2308
COMMUNICATIONS TO THE EDITOR
Vol. 72
THE OXINDOLE NATURE OF GELSEMINE’
partial structure I is present in gelsemine and that it is the reduction of a carbonyl to a methylene group that gives rise to tetrahydrodesoxygelsetained a free hydroxyl group which could be mine in which partial structure I1 must be presacetylated with acetic anhydride. All attempts ent. Oxindoles still containing a hydrogen atom to prepare an 0-acetylgelsemine have failed in in the 3-position are converted to indoles by the our hands, and other workers3 have been similarly action of lithium aluminum hydride’ and hence, unsuccessful. It can now be stated quite defi- the 3-position of the oxindole system in gelsemine nitely that gelsemine does not contain a hydroxyl must be assumed to be disubstituted. Indeed, the group. The base has been found to contain only disubstituted 3,3-dimethyloxindole gave color one active hydrogen (Zerewitinow) and its infrared changes with sulfuric acid and potassium diabsorption spectrum to show a sharp absorption chromate very similar to those given by strychband a t 3443 ern.-' indicative of an imino group. nine and gelsemine and its ultraviolet absorption Since gelsemine is an indole a l k a l ~ i dand ~ , ~ since curve was almost superposable on that of gelsethe basic nitrogen is known to be tertiary, it follows mine. Furthermore, it has been ascertained that that the one active hydrogen in the base must be 3,3-dimethyloxindole is readily reduced by lithium aluminum hydride in dioxane solution to 3,3-diattributed to the indole imino group. It is The reaction of dihydrogelsemine with lithium methyldihydroindole (m. p. 34-345’). aluminum hydride in dioxane gave rise to tetra- noteworthy that this result is contrary to the literature statement that N-unsubstituted oxindoles hydrodesoxygelsemine, prisms, m. p. 138-139’. Anal. Calcd. for CzoH260N~:C, 77.42; H, 8.39; are virtually unaffected by lithium aluminum N, 9.03; 1 act. H, 0.32. Found: C, 77.51, 77.34; hydride. The partial structures I and I1 are further supH, 8.01, 8.21; N, 9.14; act. H, 0.33. This base forms a picrate, orange needles, m. p. 216-217” ported by the strychnidine-like properties of Like strychnine it O~N~ (dec.). Anal. Calcd. for C ~ O K ~ , @ N Z . C ~: H ~tetrahydrodesoxygelsemine. C, 57.88; H, 5.38; N, 12.99. Found: C, 57.23, must contain a nitrogen atom directly attached 57.40; H, 5.23, 5.19; N, 12.80. The new base to an aromatic nucleus with a free and reactive contained one strongly basic nitrogen atom (PK para-position. When coupled with diazotized 8.9) and one weakly basic (PK 3.4). The infrared sulfanilic acid, it gave rise to a methyl orange absorption spectra6 of both gelsemine and di- which resembled in its indicator behavior that of hydrogelsemine showed an absorption band a t sulfobenzoazostrychnidine. 1720 cm.-l in the region of carbonyl absorption, (7) Julian and Printy, THISJOURNAL, 71, 3206 (1949). while that of tetrahydrodesoxygelsemine showed CHEMISTRY DIVISION RESEARCH COUNCIL MORRISKATES no absorption in this region, but still contained an NATIONAL CANADA L t b MARION absorption band (3414 cm.-l) indicative of a OTTAWA, RECEIVED FEBRUARY 13, 1950 weakly associated imino group. Tetrahydrodesoxygelsemine contained one active hydrogen (Zerewitinow) ; it was acetylated by boiling acetic THE SEPARATION AND IDENTIFICATION OF anhydride to an N-acetyl base, prisms, m. p. 131- KETOACIDS BY FILTER PAPER CHROMATOGRAPHY 132’. Anal. Calcd. for C22H2802N2: C, 75.00; H, 7.95. Found: C, 75.81; H, 7.44. The acetyl de- Sir: rivative formed a crystalline acetate, plates, m. p. I n the course of studies on the amino acid Anal. Calcd. for C ~ ~ H Z ~ O ~ N Y Cmetabolism ~ H ~ O ~ : of microorganisms carried out in this 182-186’. C, 69.90; H, 7.77; N, 6.80. Found: C, 69.96, laboratory, i t became desirable to isolate and 69.70; H, 7.77, 7.58; N, 6.74, 6.77. The infra- identify minute quantities of ketoacids. red absorption spectrum of the acetyl derivative It was found that mixtures of ketoacids could showed no absorption band indicative of an imino be separated by unidimensional ascending partigroup, but contained a carbonyl absorption band tion chromatography on filter paper. The posi(1650 cm.-l) obviously due to the acetyl group. tions of the ketoacids on the chromatogram It can, therefore, be postulated that were made visible by spraying the paper (E. and D. 613), with an aqueous solution containing 0.1% C C semicarbazide hydrochloride and 0.15% sodium acetate, drying a t l l O o , and inspection under an ultraviolet lamp (“Mineralight”). The semicarbazones of the ketoacids appear as dark shadI I1 ows on the faintly fluorescing paper. AS little as 5 pg. of ketoacid could be demonstrated. (1) Issued as N.R. C. Bull. No. 2082. The approximate RF values in a solvent sys(2) Moore, J . Chcm. Soc., 1231 (1911). (3) Forsyth, Marrian and Stevens, ibid., 579 (1945). tem consisting of 95 parts n-butanol, 5 parts (4) Marion, Can. J . Research, 2lB,247 (1943). formic acid, saturated by shaking with water, (5) Witkop, THISJOURNAL, 70, 1424 (1948).
Sir: It was claimed by Moore2 that gelsemine con-
(6) All infrared spectra were determined in chloroform solution by Dr. R . Norman Jones.
(1) Chargaff, Magasanik, Doniger and Vischer, THISJOURNAL, 71, 1513 (1949).
