Formation of bicyclic pyrroles from the catalytic coupling reaction of 2,5-disubstituted pyrroles with terminal alkynes, involving the activation of multiple C–H bonds

Abstract

Substituted bicyclic pyrroles are produced directly from the coupling reaction of 2,5-disubstituted pyrroles with terminal alkynes, involving the activation of multiple C–H bonds and regioselective cyclisation.

Transition metal-catalysed C–H bond activation and functionalisation reactions of nitrogen heterocyclic compounds have attracted considerable attention, in part due to their prominent role in the synthesis of natural products and pharmaceutical agents.¹ Highly regioselective catalytic C–H bond insertion reactions of nitrogen-containing aromatic compounds, such as pyridines, indoles and pyrroles, have been reported in recent years.² Direct oxidative coupling reactions of arene C–H bonds³ and the C–H bond oxidative annulation of indoles⁴ have also been achieved using Cu and Pd catalysts. Despite such remarkable progress, however, catalytic C–H bond activation methods have rarely been employed for constructing nitrogen-containing heterocyclic compounds. We recently developed a new catalytic coupling reaction between arylamines and alkynes, which involved the regioselective activation of sp² C–H bonds to yield tricyclic quinoline products.⁵ In an effort to extend the scope of catalytic C–H bond activation reactions, we have begun to explore the coupling reactions of pyrroles and indoles. This report delineates the coupling reaction between 2,5-disubstituted pyrroles and terminal alkynes, which involves multiple C–H bond activation and cyclisation steps.

Treatment of 2,5-dimethylpyrrole (1.0 mmol) with 4-ethynylanisole (2.0 mmol) in the presence of Ru₃(CO)₁₂/NH₄PF₆ (1: 3, 10 mol% Ru) in benzene (5 mL) at 95 °C for 36 h cleanly produced the cyclisation product, 1a (eqn (1)). Since 1a was found to be air sensitive, the analytically pure product was isolated by column chromatography under a nitrogen atmosphere (87% yield), and was fully characterised by both spectroscopic methods and elemental analysis.^‡ The initial catalyst activity survey showed that both Ru₃(CO)₁₂ and NH₄PF₆ were essential for catalytic activity. Other neutral and cationic ruthenium compounds, such as RuCl₃·3H₂O, (PPh₃)₃RuHCl, (PCy₃)₂(CO)RuHCl and [(PCy₃)₂(CO)(MeCN)₂RuH]⁺BF₄⁻, did not give any coupling products under similar reaction conditions. The analogous reaction of N-phenylpyrrole with 4-ethynylanisole produced a mixture of 1: 1 and 1: 2 coupling products, without forming any cyclisation product.

(1)

The coupling reaction was found to be strongly influenced by the steric and electronic nature of alkynes. In contrast to terminal alkynes with a para-electron-donating group, such as 4-ethynylanisole or 4-ethynyltoluene, which readily produced the cyclisation products 1a–1d, the coupling reaction with phenylacetylene gave a 1: 1 mixture of the cyclisation and 1: 2 insertion products, 1e and 2e. The coupling reaction with sterically demanding 2-ethynyltoluene (2 equiv.) produced a 3: 2 mixture of the coupling products 2f and 3f under similar conditions. Neither arylalkynes with an electron-withdrawing group, such as 4-ethynyltrifluorotoluene or 4-fluorophenylacetylene, nor the aliphatic terminal alkynes, gave any coupling products under similar conditions. A prolonged reaction time at a higher temperature did not convert 2 or 3 into cyclisation product 1. Instead, the cyclotrimerisation products from the homocoupling of the terminal alkynes were produced predominantly in these cases.

Since Ru₃(CO)₁₂/NH₄PF₆ was not particularly effective for the coupling reactions with electron-poor arylalkynes, we next surveyed the efficacy of gold catalysts to promote the formation of cyclisation products. When 2,5-dimethylpyrrole was treated with phenylacetylene (2 equiv.) in the presence of 5 mol% of Au(PPh₃)Cl/AgOTf (1: 1) in benzene for 24 h, 1e was formed exclusively, though the catalyst lost its activity after 60% conversion. Control experiments indicated that both Au(PPh₃)Cl and AgOTf were required for catalytic activity, and other selected gold compounds, such as AuCl₃ and NaAuCl₄, failed to catalyse the coupling reaction. When the Au(PPh₃)Cl/AgOTf (5 mol%) catalyst was treated with a 1: 1 mixture of 1e and 2e, 2e was cleanly converted to 1e to produce an 8: 1 mixture of 1e and 2e after 10 h at 95 °C. By using the combined catalytic system, Ru₃(CO)₁₂/NH₄PF₆ and Au(PPh₃)Cl/AgOTf, cyclisation product 1e was obtained from the coupling reaction of 2,5-dimethylpyrrole with phenylacetylene (>95% conversion, 81% combined yield, 1e: 2e = 85: 15). This result indicates that the gold catalyst was particularly effective in promoting the cyclisation step of the coupling reaction. While gold catalysts have been successfully utilised in C–H bond activation reactions,⁶ the synergistic effect of Ru/Au catalysts is not entirely clear at the present time.

The formation of both 1: 1 and 1: 2 products suggested that product 1 is resulted from the cyclisation of 1: 2 coupling product 2. To gain further mechanistic insights, the reaction mixture of 1e and 2e (1: 1) was periodically monitored by ¹H NMR at room temperature, after it had been heated at 95 °C in the presence of Ru₃(CO)₁₂/NH₄PF₆ (10 mol% Ru) in C₆D₆ (Fig. 1). Over time, the peaks due to 1e at δ 6.19, as well as the NH peak at δ 6.24, increased at the expense of the peaks due to 2e (δ 5.27 and 5.53 (CQCH₂)). The rate constant, k_obs=2.1 × 10⁻² h⁻¹, of the appearance of 1e was estimated from a pseudo first-order plot.

Fig. 1.

Partial ¹H NMR spectra of the reaction mixture of 1e and 2e.

The coupling reaction of 1,2,5-trimethylpyrrole with deuterium-labelled 4-ethynylanisole-d₁ (2 equiv., >99% D) in the presence of Ru₃(CO)₁₂/NH₄PF₆ (10 mol% Ru) in C₆D₆ was monitored by NMR. After 1 h of heating at 95 °C, the ¹H NMR spectrum showed that nearly 15% of the deuterium from 4-ethynylanisole had exchanged with 35% of the β-vinyl hydrogens of the unreacted 1,2,5-trimethylpyrrole, prior to the formation of the coupling products. The product, 2a-d, isolated from a preparative scale reaction of 2,5-dimethylpyrrole with 2 equivalesnts of 4-ethynylanisole-d₁, contained deuterium at both the α-methyl (33%) and vinyl (37%) positions. Also, in support of rapid H/D exchange between the two substrates, a relatively small deuterium isotope effect was observed from a separate reaction of 1,2,5-trimethylpyrrole with phenylacetylene/ phenylacetylene-d₁ when forming 1: 1 coupling product 3e. The pseudo first-order plots for the reactions gave k_obs = 1.65 × 10⁻² and 1.38 × 10⁻² h⁻¹ from phenylacetylene and phenylacetylene-d₁, respectively, which translated into k_CH/k_CD = 1.2.

These results suggest a mechanism involving sequential alkyne insertion and cyclisation steps, as outlined in Scheme 1. The sequential C–H activation and regioselective insertion of alkynes would be mediated by an electrophilic ruthenium catalyst to form 1: 2 coupling product 2. The subsequent ruthenium-mediated vinyl C–H bond activation and cyclisation steps could be facilitated by coordination of the adjacent olefin to ruthenium via the formation of alkene–hydride species 4. Cyclisation and reductive elimination would give product 1.

Scheme 1.

A possible mechanistic pathway.

In summary, the catalytic formation of bicyclic pyrroles has been achieved from the direct coupling reaction of 2,5-di-methylpyrroles with terminal alkynes. The cyclisation reaction involved three consecutive sp² C–H bond activation and insertion steps.