Introduction

N-fused heterocyclic compounds, such as pharmaceuticals, agrochemicals, plastics, and dyes (Fig. 1a), are integrated into everyday life1,2,3,4,5,6. Big data analysis shows that heterocycle synthesis is one of the most common reactions in the field of medicinal chemistry7,8. Among the best-selling therapeutics, almost a third contain fused heterocyclic structures9. Due to the high value of N-fused heterocycles, their novel, effective, flexible, general syntheses require investigation10,11,12.

Fig. 1: Examples of critial fused-heterocycles and our reaction design.
figure 1

a N-fused heterocycles are ubiquitous within critical molecules, including biologically active natural and synthetic compounds and fine chemicals for use in functional materials. b Transfer-annulation strategy for synthesis of N-fused heterocycles. c Different strategies used in acyl transfer of ketones. d Fused heterocycle synthesis in this study via aromatisation-driven acyl tranfer of heteroaryl ketones with alkyl bromides.

Full size image

Acyl transfer is a critical process in various biological transformations13. In the field of organic synthesis, acyl transfer is frequently used in formation carbonyl compounds14,15,16,17,18. A typical acyl transfer employs a reactive carboxylic acid derivative (e.g. an acyl chloride or a thioester) as an acyl source. However, whether relatively inert ketones may serve as acyl transfer agents remains unclear?

Ketones are ubiquitous functional groups that not only occur widely in drug molecules and natural products but also act as bulk feedstocks in the syntheses of fine chemicals and materials. They are stable, non-toxic, and simple to prepare via various methods, rendering them ideal synthetic precursors19. If intramolecular acyl transfer of heteroaryl ketones can be realised, a transfer-annulation strategy may be employed in N-fused heterocycle preparation (Fig. 1b). However, owing to the kinetic inertness of C–C bonds, acyl transfer of ketones largely focuses on highly strained ketones20,21,22,23,24,25,26. For unstrained ketones27,28,29,30,31,32, the most common strategy involves using directing groups to form of a stable chelate (Fig. 1c)33,34,35,36,37,38,39,40. Although effective, the use of directing groups complicates the overall synthesis and limits the scope of the accessible products. Hence, a acyl transfer of unstrained ketones for use in N-fused heterocycle synthesis is warranted.

Aromatisation, which enables delocalisation of electron density, stabilising the molecule41, is a critical thermodynamic driving force in the field of organic chemistry42,43,44,45, e.g. aromatisation-driven deacylations of ketones are prominent bond-cleavage strategies46,47,48. Therefore, we conceived a approach for the acyl transfer of unstrained heteroaryl ketones driven by aromatisation of a pre-aromatic intermediate (Fig. 1d). This strategy may be suitable for use in the syntheses of N-fused heterocycles, and, critically, the directing group is no longer required. The next challenge in this strategy is the in situ formation of special, high-energy, pre-aromatic substrates. Transition metal-catalysed dearomatisation is a straightforward strategy to prepare spirocyclic scaffolds49,50,51,52. The spirocyclic intermediates, which are formed in situ from readily available heteroaryl ketones via dearomatisations53,54,55,56, should serve as pre-aromatic precursors to facilitate rearrangement (Fig. 1d). This likely involves a Pd-catalysed dearomative spirocyclisation of a heteroaryl ketone with an alkyl bromide to generate a pre-aromatic intermediate (A), which is then intramolecularly trapped by the heterocyclic nitrogen57,58,59,60,61. The resulting intermediate (B) may subsequently lose a hydrogen, restoring aromaticity to yield the fused heterocyclic product.

Here, we report an acyl transfer-annulation of heteroaryl ketones driven by aromatisation. This method is operationally simple, scalable, and applicable to late-stage modifications of natural products and drug derivatives, which make it a valuable method for the synthesis of organic N-fused heterocycles.

Results

Reaction optimisation

To explore this strategy, we initially used a heteroaryl ketone with a tethered olefin (1), which was prepared in one step using commercially available benzimidazole and 2-vinylbenzoyl chloride, as a model substrate. Because of the unique properties of difluoromethylene group (CF2) and its critical applications in medicinal chemistry62,63,64, ethyl bromodifluoroacetate (BrCF2COOEt) was employed as the coupling partner. After systematic screening, the desired rearrangement product (2) is obtained in a 90% yield using PdCl2 in combination with 1,1-bis(diphenylphosphino)pentane (dpppent, L1) as the ligand and Na2CO3 as the base in dioxane/tetrahydrofuran (THF) (Table 1, entry 1). The structure of 2 was unambiguously determined by X-ray crystallography. In addition, the Pd catalyst appears to be critical in this reaction. Using Pd(OAc)2 or Pd2(dba)3 (dba = dibenzylideneacetone) as the catalyst results in much lower yields (Table 1, entries 2–3), and other metals, such as NiCl2 and FeCl2, are completely ineffective (Table 1, entry 4). A study of the ligand effect further suggests that bidentate phosphine ligands are generally superior, with the yield increasing with the increasing bite angle of the phosphine employed, and L1 is the only ligand that generates full conversion with the optimal yield (Table 1, entry 5). The addition of a base improves the reaction outcome appreciably, likely by neutralising the in situ-generated HBr (Table 1, entry 6). A survey of different solvents reveals that dioxane and THF are individually good, albeit generating slightly lower yields than that obtained using the mixture (Table 1, entries 7–9).

Table 1 Screening of reaction conditions.
Full size table

Substrate scope

With the conditions determined, the scope of alkyl bromides was examined first (Fig. 2). Ketone 1 is successfully coupled with various alkyl bromides, with 5-, 6-, 7-, or 12-membered cycloalkyls (3–6) generating good yields of the desired coupling products. Heterocyclic bromides, with moieties such as tetrahydropyrane (7) and THF (8), react smoothly, resulting in good yields. Remarkably, the polycyclic bromide derived from the natural steroid stanolone is also amenable to coupling under the reaction conditions (9). Linear alkyl bromides are also suitable for reaction (10–12). We then investigated substrates with a CF2 group. Bromofluoroacetate, bromodifluoromethyl ketone, perfluoroalkyl bromide, bromodifluoromethyl phosphonate, and bromodifluoromethyl sulfone effectively undergo the desired annulation (13–17).

Fig. 2: Substrate scope of alkyl bromides.
figure 2

Unless otherwise specified, all the reactions were carried out using ketone 1 (0.1 mmol, 1.0 equiv) and alkyl bromide (0.15 mmol, 1.5 equiv.), PdCl2 (10 mol%), dpppent (12 mol%) and Na2CO3 (1.0 equiv) in dioxane/THF (1:2) at 130 °C. Isolated yields after chromatography are shown.

Full size image

We further explored the rearrangements of various heteroaryl ketones with bromodifluoroacetate (Fig. 3). The rearrangement took place smoothly by using 2-acylimidazoles and 2-acylbenzimidazoles as substrates (18–41). Both electron-rich and deficient substrates are competent during the cyclization process. A range of functional groups are compatible, including aryl fluorides (28 and 40) and chlorides (20 and 39), trifluoromethyl (21 and 38), esters (23) and cyano (22), are all tolerated. Changing the nitrogen protecting group from methyl to isopropyl (30) and benzyl (31) did not significantly affect the reactivity.

Fig. 3: Substrate scope of heteroaryl ketones.
figure 3

Isolated yields after chromatography are shown. The CCDC number of 43 is 2116753, 52 is 2116752. aThe reaction was performed under optimised condition A: ketone 1 (0.1 mmol, 1.0 equiv) and ethyl bromodifluoroacetate (0.15 mmol, 1.5 equiv), PdCl2 (10 mol%), dpppent (12 mol%) and Na2CO3 (1.0 equiv) in dioxane/THF (1:2) at 120 °C for 24 h. bThe reaction was conducted under optimised condition A with a slight modification: bis(2-diphenylphosphinophenyl)ether (DPEPhos) (12 mol%) was used as ligand during the reaction. cThe reaction was performed under optimised condition B: ketone 1 (0.1 mmol, 1.0 equiv) and ethyl bromodifluoroacetate (0.15 mmol, 1.5 equiv), PdCl2 (10 mol%), dppf (12 mol%) and K2CO3 (1.0 equiv) in dioxane/THF (1:1) at 130 °C for 24 h. dppf = 1,1′-bis(diphenylphosphino)ferrocene.

Full size image

Compared to the substrate with 4,5-diphenylimidazole (32), the reactions of 4-phenylimidazole (33) and imidazole (34) yield lower conversions, indicating that aromatisation is essential to promote the reaction. Marketed drug-derived ketones, such as ketoconazole (41), also react smoothly despite the presence of several other functional groups. Significantly, numerous substrates are synthesised via direct acylation of commercially available imidazoles or benzimidazoles, with the resulting ketones directly undergoing rearrangement, which further highlights the efficiency of this process. Further, we examined other types of heterocycles, which should yield different heterocyclic cores via rearrangement. Heterocycles such as thiazole (42), benzothiazoles (43–51), benzoxazole (52), and oxazole (53) may also be incorporated, yielding pharmaceutically interesting fused-ring skeletons65,66.

Mechanistic considerations

A study was performed to investigate the reaction pathway. To determine whether an alkyl radical exists during this Pd-catalysed process, a radical inhibition study was performed. When 2,2,6,6-tetramethylpiperidinooxy (TEMPO) is added to the reaction mixture, it traps alkyl radicals, indicating that the reaction involves radical species (Fig. 4a). An electron paramagnetic resonance (EPR) study of the reaction of bromocyclopentane with the spin-trapping agent phenyl-N-tert-butylnitrone reveals the presence of spin adducts of the trapped alkyl radicals 56 and 57 (Fig. 4b), as indicated by the EPR spectrum (see supporting information). Deuterium labelling studies were conducted using the heteroaryl ketone D-1 (79% deuterium content) as a substrate under the optimised conditions, with a significant level of the deuterated product D-2 (76% deuterium content) detected, suggesting that there were no reversible hydro-metallation in this process (Fig. 4c)67,68. Finally, we synthesised an aryl Pd complex (58-[Pd]), with 12 produced instead of 59 in the presence of 58-[Pd], benzyl bromide, and 1 (Fig. 4d). Therefore, the alkyl group of the fused heterocyclic product is not derived from the migratory insertion of the Pd(II) complex. The proposed reaction pathway is thus shown in Fig. 4e. The reaction may be initiated by a single electron transfer between Pd(0) and the alkyl bromide, producing hybrid alkyl Pd(I)-radical species INT I. Subsequently, radical addition to the alkene occurs, leading to the hybrid benzylic radical INT II, which then undergoes dearomatisative-spirocyclisation to form the spiro-N-radical INT III. Aromatisation-driven intramolecular acyl transfer may then occur to form the alkyl radical INT IV. Subsequent β-H elimination at the latter yields the product with concomitant regeneration of the Pd catalyst. This proposed mechanism is also supported by X-ray photoelectron spectroscopy, which revealed the presence of three distinct Pd oxidation states (Pd(0), Pd(I), and Pd(II)) during the process, suggesting that Pd(I) species may be involved.

Fig. 4: Mechanistic studies.
figure 4

a Radical trapping study using TEMPO showing that alkyl radical species are involved in the reaction. b EPR studies also suggest that this reaction may involve alkyl radicals. c Deuterium labelling studies. d Reaction of 1 with benzyl bromide in the presence of [Ph(PPh3)2PdBr] (58-[Pd]). e A proposed reaction pathway.

Full size image

Synthetic utility

Further studies were conducted to demonstrate the viability of this acyl transfer-annulation strategy. The protocol was applied in the late-stage modifications of natural products and drug derivatives (Fig. 5a). Various complex molecules with diverse structural features, such as steroids (62 and 69), N-heteroarenes (oxazole 63 and indole 68), alkaloids (66), and carbohydrates (72), are readily converted into the corresponding products in useful yields. This strategy provides a straightforward, versatile method of generating valuable N-fused heterocyclic moieties within complex molecules. Given the ubiquity of N-fused heterocycles in pharmaceuticals, this approach may be used in the field of medicinal chemistry.

Fig. 5: Synthetic applications.
figure 5

a Using the tranfer-annulation strategy in the late-stage modifications of complex frameworks based on natural products and drug molecules. b Gram-scale synthesis and various useful transformations of 2. The CCDC number of 74 is 2131840.

Full size image

To showcase the scalability of this process, a gram-scale reaction was carried out. Gratifyingly, a satisfactory 67% isolated yield (80% yield based on recovered 1) of product 2 could be obtained without modification of the optimised conditions (Fig. 5b). The N-fused heterocyclic scaffold can readily undergo various transformations to access a range of synthetically useful scaffolds. For example, the bromination of 2 proceeded to afford 74, excellent selectivity for the 9-position was observed, which allows follow-up fused heterocycle manipulations through cross-couplings. Treatment with mCPBA, deconstruction of N-fused heterocycle was observed, which afforded 75 in 53% yield. Diazidation product 76 was afforded in 48% yield via vicinal diazidation of olefin. Moreover, the ester moiety was smoothly reduced with NaBH4, affording the corresponding alcohol 77 in 68% yield.

In conclusion, a synthetically useful, mechanistically intriguing intramolecular acyl transfer of heteroaryl ketones was developed, which was suitable for use in fused-ring synthesis. The formation of a high-energy pre-aromatic spirocyclic intermediate was critical in the successful transformation, with aromatisation the driving force that facilitated C–C bond cleavage. Given the ready availability of the ketone moiety, this strategy could be used to simplify the syntheses of complex N-fused heterocyclic systems, which are privileged structures within numerous biologically active compounds. Moreover, the protocol enabled the late-stage modifications of intricate natural products and drug derivatives and may thus facilitate heterocyclic drug discovery.

Methods

General condition A for transfer-annulation of heteroaryl ketones derived from (benzo)imidazoles

In a nitrogen-filled glovebox, an oven-dried 10 mL sealed tube equipped with a Teflon-coated magnetic stir bar was charged successively with heteroaryl ketone 1 (0.1 mmol), alkyl bromide (0.15 mmol, 1.5 equiv), PdCl2 (0.01 mmol, 10 mol%), dpppent (0.012 mmol, 12 mol%), Na2CO3 (0.1 mmol, 1.0 equiv) and dioxane/THF (1.0 mL, 1:2). The tube then was sealed with a Teflon screw cap, moved out of the glovebox, and placed on a hotplate pre-heated to 130 °C for 24–36 h. After completion of the reaction, the mixture was filtered through a thin pad of silica gel. The filter cake was washed with ethyl acetate and the combined filtrate was concentrated under vacuum. The residue was purified via silica gel chromatography.

General condition B for transfer-annulation of heteroaryl ketones derived from (benzo)thiazoles and (benzo)oxazoles

In a nitrogen-filled glovebox, an oven-dried 10 mL sealed tube equipped with a Teflon-coated magnetic stir bar was charged successively with heteroaryl ketone 1 (0.1 mmol), difluorobromoethyl ester (0.15 mmol, 1.5 equiv), PdCl2 (0.01 mmol, 10 mol%), dppf (0.012 mmol, 12 mol%), K2CO3 (0.1 mmol, 1.0 equiv) and dioxane/THF (1.0 mL, 1:1). The tube then was sealed with a Teflon screw cap, moved out of the glovebox, and placed on a hotplate pre-heated to 120 °C for 24 h. After completion of the reaction, the mixture was filtered through a thin pad of silica gel. The filter cake was washed with ethyl acetate and the combined filtrate was concentrated under vacuum. The residue was purified via silica gel chromatography.