Kenneth H. Fuhr

New Synthetic Approaches to 4(1H)-Pyridones. II. Derivatives Having Nonaromatic Substituents

Abstract The preceding communication reports on the synthesis of 3,5-diaryl-4(1H)-pyridone derivatives via enaminoketones derived from the acylation of styrylamines with substituted arylacetyl chlorides. In this paper we describe methods for the preparation of 4(1H)-pyridones bearing carboxylic acid, ester, nitrile, cycloalkyl, cycloalkenyl, and derived functional groups in the 3-position.

Selective reduction of 2'-nitroanilides.

Hydrogenation of 2’-nitro-2,2,2-trifluoroacetanilide in ether over palladiumon-charcoal gives 2’-aminotrifluoroacetanilide that readily cyclizes to 2-trifluoromethylbenzimidazole in methanol solution.’ If the reaction is performed in ethanol with palladium-on-charcoal as a catalyst, two products are obtained2-trifluoromethylbenzimidazole and 1 -hydroxy-2-trifluoromethylbenzimidazole. The latter product can be made to predominate by the addition of an acid catalyst and can be readily separated from 2-trifluoromethylbenzimidazole by its solubility in sodium bicarbonate. Until this finding, the instances of occurrence of N-hydroxyimidazole derivatives by catalytic hydrogenation in the literature were rare, though the production of N-oxides by this route has been described several times (FIGURE 1 ) . It was found ?, that hydrogenation of an o-nitro-N-alkylanilide gave the benzimidazole N-oxide, but the reaction could not be extended to N-oxides unsubstituted at the 1-position (i.e., those capable of tautomerizing to an N-hydroxybenzimidazole) . An exception to this rule was o-nitroformanilide that, in ethanol, gave a low yield of l-hydroxybenzimidazole. Other workers have reported that sodium borohydride-platinum mixtures cause the same reacti~n.~. None of the other known methods for the introduction of an N-oxy function were applicable to the production of 2-perfluoroalkyl-1 -oxyimidazole derivatives (e.g., direct oxidation and hydrosulfide reduction I ) . and this prompted us to look further into the catalytic hydrogenation of o-nitroanilides and o-nitroperfluoroalkanoanilides, in particular. In general, 1-hydroxy-2-( 2,2-difluoroalkyl) benzimidazoles and imidazopyridines can be prepared by reduction-preferably by hydrogenation-of the corresponding 2’4troanilide. They are high-melting solids, considerably more acidic than the corresponding imidazoles (TABLE 1) and less soluble in nonpolar solvents. They exhibit a characteristic broad continuous hydroxyl absorption stretching down to below 2400 cm-1 (in contrast to the imidazoles, which have a broad “saw-toothed” absorption down to about 2650 cm-l) . If the cyclodehydration to give an N-hydroxyimidazole occurs at the hydroxylamine stage (which seems probable), then the reason for the rapid ring closure in the case of perfluoroalkanoanilides is the high positive charge density on the carbon atom of the amide carbonyl group that induces attack by the nucleophilic nitrogen of the hydroxylamine group (FIGURE 2). The use of a nonpolar aprotic solvent decreases yields, and solvents such as ethanol maximize yields (FIGURE 3). Addition of mineral acid also favors hydroxyimidazole formation, presumably by protonation of the carbonyl group, giving rise to a full formal positive charge on the carbon atom. Any substitution on the benzo ring that tends to lower this positive charge density lessens the chance of N-hydroxyimidazole formation in favor of the desoxy compound. This latter reaction is believed to occur by cyclodehydration of an intermediate, 2-aminopeduoroanilide, rather than deoxygenation of a hydroxyimidazole, because