In the current paper, taking into consideration our experience with the synthesis of pyridones together with our knowledge about the application of Meldrum’s acid25–30 in synthesis and considering of available literature we would like to present a new approach for preparation of 2-pyridones.
Taking into consideration our previous experience, we considered as a key step to be obtaining a 1,3-dicarbonyl amide with an already attached fragment allowing for cyclization and obtaining a 2-pyridone with a fused hexagonal ring, i.e. tetrahydroquinolin-2-one. The following enamide 11 on Fig. 3 would meet these requirements. We also assumed at this stage that such an enamide would undergo further desired cyclization in acidic conditions.
Theoretically, having the enaminone and the acylating reagent, obtaining the appropriate enamide should not be a problem, however, the use of acyl-acetate esters is not recommended due to too low reactivity with the amino group and side reactions with the ketone fragment. However, the use of strong acylating reagents such as chlorides, anhydrides, or ketenes, as in the cited works18–21, implies problems with their preparation or storage. Therefore, we proposed an idea to exploit acylating properties of Meldrum’s acid derivatives, which allow to introduce 1,3-dicarbonyl moiety together with a possibility to introduce broad scope of side chains. It should be stressed that used acyl derivatives of Meldrum’s acid are stable compounds, easily prepared and purified from commercially available starting materials.
First, we run a few experiments between 4 eq of 5-(hydroxy(phenyl)methylene)-2,2-dimethyl-1,3-dioxane-4,6-dione (9a) with 1 Eq. 3-amino-5,5-dimethylcyclohex-2-enone 10a in various conditions to optimize the method. As a result, we observed the formation of desired conjugated enamide with yield varying from moderate 30% in case of reaction performed in boiling DCE without molecular sieves up to quantitive yield 97% with the presence of molecular sieves and at 55°C in DCE (Fig. 3).
Thus, isolated and purified enamide 11aa was subjected to cyclization toward the formation of 7,7-dimethyl-4-phenyl-7,8-dihydroquinoline-2,5(1H,6H)-dione (12aa). Our first attempts were focused on the application of mild Lewis acids as transition metals triflates, especially scandium triflate. However, this approach was unsuccessful regardless to condition applied (Fig. 4).
This failure prompted us to search for an effective catalyst for the preparation of 7,7-dimethyl-4-phenyl-7,8-dihydroquinoline-2,5(1H,6H)-dione (12aa). We focused on protic acid catalysts especially on PPA as a moderately acidic agent with a well-known ability to catalyst similar reactions including Knorr-type cyclization31–37.
We run two experiments with the cyclization of enamide 11aa in the presence of PPA. First in boiling dichlorobenzene DCB with a reaction time of 2h allowed to the formation of 7,7-dimethyl-4-phenyl-7,8-dihydroquinoline-2,5(1H,6H)-dione (12aa) with 37% yield, whereas the second was performed in boiling dichloroethane DCE within 6h with yield 34% (Fig. 5).
With these results, we put the question if those two processes, inter and intramolecular, could undergo subsequently without isolation of enamide intermediate in a kind of “one-pot” reaction. To evaluate our hypothesis we again carried out the condensation of a fourfold excess of benzoyl Meldrum’s acid 9a with an enaminone 10a. After completion of the enamide formation and disappearance of benzoyl Meldrum acid revealed with TLC analysis, PPA was added to the reaction, and the whole mixture was heated to boiling point for 6 hours. As a result, we obtained the 7,7-dimethyl-4-phenyl-7,8-dihydroquinoline-2,5(1H,6H)-dione (12aa) with 32% yield (Fig. 6). As an alternative for the usage of PPA we tested TsOH in toluene however with a weaker result since the yield after cumulating two steps was only 15%.
Encouraged by these results, we decided to perform a series of reactions using various derivatives of Meldrum's acid and enaminones to evaluate the scope and limitations of the newly developed method. With our new method, we were able to prepare a wide range of compounds with 7,8-dihydroquinoline-2,5(1H,6H)-dione scaffold quickly in one laboratory step with moderate yields (Table 1).
In the case of 5-(1-hydroxy-2-(naphthalen-1-yl)ethylidene)-2,2-dimethyl-1,3-dioxane-4,6-dione used as a Meldrum’s acid (Table 1, Run 9) formation of enamide was observed but further condensation using PPA failed. An identical situation was observed when 3-aminocyclopent-2-enone was applied. Surprisingly in the case of 5-(hydroxy(4-methoxyphenyl)methylene)-2,2-dimethyl-1,3-dioxane-4,6-dione (9c) as a Meldrum’s component (Table 1, Run 3) significantly different product compared to the rest of the results was isolated from the reaction mixture. The first problem was the elucidation of a structure of a new unexpected product. It was substantial for us during the analysis that this compound had a double fragment coming from the Meldrum derivative and only one coming from the enaminone. Based on data from NMR ad MS we proposed the following structure, which could be in equilibrium with its keto form (Fig. 7).
Thus, we decided to elucidate the reaction mechanism behind this particular reaction. First, we decided to check which step of the process is responsible for the formation of unusual product. As was mentioned unexpected product contain an “excess” of moieties originating from acyl Meldrum’s acid, so it would suggest that the excess of 5-(hydroxy(4-methoxyphenyl)methylene)-2,2-dimethyl-1,3-dioxane-4,6-dione (9c) used in “one step” process causes such an untypical reaction course. Therefore, we prepared enamide 11ca according to the stepwise procedure, isolated and purified it. In the next step, an attempt to cyclize purified 11ca using PPA was performed. In this reaction, we gained again compound 13ca with a yield of 21% (Fig. 8).
The result confirmed that the excess of acyl Meldrum’s acid used in “one pot” wasn’t responsible for the unusual course of the reaction. Compound 13ca must be formed by the interaction of two molecules of enamide 11ca. To explain the formation of product 13ca we proposed a tentative reaction mechanism presented in Fig. 9.
Obviously, we paid attention to the fact that only the reaction of p-metoxy derivatives of Meldrum’s acid with dimedone enamineone gave us an unexpected product. Thus, to explain this phenomenon we put the hypothesis that EDG causes a decrease in the electrophilicity of the keto carbonyl carbon in enamide 11ca. Thereby inhibiting the privileged process of intramolecular cyclization which would lead to the usual product 12ca, simultaneously causes that an entropically more difficult intermolecular process could be observed. To validate our hypothesis, we decided to obtain and purified enamide 11ga with furyl substituent possessing a strong M + effect (Fig. 10a), and then once again carry out the intramolecular condensation with PPA. As a result, we obtained compound 13ga which confirmed that EDG indeed affects the course of the reaction (Fig. 10b).
In the case of intermolecular (Fig. 9) and intramolecular (Fig. 5) reactions the role of the nucleophile is played by the enaminone fragment having the largest contribution to the HOMO orbital of the molecule. The discussed intermolecular condensation can theoretically occur in two directions – attack of a nucleophile on a keto-carbonyl (not shown on Fig. 9) or on amide carbonyl carbon atom. Considering the factors affecting the transition of reaction (presence of ED group) from intramolecular to intermolecular, it’s likely that the enamineone nucleophile initiates the reaction with the amide carbonyl. Otherwise, we would be dealing with intramolecular cyclization.
Moreover in searching for the most basic position of the molecule, we estimated with the "Chemaxon pKa calculator"38 pKa values for the conjugate acids of the enamide 11ca and its enol form 11ca’ which are 1.0 and 0.4, respectively (Fig. 11).
In a reaction environment with an excess of PPA, these compounds will be largely in the protonated form. Obviously, the protonated forms are more electrophilic than the non-protonated. When considering the participation of these two forms in the actual course of the reaction, the position of the keto-enol equilibrium of protonated and deprotonated forms should be also taken into account. Considering the 1H NMR spectra for EDG-substituted enamides in a non-polar solvent, the amount of the enol form is negligibly small. Thus the most electrophilic species in our reaction mixture have to be protonated enamide 11ca (Fig. 11) which additionally supports our proposed mechanism.