COMPOUNDS OF THE TRYPARSAMIDE TYPE. PART I. 479 LXXI.-Compounds of the Tryparsarnide Type. Part I. Resolution of N-Phenyl~lanine-4-ars~n~c Acid and of its Amide. GIBSON,JOHNDOBNEY .JOHNSON, By CHARLESSTANLEY ANDREW and BARNETT LEVIN. N-PHENYLGLYCINEAMIDE-4-ARSINICacid (" tryparsamide ") was first prepared by Jacobs and Heidelberger (J. Amer. Chem. ~oc. 1919 41 1589) and along with nnmerous derivatives examined GIBSON JOHNSON AND LEVIN therapeutically by (Miss) Pearce and Brown (J. Exp. Med. 1921 33,193). Tryparsamide has been described as the most effective organic arsenical for the treatment of human sleeping sickness due to infection by Tr. gambiense. In attempting to extend our knowledge of organic arsenicals belonging to this series it was decided to investigate in the first place those externally compensated compounds which can be resolved into their optically active components so that the thera- peutic value of the optical isomerides might be compared as well as that of the externally compensated compound itself.Further it is hoped to ascertain whether the marked differences in pharmaco- logical and therapeutical action found among other analogous groups of optical isomerides exist in this series also (compare Cushny “ Biological Relations of Optically Active Substances,” 1926).* NH*CHMe*CO,H NH*CHMe*CO*NH (II-) v 0 AsO(OH) AsO(OH) dl-N-Phenylalanine-4-arsinic acid (I)was prepared by the con- densation of sodium p-arsanilate (“ atoxyl ”) with a-bromopropionic acid in aqueous solution (compare D.-R.P.204,664of 1908; Meister Lucius and Bruning) and preliminary investigation indicates that it has Borne therapeutic value. It is essentially a dibasic acid although titration against sodium hydroxide and phenolphthalein indicates a slightly greater basicity than two in dilute solutions. The methyl and the ethyl ester-only the carboxyl group being esterified-of the acid were prepared and from the former but not from the latter the corresponding amide-acid dl-N-phenylalanine- amide-4-arsinic acid (11),was obtained by the action of aqueous ammonia. This amide-acid has been prepared by Jacobs and Heidelberger [Eoc. cit. p. 1597 ; who name it N-(phenyl-4-arsonic acid)- a-amino-propionamide] and by Fourneau and Nicolitch (loc.cit. p. 1211)by the action of a-bromopropionamide on p-arsanilic aoid. The decomposition point of this dl-amide-acid recorded by Jacobs and Heidelberger (darkening at 255” and decomposing at * The present work was begun some two years ago. It was almost com-pleted when Fourneau and Nicolitch announced the successful resolution of dl-N-phenylalanineamide-4-arsinicacid (phenylmethylglycineamidearsinic acid) and the application of the optically active enantiomorphs to the resolution of synthetic ephedrine. In correspondence M. Fourneau encouraged us to continue our work and sent us a copy of the thesis of M. Nicolitch. This has since been published (Fourneau and Nicolitch BUZZ.SOC.chirn. 1028 34 1232). COMPOUNDS OF THE TRYPARSAMIDE TYPE PART 1.481 282-263.5") is somewhat higher than that recorded by us (never higher than 244") but too much stress should not be laid on this since the actual temperature at which the profound decomposition sets in has been found to depend almost entirely on the rate of heating and it is not easy to record fhe same decomposition point on identical specimens. The resolution of dl-N-phenylalanine-4-arsinic acid was effected in aqueous solution 2 mols. of the dibasic Mid 2 mols. of sodium hydroxide and 2 mols. of brucine being used. The mmal brucine salt of the d-acid separated under the conditions desoribed on p. 484 and was optically pure after one recrystallisation from aqueous solution. From this pure brucine salt the pure d-acid was easily isolated.The pure l-acid was obtained by crystallising three times from water the crude acid liberated from the mother-liquor from which the above brucine salt had been removed the optically active acids being appreciably less soluble in water than the dZ-acid. The optically active monoethyl and monomethyl esters were prepared their rotatory powers being of the same sign as those of the optically active acids from which they were obtained. They have very much higher melting and decomposition points than the inmtive esters and their lower solubility especially of the methyl esters in water is very marked. The corresponding amide-acids were made by dis- solving the optically active methyl esters in aqueous ammonia their rotatory powers were of opposite sign to those of the esters (and of the optically active acids) from which they were prepared.When a solutioii of the monosodium salt of the optically active amide-acid was boiled with an excess of sodium hydroxide until ammonia ceased to be evolved the rotatory power of the remaining aqueous solution was of the opposite sign to that of the amide-acid taken but it was very much lower in value than that of the sodium salt of the pure acid in equivalent concentration. This was con- firmed in the case of both isomerides and it was evident that con- siderable racemisation had taken place during the cycle of operations d-dibaaic acid _~3 d-methyl ester -+ d-amide-acid (Zcevorotatory) 4d-dibasic acid. There was no evidence in the experimental work to indicate that racemisation had taken place during the conversion of the optically active acid into the corresponding methyl ester the latter only needed to be freed from mechanical impurities (one crystallisation) to be optically pure.It wits possible however that some racemis- ation had occurred during amide-acid formation (compare Wren J. 1909 95 1596) and possibly also during the hydrolysis of the amide-acid. GIBSON JOHNSON AND LEVIN The pure optically active amide-acids were prepared by the resolution of dZ-~V-phenylalanineamide-4-arsinic acid which was easily effected in aqueous solution with 2 mols. of the amide-acid and 1 mol. each of sodium hydroxide and quinine under the con- ditions described in the experimental portion.The quinine salt of the I-amide-acid needed only one crystallisation from water for complete purification and from this pure quinine salt the pure I-amide-acid was isolated. The pure d-amide-acid was obtained by recrystallising three times from water the crude d-amide-acid isolated from the mother-liquor from the above quinine salt of the I-amide-acid. This method of resolution of the dZ-ainide-acid is somewhat different from that employed by Pourneau and Nicolitch who used equimolecular quantities of the dl-amide-acid and quinine and appear to have had greater difficulties than have been en-countered in the present work in obtaining the pure optically active amide-acids. Attempted resolutions using the half-molecule method with brucine and strychnine were unsuccessful; in the latter case a crystalline strychnine salt was obtained but this proved to be a partial racemate.These optically pure amide-acids had distinctly higher rotatory powers than those obtained from the optically active methyl esters by the action of ammonia. When the pure I-arnide-acid (deztro-rotatory) was hydrolysed by boiling its aqueous solution with an excess of sodium hydroxide until ammonia ceased to be evolved the resulting solution was bavorotatory and the rotatory power was very much lower than that of the pure Iavorotatory dibasic acid in equivalent concentration. This proved that in the above cycle of operations racemisation takes place not only during the con- version of the optically active methyl ester into the optically active amide-acid of the opposite sign but also during the hydrolysis of the optically active amide-acid to the optically active dibasic acid again of the opposite sign.Since the dextrorotatory dibasic acid has been proved to be the parent substance of the Zmvorotatory amide-acid it seems aclvisablc to adopt Fischer’s system of nomenclature for configuration and sign of rotation. The above cycle of operations is therefore more completely expressed as follows D. d-Dibasic acid -+ D. d-Ester acid rammisation& /pwmitaticm) D. I-hide-acid The following is a rhumb of the optical rotation constants ([aJwsl) of the pure compounds described in the present coniinunication COMPOUNDS OF THE TRYPARSAMIDE TYPE.PART I. 483 Brucine sslt of D. d-dibasic mid (water) ............... -10.61" Quinine salt of D. Z-amide-acid (water) ............... -123.8 Dextro. Laevo. N-Phenylalanine-4-tinic acid (disodium salt water) + 56-40" -55-94' N-Phenylalanine-4-arsinic acid (ethyl ester) (alcohol) + 127.9 -125.8 N-Phenylalanine-4-arsinicacid (ethyl ester) (sodium salt water) ................................................ +lo30 -102.8 N-Phenyldanine-4-arsinir:acid (methyl ester) (sodium salt water)....................................... + 176.6 -116.3 N-Phenylalanineamide-4-arsinicacid (sodium salt water) ......................................................... -17.SS -1-16.47 E x P ic R IM EN TA L. cll-N-YhenyZuZanii~e-4-arsinic Acid (I).-A solution of 6-bromo-propionic acid (35 g.) in water (38 c.c.) was added to a hot solution of sodium p-aminophenylarsinate (" atoxyl " or '' soamin " con-taining 5H,O ; 50 g.) in water (165 c.c.) and the mixture boiled for 8 hours.The product crystallised in somewhat yellowish needles when the resulting solution was kept for 1640 hours in the ice- chest. dl-N-Phenyluhnine-4-arsinicacid,recrystallised from a large volume of boiling water decolorising charcoal being used to remove the small amount of colour formed colourless needles decomp. 207-210". 16.75G. were consistently obtained in every preparation (Pound As 25.9 26.1. C,H1,05NAs requires As 25.9%). The acid is soluble to the extent of about 0.5% in cold water and about 6% in boiling water. It dissolves readily in dilute mineral acids and in acetic acid; it is readily soluble in hot methyl and ethyl alcohols and moderately easily soluble in the cold solvents.It is also slightly soluble in acetone but insoluble in benzene and ether. In aqueous solution it reduces ammoniacal silver nitrate a silver mirror being produced and thus behaves similarly to o-and p-tolylglycines (Staats and Ehrlich Ber. 1883 16 204 ; &sack ibid. 1880 13 1091). The acid itself behaves on titration as slightly more than a dibasic acid (phenolphthalein) ; the end-point is however not well defined and varies with the dilution. Ethyl Ester of dl-N-Phenylalanine-4-arsinic dcid.-A mixture of the preceding acid (5.0 g.) absolute ethyl alcohol (30 c.c.) and concentrated sulphuric acid (0.5 c.c.) was boiled for 2 hours and poured into water (100 c.c.).The ethyl ester which separated on standing crystallised from dilute ethyl alcohol (30%) in colourless doubly-refracting prisms m. p. 175-177" (decomp.) (Found As 23.6. Cl1Hl,O5NAs requires As 2306%). The ester is readily soluble in ethyl alcohol and in hot water and behaves on titration as a slightly more than monobasic acid (phenolphthalein) the end-point not being sharply defined. Methyl ester of dl-N-phenylalanine-4-arsinic acid was prepared in an GIBSON JOHNSON AND LEVIN analogous manner to the ethyl ester pure methyl alcohol being used. After distillation of some 50% of the excess of alcohol the reaction product was ponred into a cold saturated aqueous solution of ammonium sulphate.The separated ester containing a little ammonium sulphate wa~ completely extracted with pure methyl alcohol the methyl alcohol distilled off and the residue crystallised from the minimum quantity of hot water. It was obtained in colourless acicular needles m. p. 181" (slight decomp.); yield 70% (Found As 24.3. CloH1,05NAs requires As 24.75%). This methyl ester is distinctly more soluble than the ethyl ester. dl-N-Phenylalanineamide-4-arsinic Acid (11).-This amide was prepared in two ways (a)The methyl ester (4 g.) was added gradually to an aqueous solution of ammonia (d 0.880; 12 c.c.) cooled in ice. It dissolved fairlyreadily and the solution was kept at the ordinary temperature for 48 hours the excess of ammonia then being removed by leaving the mixture over sulphuric acid under reduced pressure.The residue a thick gum was stirred with a little water and the amide precipitated by the addition of a slight excess of acetic acid. The product crystallised from boiling water in colourless needles m. p. 233-2440' (decomp.) (Found N 9-5; As 25.5. Calc. for C,H1,04N2As N 9.7; AS 26.0%). (b) a-Bromopropionamide (70 g.) was boiled with a solution of '' atoxyl "or " soamin "(containing 5H,O ; 115 g.) in water (325 c.c.) for 1 hour. The method of working up the product was similar to that described by Fourneau and Nicolitch (loc. cia. p. 1241); 75 g. were obtained m. p. 244" (decomp.) (Found N 9.4%) (compare Jacobs and Heidelberger Zoc. cia.).Attempted Reduction of dl-N-Phenylalanine-4-arsinicAcid.-Reduction of the arsinic acid with sodium hydrosulphite with or without magnesium chloride gave a yellow material soluble in alkaline solutions which contained much sulphur but too little arsenic for an arseno-compound. Reduction of a saturated solution of the acid in concentrated hydrochloric acid containing a trace of hydriodic acid at a low temperature with sulphur dioxide gave a crystalline product which could not be isolated on account of its ready solubility in the mixture at the ordinary temperature. Resolution of dl-N-Phenylalaniize-4-arsinicAcid.-To a boiling solution of the acid (70 g.) in 0~516N-sodium hydroxide (469.5 c.c.) and water (565-5 c.c.) brucine (113g.) was added and the boiling was continued until almost the whole had dissolved.Crystallisation began almost immediately and after standing at the ordinary temperature for 10 hours the crystalline material was filtered off washed with a little water and recrystallised from 2 litres of boiling COMPOUNDS OF THE TRYPARSAMIDE TYPE. PAST I. 485 water. After one further recrystallisation under similar conditions its rotatory power * in aqueous solution was constant c = 1.1446 I = 4 a = -0-48",whence [E] = -10.61". The brucine salt of d-N-phenylalanine-4-arsinicacid crystallises from water in large colourless plates containing 7H20 (Pound for air-dried material H20 10.6; N 6.0. C,H,20,NAs,2C2,H,,04N2,7H20 requires H20 10.5; N 5.8%. Pound for anhydrous material As 6.9.C,H,,05NAs,2C23H2,0,N2requires As 6.9%). d-and l-N-Phen~laEanine-4-arsinic Acids.-To obtain the d-acid the above brucine salt was ground three times with small quantities of a concentrated aqueous solution of ammonia the brucine filtered off and the filtrate extracted thoroughly with chloroform to remove any dissolved brucine. The aqueous solution after evaporation on the water-bath to about 100 C.C. was acidified with hydrochloric acid (Congo red) and the precipitated acid filtered off from the cold mixture. After one recrystallisation from water its rotatory power was constant and it was obtained in colourless needles m. p. 220-221" (decomp.) (Found As 26.0. C9HI2O,NAs requires As 25.9%). For its rotatory power determination the acid was treated with the calculated quantity of a standard aqueous solution of sodium bicarbonate to form the normal salt and this solution made up to volume with water c = 0.8562 I = 4 a = +1.93" whence [a]= +56.40".The 1-acid was isolated from the mother-liquor remaining after the separation of the above brucine salt of the d-acid. The solution was made alkaline with aqueous ammonia the brucine filtered off and the solution extracted thoroughly with chloroform and evaporated to small bulk on the water-bath. The l-acid was precipitated with hydrochloric acid and after three crystallisations from water its rotatory power was constant. It resembled the d-acid in appearance and had the same decomposition point (Pound As S5.S~/0). As in the case of the d-acid its rotatory power was determined in aqueous solution containing exactly sufficient sodium bicarbonate to form the disodiurri salt c = 0.8942 I -;4 --2.00" whence [a]= -55.94".15.5 G. of the pure d-acid and 10.0 g. of the pure Z-acid were obtained. The partly resolved acid recovered on working up the mother-liquors was submitted to further resolution the quantities of brucine etc. employed being adjusted in accordance with the rotatory power of the partly resolved acid. Ethyl Esters of d- and 1-N-Phenylalanine-4-arsinic Acids.-These were prepared from the corresponding optically active acids by the * All rotatory power determinations were carried out at 20" with the mercury-green(A = 6461) line. S GIBSON JOHNSON AND LEVIN method used for the preparation of the ethyl ester of the dl-acid.Being somewhat less soluble they are even more easily' obtained. In appearance they resemble the dl-ester but their decomposition point is very much higher 275-276" [Found (for the d-ester prepared from the d-acid) As 23.5; (for the 1-ester prepared from the l-acid) As 23.75. CllH1,O,NAs requires As 23.6%]. The following rotatory power determinations were made in pure ethyl-alcoholic solutions d-ester c = 0.4464 I = 4 a = +2.28" whence [a]= +127-9". Lester c = 0.7132 I = 4 a = -3.50" whence [a]= -125.8". The following rotatory power determinations were made in aqueous solutions containing the exact quantity of sodium bicarbonate to form the sodium salt d-ester c = 0.4118 I = 4 c( = +1.70" whence [a]= +103-0".Lester c = 0.3960 I = 4 a = -1~63"~whence [a]= -102.8". Methyl Esters of d-and 1-N-Phenylalanine-4-arsinic Acids.-These were made in a similar manner to the methyl ester of &I-N- phenylalanine-4-arsinic acid. A preliminary experiment indicated that the optically active ester is very much less soluble in water than the optically inactive one ; the esterification mixture therefore was poured into water instead of into a saturated solution of ammonium sulphate. After one recrystallisation from water the esters were optically pure a yield of over 90% being obtained. The two esters crystallised in long colourless needles m. p. 277-278" (decomp.). They are very sparingly soluble in water and are conveniently recrystallised from boiling water (6 g.to 200 c.c.) [Found (d-ester) As 25.1 ; (1-ester) As 24.6. CloHl,O,NAs requires As 24*75%]. The rotatory power determinations were done in aqueous solutions containing the calculated quantity of sodium bicarbonate to form the sodium salt d-ester c = 0.6194 Z = 4 a = +2.92" whence {a]= +117.6". 1-ester c = 0,4982 I = 4 a = -2.32" whence [E] = -116.3". Each of these optically active methyl esters was converted into the corresponding optically active amide-acid under precisely the game conditions its those employed in the preparation of the dE-amide-acid from the dl-methyl ester. In each case the product was recrystallised once from water and obtained in colourless needles m. p. 242-243" (decomp.) [Found (amide-acid from d-methyl ester) N 9.9; (amide-acid from I-methyl ester) N 9.6.Calc. for C,H,,04N~s N 9*7%]. The rotatory powers were determined in tLyueous solution con- COMPOUNDS OF THB TRYPARSAMIDE TYPE. PART I. 487 taining the exact quantity of sodium bicarbonate to form the sodium salt (a) Amide-acid from d-methyl ester c = 0.6114 1 = 4 a = -0.33" whence [a]= -13.3". (b) Amide-acid from I-methyl ester c = 0.6536 I = 4 a = +0*36O whence [a]= +13-9". The solution used in (a) (40 c.c.) was treated with sodium hydr- oxide (1 g. in 2 C.C. of water) and boiled for 10 minutes; ammonia then ceased to be evolved. The resulting solution was cooled and made up to 50 c.c. and its rotatory power determined. It had [a]= f-26.6".The solution used in (b) (40 c.c.) was treated in an exactly similar manner and the final rotatory power was [a] == -29.5". Resolution of dl-Phenylu.Zuninearnide-4-arsinic Acid .-The amide-acid was prepared by the method described by Jacobs and Heidel- berger (Zoc. cit. ; compare Fourneau and Nicolitch Ioc. cit.). The amide-acid (36 g.) was dissolved in boiling water (2000 c.c.) and aqueous sodium hydroxide (0-516N; 121 c.c.) and quinine (22.3 g. allowing 10% excess on account of water content) added. The whole was brought into solution and allowed to cool during 20 hours. The salt crystallised in soft needles and was separated from the mother-liquor. The salt was recrystallised from boiling water (2000c.c.) and another recrystallisation from water (1500 c.c.) failed to affect the rotatory power.The salt was anhydrous (Found N 9.1 ; As 12.2. Calc. for C,H,,04N2As,C20H2402N2 N 9.15; As 12.3%). The rotatory power was determined in aqueous solution c = 0.2412 I = 4 a = -1.20" whence [a]= -123.8". 1-and d-N-Phenylalanineamide-4-arsi.nic Acids.-The above quinine salt was decomposed by grinding with aqueous ammonia (d 0-SSO) and the liquid filtered. The separated quinine was re-extracted twice with further quantities of ammonia solution. The filtrate was acidified with concentrated hydrochloric acid (Congo red) and the precipitated amide-acid recrystallised from boiling water (decolorising charcoal) ; its rotatory power was then constant. The pure I-amide-acid crystallised in colourless needles m.p. 247" (decomp.) (Found N 9.4. Calc. for C9HI8OPN2As N 9.7%). Its rotatory power was determined in aqueous solution containing the exact quantity of sodium bicarbonate to form the sodium salt c = 0.5868 1 = 4 a = -0*42" whence [a] = -17*88O. The mother-liquor after separation of the above salt was evapor- ated to about half its original volume and the quinine precipitated by addition of a slight excess of an aqueous solution of ammonia. SUGDEN THE TERNARY SYSTEM After filtration the liquid was acidified with concentrated hydro- chloric acid (Congo red) and the amide-acid precipitated as before. This was recrystallised three times from boiling water ; its rotatory power was then constant. In appearance and general behaviour the pure d-amide acid was similar to its enantiomorphous isomeride and it had the same melting-decomposition point (247") (Found N 9.5%).Its rotatory power was determined as in the previous case c = 0-7770 I = 4 a = +0.51" whence [a]= 3-16-5". Before the effect of hydrolysis on these pure optically active amide- acids was examined the rotatory power of another specimen of the pure d-amide- acid was determined under the same conditions as the above c = 0.7050 I = 4 a = +0.46" whence [a]= +16.4". This solution (40 c.c.) was treated in an exactly similar way to those in the previous experiments on the hydrolysis of the optically active amide-acids the rotatory power of the final solution was [a]= -28.5". The authors wish to express their grateful thanks to the Govern-ment Grant Committee of the Royal Society and to Messrs.Imperial Chemical Industries Ltd. for grants which have been of great assistance in the purchase of chemicals and apparatus. One of them (B.L.)is indebted to the Department of Scientific and Industrial Research for a grant which enabled him to take part in the work. GUY'S HOSPITAL (UNIVERSITY MEDICALSCHOOL OF LONDON), LONDON,S.E.1. [Received,January 26th 1929.1