Aziridine Group Transfer via Transient N-Aziridinyl Radicals

Abstract

Aziridines are the smallest nitrogen-containing heterocycles. Strain-enhanced electrophilicity renders aziridines useful synthetic intermediates and gives rise to biological activity. Classical aziridine syntheses—based on either [2 + 1] cycloadditions or intramolecular substitution chemistry—assemble aziridines from acyclic precursors. Here, we introduce N-aziridinyl radicals as a reactive intermediate that enables the transfer of intact aziridine fragments in organic synthesis. Transient N-aziridinyl radicals are generated by the reductive activation of N-pyridinium aziridines and are directly characterized by spin-trapped EPR spectroscopy. In the presence of O₂, N-aziridinyl radicals are added to styrenyl olefins to afford 1,2-hydroxyaziridination products. These results establish aziridinyl radicals as new reactive intermediates in synthetic chemistry and demonstrate aziridine group transfer as a viable synthetic disconnection.

Aziridines are important synthetic building blocks and represent electrophilic pharmacophores in a variety of organic small molecule therapeutics and natural products (Figure 1a).¹ Aziridines are typically constructed via [2 + 1] cycloadditions between either olefins with nitrene equivalents or imines with carbene equivalents, or via intramolecular nucleophilic substitution chemistry within prefunctionalized substrates (Figure 1b).²N-Alkylation and metal-catalyzed C–N cross-coupling reactions provide opportunities to functionalize the exocyclic N–H valence of preformed aziridines;³ however, ring-opening chemistry to deliver 1,2-aminofunctionalization products is often observed during these transformations.⁴ Methods to transfer intact aziridines to relatively unfunctionalized substrates, such as C–H bonds or olefins, are currently unavailable.

Figure 1.

(a) Selected aziridine-containing natural products. (b) Classical synthetic disconnections for aziridines and N-functionalization chemistry. (c) Here, we report the synthesis and reactivity of N-aziridinyl radicals, which engage in intermolecular olefin addition chemistry. PC = photocatalyst.

Aziridine transfer to olefins via 1,2-aziridine functionalization would complement extant nitrene transfer reactions and provide new disconnections in aziridination chemistry. Based on the burgeoning literature of N-centered radical addition to olefins,^5,6 we envisioned that access to N-aziridinyl radicals would enable aziridine transfer chemistry. Given the strength of aziridine N–H bonds (∼92 kcal/mol)⁷ in comparison to the C–N bonds of the strained three-membered ring (∼54 kcal/mol),⁸ we viewed direct generation of N-aziridinyl radicals from N–H precursors as unlikely. Inspired by strategies to access N-centered radicals by single-electron transfer (SET) between R₂N–X (LG = −Cl, −Br, −O₂CR) reagents and either transition metal catalysts⁹ or photocatalysts,¹⁰ we speculated that an N-substituted aziridine featuring a (photo)cleavable N-substituent could provide selective entry to aziridine radical chemistry.

We previously developed N-pyridinium aziridines as electrophiles in C–N cross coupling chemistry.¹¹ We envisioned facile access to N-aziridinyl radicals from these precursors via reductive activation of the N–N bond.^12,13 Here, we demonstrate that the reductive photoactivation of N-pyridinium aziridines generates N-aziridinyl radicals. DFT studies indicate the N-aziridinyl radical is planar with the unpaired spin in a p-orbital, which we hypothesize results in electrophilic reactivity.¹⁴ In the synthetic context, the transient N-aziridinyl radicals can be trapped with olefinic substrates in the presence of O₂ to afford 1,2-hydroxyaziridination products (Figure 1c). Together, these results establish N-aziridinyl radicals as a new reactive intermediate for synthetic chemistry and demonstrate aziridine group-transfer as a viable synthetic disconnection.

We initiated the development of aziridine transfer chemistry with N-pyridinium aziridine 2a (tpp = triphenylpyridinium), which displays a reductive electrochemical feature at −0.85 V vs Fc⁺/Fc.¹⁵ We hypothesized that reductive quenching of an appropriate photocatalyst would reveal an N-aziridinyl radical. Accordingly, photolysis of a MeCN solution of 2a in the presence of Ir(ppy)₃ (Ir(III)*/Ir(IV) = −1.73 V vs Fc⁺/Fc),¹⁶ triethyl amine, and radical acceptor 1 afforded 3, the product of aziridine transfer (eq 1).

1

With the conditions in hand to promote N-aziridinyl radical transfer, we sought to engage this novel fragment in olefin addition chemistry. To this end, photolysis (blue LED) of 2a, styrene, Ir(ppy)₃ (1.0 mol %), and Et₃N in an O₂-saturated H₂O:MeCN solution afforded compound 5a, the product of olefin 1,2-hydroxyaziridination, in 35% yield (1:1 dr, Table 1, entry 1). Addition of LiBr increased the yield of hydroxyazirdine 5a to 42% (50 mol % LiBr, entry 2) and 69% (1.0 equiv LiBr, entry 3). LiOTf and LiBF₄ also have a positive impact on the efficiency of hydroxyaziridination; addition of [TBA]Br has no impact on reaction efficiency. These observations suggest Li⁺ serves as a Lewis acid while Br^– does not impact the developed reaction. No aziridine transfer products were obtained in the absence of Et₃N, a photocatalyst, or light (entries 4–6). Changes to the relative stoichiometries of olefin 4a and aziridine radical precursor 2a did not improve the efficiency of olefin hydroxyaziridination (entries 7 and 8). Using unsubstituted N-pyridinium aziridine 2a′ afforded only a 29% yield of 5a, which is consistent with the more negative reduction potential of unsubstituted pyridinium aziridines as compared to 2,4,6-triphenylpyridinium aziridines.¹⁵ In the absence of water, the reaction afforded 62% 5a but was less clean which complicated purification, and thus the optimized conditions utilized a 1:1 H₂O:MeCN mixture (see Tables S1–S5 for optimization details). Compound 5a can be envisioned as the product of epoxide opening with an aziridine nucleophile, which, to our knowledge, is unknown.

Table 1. Olefin 1,2-Hydroxyaziridination^a.

entry	variation from standard conditions	yieldb (%)
1	no LiBr	35
2	LiBr (50 mol %)	42
3	none	69 (65)
4	no PC	0
5	no Et3N	0
6	no light	0
7	4a (1.0 equiv), 2a (1.5 equiv)	41
8	4a (3.0 equiv) of styrene	45
9	py-aziridine was used instead of tpp-aziridine	29
10	MeCN	62

^aOptimized conditions: 4a (0.15 mmol), 2a (0.10 mmol), Ir(ppy)₃ (1.0 mol %), Et₃N (0.2 mmol), LiBr (0.10 mmol) in MeCN:H₂O (1:1, 2.0 mL) under blue LED irradiation.

^bNMR yields (isolated yield).

With conditions for N-aziridinyl radical transfer, we used radical precursor 2a to canvass the reactivity of this fragment against various olefinic partners (4a-4o) (Figure 2). Substrates with electron-donating groups such as 4-Me– (4b) and 4-OMe– (4c) substituents afforded 5b and 5c in 58 and 49% yield, respectively. Para-fluorinated (4d) and -chlorinated (4e) substrates engage in efficient hydroxyaziridination, and ortho-brominated 4f affords the corresponding hydroxyaziridine in 56% yield, which evidence the compatibility of the protocol with large ortho substituents. Electron-deficient substrates such as 4g (−CO₂Et), 4h (−CN), and 4i (−CF₃) are hydroxyaziridinated to 5g, 5h, and 5i in 56–64% yield; 3-nitrostyrene furnished the corresponding aziridine-addition product 5j in 81% isolated yield. Heterocycle-containing substrates are also compatible with the reaction conditions, with 4-vinylpyridine delivering product 5k in 40% yield. 1,1-Disubstituted styrenes were competent substrates: α-Methyl- and phenyl-substituted styrene (4l and 4m) gave the hydroxylated products (5l and 5m) with moderate yields (41% and 44%, respectively); 1,2-disubstituted derivatives were not productive coupling partners. Styrenes derived from pharmaceuticals such as indomethacin and ibuprofen also provided the desired products (5n and 5o) in 48% and 54% yield, respectively. Consistent with the high N–H BDE and proclivity of aminyl radicals to engage in H atom abstraction reactions, N–H aziridines were often observed as byproducts of the developed aziridine transfer chemistry.¹⁷ Finally, attempts to translate this chemistry to aliphatic olefins or enol ether derivatives were unsuccessful (Figure S1).

Figure 2.

Styrene scope. Conditions: 2a (0.1 mmol), 4a–4o (0.15 mmol), Et₃N (0.2 mmol), Ir(ppy)₃ (1.0 mol %), LiBr (0.1 mmol), blue LED, 23 °C in MeCN:H₂O (1:1, 2.0 mL). Aziridine scope. Conditions: 2p–2ad (0.1 mmol), 4a (0.15 mmol), Et₃N (0.2 mmol), Ir(ppy)₃ (1.0 mol %), LiBr (0.1 mmol), blue LED, 23 °C in MeCN:H₂O (1:1, 2.0 mL). ^a0.2 mL of CH₂Cl₂ was added as the olefin 4n is sparingly soluble in MeCN. Isolated yields.

Diverse N-aziridinyl radical precursors (2p-2ad) are also accommodated in the hydroxyaziridination protocol. Electron-neutral (5p-5r) and electron-deficient (5s-5v) radical precursors delivered addition products in moderate to good yields. N-Pyridinium aziridines with electron-donating substituents, such as 2w and 2x, were less efficient, delivering hydoxyaziridines 5w and 5x in 46% and 42% yield, respectively. The less efficient coupling of electron-rich precursors is consistent with a more challenging one-electron reduction of these substrates. Photoactivation of naphthylstyrene-derived N-pyridinium aziridine salt 2y in the presence of 3-nitrostyrene delivered product 5y in 69% isolated yield. Aziridine 2z, derived from 2-vinyl benzothiophene, is also compatible with the aziridine-transfer protocol, delivering 5z with a 57% yield. Moreover, N-pyridinium aziridines derived from pharmaceutical scaffolds such as indomethacin (2aa), tufnil (2ab), ibuprofen (2ac), and probenecid (2ad) all engage in efficient aziridine transfer chemistry (52–65% yields).

Photoactivation of N-pyridinium aziridines derived from aliphatic olefins is less efficient than those derived from styrenes (Figure 3). Photoactivation of cyclohexene-derived N-pyridinium aziridine 6 in the presence of styrene (4a) afforded hydroxyaziridinated product 7a in 42% yield. A small family of substituted styrene derivatives were treated with 6 under blue-light irradiation and all afforded the corresponding aziridine-transfer products (7b-7d, 33–48% isolated yield). Complex styrene derivatives, such as indomethacin derived 4p, could be engaged similarly, albeit in a low yield: Product 7e was isolated in an 18% yield. Finally, access to ethylene-derived N-pyridinium aziridine 8 provided the opportunity to evaluate the transfer of the simplest, completely unsubstituted N-aziridinyl radical (Figure 3b). 1,2-Hydroxyaziridination of styrene with an unsubstituted aziridine radical affords 2-hydroxy-2-phenyl-1-aziridinoethane (9, HPAE), which is currently being used as an experimental anticancer agent against neuroblastoma,¹⁸ in 22% yield.

Figure 3.

(a) Conditions: 6 (0.1 mmol), 4 (0.15 mmol), Et₃N (0.2 mmol), Ir(ppy)₃ (1.0 mol %), LiBr (0.1 mmol), blue LED, 23 °C in MeCN:H₂O (1:1, 2.0 mL). ^a0.2 mL of CH₂Cl₂ was added as the olefin 4p was sparingly soluble in MeCN. Isolated yields. (b) Unsubstituted aziridine was engaged into a photocatalytic reaction to get hydroxyazirinated product 9 involving the simplest aziridinyl radical.

The development of hydroxyaziridination chemistry was predicated on the reductive activation of N-pyridinium aziridines to afford transient N-aziridinyl radicals. DFT optimization of the parent N-aziridinyl radical indicates a planar geometry with the unpaired spin in a p-orbital (Figure 4a).¹⁴ Under O₂, hydroxyaziridination of 4q afforded hydroxyaziridination product 10; under N₂, hydroxyaziridination of 4q afforded cyclopropyl ring opened product 11 (Figure 4b).¹⁹ Further, hydroxyaziridination of 4a in the presence of H₂¹⁸O afforded 5a without significant ¹⁸O incorporation, indicating that O₂ is the source of the hydroxyl group in this reaction.²⁰ Finally, the addition of TEMPO completely inhibited the hydroxyaziridination of 2a. Together, these observations are consistent with the intermediacy of bona fide N-aziridinyl radical intermediates.

Figure 4.

(a) Optimized geometry and SOMO of the ethylene-derived N-aziridinyl radical. (b) Radical clock reaction was carried out between 2a and 4q. (c) Isotope labeling experiment with H₂¹⁸O. (d) EPR data of PBN-trapped N-aziridinyl radical.

To directly evaluate the intermediacy of N-aziridinyl radicals, we carried out the photolysis of a MeCN solution of 2a, Ir(ppy)₃, and Et₃N in the presence of N-tert-butyl-α-phenylnitrone (PBN). PBN is an attractive spin trap, because unlike BHT or TEMPO, radical intermediates form kinetically persistent covalent adducts with PBN that can be characterized by a combination of electron paramagnetic resonance (EPR) spectroscopy and mass spectrometry. The EPR spectrum following photolysis of a mixture of 2a and PBN under the conditions described above (i.e., Ir(ppy)₃ (1.0 mol %), Et₃N, blue LED) displayed a triplet of quartets attributed to PBN-trapped aziridnyl radical with a_N(PBN) = 14.0 G, a_H = 1.8 G, and a_{N(aziridinyl)} = 2.1 G (Figure 4d). The apparent triplet of the quartet is due to unresolved hyperfine coupling from a_H and a_{N(aziridinyl)}. Formation of PBN-trapped aziridinyl radical was further confirmed by mass spectrometry of the EPR sample: Calculated for [M + H]⁺ = 295.1805, observed [M + H]⁺ = 295.1797.

Stern–Volmer quenching studies as a function of both [2a] and [Et₃N] indicate that 2a is the primary quencher for this reaction (Supporting Information Section D.5). As the hydroxyaziridination reaction is carried out under O₂, one might envision that reductive quenching could also be accomplished by O₂. To evaluate this possibility, we collected cyclic voltammetry (CVs) of 2a. Under N₂, compound 2a displays an irreversible reductive event at −0.85 V vs Fc⁺/Fc; under an aerobic atmosphere, in addition to the reductive wave at −0.85 V, an O₂ reduction wave is observed at −1.2 V vs Fc⁺/Fc. From these data, 2a appears to be a more competent electron transfer partner than O₂, although given the high concentration of O₂ during olefin functionalization, the O₂-mediated reduction of 2a may contribute to the overall observed catalytic rates (for more details see the Supporting Information).

The available data are consistent with the mechanism illustrated in Figure 5. Single-electron transfer from the Ir(III) catalyst to 2a can generate aziridinyl radical I. Examination of the Frontier orbitals of reaction between styrene and the N-aziridinyl radical derived from 2a indicates energetic matching of the styrene HOMO (i.e., C=C π-bond) with the partially occupied N-orbital, which is consistent with electrophilic addition (Figure S2). Further, the radical polarity (ω) of N-aziridinyl radicals was calculated using the workflow recently introduced by Nagib et al.,²¹ to be 1.58 eV, which is consistent with a weakly electrophilic radical.²² Reaction of I with styrene generates benzylic radical II,²³ which is trapped with O₂ to afford hydroxyaziridine 5a.

Figure 5.

Potential photocatalytic mechanism for the N-aziridinyl radical generation and transfer.

In summary, N-aziridinyl radicals are novel reactive intermediates that enable transfer of intact aziridines to olefinic substrates. N-Aziridinyl radicals are generated under mild photochemical conditions, which enables the strained fragment to engage in intermolecular addition chemistry without an appreciable ring opening. Olefin 1,2-hydroxyaziridination was demonstrated with a variety of aziridine precursors, including the simplest unsubstituted fragment. Radical trapping experiments supported the formation and intermediacy of freely diffusing N-aziridinyl radical intermediates. Together, these results provide new disconnections for aziridines in functional organic molecules and demonstrate the accessibility of strained aminyl radical intermediates in synthesis.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.4c14169.

• Experimental methods, optimization data, spectral data and experimental spectra, diffraction data. (PDF)

The authors declare no competing financial interest.