An Efficient Synthesis of 2-CF3-3-Benzylindoles

The reaction of α-CF3-β-(2-nitroaryl) enamines with benzaldehydes afforded effectively α,β-diaryl-CF3-enones having nitro group. Subsequent reduction of nitro group by NH4HCO2-Pd/C system initiated intramolecular cyclization to give 2-CF3-3-benzylindoles. Target products can be prepared in up to quantitative yields. Broad synthetic scope of the reaction was shown. Probable mechanism of indole formation is proposed.

In continuation of the investigation of α,β-diaryl-CF3-enones chemistry, in this article, we report synthesis of 2-CF3-3-benzylindoles by reduction of nitro group in α-(2nitroaryl)-β-aryl-CF3-enones followed by intramolecular cyclization (Figure 2). In continuation of the investigation of α,β-diaryl-CF 3 -enones chemistry, in this article, we report synthesis of 2-CF 3 -3-benzylindoles by reduction of nitro group in α-(2-nitroaryl)β-aryl-CF 3 -enones followed by intramolecular cyclization (Figure 2). It should be noted, that 2-CF3-3-benzylindoles are quite a rare type of indoles. The approaches to the synthesis of these indoles were not studied systematically and have been not in the main focus of the publications. As a result, syntheses of only few 2-CF3-3benzylindoles were reported. Thus, prepared in three steps, N-[2-(1alkynyl)phenyl]trifluoroacetimidoyl iodides were transformed into desired indoles by the tin-radical promoted cyclization of N-[2-(1-alkynyl)phenyl]trifluoroacetimidoyl iodides as reported by Uneyama [65]. The copper-catalyzed C(sp2)-H trifluoromethylation of N,N-disubstituted hydrazones using the Togni's reagent followed by Fischer indole It should be noted, that 2-CF 3 -3-benzylindoles are quite a rare type of indoles. The approaches to the synthesis of these indoles were not studied systematically and have been not in the main focus of the publications. As a result, syntheses of only few 2-CF 3 -3-benzylindoles were reported. Thus, prepared in three steps, N-[2-(1-alkynyl)phenyl]trifluoroacetimidoyl iodides were transformed into desired indoles by the tin-radical promoted cyclization of N-[2-(1-alkynyl)phenyl]trifluoroacetimidoyl iodides as reported by Uneyama [65]. The copper-catalyzed C(sp2)-H trifluoromethylation of N,N-disubstituted hydrazones using the Togni's reagent followed by Fischer indole cyclization of CF 3hydrazones formed was described by Monteiro and Bouyssi [66]). N-Methylmorpholine mediated direct trifluoromethylation of 3-benzylindole with Umemoto's reagent was reported by Ma and Yu [67]. In spite of the mentioned methods allowed to prepare 2-CF 3indoles in good yields (59-64%), low atom efficiency and high price of some used reagent should be taken into account ( Figure 2).

Results
To start our investigation, we prepared a set of α-(2-nitroaryl)-β-aryl-CF 3 -enones using recently elaborated by us synthetic protocol [64]. Condensation of α-CF 3 -β-(2nitroaryl)enamines 1 with arylaldehydes 2 in acetic acid at 80-90 • C led to the corresponding α-(2-nitroaryl)-β-aryl-CF 3 -enones 3 in good to high yields. The reaction is very general, almost no limitations were found to give variety of such enones with a possibility to have different substituents in both aromatic rings. Moreover, some heterocyclic derivatives can be prepared as well (Scheme 1). Next, we investigated the reductive cyclization of ketone 3a in various conditions (Scheme 2). Firstly, we employed standard conditions of Leimgruber-Batcho [68] and Reissert [69] synthesis of indoles, which involve the reduction of nitro group followed by intramolecular cyclization of aniline formed. Thus, heating of ketone 3a using Fe-AcOH-H2O, Zn-EtOH-HCl and SnCl2•2H2O-EtOH systems led to the formation of a variety of hardly identifiable products, in which we were able to identify only 2-CF3-3benzylindole 4a and its acetoxy-derivative 5a (by 19 F NMR, Scheme 2). Better results were achieved when Zn-AcOH system was used. In this case, indoles 4a and 5a were isolated in 20% and 47% yield correspondingly ( Table 1, entry 1). Further heating of this reaction mixture with additional amount of Zn led to a partial transformation of acetoxy indole 5a into indole 4a (Table 1, entry 2). In Zn-AcOH-MeOH system methoxy-indol 6a became the main product, which was isolated in 77% yield (Table 1, entry 3). Further improvements in terms of chemoselectivity were made using catalytic hydrogenation on Pd/C in MeOH. Thus, reduction using H2 at room temperature or NH4HCO2 (hydrogen surrogate) at 65 °C afforded 2-CF3-3-benzylindole 4a in about 90% yield. In both cases methoxy-substituted indole 6a was formed as a byproduct in less than 1% yield (Table 1, entries 4,5). Ultimate selectivity of the reaction was achieved by the reduction with 5 Scheme 1. Synthesis of α-(2-nitroaryl)-β-aryl-CF 3 -enones 3.
Next, we investigated the reductive cyclization of ketone 3a in various conditions (Scheme 2). Firstly, we employed standard conditions of Leimgruber-Batcho [68] and Reissert [69] synthesis of indoles, which involve the reduction of nitro group followed by intramolecular cyclization of aniline formed. Thus, heating of ketone 3a using Fe-AcOH-H 2 O, Zn-EtOH-HCl and SnCl 2• 2H 2 O-EtOH systems led to the formation of a variety of hardly identifiable products, in which we were able to identify only 2-CF 3 -3-benzylindole 4a and its acetoxy-derivative 5a (by 19 F NMR, Scheme 2). Better results were achieved when Zn-AcOH system was used. In this case, indoles 4a and 5a were isolated in 20% and 47% yield correspondingly ( Table 1, entry 1). Further heating of this reaction mixture with additional amount of Zn led to a partial transformation of acetoxy indole 5a into indole 4a (Table 1, entry 2). In Zn-AcOH-MeOH system methoxy-indol 6a became the main product, which was isolated in 77% yield (Table 1, entry 3). Further improvements in terms of chemoselectivity were made using catalytic hydrogenation on Pd/C in MeOH.
Thus, reduction using H 2 at room temperature or NH 4 HCO 2 (hydrogen surrogate) at 65 • C afforded 2-CF 3 -3-benzylindole 4a in about 90% yield. In both cases methoxy-substituted indole 6a was formed as a byproduct in less than 1% yield (Table 1, entries 4,5). Ultimate selectivity of the reaction was achieved by the reduction with 5 equivalents of NH 4 HCO 2 on Pd/C in MeOH at room temperature. In this conditions 2-CF 3 -3-benzylindole 4a was isolated in almost quantitative yield while byproduct 6a was not formed at all (Table 1, entry 6). It is worth noting that the reaction with NH 4 HCO 2 (Table 1, entries 5,6) leads to a mixture of indole 4a and indolinol D, which structure is proved by NMR spectra of the reaction mixture. However, indolinol D eliminates water instantly followed by aromatization after addition of an acid (Schemes 2 and 3).  Careful analysis of results of experiments (Table 1) forced us to propose that the reaction can proceed via the formation of cyclic hemiaminal B (Scheme 3). To confirm our preposition, we performed the reduction of 3a with 3.3 equivalents of NH4HCO2 (the precise amount needed for NO2 reduction only). Heating of the reaction mixture for 1h at 60 °C led highly selectively to assembling of methoxy-substituted indole 6a in 86% yield (Table 1, entry 7). We have also found, that using THF as a solvent instead of methanol allowed to stop the reaction at the step of intermediate unsaturated indolinol B. Compound B is stable enough to be isolated in crude form (by evaporation of the solvent). The structure of B was confirmed by NMR and HRMS spectra (Scheme 3). It was also found that compound B eliminates water slowly at standing in CDCl3 solution (directly in NMR tube). Thus, NMR spectra of this solution measured after about a month (36 days) showed the complete transformation of B into C (Scheme 3). An attempt to perform acid catalyzed elimination of water from B in THF the solution and isolate C was failed. Thus, the addition of pTSA to the THF solution of B followed by evaporation of the solvent led to severe tarring immediately. However, the addition of pTSA to solution of B in methanol led to desired elimination of water followed by the conjugated addition of methanol to form methoxy-indole 5a ( Table 1, entry 8). Similarly, the addition of methanol to CDCl3 solution of C (obtained by standing in NMR tube, see above) led to the transformation of C into 5a (by 19 F NMR). So, we have successfully confirmed the mechanism of the reaction. Thus, reduction of the nitro group in indole 3a led to aniline A, which cyclizes to unsaturated indolinol B. Elimination of water from B afforded conjugated imine C, which is a strong Michael acceptor due to aromatization facilitating  Careful analysis of results of experiments (Table 1) forced us to propose that the reaction can proceed via the formation of cyclic hemiaminal B (Scheme 3). To confirm our preposition, we performed the reduction of 3a with 3.3 equivalents of NH4HCO2 (the precise amount needed for NO2 reduction only). Heating of the reaction mixture for 1h at 60 °C led highly selectively to assembling of methoxy-substituted indole 6a in 86% yield (Table 1, entry 7). We have also found, that using THF as a solvent instead of methanol allowed to stop the reaction at the step of intermediate unsaturated indolinol B. Compound B is stable enough to be isolated in crude form (by evaporation of the solvent). The structure of B was confirmed by NMR and HRMS spectra (Scheme 3). It was Careful analysis of results of experiments (Table 1) forced us to propose that the reaction can proceed via the formation of cyclic hemiaminal B (Scheme 3). To confirm our preposition, we performed the reduction of 3a with 3.3 equivalents of NH 4 HCO 2 (the precise amount needed for NO 2 reduction only). Heating of the reaction mixture for 1h at 60 • C led highly selectively to assembling of methoxy-substituted indole 6a in 86% yield (Table 1, entry 7). We have also found, that using THF as a solvent instead of methanol allowed to stop the reaction at the step of intermediate unsaturated indolinol B. Compound B is stable enough to be isolated in crude form (by evaporation of the solvent). The structure of B was confirmed by NMR and HRMS spectra (Scheme 3). It was also found that compound B eliminates water slowly at standing in CDCl 3 solution (directly in NMR tube). Thus, NMR spectra of this solution measured after about a month (36 days) showed the complete transformation of B into C (Scheme 3). An attempt to perform acid catalyzed elimination of water from B in THF the solution and isolate C was failed. Thus, the addition of pTSA to the THF solution of B followed by evaporation of the solvent led to severe tarring immediately. However, the addition of pTSA to solution of B in methanol led to desired elimination of water followed by the conjugated addition of methanol to form methoxy-indole 5a ( Table 1, entry 8). Similarly, the addition of methanol to CDCl 3 solution of C (obtained by standing in NMR tube, see above) led to the transformation of C into 5a (by 19 F NMR). So, we have successfully confirmed the mechanism of the reaction. Thus, reduction of the nitro group in indole 3a led to aniline A, which cyclizes to unsaturated indolinol B. Elimination of water from B afforded conjugated imine C, which is a strong Michael acceptor due to aromatization facilitating addition of nucleophiles. Hydrogenation of the double bond of B leads to saturated indolinol D. Elimination of water from D finalizes the process to afford indole 4a.
Next, we investigated the synthetic scope of the synthesis of CF 3 -indoles 4. Using the optimal reaction conditions, we performed a reduction of a number of ketones 3 to afford corresponding indoles 4 in high to quantitative yields (Scheme 4.). Next, we investigated the synthetic scope of the synthesis of CF3-indoles 4. Using the optimal reaction conditions, we performed a reduction of a number of ketones 3 to afford corresponding indoles 4 in high to quantitative yields (Scheme 4.). The reaction has a wide synthetic scope, allowing preparing indoles having both electron-donating and electron-withdrawing groups as well as bulky ortho-substituents and naphthyl fragment. It should be noted, that ketones 3j-l bearing the additional nitro groups were transformed into amino-substituted indoles 4j-l. These indoles are interesting objects for the further modifications at NH2-group to give promising derivatives in terms of drug design. In the case of bulky ketone 3o having 1-naphthyl substituent reduction in standard conditions (5 equivalents of NH4HCO2) led to the formation of admixture of methoxy-indole 6b (about 28%). Probably, the rate of hydrogenation of the double bond of unsaturated indolinol B is lower due to its steric hindrance and the reaction cannot be completed because of full decomposition of NH4HCO2 on Pd/C during the reaction course. Nevertheless, using of 8 equivalents of NH4HCO2 allowed to over- The reaction has a wide synthetic scope, allowing preparing indoles having both electron-donating and electron-withdrawing groups as well as bulky ortho-substituents and naphthyl fragment. It should be noted, that ketones 3j-l bearing the additional nitro groups were transformed into amino-substituted indoles 4j-l. These indoles are interesting objects for the further modifications at NH 2 -group to give promising derivatives in terms of drug design. In the case of bulky ketone 3o having 1-naphthyl substituent reduction in standard conditions (5 equivalents of NH 4 HCO 2 ) led to the formation of admixture of methoxy-indole 6b (about 28%). Probably, the rate of hydrogenation of the double bond of unsaturated indolinol B is lower due to its steric hindrance and the reaction cannot be completed because of full decomposition of NH 4 HCO 2 on Pd/C during the reaction course. Nevertheless, using of 8 equivalents of NH 4 HCO 2 allowed to overcome this obstacle to give selectively indole 4o in 87% yield. The reduction of ketones 3p and 3q having additional methoxy group in nitro-aryl fragment led to 5-and 8-methoxyindoles correspondingly.
Ketones 3r,s having heterocyclic substituents were also involved in the transformation. It should be noted that reduction of thiophene derivative 3r proceeded much more slowly compared to other substrates, which can be explained by poisoning of palladium by thiophene moiety [70]. Thus, attempt to perform the reaction in standard conditions led mostly to methoxy-indole 6c. However, increasing of the amount of NH 4 HCO 2 to 15 equivalents and prolongation of the reaction time to 5 days allowed to prepare desired indole 4r in good yield. Separation of admixture of 6c from target indole 4r was carried out by column chromatography. It should be noted, that it is one of few cases, then column chromatography was used for purification of the products (4l,r,s). All other indoles were isolated in pure form just after separation from the inorganic admixtures (Pd/C and NH 4 Cl). Due to the low stability of pyridine derived ketone 3s the reduction of this compound was performed without its isolation. An attempt to use NH 4 HCO 2 in AcOH afforded a complex mixture of products. However, using HCO 2 H instead of NH 4 HCO 2 showed much better results. Indole 4s having pyridine substituent was isolated in 21% yield from enamine 1a. Taking into account moderate yield at first step of the reaction sequence (30% for the formation of 3v) the yield at the reduction step can be estimated as 70% (Scheme 5). cases, then column chromatography was used for purification of the products (4l,r,s). All other indoles were isolated in pure form just after separation from the inorganic admixtures (Pd/C and NH4Cl). Due to the low stability of pyridine derived ketone 3s the reduction of this compound was performed without its isolation. An attempt to use NH4HCO2 in AcOH afforded a complex mixture of products. However, using HCO2H instead of NH4HCO2 showed much better results. Indole 4s having pyridine substituent was isolated in 21% yield from enamine 1a. Taking into account moderate yield at first step of the reaction sequence (30% for the formation of 3v) the yield at the reduction step can be estimated as 70% (Scheme 5).
Scheme 5. Reduction of ketones 3, having heterocyclic substituents. . Column chromatography was performed on silica gel. Melting points were determined on an Electrothermal 9100 apparatus (Electrothermal, Stone, Staffordshire, UK). All reagents were of reagent grade and were used as such or were distilled prior to use. Starting α-CF3-β-aryl enamines 1 were synthesized using previously reported procedures by the reaction with 10 equivalents of pyrrolidine in neat [71].

Materials and Methods
General Remarks 1 H, 13 C and 19 F NMR spectra were recorded on Bruker AVANCE 400 MHz spectrometer (Bruker Corp., Carlsruhe, Germany) in CD 3 CN and CDCl 3 at 400, 100 and 376 MHz respectively. Chemical shifts (δ) in ppm are reported with the use of the residual CHD 2 CN and chloroform signals (1.94 and 7.25 for 1 H and 1.30, 77.0 for 13 C) as internal reference. The 19 F chemical shifts were referenced to C 6 F 6 , (−162.9 ppm). ESI-MS spectra were measured with an Orbitrap Elite instrument (Thermo Fisher Scientific, Waltham, MA USA). TLC analysis was performed on "Merck 60 F 254 " plates (Merck, Darmstadt, Germany). Column chromatography was performed on silica gel. Melting points were determined on an Electrothermal 9100 apparatus (Electrothermal, Stone, Staffordshire, UK). All reagents were of reagent grade and were used as such or were distilled prior to use. Starting α-CF 3 -β-aryl enamines 1 were synthesized using previously reported procedures by the reaction with 10 equivalents of pyrrolidine in neat [71].

Synthesis of α-CF 3 -β-(2-nitroaryl)enamines 1 by the Reaction with Pyrrolidine in Neat (General Procedure).
One neck 25 mL round-bottomed flask was charged with dry pyrrolidine (8.5 mL, 100 mmol), cooled down to −18 • C and the corresponding styrene (10 mmol) was added in one portion with vigorous stirring. The reaction mixture was stirred at room temperature for 1-3 h until starting styrene was consumed (TLC or NMR monitoring). The excess of pyrrolidine was evaporated in a vacuum, the viscous residue was dissolved in CH 2 Cl 2 (50 mL), washed with 10% K 2 CO 3 solution (2 × 50 mL) and dried over Na 2 SO 4 . CH 2 Cl 2 was removed in vacuo to give crude enamine, which was used without further purification. For characterization data of enamines 1 see [64]. Synthesis of ketones 3 by the reactions of α-(trifluoromethyl)enamines with aromatic aldehydes (general procedure). One-necked 50-mL round bottom flask (or 12 mL vial) was charged with enamine 1 (5 mmol), aromatic aldehyde 2 (5.75 mmol) and glacial acetic acid (15 mL or 5 mL for reaction in the vial). Reaction mixture was kept at 80-90 • C (hotplate stirrer) under stirring for 6-10 h until consumption of aldehyde and corresponding benzyl ketone formed by the hydrolysis of enamine ( 1 H NMR control). Volatiles were evaporated in vacuo, the residue was dissolved in CH 2 Cl 2 (50 mL), washed with water (2 × 20 mL) and dried over Na 2 SO 4 . Volatiles were evaporated in vacuo, the residue was purified by column chromatography, using mixtures of hexane and CH 2 Cl 2 (3:1, 1:1), CH 2 Cl 2 , mixture of CH 2 Cl 2 and MeOH (100:1) as eluents. For characterization data of ketones 3 see [64].

Conclusions
In conclusion, we elaborated a novel two-step pathway towards 2-CF 3 -3-benzylindoles. Based on condensation of α-CF 3 -β-(2-nitroaryl) enamines with benzaldehydes the first step leads effectively to nitro-substituted α,β-diaryl-CF 3 -enones. The second one is a reduction of nitro group by NH 4 HCO 2 -Pd/C system followed by intramolecular cyclization to 2-CF 3 -3-benzylindoles in up to quantitative yields. High selectivity and the reaction yield of all steps are the distinct advantages of the method. Combining the experimental observations and the data of the NMR monitoring of the reaction mixtures, possible scheme of the transformation is evaluated and discussed.