Rhodium (II)-Catalyzed Synthesis of Tetracyclic 3,4-Fused Indoles and Dihydroindoles

: An e ﬃ cient synthetic method of tetracyclic 3,4-fused indoles and dihydroindoles via rhodium-catalyzed (3 + 2) cycloaddition of N -tosyl-4-(2-phenoxyphenyl)-1,2,3-triazole was described. The aromatized xanthene derivatives can be achieved in a one-pot synthesis starting from 1-ethynyl-2-phenoxybenzene. The xanthene-based fused heterocycles were considered as the valuable ﬂuorophore.


Introduction
Xanthene-based fluorescent dyes largely containing fluorescein and rhodamine have attracted continuous attention from researchers because of their good photophysical properties such as high absorption coefficient, high photostability, and high fluorescence quantum yield [1][2][3][4][5][6][7][8]. However, absorption and emission wavelengths of many xanthene derivatives are in the ultraviolet-visible light range below 600 nm, which makes them unsuitable for bioimaging in living systems [9]. An important modification to the dyes is the introduction of a fused aryl ring into the xanthene skeleton. This modification brings a remarkable bathochromic shift in excitation and emission wavelengths [10].

Results and Discussion
Firstly, the substituted phenylboronic acids were coupled with 2-iodophenol in the presence of Cu(OAc)2 to provide 1-iodo-2-phenoxybenzene 1. Iodine was converted to a trimethylsilylacetylene (TMSA) functional group using PdCl2(PPh3)2 and CuI as the catalysts, and then trimethylsilyl was removed in CH2Cl2 solution containing K2CO3 to give 1-ethynyl-2-phenoxybenzene 3 [48]. Subsequently, the key intermediate N-tosyl-4-(2-phenoxyphenyl)-1,2,3-triazole 4 was obtained via (3 + 2) cycloaddition of 3 with tosyl azide in the presence of CuTc. When the triazole 4 reacted with rhodium(II) octanoate dimer in toluene at 80 °C for 4 h, the desired tetracyclic compound 5 was obtained after chromatography purification. The results delineated the scope of the (3+2) annulation reaction as shown in Scheme 1. Substrates possessing the electron-withdrawing groups smoothly reacted, and the corresponding products 5c-5f were isolated in yields ranging from 84% to 92%. The unsubstituted and electron-donating group substituted substrates also successfully involved in the transformation with yields of 72% and 70%, respectively (5a and 5b). The stereochemistry of the tetracyclic products was analyzed by 1 H NMR data of 5a-5f. H atoms on the methenyl group showed relatively large coupling constants (J = 14.0-15.8 Hz). Narasaka K. et al. reported a series of cis-and trans-fused (4,5,6,7-η)-3a,7a-dihydro-3H-indoles, in which the J of the cis-fused isomer was much larger than the trans-one [49]. Similarly, the coupling constants of cis-fused indole derivatives in Murakami's report were also up to 14 Hz [42]. Thus, the compounds 5a-5f were considered as cis-fused isomers.

Results and Discussion
Firstly, the substituted phenylboronic acids were coupled with 2-iodophenol in the presence of Cu(OAc)2 to provide 1-iodo-2-phenoxybenzene 1. Iodine was converted to a trimethylsilylacetylene (TMSA) functional group using PdCl2(PPh3)2 and CuI as the catalysts, and then trimethylsilyl was removed in CH2Cl2 solution containing K2CO3 to give 1-ethynyl-2-phenoxybenzene 3 [48]. Subsequently, the key intermediate N-tosyl-4-(2-phenoxyphenyl)-1,2,3-triazole 4 was obtained via (3 + 2) cycloaddition of 3 with tosyl azide in the presence of CuTc. When the triazole 4 reacted with rhodium(II) octanoate dimer in toluene at 80 °C for 4 h, the desired tetracyclic compound 5 was obtained after chromatography purification. The results delineated the scope of the (3+2) annulation reaction as shown in Scheme 1. Substrates possessing the electron-withdrawing groups smoothly reacted, and the corresponding products 5c-5f were isolated in yields ranging from 84% to 92%. The unsubstituted and electron-donating group substituted substrates also successfully involved in the transformation with yields of 72% and 70%, respectively (5a and 5b). The stereochemistry of the tetracyclic products was analyzed by 1 H NMR data of 5a-5f. H atoms on the methenyl group showed relatively large coupling constants (J = 14.0-15.8 Hz). Narasaka K. et al. reported a series of cis-and trans-fused (4,5,6,7-η)-3a,7a-dihydro-3H-indoles, in which the J of the cis-fused isomer was much larger than the trans-one [49]. Similarly, the coupling constants of cis-fused indole derivatives in Murakami's report were also up to 14 Hz [42]. Thus, the compounds 5a-5f were considered as cis-fused isomers. We next speculated the possible mechanism for the production of the tetracyclic 3,4-fused dihydroindole 5 (Scheme 2). An α-diazo imino A was formed by reversible tautomerization from the corresponding triazole 4. The intermediate α-imino rhodium carbene B was obtained with the release of N2 (g) when A rapidly reacted with rhodium (II). Subsequently, the intramolecular electrophilic reaction of B occurred to form the zwitterionic intermediate C. The anionic rhodium released bonding electrons, completing the second cyclization [42]. The intermediate A acts as a 1,3-dipole equivalent due to the nucleophilic character of the imino nitrogen. According to Davies report [47], alternatively, α-diazoimine A undergoes thermal decomposition to generate free carbene D, which could cyclopropanate an arene to give norcaradiene E. The intermediate E does not have the correct geometry to undergo a (3,5)-sigmatropic rearrangement to give 5.

Scheme 2. Proposed mechanistic pathways.
It is noted that it was a bit hard to isolate compound 5 after the reaction because we found these tetracyclic structures were easily oxidized in air, giving the corresponding aromatized pyrrole-fused xanthene derivative 6. According to Murakami's report, a special oxidant (such as MnO2) was needed to complete further oxidative aromatization [42]. However, in our case, thorough aromatization can be achieved by stirring in air for some hours. Thus, we subsequently tried to directly synthesize aromatized xanthene derivatives. We next speculated the possible mechanism for the production of the tetracyclic 3,4-fused dihydroindole 5 (Scheme 2). An α-diazo imino A was formed by reversible tautomerization from the corresponding triazole 4. The intermediate α-imino rhodium carbene B was obtained with the release of N 2 (g) when A rapidly reacted with rhodium (II). Subsequently, the intramolecular electrophilic reaction of B occurred to form the zwitterionic intermediate C. The anionic rhodium released bonding electrons, completing the second cyclization [42]. The intermediate A acts as a 1,3-dipole equivalent due to the nucleophilic character of the imino nitrogen. According to Davies report [47], alternatively, α-diazoimine A undergoes thermal decomposition to generate free carbene D, which could cyclopropanate an arene to give norcaradiene E. The intermediate E does not have the correct geometry to undergo a (3,5)-sigmatropic rearrangement to give 5. We next speculated the possible mechanism for the production of the tetracyclic 3,4-fused dihydroindole 5 (Scheme 2). An α-diazo imino A was formed by reversible tautomerization from the corresponding triazole 4. The intermediate α-imino rhodium carbene B was obtained with the release of N2 (g) when A rapidly reacted with rhodium (II). Subsequently, the intramolecular electrophilic reaction of B occurred to form the zwitterionic intermediate C. The anionic rhodium released bonding electrons, completing the second cyclization [42]. The intermediate A acts as a 1,3-dipole equivalent due to the nucleophilic character of the imino nitrogen. According to Davies report [47], alternatively, α-diazoimine A undergoes thermal decomposition to generate free carbene D, which could cyclopropanate an arene to give norcaradiene E. The intermediate E does not have the correct geometry to undergo a (3,5)-sigmatropic rearrangement to give 5.

Scheme 2. Proposed mechanistic pathways.
It is noted that it was a bit hard to isolate compound 5 after the reaction because we found these tetracyclic structures were easily oxidized in air, giving the corresponding aromatized pyrrole-fused xanthene derivative 6. According to Murakami's report, a special oxidant (such as MnO2) was needed to complete further oxidative aromatization [42]. However, in our case, thorough aromatization can be achieved by stirring in air for some hours. Thus, we subsequently tried to directly synthesize aromatized xanthene derivatives.

Scheme 2. Proposed mechanistic pathways.
It is noted that it was a bit hard to isolate compound 5 after the reaction because we found these tetracyclic structures were easily oxidized in air, giving the corresponding aromatized pyrrole-fused xanthene derivative 6. According to Murakami's report, a special oxidant (such as MnO 2 ) was needed to complete further oxidative aromatization [42]. However, in our case, thorough aromatization can be achieved by stirring in air for some hours. Thus, we subsequently tried to directly synthesize aromatized xanthene derivatives.
The construction of these pyrrole-fused xanthene compounds was successfully integrated into a one-pot synthesis directly from 1-ethynyl-2-phenoxybenzene 3 (Scheme 3). For example, 3g (1.0 equiv), Catalysts 2020, 10, 920 4 of 9 tosyl azide (1.0 equiv), CuTc (0.05 equiv), Rh 2 (oct) 4 (0.02 equiv), and toluene were mixed together in a flask. The above mixture was stirred at 25 • C for 12 h, during which 3g was converted to the corresponding triazole 4g. The mixture was then stirred at 80 • C for additional 4 h. After being cooled to room temperature, the mixture was further stirred for 4 h in air. Finally, preparative thin-layer chromatography was used to afford 6g in 73% yield based on 3g. The all-in-one-pot procedure showed that the catalysts and reagents requisite in each step barely interfered with each other.
N-methyl acridone derivatives [50]. The triazole intermediate was converted to acridone by rhodium catalysis via a single cycloaddition.
In addition, a special case of the rhodium catalysis process was obtained for the O-methoxy substrate. The (3+2) annulation reaction of N-tosyl-4-(2-phenoxyphenyl)-1,2,3-triazole 4m took place by Rh2(oct)4 in toluene under the same conditions. However, oxidative aromatization directly occurred during the process of annulation, and also the O-methoxy group was simultaneously removed, affording the aromatized pyrrole-fused xanthene 8 with 78% yield (Scheme 5). Due to the rotatable C-O bond in diaryl ether, the carbenoid carbon of B is electrophilic to react with methoxyl site. When intramolecular attack of the phenyl ring in the methoxyl site occurs to furnish the zwitterionic intermediate, methoxyl as a good leaving group could be removed. In the tested substrates, only O-methoxy-substituted triazole could be aromatized, retaining a tosyl group in the oxidation process. Scheme 3. One-pot synthesis starting from 1-ethynyl-2-phenoxybenzene. Scheme 3. One-pot synthesis starting from 1-ethynyl-2-phenoxybenzene.
The present reaction was also used to synthesize xanthone 7 from triazole diaryl ether 4, as shown in Scheme 4. When N,N-diethyl substituted 4k in toluene and was catalyzed by Rh 2 (oct) 4 at 80 • C, no pyrrole-fused xanthene skeleton was obtained. Instead, the corresponding xanthone 7k was obtained in 75% isolated yield. After the first cycloaddition to form intermediate C, the strong electron-donating substituent N(C 2 H 5 ) 2 weakened the electropositivity of the allyl position, which was unfavorable for the second cyclization with the imino nitrogen. The tosyl amide was further oxidized to a carbonyl group in air, affording a stable xanthone derivative. Similarly, the substrate 4l' with an intense electron-donor group diethyl amine also underwent a single cycloaddition reaction to give xanthone 7l' in 80% yield. Shen et al. reported a similar process for the synthesis of N-methyl acridone derivatives [50]. The triazole intermediate was converted to acridone by rhodium catalysis via a single cycloaddition.
In addition, a special case of the rhodium catalysis process was obtained for the O-methoxy substrate. The (3+2) annulation reaction of N-tosyl-4-(2-phenoxyphenyl)-1,2,3-triazole 4m took place by Rh 2 (oct) 4 in toluene under the same conditions. However, oxidative aromatization directly occurred during the process of annulation, and also the O-methoxy group was simultaneously removed, affording the aromatized pyrrole-fused xanthene 8 with 78% yield (Scheme 5). Due to the rotatable C-O bond in diaryl ether, the carbenoid carbon of B is electrophilic to react with methoxyl site. When intramolecular attack of the phenyl ring in the methoxyl site occurs to furnish the zwitterionic intermediate, methoxyl as a good leaving group could be removed. In the tested substrates, only O-methoxy-substituted triazole could be aromatized, retaining a tosyl group in the oxidation process. Encouraged by the straightforward synthetic pathway described above, we tried to investigate the fluorescent properties of the pyrrole-fused xanthene skeleton. The excitation (λex) and emission (λem) wavelengths of the typical structures were tested as shown in Table 1. The unaromatized 5a showed spectral characteristics that were comparable to rhodamine with λem = 443 nm and Φ = 0.31. The entirely aromatized structures via oxidation, such as 6g-6j, exhibited longer emission wavelengths at 471, 480, and 552 nm, respectively. When the substituent was an electron-withdrawing NO2 group, the fluorescent properties including emission wavelength and quantum yield remarkably increased. Two xanthone-based products, 7k and 7l', were also analyzed, giving λem = 426 nm and λem = 457 nm with acceptable quantum yields, respectively. These results indicated that the pyrrole-fused xanthene or imino-modified derivatives in pyrrole were used as a potential fluorophore to develop new applications.  Encouraged by the straightforward synthetic pathway described above, we tried to investigate the fluorescent properties of the pyrrole-fused xanthene skeleton. The excitation (λ ex ) and emission (λ em ) wavelengths of the typical structures were tested as shown in Table 1. The unaromatized 5a showed spectral characteristics that were comparable to rhodamine with λ em = 443 nm and Φ = 0.31. The entirely aromatized structures via oxidation, such as 6g-6j, exhibited longer emission wavelengths at 471, 480, and 552 nm, respectively. When the substituent was an electron-withdrawing NO 2 group, the fluorescent properties including emission wavelength and quantum yield remarkably increased. Two xanthone-based products, 7k and 7l', were also analyzed, giving λ em = 426 nm and λ em = 457 nm with acceptable quantum yields, respectively. These results indicated that the pyrrole-fused xanthene or imino-modified derivatives in pyrrole were used as a potential fluorophore to develop new applications.  Figure S1 for details of emission (λ em ) wavelength and quantum yield (Φ) for the typical compounds.

Materials
Unless specifically mentioned, all chemicals were purchased from Beijing Ouhe Technology Co. Ltd., Beijing, China, or J&K Scientific Ltd., Beijing, China, and used without further purification.
Compounds 4b-f were synthesized using a similar route according to 4a.

Tetracycle 5 Synthesis by Rh(II) Catalyst
Compound 4a (0.65 g, 1.66 mmol) and Rh 2 (oct) 4 (25.86 mg, 0.03 mmol) were dissolved in dry toluene (10 mL) in a Schlenk tube. The solution was heated for 4 h at 80 • C. Then, the mixture was evaporated under vacuum to give the crude product, which was purified by silica gel column chromatography hexane/ethyl acetate (20:1) to give 5a (72% yield). Compounds 5b-f were synthesized using a similar route according to 5a.

One-Pot Synthesis of 6 Starting from 1-ethynyl-2-phenoxybenzene
Compound 3g (0.8 g, 3.84 mmol), TsN 3 (0.76 g, 3.84 mmol), CuTc (36.62 mg, 0.19 mmol), and Rh 2 (oct) 4 (59.82 mg, 76.83 µmol) were dissolved in dry toluene (10 mL) in a Schlenk tube. The solution was stirred at room temperature for 12 h and then heated to 80 • C for 4 h. After being cooled to room temperature, the solution was further stirred at room temperature for 4 h in air. Then, the mixture was evaporated under vacuum to give the crude product, which was purified by silica gel column chromatography hexane/ethyl acetate (20:1) to give 6g (60% yield). Compounds 6h-j were synthesized using a similar route according to 6g.

Calculation of the Fluorescence Quantum Yield
The quantum yield of the fluorophore was calculated according to Equation (1): where ϕ s is the quantum yield of the standard, F is the area under the emission spectra, A is the absorbance at the excitation wavelength, and η is the refractive index of the solvent used. U subscript denotes unknown, and S means standard. Rhodamine B was chosen as the standard.

Conclusions
In conclusion, we have described the intramolecular (3+2) annulation of α-imino rhodium carbene complexes to construct tetracyclic 3,4-fused indoles and dihydroindoles. Of note is that the reaction is illustrated by its successful integration into a one-pot synthesis directly from 1-ethynyl-2-phenoxybenzene, giving a series of xanthene derivatives. Xanthenes are important in fluorescent dye and medicinal chemistry, and we think that the current approach is an appealing choice for the construction of molecular libraries for diversity-oriented synthesis.