Although aza-β-lactams have attracted interest due to their biological activity[1] and to their utility as intermediates in organic chemistry (e.g., for the generation of α-amino acids[2] and of hydantoins[3]),[4] only limited progress has been described with regard to the enantioselective synthesis of this family of heterocycles.[5] One attractive, convergent approach to the formation of aza-β-lactams is the [2+2] cycloaddition of a ketene with an azo compound [Eq. (1)].[6] To the best of our knowledge, no stereoselective variants of this process have yet been reported. (1) We have been exploring the use of chiral derivatives of PPY (PPY = 4-(pyrrolidino)pyridine; e.g., 1 and 2) as enantioselective catalysts for an array of transformations,[7] including couplings of ketenes with imines[8] and with aldehydes.[9,10] Although there are no reports of nucleophilic catalysis of [2+2] cycloadditions of ketenes with azo compounds, we were intrigued by the possibility that our planar-chiral pyridines might be effective in this role. In this investigation, we establish that PPY derivative 1 can achieve the first catalytic asymmetric synthesis of aza-β-lactams, via [2+2] cycloadditions of ketenes with azo compounds [Eq. (2)]. (2) As part of our initial study, we examined the cycloaddition of phenyl ethyl ketene with dimethyl azodicarboxylate (1.0 equiv). We were pleased to determine that planar-chiral PPY derivative 1 serves as an effective catalyst for the desired coupling, generating the aza-β-lactam in good ee and yield (Table 1, entry 1; in the absence of a catalyst, there is no reaction: entry 2). Under the same conditions, a related catalyst (2), as well as a variety of chiral phosphines and cinchona alkaloids, provide poor enantioselectivity or little of the cycloaddition product (e.g., entries 3–5).[11,12] The substituents of the azo compound have a significant impact on the ee and the yield, with the methoxycarbonyl group affording the best results (entry 1 vs. entries 6–9). If ClCH2CH2Cl, rather than CH2Cl2, is employed as the solvent, then formation of the aza-β-lactam proceeds less efficiently (entry 1 vs. entry 10). The reaction temperature of choice appears to be −20 °C (entry 1 vs. entries 11–12).[13] Table 1 Nucleophile-catalyzed enantioselective synthesis of aza-β-lactams: Effect of reaction parameters. The optimized conditions can be applied to the enantioselective synthesis of aza-β-lactams from a variety of ketenes (Table 2). If the alkyl group is small (Me or primary), the desired heterocycle is generally produced with good, not excellent, enantioselectivity (~85% ee; entries 1–7). Fortunately, the ee of the aza-β-lactams is readily enhanced by recrystallization (e.g., the product generated from phenyl ethyl ketene can be obtained in >99% ee after a single crystallization; see entry 2 of Table 2). In the case of ketenes that bear a secondary alkyl group, catalyst 1 typically furnishes the aza-β-lactam with very good enantioselectivity and yield (>90% ee; entries 8–13).[14] Table 2 Nucleophile-catalyzed enantioselective synthesis of aza-β-lactams (for the reaction conditions, see [Eq. (2)]). A plausible mechanism for this new nucleophile-catalyzed method for the synthesis of aza-β-lactams is illustrated in Figure 1. Interestingly, the configuration at the quaternary stereocenter is different from that produced in Staudinger reactions that are catalyzed by PPY derivative 1 [Eq. (3)],[8b] which are believed to proceed through a similar pathway.[15] Figure 1 Possible mechanism for the nucleophile-catalyzed synthesis of aza-β-lactams. (3) In conclusion, we have developed a new process, the nucleophile-catalyzed [2+2] cycloaddition of ketenes with azo compounds to generate aza-β-lactams. In addition, we have established that planar-chiral PPY derivative 1 achieves this convergent transformation with good enantioselectivity, thereby providing the first catalytic asymmetric synthesis of this useful family of heterocycles.
Read full abstract