Three related diterpenoid quinones, royleanone, 9-acetoxyroyleanone, and Y-dehydroroyleanone, have been isolated from the roots of I n z~l a royleana D.C. These have been assigned structures on the basis of degradation studies and the synthesis of royleanone from ferruginol.The plant I n~~l a royleana D.C. is a perennial shrub that grows in the western Himalayas a t altitudes of 7,000 to 11,000 ft above sea level. I t is considered to be poisonous and has been used as an insecticide and a disinfectant (I). Its roots have been shown to be rich in derivatives of the diterpeiloid alkaloid lycoctonine (2-4). In addition, the presence of a yellow pigment has been reported (5). This piginent we now show to consist n~ainly of a mixture of diterpenoid quinones which we call 'royleanones'. The inajor con~ponent was an acetoxyroyleanone (I) (C22H3005). Small quantities of royleanone (IV) (CzoH28O3) and a dehydroroyleanone (111) (C2OH2603) were also present. T h a t these coinpounds were structurally closely related will become apparent in the sequel.We first concerned ourselves with establishing the nature of the chrolnophoric system in royleanone. The compound was acidic (pK, 8.5 in 50% aqueous methanol) and its alkaline solution was magenta in color. I t reacted with diazoinethane to give a monomethyl derivative. Royleanone took up 1 inole of hydrogen on catalytic hydrogenation and the colorless product was converted back to royleanone on standing in the presence of air. These properties, and the ultraviolet spectrum of royleanone, which was very similar to those of hydroxybenzoquinoi~es (G), suggested that royleanone contained this unit. The hydrogen-bonded hydroxyl band (vGFF13 3350 cm-I) and the two low-frequency carbonyl bands (vfiE2'3 1672 and 1632 cm-') in the infrared spectrum fitted this assignment. Finally, the reductive methylation of royleanone methyl ether (IX) to a triinethoxybenzene (VII) (Xgk:H 208 mp, E 49,000, 281 mp, E 740) furnished conclusive evidence that royleanone was a hydroxybenzoquinone.The structural relationship between royleanone, the dehydroroyleanone, and the acetoxyroyleanone was established as follows. Catalytic hydrogenation of the dehydroroyleanone (2 moles of hydrogen were talten up) followed by oxidatioil of the product by air gave royleanone. This indicated that dehydroroyleanone differed froin royleanone only in having an additional double bond. A comparison of the ultraviolet spectra of royleanone (XgIsH 277, 283 (shoulder), and 403 mp) and the dehydroroyleanone 212, 247 (shoulder), 330, and 459 mp) showed that this double bond was conjugated with the quinone ring.The similarity between the ultraviolet absorption spectra and inally features of the infrared absorption spectra of acetoxyroyleanone and royleailone led us t o believe that these compounds were related. The presence of an ester grouping in the acetoxyroyleanone was indicated by the carbonyl band a t 1748 cm-' in its infrared spectrum. In addition 'Issz~ed as 1V.IZ.C. No. 6030.

The alkaloids tuberostemonine, oxotuberostemonine, and the new base t~tberostemonine-X have been isolated from the rhizomes of Stentona sessilifolia. Tuberostemonine has beet1 oxidized to oxotuberostemonine and tuberostemonine-A has been related to tuberostemonine through common oxidation products. Study of transformation products of tuberostemonine has enabled a s s i g~l n~e~l t of partial structure V to the alkaloid.The work of Schild (1) and a group of Japanese chemists* has resulted in an array of trallsforinatioil products of the alkaloid tuberosteinonine. Their efforts have indicated the presence of an N-substituted pyrrolidine, a C-ethyl group, and two 7-lactones in the molecule. In the hope of obtaining other members of this interesting group of alkaloids for complementary investigation, we examined the bases from the rhizomes of Stemona sessilifolia Miq. (1). However, the main alkaloid proved to be tuberostemonine, the inail1 alkaloid of Stemona tuberosa,t with a much smaller quantity of a new base and traces of oxotuberosteilloiline (4).We hence coinrnenced a thorough exaizlination of the reactioils of the nitrogencontaining ring of tuberostemonine. Oxidation of the sodium salt of partially hydrolyzed tuberostemonine with permanganate gave, as reported earlier ( I ) , a lactain (vC,".23 1677 cm-l) which still contained a 7-lactone ring (vC,"z3 1770 cm-l). This analyzed correctly for C17H25N03 and hence had lost a five-carbon group containing one lactoile ring. Since no new oxygen function other than the lactarn carbonyl appeared in the Cl7 compound, the C5 unit must have been attached to one of the carbons a to the nitrogen. The forination of I-methylsuccinic acid was observed during an ailalogous oxidation ( I , ?), suggesting that this represented the Cg unit lost.Further insight into the nature of the C5 unit came from a dehydrogenation of the pyrrolidine ring. We were able to obtain an approxiinately 40% yield of the previously described bisdehydro coillpound by silver oxide oxidation of tuberostemonine (1, 5 ) . The ultraviolet spectrum (XE,t,q,H 236 mp (E 9200)), the positive Ehrlich test (I), and coupling with diazonium salts were all consistent with this being a pyrrole. We were able to confirm Schild's observation (1) that this underwent hydrogenolysis involving 1 mole of hydrogen, to give an acid. The latter was characterized for the first time. The hydrogeilolysis and formation of the C17 lactain require that the pyrrole ring and one lactone have the relationship expressed in I.