Photoinduced Azidosulfonylative Cyclization of 1,6-Dienes with Sulfonyl Azides

Abstract

Electrophilic azides serve as valuable reagents in visible-light-induced photocatalytic processes. While the reactivity of triplet nitrenes formed via energy transfer to electrophilic azides has been relatively well studied, the synthetic potential of nitrene radical anions generated through single electron transfer remains less explored. Herein, we report a photoinduced azidosulfonylative cyclization of 1,6-dienes with sulfonyl azides in the presence of 9,10-diphenylanthracene as a photoreductant, providing access to functionalized five-membered heterocycles. Mechanistic studies suggest that nitrene radical anions are responsible for the generation of sulfonyl radicals through hydrogen atom transfer from N,N-dimethylformamide (DMF) solvent.

Introduction

Organic azides are an important class of compounds due to broad applications of azide–alkyne [3 + 2] cycloaddition and Staudinger ligation in various fields, including chemical biology, medicinal chemistry, and materials science.^1–4) From a synthetic point of view, azides serve as valuable precursors to prepare nitrogen-containing molecules.

The introduction of electron-withdrawing groups, such as sulfonyl and acyl groups, enhances the electrophilicity of the azides, leading to unique reactivities toward nucleophiles.⁵⁾ Electrophilic azides have also found utility in visible-light-mediated photocatalysis, which has emerged as a new paradigm in organic synthesis. The most notable examples involve the aziridination of alkenes with triplet nitrenes formed via energy transfer (EnT) from the triplet excited state of a photocatalyst to electrophilic azides (Chart 1A, top). The pioneering work of Yoon demonstrated that EnT to 2,2,2-trichloroethyl azidoformate (TrocN₃) enabled the aziridination of a wide range of alkenes.⁶⁾ In addition to other azidoformates, such as CbzN₃ and BocN₃, acyl, sulfonyl, and trifluoromethyl azides have also been employed in triplet-nitrene-mediated alkene aziridinations.^7–9) Triplet nitrenes have also been proposed to undergo C(sp³)–H insertion.^10,11) Building on these developments, our group recently reported the synthesis of N-sulfonyl sulfilimines by imination of sulfides with triplet nitrenes generated from sulfonyl azides¹²⁾ (Chart 1A, bottom).

Chart 1. Electrophilic Azides in Photocatalysis

Another strategy for photocatalytic activation of electrophilic azides relies on single electron transfer (SET) reduction, enabling access to nitrene radical anion intermediates. However, their synthetic application has been less explored compared with that of triplet nitrenes. Very recently, Bouwman and Codée demonstrated that alkene aziridination can also proceed via nitrene radical anions generated from sulfonyl azides through reductive SET, offering a mechanistic alternative to triplet nitrenes.¹³⁾ Additionally, nitrene radical anions derived from iminoiodinanes have also been employed in alkene aziridinations.^14–16)

In this work, we aimed to expand the synthetic utility of nitrene radical anions and initially sought to develop a synthesis of fused bicyclic pyrrolidine A from 1,6-dienes via a radical–polar crossover mechanism, motivated by the recognized importance of sp³-rich scaffolds in medicinal chemistry¹⁷⁾ (Chart 1B). In our envisaged mechanism, addition of the nitrene radical anions to the terminal alkene in 1,6-dienes followed by the 5-exo-trig radical cyclization was expected to give alkyl radical C. Subsequent oxidation of C by the oxidized state of the photocatalyst (PC^•+) would form carbocations, which could undergo intramolecular cyclization to furnish fused pyrrolidine A. Notably, Koenigs recently developed a related [2 + 2 + 1] cycloaddition of 1,6-dienes with triplet nitrenes generated via direct photolysis of iminoiodinanes.¹⁸⁾ However, our preliminary attempts led to the formation of monocyclic product B, which appears to arise through a different mechanism involving the addition of sulfonyl radicals to the 1,6-dienes to form tertiary alkyl radical D. Intrigued by this unexpected result, we proceeded to investigate the substrate scope and elucidate the underlying mechanism, particularly the generation pathway of sulfonyl radicals. Similar azidosulfonylation of 1,6-dienes has been previously reported by Renaud, employing a classical radical generation with substoichiometric amounts of a radical initiator under irradiation with a high-energy sun lamp.¹⁹⁾ Herein, we describe a photoreductive approach to the azidosulfonylative cyclization of 1,6-dienes through a distinct mechanism involving nitrene radical anions.

Results and Discussion

We investigated the reaction of diene 1a with p-toluenesulfonyl azide (2a) in the presence of 9,10-diphenylanthracene (DPA) as a photocatalyst under 390 nm light-emitting diode (LED) irradiation (Table 1). We chose this organic photocatalyst because of the capability for single electron reduction of 2a (E_1/2(DPA^•+/DPA*) = –1.77 V and E_1/2(2a/2a^•−) = –1.22 V vs. SCE in MeCN)^12,20) and lower economic cost compared with transition metal-based photocatalysts. When the reaction was carried out in acetonitrile, pyrrolidine 3aa was obtained in 39% yield as an inseparable 1 : 1 mixture of diastereomers (entry 1). A screening of the solvents revealed that solvents had remarkable impact on the product yields (entries 2–5). Among the tested solvents, N,N-dimethylformamide (DMF) was identified as the most suitable solvent, giving 3aa in 89% yield with a 2 : 1 diastereomeric ratio (entry 5). Nuclear Overhauser effect (NOE) spectroscopy revealed that the major diastereomer has a cis geometry (Supplementary Chart S4), which can be attributed to favorable orbital or electrostatic interactions in the transition state.²¹⁾ The increased concentration from 0.1 to 0.2 M did not affect the product yield, which was determined as the optimal reaction conditions (entry 6). No formation of 3aa was observed when the reaction was performed in the dark, confirming that light irradiation was essential for this reaction (entry 7). Even without the photocatalyst, the product was obtained in moderate yield (entry 8); however, prolonged irradiation time did not improve the yield significantly (entry 9), thus the photocatalyst is required to achieve high yields.

Table 1. Optimization of Reaction Conditions

Entry	Deviations from above conditions	3aa		Remaining 1a (%)a)
		Yield (%)a)	Dr
1	MeCN (0.1 M)	39	1 : 1	27
2	CH2Cl2 (0.1 M)	7	1 : 1	72
3	THF (0.1 M)	60	2 : 1	10
4	MeOH (0.1 M)	80	1 : 1	5
5	DMF (0.1 M)	89	2 : 1	ND
6	None	89 (83)	2 : 1	ND
7	In the dark	ND	1 : 1	>95
8	Without DPA	51	2 : 1	22
9	Without DPA, 24 h	54	2 : 1	5

Reaction conditions: 1a (0.2 mmol), 2a (2 equiv.), DPA (2 mol%), solvent (1 mL), 390 nm Kessil LED, under N₂, rt, 15 h. a) Yields were determined by ¹H-NMR using dimethyl sulfone as an internal standard. Isolated yield is provided in the parentheses. ND: not detected. Ts: p-toluenesulfonyl.

With the optimized conditions in hand, we investigated the substrate scope with respect to sulfonyl azides 2 using diene 1a as a model substrate (Fig. 1A). Unsubstituted benzene sulfonyl azide provided pyrrolidine 3ab in 93% yield with a 2 : 1 diastereoselectivity. Introduction of various functional groups onto the benzene ring were tolerated, including bromo (3ac), methoxy (3ad), trifluoromethyl (3ae), ester (3af, 3ah, and 3ai), and amide (3ag). Use of 3-pyridyl sulfonyl azide provided 3aj in 83% yield. However, employment of benzylsulfonyl azide did not afford the desired product (3ak), and the starting material (1a) was recovered almost quantitatively. This result can be ascribed to the instability of the benzylsulfonyl radical intermediate, which fragments into benzyl radical and sulfur dioxide.²²⁾

Fig. 1. Substrate Scope

Reaction conditions: 1 or 4 (0.2 mmol), 2 (2 equiv.), DPA (2 mol%), DMF (1 mL), 390 nm Kessil LED, under N₂, rt, 15 h. a) Sulfonyl azide 2 (3 equiv.). b) DPA (5 mol%). c) 36 h. ND: not detected.

We next evaluated the scope of the 1,6-dienes (Fig. 1B). Our investigation focused on C7-dimethylated 1,6-diene derivatives, guided by the previous work of Mantrand and Renaud on a similar transformation.¹⁹⁾ In their report, a clear trend in product yields was observed based on the C7-substitution pattern. Specifically, high yields were obtained from C7-dialkylated 1,6-dienes, whereas the use of mono-alkylated ones resulted in lower yields, and 1,6-dienes lacking alkyl substituents at the C7 position failed to produce any desired product. This superiority of the C7-dialkylated 1,6-dienes was attributed to the formation of a nucleophilic tertiary alkyl radical, which undergoes an efficient reaction with the electrophilic sulfonyl azide.

The replacement of the sulfonamide moiety of 1a with an oxygen atom afforded tetrahydrofuran 3ba in 47% yield. A substrate with a geminal diester tether afforded cyclopentane 3ca in 67% yield with a higher diastereoselectivity (dr 6 : 1). The installation of a methyl group at the C2 position of 1a did not affect the product formation, including the diastereomeric ratio, which provided the desired product 3da in high yield. This cyclization was also applicable to 1,6-enynes 4 by using higher catalyst loading, which provided pyrrolidine 5aa and tetrahydrofuran 5ba in moderate yields. The Z-configurations of the products were unambiguously determined by NOE spectroscopy (Supplementary Charts S5, S6). On the other hand, simple olefins such as styrene and methylenecyclohexane did not undergo the azidosulfonylation, and the starting materials were recovered. This inertness is presumably due to the fast fragmentation of β-sulfonyl alkyl radical intermediates, which reverts to the starting alkenes and sulfonyl radicals.¹⁹⁾

To gain a better understanding of the origin of sulfonyl radicals, we performed density functional theory (DFT) calculations (Fig. 2). We considered two possible decomposition pathways of radical anion 2a^•−, which is generated through single electron reduction of p-toluenesulfonyl azide (2a) by the photoexcited catalyst. Radical anion 2a^•− can directly generate sulfonyl radical 6a with the release of azide anion via TS1 (ΔG^‡ = 14.5 kcal/mol). Alternatively, radical anion 2a^•− can fragment into nitrene radical anion 7a and nitrogen gas, which proceeds via TS2 with a low energy barrier (ΔG^‡ = 2.2 kcal/mol). The latter pathway is both kinetically and thermodynamically favored (ΔΔG^‡ = –12.3 kcal/mol, ΔΔG = –21.3 kcal/mol), suggesting that sulfonyl radicals are formed indirectly through the intermediacy of nitrene radical anions.

Fig. 2. Computational Investigations on the Decomposition Pathways of Radical Anion 2a^•–

DFT calculations were performed at the SMD(DMF)-UB3LYP-D3/6-311++G(2d,p)//UB3LYP-D3/6-31 + G(d) level of theory.

Based on our experimental results, a plausible reaction mechanism is shown in Chart 2. Sulfonyl azide 2 undergoes SET from highly reducing photoexcited DPA, generating radical anion 2^•–. Considering that DPA has lower triplet energy (40.8 kcal/mol) than sulfonyl azide 2a (49.6 kcal/mol),^12,20) we postulated that an EnT pathway was unlikely. Radical anion 2^•– fragments into nitrogen gas and nitrene radical anion 7. This species then would engage in hydrogen atom transfer (HAT) from the DMF solvent, abstracting a hydrogen atom from either the methyl or formyl group to yield α-amino alkyl radical 8 or carbamoyl radical 8′, respectively. These nucleophilic radicals subsequently react with sulfonyl azide 2 to generate sulfonyl radical 6 along with the formation of alkyl azide 9 or carbamoyl azide 9′. A species with the expected molecular weight corresponding to these isomers was detected by LC/MS analysis of a crude reaction mixture of 1a with 2a (Supplementary Chart S1). Although we assume that all tested solvents (Table 1) underwent hydrogen atom abstraction by radical anion 2^•−, the use of solvents with hydridic C–H bonds (i.e., α-amino and α-oxy C–H bonds) such as DMF, methanol, and tetrahydrofuran (THF) was crucial to achieving high product yields. This observation can be explained by the high nucleophilicity of the resulting α-heteroatom alkyl radicals (e.g., 8), which react smoothly with electrophilic sulfonyl azide 2. Given that the reaction proceeded to some extent even in the absence of the photocatalyst (Table 1, entry 8), direct photolysis of sulfonyl azide 2 could generate triplet nitrene 10, which also abstracts a hydrogen atom from DMF. This reactivity is consistent with a precedent where triplet nitrene 10 abstracts a hydridic hydrogen from solvents such as THF.²³⁾ Addition of sulfonyl radical 6 to diene 1 at the sterically less hindered position gives β-sulfonyl alkyl radical 11. The subsequent 5-exo-trig radical cyclization produces tertiary alkyl radical 12. Azidation of 12 with sulfonyl azide 2 affords product 3 and regenerates sulfonyl radical 6 to establish a radical chain process.²⁴⁾

Chart 2. Proposed Mechanism

To investigate the chain nature of the reaction, we conducted a light on/off experiment using 1a and 2a, which showed that formation of 3aa occurred only during photoirradiation (Supplementary Chart S2). However, such an experiment is often considered unreliable for ruling out a chain mechanism.²⁵⁾ Therefore, we next determined that the quantum yield of the model reaction of 1a with 2a is 0.28 (see Supplementary Materials). Although this relatively small value does not confirm a chain mechanism, it does not rule out the possibility either, particularly when the radical initiation step is inefficient.²⁶⁾ The initiation of sulfonyl radical 6 involves multiple steps from sulfonyl azide 2, which could account for the observed low quantum yield.

Conclusion

In summary, we have developed a photoinduced azidosulfonylative cyclization of 1,6-dienes with sulfonyl azides, providing access to functionalized five-membered heterocycles. Mechanistic studies suggest that the transformation proceeds via SET reduction of sulfonyl azides to nitrene radical anions, which subsequently generate sulfonyl radicals through HAT from the DMF solvent. We anticipate that this method will inspire further exploration of nitrene radical anion chemistry toward the development of new synthetic transformations.

Experimental

General Procedure

A dried 5 mL microwave vial equipped with a stirrer bar was charged with diene 1 (0.20 mmol, 1.0 equivalent (equiv.)), sulfonyl azide 2 (0.40 mmol, 2.0 equiv.), and 9,10-DPA (4.0 µmol, 2 mol%). The vial was sealed with a septum cap, evacuated and backfilled with N₂ at least three times, and degassed DMF (1.0 mL) was added via syringe. The vial was additionally sealed with parafilm and stirred under the irradiation with a 52 W Kessil 390 nm LED (max intensity) at room temperature for 15 h. The resulting mixture was concentrated under reduced pressure and purified by flash column chromatography on silica gel to afford the desired product.

Acknowledgments

This work was supported by JSPS KAKENHI (Grant Numbers: JP22K15248 and JP25K09894), JST SPRING (JPMJSP2110), and Research Support Project for Life Science and Drug Discovery (Basis for Supporting Innovative Drug Discovery and Life Science Research (BINDS)) from the Japan Agency for Medical Research and Development (AMED) (Grant Numbers: JP22ama121034 and JP22ama121042).

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

References

• 1)  Jewett J. C., Bertozzi C. R., Chem. Soc. Rev., 39, 1272–1279 (2010).
• 2)  Lallana E., Riguera R., Fernandez-Megia E., Angew. Chem. Int. Ed., 50, 8794–8804 (2011).
• 3)  Thirumurugan P., Matosiuk D., Jozwiak K., Chem. Rev., 113, 4905–4979 (2013).
• 4)  Bednarek C., Wehl I., Jung N., Schepers U., Bräse S., Chem. Rev., 120, 4301–4354 (2020).
• 5)  Xie S., Sundhoro M., Houk K. N., Yan M., Acc. Chem. Res., 53, 937–948 (2020).
• 6)  Scholz S. O., Farney E. P., Kim S., Bates D. M., Yoon T. P., Angew. Chem. Int. Ed., 55, 2239–2242 (2016).
• 7)  Meyer A. R., Popescu M. V., Sau A., Damrauer N. H., Paton R. S., Yoon T. P., ACS Catal., 14, 12310–12317 (2024).
• 8)  Dam D., Lagerweij N. R., Janmaat K. M., Kok K., Bouwman E., Codée J. D. C., J. Org. Chem., 89, 3251–3258 (2024).
• 9)  Baris N., Dračínský M., Tarábek J., Filgas J., Slavíček P., Ludvíková L., Boháčová S., Slanina T., Klepetářová B., Beier P., Angew. Chem. Int. Ed., 63, e202315162 (2024).
• 10)  Jung H., Keum H., Kweon J., Chang S., J. Am. Chem. Soc., 142, 5811–5818 (2020).
• 11)  Bouayad-Gervais S., Nielsen C. D.-T., Turksoy A., Sperger T., Deckers K., Schoenebeck F., J. Am. Chem. Soc., 144, 6100–6106 (2022).
• 12)  Arichi N., Amano T., Wu S., Inuki S., Ohno H., Chem. Eur. J., 30, e202401842 (2024).
• 13)  Dam D., Schoenmakers J., Bouwman E., Codée J. D. C., J. Org. Chem., 90, 6577–6583 (2025).
• 14)  Guo Y., Pei C., Jana S., Koenigs R. M., ACS Catal., 11, 337–342 (2021).
• 15)  Guo Y., Pei C., Koenigs R. M., Nat. Commun., 13, 86 (2022).
• 16)  Srivastava O. S., Anand V., Rastogi N., Asian J. Org. Chem., 12, e202300415 (2023).
• 17)  Lovering F., Bikker J., Humblet C., J. Med. Chem., 52, 6752–6756 (2009).
• 18)  Li F., Zhu W. F., Empel C., Datsenko O., Kumar A., Xu Y., Ehrler J. H. M., Atodiresei I., Knapp S., Mykhailiuk P. K., Proschak E., Koenigs R. M., Science, 383, 498–503 (2024).
• 19)  Mantrand N., Renaud P., Tetrahedron, 64, 11860–11864 (2008).
• 20)  Neumeier M., Chakraborty U., Schaarschmidt D., de la Pena O’Shea V., Perez-Ruiz R., Jacobi von Wangelin A., Angew. Chem. Int. Ed., 59, 13473–13478 (2020).
• 21)  Beckwith A. L. J., Tetrahedron, 37, 3073–3100 (1981).
• 22)  Truce W. E., Heuring D. L., J. Org. Chem., 39, 245–246 (1974).
• 23)  Zhu S., Pathigoolla A., Lowe G., Walsh D. A., Cooper M., Lewis W., Lam H. W., Chem. Eur. J., 23, 17598–17604 (2017).
• 24)  Ollivier C., Renaud P., J. Am. Chem. Soc., 122, 6496–6497 (2000).
• 25)  Cismesia M. A., Yoon T. P., Chem. Sci., 6, 5426–5434 (2015).
• 26)  Scaiano J. C., Chem. Soc. Rev., 52, 6330–6343 (2023).