It has been recently recognized that the indenobenzazepines are a distinct group of alkaloids, probably derived biogenetically from the rearrangement of spirobenzylisoquinolines (1). To date, only five naturally occurring indenobenzazepines were known (1-5). We now wish to describe a new indenobenzazepine, namely the colorless and optically inactive bulgaramine (l), C, ,H, 'NO4, obtained from Herba Fumaria Officinalis where it is accompanied by 0-methylfumarophycine (2) (6) and other spirobenzylisoquinolines (7).Bulgaramine (1) shows a complex uv absorption pattern with a strong maximum at 34 1 nm suggestive of a stilbenoid system (see Experimental section) (8). The mass spectrum exhibits a molecular ion peak mlz 35 1, which is also the base peak.Substantive evidence for the structure of bulgaramine was derived from the 360 MHz 'H-nmr spectrum in CDCl,, which has been summarized around expression 1. Salient features of the 'Hnmr spectrum are the broad singlet centered at 6 3.83 and representing the C-13 methylene protons. Long-range coupling obtains between the C-13 protons and H- 12. Because H- 12 is also split by H- 1 1, it appears at 6 6.90 as a doublet of triplet v0=7.7 Hz, J12,13A=0.9 Hz, Jiz,13~=0.9 Hz.An accompanying 'H-nmr nOeds (nuclear Overhauser enhancement difference study) (9), summarized in expression la, provided an interlocking system relating all of the protons in the alkaloid with the exception of those for the ring D methylenedioxy substituent which, as expected, do not show any nOe. The numerical values quotedaround diagram la represent the apparent nOe percentages actually observed. Irradiation of the 6 3.83 C-13 protons resulted in a 32% nOe of H- 1 (6 7.02), as well as a 16% enhancement of the H-12 absorption (6 6.90). In turn, irradiation of H-1 caused an 18% increase in the area of the H-13 protons (6 3.83), as well as a 15% increase of the C-2 rnethoxyl signal (6 3.94).Bulgaramine (1) had been previously prepared semisynthetically by the pyrolysis of 0-methylfumarophycine (2) (IO), and our natural and semi-synthetic samples proved to be identical. Several experiments were presently carried out to demonstrate that natural bulgaramine was not derived from the accompanying alkaloid 0-methylfumarophycine (2) during the isolation process. In one such experiment, 0-methylfumarophycine was allowed to stand overnight in 5% HC1 at room temperature. No bulgaramine could be detected after workup. In another experiment, O-methylfumarophycine was kept for several hours at near 90" (water bath). Such heating did not produce any bulgaramine.Bulgaramine (1) occupies a central position in the biogenetic scheme linking the spirobenzylisoquinolines with the indenobentazepines. It has been suggested that the ketonic spirobenzylisoquinoline alkaloid parfumidine is reduced in the plant to the alcohol fumaricine (3), which can supply the aziridinium cation 4 (1). Proton loss from 4 with fission of the aziridinium ring provides bulgaramine (1). Oxidation and aromatization of 1 would then lead to lahoramine, and indeed a very similar scheme has been duplicated in vitro (1). Besides O-methylfumarophicine (2), both parfumidine and fumaricine (3) occur in Herba Fumaria Officinalis (7). The isolation of bulgaramine (1) as a natural product thus lends further credence to the thesis that indenobenzazepines are derived biogenetically from spirobenzylisoquinolines. Characteristically, bulgaramine bears a methylenedioxy substituent in the bottom ring, as do all known spirobenzylisoquinoline and indenobenzazepine alkaloids.