A Light-Promoted Innate Trifluoromethylation of Pyridones and Related N-Heteroarenes

We report a practical, light-mediated perfluoroalkylation using Langlois’ reagent (sodium trifluoromethylsulfinate) that proceeds in the absence of any photocatalyst or additives. This method has allowed for the facile functionalization of pyridones and related N-heteroarenes such as azaindole. This protocol is operationally simple, uses readily available materials, and is tolerable for electron-neutral and -rich functional pyridones. Cyclic voltammetry was utilized as a mechanistic probe, and preliminary data suggest the reaction may involve an electrophilic radical mechanism.

P erfluoroalkyl groups impart unique electronic effects to small molecules, which has led to their widespread incorporation in said molecules. 1,2 In drug discovery, these groups have been shown to increase metabolic stability, lipophilicity, and overall activity. 3 Perfluoroalkylated functionalities are often added to arenes and N-heteroarenes, necessitating the development of strategies for further incorporating these groups. Recent developments in perfluoroalkylation methods have been employed by reagents such as, but not limited to, Prakash's reagent, 4 Langlois' reagent, 5 Togni's reagent, 6 Umemoto's reagent, 7 Baran's reagent, 8 Ritter's reagent, 9 CF 3 I, 10 triflyl chloride, 11 triflic anhydride, 12 and trifluoroacetic acid (TFA). 13 These reagents can be either oxidized or reduced to generate a perfluoroalkyl radical with electrophilic or nucleophilic behavior. 14−16 While many of the methods developed alongside these reagents have been shown to functionalize a range of N-heteroarenes, they often require harsh conditions, precious metal catalysts, or prefunctionalization of substrates, which can limit their practicality on a large scale and limit their utility for late-stage functionalization (LSF) of more complex N-heteroarenes.
There are fewer methodologies that afford pyridone perfluoroalkylation compared with other arenes and heteroarenes. Pyridones are a privileged scaffold and are common bioisosteres for a variety of functional groups such as amides, pyranones, pyrimidines, pyrazines, and phenols. 17−19 There are several biologically relevant perfluoroalkylated pyridones, including FDA-approved drug Pifeltro, an HIV-1 medication for use in combination with other antiretroviral medicines, and Fusilade DX, an EPA-approved herbicide used to treat weeds for cotton and soybeans. Despite their ubiquity, there are few direct, mild synthetic methods toward trifluoromethylated pyridones. They are often synthesized on a scale from the corresponding methylated pyridine, which then undergoes a chlorination and subsequent fluorination before transforming it into a pyridone (Scheme 1a). 20−22 In the context of Pifeltro, the perfluoroalkyl group was preinstalled onto a halogenated pyridine, which was subsequently reacted in an S N Ar fashion followed by oxidation to produce the trifluoromethylated pyridone moiety. 23 Another common route for achieving fluorinated pyridones involves a deoxyfluorination from the corresponding hydroxypicolinic acids from the Mykhailiuk 24 group (Scheme 1b). Direct strategies for incorporating perfluoroalkyl groups onto pyridones are highly desirable. 25 Current strategies for directly obtaining trifluoromethylated pyridones and similar scaffolds include an iron(II)-mediated trifluoromethylation using CF 3 I and hydrogen peroxide from the Yamakawa 10 group (Scheme 1c), photoredox trifluorome-thylations using triflyl chloride from the MacMillan 11 group (Scheme 1d), and TFA from the Jin 13 group (Scheme 1e). Herein, we report a simple light-mediated radical trifluoromethylation strategy that circumvents the need for harsh, expensive conditions that do not use any photocatalyst or oxidative additive on pyridones (Scheme 1f). This methodology has also been extended to include N-heteroarenes as well as some examples of late-stage functionalization (LSF).
While studying the perfluoroalkylation of aromatics, we noticed a unique reactivity where pyridones such as 1a could undergo a trifluoromethylation at position three using Langlois' reagent 2a, DMSO, and 390 nm LEDs (Table 1, entry 1) without the need for any photocatalyst or additive under an ambient atmosphere. The functionalization is occurring at the most nucleophilic site on 1a, and we postulate that this proceeds through an electrophilic perfluoroalkylation mechanism. 26 We found this noteworthy as there was no need for any photocatalyst or terminal oxidant, as is typically needed for such transformations. In the absence of light, we observe no conversion (Table 1, entry 2); however, reactivity can be regained in the absence of light by using K 2 S 2 O 8 as a terminal oxidant (Table 1, entry 3), which would be expected on the basis of work by Baran and co-workers. 14 27 We next evaluated the optimal conditions across a variety of pyridones to demonstrate the versatility of this approach. As shown in Scheme 2, the reaction is tolerable among different pyridones, including aryl-alkylated (3b and 3c) and Nalkylated pyridones (3e and 3f) in good yields ranging from 56% to 93%. Interestingly, when position 3 is methylated, there is a considerably lower yield (12%) of the 5-trifluoromethylated product (3d), versus 19% when the reaction mixture is sparged with O 2 , which suggests that the electronics of the substrate are important in determining whether the reaction can occur. Swapping Langlois' reagent for other difluorosulfinate salts such as NaSO 2 CF 2 CH 3 (2c) resulted in a 90% yield of 3h. NaSO 2 CF 2 H (2d) resulted in a 9% yield of 3h. A yield of 15% was achieved when the reaction mixture was sparged with O 2 . NaSO 2 CF 2 C 8 H 6 Br (2e) resulted in a 79% yield of 3i. 4-Pyridone, another bioactive pyridone scaffold, was also subjected to this methodology and produced a mixture of mono-and ditrifluoromethylated products in 51% (3j) and 23% (3k) yields, respectively. While we initially thought that pyridones were unique to this methodology, we were pleasantly surprised that other N-heteroarenes were also perfluoroalkylated. We were able to obtain trifluoromethylated uracil (3l) and azaindole (3m) in 60% and 70% yields, respectively. 2-Quinazoline was also perfluoroalkylated using these conditions, generating product 3n in ≤65% yield. We evaluated different pyridone-containing ligands as seen in the literature for use in catalysis, 28−30 and we obtained 3o and 3p in moderate yields of 49% and 42%, respectively. We did observe some substrate limitations, as we were unable to perfluoroalkylate 5-nitro-2-pyridone 1q and mesitylene 1r. We believe that in the case of 1q, the pyridone ring is too electron Scheme 1. Preparation of Trifluoromethylated Pyridones Organic Letters pubs.acs.org/OrgLett Letter poor due to the nitro substituent. However, in the context of 1r, it appears the photocatalyst and oxidant-free chemistry necessitate a heteroatom bearing lone pairs, which is believed to interact with the sulfinate (vide infra) despite sparging the reaction mixture with O 2 , which we note has a positive effect on yields. Inspired by the ability to perfluoroalkylate pyridonecontaining ligands, we believed that this methodology would also be suitable for the late-stage functionalization of bioactive molecules. To evaluate this strategy, we selected bioactive pyridones, including pirfenidone (1s), used in the treatment of idiopathic pulmonary fibrosis, cicloprox (1t), used as a treatment for fungal infections, cytisine (1u), a former therapy for smoking cessation in Europe and Asia, and trifluridine (1v), an ophthalmic treatment for viral infections. We were able to perfluoroalkylate these bioactive molecules in isolated yields of 45% (3s), 35% (3t), 33% (3u), and 56% (3v), respectively. We were able to increase the yield of 3t when the reaction mixture was sparged with oxygen to 46%. In the case of 1u, the secondary amine was identified to be an issue; however, upon Boc protection, we were able to produce 3u in 33% yield.
To gain more insight into the reaction mechanism, we performed the following experiments to elucidate the order of transformation. First, upon addition of TEMPO to the reaction, no product 3a was formed, suggesting that the trifluoromethylation occurred via a radical pathway (Scheme 3a). 31 Next, we wanted to investigate the role of the light by performing a light on/off experiment to determine whether the reaction occurs through a chain propagation mechanism and whether it could be initiated in the absence of light. 32 The reaction was sluggish using 20 min intervals, which were increased to hourly intervals. We observed that when light was removed, the reaction ceased and would occur only in the presence of light (Scheme 3b). To probe the formation of a charge-transfer complex, all reagents were then analyzed using UV−vis spectroscopy and were found to absorb in the UV-A and UV-B regions (see the Supporting Information). The reaction mixture was measured, and no significant bathochromic shift was observed, suggesting another mode of radical generation, possibly via n, π* as postulated by Mi, Li, and co-workers. 33 To better visualize the reaction, we performed cyclic voltammetry (CV) experiments to observe the presence of all reagents in the reaction mixture and to study their redox potentials. We found that pyridone 1a has an oxidative potential of 1.94 V and a reductive potential of −2.16 V (page S108 of the Supporting Information) and Langlois' reagent 2a has oxidative and reductive potentials of 1.65 and −0.76 V, respectively (page S109 of the Supporting Information). We were unable to introduce DMSO in the CVs to study its effect, as DMSO is oxidized at the same potential as 2a, which prevented analysis. When both reagents are in solution, we can see distinct peaks depending on the concentration; 100 mM was optimal for irradiation, and all CV

Organic Letters pubs.acs.org/OrgLett
Letter was performed at an overall concentration of 1 mM (Scheme 3c). At 0 h, we observe an ∼1:1 ratio of reagents; however, after irradiation of the solution at 1 h, we notice a significant decrease in relative peaks between 1a and 2a, suggesting the consumption of 2a initially. This is in line with finding from the Kim 34 group that 2a would be easier to oxidize than 1a.
When comparing the data at 7 and 15 h, we notice a decrease in the levels of both 1a and 2a, and the distinct formation of 3a, which has an oxidative potential of 2.41 V and a reductive potential of −1.83 V (page S109 of the Supporting Information). On the basis of our experimental probes, we postulate that 2a undergoes a light-mediated oxidation to generate a CF 3 radical, which then reacts with the N-aryl substrate, generating a trifluoromethylated intermediate bearing a radical, which then undergoes oxidation to the cationic product, followed by rearomatization. On the basis of the results in Table 1 (entries 8 and 9) we note the dependence of oxygen; thus, it may be plausible that these conditions are exciting oxygen, which can then oxidize 2a or participate in the oxidation to rearomatize the product. 35,36 Further mechanistic studies to better understand the full breadth of the mechanism are currently underway and will be the subject of future work.
In summary, we have described a new, light-promoted method for the trifluoromethylation of pyridones and related N-heteroarenes that does not need any photocatalyst or oxidant. This methodology is also inclusive of using a variety of sulfinate salts bearing two or more fluorine substituents. This approach is operationally simple and minimally affected by inert gases or water in the reaction. While Langlois' reagent can generate a CF 3 radical via reduction or oxidation, we offer evidence suggesting an oxidative mechanism for the electrophilic trifluoromethylation of pyridones and related Nheteroarenes.

■ ASSOCIATED CONTENT Data Availability Statement
The data underlying this study are available in the published article and its Supporting Information.
Experimental procedures, characterization of all new and existing compounds, UV−vis, cyclic voltammetry details, and their corresponding NMR and HRMS spectra (PDF)