c by Cp E -rhodium(III) species A affords naphthylrhodium B. Next, insertion of 3a into B gives alkenylrhodium C. The subsequent insertion of 3a into C gives dienylrhodium D. The second alkyne insertion may not proceed in the case of sterically demanding phenyl-substituted alkynoate 3f, thus generating [3+2] annulation product 6cf′ (Table 2, entry 6). Electrophilic metalation of the electron-rich 1-aminonaphthalene ring produces cationic spiro rhodacycle E, in which the carbocation is stabilized by the neighboring phenyl and acetylamino groups. Nucleophilic attack of the dienylrhodium moiety to the electrophilic 1-position of the 1-aminonaphthalene ring gives π-pentadienyl complex F [29]. Importantly, nucleophilic attack of the dienylrhodium moiety to the 3-position of the 1-aminonaphthalene ring in spiro rhodacycle E′, leading to π-pentadienyl complex F′, would be unfavorable due to the unstable dearomatized quinodimethane structures in E′ and F′. β-Nitrogen elimination [30,31] from intermediate F affords 5ca and the catalytically active rhodium(III) species A. However, the copper(II) co-catalyst was necessary in order to reoxidize rhodium(I) species generated through the competing oxidative [3+2] annulation giving 6ca. In this reaction, dearomatized spiro compound 7ca was isolated as a by-product. This compound may be generated by reductive elimination and deprotonation from Importantly, an interesting by-product was generated in the reaction of 2c and 3a as shown in Scheme 3. When using 3 equiv of 3a, a very small amount of dearomatized spiro compound 7ca was detected in the crude reaction mixture (3% yield). An isolable amount (13% yield) of 7ca was generated when using an excess amount of 3a. Importantly, an interesting by-product was generated in the reaction of 2c and 3a as shown in Scheme 3. When using 3 equiv of 3a, a very small amount of dearomatized spiro compound 7ca was detected in the crude reaction mixture (3% yield). An isolable amount (13% yield) of 7ca was generated when using an excess amount of 3a. Scheme 3. Formation of dearomatized spiro compound 7ca. The yields were determined by 1 H NMR yield using C6Me6 as an internal standard.
Plausible mechanisms for the formation of 5ca, 6ca and 7ca from 2c and 3a are shown in Scheme 4. First, C-H bond cleavage of 2c by Cp E -rhodium(III) species A affords naphthylrhodium B. Next, insertion of 3a into B gives alkenylrhodium C. The subsequent insertion of 3a into C gives dienylrhodium D. The second alkyne insertion may not proceed in the case of sterically demanding phenyl-substituted alkynoate 3f, thus generating [3+2] annulation product 6cf′ (Table 2, entry 6). Electrophilic metalation of the electron-rich 1-aminonaphthalene ring produces cationic spiro rhodacycle E, in which the carbocation is stabilized by the neighboring phenyl and acetylamino groups. Nucleophilic attack of the dienylrhodium moiety to the electrophilic 1-position of the 1-aminonaphthalene ring gives π-pentadienyl complex F [29]. Importantly, nucleophilic attack of the dienylrhodium moiety to the 3-position of the 1-aminonaphthalene ring in spiro rhodacycle E′, leading to π-pentadienyl complex F′, would be unfavorable due to the unstable dearomatized quinodimethane structures in E′ and F′. β-Nitrogen elimination [30,31] from intermediate F affords 5ca and the catalytically active rhodium(III) species A. However, the copper(II) co-catalyst was necessary in order to reoxidize rhodium(I) species generated through the competing oxidative [3+2] annulation giving 6ca. In this reaction, dearomatized spiro compound 7ca was isolated as a by-product. This compound may be generated by reductive elimination and deprotonation from Scheme 3. Formation of dearomatized spiro compound 7ca. The yields were determined by 1 H NMR yield using C 6 Me 6 as an internal standard.
Plausible mechanisms for the formation of 5ca, 6ca and 7ca from 2c and 3a are shown in Scheme 4. First, C-H bond cleavage of 2c by Cp E -rhodium(III) species A affords naphthylrhodium B. Next, insertion of 3a into B gives alkenylrhodium C. The subsequent insertion of 3a into C gives dienylrhodium D. The second alkyne insertion may not proceed in the case of sterically demanding phenyl-substituted alkynoate 3f, thus generating [3+2] annulation product 6cf (Table 2, entry 6). Electrophilic metalation of the electron-rich 1-aminonaphthalene ring produces cationic spiro rhodacycle E, in which the carbocation is stabilized by the neighboring phenyl and acetylamino groups. Nucleophilic attack of the dienylrhodium moiety to the electrophilic 1-position of the 1-aminonaphthalene ring gives π-pentadienyl complex F [29]. Importantly, nucleophilic attack of the dienylrhodium moiety to the 3-position of the 1-aminonaphthalene ring in spiro rhodacycle E , leading to π-pentadienyl complex F , would be unfavorable due to the unstable dearomatized quinodimethane structures in E and F . β-Nitrogen elimination [30,31] from intermediate F affords 5ca and the catalytically active rhodium(III) species A. However, the copper(II) co-catalyst was necessary in order to reoxidize rhodium(I) species generated through the competing oxidative [3+2] annulation giving 6ca. In this reaction, dearomatized spiro compound 7ca was isolated as a by-product. This compound may be generated by reductive elimination and deprotonation from spiro rhodacycle E. Increasing the amount of 3a may facilitate the reductive elimination, which increased the yield of 7ca. This result may support the intermediacy of spiro rhodacycle E in the catalytic cycle. Interestingly, the regioselectivity of both the present Cp E -rhodium(III)-catalyzed [2+2+2] and [3+2] annulations of N-(1-naphthyl)acetamide (2c) is opposite to that of our previously reported Cp E -rhodium(III)-catalyzed [2+2+2] and [3+2] annulations of acetanilides that proceed presumably via alkenylrhodium G [23], although the reason is not clear at the present stage.
We considered that it is possible to regenerate spiro rhodacycle E by oxidative addition with a neutral rhodium(I) complex and protonation of 7ca. Therefore, 7ca was treated with acetic acid and an in situ generated neutral Cp E -rhodium(I) complex, prepared by the reaction of 1, AgNTf 2 , NaOAc, 2a and 3a. However, no reaction was observed, thus excluding the intermediacy of 7ca in the catalytic cycle (Scheme 5). annulations of acetanilides that proceed presumably via alkenylrhodium G [23], although the reason is not clear at the present stage. We considered that it is possible to regenerate spiro rhodacycle E by oxidative addition with a neutral rhodium(I) complex and protonation of 7ca. Therefore, 7ca was treated with acetic acid and an in situ generated neutral Cp E -rhodium(I) complex, prepared by the reaction of 1, AgNTf2, NaOAc, 2a and 3a. However, no reaction was observed, thus excluding the intermediacy of 7ca in the catalytic cycle (Scheme 5).

General Information
Anhydrous acetone (No.27,072-5) and CH2Cl2 (No. 130-02457) were obtained from Aldrich (St. Louis, MO, USA) and Wako (Osaka, Japan) and used as received. Solvents for the synthesis of substrates were dried over Molecular Sieves 4Å (Wako) prior to use. Anilides 2d [32] and 2e [33] were prepared according to the literature. Internal alkynes 3d [34] and 3e [35] were prepared according to the literature. All other reagents were obtained from commercial sources and used as received. 1 H and 13 C data were collected on a Bruker AVANCE III (Billerica, MA, USA) HD 400 (400 MHz) at ambient temperature. HRMS data were obtained on a Bruker micro TOF Focus II (Billerica, MA, USA). A single crystal X-ray diffraction measurement was made on Rigaku XtaLAB mini II diffractometer (Akishima, Japan) using graphite monochromated Mo-Kα radiation. All reactions were carried out in oven-dried glassware with magnetic stirring.

General Information
Anhydrous acetone (No. 27,072-5) and CH 2 Cl 2 (No. 130-02457) were obtained from Aldrich (St. Louis, MO, USA) and Wako (Osaka, Japan) and used as received. Solvents for the synthesis of substrates were dried over Molecular Sieves 4Å (Wako) prior to use. Anilides 2d [32] and 2e [33] were prepared according to the literature. Internal alkynes 3d [34] and 3e [35] were prepared according to the literature. All other reagents were obtained from commercial sources and used as received. 1 H and 13 C data were collected on a Bruker AVANCE III (Billerica, MA, USA) HD 400 (400 MHz) at ambient temperature. HRMS data were obtained on a Bruker micro TOF Focus II (Billerica, MA, USA). A single crystal X-ray diffraction measurement was made on Rigaku XtaLAB mini II diffractometer (Akishima, Japan) using graphite monochromated Mo-Kα radiation. All reactions were carried out in oven-dried glassware with magnetic stirring.

Conclusions
In summary, we have established that an electron-deficient cationic Cp E -rhodium(III) complex catalyzes the non-oxidative [2+2+2] annulation of N-(1-naphthyl)acetamide with two alkynoates via cleavage of the adjacent C-H and C-N bonds to give densely substituted phenanthrenes under mild conditions (at 40 • C under air). Importantly, a dearomatized spiro compound was isolated in this reaction, which may support the formation of a spiro rhodacycle intermediate in the catalytic cycle. The use of N-(1-naphthyl)acetamide in place of acetanilide switched the reaction pathway from the oxidative tandem [2+2+2] annulation-lactamization involving cleavage of adjacent two C-H bonds to the non-oxidative [2+2+2] annulation involving cleavage of the adjacent C-H and C-N bonds. This chemoselectivity switch may arise from stabilization of the carbocation in the above cationic spiro rhodacycle by the neighboring phenyl and acetylamino groups, resulting in nucleophilic attack of the dienylrhodium moiety to this carbocation followed by β-nitrogen elimination.
Supplementary Materials: The following are available online: 1 H and 13 C NMR spectra of 5ca-5ce, 6ca-6ce, 6ca -6cf and 7ca and crystal data and data collection parameters of 5ce.
Funding: This work was supported partly by Grants-in-Aid for Scientific Research (No. JP26102004), for Research Activity Start-up (No. 15H06201) and for Young Scientists (No. 17K14481) from Japan Society for the Promotion of Science (JSPS), Japan. We also thank Umicore for generous support in supplying the rhodium and palladium complexes.