Reaction of indoles with aromatic fluoromethyl ketones: an efficient synthesis of trifluoromethyl(indolyl)phenylmethanols using K2CO3/n-Bu4PBr in water

A new, mild and efficient protocol for the synthesis of trifluoromethyl(indolyl)phenylmethanols by the reaction of indoles with a variety of aromatic fluoromethyl ketones in the presence of K2CO3 (15 mol %) and n-Bu4PBr (15 mol %) in water. The desired products were obtained in good to excellent yields without requiring a column chromatographic purification. The reusability of the catalytic system and large-scale synthesis of indolyl(phenyl)methanols, which would further transform into biological active indole-derived compounds, are further advantages of this protocol.

Organofluorine compounds have attracted much attention due to their potential biological applications in medicinal and agricultural sciences. Introducing fluoro groups into organic molecules can dramatically influence their physiochemical and bio- logical properties in comparison with non-fluorinated analogs [18] (see compounds I and II in Figure 1). Many pharmaceuticals and agrochemicals developed in recent decades have either a fluorine atom or a trifluoromethyl group [19][20][21][22]. Given this, the development of a method for the incorporation of fluorine or trifluoromethyl group into organic molecules perhaps remains a current challenge in organic chemistry methodology. For example, trifluoromethyl-substituted (1H-indol-3-yl)methanol derivatives were reported for their promising anti-HIV inhibitory activities [23], see for example compounds III and IV ( Figure 1). However, these compounds were synthesized from multiple synthetic routes with the involvement of air sensitive conditions.
Trifluoromethyl-substituted (1H-indol-3-yl)methanol derivatives can be synthesized by Friedel-Crafts hydroxyalkylation reactions of indoles with trifluoromethyl ketones in the presence of either Lewis/Bronsted acid catalysts. Bandini et al. reported the trifluoromethyl hydroxyalkylation of indoles catalyzed by an organic base 2-tert-butyl-1,1,3,3-tetramethylguanidine (TMG), also known as Barton's base, in excellent yields ( Figure 2A) [24]. Recently, Liu and co-workers reported this reaction in the presence of cesium carbonate in acetonitrile ( Figure 2B) [25]. Dinuclear zinc [26], cinchonidine catalysts, and solvent-free conditions [27] have been also utilized for this reaction. The base-catalyzed reaction has a benefit because it may avoid the formation of diindolylmethane and biindoles as byproducts [28].
Although these methods were useful, they have limitations and drawbacks, which are the use of organic solvents [24], difficulties in recyclability of the catalyst [24,25], multiple reaction steps involving air sensitive/unhydrous conditions [23], genera-tion of a large volume of waste liquids for the compound separation and column chromatographic purification [24,25], and moderate substrate scope of the reaction. Thus, finding an alternative method with broad substrate scope, functional group tolerance, and simple purification technique is highly desired.
In our continuous effort of synthesizing indole derivatives [29][30][31], herein, we report an efficient synthesis of multiple halogensubstituted (1H-indol-3-yl)methanol derivatives in the presence of potassium carbonate and tetrabutylphophonium bromide, which mediate the reaction in water through the formation of the interface between organic and aqueous phases. The advantageous of this reaction include high yields, no column chromatography, broad substrate scope, multiscale synthesis, and recyclable of the catalyst ( Figure 2C).

Results and Discussion
The optimization studies were carried out with model substrates 5-methoxyindole (1a, 3.4 mmol) and 2,2,2-trifluoroacetophenone (2a, 3.70 mmol) in water (5 mL, see Table 1 for results). Using water as a solvent has the advantage that the formed product 3a would not be soluble, which makes the purification easier through a simple filtration. In a first attempt, the reaction did not initiate at all without any base or catalyst ( , and it was observerd that reactant 2a separated from the aqueous phase. This makes us add quaternary salts into the reaction, which could make the interface between organic and aqueous phase. Very interestingly, the combination of 20 mol % of tetrabutylammonium bromide (TBAB) and NaOH (20 mol %) yielded the desired product, 2,2,2-   trifluoro-1-(1H-indol-3-yl)ethan-1-ol (3a) in 93% yield ( Table 1, entry 4). The product 3a was isolated by filtration and confirmed by NMR and X-ray crystallography analysis (CCDC-1973322, see Supporting Information File 1 for detailed crystallographic data). It is noteworthy to mention that the formation of the product was even increased in the presence of tetrabutylphosphonium bromide (TBPB) to 96% ( To find the best catalytic system, other quaternary salts were invested in the reaction. When tetrabutylammonium chloride (TBAC) or tetrabutylammonium fluoride (TBAF) was employed in the presence of K 2 CO 3 , the formation of the product was reduced to 72% (Table 1, entry 13) and 61% (Table 1, entry 14), respectively. These results suggest that the combination of K 2 CO 3 /TBPB could be the best catalytic system for this reaction.
Next, the amount of catalytic system K 2 CO 3 /TBPB was reduced from 20 mol % to 15 mol % to see if there any change in the yield. As indicated in Table 1, entry 15, the product 3a was isolated in >99%. However, further reduction to 10 mol % or 5 mol % slowed the reaction rate, and only 90% ( Having the optimized conditions in hand, the substrate scope of the reaction was explored. At first, different trifluoromethyl ketones were investigated, and the results are summarized in Table 2. Trifluoroacetophenones having a halogen substituent at the para-position of the phenyl ring such as p-F (2b), p-Cl (2c), and p-Br (2d) provided the corresponding trifluoro-1-(1H-indol-3-yl)ethan-1-ols 3b, 3c, and 3d in 97, 92 and 89% yields, respectively. Similarly, trifluoroacetophenones having a p-methyl (2e) and p-methoxy (2f) group at the phenyl ring resulted in 3e, and 3f with excellent yields of 98 and 93%, respectively. These results suggest that the electronic properties of the substituent on the phenyl ring of the trifluoroacetophenones did not significantly influence the yield of the reaction.
Next, the role of fluorine in the reactants was explored by subjecting the reaction of 2,2-difluoro-1-phenylethan-1-one (2i) with 1a. The desired product 3i was obtained in 63% yield. However, no product was formed when the reaction was carried out with 2-fluoro-1-phenylethan-1-one (2j) or acetophenone (2k). The reason could be due to either reducing the electrophilicity of the ketone or the presence of enolizable protons in α-position to the keto group in basic medium. Supporting this hypothesis, the reaction of 1a with 1,1,1-trifluoro-3-phenyl-2propanone (2l), which has a 3-methylene group also did not proceed at all.
The recyclability of the catalytic system n-Bu 4 PBr/K 2 CO 3 of this protocol was investigated in the preparation of 3a using 1a (3.40 mmol), 2a (3.70 mmol), K 2 CO 3 (0.5 mmol) and n-Bu 4 PBr (0.5 mmol) in distilled water (5 mL) at room temperature for 12 h (Figure 4). The product 3a was filtered after completion of the reaction. The resulting filtrate was recovered, washed with ethyl acetate to remove any organic impurities, and reused for the next cycle. This procedure was followed for each cycle. Interestingly, the catalytic system was efficient to produce the product 3a in excellent to good yields up to 4 cycles (99-84%). In the fifth cycle, the yield of 3a was reduced to 67%. This could be due to the dilution of the catalytic system in each cycle.
3-Indolylmethanols are versatile pre-electrophiles for C-C functionalization at the 3-position of indoles. Particularly, the Friedel-Crafts alkylation of 3-indolylmethanols with indoles has become a useful method for the preparation of 3,3'-, and 3,6'-DIMs, which are known to possess a wide variety of biological activities, including anti-inflammatory, and anticancer effects. Therefore, we decided to synthesize DIMs from 3a, which reacted with indole (1b) or 2-phenylindole (1k) in the presence of Ga(OTf) 3 in ACN at room temperature or at 80 °C as reported by Y. Ling et al. [32]. As indicated in Scheme 2, the desired unsymmetrical 3,3'-DIM (9: 81%) and 3,6'-DIM (10: 77%) with quaternary center were afforded in good yields. Besides, the protocol reported by S. Sasaki et al. [33] was also employed for the synthesis of 3,3'-DIMs in good yield, see for example product 11 (80%, Scheme 2).
Next, a plausible mechanism for the preparation of multihalogen-alkylated 1-(1H-indol-3-yl)-1-phenylethan-1-ols is proposed as indicated in Scheme 3. Based on the literature [34], this reaction initiates by the formation of n-Bu 4 P + KCO 3 − salt (A) from the interaction of n-Bu4PBr and K 2 CO 3 . Intermediate A makes a hydrogen bond interaction with the NH of the 5-methoxyindole (1a) and form the adduct B. This interaction assists 1a reacting with an electrophilic ketone (2a) to form the intermediate D via C-C bond formation (C). Re-aromatization of D generates E, which then protonates to form the desired product 3a by excluding A for the next catalytic cycle. Figure 4: Recyclability of the catalytic system n-Bu 4 PBr/K 2 CO 3 for the preparation of 2,2,2-trifluoro-1-(5-methoxy-1H-indol-3-yl)-1-phenylethan-1-ol (3a).

Scheme 3:
Proposed mechanism for the preparation of 3a as an example.

Scheme 4: Control experiments.
Further, to prove this hypothesis the following control experiments were performed. The reaction of 1-methylindole (1l) with 2a failed to provide the product (Scheme 4). It suggests that the indole having a free NH-functionality is important to interact with the base, thereby initiating the reaction. To find the impor-tance of electrophilicity of ketones, different enolizable and nonenolizable ketones were screened with the reaction of 5-methoxyindole (1a). The enolizable ketones 2j-l failed to provide the products (see Table 2). Similarly, nonenolizable ketones 2p and 2q (Scheme 4) failed to provide the products. This observation suggests that the multihalogen-substitution enhanced the electrophilicity of the ketone for the Friedel-Crafts hydroxyalkylation reaction of indole.

Conclusion
In conclusion, we have developed an efficient and practical protocol for the preparation of trifluoromethyl(indolyl)phenylmethanols, which are of significant interest serving as pre-electrophiles for C-C functionalization at the 3-position of indoles. Particularly, the Friedel-Crafts alkylation of 3-indolylmethanols with indoles has become a useful method for the preparation of 3,3'-, and 3,6'-DIMs, which are known to possess a wide variety of biological activities, including anti-inflammatory, and anticancer effects. Additionally, trifluoromethyl(indolyl)phenylmethanols itselft have various biological properties including anti-HIV activity. The developed new synthetic protocol for the preparation of trifluoromethyl(indolyl)phenylmethanols is operationally simple and provided products in high yields without requiring silica gel column chromatography. The reaction has a broad substrate scope and proceeds with high regioselectivity. The recovery and reusability of the catalytic system and large-scale synthesis of products, which would further transform into biologically active indole-derived compounds, are further advantages of this protocol.

Supporting Information
Materials and methods and detailed synthetic procedures and spectroscopic data of all compounds. Figure S1: ORTEP-type plot of the molecular structure of 3a, Figures S2-S25: NMR spectra, Tables S1-S3: Crystal data and structure refinement for compound 3a.

Supporting Information File 1
Experimental and analytical data.