Deacylative Thiolation by Redox-Neutral Aromatization-Driven C−C Fragmentation of Ketones

Abstract

Herein we report the development of deacylative thiolation of diverse methyl ketones. The reaction is redox-neutral, and heavy-metal-free, which provides a new way to introduce thioether groups site-specifically to unactivated aliphatic positions. It also features excellent functional group tolerance and broad substrate scope, thus allowing late-stage derivatization. The process benefits from efficient condensation between the activation reagent and ketone substrates, which triggers aromatization-driven C−C fragmentation and rapid radical coupling with thiosulfonates. Experimental and computational mechanistic studies suggest the involvement of a radical chain pathway.

Entry for the Table of Contents

A redox-neutral, heavy-metal-free approach has been developed to realize deacylative thiolation of common ketones, which provides a unique method to introduce thioether groups site-specifically to unactivated aliphatic positions.

Social media promotion:

Title: Deacylative Thiolation by Redox-Neutral Aromatization-Driven C−C Fragmentation of Ketones (Zhou et al.) https://twitter.com/gbd_lab

Ketones commonly exist in bioactive natural products, pharmaceuticals, feedstock chemicals, and synthetic intermediates.^[1] Among various transformations of ketones, those involving fragmentation of α C−C bonds, such as Baeyer–Villiger oxidation, Beckmann rearrangement, and Schmidt reaction, have been strategically significant in organic synthesis,^[2] as they can convert a common alkyl−acyl bond into a useful alkyl−O or alkyl−N bond (Scheme 1A). However, to the best of our knowledge, there has been no direct method to transform an acyl group in a regular alkyl ketone into a thio group, which, if successful, could offer a new bond-disconnecting strategy for preparing alkyl thioethers. On the other hand, alkyl thioethers have been important pharmacophores (Figure 1) and can be used to regulate lipophilicity, solubility, and metabolism of drug molecules.^[3] In particular, trifluoromethyl thioethers exhibit high Hansch lipophilicity (π = 1.44),^[4] suggesting that SCF₃-containing molecules could be easily transported through a lipid membrane. In addition, sulfoxides and sulfones, readily accessible from thioethers, are also frequently found in bioactive compounds and can serve as bioisosteres of ketones and carboxylic acids.^[5]

Scheme 1.

Thiolation via C−C fragmentation of ketones. Ts = tosyl; Ac = acetyl; Py = 2-pyridyl.

Figure 1.

Representative bioactive thioethers.

Recently, transition metal-catalyzed C−C bond activation emerges as a unique approach to functionalize unstrained ketones;^[6] however, the corresponding deacylative thiolation reaction remains elusive, likely owing to the strong coordination of sulfur moieties to transition-metal catalysts. As a complementary C−C cleavage approach, we have been engaged in developing deacylative transformations capitalized on aromatization as a driving force since 2019 (Scheme 1B).^[7] More recently, an effective activation reagent, N’-methylpicolinohydrazonamide (MPHA),^[7c] was discovered, which can efficiently condense with various alkyl ketones to form a pre-aromatic intermediate (PAI). The subsequent N−H bond cleavage triggers C−C homolytic scission to afford a carbon-center radical species (R2),^[8] which can be oxidized to give an alkene^[7c] or undergo hydrogenation/deuteration.^[7d] However, it remains unclear if deacylative C−heteroatom bond formation through trapping R2 with a polar coupling partner could be achieved in this system. Inspired by the excellent stability of sulfonyl radicals, we questioned if easily accessible thiosulfonates can serve as a bifunctional reagent to first trap the R2 radical to form C−S bond,^[9] and then to abstract the NH hydrogen from the PAI by the resulting Ts radical to regenerate R2 via C−C cleavage (Scheme 1C). Comparing to the prior art on thiolation mediated by C−C cleavage (Scheme 1D),^[10] in which highly strained substrates, oxidaitve conditions, or photoirrdidations were typically required, here we describe the development of a redox-neutral deacylative thiolation of diverse methyl ketones. In this reaction, light, stoichiometric oxidants, reductants, strong acids or bases are avoided.

To verify the proposal, synthesis of trifluoromethyl thioether 2a from methyl ketone 1a was explored first (Table 1), given the general biomedical interests and synthetic challenges associated with alkyl trifluoromethyl thioethers.^[4a,11] Catalyzed by 1-AdCO₂H, the condensation between MPHA and 1a, using Al₂O₃ as the dehydrating agent,^[12] undergoes smoothly with nearly full conversion. After tuning various reaction parameters (for additional control experiments, see Supporting Information), the desired deacylative trifluoromethylthiolation product (3a) was obtained in 74% yield after treating the condensation mixture with 1.2 equiv of TsSCF₃ and 1.0 equiv of NaOAc in toluene (0.1 M) at 140 ^oC for 3 hours (entry 1). Excess TsSCF₃ is not necessary, and the yield dropped to some extent when using 2 equiv of TsSCF₃ (entry 2). Compared to the previous methods of forming alkyl trifluoromethyl thioethers,^[13] this strategy does not require stoichiometric heavy metals, oxidants, excess costly reagents or light.^[14]

Table 1.

Selected optimization studies.

entry	Variation from ‘standard condition’	Yield of 3a[a]	entry	Variation from ‘standard condition’	Yield of 3a[a]
1	None	74(70[b])%	7	120 oC instead of 140 oC	62%
2	2 equiv. TsSCF3	60%	8	1,4-dioxane (0.1 M)	67%
3	w/o 1-AdCO2H	56%	9	THF (0.1 M)	43%
4	w/o Al2O3 acidic	62%	10	DMSO (0.1 M)	15%
5	w/o NaOAc	59%	11	toluene (0.5 M)	60%
6	Additional 30 mol% Cu(OAc)2	17%	12	toluene (0.05 M)	72%

^[a]Based on ¹H NMR analysis of the crude reaction mixture.

^[b]Isolated yield. 1-AdCO₂H = 1-adamantanecarboxylic acid; w/o = without

In the absence of 1-AdCO₂H or running the reaction without Al₂O₃ (entries 3 and 4), the yield only decreased slightly, suggesting a robust condensation process. The addition of 1 equiv of NaOAc, likely to buffer the sulfinic acid (Ts-H) generated, improved the yield (entry 5). Unlike the previous deacylative hydrogenation or alkene formation with this system,^[7c,d] adding a copper catalyst was detrimental to the reaction (entry 6), as copper likely accelerated decomposition of TsSCF₃.^[15] Lowering the reaction temperature or changing to more polar solvents gave worse results (entries 7–10). While more concentrated conditions led to lower yield owing to more side-reactions, such as hydrolysis of the PAI and dimerization of radical intermediates, decreasing the concentration afforded comparable yield (entries 11 and 12).

With the optimal conditions in hand, the substrate scope was next examined (Table 2). A diverse range of methyl ketones^[16] smoothly underwent the desired deacylative trifluoromethylthiolation. The C−C cleavage took place site-selectively at the non-methyl side. Broad functional groups (FGs), including trimethylsilyl (3b), aryl iodide (3c), bromide (3d, 3f), boronate (3g), acetal (3h), ester (3o), ether (3r, 3t, 3u, 3v) and N-Boc group (3p), were all tolerated. Interestingly, the reaction selectively occurred at the dialkyl ketone moiety in the presence of an aryl methyl ketone (3e). Additionally, α-substituted ketones also worked well and gave comparable yields (3l-3m). Heteroarenes, such as dibenzothiophene (3j), dibenzofuran (3m), isoquinoline (3q), N-methylindole (3w), and protected piperidine (3x), and polyarenes, such as anthracene (3i) and phenanthrene (3l), were also tolerated. Note that good yield can also be obtained on a larger scale (3a, 1 mmol, 67%).

Table 2.

Substrate scope of ketones.^[a,b]

^[a]Unless otherwise noted, the reactions were conducted on a 0.1 mmol scale. All yields are isolated yields.

^[b]1.0 mmol scale.

^[c]0.2 mmol scale. Boc = tert-butyloxycarbonyl; Bn = benzyl; TMS = trimethylsilyl; Bpin = 4,4,5,5-tetramethyl-1,3,2-dioxaboryl.

Beyond forming trifluoromethyl thioethers, the deacylative coupling can be extended to install other alkyl- or aryl thio groups simply by using different thiosulfonate reagents (Ts-SR) (Table 3). To our delight, the reaction appears to be quite robust and general. A number of different alkyl- and aryl thio groups can be efficiently installed in good yields. For example, the SMe-based reagent gave 79% isolated yield of the corresponding methyl thioether (4a). When forming aryl thioethers, the electronic effect of the aryl group did not significantly impact the reaction (4e-4n). Nitro group (4h) survived under the standard conditions. Heteroarene-derived reagents, e.g., pyridine (4m)- and thiophene (4n)-based ones, also worked well.

Table 3.

Scope of forming other thioethers.^[a]

^[a]The reactions were conducted on a 0.1 mmol scale. All yields are isolated yields.

The synthetic utility of the deacylative thiolation was further investigated (Scheme 2). Ketone moieties can serve as a traceless handle to access interesting products. As shown in Scheme 2A, oxidative cleavage of the alkene in (+)-Δ3-carene,^[17] followed by this reaction, led to ring-opening of the cyclopropane and trifluoromethylthiolation at the methylene position. In addition, fluorination at the ketone α-position, followed by applying this method, resulted in a monofluoroalkyl trifluoromethyl thioether (9) that is hard to prepare by existing approaches. Through merging with metal-hydride H-atom transfer (MHAT) between a 1,1-disubstituted olefin and butenone,^[18] deacylative thiolation of the resulting ketone afforded trifluoromethyl thioether 12 containing a quaternary carbon. On the other hand, methyl ketones derived from various natural products allowed smooth installation of the trifluoromethyl thioether moieties (Scheme 2B), despite the existence of labile or reactive FGs, e.g., ketal (13), epoxide (15), carbonyl (17) and xanthine (19). Thus, this method shows potential for use in late-stage modifications of bioactive compounds. Considering that diverse aryl-substituted methyl ketones are easily accessible,^[19] the combination with this method allows flexible preparation of various thioethers that contain a tethered or fused aryl group (Scheme 2C).

Scheme 2.

Synthetic applications of the deacylative thiolation.

To obtain some mechanistic insights (Scheme 3), the PAI derived from substrate 1a was synthesized separately. First, the control experiments show that the reaction between PAI and TsSCF₃ can take place even at room temperature,^[20] and good yield of the desired product can be obtained in the dark, which excludes light-mediated electron transfer pathways (Scheme 3A). Another question is what the fate of the Ts radical is. The major byproduct isolated from this reaction is thiosulfonate 2g. Given that sulfinic acids are known to undergo self-condensation and disproportionation to form thiosulfonates,^[21] it could be reasonable to assume that the reaction initially formed sulfinic acid Ts-H (Scheme 2B). The presence of the Ts-H intermediate is further supported by a crossover experiment with added phenyl sulfinic acid, in which the crossover byproduct (TsSPh) 2b was detected by ¹H NMR and LCMS. Moreover, the radical trapping experiment was carried out by adding TEMPO (Scheme 2C). Indeed, by raising the loading of TEMPO, the trifluoromethylthiolation was gradually inhibited and the yield of the TEMPO-trapped product (34) increased, indicating that the reaction should involve a carbon-center-radical intermediate.

Scheme 3.

Mechanistic explorations. TEMPO = 2,2,6,6-tetramethylpiperidine 1-oxyl.

To gain a deeper understanding on how the radical reaction is initiated and propagated, a density functional theory (DFT) study was performed (Figure 2). The computational results show that, Ts−SCF₃ first undergoes facile S−S bond homolytic cleavage to yield SCF₃ and Ts radicals. Although both radical intermediates could abstract the NH hydrogen from PAI, the H atom transfer to the SCF₃ radical appears to be much faster (TS 1 versus TS 1’). Driven by aromatization, the resulting delocalized radical int 2 can then undergo barrierless C−C homolytic fragmentation to yield the triazole byproduct and alkyl radical int 3. The following S_H2^[22] reaction between int 3 and Ts−SCF₃ occurs with a low activation barrier (1.7 kcal/mol) to provide product 3a and the generated Ts radical then reinitiates the chain reaction by abstracting the NH hydrogen from PAI via TS 1. Note that during the chain propagation, the overall reaction is thermodynamically favored by 22.2 kcal/mol.

Figure 2.

DFT calculation of the deacylative trifluoromethylthiolation process. ^[a]Computations were performed with Gaussian 16 at the B3LYP(d3)- 6–31G(d,p) level for geometry optimizations and the M06(d3) (SMD(Toluene))-SDD/6–311+G(d,p) level for single-point energies. ^[b]Structures of intermediates and transition states were illustrated using CYLview.

In summary, we have developed the first deacylative thiolation of diverse methyl ketones, which provides a straightforward approach to install thio groups site-specifically to unactivated C(sp³) positions. Given the versatile reactivity and easy accessibility of ketones, this method offers an efficient and unusual strategy to prepare functionalized thioethers. The broad scope and FG tolerance make this method attractive for late-stage functionalization of complex molecules. The mechanistic insights gained here should provide implications for developing other aromatization-driven C−C cleavage reactions.