Trifluoromethylthiolation, Trifluoromethylation, and Arylation Reactions of Difluoro Enol Silyl Ethers

This study reports a new application area of difluoro enol silyl ethers, which can be easily obtained from trifluoromethyl ketones. The main focus has been directed to the electrophilic fluoroalkylation and arylation methods. The trifluoromethylthiolation of difluoro enol silyl ethers can be used for the construction of a novel trifluoromethylthio-α,α-difluoroketone (−COCF2SCF3) functionality. The −CF2SCF3 moiety has interesting properties due to the electron-withdrawing, albeit lipophilic, character of the SCF3 group, which can be combined with the high electrophilicity of the difluoroketone motif. The methodology could also be extended to difluoro homologation of the trifluoromethyl ketones using the Togni reagent. In addition, we presented a method for transition-metal-free arylation of difluoro enol silyl ethers based on hypervalent iodines.


■ INTRODUCTION
Organofluorine compounds have found many important applications in several areas of life sciences. The high metabolic stability, ability to modify the lipophilicity, acid−base properties, and overall reactivity/bioavailability of small molecules 1 led to a widespread application of organofluorines especially in medicinal chemistry 2 and agrochemistry. 3 As a consequence, expanding the chemical space of new organofluorine compounds has attracted great attention in industrial and academic research. 4 An important class of druglike molecules is based on α-fluorinated ketone motifs ( Figure 1). Due to the strong electron-withdrawing character of di-, tri-, and perfluoro alkyl groups, the electrophilicity of the neighboring keto functionality is enhanced. As a consequence, nucleophilic functionalities of enzymes, such as the hydroxy group of serine or other residues, readily interact with the low lying π*(CO) orbitals forming (covalently bound) ketal/hemiketal-type products. 5 Thus, α-fluorinated ketones are important pharmacophores for serine proteases and related enzymes (Figure 1). For instance, small molecules with β-amino-α,α-difluoroketone motifs (Figure 1a) are selective inhibitors of human renin (protease regulating blood pressure). 6 Certain types of α,αdifluoroketones ( Figure 1b) interact with the hydroxy methylglutaryl binding domain of coenzyme A (HMG CoA) reductase and are thus efficient inhibitors of these enzymes. 7 Phospholipase A 2 (iPLA 2 ) enzymes, which are involved in inflammatory disorders (such as arthritis and autoimmune diseases), can be efficiently inhibited by perfluoroethyl ketones (Figure 1c). 8 Human neutrophil elastase (HNE) is a serine protease, which can also be involved in pathophysiological states (such as cystic fibrosis and chronic bronchitis) and inhibited by trifluoromethyl ketones (Figure 1d). 9 The binding properties to HNE can be improved by variation of the fluorinated carbonyl activating group. 10 Thus, perfluoroethyl ketones ( Figure 1e) have also been considered as HNE inhibitors. 9b, 10,11 Application of the α-trifluoromethylthio ketones is less common in druglike molecules than the fluoro or fluoroalkyl analogues, which might be the consequence of synthetic limitations. This pharmacophore occurs, for example, in cefazaflur (Figure 1f), which is a cephalosporin antibiotic. 12 Difluoro enol silyl ethers are useful synthons for introduction of the α,α-difluoro carbonyl functional group. 13 These compounds can be prepared from trifluoroketones (1) by Mg-mediated cleavage of one of the C−F bonds ( Figure  2a). 14 This reaction affords fairly stable difluoro enol silyl ethers (2), which are difficult to isolate, and therefore, derivatives of 2 are usually reacted with various electrophiles without purification. The standard applications 13a involve aldol reactions, 14a,15 Mannich reactions, 16 protonation, 14b halogenation, 17 and arylation 18 reactions ( Figure 2b).
As a part of our organofluorine chemistry program, 19 we sought to expand the reagent scope of difluoro enol silyl ethers (2) to reactions with new types of electrophiles. Our interest was 2-fold: (i) testing electrophilic reagents for the introduction of (S)CF 3 groups and (ii) studying the application of hypervalent iodine-based electrophiles ( Figure  2c) in C−C bond-forming reactions with 2.
Many recent efforts have been undertaken to find new methodologies for selective introduction of the SCF 3 group. 4b,20 This group frequently occurs in drug molecules (e.g., Figure 1f) because of its excellent pharmacochemical properties, such as the strong electron-withdrawing character and an exceptionally high Hansch lipophilicity parameter (π = 1.44). 21 Although α-trifluoromethylthiolation of ketones is described in the literature, 22 introduction of the SCF 3 group in difluoro enol silyl ethers 2 has never been reported. We hypothesized (Figure 3) that reacting 2 with sufficiently reactive electrophilic SCF 3 transfer reagents, such as benzenesulfonimide 3 , 2 3 a novel perfl uoroalkyl (−COCF 2 SCF 3 ) functionality could be constructed (4). In this way, the difluoro ketone group could be equipped with an electron-withdrawing but lipophilic group allowing, for example, extension of the pharmacochemical space of αfluorinated ketones (Figure 1). In fact, very few synthetic methods have been reported for construction/introduction of the −CF 2 SCF 3 functionality, 24 while none of these literature methods were suitable for the preparation of −COCF 2 SCF 3containing molecules.

■ RESULTS AND DISCUSSION
In line with our hypothesis (Figure 3), when N-trifluoromethylthiodibenzenesulfonimide 3 23 was reacted with difluoro enol silyl ether 2a (freshly prepared 14 from trifluoroacetophenone 1a with Mg and TMSCl) in the presence of KF, −COCF 2 SCF 3 -functionalized product 4a was obtained ( Table  1). The optimal conditions involved application of 2a, 3, and KF in equimolar amounts in acetonitrile at room temperature for 3 h. Under these conditions, 4a was obtained in 78% (NMR) yield along with 3% of the protonated analogue 5a (Table 1, entry 1). Deviation from these conditions led to lower yields and/or extensive formation of 5a.
When KF was replaced by CsF, the yield slightly decreased (entry 2). Using TBAF as an activator instead of KF proved to be even less effective (entry 3). The reaction proceeded with poor yield (15%) and extensive formation of protonated side product 5a. Application of TBAT additive (entry 4) led to a higher yield (41%) than use of TBAF. However, the yield was still lower than with KF. Since tertiary amines have been successfully used as activators in aldol reactions of 2, 15d we attempted to replace KF with DABCO. However, we could not detect formation of 4a in the presence of DABCO (entry 5).   We have also attempted to use FeCl 2 as an activator, which proved to be efficient in the analogue trifluoromethylation reaction (see below). However, the trifluoromethylthiolation reaction proceeded with only 8% yield (entry 6). When acetonitrile was replaced with dichloromethane, formation of product 4a was not observed (entry 7). A possible explanation is the poor solubility of KF in dichloromethane. In the absence of any activator, only traces of 4a were formed (entry 8).
With the optimized conditions in hand, we explored the substrate scope of the trifluoromethylthiolation of 2 ( Table 2).
As mentioned above, the reaction of 2a proceeds with high NMR yield (78%), but because of the volatility of 4a only 25% of the product could be isolated. The volatility of the products could not be decreased by hydration of the difluoro-keto groups. p-Phenyl-substituted ketone 4b formed smoothly and could be isolated in 78% yield. The reaction of naphthyl enol silyl ether 2c afforded 4c in 84% yield within 4 h. Reactions using substrates with electron-withdrawing substituents in the aromatic ring gave the corresponding products 4d,e with 67% and 62% yield, respectively. The difluoro enol silyl ether with an electron-withdrawing fluoro group also reacted smoothly, affording product 4f in 57% NMR yield. However, because of the volatility of 4f, it could be isolated only in 35% yield. Alkenyl difluoromethyl enolates could also be trifluoromethylthiolated with good yields, as 4g and 4h are obtained in 70% and 60% yields, respectively. We were also able to prepare the −CF 2 SCF 3 analogue of the synthetic intermediate of the HNE inhibitor 25 presented in Figure 1d. Compound 4i formed with 35% yield. The reaction could be scaled up to 1.0 mmol scale (4b) without significant change of the yield (71%).
We were able to extend the above methodology to trifluoromethylation of difluoro enol silyl ethers 2 (Table 3). Using Togni reagent 4c 6a (instead of 3), difluoro homologation of trifluoromethyl ketones (1) could be achieved. The trifluoromethylation of difluoro enol silyl ether 2a proceeded with poor NMR yield (35%) in the presence of KF as the mediator (Table 3). However, when catalytic amounts of DABCO or FeCl 2 were used the yield could be improved to 43% and 62%, respectively. Application of Sc(OTf) 3 led to formation of 7a with 2% yield. When the other type of Togni reagent 4c (1-trifluoromethyl-3,3-dimethyl-1,2-benziodoxole, 6b) was used, the yield dropped to 20%. With Umemoto reagent 5-(trifluoromethyl)dibenzothiophenium tetrafluoroborate, 26 only 1% of 7a was obtained. Considering these results, further reactions were carried out with 6a in the presence of FeCl 2 catalyst. 27 Similar to the SCF 3 analog (4a), compound 7a was volatile, and therefore, the yield was not determined. However, perfluoroethyl compound 7b with a phenyl substituent on the aromatic ring could be isolated in 85% yield. The naphthyl analogue was obtained in 33% yield. However, 7e, with the electron-donating EtS group on the aromatic ring, formed in high yield (70%). In the presence of the electron-withdrawing fluoro substituent, the reactions still proceeded smoothly, affording 7f with 65% NMR yield. Product 7f was also volatile, and therefore, it was not isolated. Trifluoromethyl vinyl ketone 1h also underwent difluorohomologation via 2h, affording 7h in 20% yield. Similarly to the trifluoromethyltiolation (4i), the drug intermediate of HNE inhibitors (Figure 1d), 1i, could be converted to its perfluoroethyl analogue 7i in 33% yield.
Interestingly, the above trifluoromethylation and trifluoromethylthiolation reactions could also be performed in a sequence ( Figure 4). Thus, 7b (obtained from 1b by trifluoromethylation) may undergo a subsequent trifluoromethylthiolation with 3 via enolate 2j to afford 4j in 86% yield. Table 2. Substrate Scope of the Trifluoromethylthiolation Reaction of Difluoro Enol Silyl Ethers a a General procedure: 2 (0.1 mmol), 3 (0.1 mmol), and KF (0.1 mmol) were stirred in MeCN (0.5 mL) at room temperature for 3 h. b19 F NMR yield using PhCF 3 as the internal standard. c Reaction conducted on 1.0 mmol scale. The interesting feature of 4j is that all three functional groups, which are widely used in organofluorine chemistry, are attached to the same carbon center. The above results with the Togni reagent 6a suggested that other hypervalent iodines 28 can also be useful electrophiles for the functionalization of difluoro enol silyl ethers 2. Indeed, we have found that 2 reacted smoothly with diphenyl iodonium salt 8a (Table 4) in the presence of KF under conditions very similar to those of the above trifluoromethylthiolation reactions ( Table 2). The reaction afforded 9a with 61% yield using KF as mediator, while formation of 9a was not observed with DABCO or FeCl 2 as mediators (instead of KF). The bromo analogue 8b reacted with an even higher yield of 72%. The synthesis of aryl bromide products proceeded smoothly under the applied transition-metal-free conditions. Thus, both phenyl-and bromophenyl-containing difluoroketones 9c−9g could be easily obtained in 50−71% yields. Not only aryl ketones but also vinyl ketone 1h could be converted to bromophenyl derivative 9h (69%) via enolate 2h. We attempted to use alkynyl derivatives of hypervalent iodine reagents, 29 but formation of the corresponding difluoroketone product was not observed.

■ CONCLUSIONS
In summary, we have presented a new method for the functionalization of difluoro enol silyl ethers with electrophilic alkyl fluoride and aryl-transfer reagents. Using trifluoromethylthiolation reagent 3, −COCF 3 functionalities could be converted to −COCF 2 SCF 3 groups. The electron-withdrawing but lipophilic −CF 2 SCF 3 group is an interesting new modifier for the reactivity of the keto groups in bioactive compounds ( Figure 1). We have also presented two extensions of the methodology using hypervalent iodine-based electrophiles. By use of Togni reagent 6a, difluoro homologation of the −COCF 3 group can be performed. This way of construction of a perfluoroethyl group can be used as an alternative method for the introduction of −CF 2 CF 3 moiety to ketones. 30 In addition, we have shown that transition-metal-free arylation of difluoro enol silyl ethers can be performed using hypervalent iodines as aryl source. This method complements the reported transition-metal-catalyzed cross-coupling and other related arylation reactions to obtain −COCF 2 Ar functionality. 18,31 Overall, the new methodology is suitable for the expansion of the synthetic space of difluoro ketone based compounds including bioactive molecules.

■ EXPERIMENTAL SECTION
General Information. The difluoro enol silyl ethers 2 were freshly prepared (and used without purification) from the corresponding trifluoroketones according to the literature procedures reported by the Olah/Prakash 14b and the Uneyama 14a,32 groups. Trifluoromethylthiolating reagent 3 was synthesized according to a procedure by Shen and co-workers. 23 Trifluoromethylating reagent 6a (trifluoromethylbenziodoxolone) and 6b (trifluoromethylbenziodoxole) were synthesized according to the reported literature procedures. 27,33 Anhydrous acetonitrile was purchased from Sigma-Aldrich and stored in an argon-filled glovebox. KF was dried by heating with a heat gun (550°C) for about 2 min under vacuum and kept under Ar before use in the glovebox. All other chemicals, including diaryliodonium salts 8a and 8b, were obtained from commercial sources and used as received. 1 H, 13 C, and 19 F NMR spectra were recorded in CDCl 3 (internal standard: 7.26 ppm, 1 H; 77.0 ppm, 13 C) using 400 or 500 MHz spectrometers. High-resolution mass data (HRMS) were recorded on a Bruker microTOF ESI-TOF mass spectrometer. For column chromatography, silica gel (35−70 μm) was used. Unless otherwise stated, the reactions were conducted under Ar atmosphere.
2,2-Difluoro-1-phenyl-2-((trifluoromethyl)thio)ethan-1-one (4a). The title compound 4a was prepared according to the general procedure A. The NMR yield (78%) was determined by 19 F NMR (with 1 H decoupling, using 3s of delay time in 32 scans) with PhCF 3 as the internal standard. The product was purified by silica gel column chromatography (eluent: pentane) to afford 4a as a volatile pale yellow oil ( Synthesis of 4b on 1.0 mmol Scale. To a mixture of the difluoro enol silyl ether 2b (0.30 g, 1.0 mmol), N-trifluoromethylthiodibenzenesulfonimide 3 (0.40 g, 0.1 mmol, 1.0 equiv), and KF (58 mg, 1.0 mmol, 1.0 equiv) was added 5.0 mL of dry MeCN in a glovebox. The reaction mixture was stirred at room temperature for 3 h. After evaporation, the product was isolated by silica gel column chromatography (eluent: pentane) to afford 4b as a white solid (0.24 g, 71% yield).