Regiospecific de mayo reactions of enamino ketones

Abstract

The de Mayo reaction is a powerful photochemical method for skeletal editing, enabling rapid synthesis of 1,5-dicarbonyl compounds through formal addition of 1,3-dicarbonyl units across olefins. However, its regio- and chemoselectivity remain limited. Distinct aryl groups cannot be selectively installed on 1,5-diketones; substituents in 1-aryl-5-alkyl diketones favor the α-position of the aryl ketone; and electron-withdrawing groups from deficient olefins are generally restricted to the 3-position. Such olefins also favor radical addition and formal redox pathways over the desired reaction. Here, we introduce de Mayo reactions of readily accessible enaminoketones that overcome these limitations and directly furnish valuable cyclohexenone derivatives. Tuning the N-protecting group and using a chiral Brønsted acid catalyst enables proof-of-concept kinetic resolution of the imine intermediate formed after [2 + 2] cycloaddition. This regiospecific platform expands precise access to 1,5-diketones and related motifs relevant to pharmaceutical discovery, while broadening the synthetic scope of this established photochemical transformation under practical conditions.

The de Mayo reaction is a valuable photochemical method for skeletal editing, but longstanding regio- and chemoselectivity limitations restrict its synthetic scope. Here the authors report enaminoketone-enabled de Mayo reactions that overcome these constraints, providing regiospecific access to cyclohexenones and 1,5-diketone motifs.

Introduction

The rapid and efficient discovery of drugs and materials serves as a central driving force in advancing the development of novel and robust synthetic methodologies. In this regard, photochemical skeletal editing has emerged as a transformative approach by enabling facile late-stage diversification of high-value molecules, thereby significantly reducing the time, cost, and resources typically associated with de novo synthesis^1–3. Notably, the de Mayo reaction—first reported by Paul J. de Mayo⁴ in 1962—has recently regained attention because it uniquely harnesses readily available 1,3-dicarbonyl compounds to edit olefinic double bonds into saturated, carbon-chain 1,5-dicarbonyl frameworks, thereby providing direct access to non-natural bioactive molecules and natural products (Fig. 1A)^1–26. Mechanistic investigations have demonstrated that enol formation is critical for the success of this transformation, as the enol tautomer can be promoted to its triplet excited state and participate in productive reactions with olefins. The resulting 1,4-biradical intermediates also play a key role^27,28; inefficient radical–radical coupling—dictated by the electronic and steric characteristics of the radical centers—may lead to fragmentation back to the ground-state starting materials rather than productive cyclization. Furthermore, direct photoexcitation of the keto tautomer can trigger n–π* transitions, promoting undesired side reactions such as oxidation and hydrogen atom transfer (HAT)²¹. To overcome these limitations, several strategies—including acid catalysis and energy-transfer (EnT)-based photocatalysis—have been recently developed to improve both reactivity and chemoselectivity^17–23.

Fig. 1. Outline of This Work.

A Mechanistic profile of De Mayo reactions; B Challenges in regioselectivity for 1,3-diketones; C De Mayo reactions of enolate imines; D Design plan: enamino ketones via energy transfer catalysis; E)This work.

Nevertheless, despite the value of 1,5-diketones as key intermediates en route to cyclohexenone derivatives, achieving high levels of regioselective control with unsymmetrical 1,3-diaryl ketones remains a long-standing, unresolved challenge (Fig. 1B). For such substrates, the comparable reactivity of the two carbonyl groups gives rise to competing enol isomers and, upon addition to olefins, distinct 1,4-biradical intermediates, culminating in poor regioselectivity. To date, no de Mayo reactions have been reported that deliver high regioselective fidelity under these conditions. For 1-aryl-3-alkyl diketones, enolization preferentially occurs at the aryl ketone, such that olefin-derived substituents are typically incorporated at the α-position of the aryl ketone in the final products^{17,20,23–26}. A further—and equally consequential—limitation is encountered with electron-deficient olefins, which generally fail to furnish the desired de Mayo products. Because α-ketone radicals are strongly electrophilic, their addition to electron-poor olefins is intrinsically disfavored. Even when addition occurs, the resulting 1,4-biradicals often undergo formal intramolecular redox events^29,30 to give simple addition adducts rather than productive de Mayo reaction. Consistent with this reactivity, de Mayo reported only a single example with methyl acrylate in 1969, and poor chemoselectivity was observed⁶. Of note, replacing one ketone with an ester reverses—and improves—regioselectivity, placing the electron-deficient substituent at the 3-position of the resulting 1,5-dicarbonyl products²¹. This outcome has been ascribed to the weaker electron-withdrawing character of esters relative to ketones, which enables an alternative pathway involving ketyl radical addition to the olefin followed by radical–radical coupling³¹. By contrast, when both carbonyl groups are ketones, coupling of the 1,4-biradicals is disfavored, leading to diminished reactivity and eroded regioselective control.

Notably, the construction of nitrogen-containing cyclic frameworks has been advanced through photoexcited intramolecular de Mayo reactions of enamino ketones with olefins^13–16. By contrast, intermolecular variants have remained elusive, with no detectable product formation under the reported conditions³². Although the origin of this limitation remains unclear, we speculate that this may be attributed to the higher nucleophilicity of α-amino radicals compared with ketyl radicals, which diminishes the propensity of the resulting 1,4-biradical manifold to undergo productive bond formation. To address this challenge, Itoh and co-workers introduced difluoro(aminoenolato)boron complexes as surrogates for 1,3-diketones (Fig. 1C)^33,34. However, these reactions still exhibit poor regioselectivity with olefins, and substrate isomerization further erodes chemoselectivity. In addition, the imine unit can itself engage in cycloaddition with olefins to form azacyclobutanes, leading to complex reaction outcomes. Motivated by these limitations, we revisited enamino ketones (Fig. 1D). Relative to their enol counterparts, enamino ketones bearing distinct 1- and 3-substituents are more readily accessible^35–37—an essential prerequisite for achieving high regiocontrol in de Mayo-type transformations. We further reasoned that the considerably higher stability of enamino ketones compared to the corresponding enol imines would suppress isomerization, thereby further ensuring regioselectivity³⁸. Moreover, the greater nucleophilicity of α-amino radicals suggests that enamino ketones may engage not only electron-neutral and electron-rich olefins but also electron-deficient partners while preserving regiochemical fidelity. Finally, and importantly, the modularity of the N-protecting group provides an attractive handle for the rational tuning of both reactivity and regioselectivity.

Herein, we report the successful implementation of this exploratory strategy (Fig. 1E). In the presence of 4CzIPN and under irradiation with a 3 W blue LED, a broad range of aryl-substituted enamino aryl ketones—bearing diverse substituents and substitution patterns on both aryl rings—react with a variety of olefins with high chemo- and regioselectivity, delivering the corresponding 1,5-dicarbonyl products in satisfactory yields. Furthermore, alkyl-substituted enamino aryl ketones react efficiently in the absence of a photocatalyst, where direct photoexcitation enables access to de Mayo products bearing substituents selectively at the α-position of the alkyl ketone in high yields. Notably, both classes of synthons are compatible with electron-deficient olefins, affording de Mayo products while maintaining the regiochemical outcomes observed with other olefin partners. Moreover, a modest adjustment of the reaction conditions provides direct access to cyclohexenone derivatives, and replacing racemic acids with chiral phosphoric acids (CPAs) enables proof-of-concept kinetic resolution to furnish enantioenriched products, underscoring the synthetic utility and potential pharmaceutical relevance of this method.

Results

We initiated our studies with prop-1-en-2-ylbenzene (1) and the simple (Z) − 3-amino-3-phenyl-1-(p-tolyl)prop-2-en-1-one (2), bearing a primary amine, as model substrates (Table 1A). In the presence of 0.5 mol% 4CzIPN^39,40 and under irradiation with a 3 W blue LED (λ^em_max = 445 nm), the reaction afforded a complex mixture, and no major product was identified. We then evaluated the corresponding secondary amine (3) bearing an N-phenyl substituent; no conversion was observed. We next examined secondary amine derivatives in which the N-protecting group was varied to modulate reactivity. Gratifyingly, the N-benzyl substrate 4 generated an imine intermediate (A), which could be hydrolyzed in one pot with HCl to furnish the desired product 13 in 56% yield. The appreciable yield and the excellent regioselectivity observed prompted us to survey a series of enamino ketones (5–12) bearing diverse N-alkyl substituents. Under the optimized conditions, the N-methyl analogue 5 delivered 13 in 78% yield, and the 2-(pyridin-4-yl)ethyl substituent (10) further improved the yield to 81%, representing the best outcome. By contrast, the N-tert-butyl substrate 7 reacted sluggishly, presumably owing to increased steric hindrance. Moreover, substrates bearing electron-withdrawing N-substituents (for example, 11 and 12) showed little to no reactivity under these conditions. While the precise origins of how different N-substituents—such as aryl, alkyl, and electron-withdrawing groups—affect reactivity remain under investigation, our experimental and computational studies provide valuable insight. Specifically, both measured and calculated triplet energies of representative enamino ketones indicate efficient EnT from the triplet 4CzIPN sensitizer to these substrates. Given the structural similarity among the α-ketone radicals generated from these substrates, their initial addition to olefin 1 is expected to proceed with comparable facility. We therefore propose that the conjugative effects of aryl groups—and the electron-withdrawing character of carbonyl and pyridyl moieties—may subtly influence the efficiency of the key radical coupling step that forms the cyclobutane intermediates. Accordingly, these findings underscore the pivotal role of the N-substituent in enabling successful transformation—offering both strong support for our design rationale and a meaningful advance in its realization.

Table 1.

Reaction optimization^a

Entry	Alteration to conditions	Yield (%)a	Entry	Alteration to conditions	Yield (%)a
1	IrIII complex (ET = 60.8 kcal/mol, E1/2 (PC*/PC −) = + 1.21 V, E1/2 (PC*/PC −) = − 0.89 V) as PC	41	6	1.0 mol% 4CzIPN	81
2	fac-Ir(ppy)3 (ET = 55 kcal/mol, E1/2 (PC*/PC −) = + 0.31 V, E1/2 (PC*/PC −) = − 1.73 V) as PC	61	7	without 4CzIPN	no reaction
3	DPZ (ET = 47.5 kcal/mol, E1/2 (PC*/PC −) = + 1.42 V, E1/2 (PC*/PC −) = − 1.07 V) as PC	no reaction	8	without 4CzIPNb	5.0
4	Ru(bpy)2Cl2 (ET = 46.5 kcal/mol, E1/2 (PC*/PC −) = + 0.77 V, E1/2 (PC*/PC −) = − 0.81 V) as PC	no reaction	9	under air	no reaction
5	Mes-Acr+ (ET = 44.5 kcal/mol, E1/2 (PC*/PC −) = + 2.18 V, E1/2 (PC*/PC −) = − 0.49 V) as PC	no reaction	10	no light	no reaction

^aIsolated yield,^b 3 W LED (λ_max = 395 nm instead of 420 nm. Ir^III complex = Ir[dF(CF₃)ppy]₂(dtbbpy)PF₆; dtbbpy = 4,4-di-tert-butyl-2,2’-bipyridine; bpy = 2,2’-bipyridine.

Beyond the modestly higher yield, we selected 2-(pyridin-4-yl)ethyl as the N-substituent for the enamino ketone (10) because the corresponding imine intermediate is sufficiently robust to withstand purification by silica-gel column chromatography. By contrast, when an N-methyl substituent (5) was used, partial hydrolysis was evident even after the photochemical step, and small amounts of the hydrolysis product 13 were detected directly in the crude reaction mixture. Subsequent chromatographic purification would be expected to exacerbate this loss, which is synthetically undesirable for the downstream kinetic resolution to furnish enantioenriched products (vide infra). Moreover, the N-methyl variant delivered markedly lower yields than the 2-(pyridin-4-yl)ethyl analogue across a range of substrates [see the Supplementary Information (SI) for selected examples]. Accordingly, we conducted a series of control experiments for the reaction between enamino ketone 10 and olefin 1 to delineate the key parameters (Table 1B). We first examined alternative photosensitizers in place of 4CzIPN (entries 1–5, Table 1B). Both Ir[dF(CF₃)ppy]₂(dtbbpy)PF₆ and fac-Ir(ppy)₃ afforded product 13 in diminished yields (entries 1 and 2, Table 1B). By contrast, no reaction was observed with DPZ, Ru(bpy)₃Cl, or Mes-Acr⁺—photosensitizers with triplet energies lower than that of fac-Ir(ppy)₃ (entries 3–5, Table 1B). Increasing the loading of 4CzIPN to 1.0 mol% did not improve the yield (entry 6, Table 1B), whereas omission of 4CzIPN completely suppressed reactivity (entry 7, Table 1B). When higher-energy irradiation was employed, the reaction proceeded with poor conversion and compromised chemoselectivity, delivering 13 in only 5% yield (entry 8, Table 1B). Finally, the reaction was found to require both an oxygen-free atmosphere and light irradiation; omission of either condition shut down the transformation (entries 9 and 10, Table 1B).

With the optimal conditions in hand, we next evaluated the substrate scope to delineate the generality and synthetic potential of this new de Mayo reaction platform. We first examined a series of 1,3-diaryl enamino ketones with olefin 1 (Fig. 2). To rigorously probe the regioselectivity, we initially tested pairs of enamino ketones in which the two aryl substituents were interchanged between the 1- and 3-positions. Under the standard conditions, products 13–48 were obtained in satisfactory yields. Across this set, the two substituents derived from olefin 1 were consistently incorporated at the α-position of the aryl ketone originating from the enamine moiety, irrespective of whether electron-withdrawing or electron-donating substituents were present on either aryl ring and regardless of substitution pattern. Likely owing to steric congestion, an ortho-methoxy substituent on the aryl ring of the enamine fragment suppressed product formation. Importantly, pharmaceutically relevant azaarenes^39–44, including pyridine and quinoline, were readily accommodated on either ketone moiety of the resulting 1,5-diketones (45–50). Notably, in previously reported de Mayo approaches to these targets, only a single regioisomer was accessible, in which the olefin-derived substituents were installed on the azaarene-substituted ketone (for example, 45, 47, and 49)²⁵. We further evaluated enamino ketones bearing fragments of bioactive molecules, including dicamba, adapalene, estrone and clofibrate, on one aryl ring. These substrates performed well, furnishing the corresponding products 53–56 in 62–95% yield.

Fig. 2. de Mayo Reactions of 1,3-Diaryl Enamino Ketones with Olefin 1.

Reaction conditions: 10 (0.1 mmol), 1 (3.0 equiv.), 4CzIPN (10 mol%), toluene (2 mL), 1 N HCl (2 mL), toluene (2 mL), 24 h; 1 N HCl (2 mL), 2 h. Isolated yield.

We next examined enamino ketones with a range of α-methyl styrenes bearing either electron-withdrawing (57–60) or electron-donating substituents (61–66) on the aryl ring (Fig. 3). The corresponding products 57–66 were obtained in 67–95% yield, with uniformly high and site-selective regioselectivity. Comparable performance was observed for olefins in which the phenyl group was replaced by polycyclic aromatics (67 and 68) or heteroaromatics (69–72). Encouraged by these results, we evaluated the capacity of this method to merge multiple bioactive scaffolds within a single product. As a representative example, product 73—incorporating key structural motifs from celestolide, probenecid, and bezafibrate—was obtained in 54% yield, underscoring the potential of this platform for medicinal chemistry, given its broad scope, functional-group tolerance, and predictable regiochemical outcome. We then replaced the α-methyl group with an acetoxy substituent (74 and 75) or a p-methylbenzamido group (76 and 77), representative electron-rich olefins, and obtained products 74–77 in 60–90% yield. To probe olefins with modestly electron-deficient character while enabling installation of a valuable trifluoromethyl group, we next examined (3,3,3-trifluoroprop-1-en-2-yl)benzene, which furnished products 78 and 79 in moderate yields. Finally, the efficient formation of 80 and 81 demonstrated that this method can be used to construct tertiary carbon stereocentres in high yield.

Fig. 3. de Mayo Reactions of 1,3-Diaryl Enamino Ketones with Various α-Substituted Styrenes.

Reaction conditions: 10 (0.1 mmol), Olefin (3.0 equiv.), 4CzIPN (10 mol%), toluene (2 mL), 24 h; 1 N HCl (2 mL), 2 h. Isolated yield. ^a2.0 mol% 4CzIPN was used. ^b5.0 equivalents of olefin was used. ^c48 hours.

Reaching products 78 and 79 provided an important indication that this method is likely to be applicable to electron-deficient olefins (Fig. 4A). We therefore evaluated 2-(prop-1-en-2-yl)pyridine, in which the pyridyl group is electron-withdrawing, albeit less so than carbonyl substituents. Gratifyingly, products 82 and 83 were obtained in high yields and with the same regioselectivity observed for electron-neutral olefins (Figs. 2 and 3). We next examined additional azaarene-substituted olefins, affording the corresponding products 84–91 in 46–77% yield. We then further challenged the method with a range of carbonyl-substituted olefins. Regardless of whether esters (92, 93), amides (94, 95), or ketones (96, 97) were employed, the desired products were obtained in satisfactory yields. Notably, the moderate or poor yields observed across different carbonyl-containing olefins arise from distinct underlying issues. For instance, esters are prone to minor addition by-products; in the case of product 92, the corresponding addition product was obtained in 24% yield. For amides (e.g., 94), the reaction proceeds sluggishly, yet no side products were detected. When a ketone substituent is present (e.g., 96), several unidentified side products were observed. These results collectively underscore the intrinsic challenge of engaging electron-deficient olefins in the de Mayo reaction.

Fig. 4. de Mayo Reactions of 1,3-Diaryl Enamino Ketones with Electron-Deficient, Cyclic, or Complex Olefins.

A Electron-deficient olefins scope; B exocyclic alkenes scope. Reaction conditions: 10 (0.1 mmol), Olefin (3.0 equiv.), 4CzIPN (10 mol%), toluene (2 mL), 24 h, 25°C; 1 N HCl (2 mL), 2 h. Isolated yield.

Subsequently, we further expanded the olefin scope by examining substrates bearing exocyclic alkenes (Fig. 4B). As a representative example, 1-methylene-1,2,3,4-tetrahydronaphthalene delivered products 98 and 99 in 50% and 87% yield, respectively. We then examined 2,3,4,5-tetrahydro-1,1’-biphenyl, which contains a cyclic, non-terminal olefin. This class of substrates is expected to be particularly challenging in terms of reactivity, as non-terminal alkenes can impede the initial radical addition step. Pleasingly, products 100 and 101 were obtained in moderate yields. We next evaluated complex substrates bearing multiple unsaturated functionalities. Excellent chemoselectivity was observed: for diolefin substrates (102, 103), only one alkene unit participated in the de Mayo reaction with enamino ketones, whereas alkynes (104, 105) remained unreacted under the standard conditions. It is worth noting that we also tested 1,2-olefins in this protocol (see SI for details). Using (E)-prop-1-en-1-ylbenzene with acetonitrile as the solvent led to a reaction; however, the product was not the desired de Mayo adduct but an attractive fully substituted pyrrole derivative, obtained in 43% yield. Nevertheless, this result still demonstrates excellent regioselectivity, as the transformation was initiated by regiospecific addition to the methyl side of the olefin. When (E) − 1,2-diphenylethene was attempted under otherwise identical conditions, no reaction was observed. These results indicate that internal olefins are less reactive than their terminal counterparts.

As noted above, 1-aryl-3-alkyl diketones preferentially afford 1,5-diketones in which the olefin-derived substituent is installed at the α-position of the aryl ketone; accessing the complementary regioisomers bearing this substituent at the α-position of the alkyl ketone therefore remained an important objective. To address this limitation, we sought to employ alkyl-based enamino ketones. However, using 2-(pyridin-4-yl)ethyl as the enamine N-protecting group in the reaction with hex-1-ene proved ineffective: no conversion was observed under either the established photocatalytic conditions or direct photoexcitation (see SI for details). We next evaluated other N-substituents, including methyl, benzyls, phenyl, OAc, and several electron-withdrawing groups (e.g., 4-pyridyl, tert-butyloxycarbonyl, trifluoroacetyl, and benzoyl) (Table S3 in the SI). Among these, only benzoyl delivered the desired product 106, albeit in 26% yield. We reasoned that the reactivity is highly sensitive to the stability and electrophilicity of the α-amino radical within the resulting 1,4-biradical intermediates following the initial addition step. In particular, N-electron-withdrawing substituents render such radical species more stable and more electrophilic compared with alkyl substituents. Guided by this rationale, we explored substituents with stronger electron-withdrawing ability and found isonicotinoyl to be optimal. Specifically, under irradiation with a 3 W LED (λem max = 395 nm), a stable cyclobutane intermediate (B) was detected; this intermediate could be readily hydrolyzed with HCl (70 °C, 1 h) to furnish the desired product 106 in 60% yield (Fig. 5). We next examined the reactions of this enamino ketone with a range of alkyl-substituted olefins, including substrates bearing linear (107, 109–111) and cyclic (108) substituents, delivering products 107–111 in moderate yields. Notably, ether (109) and silyl (110) groups were well tolerated. In addition, prop-2-en-1-ol enabled installation of a terminal alkene at the α-position of a methyl ketone, presumably via acid-promoted reverse Mannich ring opening followed by dehydration in the presence of HCl. The aryl group could also be replaced by azaarenes (for example, pyridine; 112), and styrene was likewise a competent coupling partner. We also explored the synthesis of 1,5-dialkyl diketone derivatives; the strategy showed good functional-group tolerance, affording products 114 and 115 in satisfactory yields.

Fig. 5. de Mayo Reactions of Alkyl-Based Enamino Ketones with Various Olefins.

Reaction conditions: Enamino Ketone (0.1 mmol), Olefin (3.0 equiv.), CH₂Cl₂ (5 mL), 24 h; 1 N HCl in 1,4-dioxane (2 mL), 2 h. Isolated yield.

Alkenyl ethers and alkenyl sulfides were next evaluated as electron-rich olefins, furnishing products 116–118 in 52–70% yield. Importantly, our studies further established that electron-deficient olefins are also viable partners: the corresponding 1,5-diketones 119 and 120, bearing amide substituents at the α-position of the methyl ketone, were obtained successfully. We then examined a series of 1,1-dialkyl-substituted terminal olefins, delivering products 121–126 in 32–60% yield. Notably, beyond assembling two distinct alkyl groups to form an all-carbon quaternary centre at the α-position of the methyl ketone, a range of structurally appealing saturated carbocycles—including cyclobutane (124), cyclododecane (125) and (–)-β-pinene (126)—could be incorporated. Encouraged by these results, we evaluated cyclic olefins, affording products 127–129 in 41–50% yield with excellent diastereoselectivity. Internal linear olefins were also competent substrates, providing products 130 and 131 in moderate yields. Beyond their lower reactivity relative to terminal olefins, the non-terminal alkenes showed poor chemoselectivity, with several unidentified side products being observed. Given this broad olefin scope, we finally probed complex olefins derived from bioactive molecules, including (S)-citronellol, pregnenolone, tigogenin, and dihydrocholesterol. The corresponding products 132–135, each incorporating a newly installed 1,5-diketone motif, were obtained in 41–48% yield, underscoring the synthetic utility and potential pharmaceutical relevance of this photochemical skeletal-editing protocol.

One important application of the de Mayo reaction is the synthesis of cyclohexenone derivatives from the adducts formed between 1-aryl-3-alkyl diketones and olefins. We therefore sought to integrate this photoexcitation platform with a subsequent, one-pot aldol condensation. As shown in Fig. 6A, a series of olefins were reacted with a methyl-based enamino phenyl ketone under the standard conditions. After 24 hours, the crude reaction mixture was directly subjected to hydrolysis and condensation by the addition of concentrated HCl and an extra portion of solvent (CHCl₃/MeOH = 1:1) at 60 °C. After 2 h, the corresponding cyclohexanone products 136–140 were obtained, generally in moderate yields. Notably, this sequence enabled rapid assembly of an attractive cyclohexanone framework from a lanosterol-derived olefin; the resulting ketone and olefin functionalities provide versatile handles for downstream derivatization. Moreover, the formation of 140 from benzothiophene represents, to our knowledge, the first example of a dearomative de Mayo reaction.

Fig. 6. Direct Synthesis of Cyclohexenone Derivatives and Kinetic Resolution of Imine Intermediates.

Reaction conditions: A Enamino Ketone (0.1 mmol), Olefin (3.0 equiv.), CH₂Cl₂ (5 mL), 24 h; 12 N HCl in CHCl₃: MeOH = 1:1 (5 mL), 2 h. Isolated yield. B A2 (0.05 mmol), C1 (10 mol%), NaH₂PO₄ (1 equiv.), p-xylene: H₂O = 4:1 (5 mL), 8 h. Isolated yield.

In recent years, substantial efforts have been devoted to asymmetric de Mayo reactions^24–26. Nevertheless, assembling all-carbon quaternary stereocentres from simple 1,3-diaryl ketones remains highly challenging. For unsymmetrical 1,3-diaryl ketones, poor regioselectivity has proved a major barrier to enantioselective synthesis. In our system, we observed that imines A (Table 1A) are formed as isolable intermediates. Notably, employing 2-(pyridin-4-yl)ethyl as the enamine N-protecting group enabled these intermediates to be purified by column chromatography, likely because the basic 4-pyridyl unit suppresses silica gel–promoted hydrolysis. Given that HCl hydrolysis provides the corresponding 1,5-diketones, we envisioned a kinetic resolution using chiral Brønsted acids to access both enantiomers. Using A2 as a model substrate, we found that in the presence of 10 mol% CPA C1, hydrolysis delivered ent-13 in 46% yield with 80% ee. In parallel, the recovered imine A2 (that is, ent-A2) was obtained in 42% yield with 80% ee, and could be further hydrolyzed with HCl to afford ent-13′. We then applied this strategy to A3, in which the two aryl substituents are exchanged relative to A2, enabling isolation of ent-30 and ent-30′ in satisfactory yields and enantioselectivities.

The direct photoexcitation reactions of alkyl-based enamino ketones, together with structural identification of intermediate B, indicate that these transformations proceed through a canonical de Mayo pathway. We therefore carried out a series of experiments to probe the mechanism for aryl-based enamino ketones under 4CzIPN photocatalysis, for which two activation manifolds—single-electron transfer (SET) or EnT—are plausible (for DFT study, see Source Data in SI). Comparison of the triplet energies and redox properties of 4CzIPN (E_T = 57.1 kcal/mol, E_1/2 (PC*/PC^•‒) = + 1.43 V vs. SCE) and model substrate 10 (E_T = 58.0 kcal/mol, E_p/2 = + 0.73 V vs. SCE) suggests that either pathway could, in principle, operate (Fig. S1–S11). However, several lines of evidence support EnT as the dominant mechanism. First, addition of pyrene (3.0 equiv.; E_T = 47.0 kcal/mol) as an energy quencher completely suppressed product formation. Second, as shown in Table 1B, photosensitizers with insufficient triplet energy failed to promote the reaction, even though their excited-state oxidizing power should be adequate to engage 10 via SET (entries 3–5, Table 1B). Stern–Volmer quenching studies further revealed that the excited state of 4CzIPN is quenched by 10, but not by olefin 1 (Fig. S12). Finally, transient absorption measurements on a mixture of 1a and 4CzIPN provided additional support for an EnT process (Fig. S13). Collectively, these data are consistent with an EnT-initiated pathway in which triplet 4CzIPN sensitizes the enamino ketone to generate its triplet excited state. The lower energy of the enamino ketone relative to its imine–enol isomer is expected to favor the selective formation of the triplet enamino ketone species. Subsequent radical addition to the olefin, followed by intersystem crossing and key radical–radical coupling, furnishes the cyclobutane intermediates A, which, upon hydrolysis, deliver the corresponding 1,5-diketone products.

Discussion

In summary, regiospecific de Mayo reactions have been developed by employing readily accessible enaminoketones as surrogates for 1,3-diketones. This approach enables the precise incorporation of a wide range of olefins into the skeleton of 1,5-diaryl-substituted diketone products with excellent and strict regioselectivity. Notably, the regioselectivity is independent of the electronic properties or substitution patterns of the two aryl substituents, being governed solely by the position of the enamine moiety within the substrate. Moreover, 1-aryl-5-alkyl diketone products exhibit regioselectivity distinct from that of classical de Mayo reactions, providing a valuable route to these important scaffolds. Furthermore, by tuning the reaction conditions, cyclohexenone derivatives can be accessed directly. Significantly, replacing enol ketones with enaminoketones allows facile modulation of the N-protecting group on the enamine unit, thereby enabling the use of electron-deficient olefins and therefore substantially expanding the substrate scope of the de Mayo reaction. Remarkably, the regioselectivity with electron-poor olefins remains identical to that observed with electron-neutral and electron-rich olefins—an outcome that is not only unprecedented but also underscores the exceptional regiospecificity of this strategy. The synthetic utility of this method is further extended by the kinetic resolution of imine intermediates via chiral Brønsted acid catalysis. We anticipate that this regiospecific protocol will find broad application in the synthesis of diverse 1,5-diketones and their derivatives, including enantioenriched compounds, thereby emerging as a valuable tool for drug discovery.

Methods

Transformations of 1,3-diaryl enamino ketones with olefins

For 13-105: 1,3-Diaryl enamino Ketone (0.1 mmol, 1.0 equiv.), olefin (0.3 − 0.5 mmol, 3.0 − 5.0 equiv.), 4CzIPN (0.0005 – 0.002 mmol, 0.5 − 2 mol%) and toluene (2.0 mL) were sequentially added into a 25 mL Schlenk tube, degassed three times by freeze-pump-thaw method. The reaction mixture was stirred under an argon atmosphere at 25 °C, then irradiated by a 3 W blue LED (λ = 450 − 455 nm) from a 1.0 cm distance for 24 − 48 h. After completion of the reaction, 1.0 N hydrochloric acid (HCl) (2.0 mL) was added to the reaction mixture at 0 °C, and the resulting mixture was stirred at 25 °C for an additional 2 h. The reaction mixture was then extracted with ethyl acetate (3 × 5.0 mL). The combined organic layers were washed with brine, dried over anhydrous sodium sulfate, filtered, and concentrated under reduced pressure. The crude product was purified by flash column chromatography on silica gel (petroleum ether/EtOAc, 20/1‒5/1 ratio) to afford the desired 1,5-diketone product.

Transformations of alkyl-based enamino ketones with olefin

For 106-113, 116-135: 1,3-Alkyl-based enamino Ketone (0.1 mmol, 1.0 equiv.), olefin (0.3 mmol, 3.0 equiv.) and dichloromethane (5.0 mL) were sequentially added into a 25 mL Schlenk tube, degassed three times by freeze-pump-thaw method. The reaction mixture was stirred under an argon atmosphere at −67 °C, then irradiated by a 3 W purple LED (λ = 390 − 395 nm) from a 1.0 cm distance for 24 h. After completion of the reaction, the resulting mixture was concentrated under reduced pressure. The residue was treated with 1.0 N hydrochloric acid (HCl) in 1,4-dioxane (2.0 mL), then heated to 70 °C and stirred for 1 h. The mixture was then extracted with ethyl acetate (3 × 5.0 mL). The combined organic layers were washed with brine, dried over anhydrous sodium sulfate, filtered, and concentrated under reduced pressure. The crude product was purified by flash column chromatography on silica gel (petroleum ether/EtOAc, 20/1‒5/1 ratio) to afford the desired 1,5-diketone product.

For 114, 115: 1,3-Alkyl-based enamino Ketone (0.1 mmol, 1.0 equiv.), olefin (0.3 mmol, 3.0 equiv.) and dichloromethane (5.0 mL) were sequentially added into a 25 mL Schlenk tube, degassed three times by freeze-pump-thaw method. The reaction mixture was stirred under an argon atmosphere at −50 °C, then irradiated by a 3 W LED (λ = 360 − 365 nm) from a 1.0 cm distance for 48 h. After completion of the reaction, the resulting mixture was concentrated under reduced pressure. The residue was treated with 1.0 N hydrochloric acid (HCl) in 1,4-dioxane (2.0 mL), then heated to 70 °C and stirred for 1 h. The mixture was then extracted with ethyl acetate (3 × 5.0 mL). The combined organic layers were washed with brine, dried over anhydrous sodium sulfate, filtered, and concentrated under reduced pressure. The crude product was purified by flash column chromatography on silica gel (petroleum ether/EtOAc, 20/1‒10/1 ratio) to afford the desired 1,5-diketone product.

For 136-140: 1,3-Alkyl-based enamino Ketone (0.1 mmol, 1.0 equiv.), olefin (0.3 mmol, 3.0 equiv.) and dichloromethane (5.0 mL) were sequentially added into a 25 mL Schlenk tube, degassed three times by freeze-pump-thaw method. The reaction mixture was stirred under an argon atmosphere at −67 °C, then irradiated by a 3 W purple LED (λ = 390 − 395 nm) from a 1.0 cm distance for 24 h. After completion of the reaction, the resulting mixture was concentrated under reduced pressure. The residue was treated with 12 N hydrochloric acid (HCl) in CHCl₃: CH₃OH(1:1 v/v, 5.0 mL), then heated to 70 °C and stirred for 1 h. The mixture was then extracted with ethyl acetate (3 × 5.0 mL). The combined organic layers were washed with brine, dried over anhydrous sodium sulfate, filtered, and concentrated under reduced pressure. The crude product was purified by flash column chromatography on silica gel (petroleum ether/EtOAc, 20/1‒5/1 ratio) to afford the desired 1,5-diketone product.

General procedure for the kinetic resolution of imine intermediates

Imine intermediates (0.05 mmol, 1.0 equiv.), C1 (0.005 mmol, 0.1 equiv.), sodium dihydrogen phosphate (NaH₂PO₄) (0.05 mmol, 1.0 equiv.) and p-xylene: water (4:1 v/v, 5.0 mL) were sequentially added into a 25 mL Schlenk tube. The reaction mixture was stirred for t h. After completion of the reaction, the reaction mixture was then extracted with toluene (3 × 5.0mL). The crude product was purified by flash column chromatography on silica gel (petroleum ether/EtOAc, 20/1‒2/1 ratio). The intermediate ent-A was then treated with 1.0 N hydrochloric acid (HCl) in toluene (2.0 mL) at 0 °C, and the reaction was stirred at 25 °C for an additional 2 h. The mixture was then extracted with ethyl acetate (3 × 5.0 mL). The combined organic layers were washed with brine, dried over anhydrous sodium sulfate, filtered, and concentrated under reduced pressure. The crude product was further purified by flash column chromatography on silica gel (petroleum ether/EtOAc, 20/1‒10/1 ratio) to yield the desired enantiomerically enriched 1,5-diketone product.

Author contributions

Z.J. conceived and designed the experiments. W.Z., X.S., and D.T. performed the experiments. W.Z., X.S., X.Z., and Z.J. analysed and interpreted the results. W.Z. and Z.J. prepared the Supplementary Information. S.C. conducted density functional theory (DFT) calculations to determine the triplet energies of the relevant species. Z.J. and X.Z. wrote the paper. All authors discussed the results and commented on the manuscript.

Peer review

Peer review information

Nature Communications thanks Sanzhong Luo and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This work was supported by the National Natural Science Foundation of China (Nos. 22471064, W2441006, 22201069, 22301061, 22571072, 22401087, and 22401085), Central Government Guided Local Science and Technology Development Fund Projects (Z20231811081), Outstanding Foreign Scientists’ Studio of Henan Province (CZS2024008), and the Key Project of the Henan Provincial Natural Science Foundation (Nos. 252300421286 and 254000510005).

Data availability

All data are available from the corresponding author. The authors declare that all data supporting the findings of this study are available in the paper and its Supplementary Information files. Crystallographic data for the structures reported in this Article have been deposited at the Cambridge Crystallographic Data Centre, under deposition numbers CCDC 2531109 (33) and 2531110 (ent-141’). Copies of the data can be obtained free of charge via https://www.ccdc.cam.ac.uk/structures/. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-76455-x.