Net Oxidative Photocatalytic Strategy toward the o‑Trifluoroacetylation of N‑Heteroarenes

Abstract

The direct introduction of fluorinated functional groups into small molecules remains one of the most powerful chemical transformations toward synthesizing biologically active compounds. Heteroaromatics are particularly significant scaffolds, and fluorofunctionalization reactions are still in demand. Herein, we report a protocol to synthesize heteroaromatic trifluoromethyl ketones (TFMKs), installed through photomediated Minisci coupling, followed by subsequent deketalization. We have demonstrated the key Minisci reaction in more than 10 unique heterocyclic classes, notably including pyridines, pyrazines, pyridazines, pyrimidines, quinolines, and polycyclic systems, alongside select deketalizations of high-value aromatics.

N-Heteroarenes are some of the most frequently represented scaffolds in chemistry, appearing often in both natural products and synthetic medicines. Aromatic N-heterocycles such as pyridines, diazines, and quinolines are of particular interest to synthetic chemists as they have been recognized as the most prevalent heterocyclic scaffolds among FDA-approved compounds in the past decade (Figure A). Functionalization protocols for these electron-deficient heteroarenes are well-studied, with the Minisci reaction remaining one of the most powerful tools for accessing medicinally viable heterocyclic candidates.

1.

Importance and challenges of trifluoroacetylation of heteroaromatics through Minisci functionalization.

Alongside evaluating heterocycles, , the introduction of fluorine atoms has also become a prominent strategy in synthesizing therapeutic lead compounds. Considering the prevalence of both functional groups, the incorporation of fluorinated motifs into heterocycles is an extremely desirable transformation. Trifluoromethyl ketones (TFMKs) are of particular interest for their roles as enzymatic inhibitors, as they can act as electrophilic warheads through reversible covalent bonds (Figure B). − Additionally, TFMKs are known to be employed as building blocks toward fluorinated heterocycles and as functional group substitutes for carboxylic acids. −

Minisci reactions are driven by the addition of electron-rich, carbon-centered radicals into electron-deficient, heteroaromatic rings. As such, the direct installation of TFMKs is challenging due to the inherent polarity mismatch of the electron-poor trifluoromethylacyl radical (Figure C). We propose that we can overcome this polarity mismatch by employing a nucleophilic radical. Our previously reported masked acetylating reagent yields a fluorinated acetal radical, which would enable the necessary reactivity parameters for Minisci coupling.

Others have demonstrated addition of the masked ketone into select heterocyclic classes of quinoxalinones, , quinolines, and azauracils (Figure D). These examples are powerful precedents for the functionalization of specific heterocyclic classes. However, these protocols do not address more common heterocyclic scaffolds in medicinal chemistry, such as pyridines and diazines. Herein, we developed a net oxidative approach toward the trifluoroacetylation of heteroaromatics, first through photomediated C–H functionalization, followed by deketalization to yield the desired N-heterocyclic TFMK (Figure E).

Our seminal work focused on the design of masked acyl reagents to install TFMKs into electron-deficient olefins in a redox-neutral approach. In contrast, a Minisci-type reaction necessitates a net oxidative protocol for the incorporation of masked TFMKs into heterocycles by using our proposed reagent design (Table ). We selected 4-(trifluoromethyl)­pyridine as our model substrate and evaluated a variety of differentially modified TFMK precursors (1-CO ₂ H–4-CO ₂ H) to study both reactivity and deketalization. As expected and in agreement with our previous study, no desired product was observed in the reaction with trifluoropyruvic acid 1-CO ₂ H (entry 1). In our prior work, N,O-acetals were used as an alternative to the dioxolane scaffold as these could be readily deketalized with aqueous HCl. However, when we began our initial screenings, to our surprise, the use of N,O-acetal reagent 2-CO ₂ Na was unsuccessful (entry 2) and resulted in no desired product, even with the addition of an acid activator (entry 3). Acylated oxazolidine derivative 3-CO ₂ H was also unsuccessful (entry 4). With the optimized conditions of photocatalyst (Ir­[dF­(CF₃)­ppy]₂(dtbbpy))­PF₆ and oxidant (NH₄)₂S₂O₈ in a 1:1 solvent mixture of DMSO and water irradiated with 34 W Kessil 456 nm LEDs for 18 h, we obtained the best reactivity using O,O-acetal 4-CO ₂ H (77% isolated yield, entry 5). It can be noted that while organophotocatalyst 4CzIPN gave similar results, we opted to move forward with the iridium photocatalyst for ease of purification.

1. TFMK Reagent Screen and Corresponding Results.

entry	conditions	yield (%)
1	1-CO2H instead of 4-CO2H	0
2	2-CO2Na instead of 4-CO2H	0
3	entry 2 with 3.0 equiv of H2SO4	0
4	3-CO2H instead of 4-CO2H	0
5	as shown	90 (77)

^aYields determined by ¹⁹F NMR using trifluorotoluene as an internal standard. The yield in parentheses is the isolated yield after flash column chromatography. See the Supporting Information for additional reaction optimization.

To further understand our experimental outcomes, we used computational methods to evaluate the corresponding electrophilicity indices (ω) of the radicals derived from compounds 1-CO ₂ H–4-CO ₂ H (Figure ). , The values of the radicals from reagents 1-CO ₂ H and 4-CO ₂ H had been previously reported, and our calculated polarities for 1 ^• and 4 ^• were consistent with the literature, calculated as 1.71 and 0.97 eV, respectively. These values serve both to reinforce the mismatched polarity of 1 ^• toward electron-poor partners and to quantify the required nucleophilicity of 4 ^• that ensures excellent reactivity. We found the value of radical 3 ^•, from acylated N,O-derivative 3-CO ₂ H, to be 1.31 eV, which appears to be too electrophilic for addition to N-heterocycles. Radical 2 ^• derived from reagent 2-CO ₂ Na was determined to have a value of 0.80 eV, confirming enhanced nucleophilicity compared to model O,O-acetal derivative 4 ^•. Considering that radical 2 ^• is the most nucleophilic of all radicals and had previous success in coupling to electron-deficient olefins, we hypothesized that our contradictory results were due to differences in reaction conditions compared to our initial study. Our optimal conditions include the use of persulfate that decomposes to the sulfate radical anion (SO₄ ^–•), which is known to possess exceptional single-electron oxidizing power (∼2.25–2.85 V vs SCE). It is known that the more nucleophilic a radical, the more easily oxidized it is to a stabilized carbocation. We propose that this is true in the case of our N,O-acetal reagent. Under oxidative conditions, α-amino radical 2 ^• rapidly oxidizes to carbocation 2 ⁺ , which yields the fluorinated dihydrooxazole species upon deprotonation (Figure ). We were able to experimentally confirm the presence of 2-(trifluoromethyl)-4,5-dihydrooxazole through both independent synthesis and NMR studies under DMSO-only solvent conditions. We anticipated that this dihydrooxazole would be hydrolyzed under the standard aqueous solvent conditions and were able to confirm that we had indeed generated trifluoroacetic acid (TFA) as a major fluorinated byproduct (see the Supporting Information for further details). We believe that these data are key evidence of carbocation formation and further confirm that the use of reagent 2-CO ₂ Na is not feasible under oxidizing conditions.

2.

Electrophilicity indices (ω) of masked radicals and plausible oxidative decomposition pathway.

Following the conclusion of our optimization studies, we went forward with highest-yielding masked reagent 4-CO ₂ H as our TFMK precursor and demonstrated the success of our reaction conditions on a wide scope of heterocycles (Figure ). While our iridium-based photocatalyst conditions worked well for a majority of our substrate scope (Figure , condition A), we found that some extended heteroaromatics demonstrated increased reactivity when subjected to photomediated hypervalent iodine conditions (Figure , condition B) using (bis­(trifluoroacetoxy)­iodo)­benzene (PIFA). With our optimized conditions in hand, we were able to showcase our method on a variety of 4-, 3-, and 2-substituted and electronically diverse pyridines in 13–79% yields (5–12). Notably, functionalization of biologically active (−)-nicotine was showcased in 60% yield with only one regioisomer (9). In the case of 3-phenylpyridine, we were able to successfully isolate the Minisci product as a single regioisomer (10), with a crystal structure confirming regioselectivity at sterically favored position 6 (see the Supporting Information for the crystal structure).

3.

Substrate scope of masked heterocycles. See the Supporting Information for experimental details.

Excitingly, we were able to highlight the coupling of our masked acyl reagent to a wide scope of coveted diazine heterocycles, such as pyrimidines and pyridazines. Fluorofunctionalizations (13–15) of ethyl pyrimidine-5-carboxylate, 4-phenylpyrimidine, and 4,6-dimethylpyrimidine were demonstrated in 44%, 45%, and 78% yields, respectively. Installation of the masked TFMK on 4-methylpyridazine was achieved in an excellent 91% yield (16), with decorated pyridazines 3,6-dichloropyridazine-4-carbonitrile and methyl 6-chloropyridazine-4-carboxylate being accessed in modest yields of 52% (17) and 25% (18), respectively. To probe the extent of our methodology, we screened a variety of extended azole-fused pyridazines, achieving 60% yields for both decorated imidazo­[1,2-b]­pyridazines (19 and 21). We were also successful in the fluorofunctionalization of triazolo­[4,3-b]­pyridazine accessed in a modest 40% yield (20).

Further screening revealed success with quinolines and isoquinolines, both recognized as traditional Minisci coupling partners. We obtained an isolated 63% yield of functionalized 4-methylquinoline (22), a common model substrate for such reactions. We were also able to install our TFMK precursor on 4-chloroisoquinoline and ethyl isoquinoline-3-carboxylate in 42% yield (23) and 54% yield (24), respectively. As our method works exceptionally well for pyrimidines and pyridazines, we were also able to showcase fully substituted pyrazine in 28% yield (25). We achieved a 62% yield (26) of benzylated quinoxalin-2­(1H)-one and a 67% yield (27) of unsubstituted quinazoline. We were also able to achieve an exemplary 92% coupling yield (28) by employing benzothiazole, a traditionally accessible Minisci reagent. Constitutional isomers, phenanthridine and acridine, were functionalized in 83% (29) and 46% (30) yields, respectively. Overall, we were able to install our masked TFMK reagent into a large and diverse scope of N-heterocycles, representing more than 10 heterocyclic classes.

With our trifluoromethylketalization scope in hand, we were excited to showcase select deketalizations to the highly desired trifluoromethyl ketones. Unlike 1,3-dioxolane, the trifluoroketal motif is extremely stable and does not decompose under traditional acidic conditions. , Our understanding of the nature of the enhanced stability of trifluoromethyl ketals is reinforced by our synthesis of 4-CO ₂ H, which necessitates the use of bromoethanol rather than traditional ethylene glycol for its formation. With this insight in hand, we postulated that the deprotection should also proceed through a microscopic reverse reaction to yield a bromoethanol-like byproduct. On this premise, we proposed the use of BBr₃ to be the most reasonable reagent to use for the general deketalization protocol (Figure ). After many trials, we found that rigorously dried, oxygen-free conditions were necessary for continued success and reproducibility. The reactions were performed inside of a glovebox using neat boron tribromide, followed by an aqueous quench outside of the glovebox.

4.

Substrate scope of unmasked heterocycles and proposed deketalization steps. Thermal ellipsoids are shown at the 50% probability level.

With these conditions in hand, we were able to access hydrated TFMK pyridine and pyrimidine derivatives 31-H ₂ O and 32-H ₂ O in 69% and 42% yields, respectively. The deketalization of benzothiazole derivative 28 was executed in an exciting 92% yield (33-H ₂ O). In agreement with previous reports, we obtained a crystal structure of 33-H ₂ O that confirms hydrated TFMK. Finally, phenanthridine-coupled product 29 was unmasked to afford hydrated product 34-H ₂ O in 38% yield. Interestingly, during our deketalization study with substrate 29, we observed a new species in the crude reaction mixture that did not correspond to the final product. From the crude reaction, we were able to fully and unambiguously characterize this new species by crystal structure as hemiacetal complex 29-BBr ₃ -int. Excitingly, this intermediate contains a bromoethanol motif, again supporting the unusual stability of fluorinated hemiacetals and our microscopic proposal for ease of deketalization.

Based on our observations, we believe that after the addition of boron tribromide boron will form a complex with the starting material to yield proposed intermediate 29-BBr ₃ . This intermediate can rearrange to form stabilized hemiacetal 29-BBr ₃ -int, which progresses to final product 34-H ₂ O after silica gel workup. The isolation and identification of this stable boron intermediate were key in understanding the challenges associated with deketalizations on nitrogen-containing scaffolds.

In conclusion, we have developed a method to synthesize N-heterocyclic trifluoromethyl ketones (TFMKs) by employing our masked acyl radical reagent. We harnessed photoredox catalysis to install our TFMK precursor into more than 20 substrates encompassing more than 10 heterocyclic classes, including pyridines and diazines. We have also elucidated the potential intermediates in the Lewis acid-mediated deketalization of our trifluoromethylketal masking group. We are excited to provide the greater synthetic community with a general method to install TFMKs into biologically relevant small molecule heterocyclic scaffolds.

The data underlying this study are available in the published article and its .

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.orglett.6c00477.

• Experimental procedures, characterization data, HRMS, XRD data, and ¹H, ¹³C, and ¹⁹F nuclear magnetic resonance spectra of all compounds (PDF)

• Radical polarity computational data (XLSX)

The authors declare no competing financial interest.