NIERENSTEIN CATECHUTANNINS. PART I. 23 IV.-Catechutannins. Part I . Paullinia Tannin. By MAXIMTT~AN NIERENSTEIN. IT has recently been suggested by Freudenberg (Ber. 1920 53, [B] 1417; compare Lowe J. pr. Chem. 1868 [i] 105 33 75; Etti Nonutsh 1889,10,467) that the catechutannins are amorphous dehydration products of catechin. As evidence in support of this view Freudenberg quotes Kirmsse (Arch. Phurm. 1895 236 122), who during the course of an investigation of guarana a paste obtained from the seeds of Puulli~zia cupuna H. B. & K. was able to isolate a crystalline catechin in addition to amorphous guarana tannin.* A deduction based on this observation cannot * For a summary of the literature of catechutannins in general and of guarana and paullinia tannins in particular see Perkin and Everest “The Natural Organic Colouring Matters,” 1918 and Dekker “ Die Gerbetoffe,” 19 13 respectively 24 NIERENSTEIN CATECHUTANNINS.PART I. be taken seriously since the manufacture of guarana involves a series of vigorous processes including the roasting of the seeds (compare Planchon and Collin ‘‘ Les Drogues simple d’origine v&g&tale,” 1898 2 576). It is therefore very probable that the catechin found by Kirmsse in guarana is a disintegration product of guarana tannin. The results obtained by the author are in every respect opposed to Freudenberg’s theory since paullinia tannin is found to be a crystalline substance consisting of one molecule of dextrose and two molecules of gambier-catechin * carboxylic acid. The latter acid is not identical with the synthetic gambier-catechin carboxylic acid previously described (Annalen 1913 396 194).As these two acids are probably isomerides it is proposed provisionally, to employ the prefix cc for the synthetic acid and for the acid occurring in paullinia tannin. The present investigation shows that the two molecules of P-gambier-catechin carboxylic acid in paullinia tannin form a depside (compare Fischer Ber. 1913 46 3253) since the methylo-deriv-ative prepared by the action of diazomethane on paullinia tannin, yields tetramethyl glucose on hydrolysis (compare T. 1921 119, 284). As paullinia tannin contains a free carboxyl group and when hydrolysed with emulsin yields both P-gambier-catechin carboxylic acid and dextrose it becomes also evident that this tannin is a normal glucoside (compare Fischer and Strauss Ber.1912 45, 2467) and not an acyl derivative of glucose as assumed by Fischer in the case of gallotannin. Our incomplete knowledge of gambier-catechin makes it impossible to discuss the constitutions of p-gambier-catechin carboxylic acid and paullinia tannin hence a t present an empirical formula only is assigned to the latter : QeC,H1l 0 5 C,,H,O( OH) ,*CO*O-C 15HsO( OH),*CO,H It should be noted that the methylo-derivative of paullinia tannin, on hydrolysis yields in addition to tetramethyl glucose penta-methoxy- and trirnethoxy- p-gambier-catechin carboxylic acid. The * The catechin found by Kirmsse in guarana is obviously gambier-catechin although he gives no melting point for his product (compare A.G. Perkin and Yoshitake T. 1902 81 1163 1173). Freudenberg’s suggestion (“ Die Chemie der natiirlichen Gerbstoffe,” 1920) of the special name paullinia-catechin for this substance is consequently not required. t The graphical formula previously assigned to a-gambier-catechin carb-oxylic acid (Zoc. cit. p. 195) is withdrawn as it was based on the catechin formula of Kostanecki and Lampe (Ber. 1906 39 4007) which has since become untenable (T. 1920 117 973) NIERENSTEIN CATECHUTANNINS. PART I. 25 latter acid gives a pronounced green coloration with alcoholic ferric chloride thus showing that in paullinia tannin the catechol nucleus in one molecule of p-gambier-catechin carboxylic acid is combined directly with the other molecule of p-gambier-catechin carboxylic acid and also with dextrose.E X P E R I M E N T A L . Preparation of Paullinia Tannin. The seeds of Paullinia cupana are extracted in a Soxhlet apparatus with (1) ether and chloroform to remove chlorophyll and fats and ( 2 ) ethyl acetate. The tannin is removed from the latter extract by shaking with a solution of sodium hydrogen carbonate saturated with carbon dioxide. The acidified aqueous solution is again extracted with ethyl acetate the extract dried over anhydrous sodium sulphate and the residue obtained on evaporation dissolved in hot water. The cold solution is shaken with 3 to 5 grams of fat-free caseinogen in order to remove any amorphous tannin (compare T. 1919 115 1330). The filtrate to which a few drops of chloroform are added is allowed to stand in the cold.Crystal-lisation commences after three to four days paullinia tannin separating in small needles. These are filtered and recrystallised several times from water chloroform being added each time. The average yield from 500 grams of the seeds is 12 grams which may be increased to 16 grams on concentrating the filtrates (Found * : C,,H,,0,8*C0,H requires C = 56-2 ; H = 4-4 ; C0,H = 5.5 per cent. Found fM in alcohol = 795,820,516 843 ; M in water = 956, Paullinia tanizin crystallises from water in gray-coloured needles, which melt a t 199-201" carbon dioxide being evolved. At 130°, i t loses two molecules of water of crystallisation (Found : H,O = 4.2 4.3. Calc. H,O = 4.2 per cent.) ; the anhydrous product melts a t 259-261" with evolution of carbon dioxide and decomposition.The tannin is soluble in alcohol ethyl acetate, or acetio acid but not in any other organic solvent. Paullinia tannin gives all the typical colour tests of the catechutannins and C = 55.9 56.3 56.0; H = 4.6 4.4 4.5; C0,H 'f = 5-7 5.6 5.3. 872 840 536. Calc. fM = 812). * Dried at 130". t By the method of Baumann and Kux (2. anal. Chem. 1893 32 129). Preliminary experiments with phloroglucinolcarboxylic acid protocatechuic and 2 3-dihydroxybenzoic acids gave good results by this method which was however found not to be trustworthy in the case of gallic acid and pyrogallolcarboxylic acid 26 NIERENSTEZN CATECHUTANNMS. PART I. is precipitated by gelatin * and alkaloids. The crystalline sodium and potassium salts are obtained by neutralisation of the tannin with the corresponding carbonates.They crystallise from water ; they do not melt below 320" (Found Na = 3-2,2*9. Calc. Na; = 2.7 per cent. Found K = 6.1. Calc. K = 5-7 per cent.). The dextrose content of paullinia tannin is estimated in the same manner as in gallotannin (T. 1921 119 279) (Found dextrose = 21.6, 21.8 22.3 21.9. Calc. dextrose = 22-2 per cent.). The dex-trosazone (m. p. and mixed m. p. 202-204") was also prepared (Found N = 15.9. Calc. N = 15.6 per cent.). Paullinia tannin gives the following optical data t [a]: - 74.4" (10 per cent. solution in water) ; [a?; - 39.1" (8 per cent. solution in alcohol) ; [a]:; - 48-1" (6 per cent. solution in acetone) ; [a] - 56.8" (8 per cent.solution in pyridine). All solutions exhibit muta-rotation which is most evident in the case of the pyridine solution. The rotation falls from [a]" - 56.8" to - 11.4" in sixteen minutes and to - 5.6" in forty minutes when it becomes constant. Action of Emulsin on Paullinia Tannin. To a solution of 10 grams of paullinia tannin in 200 C.C. of water 0.3 gram of emulsin (from bitter almonds) in 75 C.C. of water and 5 C.C. of chloroform are added. The solution on keeping a t 37" for eight days deposits a mass of crystals which are filtered off. The filtrate freed from unchanged paullinia tannin with the aid of lead acetate and hydrogen sulphide contains dextrose which is identified as the dextrosazone (m. p. and mixed m. p. 201-203"). The solid formed by the action of emulsin crystallises from water in small pointed needles and is optically inactive p-gambier-catechin carboxyEic acid which melts at 252-253" carbon dioxide being evolved (Found j C = 57.3 57.5; H = 4-8 4.4.C,,H,,O, requires C = 57-5; H = 4.2 per cent.). The acid is soluble in alcohol ethyl acetate or acetic acid but not in any other of the usual organic solvents. On heating an alcoholic solution of the acid with pyridine for several hours (compare T. 1916,109,593) gambier-catechin is readily obtained. The latter on crystallisation from distilled water melts a t 175-177" which is the melting point * Freudenberg'B statement (Ber. 1920,53 [B] 236) that gambier-catechin is precipitated by gelatin is not correct as repeated observations in this laboratory have shown.His observation is also contrary to all previous data on this point (compare for example Procter " Leather Industries Laboratory Book," 1908 p. 130). j- I 2 4.4 for all optical observations given in this paper. $ Dried at 100" in a vacuum NIERENSTEIN CATECHUTANNINS. PART I. 27 given by A. G. Perkin and Yoshitake (T. 1902 81 1163) for gambier-catechin. This melting point is not depressed when the substance is mixed with authentic gambier-catechin (Found * : C = 62.0; H = 4-9. Calc. C = 62.1 ; H = 4.8 per cent.). The acetyl derivative was also prepared. It melted correctly a t 126125" and showed no depression when mixed with authentic penta-acetyl-gambier-catechin. Diazomethane converts p-gambier-catechin-carboxylic acid into the corresponding pentamethoxy methyl ester which crystallises from light petroleum in long needles melting a t 74" (Found t : C = 63.3; H = 6.5.C,,H,,O requires C = 63.2; H = 6.2 per cent.). When oxidised with potassium permanganate in alkaline solution veratric acid is obtained (m. p. and mixed m. p. 179-180') (Found C = 59.1 ; H = 5.8. Calc. C = 59.3; H = 5.5 per cent.). I n this connexion it is interesting to note that the methyl ester of pentamethoxy-a-gambier-catechin carboxylic acid (m. p. 92") previously described (Eoc. cit.) gives under similar con-ditions m-hemipinic acid and not veratric acid. The inactive p-gambier-catechin carboxylic acid yields a crystal-line strychnine salt long pointed needles from water which melt at 222-223" (Found N = 4.5. Calc. N = 4.2 per cent.).The quinine salt however was found to be the most suitable for the resolution of the acid. The two optically active forms are obtained on fractional crystallisation the salt of the d-form being less soluble than that of the 1-form. The free acids are obtained by decompos-ing thc qiijnine salts with sodium hydroxide in the cold. d- ~-G'umtier-catehin carboxylic acid crystallises from water in microscopic needles which melt a t 249-251" carbon dioxide being evolved (Found $ C = 57.6 ; H = 4.5. Calc. C = 57.7 ; H = 4-2 per cent.). The rotatory power in water (5 per cent. solution) is [ctK + 12.6"; in alcohol (7 per cent. solution) [a]r + 17.6". I-p-Gambier-catechin carboxylic acid crystallises from water in fairly large needles which melt a t 258-261" carbon dioxide being evolved (Found $ C = 57-3; H = 4.3.Calc. C = 57.5; H = 4.2 per cent.). The rotatory power in water (5 per cent. solution) is [ c c ] ~ " - 22.4'; in alcohol (6 per cent. solution) [ a ] r - 31.6". Methylated Paullinia Tannin. Thc methylo-derivative is obtained in excellent yield by the action of diazomethane on an ethereal suspension of paullinia tannin. It crystallises from dilute alcohol or acetone in small needles which melt a t 126-127" [Found t C = 60.2 60.3 60.1 ; H = 6.6 6.5, t Dried in a vacuum over paraffin. * Dried at 100'. $ Dried at 130' 28 NIERENSTEIN CATECHUTANNINS. PART I. 6.4 ; OMe = 43-2,43*0 43.1. C38H2307 (OMe)13 requires C = 60.6 ; H = 6.3; OMe = 43.1 per cent.]. It is soluble in all the usual reagents with the exception of water.The rotatory power in alcohol (5 per cent. solution) is [alga - 31.8" ; in benzene (8 per cent. solution) [a]:6' - 26.4" ; in chloroform (6 per cent. solution) [a]:' - 37.7"; in s-tetrachloroethane (7 per cent. solution) Hydrolysis of Methylated Paullinia Tannin.-Eight grams of the methylo-derivative are hydrolysed in a sealed tube with 100 C.C. of glacial acetic acid as described in the case of methylafed gallo-tannin (T. 1921 119 284). The semi-crystalline solid formed consists mainly of trimethoxy-p-gambier-catechin carboxylic acid, which crystallises from alcohol in small needles melting at 174-175" (Found * C = 60.5; H = 5.5. C1,H2,0 requires C = 60.6; H == 5.3 per cent.). On methylation with diazo-methane it is quantitatively converted into methyl pentamethoxy-(3-gambier-catechin carboxylate (m.p. and mixed m. p. 74"). The acetic acid filtrate gives on addition of water a bulky precipitate, which consists mainly of pentamethoxy-P-gambier-catechin carboxylic acid. It crystallises from alcohol acetic acid or ethyl acetate in pointed needles which melt a t 196-197" (Found * C = 62.2; H = 6.3. C2,M,,08 requires C = 62.4; H = 6.0 per cent.). Diazomethane converts it quantitatively into the methyl ester (m. p. and mixed m. p. 74"). The remaining filtrate contains tetramethyl glucose which is isolated in exactly the same manner as used in the case of methylated gallotannin. It melts at 91-93" and the melting point is not depressed by admixture with authentic tetramethyl a-glucose (Found OMe = 52-2. Calc. OMe = 52.6 per cent.). [E]F - 47.1". This investigation has with many interruptions been in progress since 1908 during which time the author has had the collaboration of Messrs. C. W. Spiers M.Sc. and R. Barr B.Sc. for which he wishes to thank them. He also wishes to acknowledge the help of the late K. C. R. Daniel B.Sc. The author is also indebted to the Director of the Servica Sanitario of Maneos North Brazil for a liberal supply of material used in this investigation and t o the Colston Society of the University of Bristol for a grant in aid of this research. THE UNIVERSITY BRISTOL. [Received November 22nd 1921.1 * Dried at 130"