Pyridines are privileged cores frequently found in a wide-range of pharmaceuticals, functional materials, and natural products [
1]. Accordingly, myriads of synthetic methods, such as encountered in Minisci-type reactions, have been developed to rapidly functionalize such scaffolds, and have provided ready access to the core structures of many medicinally important derivatives [
2-
4]. Despite their synthetic versatility, these methods demand stoichiometric amounts of Brønsted acids, and frequently resulted in mixtures of undesired byproducts due to competitive overalkylation [
5,
6]. Therefore, there is a growing demand for methods that can functionalize pyridine derivatives in a predictable and controllable manner. In this context, the advent of photocatalyzed reactions utilizing the bench-stable
N-alkoxypyridinium salts as versatile pyridine surrogates marked a significant advance, which offers direct access to functionalized pyridine derivatives [
7-
10]. Relative to simple pyridines, the readily available
N-alkoxypyridinium salts with low-lying LUMOs exhibit enhanced reactivity under acid-free conditions [
11,
12], and they have emerged as promising alkoxy radical precursors [
13,
14]. Indeed, efficient methods for functionalization by using
N-alkoxypyridinium salts have been developed, and so demonstrating that the desired radical processes can be initiated by alkoxy radicals generated
via light-promoted homolysis of the N–O bonds contained within
N-alkoxypyridinium systems (
Scheme 1a) [
15-
20]. For examples, Hong and co-workers reported various alkoxy radical-mediated intermolecular hydrogen atom transfer (HAT) strategies for the site-selective functionalization pyridines [
21,
22]. We recently disclosed a photochemical synthesis of unsymmetrical methylene-bridged bis-heterocycles involving an
in situ generated O-centered radical-mediated HAT process using
N-methoxypyridinium salts [
23]. To broaden the scope of the synthetic strategy encapsulated by these pioneering studies, we considered whether methoxy radicals are capable of engaging in a radical relay process with trimethylsilyl azide (TMSN
3) to generate azidyl radical, which could promote a multicomponent Minisci reaction.
Organic azides are used as highly powerful and valuable building blocks in synthetic organic chemistry because they can undergo a variety of chemical transformations [
24], such as Staudinger reduction [
25,
26], “click” reactions [
27], and others [
28]. Traditional approaches for the generation and use of azidyl radical often depend on stoichiometric oxidant [
29-
33]. In this context, the development of an environmentally friendly and atom-economical process for regioselective azidation is greatly important [
34,
35]. Inspired by our recent interest in developing alkoxy radical-relay process promoted by visible light [
36]. We speculated, as shown in
Scheme 1b, that the reactive azidyl radical might be generated from TMSN
3 via Si−N
3 bond activation mediated by
in situ generated methoxy radical. The alkyl radical intermediate was then formed as a result of the addition of azidyl radical to the alkene, which may set the stage for a second radical addition to the pyridine ring and so representing a potentially versatile extension of the classic Minisci reaction. To apply the strategy for such a reaction sequence, the alkyl radical intermediate should need to react preferentially with the pyridyl moiety of
N-methoxypyridinium salts over radical−radical coupling with azidyl radical to generate the diazidation byproduct. In our working hypothesis, there is a large dynamic driving force for the formation of azidopyridylation prodcuts because of the relatively high concentration and strong electrophilic nature of
N-methoxypyridinium salts [
9,
37]. Herein we detail the successful implementation of such a sequence. This method offers a facile access to valuable azidyl- and pyridyl-containing frameworks under mild reaction conditions.
To investigate the viability of the proposed radical relay reaction sequence, unactivated alkene
1a,
N-methoxypyridinium salt
2a, and TMSN
3 were selected as the model substrates under the irradiation of Green LEDs (
λmax = 525 nm) to identify workable reaction conditions (see Supporting information for further details). After evaluating various photosensitizers, as shown in
Table 1, we were pleased to find that the use of (Ir[dF(F)ppy]
2(dCF
3))PF
6 as the photocatalyst in acetonitrile at room temperature afforded the
β-pyridyl azide
4a in 48% yield (entry 6), thus indicating that the proposed overall process involving a radical azidation and subsequent pyridylation sequence can be achieved effectively. Among the solvents screened, acetonitrile was optimal since it showed higher efficiency (entries 6-8). The choice of base was critical for the reaction efficiency, and NaHCO
3 was proved to be the optimum one by delivering
4a in 69% yield (entry 11). From the viewpoint of atom economy, oxidant free and operationally simple protocols are highly appealing. Thus, entry 11 was chosen as the optimal reaction conditions, even though the addition of oxidant resulted in a higher yield of
4a (entry 12). Interestingly, a low yield of product
4a (12%) was also obtained in the absence of the photocatalyst, indication that electron-donors might form EDA complexes with pyridinium salts and ultimately leading to the azidopyridylation of alkenes (entry 14) [
38-
41]. A control experiment confirmed that visible light is crucial (entry 15). Consistent with the proposed radical mechanism for the reaction, no desired product was observed when the reaction was conducted in the presence of TEMPO (entry 16).
Once optimal reaction conditions had been established, we investigated the substrate scope to demonstrate the versatility of the current photocatalytic tandem reaction. Initially, a variety of alkene substrates were subjected to the optimized reaction conditions, as revealed in
Scheme 2. Substrates bearing either electron-withdrawing or electron-donating groups on the aromatic ring proceeded evenly in the reaction with
2a and afforded the expected
β-pyridyl azides (
4a−
4i). In addition, unactivated alkenes bearing esters (
1j−
1m), ethers (
1n−
1o), oxy-isoindolinedione (
1p), phenyl (
1q), alkyls (
1r−
1t), ketone (
1u), amides (
1v−
1x), and trimethylsilane (
1y) were well tolerated in this reaction. Later, the molecular structure of
4p was unambiguously confirmed by single-crystal X-ray analysis (see Supporting information for details). Gratifyingly, bromide was tolerable in this transformation to provide the azidopyridylation product
4z, thus allowing further modification. Notably, this methodology was found to be successful with enol ether (
1aa) to yield the desired product. Moreover, to demonstrate the broad applicability of our developed method, the utility of the protocol in late-stage functionalisations was explored. For example, theobromine, fenofibric acid, (-)-menthol, 4-methylumbelliferone, a flavone-derivative, and (+)-sulbactam derivatives were all successfully engaged to afford the corresponding products (
4ab-
4ag), indicating the potential of this method for late-stage application.
To further test the applicability of the radical relay process, a range of other heteroarenium salts to serve as coupling partners was explored, and the results are presented in
Scheme 3. Reactions of pyridines bearing C-4 substituents (cyano, trifluoromethyl, and methyl ketone) proceeded smoothly to furnish the desired products (
5a–
5c). Pyridine bearing ester groups at the C-3 position deliver a 3:1 mixture compound
5d and its C-2 regioisomer in 71% combined yield. When substrates bearing electron-withdrawing groups on the pyridyl ring at the C-2 position (cyano, and trifluoromethyl) were subjected to the reaction, C-4 alkylation products
5e and
5f were exclusively obtained in 75% and 72% yields, respectively. Moreover, expanding the scope from the pyridines to the quinoline and isoquinoline systems were also adaptable. Specifically, quinolinium substituted with the cyano, methyl ketone, methoxy, fluoro, bromo, and phenyl at various positions were tolerated in this transformation, leading to the formation of the desired products (
5h–
5n). When isoquinolinium was used as the heteroarene source, the C-1 alkylation products
5o and
5p were formed, highlighting the broad functional group tolerance.
With this convenient approach to accessing
β-pyridyl azides, further transformations of the product were surveyed to demonstrate the synthetic utility of this radical azidopyridylation protocol (
Scheme 4). For instance, triazole
6 could be easily achieved from the derived
β-pyridyl azide
4a and phenylacetylene through a copper catalyzed azide−alkyne [3 + 2] click reaction (Eq. 1) [
42]. Treatment of
4a with acetylacetone in the presence of K
2CO
3 then delivering triazole product
7 in 82% yield (Eq. 2). Further,
β-pyridyl azide
4a could be converted into
β-pyridyl amine
8 via a phosphine-mediated Staudinger reduction (Eq. 3) [
26].
The ready access to the above-mentioned heterocycles offered by the protocols reported herein presented an opportunity to examine the cytotoxic effects of these selected compounds against HeLa, an immortalized cell line. As revealed in
Fig. 1 (see Supporting information for further details), the compounds
4a, 6, 7 and
8 displayed IC
50 values of 0.12 ± 0.02, 0.60 ± 0.08, 11.21 ± 0.31, and 2.78 ± 0.09 µmol/L, respectively. Therefore, these preliminary cell viability data suggest that these readily-derived compounds have potential as cancer therapeutic candidates.
To elucidate the reaction mechanism, a series of Stern−Volmer quenching experiments were performed [
43,
44], revealing that the quenching rate was directly proportional to the concentration of pyridinium salt
2a. On the other hand, the photoexcited Ir*catalyst was not quenched by
1a, TMSN
3, and NaHCO
3. The model reaction quantum yield was found to be
= 0.8, which indicates that the closed photoredox catalytic cycle pathway is particularly efficient (see Supporting information for details).
A possible mechanistic pathway is depicted in
Scheme 5. Initially, the generated excited-state Ir* species facilitates the single-electron transfer (SET) reduction of pyridinium salt
2 to liberate the methoxy radical
Ⅰ, which appears to react with TMSN
3 to generate the reactive azidyl radical
Ⅱ. Subsequently, the azidyl radical engages in an addition to the alkene
1 to provide radical intermediate
Ⅲ. The resultant radical
Ⅲ attacks pyridinium salt
2 in a Minisci-type process to form unstable radical cation
Ⅳ. Finally, the radical cation
Ⅳ undergoes deprotonation followed by homolytic cleavage of the N–O bond to furnish the observed
β-pyridyl azide and methoxy radical
Ⅰ. The released methoxy radical
Ⅰ then promotes the radical-chain process with TMSN
3, thus continuing the cycle.
In conclusion, we have presented a visible light-mediated three-component azidopyridylation of unactivated alkenes with accessible TMSN3 and pyridinium salts via an in situ generated methoxy radical-initiated relay reaction. This modular protocol features mild conditions, exhibits broad compatibility of functional groups and substrate scope, and highlights its potential in the late-stage diversification of biorelevent molecules. Employing this approach, a variety of synthetically useful azidyl- and pyridyl-containing building blocks were prepared. The resulting β-pyridyl azides could be efficiently converted into triazoles and β-pyridyl amine. This advancement is anticipated to have significant application potential in organic synthesis. Further mechanistic research and synthetic utilization of this chemistry are underway in our laboratories.
PDF
Full
64
36
关闭全屏
No. 86 Xueyuan South Road, Haidian District, Beijing
010-62662699(Operation and Technical Support)
qkjq@cast.org.cn
