LECITHIN. I. HYDROLECITHIN AND ITS BEARING ON THE CONSTITUTION OF CEPHALIN. BY P. A. LEVENE AND C. J. WEST. (From the Laboratories oj The Rockefeller Institute for Medical Research.) (Received for publication, December 1, 1917.) The recent investigations on the chemical structure of lecithin have resulted in many important contributions, all of which point to the correct,ness of the generally accepted view of its molecular structure, H,C 0 CO &H~~ I HC I 0 COC,,HsI H&OP = 0 OH I 0 CHz CH, N = (CH,), However, a scrutiny of all the work on lecithin reveals a remarkable incompleteness of each individual investigation. A rigorous proof of the accepted theory requires for lecithin an element,ary composition of C = 65.60, H = 10.79, N = 1.74, P = 3.86. It further requires that the nitrogen of the molecule should be composed entirely of choline. Hence, lecithin should not contain even a part of its nitrogen in the form of free apino groups. Still further, it requires a proof of the identity of the fatty acids with those accepted by theory, and finally it requires the isolation of the glycerophosphoric acid. The work up to the present has satisfied many of the requirements. The fatty acids and the glycerophosphoric acid have been ident.ified; and MacLean has prepared, at least once, a sample of lecithin that was free of amino nitrogen. However, this one sample has been incomplet,ely analyhed. All other samples of lecithin prepared by various workers contained amino 111
112 Lecithin. 1 nitrogen in their molecule, and from the standpoint of elementary analysis showed a marked disagreement with the theory. This is well illustrated by the following table. Author. SOUW3. c H N P ( ~holinc. I I Thudichum.... Baskoff... Hefftera... Stern and Thierfelder.... MacLeans.... Erlandsenc... MacLean.... Eppler*.... MacLean.......... Brain. Liver. Egg. I Heart. I Not given. From CdC12 salt. 1, er cent I per cent I per cent I / 64.63 10.96 1.79 64.18 10.60 1.87 66.29 10.17 1.87 66.27 10.32 J.85 66.46 10.69 1.87 1.89 1.85 1.87 wr crrr/ 4.00 4.00 3.95 3.95 3.95 3.97 4.03 4.04 4.00 4.15 66 42 41.4 68 66 98.7 The analytical data obtained by Ritterg on hydrolecithin showed better agreement with the theory, and one might be inclined to regard the material of Ritter as such that it contained all the necessary proof in favor of the conventional theory. Unfortunately, Ritter did not determine the amino nitrogen content of the reduced lecithin, and hence failed to furnish definite proof of its purity. Indeed, the present report contains data unmistakably proving t,hat hydrolecithin of an elementary composition fully harmonizing with the theory may be and generally is impure, containing between 10 and 20 per, cent of its nitrogen in the form 1 Thudichum, J. L. IV., The Chemical Constitution of the Brain, London, 1884. 2 Baskoff, A., Z. physiol. Chem., 1907, Ivii, 395. * Heffter, A., Arch. ezp. Path. u. Pharm., 1891, xxviii, 100. * Stern, M., and Thierfelder, H., Z. physiol. Chem., 1907, liii, 381. &MacLean, H., Z. physiol. Chem., 1908, Iv, 360; 1909, lix, 223; Biochem. J., 1909, iv, 38, 240. 6 Erlandsen, A., Z. physiol. Chem., 1907, li, 71. 7 MacLean, Biochem. J., 1915, ix, 364. 8 Eppler, J., Z. physiol. Chem., 1913, lxxxvii, 241. 0 Ritter, F., Ber. them. Ges., 1914, xlvii, 530. Cf. Reidel s Ber., 1913, Ivii, 20; 1914, Iviii, 15.
I. A. Levene and C. J. West 113 of amino nitrogen. Thus the task of the preparation of lccit,hin having a composit,ion required by the t.heory and at the same time free of impurities has not yet been accomplished. l?fforts in this direction are now in progress in this laboratory. However, the present finding has a great significance because of its bearing on the structure of cephnlin, and the work is presented in its present incomplete state because of t,his significance. The retnarks made earlier in t,his communication regarding lecithin apply also to cephalin. On the basis of recent work on the hydrolytic products of the substance, a certain structural formula has been assm~led. This formula requires an elementary composition of C = 66.17, H = 10.57, N = 1.88, and P = 4.1 7. However, all samples analyzed, beginning with Tl~uclichum and up to the prcscnt by the most recent investigators,10 consistently had the average composition of C = 60.00, H = 9.30, N = 1.80, and I = 3.80. On the basis of these considerations, one may argue that if cephalin and lecithin both had the composition required for them by the theory, then a mixture of the two should possess practically t.he same clemcntary composition as either one of them in the pure state. On the other hand, if lecit.hin possessed t.he composition assumed by t,he theory and cephalin that found empirically, t,hen a mixture containin, (r 80 per cent of the former and 20 per cent of the latter shoultl psscss :L carbc,n cont,ent of 64.56 per cent. instead of 65.35 per ccni. Conversely, if a mixture of the two rctluced subst.anc.cs pcsso.; (~1 an clonlentary analysis of C = 65.30, II = 11.20, N = 1.75, 1 = 3.85, it would justify the conclusion that both lecithin ant1 caephalin possess the composition assmnctl for t)lrcm by the theory. The material unalyzcld by us contained 80 per cent of lecithin and 20 per cent of an impurity. It was found that, the material yielded on hydrolysis ljesides the choline also ihe base aminoethanol. Hence it was reasonal)lc to assume that, the 20 per cent of impurity consisted of cephalin. If cephalin had the composition found by experience then a substance consisting of 80 per cent of hydrolecithin and 20 per cent of cephalin should have an elementary corn posit ion of C = 64.56, II = 10.4!1,!V = 1.75, P =
114 Lecithin. I 3.84. On the other hand, if both lecithin and cephalin possess the struct,ure assigned to them by theory then the above mixture should have the elementary composition found by experiment. Thus the facts prcsent.ed in this report furnish evidence in favor of the prevailing theory of the molecular structure of lecit.hin and of cephnlin; they also indicate the method by which the pure reduced cephalin may eventually be obtained. ICfforts in this direction are now in progress. In this connection it may be recalled that the product obtained on reduct,ion of cephalin with hydrogen ga,s in the presence of palladium contjained fiyjll1 6:! to 63 per cent of carbon in its molecule, thus approaching nearer than the nonreduced cephalin the theoretical value. The nonreduced cephnlin generally obtaincd is undoubtedly an altered and perhaps oxidized form of the original substance. The nature of the alteration is not known :1s yet. F XI EIIIBfENTAL. 1 Hydrolecithin was first, prepared by l aal and Oehme;l! they reduced an alcoholic solution of egg lecit,hin with hydrogen and colloidal palladium. The product was obtained as microscopic compact q&as, which sintered at 8384, and decomposed over 150 with blackening. Upon hydrolysis the hydrolecithin gave a mixture of stenric, pnlmitic, and probably myristic acids. The following year Ijitter reduced lecithin which had been prepared from fresh, dry egg yolk as foliows: The egg yolk was first extracted with petroleum ether and then with ether; the ethercal extract was concentrated and the residue was then extracted with methyl alcohol. The hydrolecithin prepared in this way proved t:, IX distearylhydrolecithin; that is, on hytlrolysis it yielded only stearic acid. Hydrolecithin has been prepared in our laboratory, by Pnal s method, from various fra.ct,ions of egg lecithin, and from lecil hin of brain and other organs. During the course of the work it has been noted that, lecithin, which has been washed according to MacLc:m si method, in which the lecithin is ground up wit,h a little water and precipitated with acetone, is reduced more rapidly and more completely than unwashed lecithin. Also, it was 11 Paal, C., and Oehme, H., Ber. chem. Ges., 1913, xlvi, 1297.
P. A. Levene and C. J. West 115 found that the addition of 1 to 2 per cent of acetic acid to the alcoholic solution facilitated the reduction. With fairly concentrated solutions the hydrolecithin separates out during the course of the reduction. This product was brought into solution by warming, filtered from the coagulated palladium, cooled to O, and the material which separated out, crystallized from dry methyl et,hyl ketone until the composition was constant. In most cases it was easily possible to obtain material with correct analytical figures for carbon and hydrogen. In some cases, however, this was impossible, even after repeated crystallizat,ion. Hydrolecithin crystallizes well from methyl ethyl ketone, in which it is insoluble in the cold; it softens between 80 and 90, turns brown about 100, starts to melt about 200, and runs down the tube, giving a dark red liquid, at 235. The figures vary somewhat, depending upon the rate of heating. The optical activity was determined in chloroform solution and varied between + 5.2 and + 5.4. The presence of the amino nitrogencontaining body did not seem to affect the value for [a]:. Sample 400 [a]: = SE = t5.4.. Analysis of these samples gave the following values: Samrh c: H N P 1 Ash. ~~ pm WILL per cent prr cent pw cent per cent 400 65.20 10.89 2.00 4.00 399 65.59 11.16 2.37 3.85 429 65.60 11.03 1.96 3.79 508 65.50 11.30 1.80 3.90 492 1.98 3.98 Theory. 65.37 11.23 1.74 3.84! I I I I 10.61 10.38 9.75 9.66 NH?Pi 7.00 20.00 6.00 0
116 Lccit,hin. I Hydrolysis of Hydrolecithin. A large quantity (200 gm.) of hydroiecithin was prepared for the purpose of studying its hydrolyt ic products. This analyzed as follows : 100 gl11. of this material were hydrolyzed by boiling with 1 liter of 3 per cent sulfuric acid for 8 hours. The fatty acid fraction was filtered off and recrystallized from acet.one, after boiling about 2 hours with animal charcoal. Again recrystaiiized from acetone, the acid melted at 6%TO, and on combustion and (.itrat,ion gave figures for stearic acid. This confirms llitter s observation that. it is possible to obtain pure distearylhydrolecithin. The aqueous filtrate was freed from sulfuric acid by the addition of barium hydroxide, concent.rated in IXKXO, precipitat.ed with basic lead ncetatc, the filtjrnte freed from Iead, and used for the determination of aminoethyl :~lcol~ol according to Thierfelder and Schulze. The ctheresl extract was concentrated, taken up in water, and the amino nitrogen determined. The theoretical amount, of gold chloride was added to the acidified solution. The gold salt separated as long needles after standing 2 days in a desiccator ovet sulfuric arid. It melts at 184186. TrierI gives 186187 for the gold chloride salt of aminoethyl alcohol 12 Thierfelder, H., and Schulae, O., %. physiol. Chem., 1916, xcti, 196. This method depends upon the fact that calcium oxide does not liberate choline from its hydrochloritlc, but does free aminoethyl alcohol. The concentrated solution of the mixed hydrochlorides is rubbed up with pure calcium oxide until it is :L dry povder and extracted with ether in a Soxhlet apparatus; the flask should contain 0.1 x sulfuric acid to bind the base, ot,herwise considerable loss occurs. lifter 27 hours about 96 per cent of the alcohol has been extracted by the ether. The choline may then bc extracted wi6h hot, alcohol. 13 Trier, G., %. physiol. Chem., 1911, lxxiii, 383; 191112, lxxvi, 496.
P. A. Levene and C. J. West 117 hydrochloride, and I<norr14 gives about 190 for the synthetic product. It was analyzed by heating to constant weight. 0.2008 gm. substance gave 0.0990 gm. Au. 0.2075 0.1016 Au... Calculated: 49.li Found: 49.35 48.96 I4 Know, L., Bcr. them. Ges., 1597, xxx, 913.