MOORE : THE cmsrI‘rumn OF GELSEMIUM. 2223CCXXXI1.-The Constituents o f Gelserniuna.By CHARLES WATSON MOORE.UNDER the title of “gelsemium” several of the pharmacopaeiasrecognise the dried rhizome and roots of Gelsemiurn sempervirens,Aiton, commonly known as the “ yellow jessamine.”The medicinal value of the plant is due to the presence of certainalkaloids, only one of which, however, has been obtained in a crystal-line condition.Among the earlier investigations of gelsemiurn there may be notedthat of Wormley (Amer. J. Pharm., 1870, 42, l), who isolated animpure alkaloidal product to which he gave the name of “ gelseminine.”This base was afterwards investigated by Sonnenschein ( B e y . , 1876,9, 1182) and Gerrard (Pharm. J., 1883, 13, [iii], 641), who assignedto it the formulse C,,H,,O,N, and C2,H,,0,N2 respectively.The last-mentioned investigator was the first to obtain gelsernine and its saltsi n a crystalline state. Thompson (Jahresbey., 2887,2218), who ascribedto gelsemine the formula C,,H,,O,,N,, showed that i t was accompaniedin the plant by a second alkaloid, which he obtained in an amorphouscondition, and which he designated as ‘‘ gelseminine.” Both gelsemineand gelseminine have more recently been examined by Cushny(Bey., 1893, 26, 1725), who proposed the formulae C,,H,,O,,N, andC,,H470,4N, respectively for the two bases. Spiegel (Be?*., 1893, 26,1045) suggested the formula C,,H,,O,N, for the crystalline base, whichwas confirmed by Gceldner (Ber. deut. pharm. Ges., 1895, 5, 330), whoobtained it in colourless crystals, melting at 160’.Some confusion has arisen as to the nomenclature of the two basesisolated from gelsemium; thus in the English literature the crystal-line base is referred to as gelsemine, and the amorphous product asgelseminine, whilst most of the German investigators, for example,Spiegel (Zoc.cit.) and Gceldner (Zoc. cit.), use these names in the oppositesense. I n this communication the English nomenclature is adhered to.The present investigation has resulted in the isolation of the tzlkaloidgelsemine in a pure crystalline condition. The base is found to meltconsiderably higher than has hitherto been recorded (m. p. 178”, instea2224 MOORE : THE CONSTITUENTS OF GELSEMIUM.of 160°), and it has been conclusively shown to possess the formulaC2,,H2,0,N2.Besides gelsernine and gelseminine, the presence of athird alkaloidal substance in gelsemiurn has been established. Thissubstance is weakly basic and amorphous, but possesses strongly toxicproperties.It was shown by Wormley (Zoc. cit.) that gelsemine was accompaniedin the plant by an acidic substance, which he called I‘ gelseminic acid,”an observation which has been confirmed by the present author.Gelseminic or “gelsemic ” acid has been shown by Schmidt (Arcii.Phurm., 1898, 236, 236) to be a monomethyl ether of t-esculetin(4 : 5-dihydroxycoumarin). Two sesculetin monomethyl ethers areknown, which have been incorrectly termed a- and P-methylmculetinrespectively (compare Beilstein’s Handbuch, III., 5 68), the compoundfrom gelsemiurn having been given by Schmidt the latter designation.It is evident, however, that the names a- a.nd P-methylsesculetin canonly be correctly applied t o substances possessing the following formulsrespectively (Pechmann and Kraft, Ber., 1901, 34, 423) :CH C*CH,Gelseminic acid is, therefore, xsculetin 4-(or-5)monomethyl ether,and it is considered desirable to retain for this substance the name‘( scopoletin,” as proposed by Eykman (Rac.trccv. chirn., 1884, 3, 171),who first obtained it from the rhizome of Xcopolicc japonica. Thefluorescent substance, known as P-methyl~esculetin, which is containedin the bark of Prztnus serotina and in jalap (Trans,, 1909, 95, 243;J. Amer. Chem. ~ o c . , 1910, 32, 93) would accordingly be moreappropriately termed scopoletin.A summary of the results of the complete investigation of gelsemiurn,is given a t the end of this paper.E XPE R I MEN TAL.The material employed in this investigation consisted of the driedrhizome and roots of Gelsemiurn senapemipcens, Aiton.A portion (20 grams) of the crushed drug was extracted successivelyin a Soxhlet apparatus with various solvents, when the followingamounts of extract, dried a t looo, were obtained :Petroleum (b.p. 35-50”) extracted 0.39 gram = 1’95 per cent.Ether ,, 0.16 y y 0.80 ,)Chloroform ), 0.34 ,, 1-70 y yEthyl acetate ,, 0.16 ,, 0’80 ,)Alcohol ?, 1’63 $ 7 8.15 ,,- -Total ,........... 2‘68 grams=13.4MOORE : THE COXSTITVENTS OF OELSEMIUM. 2225For the purpose of a complete examination, 49.44 kilograms of theground material were completely extracted with hot alcohol, Afterthe removal of the greater portion of the alcohol, a viscid, dark-coloured extract was obtained, amounting to 9.20 kilograms.Distillation of the Extract with Steam.A quantity (2 kilograms) of the above-mentioned extract, represent-ing about 10.75 kilograms of the drug, was mixed with water, andsteam passed through the mixture for some hours.The distillate,which amounted to 5 litres, contained some drops of oil floating on thesurface. It was extracted with ether, the ethereal liquid being driedand the solvent removed, when a small quantity of an essential oil masobtained, This was a very pale yellow liquid, and amounted to about2 grams, being thus equivalent to about 0.019 per cent.of the weightof the drug.Non-volatile Constituents of the Extract.After the distillation of the extract with steam, as described above,there remained in the distillation flask a quantity of a brown resin(A) and a dark-coloured aqueous liquid (B). The resin was collected,and repeatedly washed with water until nothing further was removed,the washings being added to the a bove-mentioned aqueous liquid.This resin was a brown, viscid solid, and amounted to 412 grams. Itwas dissolved in alcohol and mixed with purified sawdust, thethoroughly dried mixture being then successively extracted in aSoxhlet apparatus with petroleum (b. p. 35-50"}, ether, chloroform,ethyl acetate, and alcohol.Petroleum Extract of the Resin (A).Isokation of Pentatriacontane, C35K72, and Emodim Monomethyl Ether.The petroleum extract, which formed a brown, semi-liquid mass andamounted to 224 grams, was dissolved in 2 litres of warm ether andthe solution kept for some days, when a small quantity of an almostcolourless substance separated, This was collected and washed with alittle ether, after which it was dihtilled under diminished pressure.The distillate, which rapidly solidified, was crystallised from ethylacetate, when it was obtained in small, colourless, glistening leaflets,melting at 75'.(Found, C = 84.9 ; H = 14.5. Calc., C = 85.4 ; H = 14.6per cent.)This substance was therefore pentatriacontane.The ethereal liquid, from which the pentatriacontane had beenremoved as above described, was extracted with successive portions ofan aqueous solution of sodium carbonate, and finally washed wit2226 MOORE : THE CONSTITUENTS OF GELSEMLUM.water. The alkaline liquids and washings were united, acidified, andextracted with ether, when 15 grams of a viscid, oily liquid wereobtained.On distilling this liquid under diminished pressure, i tpassed over between 245Oand 255"/25 mm., and then became almostsolid. It consisted of a mixture of fatty acids, which were examinedin connexion with a similar product obtained from the non-acidicportion of. the petroleum extract after its hydrolysis.The ethereal liquid, from which the pentatriacontane and free fattyacids had been removed, as above described, was subsequently shakenwith a solution of sodium hydroxide.The alkaline extracts, whichhad assumed a red colour, were acidified and extracted with ether, whena very small quantity of an orange-yellow substance was obtained.This when crystallised from ethyl acetate formed orange-red prisms,which melted a t about 190°, and when mixed with a little ernodinmonomethyl ether, fusion occurred at the same temperature. Thequantity so obtained was too small for analysis, but the substanceappeared to be emodin monomethyl ether (m. p. 195O), since on heatingfor a short time with concentrated sulphuric acid it gave a substancesoluble in aqueous sodium carbonate and agreeing in its propertieswith emodin.Isolation of GL Phytosterol, C,lH,,O.The ethereal liquid which had been extracted with alkalis, as abovedepcribed, was evaporated, when a quantity of an oily product wasobtained. This was hydrolysed by heating with an alcoholic solutionof potassium hydroxide, the alcohol removed, water added, and thealkaline liquid extracted with ether.The ethereal solution waswashed, dried, and the solvent removed, when a quantity of brownresinous material was obtained. This was extracted with cold absolutealcohol, in which only a small portion dissolved. The alcoholic solu-tion was concentrated, and a little water added, when, on keeping, asubstance separated in flat needles, which after recrystallisation froma mixture of dilute alcohol and ethyl acetate formed glistening,flat needles, melting at 136".The amount of this substance was1.5 grams :0.1600, on heating at 1 loo, lost 0.0072 H,O.0*1336* gave 0.4110 CO, and 0;1455 H,O.H,O = 4.5.C = 83.9 ; H = 12.1.C,7H,60,H,0 requires H,O = 4.5 per cent.C,7H,60 requires C = 83.9 ; H = 11.9 per cent.This subst.ance thus agrees in composition with a phytosterol, and i tyielded the colour reaction of that class of compounds. A determinationof its rotatory power gave the following result :* Anliydrous substanceMOORE : TEE CONSTITIJENTS OF GELSEAHUM. 22270.2393, made up t o 20 C.C. with chloroform, gave aD -0'58' in a2-dcm. tube, whence [a]D -40.4'.The acetyl derivative, when crystallised from acetic anhydride,separated in needles melting at 125-127'.The brown resinous material, from which the phytosterol had beenremoved by extraction with alcohol, as above described, was thoroughlyexamined, but nothing definite could be isolated from it.I t appearedto consist of a mixture of hydrocarbons.IdentiJication of t?k Fatty Acids.The alkaline aqueous solution of potassium salts, from which thephytosterol had been removed by extraction with ether, as abovedescribed, was acidified and again extracted with ether, the etherealsolution being washed, dried, and the solvent removed. A quantity(10 grams) of fatty acids was thus obtained, which, when distilledunder diminished pressure, passed over between 240" and 260'/25 mm.As these acids distilled within the same range of temperature as thosepreviously obtained, which existed in the drug in the free state, for thepurpose of their examination the two portions were mixed.Twenty grams of the mixed acids were converted into their Ieadsalts, and the latter digested with ether, when a portion dissolved.Both the soluble and insoluble portions were decomposed by hydro-chloric acid, and the regenerated fatty acids purified by distillationunder diminished pressure. The soluble portion of the lead saltsyielded 11 grams of liquid acids, while the insoluble portion gave8 grams of solid acids.The Liquid Acids.-These acids, when distilled under diminishedpressure, passed over at about 225'/15 mm.An analysis and adetermination of the iodine value gave the following results :0.1430 gave 0,4030 GO, and 0.1518 H,O.0.4224 absorbed 0,6783 iodine.C= 76.8 ; H = 11.8.Iodine value = 160.C18H3402 requires C = 76.6 ; H = 12.1 per cent.Iodine value = 90.1.C16H3202 ,, C=77*1 ; H.= 11.4 ,, ,) ,) =181.4.In order to obtain more definite information respecting the natureof the above mixture, a quantity of it was oxidised according to themethod described by Lewkowitsch (Chemiccd Technology and Analysisof Oils, Fats, and Waxes, 1904, Vol. I,, p. 360). This resulted in theformation of tetrahydroxystearic acid (m. p. 157-160') and a smallquantity of dihydroxystearic acid (m. p. 125-127'). It may thus beconcluded that the liquid acids consisted chiefly of a mixture of oleicand linolic acids, the latter in predominating amount.The Solid Acids.-These acids melted at about 55', and on analysisgave the following result 2228 MOORE : THE CONSTITUENTS OF GELSEMIUM.0.1383 gave 0,3842 CO, and 0.1590 H20.C = 75.8 ; H = 12.7.ClsH,,02 requires C = 75.0 ; H = 12.5 per cent.C18H3602 ,, C== 76.1 ; H= 12.7 ,,From this result it would appear that the solid acids consisted of amixture of palmitic and stearic acids, the latter predominating.Ethereal ExtructIof the Resin.Ierolation of Ipurunol, C23H3802(OH)2.This extract was a brown, amorphous mass, and amounted to10 grams. It was redissolved in about 500 C.C. of warm ether andkept for some days, when a small quantity of an almost colourless,amorphous substance separated. This was collected and crystallisedfrom a mixture of pyridine and dilute alcohol, when it formedmicroscopic needles, melting a t 290'.(Found, C = 72.3 ; H = 10.5.Calc., C = 72.6 ; H = 10.5 per cent.)This substance was thus identified as ipuranol, and when treatedwith sulphuric acid and acetic anhydride it yielded the colour reactionshown by this coqpound. From it was also prepared diacetylipuranol,which separated from acetic anhydride in glistening leaflets, meltingat 162'.The ethereal solution from which the ipuranol had been separated,as above described, was examined, but nothing definite was isolatedfrom it.The chloroform, ethyl acetate, and alcohol extracts of the resinamounted to 35, 36, and 95 grams respectively, and consisted entirelyof amorphous products.Examination of the Aqueous Liquid (B).Isolation of Scopoletin.This liquid, as already indicated, represented that portion of theoriginal alcoholic extract of the drug which was soluble in cold water,and from which the previously-described resin (A) had been removed.It was thoroughly extracted with chloroform, these extracts beingwashed, dried, and the solvent removed.A quantity (about 5 grams)of a crystalline compound was thus obtained, which, after recrystallisa-tion from alcohol, formed long, almost colourless needles, melting at204'. Its alkaline solution showed a fine blue fluorescence,0.1430 gave 0.3286 CO, and 0.0550 H20. C = 62.6 ; H = 4.2.CIoH,O, requires C = 62.5 j H = 4.2 per centMOORE : TEE CONSTITUENTS OF GELSEMIUM. 2229A methoxyl determination by means of Perkin's modification of the0.2132 gave 0.2584 AgT.The substance is thus identi6ed as scopoletin, a methyl ether ofasculetin.Its acetyl derivative separates from acetic anhydride in colourlessleaflets, melting at 177".Dibromoscopoletin, C,,R,O,Br,.-Five grams (six atoms) of brominewere added t o a solution of scopoletin (2 grams) in about 50 C.C.ofchloroform. Hydrogen bromide was slowly evolved, but the liquid did notbecome colourless. After keeping some hours, a crystalline substanceseparated, which was removed and recrystallised from alcohol, when i tformed yellow, glistening plates, melting a t 249' :Zeisel method gave the following result :OMe= 16.0.C,HBO,*OMe requires OMe = 16*1 per cent.0,1682 gave 0.1800 AgBr. Br=45.5.This substance is therefore a dibromoscopoletin.Dibromoscopoletin is sparingly soluble in ether, chloroform, oralcohol, and its solution in alkalis shows a very intense green fluores-cence.The two bromine atoms in dibromoscopoletin appear to be in thebenzene nucleus, as this Substance instantly decolorises a cold alkalinesolution of potassium permanganate, and, therefore, still contains adouble linking.I n this respect it resembles the dibromocoumarindescribed by Perkin (Trans., 1870, 23, 371).On heating dibromoscopoletin with acetic anhydride, it is readilyacetylated. The acetyl derivative forms colourless prisms, meltinga t 224O.Isolation of Gelsemine, C2-,H2202NpC,,H,0,Br2 requires Br = 45.7 per cent.The aqueous liquid from which the scopoletin had been removed, asabove deFcribed, was extracted with successive portions of amylalcohol. This, however, only removed small quantities of an amorphousnitrogenous product, which was non-basic, and from which nothingdefinite could be isolated.The liquid was accordingly renderedalkaline with sodium carbonate and thoroughly extracted with ether,the combined ethereal extracts being washed, dried, and the solventremoved. A quantity of a pale yellow product was thus obtained,which crystallised very readily from acetone in handsome, glis teniogprisms, melting a t 175-178? After recrystallisation from the samesolvent, its melting point was constant a t 1'78'. The quantity isolatedamounted to 12 grams. It gave all the usual reactions characteristicof alkaloids 2230 MOORE : THE CONSTITURNTS OP GELSEMIOM.1.1448, when heated at 120°, lost 0.1774 acetone.0.1594" gave 0.4353 CO, and 0.0980 H,O.0*345S* ,, 27*5 C.C.N, at 27' and 764 mm. N=S*7.C3H60= 15.5.C = 74.5 ; H = 6.8.C,,H,,O,N, requires C = 74.5 ; H = 6.8 ; N = 8.7 per cent.C,,H,,0,N,,C3H60 requires C,H,O = 15.3 per cent.This substance, therefore, corresponds with the crystalline alkaloid,gelsemine, which has previously been isolated from gelsemium, and forwhich, as already mentioned, several empirical formula? have beensuggested, The fact that gelsemino crystallises from acetone with onemolecule of this solvent (see above) was con6rmed by mixing 1 gramof the air-dried preparation with 20 C.C. of water and distilling theliquid. On adding p-bromophenylhydrazine t o the distillate, a crystal-line precipitate was formed, melting a t 03', which corresponded in allrespects with acetone-p-bromophenglhydrazono.The molecular weight of gelsemine was determined by the cryoscopicmethod in acetic acid solution :0*5250*, in 24.90 acetic acid, gave At = - 0*270°.C,,H2,0,N, requires M.W. = 322.In benzene solution association occurs, and numbers corresponding0*6340*, in 20.70 benzene, gave At= - 0*248O.(C2,H,,02N2), requires M.W. = 644.In order to ascertain whether gelsemine is homogeneous, a quantitywas converted into its hydrochloride, and this salt recrystal-lised, first from dilute alcohol and then from water. The base wasthen regenerated, and, after crystallisation from acetone, againanalysed :M.W.= 305.with twice this molecular weight are obtained :M.W. = 605.0.1414" gave 0.3866 CO, and 0.0880 H,O.For further confirmation of the purity of the material, the base wasconverted into its nitrate. This salt, which forms glistening prisms,melting above 280", was recrystallised from water, and .the baseregenerated from it. The product so obtained, after crystallisationfrom acetone, gave the following results on analysis :C= 74.5; H=6*9.C,,H2,0,N2 requires C = 74.5 ; H = 6.8 per cent.0.1462" gave 0.3980 GO, and 0.0906 H,O.The formula of the base deduced from these analyses is in harmonyGelsemine forms a monohydrochloride cry stallising in small0.5614 gave 0*2310 AgCI.C = 74.2 ; H= 6%C,,H,,O,N, requires C = 74.5 ; H = 6.8 per cent.with the result ohtained from the analysis of the hydrochloride.prisms, melting indefinitely at about 300° :C1= 10.1.C,,H,,O,N,,HCl requires C1- 9.9 per cent,* Constant at 120"MOORE : THE CONSTITUENTS OF CIELSEMIUM.2231A determination of its specific rotatory power gave the following0.3100, made up to 20 C.C. with water, gave [.ID + 0'5' in a 2-dcmThe close agreement of these results shows conclusively that theA determination of its specific rotatory power gave the following0.4066*, made up to 20 C.C. with chloroform, gave [a]D +Oo39' in aresult :tube, whence [a], + 2.6'.empirical formula of gelsemine is C,,H,,O,N,.result :2-dcm. tube, whence [a]. + 15.9'.Examination of the Amorphous Alkaloidal Products.The alkaline, aqueous liquid from which the gelsemine had beenremoved by extraction with ether, as above described, was repeatedlyextracted by means of amyl alcohol, when a relatively small quantityof an amorphous, basic product was obtained.This appeared toconsist of a mixture, and two alkaloidal products were found to bepresent, one of which was much more strongly basic than the other.It was dissolved in chloroform, and extracted several tinies with1 per cent. aqueous hydrochloric acid, which removed the more stronglybasic product. The material obtained on rendering the acid extractsalkaline was isolated by means of chloroform, when it formed anamorphous, brown-coloured product. Neither the free base nor any ofits salts could be obtained in a crystalline condition.This morestrongly basic product appears to correspond with the amorphousalkaloid to which the name '( gelseminine " has been given.The chloroform solution from which the '( gelseminine " had beenremoved by means of 1 per cent. acid, as above described, was shakenmany times with 10 per cent. aqueous sulphuric acid, which slowlyremoved a small quantity of (L. very weakly basic substance. As inthe case of (' gelseminine," neither the free base nor its salts could beobtained in a crystalline condition. This substance responds to theusual alkaloid reagents, but appears to be stable only in the form of itssalts, as on keeping a chloroform solution of the base for some timethe product becomes insoluble in acids.The alkaline aqueous liquid from which the alkaloidal products hadbeen removed, as above described, was neutralised by means of aceticacid and treated with a solution of basic lead acetate.This produceda voluminous yellow precipitate, which was collected, washed, andthen suspended in water and decomposed by hydrogen sulphide. Onfiltering the mixture, a liquid was obtained which gave a bluish-black* Constant at 120"2232 MOORE : THE CONSTITUENTS OF GELSEMIUM,coloration with ferric chloride, and evidently contained a quantityof tannin, but no definite products could be isolated from it.The filtrate from the basic lead acetate precipitate was treated withhydrogen sulphide for the removal of the excess of lead, and thefiltered liquid concentrated under diminished pressure to a volume ofabout 2 litres. The Concentrated liquid contained a considerablequantity of a sngar, as i t readily reduced Fehling's solution, andyielded d phenylglucosazone, melting at 808-21 0'.One-fifth of the total liquid was diluted with water to 1 litre, about60 grams of concentrated sulphuric acid, diluted with a n equal weightof water, added, and the liquid repeatedly extracted with chloroformwith the object of isolating any organic acids present.As thisoperation removed only a small quantity of acetic acid, the acidaqueous liquid was boiled for an hour and again extracted with chloro;form, when nearly a gram of scopoletin was obtained. It thus appearsprobable that a glucoside of scopoletin was present in the originalaqueous liquid, but all attempts t o isolate this substance wereunauccessf ul.Pl~ysiological Teats.The following physiological tests were conducted in the WellcomePhysiological Research Laboratories by Dr.H, H. Dale, t o whom theauthor now wishes to express his thanks :A quantity (0.1 gram) of gelsemine hydrochloride, when injectedintravenously into a rabbit, caused practically no effect, a resultwhich is in agreement with an observation by Cushny.One milligram of the hydrochlorides of both the amorphous bases,when injected intravenously into rabbits, caused death from respiratoryfailure in about twenty-five minutes, preceded by convulsions.fluntmnry .The results of this investigation may be summarised as follows :The material employed consisted of the dried rhizome ,and roots ofGelsemiurn sempervirena, Ai ton.An alcoholic extract of the drug, when distilled with steam, yieldeda small amount of an essential oil. The nonLvolatile constituents, asobtained after treating the alcoholic extract with steam, consisted of abrown resin insoluble in water, and material which remained dissolvedin the cold aqueous liquid. The resin, amounting to about 3.8per cent. of the weight of the drug, yielded pentatriacontane ; traces ofemodin monomethyl ether; a phytosterol, C,7H,,0 (m. p. 136';[a], -40.4'); a small amount of ipuranol, C,,H,,O,(OH),; and amixture of fatty acids, consisting of palmitic, stearic, oleic, and linolicacids. The portion of the alcoholic extract of the drug which waEVANS: THE DISTILLATION OF MIXTURES, ETC. 2233soluble in water, and from which the above-described resin had beenremoved, contained scopoletin (admonomethyl ether of asculetin), whichwas present in the free state, and also in the form of a glucoside,together with a quantity ,of sugar, It yielded, furthermore, threealkaloidal products, one of which, gelsemine, has been obtained ina pure crystalline state, melting considerably higher than has hithertobeen recorded (1789 instead of 160°), and which has been conclusivelyshown to possess the formula C2,H,,0,N,. The other alkaloidalproducts, one of which corresponds with the so-called ‘( gelseminine ” ofThompson (Zoc. cit.) and Cushny (Zoc. cit.), were amorphous, andno crystalline derivative could be obtained from them.THE WELLCOME CHEMICAL RESEARUH LABORATORIES,LONDON