PHYSOSTIGMINE (ESERTNE). PART m. 247 XLI1.-Ph ysostigmine (Eserine). Part I I I . By EDGAR STEDXAN and GEORGE BARGER. CERTAIN details of the structure of physostigmine have been elucidated (J. 1923 123 758; 1924 125 1373) but the evidence which was available was insufficient to permit the construction o 248 STEDMAN AND BBRGER: a formula which satisfactorily represented the known facts con-cerning the chemistry of this alkaloid. Further experimental evidence is now advanced and this combined with previously published work appears to leave no doubt as to the structure of eseret’hole and hence of physostigmine which are respectively the ethyl ether and methylcarbamido-derivative of eseroline. The presence in eseroline of the grouping I has been recognised since the work of Straus (AnnuZen 1913 401 350; 1914 406, 332) who obtained a phenolic indole compound (physostigmol) by degradation of eseroline methiodide.It was shown by one of us (E. S. ; Zoc. cit. 1373) that physostigmol is 5-hydroxy-1 3-dimethyl-indole (11) and further that its ethyl ether may be obtained in a yield of 66% by distillation of eserethole methiodide. This estab-lished the position of the hydroxyl group in eseroline and also confirmed Straus’s supposition that physostigmol contains a methyl group attached to position 3 of the indole ring. The formation of an indole compound by the somewhat violent decompositicn of an alkaloid does not in general permit the conclusion to be drawn that a preformed indole skeleton exists in the latter. Nevertheless, it appears justifiable in view of the comparat,ive completeness with which eserethole methiodide is degraded into physostigmol ethyl ether to assume even without further evidence the presence of the grouping I11 in eseroline.Additional evidence pointing to the presence of an indoline grouping in the molecule is not however, wanting. Thus the feebly basic properties of one of the nitrogen atoms in physostigmine (Straus Zoc. cit.) and the properties of etheserolene obtained by Max and Michel Polonovski (BUZZ. Xoc. chim. 1918 [iv] 23 335; 1923 [iv] 33 969) by the exhaustive methylation of eserethole point in this direction. The presence of the grouping I11 may thus be regarded as estab-lished and the problem resolves itself into determining the manner in which the remainder of the molecule (C,H,N) is linked to this grouping The nitrogen atom the position of which has still to be determined is the one with the stronger basic properties.Straus has shown that one methyl group is attached to this nitrogen atom whilst Salway (J. 1912 101 978) had previously demon-strated its tertiary character. Formula IV which was suggested to us by Professor R. Robinson F.R.S. conforms to these con-ditions and evidence will be adduced in the following discussio PHYSOSTIGMINE ( ESERINE). PART rn. 249 to show that this actually represents the structure of eseroline. The constitution of physostigmine itself will accordingly be repre-sented by formula V. Me CM Me CH, HOf>-?/\$!H I NHMe*CO*ON\ &/\?€I2 (v. ’ “-t!H -l<AIe \/\,,4H--NMe \/\/ (IV.1 NMe NMe The considerations which have led to the adoption of formula I V for eseroline are largely based on the behaviour of eserethole towards reducing agents and on the properties of eseretholemethine. Polo-novski (Bull. Xoc. chim. 1918 [iv] 23 357) has shown that esere-thole on reduction with zinc and hydrochloric acid takes up two atoms of hydrogen a result which has now been confirmed using a catalytic method of reduction. That this reduction is not due to the presence of a double bond follows from the fact first pointed out by Max and Rfichel Polonovski (Compt. rend. 1024 178 2078), that whereas eserethole is a tertiary base dihydroeserethole is a secondary base. Reduction evidently opens a ring between the more strongly basic nitrogen atom and one of its adjacent carbon atoms.On the basis of IVY dihydroeserethole may thus be repre-sented by 171 the point a t which tjhe ring is ruptured being deter-Me Me I -$XH,*CH,*N,\le, NMe (T;II.) \ /\,/CJH.OH Etoli Et 0’ \--~-CH,*CH,*NH?IIe I i II \ / \ p H 2 (171.) NXe mined from Polonovski’s observation that the substance obtained by reduction of etheserolcne is identical with the product of the exhaustive methylation of dihydroeserethole. Eseretholemethine was ubtained by Max and Michel Polonovski (Zoc. cit.) by treating eserethole methiodide with sodium hydroxide. Its pseudo-basic character was not recognised until in Part I cf this series attention was directed to the fact that it is reconvertcd into eserethole rnethiodide on treatment with hydriodic acid m d that its behaviour in this respect was analogous to that of certain indoline compounds prepared by Brunner.As a result of this observation Polonovski revised the molecular formula of eserethole-methine and denionstrated that it was produced from methyl-eseretholinium hydroxide by a tautomeric change and not by loss of water. Experimental results which are enumerated in the succeeding paragraph have now been obtained which demonstrate that eseretholemethine is in fact a substituted indoline nit1 250 STEDMAN AND BARGER: hydroxyl group in the a-position (VII); its formation may accord-ingly be represented by the following scheme : tautomeric -CMe.CH 2 2 -CH *NMe CH2 4 /\ NaOH /\ change I -?Me yH2 -+ -$!Me $!H2 -+ -CH*OH -CH-NMe21 -CH-NMe,*OH Eserethole methiodide Intermediate Eseretholemethine quaternary hydroxide The position assigned to the basic nitrogen atom in V thus corre-sponds with this behaviour.Assuming the correctness of this scheme it is evident that while the conversion of eseretholemethine into a quaternary salt by treatment with methyl iodide should prevent the closure of the pyrrolidine ring on treatment with acids the compound should nevertheless retain its properties as a pseudo-base in virtue of its indolinol structure. This consequence has been verified. Treat-ment of eseretholemethine methiodide with picric acid results in the elimination of the hydroxyl group with the formation of a diquaternary picrate : \G-~Me*CH,*CH,=NMe,l \G-~Me*CH,°CH2*NMe3*OoC6H,06~3 ,,C CH*OH ,C CH ’ \/ ++ \/ /N\ NMe Me O*C6H,06N3 Still more conclusive are the results obtained by oxidation of eseretholemethine.Brunner (Nonatsh. 1896 17 253) has shown that 1 3 3-trimethyl-2-indolinol is oxidised by ammoniacal silver nitrate in alcoholic solution to 1 3 3-trimethyl-2-indolinone. By subjecting eseretholemethine to the same treatment a compound the composition of which corresponds with VIII has been obtained. The same substance has also been prepared by oxidising eserethole-methine with potassium ferricyanide a method which Decker ( J . pr. Chem. 1893 47 28) has employed for the oxidation of pseudo-bases. When this compound is degraded by the method of Et 0’ \-~MeCH2=CH,*NMe2 Et ON\-FMe-CH:CH2 I II \ \ / \ P O < II NMe W.1 \ \ / \ P O (VIII.) NMe Etd()-?MeEt (X.1 NMe \ / \ P PHYSOSTIGMINE ( ESERINE). PART III. 251 exhaustive methylation trimethylamine is evolved and an un-saturated compound (IX) obtained which is practically devoid of basic properties and forms a deep crimson picrate. On reduction, this takes up two atoms of hydrogen with the formation of a substance which is considered to be 5-ethoxy-1 3-dimethyl-3-ethyl-2-indolinone (X). The chemical properties of eserethole-methine thus correspond entirely with those predicted on the assumption that V correctly represents the structure of physos-tiginine ; the constitution of this alkaloid may therefore be regarded as established. As an additional confirmation the synthesis of the substance represented by X has been undertaken and it is hoped that it will be possible to communicate the results shortly.An explanation of the mcchanism of the formation of eserethole-methine and of dihydroeserethole identical with that advanced above has recently been put forward by Max and Michel Polonovski (Compt. rend. 1924 178 2078). These authors therefore consider that the structures of eserethole and eseretholemethine are repre-sented by XI and XII respectively; the former differs from that proposed by Straus (Zoc. cit.) for eseroline only in the position of the nitrogen atom contained in the piperidine ring. Although the reactions described in this communication could be formulated with equal facility upon the basis of these structures it is evident that they do not explain the formation of physostigmol; the ease with which this substance is formed demands the presence of a preformed methyl group in the P-position of the pyrrolinc ring.CH, Consideration of the possible mechanism of the phytochemical synthesis of the ring system present in physostigmine speaks no less convincingly in favour of formula V. Perkin and Robinson (J. 1919 115 944) have shown how harmine may be elaborated in the plant from tryptophan. If the assumption is made that the methylation of an indole nucleus may proceed in the plant in the manner in which it is known to take place in the laboratory a relation between physostigmine and this amino-acid at once becomes evident. By decarbosylation and methylation followed by a, ring closure the ring system of physostigmine would be readily formed 252 STEDMAN AND BAROER: (NH,)*CO,H decarboxylation f)-G*CH2*CH2*NH, -5- -NH / NH Me \/\/CH-fiMe NMe The possible mechanism of the phytochemical synthesis of the ring system present in XI cannot be similarly represented in a simple manner from known naturally-occurring substances.During the course of this investigation a small quantity of etheserolene was prepared. Contrary to Polonovski’s statement, it readily formed a yellow crystalline picrate. If etheserolene is formed by a straightforward degradation of eseretholemethine in a manner similar to that of the indolinone compound now described, one would feel inclined to attribute to it the structure represented by XIII. On this basis however it should possess the properties of a pseudo-base.But analysis of the picrate shows that this is not the case. It therefore appears that some isomeric change has taken place during its formation. This would correspond with the results of Polonovski who has recently stated (Cmpt. rend. 1924, 179 178) that etheserolene on reduction absorbs only two atoms of hydrogen and at the same time loses an atom of oxygen. Whether his view that this oxygen atom is not present as a hydroxyl group but forms part of a ring is correct cannot be stated with certainty a t present, Trinitroeserethole (XIV) and an oxidation product of eserethole which have been prepared during the course of this investigation and the results of a reinvestigation of the degradation of one of th PHYSOSTIGMINE (ESERINE).PART III. 253 esoline compounds described in Part I of this series are described in the experimental portion. E x P E R I M E N T A L. Dihydroeserethole (VI) .-Attempts to reduce eserethole catalytic-ally with a palladium sol in either neutral or acid solution with or without the addition of gum arabic as protective colloid werc unsuccessful owing to flocculation of the sol. With platinum black prepared by Willstatter's method (Ber, 1921 54 113), reduction took place readily. Platinum black (0.1 g.) was washed into a hydrogenation apparatus similar to that described by Hess (Ber. 1913 46 3113) air was completely removed a solution of 1 g. of eserethole in glacial acetic acid introduced and the mixture shaken; 52 C.C. of hydrogen were absorbed during the first hour, and 92 C.C.in all (calculated for two atoms 91 c.c.). The solution was filtered made alkaline with sodium hydroxide and the oil thus thrown down extracted with ether. From the extract dried over potassium carbonate a yellow oil was obtained which became almost colourless when distilled under the vacuum of a rotatory oil pump (bath at about 200°). When treated with alcoholic oxalic acid it formed an oxalate (sheaves of fine needles m. p. 204" from aqueous alcohol) identical with that obtained from dihydro-eserethole prepared by Polonovski's method (Zoc. cit.). Dihydroeserethole (1 g.) was heated in a sealed tube with excess of methyl iodide for 1 hour a t loo" the excess of methyl iodide evaporated and an alcoholic solution of the product treated with alcoholic picric acid.The yellow picrate precipitated on addition of water crystallised from aqueous alcohol in prisms m. p. 204" (Found C = 47.0 ; H = 4.7. C2,H,,Q1,N8 requires C = 47.3 ; H = 4.6%). The substance was thus a dipicrate and two methyl groups had been introduced into the molecule during methylation, indicating that in contrast Go eserethole dihydroeserethole is a secondary base. Behaviour of' Eseretholemethine Methiodide as a $-Base.-The orange oil which separated on mixing the salt (0-2 g.) and picric acid in alcoholic solution changed on warming into a crystalline solid (0.27 g.) which was obtained from aqueous alcohol in yellow prisms m. p. 170" (Pound C = 47.7; H = 4.7. C,,H,2015N,s requires C = 47.5; H = 4.4%). In the formation of this di-quaternary picrate water has been eliminated from the molecule, t l t u s inclienting the pscudo-basic function of the oxygen atom.L)eh~/clroeserethoZernethilze (VIII) .-The methine (1 g.) was refluxed for 10 hours with an ammoniacal solution of 3 g. of silver nitrate in methyl alcohol. After filtration evaporation and dissolutio 254 STEDMAN AND BARGER: of the residue in water ether extracted an insignificant amount of an oil which formed a crystalline picrate but was not investigated further. The hot aqueous solution made faintly alkaline with ammonia, was treated with hydrogen sulphide to remove excess of silver atered boiled to expel ammonia and hydrogen sulphide, and treated with alcoholic picric acid. The viscous precipitate first formed became crystalline on warming and by fractional crystallisation from aqueous alcohol gave two picrates the more soluble one forming needles m.p. 211" (Found C = 47.0; H = 4.5. Cl6H2,ON2,2CBH,0,N requires C = 46.7; H = 4.4%) and the less soluble fraction consisting of plates contaminated with some needles. The latter crystallised from acetone in plates m. p. 199" (Found C = 52.8 ; H = 5.4. CI6H2,O2N2,C6H,0,N requires C = 52-3; H = 5.4%). A different result was obtained by the following procedure An alcoholic solution of eseretholemethine (5 g.) was added to 15 g. of silver nitrate in alcohol made strongly alkaline with ammonia (d 0.88), the total volume being brought to about 100 C.C. A silver mirror formed on the sides of the flask almost immediately.The mixture was heated on the water-bath for 5 hours filtered and excess of silver removed by addition of hydrochloric acid; from the filtrate, made alkaline with sodium hydroxide ether extracted 4.45 g. of an oil which treated with alcoholic picric acid yielded 5.35 g. of the picrate m. p. 199" described above. No trace of the picrate of higher melting point was obtained in this experiment. It seems evident that this substance was formed by reduction of an indolinone to an indoline compound the reducing agent being the hydrogen sulphide used to remove excess of silver. Oxidation was also effected by means of potassium ferricyanide. A solution of eseretholemethine (3 9.) in a small quantity of alcohol was boiled with a strongly alkaline solution of 7-2 g.of potassium ferricyanide in 100 C.C. of water for about 5 minutes and after cooling extracted with ether. From this extract 4 g . of the methiodide which had previously been prepared in the manner described below were obtained. The picrate (4.19 g.) m. p. 199" was treated with sodium hydroxide and the oily base which separated extracted with ether. The extract was dried over sodium sulphate filtered concentrated, and methyl iodide added. A crystalline methiodide slowly formed (3.3 g.). This separated from acetone in transparent prisms but the method was wasteful owing to its large solubility in this solvent. It was therefore crystallised by addition of dry ether to the acetone solution. With dry solvents the methiodide was obtained as a colourless substance m.p. about 131"; the fact that the meltin PHYSOETIGMINE ( ESERINE). PART III. 256 point was not absolutely sharp was no doubt due to the water of crystallisation which it contained (Found for air-dried material, H,O = 3.7 ; for material dried at 110" I = 30.0. C1,HZ70,N,I,H,O requires H,O = 4.1 ; C1,H,,O,N,I requires I = 30.1%). When, however the solvents were not perfectly dry the transitory form-ation of a blue iodine adsorption compound was consistently observed ; its disappearance no doubt coincided with the trans-formation of an a t first amorphous precipitate into crystalline form. Degradation of Deh ydroeseretholemethine i!fethiodide by Hof I)Z an n ' s Nethod to Compound IX.-An aqueous solution of 1.9 g. of the methiodide was shaken with a suspension of silver oxide prepared from 2 g.of silver nitrate. After filtration and removal of the water by evaporation under diminished pressure the residual oil was distilled a t 12 mm. After evident decomposition with the evolution of a gas a slightly brown oil distilled when the tem-perature of the bath was between 200" and 210". The distillate would not crystallise. It was therefore treated with an alcoholic solution of picric acid when a small quantity of a crystalline picrate m. p. 199" identical with that described above was obtained. The residue from the mother-liquors was treated with sodium hydroxide and extracted with ether and the extract washed with water until the sodium picrate was completely removed. After drying over sodium sulphate the ether was evaporated, when a colourless oil which crystallised spontaneously was obtained.Recrystallised from aqueous alcohol it formed colourless prisms, m. p. 62" (Found C = 72.3; H = 7.3. C14H1702N requires C = 72-7; H = 7-4:/,). This substance was practically devoid of basic properties; it was insoluble in dilute hydrochloric acid but formed a picrate which at first separated from alcohol in a yellow form and quickly changed into a deep crimson one; m. p. 103". Reduction of Compound I X to Compound X.-Hydrogenation was effected in the apparatus mentioned above using a solutim of colloidal palladium as catalyst and gum arabic as protective colloid; 0.16 g. of substance was used. Hydrogen was readily absorbed but as a result of a mishap the amount could not be measured.After filtration from the palladium and evaporation of the alcohol (aqueous alcohol was used as solvent) the solution, after addition of a small quantity of sodium hydroxide to ensure alkalinity was extracted with ether. After drying and evaporation, this yielded a crystalline product which when crystallised from aqueous alcohol formed cubes m. p. 68" (Fcund C = 71.8 H = 8-2. It is evideat from the analysis that two atoms of hydrogen were absorbed and C1,H1,O,N requires C = 72.1 ; H = S-2:/,) 256 STEDMU AND BARGER: that the oxygen atom present as a carbonyl group was retained in the molecule. Trinitroeserethote.-Among the diverse experiments which have been carried out with a view to effect the oxidation of eserethole, the action of nitric acid on this substance has been studied.Cold dilute nitric acid has little if any action but the concentrated acid (d 1-4) reacts vigorously. Eserethole (1 g.) was added drop by drop to 3 C.C. of concentrated nitric acid cooled in a freezing mixture and vigorously stirred. The product was warmed on the water-bath for 8 hour cooled diluted with waber and the yellow crystalline precipitate (0.3 g.) recrystallised from aqueous alcohol, trinitroeserethole separating in orange rectangular plates m. p. 152" (Found C = 47.1 ; H = 5.0; N = 18.7. C1,H1,O,N requires C = 47-2; H = 5.0; N = 18.4%). It is insoluble in alkalis and dilute acids but dissolves in concentrated acids. Oxidation of Eserethole with Potassium Permanganate in Acetone Solution.-" solution of eserethole (1 g.) in 100 C.C.of acetone was maintained at about -10" while finely powdered potassium per-manganate (representing 6 0 ) was added during 2 days. The solution was then warmed with methyl alcohol to remove excess of permanganate filtered and the solvent evaporated. The red s p p thus obtained would not crystallise even after distillation in a high vacuum but on treatment with alcoholic picric acid it readily yielded a picrate (0.9 g.) which crystallised from alcohol in rhomb-shaped prisms m. p. 166" (Found C = 51.8; H = 4-9. Cl,H2,0,N,,C8H30,N3 requires C = 51.5; H = 4.7%). The base was recovered from this picrate by dissolving the latter in glacial acetic acid pouring into water and extracting the picric acid with ether making alkaline with sodium hydroxide and again extracting with ether.The oily base obtained on evaporation of the ether was warmed with methyl iodide and the product dissolved in hot methyl alcohol. On cooling the methiodide crystallised in needles, m. p. 198-199" (Found C =48*1; H = 5.8; I = 31-6. C1,H2,0~2,CH,I requires C = 4743; H = 5.8; I = 3106%). This methiodide resembles eserethole methiodide in its behaviour towards alkalis. It is stable towards sodium carbonate but when treated with sodium hydroxide yields an ether-soluble base. The latter could not however be crystallised neither could its methiodide be obtained in a crystalline condition. Etheserolene picrate was prepared incidentally during an experi-ment designed to test whether the presence of ethyl alcohol or sodium ethoxide was an essential condition for the formation of the esoline compounds described in Part I of this series.Starting from eserethole methiodide the experimental conditions mer PHYSOSTIGMINE ( ESERINE). PART III. 257 exactly similar to those described in that paper for the preparation of esoline ethyl ether dimethiodide except that a solution of sodium in n-propyl alcohol was used in place of one in ethyl alcohol. No solid separated however from the solution. The propyl alcohol was therefore evaporated the oily residue dissolved in water made strongly alkaline with solid sodium hydroxide and the solution refluxed for several hours tetramethylammonium iodide and an oil volatile in steam separating. The latter isolated from the mother-liquors by extraction with ether crystallised in a few hours, The solid m.p. about 45" (etheserolene according to Polonovski, melts at 48") was dissolved in alcohol and treated with alcoholic picric acid when a yellow picrate separated which crystallised from alcohol in stout prisms m. p. 98" (Found C = 52.1 ; H = 4.9. Calc. for Cl,H1,02N C6H307N3 C = 52.0; H = 4-b%). From the melting point of the original substance and the analysis of the picrate it seems evident that this was etheserolene picrate. It is also evident that the oxygen atom has no pseudo-basic function in this compound. Degradation of Esoline Ethyl Ether Dimethiodide.-The preliminary results of the degradation of this substance by Hofmann's method reported in Part I of this series have been confirmed. Since iden-tical results have been obtained by distillation of the product of the action of silver oxide on the iodide and by treating it with potassium hydroxide only the latter method need be described.The oil obtained by treating 4 g. of the iodide with silver oxide was refluxed for $ hour with 10 C.C. of SOY; potassium hydroxide. The solution was then steam-distilled until the oil had been com-pletely carried over. The distillate was extracted with ether upon evaporation of which an oily residue was obtained. This was dissolved in alcohol and the solution treated with picric acid. A viscous picrate separated which slowly solidified on warming. After filtration this was suspended in a relatively large volume of alcohol boiled and filtered. On cooling the filtrate a picrate separated in short stout prisms m.p. 156" unchanged on recrystallisation from aqueous alcohol (Found C = 46.7 ; H = 4.7. C,6H2,0N2,2C6H30,N3 requires C = 46.5; H = 4.7. soluble picrate which was separated in the manner described above, crystallised from aqueous alcohol in rhomb-shaped prisms m. p. quires C = 47.1 ; H = 4.7. C18H2g02N,,2C6H307N requires C = 47-2 ; H = 4-57;>. Both picrates were converted into methiodides by the following treatlment. The picrate was dissolved in glacial acetic acid poured into water extracted with ether to remove C16H260N2,2C6H30,N3 requires c = 46.7 ; H = 4.4%). The less 199" (Found C = 46.9 ; H = 4.7. C1&3,O2N2,2C6H,O7N3 re-VOL. CXXVII. 258 RIDEAL AND WILLIAMS THE ACTION OF LIGHT ON picric acid made alkaline with ammonia and the base extracted with ether.Addition of methyl iodide to the ethereal solution (dried over sodium sulphate) caused the methiodide to separate. The picrate melting a t 156" thus yielded a methiodide which crystallised in plates melted a t 169-170" and was identical with a methiodide which had been obtained directly from the crude product of the distillation of the quaternary hydroxide without purification through the picrate (Found C = 50.2; H = 7.4. C1,H,lON,I requires C = 50.2; H = 7.6. C1,H,,ON,I requires C = 50-5; H = 7.2%). The picrate melting at 199" yielded a methiodide m. p. 141"; when this was heated a t 100" in a sealed tube for 4 hour with excess of methyl iodide a second methiodide, m. p. 188" was obtained but no esoline ethyl ether dimethiodide was formed.Apparentlly this base was not as was originally believed the ethyl ether of bhe hypothetical esoline base. The authors desire to express their thanks to the administ'rators of the Carnegie Trust for the Scottish Universities for a grant which has alone rendered this research possible. DEPARTMENT OF MEDICAL CHEMISTRY, UNIVERSITY OF ED IN BURG^. [Received Notiember 18th 1924. PHYSOSTIGMINE (ESERTNE). PART m. 247 XLI1.-Ph ysostigmine (Eserine). Part I I I . By EDGAR STEDXAN and GEORGE BARGER. CERTAIN details of the structure of physostigmine have been elucidated (J. 1923 123 758; 1924 125 1373) but the evidence which was available was insufficient to permit the construction o 248 STEDMAN AND BBRGER: a formula which satisfactorily represented the known facts con-cerning the chemistry of this alkaloid.Further experimental evidence is now advanced and this combined with previously published work appears to leave no doubt as to the structure of eseret’hole and hence of physostigmine which are respectively the ethyl ether and methylcarbamido-derivative of eseroline. The presence in eseroline of the grouping I has been recognised since the work of Straus (AnnuZen 1913 401 350; 1914 406, 332) who obtained a phenolic indole compound (physostigmol) by degradation of eseroline methiodide. It was shown by one of us (E. S. ; Zoc. cit. 1373) that physostigmol is 5-hydroxy-1 3-dimethyl-indole (11) and further that its ethyl ether may be obtained in a yield of 66% by distillation of eserethole methiodide.This estab-lished the position of the hydroxyl group in eseroline and also confirmed Straus’s supposition that physostigmol contains a methyl group attached to position 3 of the indole ring. The formation of an indole compound by the somewhat violent decompositicn of an alkaloid does not in general permit the conclusion to be drawn that a preformed indole skeleton exists in the latter. Nevertheless, it appears justifiable in view of the comparat,ive completeness with which eserethole methiodide is degraded into physostigmol ethyl ether to assume even without further evidence the presence of the grouping I11 in eseroline. Additional evidence pointing to the presence of an indoline grouping in the molecule is not however, wanting. Thus the feebly basic properties of one of the nitrogen atoms in physostigmine (Straus Zoc.cit.) and the properties of etheserolene obtained by Max and Michel Polonovski (BUZZ. Xoc. chim. 1918 [iv] 23 335; 1923 [iv] 33 969) by the exhaustive methylation of eserethole point in this direction. The presence of the grouping I11 may thus be regarded as estab-lished and the problem resolves itself into determining the manner in which the remainder of the molecule (C,H,N) is linked to this grouping The nitrogen atom the position of which has still to be determined is the one with the stronger basic properties. Straus has shown that one methyl group is attached to this nitrogen atom whilst Salway (J. 1912 101 978) had previously demon-strated its tertiary character. Formula IV which was suggested to us by Professor R.Robinson F.R.S. conforms to these con-ditions and evidence will be adduced in the following discussio PHYSOSTIGMINE ( ESERINE). PART rn. 249 to show that this actually represents the structure of eseroline. The constitution of physostigmine itself will accordingly be repre-sented by formula V. Me CM Me CH, HOf>-?/\$!H I NHMe*CO*ON\ &/\?€I2 (v. ’ “-t!H -l<AIe \/\,,4H--NMe \/\/ (IV. 1 NMe NMe The considerations which have led to the adoption of formula I V for eseroline are largely based on the behaviour of eserethole towards reducing agents and on the properties of eseretholemethine. Polo-novski (Bull. Xoc. chim. 1918 [iv] 23 357) has shown that esere-thole on reduction with zinc and hydrochloric acid takes up two atoms of hydrogen a result which has now been confirmed using a catalytic method of reduction.That this reduction is not due to the presence of a double bond follows from the fact first pointed out by Max and Rfichel Polonovski (Compt. rend. 1024 178 2078), that whereas eserethole is a tertiary base dihydroeserethole is a secondary base. Reduction evidently opens a ring between the more strongly basic nitrogen atom and one of its adjacent carbon atoms. On the basis of IVY dihydroeserethole may thus be repre-sented by 171 the point a t which tjhe ring is ruptured being deter-Me Me I -$XH,*CH,*N,\le, NMe (T;II.) \ /\,/CJH.OH Etoli Et 0’ \--~-CH,*CH,*NH?IIe I i II \ / \ p H 2 (171.) NXe mined from Polonovski’s observation that the substance obtained by reduction of etheserolcne is identical with the product of the exhaustive methylation of dihydroeserethole.Eseretholemethine was ubtained by Max and Michel Polonovski (Zoc. cit.) by treating eserethole methiodide with sodium hydroxide. Its pseudo-basic character was not recognised until in Part I cf this series attention was directed to the fact that it is reconvertcd into eserethole rnethiodide on treatment with hydriodic acid m d that its behaviour in this respect was analogous to that of certain indoline compounds prepared by Brunner. As a result of this observation Polonovski revised the molecular formula of eserethole-methine and denionstrated that it was produced from methyl-eseretholinium hydroxide by a tautomeric change and not by loss of water.Experimental results which are enumerated in the succeeding paragraph have now been obtained which demonstrate that eseretholemethine is in fact a substituted indoline nit1 250 STEDMAN AND BARGER: hydroxyl group in the a-position (VII); its formation may accord-ingly be represented by the following scheme : tautomeric -CMe.CH 2 2 -CH *NMe CH2 4 /\ NaOH /\ change I -?Me yH2 -+ -$!Me $!H2 -+ -CH*OH -CH-NMe21 -CH-NMe,*OH Eserethole methiodide Intermediate Eseretholemethine quaternary hydroxide The position assigned to the basic nitrogen atom in V thus corre-sponds with this behaviour. Assuming the correctness of this scheme it is evident that while the conversion of eseretholemethine into a quaternary salt by treatment with methyl iodide should prevent the closure of the pyrrolidine ring on treatment with acids the compound should nevertheless retain its properties as a pseudo-base in virtue of its indolinol structure.This consequence has been verified. Treat-ment of eseretholemethine methiodide with picric acid results in the elimination of the hydroxyl group with the formation of a diquaternary picrate : \G-~Me*CH,*CH,=NMe,l \G-~Me*CH,°CH2*NMe3*OoC6H,06~3 ,,C CH*OH ,C CH ’ \/ ++ \/ /N\ NMe Me O*C6H,06N3 Still more conclusive are the results obtained by oxidation of eseretholemethine. Brunner (Nonatsh. 1896 17 253) has shown that 1 3 3-trimethyl-2-indolinol is oxidised by ammoniacal silver nitrate in alcoholic solution to 1 3 3-trimethyl-2-indolinone. By subjecting eseretholemethine to the same treatment a compound the composition of which corresponds with VIII has been obtained.The same substance has also been prepared by oxidising eserethole-methine with potassium ferricyanide a method which Decker ( J . pr. Chem. 1893 47 28) has employed for the oxidation of pseudo-bases. When this compound is degraded by the method of Et 0’ \-~MeCH2=CH,*NMe2 Et ON\-FMe-CH:CH2 I II \ \ / \ P O < II NMe W.1 \ \ / \ P O (VIII.) NMe Etd()-?MeEt (X. 1 NMe \ / \ P PHYSOSTIGMINE ( ESERINE). PART III. 251 exhaustive methylation trimethylamine is evolved and an un-saturated compound (IX) obtained which is practically devoid of basic properties and forms a deep crimson picrate. On reduction, this takes up two atoms of hydrogen with the formation of a substance which is considered to be 5-ethoxy-1 3-dimethyl-3-ethyl-2-indolinone (X).The chemical properties of eserethole-methine thus correspond entirely with those predicted on the assumption that V correctly represents the structure of physos-tiginine ; the constitution of this alkaloid may therefore be regarded as established. As an additional confirmation the synthesis of the substance represented by X has been undertaken and it is hoped that it will be possible to communicate the results shortly. An explanation of the mcchanism of the formation of eserethole-methine and of dihydroeserethole identical with that advanced above has recently been put forward by Max and Michel Polonovski (Compt. rend. 1924 178 2078). These authors therefore consider that the structures of eserethole and eseretholemethine are repre-sented by XI and XII respectively; the former differs from that proposed by Straus (Zoc.cit.) for eseroline only in the position of the nitrogen atom contained in the piperidine ring. Although the reactions described in this communication could be formulated with equal facility upon the basis of these structures it is evident that they do not explain the formation of physostigmol; the ease with which this substance is formed demands the presence of a preformed methyl group in the P-position of the pyrrolinc ring. CH, Consideration of the possible mechanism of the phytochemical synthesis of the ring system present in physostigmine speaks no less convincingly in favour of formula V. Perkin and Robinson (J.1919 115 944) have shown how harmine may be elaborated in the plant from tryptophan. If the assumption is made that the methylation of an indole nucleus may proceed in the plant in the manner in which it is known to take place in the laboratory a relation between physostigmine and this amino-acid at once becomes evident. By decarbosylation and methylation followed by a, ring closure the ring system of physostigmine would be readily formed 252 STEDMAN AND BAROER: (NH,)*CO,H decarboxylation f)-G*CH2*CH2*NH, -5- -NH / NH Me \/\/CH-fiMe NMe The possible mechanism of the phytochemical synthesis of the ring system present in XI cannot be similarly represented in a simple manner from known naturally-occurring substances. During the course of this investigation a small quantity of etheserolene was prepared.Contrary to Polonovski’s statement, it readily formed a yellow crystalline picrate. If etheserolene is formed by a straightforward degradation of eseretholemethine in a manner similar to that of the indolinone compound now described, one would feel inclined to attribute to it the structure represented by XIII. On this basis however it should possess the properties of a pseudo-base. But analysis of the picrate shows that this is not the case. It therefore appears that some isomeric change has taken place during its formation. This would correspond with the results of Polonovski who has recently stated (Cmpt. rend. 1924, 179 178) that etheserolene on reduction absorbs only two atoms of hydrogen and at the same time loses an atom of oxygen.Whether his view that this oxygen atom is not present as a hydroxyl group but forms part of a ring is correct cannot be stated with certainty a t present, Trinitroeserethole (XIV) and an oxidation product of eserethole which have been prepared during the course of this investigation and the results of a reinvestigation of the degradation of one of th PHYSOSTIGMINE (ESERINE). PART III. 253 esoline compounds described in Part I of this series are described in the experimental portion. E x P E R I M E N T A L. Dihydroeserethole (VI) .-Attempts to reduce eserethole catalytic-ally with a palladium sol in either neutral or acid solution with or without the addition of gum arabic as protective colloid werc unsuccessful owing to flocculation of the sol.With platinum black prepared by Willstatter's method (Ber, 1921 54 113), reduction took place readily. Platinum black (0.1 g.) was washed into a hydrogenation apparatus similar to that described by Hess (Ber. 1913 46 3113) air was completely removed a solution of 1 g. of eserethole in glacial acetic acid introduced and the mixture shaken; 52 C.C. of hydrogen were absorbed during the first hour, and 92 C.C. in all (calculated for two atoms 91 c.c.). The solution was filtered made alkaline with sodium hydroxide and the oil thus thrown down extracted with ether. From the extract dried over potassium carbonate a yellow oil was obtained which became almost colourless when distilled under the vacuum of a rotatory oil pump (bath at about 200°).When treated with alcoholic oxalic acid it formed an oxalate (sheaves of fine needles m. p. 204" from aqueous alcohol) identical with that obtained from dihydro-eserethole prepared by Polonovski's method (Zoc. cit.). Dihydroeserethole (1 g.) was heated in a sealed tube with excess of methyl iodide for 1 hour a t loo" the excess of methyl iodide evaporated and an alcoholic solution of the product treated with alcoholic picric acid. The yellow picrate precipitated on addition of water crystallised from aqueous alcohol in prisms m. p. 204" (Found C = 47.0 ; H = 4.7. C2,H,,Q1,N8 requires C = 47.3 ; H = 4.6%). The substance was thus a dipicrate and two methyl groups had been introduced into the molecule during methylation, indicating that in contrast Go eserethole dihydroeserethole is a secondary base.Behaviour of' Eseretholemethine Methiodide as a $-Base.-The orange oil which separated on mixing the salt (0-2 g.) and picric acid in alcoholic solution changed on warming into a crystalline solid (0.27 g.) which was obtained from aqueous alcohol in yellow prisms m. p. 170" (Pound C = 47.7; H = 4.7. C,,H,2015N,s requires C = 47.5; H = 4.4%). In the formation of this di-quaternary picrate water has been eliminated from the molecule, t l t u s inclienting the pscudo-basic function of the oxygen atom. L)eh~/clroeserethoZernethilze (VIII) .-The methine (1 g.) was refluxed for 10 hours with an ammoniacal solution of 3 g. of silver nitrate in methyl alcohol. After filtration evaporation and dissolutio 254 STEDMAN AND BARGER: of the residue in water ether extracted an insignificant amount of an oil which formed a crystalline picrate but was not investigated further.The hot aqueous solution made faintly alkaline with ammonia, was treated with hydrogen sulphide to remove excess of silver atered boiled to expel ammonia and hydrogen sulphide, and treated with alcoholic picric acid. The viscous precipitate first formed became crystalline on warming and by fractional crystallisation from aqueous alcohol gave two picrates the more soluble one forming needles m. p. 211" (Found C = 47.0; H = 4.5. Cl6H2,ON2,2CBH,0,N requires C = 46.7; H = 4.4%) and the less soluble fraction consisting of plates contaminated with some needles. The latter crystallised from acetone in plates m.p. 199" (Found C = 52.8 ; H = 5.4. CI6H2,O2N2,C6H,0,N requires C = 52-3; H = 5.4%). A different result was obtained by the following procedure An alcoholic solution of eseretholemethine (5 g.) was added to 15 g. of silver nitrate in alcohol made strongly alkaline with ammonia (d 0.88), the total volume being brought to about 100 C.C. A silver mirror formed on the sides of the flask almost immediately. The mixture was heated on the water-bath for 5 hours filtered and excess of silver removed by addition of hydrochloric acid; from the filtrate, made alkaline with sodium hydroxide ether extracted 4.45 g. of an oil which treated with alcoholic picric acid yielded 5.35 g. of the picrate m. p. 199" described above. No trace of the picrate of higher melting point was obtained in this experiment.It seems evident that this substance was formed by reduction of an indolinone to an indoline compound the reducing agent being the hydrogen sulphide used to remove excess of silver. Oxidation was also effected by means of potassium ferricyanide. A solution of eseretholemethine (3 9.) in a small quantity of alcohol was boiled with a strongly alkaline solution of 7-2 g. of potassium ferricyanide in 100 C.C. of water for about 5 minutes and after cooling extracted with ether. From this extract 4 g . of the methiodide which had previously been prepared in the manner described below were obtained. The picrate (4.19 g.) m. p. 199" was treated with sodium hydroxide and the oily base which separated extracted with ether.The extract was dried over sodium sulphate filtered concentrated, and methyl iodide added. A crystalline methiodide slowly formed (3.3 g.). This separated from acetone in transparent prisms but the method was wasteful owing to its large solubility in this solvent. It was therefore crystallised by addition of dry ether to the acetone solution. With dry solvents the methiodide was obtained as a colourless substance m. p. about 131"; the fact that the meltin PHYSOETIGMINE ( ESERINE). PART III. 256 point was not absolutely sharp was no doubt due to the water of crystallisation which it contained (Found for air-dried material, H,O = 3.7 ; for material dried at 110" I = 30.0. C1,HZ70,N,I,H,O requires H,O = 4.1 ; C1,H,,O,N,I requires I = 30.1%). When, however the solvents were not perfectly dry the transitory form-ation of a blue iodine adsorption compound was consistently observed ; its disappearance no doubt coincided with the trans-formation of an a t first amorphous precipitate into crystalline form.Degradation of Deh ydroeseretholemethine i!fethiodide by Hof I)Z an n ' s Nethod to Compound IX.-An aqueous solution of 1.9 g. of the methiodide was shaken with a suspension of silver oxide prepared from 2 g. of silver nitrate. After filtration and removal of the water by evaporation under diminished pressure the residual oil was distilled a t 12 mm. After evident decomposition with the evolution of a gas a slightly brown oil distilled when the tem-perature of the bath was between 200" and 210". The distillate would not crystallise.It was therefore treated with an alcoholic solution of picric acid when a small quantity of a crystalline picrate m. p. 199" identical with that described above was obtained. The residue from the mother-liquors was treated with sodium hydroxide and extracted with ether and the extract washed with water until the sodium picrate was completely removed. After drying over sodium sulphate the ether was evaporated, when a colourless oil which crystallised spontaneously was obtained. Recrystallised from aqueous alcohol it formed colourless prisms, m. p. 62" (Found C = 72.3; H = 7.3. C14H1702N requires C = 72-7; H = 7-4:/,). This substance was practically devoid of basic properties; it was insoluble in dilute hydrochloric acid but formed a picrate which at first separated from alcohol in a yellow form and quickly changed into a deep crimson one; m.p. 103". Reduction of Compound I X to Compound X.-Hydrogenation was effected in the apparatus mentioned above using a solutim of colloidal palladium as catalyst and gum arabic as protective colloid; 0.16 g. of substance was used. Hydrogen was readily absorbed but as a result of a mishap the amount could not be measured. After filtration from the palladium and evaporation of the alcohol (aqueous alcohol was used as solvent) the solution, after addition of a small quantity of sodium hydroxide to ensure alkalinity was extracted with ether. After drying and evaporation, this yielded a crystalline product which when crystallised from aqueous alcohol formed cubes m.p. 68" (Fcund C = 71.8 H = 8-2. It is evideat from the analysis that two atoms of hydrogen were absorbed and C1,H1,O,N requires C = 72.1 ; H = S-2:/,) 256 STEDMU AND BARGER: that the oxygen atom present as a carbonyl group was retained in the molecule. Trinitroeserethote.-Among the diverse experiments which have been carried out with a view to effect the oxidation of eserethole, the action of nitric acid on this substance has been studied. Cold dilute nitric acid has little if any action but the concentrated acid (d 1-4) reacts vigorously. Eserethole (1 g.) was added drop by drop to 3 C.C. of concentrated nitric acid cooled in a freezing mixture and vigorously stirred. The product was warmed on the water-bath for 8 hour cooled diluted with waber and the yellow crystalline precipitate (0.3 g.) recrystallised from aqueous alcohol, trinitroeserethole separating in orange rectangular plates m.p. 152" (Found C = 47.1 ; H = 5.0; N = 18.7. C1,H1,O,N requires C = 47-2; H = 5.0; N = 18.4%). It is insoluble in alkalis and dilute acids but dissolves in concentrated acids. Oxidation of Eserethole with Potassium Permanganate in Acetone Solution.-" solution of eserethole (1 g.) in 100 C.C. of acetone was maintained at about -10" while finely powdered potassium per-manganate (representing 6 0 ) was added during 2 days. The solution was then warmed with methyl alcohol to remove excess of permanganate filtered and the solvent evaporated. The red s p p thus obtained would not crystallise even after distillation in a high vacuum but on treatment with alcoholic picric acid it readily yielded a picrate (0.9 g.) which crystallised from alcohol in rhomb-shaped prisms m.p. 166" (Found C = 51.8; H = 4-9. Cl,H2,0,N,,C8H30,N3 requires C = 51.5; H = 4.7%). The base was recovered from this picrate by dissolving the latter in glacial acetic acid pouring into water and extracting the picric acid with ether making alkaline with sodium hydroxide and again extracting with ether. The oily base obtained on evaporation of the ether was warmed with methyl iodide and the product dissolved in hot methyl alcohol. On cooling the methiodide crystallised in needles, m. p. 198-199" (Found C =48*1; H = 5.8; I = 31-6. C1,H2,0~2,CH,I requires C = 4743; H = 5.8; I = 3106%). This methiodide resembles eserethole methiodide in its behaviour towards alkalis.It is stable towards sodium carbonate but when treated with sodium hydroxide yields an ether-soluble base. The latter could not however be crystallised neither could its methiodide be obtained in a crystalline condition. Etheserolene picrate was prepared incidentally during an experi-ment designed to test whether the presence of ethyl alcohol or sodium ethoxide was an essential condition for the formation of the esoline compounds described in Part I of this series. Starting from eserethole methiodide the experimental conditions mer PHYSOSTIGMINE ( ESERINE). PART III. 257 exactly similar to those described in that paper for the preparation of esoline ethyl ether dimethiodide except that a solution of sodium in n-propyl alcohol was used in place of one in ethyl alcohol.No solid separated however from the solution. The propyl alcohol was therefore evaporated the oily residue dissolved in water made strongly alkaline with solid sodium hydroxide and the solution refluxed for several hours tetramethylammonium iodide and an oil volatile in steam separating. The latter isolated from the mother-liquors by extraction with ether crystallised in a few hours, The solid m. p. about 45" (etheserolene according to Polonovski, melts at 48") was dissolved in alcohol and treated with alcoholic picric acid when a yellow picrate separated which crystallised from alcohol in stout prisms m. p. 98" (Found C = 52.1 ; H = 4.9. Calc. for Cl,H1,02N C6H307N3 C = 52.0; H = 4-b%).From the melting point of the original substance and the analysis of the picrate it seems evident that this was etheserolene picrate. It is also evident that the oxygen atom has no pseudo-basic function in this compound. Degradation of Esoline Ethyl Ether Dimethiodide.-The preliminary results of the degradation of this substance by Hofmann's method reported in Part I of this series have been confirmed. Since iden-tical results have been obtained by distillation of the product of the action of silver oxide on the iodide and by treating it with potassium hydroxide only the latter method need be described. The oil obtained by treating 4 g. of the iodide with silver oxide was refluxed for $ hour with 10 C.C. of SOY; potassium hydroxide. The solution was then steam-distilled until the oil had been com-pletely carried over.The distillate was extracted with ether upon evaporation of which an oily residue was obtained. This was dissolved in alcohol and the solution treated with picric acid. A viscous picrate separated which slowly solidified on warming. After filtration this was suspended in a relatively large volume of alcohol boiled and filtered. On cooling the filtrate a picrate separated in short stout prisms m. p. 156" unchanged on recrystallisation from aqueous alcohol (Found C = 46.7 ; H = 4.7. C,6H2,0N2,2C6H30,N3 requires C = 46.5; H = 4.7. soluble picrate which was separated in the manner described above, crystallised from aqueous alcohol in rhomb-shaped prisms m. p. quires C = 47.1 ; H = 4.7. C18H2g02N,,2C6H307N requires C = 47-2 ; H = 4-57;>.Both picrates were converted into methiodides by the following treatlment. The picrate was dissolved in glacial acetic acid poured into water extracted with ether to remove C16H260N2,2C6H30,N3 requires c = 46.7 ; H = 4.4%). The less 199" (Found C = 46.9 ; H = 4.7. C1&3,O2N2,2C6H,O7N3 re-VOL. CXXVII. 258 RIDEAL AND WILLIAMS THE ACTION OF LIGHT ON picric acid made alkaline with ammonia and the base extracted with ether. Addition of methyl iodide to the ethereal solution (dried over sodium sulphate) caused the methiodide to separate. The picrate melting a t 156" thus yielded a methiodide which crystallised in plates melted a t 169-170" and was identical with a methiodide which had been obtained directly from the crude product of the distillation of the quaternary hydroxide without purification through the picrate (Found C = 50.2; H = 7.4. C1,H,lON,I requires C = 50.2; H = 7.6. C1,H,,ON,I requires C = 50-5; H = 7.2%). The picrate melting at 199" yielded a methiodide m. p. 141"; when this was heated a t 100" in a sealed tube for 4 hour with excess of methyl iodide a second methiodide, m. p. 188" was obtained but no esoline ethyl ether dimethiodide was formed. Apparentlly this base was not as was originally believed the ethyl ether of bhe hypothetical esoline base. The authors desire to express their thanks to the administ'rators of the Carnegie Trust for the Scottish Universities for a grant which has alone rendered this research possible. DEPARTMENT OF MEDICAL CHEMISTRY, UNIVERSITY OF ED IN BURG^. [Received Notiember 18th 1924.