A catalyst-free multicomponent domino sequence for the diastereoselective synthesis of (E)-3-[2-arylcarbonyl-3-(arylamino)allyl]chromen-4-ones

Summary The three-component domino reactions of (E)-3-(dimethylamino)-1-arylprop-2-en-1-ones, 3-formylchromone and anilines under catalyst-free conditions afforded a library of novel (E)-3-(2-arylcarbonyl-3-(arylamino)allyl)-4H-chromen-4-ones in good to excellent yields and in a diastereoselective transformation. This transformation generates one C–C and one C–N bond and presumably proceeds via a reaction sequence comprising a Michael-type addition–elimination reaction, a nucleophilic attack of an enamine to a carbonyl reminiscent of one of the steps of the Bayllis–Hilman condensation, and a final deoxygenation. The deoxygenation is assumed to be induced by carbon monoxide resulting from the thermal decomposition of the dimethylformamide solvent.

The structure of compounds 5 was deduced from one and twodimensional NMR spectroscopic data, as detailed in Supporting Information File 1 for 5h as a representative example. The structure of 5h deduced from the NMR spectroscopic studies was subsequently confirmed by single-crystal X-ray crystallographic data, as shown in Figure 1 [56]. Our initial mechanistic hypothesis accounting for the formation of compounds 5 is depicted in Scheme 3. An initial Michael addition-elimination reaction, leading to the exchange between the starting arylamine 3 and dimethylamine, explains the formation of intermediates 4, which were the final reaction products in most investigated solvents. However, in DMF, the enamine moiety of 4 attacks the aldehyde group in 1 giving rise to 6. This reaction resembles one of the steps of the Bayllis-Hilman condensation and is presumably promoted by the presence of traces of formic acid as a contaminant of the solvent [57].
We established the feasibility of the first step by showing that the reaction between two of our starting enaminones, namely 2a (Ar = Ph) and 2c (Ar = 3-NO 2 C 6 H 4 ) with aniline under our standard reaction conditions (DMF, 130 °C) affords the corresponding compounds 4 after 3 h in 94% and 90% yields, respectively. As previously mentioned (Table 1), the nature of the solvent was found to be of critical importance, and the transformation of intermediate 4 into the final products 5 was found to work only in DMF, among many investigated solvents. Since the reaction between 4 and 1 is proposed to be catalyzed by acid, we carried out the multicomponent reaction in a selection of solvents (dioxane, acetonitrile, dimethyl sulfoxide, ethanol, ethylene glycol) in the presence of one equivalent of HCl, but all these attempts failed, while the same conditions were successful in DMF. These results can be explained by assuming that the initial aldol-type reaction between 1 and 4 to give 6 is reversible and is driven to completion by the reduction step, which takes place in DMF only. Thus, interestingly, the reaction does not stop at compound 6, but instead undergoes a reductive termination step, which leads to the final products 5.
This reduction step is intriguing, and a first mechanistic possibility could be a hydride transfer from formate anion. While this mechanism might partly account for the observed reduction, we acknowledge that it is problematic to rely on solvent impurities to account for a stoichiometric reduction. Alternatively, dimethylformamide itself might have been the reducing agent.
There are a few scattered literature reports on the use of DMF as a reducing agent, the first of which seems to be the reduction of diazonium tetrafluoroborate salts to arenes [58]. DMF has also been described as a reducing agent acting by hydride transfer in Pd-catalyzed processes, where the metal plays a critical role in the reduction by decomposing it and facilitating hydride transfer from a Pd intermediate [59][60][61]. However, it is not clear whether the same type of reaction may take place under our conditions. In order to test this ideas experimentally, we carried out the reaction between 3-formylchromone (1), (E)-3-(dimethylamino)-1-(4-chlorophenyl)prop-2-en-1-one and p-toluidine in DMF-d 7 . If the mechanistic hypothesis was correct, this reaction should lead to a d-7 intermediate and hence to a mixture of 5 and d-5, with the latter being major due to isotope effects in the tautomeric equilibrium. However, this reaction failed to show any incorporation of deuterium into 5, and therefore we had to abandon the hydride-transfer hypothesis.
We came up with an alternative explanation based on the wellknown fact that DMF decomposes into dimethylamine and carbon monoxide at its boiling point. Carbon monoxide can act as a deoxygenating agent [62] and thus explain the transformation of 6 into 5 as shown in Scheme 4. Because of the synthetic interest of allylic and benzylic deoxygenations, further research into this reaction is under way in our laboratories.

Scheme 4:
Alternative mechanistic proposal based on a carbon monoxide-induced deoxygenation.

Conclusion
We have developed a facile three-component diastereoselective synthesis of novel (E)-3-[2-arylcarbonyl-3-(arylamino)allyl]-4H-chromen-4-ones containing chromone and β-enaminoketone structural fragments from simple, readily available starting materials in a one-pot operation and in good to excellent yields. This transformation occurs via a domino sequence of reactions, which generates one C-C and one C-N bond.
Presumably, this transformation proceeds via a reaction sequence comprising a Michael-type addition-elimination reaction, a nucleophilic attack of an enamine to a carbonyl, and a final deoxygenation step. We propose that the deoxygenation step is induced by carbon monoxide resulting from thermal decomposition of the dimethylformamide solvent.

Experimental
General methods. Melting points were measured in open capillary tubes and are uncorrected. The 1 H NMR, 13 C NMR, DEPT, H,H-COSY, C,H-COSY and HMBC spectra were recorded on a Bruker (Avance) 300 MHz NMR instrument by using TMS as an internal standard and CDCl 3 as a solvent. Standard Bruker software was used throughout. Chemical shifts are given in parts per million (δ-scale), and the coupling constants are given in Hertz. Silica gel-G plates (Merck) were used for TLC analysis with a mixture of petroleum ether (60-80 °C) and ethyl acetate as an eluent. Elemental analyses were performed on a Perkin Elmer 2400 Series II Elemental CHNS analyzer.
General procedure for the synthesis of (E)-3-(2-arylcarbonyl-3-(arylamino)allyl)-4H-chromen-4-one derivatives 5a-5p. In a similar manner as described in [55], a mixture of 3-formylchromone (1, 1 mmol), enaminone 2 (1 mmol) and aniline 3 (1 mmol) in DMF (5 mL) was heated at 130 °C for 6-7 h. The reaction progress was monitored by TLC. After completion of the reaction, the solvent was removed and the product was purified by column chromatography with a petroleum ether-ethyl acetate mixture (4:1 v/v) as an eluent to afford compounds 5. Characterization data for representative com-pounds are given below. The characterization data for the full library can be found in Supporting Information File 1.     General procedure for the isolation of intermediates 4. A mixture of the suitable enaminone 2 (1 mmol) and aniline 3 (1 mmol) in DMF (5 mL) was heated at 130 °C for 3 h. The reaction progress was monitored by TLC. After completion of the reaction, the solvent was removed and the product was purified by column chromatography with a petroleum ether-ethyl acetate mixture (4:1 v/v) as an eluent. Characterization data for compounds 4 can be found in Supporting Information File 1.

Supporting Information
Supporting Information File 1 Experimental details, full characterization data, detailed structural characterization of compound 5h and copies of the spectra of all compounds.