350 DUNSTAK AND CARR: CONTRIBUTIONS TO OUR ~XIIiVI.-Co7bt?’ib~tio12~ to ozw Kuozr!ledge of the Aconite Alkaloids. Pcwt X I V.-O.iz Pseudaconitine. By WYNDHAM R. DUNSTAN, %LA., F.R.S., and FRANCIS H. CARR, A.I.C., Salters’ Company’s Research Pellom in the Laboratories of the Scientific Department of the Imperial Institute. 1~ previous papers communicated to this Society, an account has been given of an investigation of the principal properties and decomposition products of the alkaloid aconitine, derived from the roots of Aconitulrn Nci,peZZus. The enquiry has since been extended to a simiIar examination of the alkaloids occurring in other species and varieties of aconite. At * During the preparation of nitroethane by the action of mercurous nitrite on ethylic iodide, I have obtained mercurous iodide in the shape of bright yellow scales.See also Yvon (Cmnpt. rend., 1873, 76, p. 1607). Conipt. rend., 1883, 96, 30.KNOWLEDGE OF THE ACONITE ALKALOIDS. 351 the request of the Government of India, an investigation is being made, in the Scientific Department of the Imperial Institute, of the alkaloidal constituents of the chief kinds of aconite indigenous to India, especially of those which are highly poisonous, or are reputed t o be of medicinal value. I n this connection, Dr. H. A. D. Jowett has described (Part XI11 of this series) the principal properties and decomposition products of ntisine derived from the Aconituna hetevoplyllZuL.l?t of India. I n a previous communication, and in the present paper, we give an account of pseudaconitine, the highly poisonous constituent of the aconite occurring in Nepaul, which is usually regarded as Aconitum fei-ox, and locally known as ‘‘ bish” (bikh). Our previous knowledge of this alkaloid is almost wholly due to the researches of Alder Wright, who, in conjunction with Luff, gave an account of its properties in a paper communicated to this Society in 1878.The material employed in our work consisted of roots of the plant, which were specially collected with great care in the Himalayas under the supervision of Dr. George Watt, the Reporter on Economic Products to the Government of India. I n a preliminary notice communicated to the Society two years ago (Proc., June, lS95), the authors described some of the properties of pseudaconitine. They showed that its hydrolysis occurs in two stages, in the first of which acetic acid and a crystalline base verntrylpseud- aconine are formed, and in the second the elimination of a molecule of dimethylprotocatechuic acid takes place with the formation of pseud- aconine.It has also been shown that, when pseudaconitine is heated in the dry state, one molecular proportion of acetic acid distils over, and a base is left, to which the name pyropseudaconitine was given. This base, when hydrolysed, furnishes dimethylprotocatechuic (veratric) acid and pyropseudaconine. The present paper gives a more detailed account of the experiments which furnished these results, and also an account of other observations on the properties of the salts and derivatives of pseudaconit ine. Extvccction of the AZkdoiJ.-Several methods have been tried for the extraction of the base froin the root, involving the use of methylic, ethylic, and amylic alcohols.Finally, a mixture of methylic and amylic alcohols, in the proportion of 5 to 1, was adopted as the most efficient solvent. The metbylic alcohol is distilled from the slightly warmed percolate, under reduced pressure, when a quantity of a t separates; this removed, and the alkaloid is extracted from the amylic alcohol by making it with very dilute (1 per cent.) aqueous hydrochloric acid. The solution is then shaken with ether, to remove E l 3 9352 DUNSTAN AND CARP,: CONTRIBUTIONS TO OUR the dissolved amylic alcohol, the alkaloid liberated by the addition of dilute ammonia, and then extracted by shaking with ether in the usual manner.On evaporating the dried ethereal solution, white crystals separate, which are recrystallised by dissolving them in dry chloroform, adding dry ether, and then dry light petroleum, until a slight turbidity is produced; by this means, a considerable supply of pure pseudaconitine was obtained. Judging from the yield obtained from the roots of Aconitum feyoz, it would appear that more pseudaconitine is present in them than there is of aconitine in the roots of A . Nctpellus; but this is a question to which we shall return in a future paper. Propes.ties of Pseudc6conitine.-The pure base crystallises well. Mr. W. J. Pope has kindly examined some fairly well-defined crystals, with the following results. '' The crystals of pseudaconitine consist of small, colourless, trans- parent crystals of rhomboidal shape, having a rather vitreous lustre. Owing to the poor character of the images obtained from the various faces, the measurements given below are of no great accuracy; they would indicate that the crystals belong to the orthorhombic system.That the crystals are, however, not orthorhombic, is shown by the faces which they exhibit, and also by the interference figure observed in polarised light. Considering, for the purpose of description, that the crys- tals are really orthorhombic, the following faces are always observed : -(010), (OlO), (lll), (111), ( l i l ) , (IiT), (ili), and ( n l ) , with traces of the form (110) ; the two faces (111) and (711) are never observed. This observation was made on all the crystals examined-ten in number, belonging to two different crops-and the same faces of the forms (010) and (111) were found in every case ; this arrangement of faces is not possible in the hemihedral subdivisions of either the orthorhombic or monosymmetric system.The crystals must, there- fore, be assigned to the rare anorthic hemihedral system, two of the forms having the indices 111 being present as half-forms only, and the interaxial angles a, p, and y being equal to 90" within the rather wide limits of error incurred in the measurement of crystals such as those now described. "The crystals present the appearance shown in Fig, 1, and have the axial ratios-KNOWLEDGE OF THE ACONITE ALKALOIDS. 353 n : b : c = 0.8362 : 1 : 0.6938. FIG.1. '' The following angular measurements were obtained : No. of .Angle. measnrernen ts. ho = 010 : 111 23 61" 09 = 11_1 : 111 10 54 bo = Ol_O :211 9 117 00 = 11; : 111 18 67 00 = 111 : 111 7 111 00 = 111 :Ti: s 94 00 = 111 : 111 4 S5 pp = 110 : 110 4 fS p p =: 110 : i i o 1 Linii ts. 14'-- 62" 17' 27 - 56 46 26 -118 19 1 2 - 68 59 1 4 -111 35 1 -- 95 29 2 - 85 37 54 - SO 16 - IIean. 61" 51' 56 5 117 5s 68 34 111 24 94 35 55 20 79 20 101 4 Calculated. - 56" 12' 118 6 111 26 94 29 E5 31 I 9 4s 100 1 2 - "The crystals are very brittle, and possess a good cleavage; the latter however, could not be determined. On examining a cleavage fragment under n very wide angle objective, one optic axis is seen t o emerge a t the edge of the field; it shows that the dispersion is inclined, which is only possible in the monosymmetric or anorthic system.The hemi- hedral character of the crystals is of interest, because non-superposable hemihedrisiu is so rarely observed on crystals of the natural alkaloids, that it has been said not to occur. (Wyrouboff, Ann. C?&12. P?~ys., 1894, [vii], 1, 11). "The crystalline form of aconitine has been determined by Tutton (Trans., 1S91, 59, ZCS), who found the crystals t o be orthorhombic, but did not observe hemihedrism. Although morphotropic relationships would seem t o exist between the crystalline forms of aconitine and pseudaconitine, yet these can hardly be worked out from the data now- given for the latter alkaloid ; the following corresponding angles on the two compounds seem t o show some similarity." Aconitine.Pseudaconitinc. 100 : 121 60" 39' 010 : 111 61' 54' 010 : 121 57 42 100 : 111 55 43 001 : 121 46 33 001 : 111 47 15" The crystals melt with decomposition at 201", acetic acid gradually distilling off; the melting point is fairly sharp if the substance is put into the bath heated to 150" and the temperature slowly raised. Wright and Luff (Zoc. cit.) have recorded 104-105" as the melting point of354 DUXSTAN AND C'ARR: CONTRIBUTIONS TO OUR pseudaconitine. They state that Ohe alkaloid contains 1 H,O, which is lost at loo", but we have not been able to confirm this observation. Pseudaconitine dissolves readily in alcohol, chloroform, and acetone, less readily in ether, very slightly in water, and scarcely a t all in light petroleum. A determination of the specific totatory power, using an alcoholic solution, gave Pseudaconitine is dextrorotatory.c = 1.12 I = 2 d??E. a = 25' t = 15" 100 x 25 whence [aID=5i 1.12 + 18" 36'. The ordinary salts of pseudaconitine are Isvorotatory, and usually Combustion of the alkaloid made with soluble in water and alcohol. the material dried at 100" gave the following results. I. 0.2612 gave 0.5964 GO, and 0,1694 H,O. C = 62-99 ; H= 7.20. 11. 0.2587 ,, 0.5975 CO, ), 0.1484 H,O. C=62*96 ; H=6.37. These figures nearly correspond with those calculated from the for- mula proposed by Wright and Luff (Trans., 1878, ii, 151)) namely, C1,GH,,NO,,, for which the calculated percentages are, for carbon, 62.88 ; for hydrogen, 7.13. Like aconitine, pseudaconitine and its salts, even in very dilute solution, give rise to a persistent tingling and numbing sensation on the tongue, and are highly poisonous.From preliminary experiments on the relative toxicity of various aconite alkaloids, which have been made at our suggestion by Dr. F. W. Tunnicliff e, it mould appear that pseudaconitine is slightly more toxic than aconitine. S d t s of Pseucln conitine. Pseudnconitine / ~ y d ~ o c h h i d e , (&H4,'N0,,,HC1.-\~e have not suc- ceeded in obtaining this salt in a crystalline condition. It has been prepared by the direct action of dilute hydrochloric acid on both aqueous and alcoholic solutions of the base, but all attempts t o crystallise it from water, alcohol, or a mixture of alcohol and ether have resulted in the production of a colourless varnish.Pseudaconitine Iqch-obromicle, C1,,H,,NO1,,HBr.-This salt is pre- pared by dissolving the base in dilute hydrobromic acid and evapo- rating the solution. A colourless varnish remains, and on adding n, little alcohol to this, the mass rapidly becomes crystalline. It is best purified by dissolving it in dry alcohol and adding dry ether until a slight turbidity is produced; it then separates in large, cubical crystals often arranged in rosettes. The salt readily dissolves in alcohol and water, but is insoluble in ether and light petroleum. The crystals con tain 2H,O, which are expelled on drying at 100-103". The driedKNOWLEDGE OF THE ACONITE ALKALOIDS. 365 substance melts a t 191". by heating a t 100-103" in an air bath. 0.546 lost 0.0863 = 4.8 per cent. H,O.Determinations of the bromine in the undried and in the dried The water of crystallisation was estimated substance gave the following figures. 0.3379 nndried salt gave 0 0751 AgBr. 05197 dried salt gave 0.1227' AgBr. C36H4nNOl,,HBr + 8H,O requires H,O = 4.5. Br = 9.95 per cent. C,,H,,NO,,,HBr requires Br = 10.3 per cent. of the specific rotatory power led to the folloming result Br = 9.44. Br = 10.05. An aqueous solution of the salt is lzevorotatory ; the determination I = 2 cln8 c = 0.6635, Pseudaconitine, therefore, resembles aconitine in being a dextro- rotatory base whose salts are laevorotatory. Pseudccconitiqze hydriodide, C:36H4SN012,HI.--This salt is precipi- tated in an amorphous condition when aqueous potassium iodide is added to an aqueous solution of pseudaconitine hydrobromide.Although at first amorphous, the precipitate rapidly becomes crystalline ; it may readily be purified by recrystallisation from a mixture of alcohol and ether. Psezcdacovaiti?ze qzitrccte, C,GH,nNO,z,HNO,.-This salt was prepared by Wright and Luff (Zoc. cit.) by dissolving the base in dilute nitric acid and precipitating the nitrate by adding strong nitric acid, in which it is only sparingly soluble; this method, however, is not to be recom- mended, as strong nitric acid is very apt to decompose the alkaloid. By dissolving the alkaloid in dilute nitric acid to exact neutrality, and evaporating to dryness, the nitrate is obtained as an amorphous varnish, which crystallises a t once on the addition of alcohol ; it is readily puri- fied by crystallisation from a mixture of alcohol and ether, and when pure may be crystallised from water.The dried salt melts at 192" and effervesces at a slightly higher temperature ; the melting point is fairly sharp if the substance is put into the bath at 155" and slowly heated. The water of crystallisation was determined by heating the powdered, air-dried salt a t 100-105". 0.1975 lost 0.014 H,O. H,O= 7.0 per cent. C:3GH,,NOl,,HNOs + 3H,O requires H,O = 6.7 per cent. 5 C.C. of a solution saturated at this temperature yielded, on evaporation, 0.209 gram of salt. 100 C.C. of water at 15", therefore, dissolves 4-15 grams of salt. Its solubility in water a t 15" was determined.356 DUNSTAN AND CARR: CONTRIBUTIONS TO OUR Hydrolysis of Pseuduconiiiine.We have previously pointed out (Zoc. c i t . ) that, in addition to the pseudaconine and veratric acid, observed by Wright and Luff, acetic acid is formed by the hydrolysis of pseudaconitine, and we have also shown that the hydrolysis may occur in two stages. To determine the first stage only in the hydrolysis, namely, the elimination of acetic acid with the formation of veratrylpseudaconine, it is best to employ a process similar to that which was found to answer in the case of aconitine. A neutral aqueous solution of a pseudaconitine salt, pre- ferably the sulphate, is heated in a sealed tube at 135" for 3 hours, the amount of acetic acid formed is determined by direct titration with N/10 alkali, and the alkaloid, after being liberated by the addition of sodium carbonate, is dissolved by shaking with ether.A solution containing 0.168 gram of alkaloid (as salt), after this treatment, required for neutralisation 4.3 C.C. of N/10 alkali = 7.5 per cent. of acetic acid, which is slightly lower than that calculated for one molecular proportion, namely, 8.7 per cent. Analysis of the silver salt of this acid showed that it contained 64.56 per cent. of silver. Silver acetate contains 64.66 per cent. The formation of veratrylpseudaconine may thus be represented by the following equations. C3,H,9N0,, + H,O = C3,H,~NOll + CH,*COOH. Psendaconitine Veratrylpseudaconine Ye,.c~t,ylpseudcLconine. The pure base crystallises from ether i n large, irregular crystals, which are nearly insoluble in water and in light petroleum, but readily soluble in ether, alcohol, and chloroform. They melt at 199" when put into the bath at 150".Ade- termination of the specific rotatory power in alcoholic solution led to the following results. A solution of the base is lzvorotatory. t = 16', U = - 1'16' c = 1.5035 Yeratrylpseudaconine, therefore, unlike its analogue benznconine, exhibits rotatory power of the opposite sign t o that of its parent base. Aconitine and benzaconine are both dextrorotatory, whilst pseud- aconitine is dextrorotatory and veratrylpseudaconine lzevorotatory. Combustions of the base, dried at 100-103", futnished the following results, showing it to be a monhydrate. I. Carbon 61.44; hydrogen 7.15 per cent, 11. ,, 61-01 ; ? ¶ 7-05 99 9 , Calculated for C,,H,7NOll,H,0 : Carbon 61 -54 j hydrogen 7.34 per cent.KKOWLEDGE O F THE ACONITE ALKALOIDS.357 This alkaloid and its salts have a very bitter taste, but produce no tingling sensation, and do not appear to be poisonous. Ve1wat?.yl~seudc~conine Hyds.obronaide, C,4H,7NO11,HBr.-This salt separates from a mixture of alcohol and ether in large, prismatic crystals which contain 3H20. 0.3478 salt lost 0.0264 H,O a t 100'. 0.3478 ,, gave 0.0842 AgBr. Br = 11.21. C,,H,~NO,,,HBr + 3H20 requires H,O = 6.95 ; Br = 11.02 per cent. The salt is soluble in water, alcohol, and chloroform. Ve~~c~~~~~Zpseudaconi.ne nitmte, C,4H,7N0,1,HN03, crystallises from a mixture of alcohol and ether in rhombic prisms arranged in rosettes. In melting, two fairly sharp points may be noticed, one at 222", when softening and change commences, and at 2 3 2 O , when the salt melts sharply with decomposition.Combustion of the anhydrous salt gave C = 57.34 ; H= 6-40 per cent. : C,,H,7N0,1,HN03 requires C = 57.62 ; H = 6.77 per cent. Vercctil:?/1psezcdccconine aus.ichZoride, C,4H,~NOl,,HAuC14, is thrown down as a pale yellow, amorphous precipitate when auric chloride is added to a solution of the hydrochloride. It is insolubTe in water, ether, and light petroleum, but readily soluble in ethylic and methylic alcohols, chloroform, and acetone ; it could not be crystallised from any of the last-mentioned solvents alone, or on the addition of ariy of the former to them. Pseudaconine. H,O= 7.5. The second stage of the hydrolysis by which veratrylpseudaconine passes into veratric acid and pseudaconine, may be best effected by adding alcoholic soda to an alcoholic solution of pseudaconitine, or veratrylpseudaconine.Hydrolysis takes place rapidly, and is complete in about 2 hours. Dilute sulphuric acid is then added, the filtrate evaporated, the veratric acid extracted from the acidified solution by ether, and the pseudaconine by chloroform, after rendering the solution alkaline with ammonia. 0.2143 gram of alkaloid gave 0.0556 gram of veratric acid = 25.94 per cent, The acid melts a t 178' and exhibits the other properties of di-methyl- protocatechuic acid (veratric acid). This stage of the hydrolysis may therefore, be represented by the equation, Calculated for one molecular proportion, 26.49 per cent. C,H,7N0,, + H20 = G,H3(OCH3),*COOH + C,,H,,NO, Veratryl pseudaconine Veratric acid Pseudaconine.Pseudaconine is an amorphous, hygroscopic base readily soluble in water, chloroform, alcohol, and acetone, and less readily in ether. Its aqueous solution is strongly alkaline to litmus. All attempts to crgstallise the base uncombined with its solvent have been unsuccessful. VOL. LXXI. c c3.58 DUNSTAN AND CARR: CONTRIBUTIONS TO OUR An aqueous solution of pseudaconine is dextrorotatory. The specific rotatory power of an aqueous solution was determined with the follow- ing results. a[20°] = +32.5 ~ = 0 . 8 9 6 Z=2 dfih, 100 0.541 = +300 6'. whence [a], = __-- ~ 2 x 0.896 Yseudaconins hyd~ochlos*ide, C,,H,,NO,,HCl, wagprepared by dissolving the base in dilute hydrochloric acid to neutrality. Many attempts made t o crgstallise this salt from various solvents were unsuccessful, although, on one occasion, crystals were obtained from an alcoholic solution which had stood for six months ; these were prisms and melted at 68'.Pseudccconilze hydvobromide, C,,H,,NOs,HBr, was prepared in the same manner as the hydrochloride, but it could not be crystallised. Pseudaconine nitrrwte, C,,H,,NO,,HNO,, was prepared by the direct action of dilute nitric acid on the base, and also by double decomposi- tion between silver nitrate and the hydrochloride, and barium nitrate and the sulphate. It was always obtained in an amorphous state. Pseudaconine sulphccte, ( C,,H,,NO,),,H,SO,, was prepared by acting on pseudaconine with dilute sulphuric acid, but this salt could not be crystallised.Pseudaconine c~uu?*ichlo~icEe, C,,H,,NO,,HAuCl,, is precipitated when auric chloride is added to a concentrated solution of pseudaconine hydrochloride. It is a yellow, amorphous precipitate sparingly soluble in water, and could not be crystallised from any of its solutions. When light petroleum is used, the yellow colour of the solution is clis- charged and it becomes colourless, although no gold is precipitated. This change, in other cases, has been observed t o be due to the produc- tion of an aurichlor-derivative by loss of hydrogen chloride from the aurichloride. We have so far failed to crystallise an aurichlorpseud- aconine from this solution. Pyropseudccconitine. As previously recorded by us, when pseudaconitine is heated slightly above its melting point, it effervesces and loses acetic acid.A deter- mination of the amount of acetic acid which distils under these circumstances, proved that one molecular proportion of acetic acid is expelled ; analysis of the silver salt of the acid proved it to be silver acetate. The reaction may therefore be represented by the following equation. C,,H,,NO,, = ~,H,O, + %4H,,NO,, Pseudaconitine Pyropseudaconitine Yyropseudaconitine, an anhydride of veratrylpseudaconine, is obtained from the residue by solution in dilute acid, and is purified by fractional precipitation from this solution with dilute ammonia. The colourlessKKOWLEDGE OF THE ACONITE ALKALOIDS, 359 fractions are dissolved in dilute acid, precipitatecl with ammonia, and the pure base extracted from the alkaline solution by ether.The base so far has only been obtained as an amorphous varnish, nearly insoluble in water, but readily soluble in alcohol, chloroform, and ether. I t s salts appear to crystallise well; they have a bitter taste, but produce no tingling, and do not seem to be poisonous. The hydi-iodide crystallises in prisms. Although, in publishing our preliminary notice of psendaconitine we stated that we were engaged in a complete investigation of the alkaloid, this did not prevent Herr lllartin Freund from examining the alkaloid, and, nine months after the appearance of our paper in the Proceedings of this Society, publishing in the “Berichte ” (Bey., 29, 6,552) a n account of his and Herr Niederhoffheim’s experiments on the subject. They adopt Wright’s formula for the alkaloid, and confirm our conclusion that pseudaconitine, like aconitine, contains an acetyl group. For the rest, they record melting points which differ somewhat from those pre- viously recorded by us, but since these points are in most cases decom- posing points, and depend on the conditions under which the observations are made, no real importance attaches to these discrepancies. As to their assertion that pseudaconine is the anhydride of the aconine derived from aconitine, it is t o be observed that this statement is based solely on the numerical coincidence that the formuIa for pseud- aconine deduced from Wright’s formula for pseudaconitine, namely, C,5H39N08, differs by one molecule of water from the formula which Freund has suggested €or aconine (C,,H,,NO,). But, as we have else- where pointed out, Freund’s new formulze for aconitine and its deriva- tives cannot at present be accepted as proved, and we have so far seen no reason to depart from a formula for aconitine, differing very slightly from that originally suggested by Wright, which does not allow of pseudaconine ( C,,H3,N08) being regarded as the anhydride of aconine (C2,H3,N0,,). As a matter of fact, we have already described an anhydride of aconine (pymconine), whose properties are very different from those of pseudaconine. We have pleasure in acknowledging the skilful help afforded us in the early stages of this work, in tlie Research Laboratory of the Phar- maceutical Society, by Mr. H. T. Durant. SCIENTIFIC DEPARTJIEST, IIVlPERIzlL INSTITUTE, LOXDON. c c 2