ChemInform Abstract: Ring Enlargement of N‐Heterocycles

Abstract

This review describes our preliminary results of ring enlargement of N-beterocyc1es by N-ylide formation at a targeted site, exemplified by 55~56~58. Before development of this method, the strategy can be traced to our modified Hofmann degradation led at first to several miscarried probings on the reaction between N-halomethyl ammonium salts and halophilic reagents. 601The ring enlargement of N-heterocycles to be described at the end of this article, in the form of only a few preliminary explorations, involves insertion of an extra CH 2 group from an external source. This method promises to be attractive for naturally occurring alkaloids of certain types that respond to the experimental protocol.The idea of ring enlargement of N-heterocycles germinated from seemingly unrelated work in our research group.Several years ago, we had occasion to isolate from Spiraea japonica L f. vss.fortunei(PI.) Rhed. more than a dozen diterpene alkaloids with previously unreported structures, designated spirasines.' A quaternary salt 2 was also isolated that seemed to be an artefact derived from spirasine IX 1. The suspected culprit dichloromethane was indeed used as one solvent for extraction purposes at ambient temperature. No reaction occurred when 1 was purposely heated in CH2Ch under reflux for one week. Perhaps an unknown catalyst was operative during the extraction. In retrospect, it is also puzzling why only spirasmeIX 1, of all duerpenoid congeners, should succumb to quaternization. The reason might be a flaw during isolation that somehow failed to uncover other quaternary salts in trace amounts.To insure the structural kinship between 1 and 2, we treated compound 2 with boiling alkali in order to hydrolyze off tile chloromethyl group. The major product was 3, in 60% yield, as a result of a Hofmann-type degradation via immonium intermediate 4. The ease of modified Hofmann degradation contrasts with failure of reaction in the case of the classical methiodide of a similar alkaloid.' The extended fragmentation of halomethyl derivative 5, compared with classical counterpart 6, has two favorable features; namely that halide (X-) is a much better leaving group than the tertiary amine 2, and that the transition structure of 5 is energetically favored by the presence of 1t-1t conjugation. The two 1t-bonds in 4 are not coplanar, although partial overlap still prevails between the top lobe of one 1C-bond and the bottom lobe of the other.