Excited-State Copper-Catalyzed [4+1] Annulation Reaction Enables Modular Synthesis of α,β-Unsaturated-γ-Lactams

Abstract

Synthesis of α,β-unsaturated-γ-lactams continue to attract attention due to the importance of this structural motif in organic chemistry. Herein, we report the development of a visible-light-induced excited-state copper-catalyzed [4+1] annulation reaction for the preparation of a wide range of γ-H, -OH, and -OR substituted α,β-unsaturated-γ-lactams using acrylamides as the 4-atom unit and aroyl chlorides as the 1-atom unit. This modular synthetic protocol features mild reaction conditions, broad substrate scope, and high functional group tolerance. The reaction is amenable to late-stage diversification of complex molecular architectures, including derivatives of marketed drugs. The products of the reaction can serve as versatile building blocks for further derivatization. Preliminary mechanistic studies suggest an inner-sphere catalytic cycle involving photoexcitation of the Cu(BINAP) catalyst, single-electron transfer, and capture of radical intermediates by copper species followed by reductive elimination or protonation to give the desired γ-functionalized α,β-unsaturated-γ-lactams.

Graphical Abstract

INTRODUCTION

α,β-Unsaturated-γ-lactams are core structural elements in natural and synthetic organic compounds possessing a wide diversity of important biological activities.¹ For example, natural products, such as Ascosali Pyrrolidinone A, and marketed drugs, including Imrecoxib and Glimepiride, are used as antimalarial, anti-inflammatory, and anti-diabetic agents, respectively (Scheme 1A). The α,β-unsaturated-γ-lactam fragment also serves as a versatile synthetic building block and has been used in the preparation of functionalized γ-lactams² and pyrroles.³ Consequently, considerable effort has been devoted to the synthesis of α,β-unsaturated-γ-lactam scaffolds, leading to the development of many attractive stoichiometric and catalytic synthetic strategies.^4,5 For example, Wang et al. reported the Fe-catalyzed intramolecular cyclization of tertiary enamides followed by an intriguing 1,3-OH shift to afford γ-hydroxy α,β-unsaturated-γ-lactam derivatives (Scheme 1B).^5d More recently, Kleji and co-workers developed an elegant Pd-catalyzed domino synthesis of γ-hydrogen α,β-unsaturated-γ-lactams.⁵ⁱ Despite these advances, a catalytic protocol for the preparation of γ-alkoxyl α,β-unsaturated-γ-lactam derivatives is rare.⁶ As such, the development of an operationally simple, modular, and unified method to access a diverse range of γ-alkoxyl, γ-hydroxy, and γ-hydrogen α,β-unsaturated-γ-lactams is highly desirable.

Scheme 1.

Applications and synthesis of α,β-unsaturated-γ-lactams.

Visible light-induced excited-state catalysis is becoming a versatile tool in organic synthesis because it allows access to an excited-state reaction landscape for the discovery of novel chemical transformations.⁷ The use of ecologically benign and cost-effective copper complexes (Cu: US$0.512/mol)⁸ replacing precious metal Ru^II polypyridine (Ru: US$2,015/mol) or cyclo-metallated Ir^III complexes (Ir: US$30,282/mol) as photocatalysts has recently attracted increasing attention.^9,10,11 In addition to the economic benefits of their use, copper photocatalysts have dual catalytic reactivity and can serve as both photocatalysts and coupling catalysts,¹¹ participating in inner-sphere catalysis through direct engagement with radical intermediates. We have a continuing interest in advancing fundamental knowledge in excited-state and radical chemistry,¹² expanding the reaction profile of Cu catalysis,^12g and exploring synthetic applications of the Umpolung reactivity of acyl radicals.^{12b–d, 12g} Accordingly, we questioned if a copper catalyst could capture a radical intermediate generated from the addition of a nucleophilic acyl radical to acrylamide and undertake a reaction pathway distinct from the Ir-photocatalyst-mediated reaction (Scheme 1C).¹³ Herein, we report the realization of this goal leading to the first excited-state Cu-catalyzed [4+1] annulation reaction for the construction of a broad array of valuable γ-functionalized α,β-unsaturated-γ-lactams.

RESULTS AND DISCUSSION

We began our investigation with the evaluation of a broad set of conditions, including different catalysts, ligands, and solvents (Tables S1–S3, SI). Upon exposing N,2-diphenylacrylamide (1a), 4-fluorobenzoyl chloride (2a), and n-hexanol (3r) to visible light from 100 W blue light-emitting diodes (LED) in the presence of Cu₂O (10.0 mol%) and rac-BINAP (20 mol%) in a 1:1 mixture of p-xylene:DCE (0.100 M) at room temperature, we obtained the desired γ-hexyloxy α,β-unsaturated-γ-lactam 4a in 98% NMR yield (Table S4, SI).¹⁴ Control experiments showed that copper catalyst, the rac-BINAP ligand, and light are essential for the reaction, and the absence of any one of these afforded no desired product. It is worth noting that conventional Ru- and Ir-based photocatalysts failed to generate the desired product (Tables S1, SI), demonstrating the distinct reactivity of the copper photocatalyst.

With the optimized reaction conditions established, we explored the scope of the reaction by examining a variety of (hetero)aroyl chlorides, enamides, and alcohols (Table 1). The reaction tolerated a range of (hetero)aroyl chlorides with different electronic properties and functional groups (Table 1A). For example, electron-deficient aroyl chlorides substituted with halogen, nitrile, ester, trifluoromethoxyl, and trifluoromethyl groups reacted smoothly, furnishing the desired products (4a-4i) in 68–92% yield. Electron-neutral (4j) and electron-rich aroyl chlorides (4k-4l) are viable substrates. Although alkyl acyl chlorides failed, aroyl chlorides with extended π-conjugation such as a β-napthyl ring (4m) and heteroaroyl chlorides, including isoxazolyl (4n) and benzothiophenyl rings (4o), tolerate the reaction conditions and afford the corresponding γ-hexyloxy α,β-unsaturated-γ-lactams in 70–85% yield.

Table 1.

Scope of aroyl chlorides, acrylamides, and alcohols for the synthesis of γ-alkoxy, γ-hydroxy α,β-unsaturated-γ-lactams^a

^aSee Supporting Information for experimental detailss. Standard conditions A: 1 (1.00 equiv), 2 (2.00 equiv), 3 (5.00 equiv), Cu₂O (10.0 mol%), rac-BINAP (20.0 mol%), p-xylene:DCE (1:1, 0.100 M), 100 W blue LED, rt, 22 h. Cited yields are of isolated material following chromatography.

^bN,2-diphenylacrylamide (1a) was used as a coupling partner in presence of hexanol (3r).

^c4-Br-benzoyl chloride (2c) was used as a coupling partner in presence of 3r.

^d1a and 2c were used as the standard reactants.

^eBenzoyl chloride (2j) was used as a coupling partner.

The reaction proceeded well with a broad array of acrylamide derivatives with simple or complex molecular skeletons (Table 1B). Acrylamides bearing α-aromatic substituents with different electronic properties and substitution patterns reacted under the standard reaction conditions, affording the desired products 5a-5f in 71–84% yield. Investigations into medicinally relevant heterocyclic scaffolds demonstrated that carbazole, dibenzothiophene, and benzothiophene-derived acrylamides could be employed to generate the corresponding products (5g-5i) in good yields. α-Alkyl-substituted acrylamides (5j-5l) are also viable substrates under this protocol. Both N-aryl- and N-alkyl-protected acrylamides reacted smoothly, furnishing the corresponding products (5m, 5n) in 87% and 85% yield, respectively. Late-stage modifications of biologically active molecules are often a key to the identification of medicinal agents.¹⁵ To demonstrate the applicability of this excited-state copper-catalyzed [4+1] annulation reaction in late-stage diversification, derivatives of bio-relevant molecules, such as ibuprofen, adapalene, and piperazine were subjected to the standard reaction conditions, and delivered the desired products (5o-5q) in good yields. Thus, our strategy can increase structural diversity, synthetic efficiency, and is compatible with biologically-active pharmaceutical scaffolds.

We next sought to examine the generality of this transformation by exploring the scope of the alcohol coupling partner. As outlined in Table 1C, a wide variety of hydroxy compounds reacted efficiently in this excited-state copper-catalyzed cross-coupling protocol. Water can be used as a coupling partner to furnish the γ-hydroxy α,β-unsaturated γ-lactam (6a). A series of primary and secondary alcohols bearing trichloromethane, cyclopropane, adamantane, N-Boc piperidine, cyclohexane, a phenyl ring, and thiophene are compatible under the standard conditions, affording the desired products (6b-6m) with up to 94% yield. It is noteworthy that alkyne and alkene functionalities survive the reaction intact (6n-6o), providing useful handles for further synthetic elaboration. Naturally occurring alcohols, such as nerol and (−)-citronellol, can be used directly to furnish the corresponding products 6p and 6q in 51% and 81% yield, respectively, further demonstrating the synthetic utility of the transformation. The structure of the coupling products were unambiguously confirmed by single-crystal X-ray analysis of compound 6e (Table 1).

During the exploration of the substrate scope, 2-trifluoromethyl benzoyl chloride afforded the desired γ-hexyloxy α,β-unsaturated-γ-lactam in only 11% yield. Interestingly, careful examination of the reaction mixture led to the identification of γ-hydrogen α,β-unsaturated-γ-lactam 7a (Table 2). Fine-tuning the reaction conditions by modifying the catalysts, solvents, and removal of alcohol additives furnished the desired product 7a in 91% yield (Tables S5–S8, SI). The reaction proved to be general and tolerated (hetero)aroyl chlorides with diverse electronic and substitution patterns (Table 2A). Both electron-rich and electron-poor aroyl chlorides were converted into the corresponding α,β-unsaturated-γ-lactams (7a-7q) in good to excellent yields. Ortho-, meta-, and para-substituents on the ring are tolerated (7d-7f), although the reaction of the para-substituted substrate required the addition of 0.50 equiv of PhSiH₃ to give the desired product in high yield. Our protocol was also efficient toward substrates with the pharmaceutically relevant heteroarenes (7r-7t), trifluoromethyl (7a, 7j), fluoro (7b, 7j, 7k), and trifluoromethoxyl (7l, 7m) groups. Functional groups such as nitrile (7h), chloride (7c, 7g, 7k, 7r, and 7t), or bromide (7d-7f, 7i) that are frequently reactive in typical late transition-metal catalysis are also compatible. These groups are very valuable for further orthogonal manipulations and build-up of molecular complexity.

Table 2.

Scope of aroyl chlorides and acrylamides for the synthesis of γ-hydrogen α,β-unsaturated-γ-lactams^a

^aSee Supporting Information for experimental details. Standard conditions B: 1 (1.00 equiv), 2 (2.00 equiv), CuI (10.0 mol%), rac-BINAP (20.0 mol%), p-xylene (0.100 M), 100 W blue LED, rt, 22 h. Cited yields are of isolated material following chromatography.

^bN,2-diphenylacrylamide (1a) was used as a coupling partner.

^c0.50 equiv of PhSiH₃ was added, 36h.

^dThe reaction time was extended to 36 h.

^e2-Br-benzoyl chloride (2d) was used as a coupling partner.

The scope with respect to the acrylamides was also investigated (Table 2B). In general, acrylamides with (i) α-aryl (8a-8f, 8i, 8j), α-heteroaryl (8g-8h, 8m), and α-alkyl (8k, 8l) substituents, (ii) aryl (8a-8i, 8k-8m) or alkyl (8j) N-protecting groups, and (iii) derivatives of bioactive molecules (8m) were competent under the optimized reaction conditions, affording the desired products with up to 91% yield. It is of note that 80 of the 82 compounds reported here have not been prepared previously, demonstrating that our protocol allows access to unexplored chemical space, thus aiding the discovery of new pharmaceuticals, agrochemicals, and functional materials.

To demonstrate the synthetic utility of the α,β-unsaturated-γ-lactam products toward useful molecular scaffolds, a series of transformations were carried out (Scheme 2). Treatment of the α,β-unsaturated-γ-lactam product 7g with Pd/C in the presence of H₂ (1 atm) gave the corresponding γ-lactam 9a in 68% yield with >20:1 diastereoselectivity. Epoxidation of 7g using hydrogen peroxide at 60 °C formed a compound 9b with an oxa-azabicyclohexan-2-one ring in 61% yield. Product 7g could be converted to a TBS-protected pyrrole 9c in 89% yield using tert-butyldimethylsilyl trifluoromethanesulfonate (TBSOTf) at room temperature.

Scheme 2.

Post functionalization of α,β-unsaturated-γ-lactam 7g.

A series of mechanistic studies were conducted to gain a better understanding of the reaction mechanism. The addition of a radical scavenger, (2,2,6,6-tetramethylpiperidin-1-yl)oxyl (TEMPO), to the standard three-component reaction was found to inhibit the formation of the desired product 4a (Scheme 3A). Although we could not isolate the TEMPO-trapped adduct, we were able to obtain adduct 10a in 54% yield when 1,1-diphenylethylene was used as a radical trap (Scheme 3A). These experiments imply that the reaction is likely a radical process involving an aroyl radical intermediate. To determine whether the γ-alkoxylated product is formed through an N-acyliminium ion intermediate generated from the corresponding γ-hydroxy adduct such as 6a,^6e we carried out a standard reaction using 1b and 2c as substrates with 6a as an additive (Scheme 3B). This reaction afforded product 5a only and none of the product 4c derived from 6a, indicating that γ-hydroxy adduct is unlikely to be an intermediate leading to the γ-alkoxylated product. To identify potential reaction intermediates, we also conducted a time-dependent mass spectrometry analysis of the three-component reaction using 1a and 2c as starting materials (Scheme 3C). Analysis of isotope patterns and collision-induced fragmentation of the reaction mixture implied the formation of copper complex Int-I (Scheme 3C, Fig. S5, SI). We envisioned that the copper(III) catalyst in Int-I migrates to the γ-position and subsequent reductive elimination liberates the desired product 4c, regenerating the Cu(I)BINAP catalyst.^{11x, 11y, 16} However, when an ortho-substituted aroyl chloride is used as the substrate, this reaction pathway is disfavored by a steric interaction between the phosphine ligand and the γ-aryl ring of Int-I. In this case, Int-I is more likely to be protonated and then tautomerized, forming γ-hydrogenated products such as 7a and 7d (Scheme 3D).

Scheme 3. Mechanistic studies.

See Supporting Information for experimental details.

The formation of the γ-hydrogenated product is very intriguing, and we performed a series of deuterium labeling experiments to investigate the origin of the γ-hydrogen atom in the product (Scheme 3E). Deuterated compound d₂-1a was used as a starting material, and in this case, we obtained the desired product d₂-7a with 25% deuterium incorporation at the γ-position. Also, a cross-over experiment using d₂-1a and 1k as starting materials afforded the products d₂-7a and d-8k with 28% and 23% γ-deuterium incorporation, respectively. GCMS analysis of the reaction mixture revealed the formation of chlorinated p-xylene (Fig. S6, SI).¹⁷ Using d₈-toluene as the solvent, we obtained a product d-7a with 13% γ-deuteration. Collectively, these results suggest that the γ-hydrogen atom originates from both acrylamide and the solvent.

Light On/Off experiments showed that the reaction stops in the absence of light (Scheme 3F). The quantum yield of the standard reaction forming the γ-hydrogenated product is 0.083 (Fig. S8, SI), implying that an extended radical chain process is unlikely. The UV-Vis absorption spectra of the CuI-BINAP complex shows a peak at 397 nm with a tail into the visible region, which resembles the UV-Vis absorption spectra of the standard reaction mixture (Scheme 3G). This indicates the involvement of the CuI-BINAP photoactive complex in the reaction medium. A preformed CuI-BINAP complex was catalytically competent and afforded the desired product in 91% yield (Fig. S10, SI). Finally, 4-fluorobenzoyl chloride (2a) was shown to quench the luminescence of the CuI-BINAP complex. (Fig. S11, SI).

Based on these mechanistic insights, a plausible reaction mechanism was proposed and is depicted in Scheme 4. Blue-light photoexcitation of Cu^I-BINAP complex I generates excited *[Cu^I-BINAP] complex II, which reduces the aroyl chloride 2 (E_p of benzoyl chloride = –1.02 V vs. SCE)^13c to form copper complex III, liberating aroyl radical 2’. The addition of 2’ to the alkene of acrylamide 1 affords tertiary alkyl radical IV, which is captured by III to form intermediate V. Intramolecular condensation of V furnishes intermediate VI. In the presence of alcohol, VI undergoes ligand exchange and migration of the copper center to the γ-position followed by reductive elimination, liberating the desired γ-alkoxylated product 4 and regenerating copper catalyst I. In the absence of alcohols or if an ortho-substituted aroyl chloride is used, complex VI will proceed through protonation and tautomerization to form γ-hydrogenated product 7 and copper species VIII. Reduction of VIII by p-xylene affords copper catalyst I,¹⁷ closing the catalytic cycle.

Scheme 4.

Proposed reaction mechanisms.

SUMMARY

In summary, we report the development of an excited-state copper-catalyzed [4+1] annulation reaction. This provides a unified and modular strategy for the preparation of a wide array of γ-hydrogen, -hydroxy, and -alkoxy α,β-unsaturated γ-lactams. The reaction features mild reaction conditions, has broad substrate scope, and tolerates complex molecular architecture, including derivatives of marketed drugs. The α,β-unsaturated-γ-lactam products can serve as versatile building blocks for the synthesis of useful heterocycles. Preliminary mechanistic studies suggest an inner-sphere catalytic cycle that involves photoexcitation of the Cu(BINAP) catalyst, single-electron transfer, and trapping of radical intermediates by copper species followed by reductive elimination or protonation to give the desired coupling products. Further studies of the reaction mechanism and expansion of the substrate scope as well as the asymmetric process are currently underway.