s ner daete gh ad ee Caled Ss pg tererts ede Sj ambnae 1 Ad tet ye d Gham a if Cee =i oF Poet ie ot ee ae Ast oy ge tof Lr Paes vweaiane yi an i PO Me ite ry ots beok hmm We? rete 45 Seeite ‘ Sit Big. gs55 has Say * gee seh saschs 4 may iis G eh Tae P) Lethe 9 24 9) eevee is : ; sae ie ery seat cigs as Puy ie, a: + eats? 4 ray cata we ney ey as tig pls peti uate bei ve oaearas say ui Sif: ape > Grate rents gt Lean em noes CASALE Grersre oy yom Vetre habe | » er bve se 5 ae adrian ee Eo oo poses 624 Siew . Memo ye aed Fs ratte iar a 07 eee mm abide ii iebsadl pe Bea fetta astra) sor | Sasst3 Tie hen + pes Popeae ot. z gts iste Set a 7 BA mt % ates ehh Sa art Sele she. Datesasn 4 s2% a ase Fitiwecs A om me a by aes ereaar ; Sen aratape as: sDeatestel Reoeeas Shi hcreode Cae wpe al Fad ee Fintees le F + + Fer RMe gh) terete erect - ets sats Peeters) ‘ THE JOURNAL \A OF BIOLOGICAL CHEMISTRY FOUNDED BY CHRISTIAN A. HERTER AND BUSTAINED IN PART BY THE CHRISTIAN A. HERTER ? MEMORIAL FUND OFFICIAL ORGAN OF THE AMERICAN SOCIETY OF BIOLOGICAL CHEMISTS EDITED BY STANLEY R. BENEDICT, New York,N. Y. LAFAYETTEB. MENDEL, New Haven, Conn. HENRY D. DAKIN, Scarborough, N. Y. DONALD D. VAN SLYKE, New York, N. Y. 2 WITH THE COOPERATION OF THE EDITORIAL COMMITTEE _OTTO FOLIN, Boston, Mass. WALTER JONES, New York, N. Y. L. J. HENDERSON, Cambridge, Mass. GRAHAM LUSK, New York, N. Y. ANDREW HUNTER. Toronto, Canada THOMAS B. OSBORNE, New Haven, Conn. WALTER W. PALMER, New York, N. Y. A. N. RICHARDS, Philadelphia, Pa. L. L. VAN SLYKE, Geneva, N. Y. [qr ss VOLUME LI BALTIMORE 1922 CopyrnricurT 1922 BY THE JOURNAL OF BIOLOGICAL CHEMISTRY PUBLISHED BY THE ROCKEFELLER INSTITUTE FOR MEDICAL RESEARCH FOR THE JOURNAL OF BIOLOGICAL CHEMISTRY, INC. WAVERLY PRESS Tae Wittiams & Wimixins CoMPpANY Battimorn, U.S. A. CONTENTS OF VOLUME LI. No. 1. March, 1922. Cuark, E.P. Note on the preparation of mannose.................. MarsHatL, E. K., Jr. The effect of loss of carbon dioxide on the hydrogen ion concentration of urine......................00002 NevuwirtH, Isaac. Studies in carbohydrate metabolism. III. A study of urinary sugar excretion in twenty-six individuals...... FRIEDEMANN, W.G. The nitrogen distribution of proteins extracted by 0.2 per cent sodium hydroxide solution from cottonseed meal, Beem DEAN, andthe COGOMUT... 2... 2.52.2... eee eee eceases Wu, Hsien. Separate analyses of the corpuscles and the plasma.... Wu, Hsien. A new colorimetric method for the determination of plasma proteins...... ES RAEI s n.d . sae ee eee 213 WitiaMAN, J. J. The preparation of inulin, with special reference to artichoke tubezsas SOURCE. ..)1. .s «a+: 4h eee eee 275 LEvENE, P. A., and Simms, H.S. The unsaturated fatty acids of liver lecithin... .... es. ces -s nas ss ow os]e a5 oe eee 285 GaMBLE, JAMES L. Carbonic acid and bicarbonate in urine........ 295 No.2. April, 1922. - Witper, RusseLtt M., Bootuspy, Watter M., and BrEeLer, CaRoL. Studies of the metabolism of diabetes.....................-.+-- 311 Hitt, A. V. The interactions of oxygen, acid, and CO, in blood.... 359 Gap-ANDRESEN, K.L. A micro method for the estimation of ammonia in blood and in organie fluids: .......-5. aaneeeee eee 367 Gap-ANDRESEN, K. L. A micro urease method for the estimation of _ urea in blood, secretions, and tissues. ......- 2,2... ae 373 Four, Orro. A system of blood analysis. Supplement III. A new colorimetric method for the determination of the amino-acid nitrogen in blood............ 0: -22.+0000 6 =n-0 Seeger 377 Foun, Orro. A colorimetric determination of the amino-acid nitro- gen in normal urine,..............<2« = cue apes ee 393 Foun, Orro, and Berauunp, Hitpine. The retention and distribu- tion of amino-acids with especial reference to the urea formation. 395 Foun, Orro. Note on the necessity of checking up the quality of sodium tungstate used in the system of blood analysis...........- 419 Foxx, Orro, and Looney, JosepH M. Colorimetric methods for the separate determination of tyrosine, tryptophane, and cystine IM PROLEINB. «cs. «4+ po Seca tes «spe on eee en s.....-....... 54.32 55.18 } 47.18 Non-amino N of filtrate........... yaa | 2.93 2.94 IS 0.487 0.5235, 0.5056 N in protein, per cent............. 15.99 16.94 15.68 a 1.04 3.01 W extracted, per ceni.............. 50 and 65} 78.00 * Corrected for solubilities. 7 Jones, D. B., and Waterman, H.C., J. Biol. Chem., 1921, xlvi, 461. ft Osborne and Harris (Osborne, T. B., and Harris, I. F., J. Am. Chem. Soc., 1903, xxv, 348) obtained 7.3 per cent amide nitrogen in coconut globulin. § Johns, Finks, and Gersdorff (Johns, C. O., Finks, A. J., and Gersdorff, C.E. F., J. Biol. Chem., 1919, xxxvii, ra eee W.O2-per cent arginine nitvened in coconut saiowiin. || Total nitrogen of filtrate 51.35 per cent. The 5 per cent barium hydroxide solution to which 1 per cent alcohol was added by volume extracted 74.3 per cent of the total nitrogen of cottonseed meal. The protein precipitated from the 0.2 per cent sodium hydroxide extract and the precipit- 20 Proteins of Cottonseed, Soy Bean, Coconut able protein from the 5 per cent barium hydroxide extract had the same nitrogen distribution as that of the globulin. The nitrogen distribution of the precipitable proteins in the 0.2 per cent sodium hydroxide extract of the soy bean shows a. somewhat similar distribution to that of the globulin (glycinin). No difference in the nitrogen distribution of precipitated pro- tein containing 50, 65, and 78 per cent of the total nitrogen of the coconut was observed. The protein of the coconut precipi- tated from the 0.2 per cent sodium hydroxide extract by acidifying with dilute acetic acid differs—-markedly ome ii in.the-arginine_nitrogen-value, Sit. @ & unlant x , * f “a - yo SEPARATE ANALYSES OF THE CORPUSCLES AND THE PLASMA. By HSIEN WU. (From the Laboratory of Physiological Chemistry, Peking Union Medical College, Peking.) (Received for publication, July 29, 1921.) In spite of the varied functions of the different components of blood the value of separate analyses of the corpuscle and the plasma has been little appreciated. Unless distribution of a substance within the blood is known and the percentage of the corpuscles is taken into account, the data obtained from the whole blood analysis admit of no strictly scientific treatment. In the transport of material to and from the tissues both the corpuscles and the plasma may play a part; but their actual passage into or out of the circulation is effected only through the plasma. Hence for the solution of problems in connection with absorption, excre- tion, or similar processes, a knowledge of the concentration in the blood as a whole is insufficient, and the determination of the concentration in the plasma is a necessity. Even in clinical practice where the whole blood analysis has proved a useful method of diagnosis and prognosis, separate analyses of the corpuscles and the plasma would afford a more reliable index of certain pathological conditions. For, if the distribution of a substance within the blood is unequal, the concen- tration in the blood as a whole will fluctuate with the percentage of the corpuscles, although the concentration in the plasma and and that in the corpuscles may remain unaltered. The distribution of the better known water-soluble non-protein constituents has been a subject of considerable investigation. It is tacitly assumed by nearly all investigators that the results of the analysis of the corpuscles and the plasma obtained in the usual way truly represents the relation that exists in the circulat- ing blood. Recently, Falta and Richter-Quittner (1) advanced 21 22 Analyses of Corpuscles and Plasma the view that in the circulating blood the non-protein nitrogenous substances, the sugar, and the chloride occur only in the plasma, and that these substances enter the corpuscles only when they areinjured by sodium oxalate, fluoride, or by other manipulations prior to the separation of the corpuscles from the plasma. In explaining the unusual findings upon which they based their conclusion, they referred to the use of hirudin as the punctum saliens of their research, believing that this anticoagulant, unlike the oxalate or fluoride, did not injure the corpuscles. The experimental data presented by these authors are, however, of such a character as to indicate faulty technique,! and their findings have failed to be confirmed by some Danish investigators (2, 3, 4). The view of Falta and Richter-Quittner cannot, therefore, be accepted at present. The results of the study on distribution obtained by other inves- tigators are of considerable interest. The total non-protein nitrogen (5, 6, 7), the amino-acids (5, 6, 8, 9, 10), and the creatine (6, 7, 11) are localized in the corpuscles, the urea (3, 5, 6, 12, 13, 14) and the creatinine (6, 7, 11) are equally distributed, while the uric acid (15, 16) is sometimes more and sometimes less con- centrated in the plasma. The distribution of sugar (1, 17, 18) has long been a disputed question; in the case of human blood the preponderance of evidence now points to an approximately equal distribution. The chloride, the only inorganic constituent touched upon in the present paper, is more abundant in the plasma (19). Unfortunately, most of the studies on distribution were con- ducted on the blood of different species of animals and the data at hand do not lend themselves to correlation. Moreover, parallel analyses were made not of the corpuscles and the plasma, but of the whole blood and the plasma, frequently neglecting to mention the percentage of the corpuscles. Where this percentage is given the concentration in the corpuscles can, of course, be computed. But it is obvious that with methods capable of only moderate degrees of accuracy slight differences between the concentration 1 For instance, they reported in some cases as low as 9 mg. of non- protein nitrogen per 100 cc. of whole blood which is lower than the lowest figure for urea N we have ever found. Hsien Wu 23 in the corpuscles and that in the plasma can be shown only by direct analyses of the corpuscles and the plasma. Folin and Wu (20) developed some time ago a system of blood analysis in which the blood proteins are removed with tungstic acid, yielding a filtrate suitable for the determination of non- protein nitrogen, urea, uric acid, creatine, creatinine, and sugar, and, as shown recently by Whitehorn (21), also for the determina- tion of chloride. It is, of course, very easy to extend that system so as to include separate analyses of the corpuscles and the plasma, and by this extension systematic studies on distribution will be facilitated, and the clinical application of the knowledge gained by such studies will be encouraged. In applying the tungstic acid to the analyses of the corpuscles and the plasma we have required that the procedure employed must permit the quantitative recovery of at least 10 mg. each of uric acid and creatinine added to 100 cc. of plasma or corpuscles and that the non-protein nitrogen figures must be comparable with those of the whole blood. The latter requirement was deemed necessary in view of the fact that the non-protein nitrogen is not a very definite quantity but depends rather on the kind and amount of the precipitant used. It was found as a matter of fact, however, that the non-protein nitrogen values of the cor- puscles and of the plasma obtained with different amounts of tungstic acid were substantially the same when the amount of this precipitant used was above a certain minimum. Table I gives the result of an experiment on this point. Table II shows that the non-protein nitrogen values for the whole blood, calculated from those of the corpuscles and the plasma, agree with those determined directly. Instead of using tungstate and sulfuric acid solutions of dif- ferent strengths for the corpuscles and for the plasma, it has seemed convenient to use the same 10 per cent tungstate solution and the 2 w sulfuric acid and to retain the 1:10 dilution for the corpuscles and the plasma as for the whole blood. The method is as follows. The oxalated blood is centrifuged in graduated tubes until the volume of the corpuscles remains constant. The length of time required for this has been determined previously. After not- ing the volume of the whole blood and that of the corpuscles, 24 Analyses of Corpuscles and Plasma TABLE I. Experiment Showing Effect of Different Dilutions and Different Amounts of Tungstic Acid on Non-Protein Nitrogen of Corpuscles and Plasma. Composition of precipitated mixture. 5 ec. corpuscles + 25 cc. HO + 10 ce. 10 per cent tungstate + 10 cc. 3 N sul- furie acid wase.cissasewes tgs eee be ees 4 cc. corpuscles + 26 ec. H2O + 10 ce. 10 per cent tungstate + 10 cc. } N sul- FUTIC BCIELs ooo. aie oc na oe ate oe eee 6 cc. corpuscles + 24 cc. H2O + 10 ce. 5 cc. corpuscles + 75 cc. H2O + 10 ce. 10 per cent tungstate + 10 cc. 3 N sulfuric 10 ce. plasma + 80 ec. HO + 5 ce. 10 per cent tungstate + 5 ec. } N sulfuric acid.. 8 ce. plasma + 82 ec. H2O + 5 ec. 10 per cent tungstate + 5 cc. 2N sulfuric acid... 12 cc. plasma + 78 ec. H2O + 5 ce. 10 per cent tungstate + 5 ec. 3N sulfuric acid.. 10 ce. plasma + 30 cc. H,O + 5 ec. 10 per cent tungstate + 5 ec. 3 N sulfuric acid.. 15 ec. plasma + 25 ce. H2O + 5 ce. 10 per cent tungstate + 5 ec. 3 N sulfuric acid... | Volume of tungstate Volume of corpuscles or plasma’ Dilu- tion. 2:25 BE 2a 1:20 Non-protein nitrogen per 100 cc. mg. 49 51 49 Incomplete coagulation. 35 34 34 34 Incomplete coagulation. Hsien Wu 25 carefully pipette off the plasma without disturbing the corpuscle layer. This is best done by means of a pipette connected at the upper end with soft rubber tubing. Measure a convenient vol- ume of the plasma, dilute with 8 volumes of water, and then add 1 volume each of 10 per cent sodium tungstate solution and 3 N sulfuric acid. Stopper the flask and shake. Remove the plasma that remains above the corpuscle layer as completely as possible. Insert a blood pipette (20) into the cor- puscle layer and take out a convenient volume. Lake it with 5 volumes of water and after thorough rinsing of the pipette with the corpuscle solution add 2 volumes each of the tungstate solu- tion and the sulfuric acid. Stopper the flask and shake. The amount of plasma which cannot be removed from above the cor- TABLE II. Non-Protein Nitrogen of Corpuscles, Plasma, and Whole Blood. Non-protein nitrogen per 100 ce. Source. Corpuscles. Whole blood. Corpuscles. LO | Determined.| Calculated. per cent mg. mg mg mg pee, ies 0). 20 5a 34 39 38 SRS ried be Bands ost 26 52 36 40 40 higkenyh oe... 38 90 24 48 49 Oi =e 11 60 10 17 16 puscle layer is less than 0.1 cc., and if it is desired to economize the material the whole corpuscle layer may be used for analysis with- out appreciable error. It is then simply washed into an Erlen- meyer flask with 5 volumes of water followed by the required amounts of tungstate and sulfuric acid. The precipitated corpuscles and plasma may be filtered im- mediately. The precipitated plasma should be poured on the filter, slowly at first to allow the wetting of the filter paper before any filtrate has passed through. If for any reason the precipita- tion is incomplete and the filtrate is turbid, the analysis can be saved by adding a few drops of normal sulfuric acid to the pre- cipitate mixture. The plasma and corpuscle filtrates, like the filtrate of the whole blood, are perfectly clear, only faintly acid, 26 Analyses of Corpuscles and Plasma and suitable for the determination of all constituents included in the system of blood analysis. The method described above has been applied to a number of cases of normal human blood. These were obtained from normal persons taking the Wassermann test and from patients suffering from slight external injury such as frost-bitten fingers and lacer- ated toes. The analyses included total non-protein nitrogen, urea, uric acid, amino-acids, creatine, creatinine, sugar, and chloride. The amino-acids were determined by a new method to be published shortly from the Biochemical Laboratory of the Harvard Medical School as a supplement to the system of blood analysis. . The results of the analyses, arranged in the order of increasing urea concentration of the plasma, are shown in Table III. They constitute in the main a confirmation of previous findings, and are intended primarily for illustration. They represent, however, the first series of comprehensive analyses of the corpuscles and the plasma of human blood and some discussion may therefore not be amiss. Urea is about equally distributed, but the concentration in the plasma is sometimes a little higher than that in the corpuscles. The average value for the plasma is 19.3 mg. per 100 cc. as against 17.1 for the corpuscles. One may thus wonder whether in the determination of the so called Ambard coefficient or similar con- stants more consistent values would be obtained if the urea in the plasma instead of that in the whole blood were determined. Contrary to the finding of Bornstein and Griesbach (16) we find that concentration of uric acid in the plasma is always higher than that in the corpuscles. The values of the ratio of the former to the latter are about 2 on the average. The finding of the Hamburg investigators that uric acid is sometimes higher in the corpuscles is probably an error, although we have not studied a large number of pathological cases? to exclude that possibility. In view of the uncertainty in the determination of creatine and creatinine in the whole blood or rather in the corpuscles, as shown by Hunter and Campbell (11), any discussion on the distribution 2 We have studied the distribution of uric acid in a number of patho- logical cases. The results, not here reported, are similar to those in the case of normal blood. Hsien Wu Z'ST9 |2°608 Iz’ SOT 6°66 | 29°S 1Z4b'°6 | POT | Sb'Z | LPT | 288 | 26'S | SOE | SEL | TAT | 8°82 | &'6F | BLP *OFRIOAY OLS O26. | Sal te SIT.) -€°9)| 2:8 (CO) Pe Wea Se! Tae oS) Pal ce €% 02 bE 8P cP 02 09 STS 2 BEL" OL), “OS Gs fa CM ee eH VR a Ts «at ° sD 9 €% ST 9€ ¥¢ 8P 61 919 9c& | OTT | OOT T9|9°8 Pall Shae! Gal |» Si6sieeetoulk ay ob €% 1% ce GG LV 8T og9 Teg | SIT | 86 OF | 06 Pale SesuG: leer aks GaS) 1) Si cpa Ole CG 1G LG 19 8E LT GLO cps | OOT | &6 SO1C0L| rt) 08) FT Pr O6)T Of) Me (a6 GG Té 9S OF 9T 009 coe | COL. Sola) BSa 2-0) OL st Teele GeOD|) earlier 1% 9T 62 8P LY cI €8¢ 266 | G6 18 09|68 CO Ce AS aS G2 I Ores 13 ST 62 cP 0g iA! 629 ets | OOL | OOT | 29)F76 Sole. ona eels. GO. Bred arb IZ OT Ge OP Lg §I 0g9 org | 2Z0L | 80L | F¢gica6 Pol | Ocoee aabe | OxO | eo Gade oe ob 0% ST 6 8P €¢ ral 099 LES) “SLL |) “G00 || Osos en6 Fob lioue. Cott 2aO 2oGa BE 0% LT 0€ LP ¥¢ II OL9 OPE | 16 16 0S | 001 Lait ie ieee | “OFOa pa BT 61 ST 9% cP 6G OL 06S OLZ |. S&T Ter | Gh) 68 OT eG: ook) Pes! SO Gulh (GG 61 CT 9G LV SP 6 Z09 css | 68 16 GL1|36 UabieG 6 jet | Sceet Gal Pew 61 ST €€ PG SP 8 0€9 cae |= SOL is OOLs 04.80). Feb eas. | FE | 60s OPFa Ge 61 ST 62 Tg SP }e O8S 062 | G6 68 0g |96 Baik |e) Se: ee Paks: eS eels Oar One SI ST 92 oS 68 9 09S ThG | QOL O8 Goi lds G-b BtS OST |) SE Bal Sees), 2 °F 8T 8I 0€ 6P oS G GLS OOP ie SOLse OOLs t=O siG Pay |} 2OsBcley Ba0. |) COU s) Te ee ST GT €T GS cP OS P 919 Lee | &§8 £6 PF Se) BeObs! “Csr | Oe i Pats eee Oper ee CT €T €% €¢ €¢ € LLG 8s | SIT LOTS) SS) Fe LL) Be | See iS Pel SOs im Gane, 7 tl ol &% 6 €¢ z P29 ose | G6 601 | Gr] T6 F823 |) SPA. | 0-6 oO gaere T &T él 06 bP €€ I Bu Bu Bua Bu But Bu Bua bu Bu “pu Bu Bu Bu bu Bu Bue | uaa wad BUSEg che “BRUISE ha “BUNS bp a BUISE| eee" “BUIS] T bad “BUISE| ees “RUISt[ i oy “BUUSEL vole ‘09 00I 19g "poolg UuuntT jULON fo Dusv)_g pun sajosndsop ur oprw0jyQ pun ‘uvbng ‘saounjsqngy snoauabo.ujiN U29}0Lq-UoN 28 Analyses of Corpuscles and Plasma of these constituents would seem unwarranted. It is clear, how- ever, from Table III that the plasma is practically free from creatine. The average creatine content of the corpuscles is 5.84 mg. (as creatinine) per 100 cc., whereas that of the plasma is only 0.23 mg. This striking difference brought out by direct analysis of the corpuscles confirms and supplements the conclu- sion reached by Hunter and Campbell and by Wilson and Plass (7) that the creatine is chiefly. contained in the corpuscles. It may be of interest to point out also that the apparent creat- inine content of the corpuscles is just twice that of the plasma. The amino-acid content of the corpuscles averages almost twice that of the plasma. Similar distribution has been found in dogs (9), but so far as we are aware no figure has been reported for human blood. The concentration of sugar in the plasma is usually a little higher than in the corpuscles, but at times the reverse may be true. However, the difference in concentration in such cases is often so small as to be within the limits of error. The concentration of chloride in the plasma is about twice that in the corpuscles. There is another interesting point which the present study has brought out. It is probably a recognized fact that the non- protein nitrogen in the whole blood is not wholly accounted for by the known constituents. A normal human blood contains on the average:3 Per 100 ce. mg. Urea N.. . . oiis acces ores one eie excl ueet one ouete slels 2 else) to) senna 17-10 UWrie acid Nin .u. 5 2205 oc lee oc Ole dyelciece elle eee eee 0.78 Creatinine No ..s.05 26.0 Sis. cre actaleleia cho saieie Shere ele dele nee tet ae 0.47 Creatine N....:sses% seus eusnce oh» le Stes ole dese oeeeaeee ean Amino-acid N...2.. cwa.de ae ues oe eee Roa fe) Total non-protein N calculated. .::... ....: J..25 00s eee 26.78 2 as N determined’... .. .:/3-scc, Gece eee eee 35 .60 Undetermined’ N...2. 7. eoee eee en oe eee eee pas os 8.80 3 These figures, with the exception of the amino-acid N, are taken from a paper by Hammett (Hammett, F. 8., J. Biol. Chem., 1920, xh, 599). The amino-acid N figure is taken from Bock (10). That given by Hammett is 4.9 which seems too low. Hsien Wu 29 There are about 9 mg. of nitrogen not accounted for in the whole blood. This question has been discussed by some authors, but so far as we are aware no satisfactory explanation has yet been offered. Can the undetermined nitrogen be due to the incomplete removal of proteins? It is well known that the non-protein nitrogen values obtained with different methods do vary appreciably, and there are doubtless some protein precipitants which leave traces of protein in the filtrate. But the tungstic acid filtrate gives as low non-protein nitrogen as can be obtained with any other method without losing the known constituents and it is improbable that as much as 9 mg. of protein nitrogen can escape precipitation. TABLE IV. Distribution of Non-Protein Nitrogenous Constituents between Corpuscles and Plasma. Constituent. Corpuscles. Plasma. oS a ee 17:10. | 4 deeae 0 TE 2 0.64 1.31 Le So a = 10° 4 0.55 a ls 9.47 | 5.52 Total non-protein N calculated .............. 30.31 | 26.68 = se N determined ............. Cs Se aaa a SS 0 19.0 Pea | The present study on the distribution seems to throw much light on this question. It will be noted in Table IV, which repre- sents the averages of Table III, that the non-protein nitrogen in the plasma is within limits of error‘ all accounted for and that the undetermined nitrogen is all contained in the corpuscles. This localization shows that the undetermined nitrogen is not due to incomplete precipitation of the proteins, for otherwise we should expect the plasma also to contain some of it. To what substance then can be ascribed this undetermined nitrogen which amounts to almost 20 mg. per 100 cc. of corpuscles? ‘There are, of course, traces of non-protein nitrogenous substances other than those determined. 30 Analyses of Corpuscles and Plasma In the corpuscles there are, on the one hand, the amino-acids, and on the other hand, the proteins. In the living protoplasm of the corpuscles, as in all living cells, there is a metabolic equilibrium between the amino-acids and the proteins with all the inter- mediates; namely, the peptides and peptones. The presence of peptides and peptones in the corpuscles is indeed very probable, and pending further experimental evidence the undetermined nitro- gen in the corpuscles may be regarded as such. Deductions Concerning Clinical Blood Analysis——Since the undetermined nitrogen is all contained in the corpuscles and the amount of amino-acids in the corpuscles is larger than that in the plasma, the non-protein nitrogen of the whole blood will fluctuate with the content of the corpuscles in these constituents. As the latter is not known to vary with renal efficiency, the non-protein nitrogen of the whole blood would not be a reliable index of reten- tion. The determination of the non-protein nitrogen of the plasma should therefore be preferred to that of the whole blood. The approximately equal distribution of sugar and urea justifies the usual practice of their determination in the whole blood, al- though their determination in the plasma is theoretically better. The distribution of uric acid is so different and variable that one may wonder whether the content in the whole blood, the cor- puscles, or the plasma is the most valuable for practical purposes. In the absence of other information the concentration in the plasma would furnish the most useful data. Hunter and Campbell (11) have shown that the only substance in the plasma capable of simulating the reaction for creatinine is glucose and its influence upon the determination of creatinine is too small to have much practical importance, whereas the cor- puscles contain other interfering substances which cause high results. Creatinine should therefore be determined in the plasma rather than in the whole blood. There is thus abundant reason for substituting plasma analysis for whole blood analysis. SUMMARY. A method is described for the preparation of protein-free fil- trates of the corpuscles and the plasma suitable for all determina- tions included in the system of blood analysis of Folin and Wu. Hsien Wu 31 Sample analyses of the corpuscles and the plasma of normal human blood are given and the points of interest brought out by these analyses are discussed. It is recommended that plasma analysis be substituted for whole blood analysis. more wl BIBLIOGRAPHY. . Falta, W., and Richter-Quittner, M., Biochem. Z., 1919, c, 148; 1921, exiv, 145. . Ege, R., Biochem. Z., 1920, evii, 246. . Gad-Andresen, K. L., Biochem. Z., 1920, evii, 250. . Hagedorn, H. C., Biochem. Z., 1920, evii, 248. . Bang, I., Biochem. Z., 1915, lxxii, 104; 1916, Ixxiv, 294. Wilson, D. W., and Aaolah, E. F., J. Biol. Chem., 1917, xxix, 405. (In the case = fishes they found more urea in the whole blood than in the plasma.) . Wilson, D. W., and Plass, E. D., J. Biol. Chem., 1917, xxix, 413. (In the case of fishes they found more creatine in the plasma than in the whole blood.) . Costantino, A., Biochem. Z., 1913, li, 91; lv, 402. . Gyorgy, P., and Zunz, E., J. Biol. Chem., 1915, xxi, 511. . Bock, J. C., J. Biol. Chem., 1917, xxix, 191. . Hunter, A., and Campbell, W. R., J. Biol. Chem., 1917, xxxii, 195; 1918, xxxiii, 169. . Hammarsten, O., Text-book of phiyuisiogical chemistry, translated by Mandel, J. AS 1915, 333. . Marshall, E. K., and Davis, D. M., J. Biol. Chem., 1914, xviii, 53. . Karr, W. G., and Lewis, H. B., J. Am. Chem. Soc., 1916, xxxviii, 1615. . Benedict, 8. R., J. Biol. Chem., 1915, xx, 633. (This author found that in ox blood the uric acid was quantitatively contained in the cor- puscles, whereas in chicken blood the corpuscles were almost free from uric acid.) . Bornstein, A., and Griesbach, W., Biochem. Z., 1920, ci, 184. (The distribution of uric acid in human blood was found to be irregular by these authors.) . Wishart, M. B., J. Biol. Chem., 1920, xliv, 563. . Ege, R., Biochem. Z., 1920, cxi, 189. . Abderhalden, E., Lehrbuch der physiologischen Chemie, Berlin, 1906, 591-593. . Folin, O., and Wu, H., J. Biol. Chem., 1919, xxxviii, 81; 1920, xli, 367. . Whitehorn, J. C., J. Biol. Chem., 1920-21, xlv, 449. THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI, NO. 1 A NEW COLORIMETRIC METHOD FOR THE DETERMI- NATION OF PLASMA PROTEINS. By HSIEN WU. (From the Laboratory of Physiological Chemistry, Peking Union Medical College, Peking.) (Received for publication, November 17, 1921.) In connection with the study of a certain blood reaction in cases of kala-azar (1), we had occasion to determine the different plasma proteins. The refractometric method of Robertson (2) or the recent method of Cullen and Van Slyke (3) based on nitrogen determinations would have served our purpose; but the lack of certain facilities required in these methods compelled us to devise some simpler technique adapted to the equipment at our disposal. As a result of our endeavor we have developed a new colorimetric method which is much simpler than and, we believe, fully as accurate as, any of the other methods. There are two steps in the determination of the plasma pro- teins; wz., (a) their separation from each other, and (6) their | quantitative estimation. In the first step we have followed, in the main, the procedure of Cullen and Van Slyke. The isolation of the fibrin by recalcification of the oxalated plasma under the conditions prescribed by these authors is very satisfactory, and we have further simplified the process by whipping the fibrin out of the jelly, thus obviating the necessity of washing. For the precipita- tion of the globulin they used ammonium sulfate at half satura- tion. In our study we have found that saturation with magne- sium sulfate gives the same results as half saturation with ammo- nium sulfate, and either process can be used in the new colorimetric method. In the application of this method to clinical cases, how- ever, we have used the ammonium sulfate process which is some- what more convenient. The magnesium sulfate process was used only when parallel determinations by the Kjeldahl method were desired. 33 2 34 Determination of Plasma Proteins In the second step we make use of the color reaction (4) of pro- teins with phospho-18-molybdictungstic acid (phenol reagent). Proteins in solution react with this reagent, due largely to the tyrosine which they contain. We have made no experiment to determine what the color produced by the proteins quantitatively represents. But since this chromogenic value is a constant for any given protein, the intensity of the color produced under definite conditions can be used as a measure of the amount of the same protein. Solutions of pure serum globulin or albumin can, of course, be used for the standard. But as these are laborious to prepare and difficult to keep, we have used exclusively a solution of tyrosine in our work. A convenient standard is made by dissolving 50 mg. of tyrosine in 250 cc. of 0.1 N HCl. We have observed no change in this solution in the course of 6 months and it may keep much longer. Under the conditions described below we have found for human plasma that 1 mg. of tyrosine! = 16.4 mg. of fibrin (N X 6.25), 25.2 mg. of globulin (N X 6.25), or 27.5 mg. of albumin (N X 6.25). The tyrosine equivalents of the plasma proteins of different species of animals are probably different and should be determined when required. To do this it is only necessary to make parallel determinations on the same sample of plasma by the new colori- metric method and by the Kjeldahl method. In the method which follows the fibrin and the albumin are determined directly, while the globulin is determined by the dif- ference between the total serum proteins and the albumin. All the determinations can be made simultaneously and finished in 1 hour. Determination of the Fibrin—To 1 cc. of the plasma (from blood containing 0.2 to 0.6 per cent potassium oxalate) add 28 cc. of 0.8 per cent NaCl solution and 1 cc. of 2.5 per cent CaCl: solution. Mix and allow to stand undisturbed for 20 minutes. Break up the jelly by shaking slightly and transfer it to a dry filter. While filtering insert into the jelly a slender glass rod with a pointed end 1We used Pfanstiehl tyrosine prepared by the sea Chemicals Com- pany, Highland Park, Ill. Hsien Wu 35 and whirl gently. All the fibrin will stick to the rod.? Slip the fibrin off the rod, and press it between dry filter paper to remove as completely as possible the adhering liquid. Transfer it to a 15 ce. centrifuge tube, add 4 cc. of 1 per cent sodium hydroxide. Place the tube in a boiling water bath and stir with a slender glass rod until the fibrin lump has completely disintegrated. The fibrin has now dissolved, leaving the calcium oxalate in suspension. Add 10 ce. of water, mix, and centrifuge. Transfer the supernatant liquid to a 25 cc. volumetic flask or graduated tube. Cool under the tap. Add 1 cc. of 5 per cent HSOs, 0.5 cc. of phenol reagent, and dilute to about 20 ce. Add 3 ce. of 20 per cent Na2COs solu- tion. Shake. Add 1 drop of ether to dissipate the foam, make up to volume, and mix. The standard is prepared as follows: Measure 1 cc. of the standard tyrosine solution into a 25 cc. volumetric flask or graduated tube, add 0.5 cc. of phenol reagent, dilute to about 20 cc., and finally add 3 cc. of 20 per cent NasCO3 solution. Make up to volume and mix. The standard should, of course, be prepared at the same time as the unknown. Let stand for 15 minutes before making the color comparison. Calculation.—If the standard is set at 20 and the reading of the unknown is R, then the amount of the apparent tyrosine deter- mined is cia xX 0.2 mg. Since 1 mg. of tyrosine = 16.4 mg. of R fibrin, the amount of fibrin in 1 cc. of plasma is = xX 0.2 X 2 16.4 mg. or the percentage of fibrin = = X 0.328. Determination of Albumin.—To 1 ce. of plasma add 9 ce. of $ saturated ammonium sulfate solution® or 9 cc. of saturated mag- nesium sulfate solution and 0.3 gm. of anhydrous magnesium sulfate. Mix and allow to stand for 30 minutes. Filter. Meas- ure 1 ce. of the filtrate into a 15 cc. centrifuge tube. Add about 2 If any fibrin fails to stick to the rod it should be picked up with the tip of the rod. It requires no great skill to do this successfully. Ifthe amount of the fibrin in the plasma is very high, say 0.8 per cent or more, - the fibrin jelly will not shrink readily. In such a case it is necessary to use less of the plasma for the determination. ’ This is made by diluting 555 ec. of saturated ammonium sulfate to 1 liter. ¢ 36 Determination of Plasma Proteins 12 cc. of H.O, 1 ce. of 10 per cent sodium tungstate solution, and then 1 cc. of 2 wn sulfuric acid. The amount of the tungstate solution and the sulfuric acid need not be exactly measured, but care must be taken to see that the volume of the acid used is at least equal to that of the tungstate solution. Stir thoroughly with a slender glass rod and centrifuge. Carefully decant off the supernatant liquid as completely as possible. (The volume of the wet precipitate usually amounts to about 0.5 ce. If it is much smaller than 0.5 ecc., indicating low albumin, measure another cc. of albumin filtrate into the same tube, dilute with water and proceed as before.) Add to the precipitate in the cen- trifuge tube 1 cc. of sodium tungstate solution. Stir until the precipitate has dissolved, dilute with 13 cc. of H2O, and add 1 ce. of H.SQ,. Stir again, centrifuge, and decant off the supernatant liquid. This second precipitation is intended to remove the calcium and ammonium or magnesium so nearly completely that they cannot possibly interfere with the subsequent color reaction, although experience has shown that a single precipi- tation usually suffices. Add to the precipitate in the tube 10 cc. of H.O and 1 or 2 drops (but no more) of 20 per cent Na2COs. Stir until the precipitate has dissolved. Transfer the resulting solution to a 25 cc. volumetric flask or graduated tube. Rinse the centrifuge tube twice with 3 cc. of H2O. Add 0.5 ec. of phenol reagent and 3 cc. of 20 per cent Na2CQs solution. Shake. Add 1 or 2 drops of ether to dissipate the foam. Make up to volume and mix. Prepare a standard as in the fibrin determination and read the color after 15 minutes. Determination of Albumin and Globulin.—Measure 2 ce. of the filtrate in the fibrin determination into a 15 cc. centrifuge tube and proceed exactly as in the determination of albumin. Calculation.—In the calculation it is to be noted that the solu- tion used for the albumin determination is plasma diluted 1 to 10, while that used for the determination of albumin and globulin is plasma diluted 1 to 30. If 1 ec. of the albumin solution is used for the former determination and 2 cc. of the serum solution are used for the latter determination and the colorimeter readings are R, and R; respectively, the standard being set at 20, then the 20 total apparent tyrosine in 1 ce. of srum = 15 X R, xX 0.2 mg., ‘ t . Hsien Wu ad the apparent tyrosine of albumin in 1 cc. of serum = 10 X 0 eis. R < 0.2 mg., and the apparent tyrosine of globulin in 1 ce. of serum is Bayt. 20 _ 60 40 (15 X FX 0.2) — 0X BX 0.2) =p — pm. Since 1 mg. of tyrosine = 25.2 mg. of globulin = 27.5 mg. of albumin, 60 _ $0) 559 *, per cent of globulin = vile! x 100 ay 5 5 1,000 20 10x x 02) x 27.5 a R 20 per cent of albumin = 1,000 < 100 = R, % 544 EXPERIMENTAL. Determination of the Tyrosine Equivalent of the Plasma Proteins. Fibrin.—The fibrin in about 5 cc. of human plasma was isolated as described. It was placed in a 25 cc. volumetric flask and 20 ce. of 1 per cent NaOH were added. The flask was placed in boiling water with frequent shaking. When the fibrin lump had com- pletely disintegrated the flask was cooled under the tap and the solution was made up to volume with 5 per cent H»SO,; and mixed. It was filtered to remove the calcium oxalate. 5 cc. of the filtrate were taken for the tyrosine determination as described above. The nitrogen was determined in 2 or 3 cc. of the filtrate by the micro Kjeldahl method (5). It was found in two experiments representing two samples of plasma that 1 mg. of tyrosine = 16.8 and 16.0 mg., respectively, averaging 16.4 mg. of fibrin (N X 6.25). Albumin.—The globulin together with fibrin was precipitated from the plasma using magnesium sulfate as described. 1 ce. of the filtrate was taken for the tyrosine determination. 5 cc. of the filtrate were measured into a 10 cc. volumetric flask, diluted to volume, and mixed. 1 cc.of this diluted solution was measured into a Pyrex test-tube, 2 cc. of digesting mixture were added, and 38 Determination of Plasma Proteins the digestion was carried out in the usual way (5). 25 ce. of H.O were added and the solution was cooled and transferred to a very large test-tube or small flask. 25 cc. of 10 per cent NaOH were then added and the ammonia was aspirated and determined by Nesslerization in the usual manner. The ammonia in the reagents and the non-protein nitrogen of the plasma were deter- TABLE Il. Plasma Proteins in Normal and Pathological Blood.* Total No. Diagnosis. Fibrin. serum | Albumin. | Globulin. protein. per cent per cent | per cent | per cent 1] SNormmals aaa sot Aare 6.50 4.55 1.95 2 St oe eee eerie yaa ie ie 6.55 4.80 Lis 3 eae RY ee eee Ee ae ree 7.51 5.28 2.23 4 se Py te Wie ye ros 0.222 | 6.62 4.75 1.87 5 sh SB or am ee RR ' 7.54 4.89 2.65 6 | Syphiliss 2350 oe ea tecereeee 8.04 4.63 3.41 if eT Ae eee es 9.37 4.95 4.42 5), eal salma ee ears oie eee 9.59 3.83 5.76 9 Ms ty PEE APE ST TS 10.54 3.48 7.06 10 AO ON BBO os Sets aaa 6.78 2.90 3.88 11 eM 3 Vere Abhi) Bs y 2.94 4.43 12 “ Lapegeai See eee 0.271 | 6.99 1.82 5.17 13 £ ony vielen SAE a ee ee 0.306 | 7.82 2.35 5.47 14 + treated: 2.0 cceene 0.290 | 7.40 4.05 3.35 15 me RN is eS cs 0.292 | 7.79 2.62 5.17 16 CO eet te. eee || 708° ")) epee ee et 17 | Streptococcus septicemia........ 5.56 2.76 2.80 18 | Amebic dysentery............... 4.24 2.90 1.34 19.,|: Relapsing fever. ..2.c.<6t eee 0.485 | 5.57 2.58 2.99 20 | Nephritis with edema........... 0.630 | 3.07 0.53 2.54 7.50 5.00 2.50 21 ANVEINID,.. 22:5) <2. eC ne * Some of the figures in this table have been reported elsewhere (1). mined and corrections were made accordingly. In five experi- ments representing five specimens of plasma it was found that 1 mg. of tyrosine = 27.6, 27.4, 27.5, 27.3, and 27.8 mg., respectively; averaging 27.5 mg. of albumin (N X 6.25). Globulin.—1 cc. of the serum filtrate of the same plasma used for the albumin determination was measured into a Pyrex test-tube and the nitrogen was determined exactly as in the preceding ex- Hsien Wu 39 periment. In another 2 cc. of the filtrate the tyrosine was deter- mined. It was found by the method of calculation indicated above that 1 mg. of tyrosine = 24.9, 25.8, 24.7, 24.9, and 25.6 mg., respectively; averaging 25.2 mg. of globulin (N X 6.25). It would appear from the values of tyrosine equivalents given above that the fibrin contains much more tyrosine than the al- bumin or the globulin. But this is not the case. We have ob- served that the chromogenic value of the albumin or the globulin is greatly increased by treating with NaOH for a few minutes or with NasCO; on long standing. The apparently high tyrosine content of the fibrin is no doubt due to the action of NaOH used to bring it into solution. In Table I are shown some of the results obtained with the new method. It is not the purpose of the present paper to discuss the significance of the findings in those pathological cases reported here, but the deviation from normal is sometimes so marked that the clinical value of the determination of plasma proteins is apparent. BIBLIOGRAPHY. . Sia, R. H. P., and Wu, H., China Med. J., in press. . Robertson, T. B., J. Biol. Chem., 1912, xi, 179. . Cullen, G. E., and Van Slyke, D. D., J. Biol. Chem., 1920, xli, 587. . Wu, H., J. Biol. Chem., 1920, xliii, 208. . Folin, O., and Wu, H., J. Biol. Chem., 1919, xxxviii, 87. ok Whe » ff a STUDIES ON EXPERIMENTAL RICKETS. XVI. ADELICATE BIOLOGICAL TEST FOR CALCIUM-DEPOSITING SUBSTANCES. , By E. V. McCOLLUM anv NINA SIMMONDS, (From the Laboratory of the Department of Chemical Hygiene, School of Hygiene and Public Health, the Johns Hopkins University, Baltimore.) AND P. G. SHIPLEY anp E. A. PARK. (From the Department of Pediatrics, the Johns Hopkins University, Baltimore.) Puates 1 AnD 2. (Received for publication, January 11, 1922.) In our second communication (1) we described two diets (Lots 2638 and 2677) which we had used to prepare animals for a bio- logical test for the calcium-depositing power of various food sub- stances. VThis test depends on the power of a given substance to cause the reappearance of a provisional zone of calcification in epiphyseal cartilages of animals with very severe rickets, whose cartilages had been rendered calcium-free by faulty diets. We called this procedure the ‘‘line test,’”’? because in gross specimens or in histological preparations the new zone of calcification appears as a line of calcified tissue crossing an area which is free from lime salt deposits.Y Schmorl (2) showed that the first sign of healing in rachitic children was the reformation of the provisional zone of calcification on the epiphyseal side of the metaphysis. The value of the reaction is dependent upon the certainty with which a given diet will operate to produce bones in which the epiphyseal cartilage and\the metaphysis are calcium-free. In other words, it is not sufficient that the diet shall produce rickets. It must cause an exaggerated form of florid rickets to develop without _ fail. Our results with the above diets were not entirely satis- factory, since now and again an animal would show some evidence of calcification of the epiphyseal cartilage. Furthermore, we 41 42 Studies on Experimental Rickets. XVI found that young rats on these diets did not grow enough to produce as severe rickets as we desired. Many animals died in a short time because of the pronounced deficiencies of the diet (Lot 2677). Moreover, rats on these diets showed a marked tendency to destroy and eat their companions in the cage. This, of course, vitiated the experiments. The data presented in the earlier papers of this series leave no room for doubt that certain faulty relations between three dietary factors, vz. calcium, phosphate, and an unidentified organic sub- stance, lead to the development of the picture of rickets in young rats as it occurs in children. When a diet contains a somewhat excessive amount of calcium, decidedly less than the optimal amount of phosphate, and very little of an organic substance which is especially abundant in cod liver oil, rats which are fed on it develop rickets (3). We have pointed out that pronounced rickets develops only when the conditions are such as to permit young rats to grow. When the faults in the diet are such as to cause rickets, the severity of the disease is increased if the food is such as to insure to the animal a certain amount of growth and to keep up its vitality. VA diet which proves highly satisfactory for the purpose in hand, in that young rats fed upon it have marked rickets with an epi-_ physeal cartilage and metaphysis uniformily free from calcium, is made up as follows: Diet 3143. Whole wheat kernel. .33.0 100 gm. of this diet contains 1.221 gm. “maize CAROoaO of calcium, and 0.3019 gm. of phos- etait. 555 see 15.0 phorus. Wheat gluten......... 15.0 Weight ratio Ca:P = 4.04:1 NED) Ci legit lah ah I ag 1.0 WaACOR ase oe 3.0 Atomic ratio Ca:P = 1:0.319 The proteins of this diet are of good quality and are abundant (about 33 per cent of the food mixture). The amount of phos- phorus in the mixture is distinctly below the need of the growing rat for this element. 100 gm. of this food contain 0.3019 gm. of phosphorus. Our observations indicate that the optimal amount of phosphorus in the diet of the growing rat is not less than 0.4146 per cent. It may be considerably higher. The calcium McCollum, Simmonds, Shipley, and Park 43 content of the food mixture is 1.221 gm.—approximately double that required for optimal growth and function. The content of fat-soluble A is very low but suffices to prevent the development of xerophthalmia, and to make growth possible for a time, other factors permitting. The content of the second organic factor supplied by certain fats (if this proves to be a distinct entity) is likewise very low. Young rats which have grown rapidly to about 55 to 60 gm. (35 to:40 days old) are confined to Diet 3143 for a period of about 35 to 40 days (preparatory period). If the animals are allowed to remain too long on this diet they deteriorate physically and their hind legs become paralyzed. After being confined to the rickets-producing diet for from 35 to 40 days the rats have some difficulty in using their hind legs. The changes in the animals’ gait are not so striking that they would be noticed except by one who is familiar with the habits and movements of the rat. The gait is tottering, and the hind quarters waver slightly from side to side. When the animal starts to move off rapidly it hops, usually favoring one hind leg. We have never failed to find the cartilages of the bones of animals in this condition calcium-free. The metaphysis is usually quite free from calcium salts but occa- sionally a few calcified areas are found. When a group of rats are ready to be used for the test, some of them are given the substance which we desire to study, in their food, and the rest serve as control animals and continue to receive Diet 3143 unchanged. (The time during which the rats receive the substance which is being studied may be called the ‘‘test period.”) The test animals and controls are killed and autopsied at the end of a designated time. It has been our custom to examine the distal end of the left femur and the proximal end of the left tibia for reformation of the provisional zone of calcifica- tion. Of the two bones the tibia is perhaps the better, since the epiphyseal line is straighter in this bone than in the femur, and the metaphysis wider. The bones are split in two with a sharp scalpel and one-half of each is fixed in 10 per cent formaldehyde, decalcified in Miiller’s fluid, and embedded in parloidion. The other half of each bone is immersed in a 1 per cent solution of silver nitrate and exposed to sunlight or to the light of a Mazda lamp, and studied through a 44 Studies on Experimental Rickets. XVI binocular microscope. The slices which have been treated with silver nitrate may be cut into frozen sections or may be preserved as permanent gross specimens. For the latter purpose they are washed well with water, decolorized with a solution of sodium thiosulfite, rewashed and kept in formaldehyde. The bones of these animals are very soft. The ends are very much enlarged. On section the cartilage of the epiphysis is irregular in depth and the epiphysis is separated from the dia- physis by a zone of white tissue (the metaphysis) which may be about 0.5 em. deep. In the bones of the control animals or rats which give a negative ‘‘line test’”’ there is no calcification of the epiphyseal cartilage, and little if any in the metaphysis (Fig. 1). The shaft is incompletely calcified. We have described these animals and their bones completely in another place (8). Rats which give a positive “line test” differ from the controls in having a broad, linear deposit of calcium salts on the metaphyseal side of the epiphyseal cartilage. The band, which may not be com- plete, is separated from the shaft of the bone by the depth of the metaphysis, and from the nucleus of ossification of the epi- physis by the depth of the epiphyseai cartilage. It can be seen on the freshly cut surface of an untreated bone as a yellow line which marks the epiphyseal border of the metaphysis. The deposit is blackened by silver nitrate in fresh or fixed gross specimens. It appears like a cross-section of a black honey- comb when it is examined with a binocular microscope. The | metaphyses of these bones usually appear to be congested. ‘This deposit is in the proliferative zone of the cartilage. It may extend completely across the bone or may be interrupted or frag- mentary according to the activity of the calcium-depositing substance (Fig. 2). It is stained brown by silver nitrate and an intense blue by hematoxylin. It is separated from the calcified trabecule of the shaft by the width of the metaphysis. The honeycomb-like appearance of the line is explained by sections which show that the matrix of the cartilage only is calcified (Fig. 3). The cartilage cell is not. The deposit is granular.' Although the possibility of detecting this calcification in living “ 1In studying fresh, silvered preparations it must be borne in mind that a precipitate of silver is sometimes thrown down irregularly on the surface of the bone without relation to calcium deposits. If this precipitate is McCollum, Simmonds, Shipley, and Park 46 children by the use of the x-ray has been recognized for sometime (4), Park and Howland (5) were the first to make use of it experi- mentally. They were able to show that cod liver oil would regularly induce calcification and healing in the bones of children withrickets. Since their experiments Huldschinsky (6) and others have demonstrated the curative power of ultra-violet light in the same way. When rats are used as subjects the picture on the x-ray plate may be confirmed by microscopic study of the bones. The results obtained with this test are illustrated by the effect of the administration of cod liver oil. Within 5 days the addition of 2 per cent of cod liver oil to the food causes complete recalci- fication of the cartilage. “ Cod liver oil added to the food in amounts equal to only 0.2 or 0.4 per cent of the food mixture does not cause the line of calcium phosphate to appear in 5 days. 0.6 per cent, on the other hand, causes the appearance of an irregular calcification in some of the bones in this time. We shall describe later the effects of small doses of cod liver oil repeated for shorter or longer periods in producing the reformation of the provisional calcified zone. In conducting these tests it is necessary to keep accurate records of the food consumption by both test and control rats during the interval following the administration of the substance to be tested. This is so because starvation causes rats with rickets to deposit calcium salts in the epiphyseal cartilage, just as does cod liver oil (7). We have taken this precaution in all our tests. A positive test can be relied upon only when the animals eat enough. Since the amount of and the relation between calcium and phosphorus in the food is a very important factor in determining the ability of the skeleton to grow normally, it is essential, when employing the test here described for the examination of.a natural food (e.g. the leaves of plants, etc.), to modify the basal diet at formed on the cartilage it may be mistaken for an irregular calcification. The regular honeycomb appearance of the true “‘line’”’ of calcification dis- tinguishes it. In case of doubt the bone may be treated with sodium thios e. The blackness of the silvered calcium is unaffected by 2 to 3 minutes’ treatment with the solution, but the precipitate will be dissolved at once. We have always controlled the results of the examination of fresh tissues by the study of microscopic sections. The results obtained by the use of the two methods have never failed to correspond. ,, tal Rickets. XVI xperimen Studies on E 46 ‘qisodop repNsoAIT -- “SulpRoyy “UOTpIpUOD ssojdjoy UT se pu SFoy pury ut pozATwaud sum yvyY *UOTPVAI]S OF O[GuaeduIOy y a ee ee le 6& LE LE «1 “ud “skup ¢ ur uozBo poo ” G ” ” ” ” ” ” 0G ” G ” ” ” ” ” »” 0'% ” G ” ”» ” ” ” ”? 0% ” G ” ”? ” ” ” ” 0% ” G ” ” ” ” ” ” 0° IT ” ¢ ” ” ” ” ” ” GO ” G ” ” ” ” ” ” V0 ” ¢ ” ” ” ” ” ” 90 *“sAUP G AO] [LO JOATT poo quod a90d 0)’ T ‘[10 ANAT] POO snyd yorp poAtooor SAU ‘TOUTAVL ‘[O.LUOD) *[o1yUOD ro) C&T | 9 SP ved OTT’ 2P 0g é OOT | 9 cP re) 98 OL ce é Ost | OL gg é g8 Lg 0g é got | 29 0g é €L Lg 0G red GL 68 OV é 06 6P Gg 6 LL 6P cP oe 6P cP es LOT | 6% cP ie 66 $9 GP 6 21L | O2 09 6 86 OL 09 6 | SOT} 04 | 09 6 yh OL 09 | ‘ao sivp ‘yyeop | “gorp Aalp soy |e | UP | want VysIOM | SABC ony ‘ON yOu McCollum, Simmonds, Shipley, and Park 47 the time of making the test, so that the added food will make no change in the inorganic composition of the diet as compared with that of the food in the preparatory period. In another paper in this series we show in some detail how important it is to adhere rigidly to the composition laid down for the diet used for the preparation of the animals for this test. The test is applicable for the demonstration of the effect of light rays or other radiations, or other remedial agencies. It constitutes, therefore, a valuable aid to further experimental work in the field of bone pathology. SUMMARY. 1. A diet is described which is low in phosphate and in an undetermined organic substance and relatively high in calcium. This diet uniformly causes the bones of animals fed on it to be free from calcium as regards the epiphyseal cartilages and the metaphyses. These bones show an exaggerated type of florid rickets. 2. The reappearance of the provisional zone of calcification after the addition of any substance to the rickets-producing diet (beginning healing of the rickets) constitutes a test for the calcium- depositing power of the substance (the so called ‘‘line test’’). 3. Cod liver oil to the extent of 2 per cent by weight of the diet causes the provisional zone of calcification to be reformed in these bones in 5 days (a positive “‘line test’’). 4. 0.4 per cent of cod liver oil fails to cause the provisional zone of calcification to reappear in the bones. 1 per cent of cod liver oil for 5 days does not cause so complete a calcium salt deposition as does 2 per cent. 5. Both test and control animals must be kept under the same conditions of illumination, and since starvation causes redeposi- tion of calcium in the bones, records of the food consumption of the animals must be kept during the experiment. 6. The basal diet must be so modified when any natural foods are tested that the addition of the food to be examined will not change the inorganic content of the mixture used in making the test from that of the original diet used during the preparatory period. 7. This test may be used to determine the ability of any agency to heal rickets. 48 Studies on Experimental Rickets. XVI Cuart 1. The curves of Lots 2818 and 3143 illustrate the importance of having an optimal ratio between calcium and phosphorus in the diet of the growing rat. ~ Lot 3143 had a diet which contained an abundance of proteins (33 per cent) of good quality, but was too low in phosphorus and too high in cal- cium. The diet contained 1.221 gm. of calcium in 100 gm. This is about twice the optimal amount of this element. 100 gm. contained 0.302 gm. of phosphorus. This we know to be distinctly below the optimal amount, although the most favorable concentration of this element in the diet has not been accurately determined. In certain other diets we have been able to secure good nutrition when the food contained 0.415 per cent of phosphorus. We may safely assume that the diet under consideration is at least 30 per cent too low in this element. In addition to these defects the diet of Lots 3143 and 2818 was relatively deficient in fat-soluble A. The diet of Lot 2818 differed essentially from that of Lot 3143 in its con- tent of phosphorus. As respects protein, calcium, and fat-soluble A, they were essentially alike. Diet 2818 contained 0.4928 per cent of phosphorus, and 1.247 per cent of calcium. The calcium: phosphorus ratio in each of these rations were: Lot 3148. Lot 2818. Atomic ratio ...Ca:P::1:0.319 Atomic ratio ...Ca:P::1:0.523 Weight ratio ....Ca:P::4.04:1 Weight ratio ....Ca:P::2.51:1 The diet of Lot 3143 contained only 61.25 per cent as much phosphorus as did the diet of Lot 2818. This difference in phosphorus content was sufficiently significant for the welfare of the growing rat to make the differences in growth seen in the contrasting curves of the two groups shown in the chart. McCollum, Simmonds, Shipley, and Park 49 BIBLIOGRAPHY. 1. Shipley, P. G., Park, E. A., McCollum, E. V., Simmonds, N., and Par- sons, H. T., Studies on experimental rickets. II. The effect of cod liver oil administered to rats with experimental rickets, J. Biol. Chem., 1920-21, xlv, 348. 2. Schmorl, G., IX. Die pathologische Anatomie der rachitischen Knochen- erkrankung mit besonderer Beriicksichtigung ihrer Histologie und Pathogenese, Ergebn. inn. Med. u. Kinderheilk., 1909, iv, 403. 3. McCollum, E. V., Simmonds, N., Shipley, P. G., and Park, E. A., Studies on experimental rickets. VIII. The production of rickets by diets low in phosphorus and fat-soluble A, J. Biol. Chem., 1921, xlvii, 507. 4, Park, E. A., and Howland, J., The radiographic evidence of the influ- ence of cod-liver oil in rickets, Bull. Johns Hopkins Hosp., 1921, xxxii, 341. 5. Park, E. A., and Howland, J., The dangers to life of severe involvement of the thorax in rickets, Bull. Johns Hopkins Hosp., 1921, xxxii, 101. 6. Huldschinsky, K., Die Behandlung der Rachitis durch Ultraviolett- bestrahlung Dargestellt an 24 Fallen, Z. orthop. Chir., 1919-20, Xxxi1x, 426. 7. McCollum, E. V., Simmonds, N., Shipley, P.G., and Park, E. A., Studies on experimental rickets. XV. The effect of starvation on the healing of rickets, Bull. Johns Hopkins Hosp., 1922, xxxiii, 31. EXPLANATION OF PLATES. PLATE 1. Fig. 1. This figure shows the distal end of a section of the tibia of a control animal which had been fed our diet, No. 3143. Note the prolif- erative zone of cartilage (C) which is entirely free from calcium salts, and the wide calcium-free zone (M) between the cartilage and the end of the shaft (D) which contains partially calcified trabecule. Fig. 2. This figure shows the distal end of the tibia of an animal which had given a positive line test. This animal had been fed diet, No. 3148, but had received 2 per cent of cod liver oil for 5 days before death. Note the ‘‘line’’ (CA) of newly deposited calcium in the proliferative cartilage. This line is a newly formed zone of provisional calcification. PuaTeE 2. Fig. 3. Higher magnification, X 300, of the newly formed provisional calcified zone (CA). This picture shows the calcification of the inter- cellular matrix of the proliferative cartilage. Some of the calcified cell capsules have been broken open and invaded by capillary vessels from the shaft. a ey : RNA? 7 THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI. PLATE 1, Fig. 2. (McCollum, Simmonds, Shipley, and Park: Studies on experimental rickets. XVI.) THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI. PLATE 2. @. yd (McCollum, Simmonds, Shipley, and Park: Studies on experimental rickets. XVI.) ) bas. es CALCIUM IN EGG-SHELL FORMATION. By G. D. BUCKNER, J. H. MARTIN, W.C. PIERCH, anp A.M. PETER. (From the Department of Chemistry, Agricultural Experiment Station. University of Kentucky, Lexington.) (Received for publication, January 4, 1922.) Domestication of the chicken has intensified its egg-laying power and it is not unusual for a White Leghorn hen to lay 300 eggsin1l year. This is a large number when contrasted with the 22 to 26 eggs laid yearly by the jungle fowl, Gallus ferrugineus, the probable progenitor of the domestic fowl. For the production of eggs the Leghorn will require yearly approximately 750 gm. of CaO instead of the 32 gm. used by the jungle fowl for that purpose. . Common practice teaches that calcium in some concentrated form, such as oyster shell, must be added to the ordinary feeds fed to poultry in order to supply the large amount of calcium necessary for optimum egg production, and that grit is added to aid digestion. In connection with an experiment dealing with the Ca, Mg, and P metabolism in the laying hen, six lots, each containing ten 7 month old White Leghorn pullets which came from the Same parent stock, hatched the same day, and existing under identical conditions, were fed the following feeds: Lot1. Grains + caer + no mineral material. cere «+ + granite grit, ad libitum. ie I “ ae “ au ‘“ “ c “c +oyster shell, ad libitum. “ 4. “ + “ a “ “ “cc “ +limestone, ad libitum. art ie anak «+ limestone, ad libitum. ca ay oe «+ rock phosphate, ad libitum. Tankage (12.5 per cent of total fed) containing 6.4 per cent P.O; was fed in the dry mash. The grit used contained 2.4 per cent CaO soluble in strong HCl. 51 52 Calcium in Egg-Shell Formation This experiment was started December 1, 1920, and ended August 1, 1921. During these 8 months the pullets were con- fined in houses and at no time allowed access to the ground, thereby excluding any chance of their obtaining calcium-con- taining materials from undesired sources. No milk or green foods were fed. At the beginning of the experiment two representative birds were killed and after their femurs and tibias were dissected out and separately weighed, the CO,-free ash and the CaO contained therein were determined. A hen of the same age, having received the same treatment as all other birds in this experiment up to December 1, 1920, but which had from that time until August 1, 1921, been allowed the normal freedom of a meadow and been fed the same as Lot 3, was killed and similarly dissected as were the early controls just mentioned. All the birds remaining in the six lots on that date were killed and treated in the same way. A trap-nest record was kept of each bird confined. Among other things the CO,-free ash and the percentage of CaO in this ash were obtained in a composite sample of the shells of the first. three eggs laid each month by each hen. The results in Table I, which represents averages, show that the total weights of the four large leg bones of the hens in Lots 1, 2, 3, 4, 5, and 6 were approximately the same, while that of the first control was lighter, being 8 months younger, and that of the second control was heavier, resulting from the different physical conditions governing the 8 months over which this experiment extended. The actual weight of the CQ:-free ash of the leg bones from Lots 1 and 2 was lighter than that from Lots 3, 4, 5, and 6 which received the concentrated calcium materials, and the ash from Lot 6 which received rock phosphate was heavier than that from any of the other lots. This agrees with results published by D’Anchald! in 1904 who obtained a heavier skeleton by adding daily 4 gm. of powdered bone per hen to an ordinary poultry feed. It will be seen that Lot 1, which received no mineral matter in addition to the grains and tankage fed, laid 19.9 eggs per hen in 8 months which is slightly in excess of the number laid by 1D’Anchald, H., J. agric. prat., 1904, vii, 619. 53 Buckner, Martin, Pierce, and Peter ee ee ee 6 tS L9°9 & GE v6 TT 86°98 | “[Or}uOO pug) T » 69°26 LL 18°T Ist 0°S¢ Li-9 0°96 0G IT cL Ié 9 T » 69° 2 SPS PG % 96S o'g¢ oo 9 LYvé 69 OL pS 0& g Lk > 69° 16 96°S GPG 9°6S Gag T9°¢ GSE ST OL 62 0& ie Ly 89° 46 68°S SVG G GG Lvs 61°¢ 8 && 6¢ OL 0€ TE & | ae LV 16 66 'T 86°T 8°8& v tg 6L°P 9°16 088 G6 TE G Le ay Lv 16 61° T p81 6 61 € vs ee P b 9G L6°2 96 0& I T ‘3ny 1661 6 FS 60°& G GG £9'¢ PE CG | Osan ey T (00d “ub “m6 quao sad “Ub quaa sad “mB “mu O@6l acl a *qo] sod awe “syzUOUL g UT *souo0q So] *sou0q Fo] , 3 od MSP USMS) a0 730 | cr gee toys | 20 22d wou Bhigtene we ut aOR Us i Ses aoe af enuiene phi Oey 94-700 x one 9.AJ-OD Ul OBO o3v10A y ee vo pe wyapeas i109 ON 90'T “ed jo aa oy a JO FYSIOM a AV ODVIIAY ODVIIA YW BIBIOAW Ni ‘I ATa&VL 54 Calcium in Egg-Shell Formation Lot 6 which received the rock phosphate ad libitum. This would seem to indicate, when considered with the previous statement, that the rock phosphate can be utilized by the hen in bone forma- tion but not in the production of eggs, calcium being the limiting factor in the case under consideration. This is shown by the fact that the average weights of the CO,-free ash and the CaO in the egg-shells of Lots 1, 2, and 6 are practically the same while those in Lots 3, 4, and 5 are larger. It is interesting to note that the percentages of CaO in the CO--free ash of the egg-shells is practically constant in all lots. Elsewhere? we have shown that the quantity of CaO in the contents of the eggs governed by these conditions is practically constant. Lot 2 laid a few more eggs per hen than did Lot 1—probably due to the 2.4 per cent CaO available in the grit fed to Lot 2. The average number of eggs laid by each hen in Lots 3, 4, and 5 is approximately the same for each lot. In briefly summing up these results it would seem that the hen can utilize the calcium in calcium carbonate for the produc- tion of both egg-shell and bones but that the calcium in tricalcium phosphate can only be utilized for the growth of bone and not for egg-shell production. Finally, calcium starvation is not the determining factor in the production of shell-less eggs. | 2 Buckner, G. D., and Martin, J. H., J. Biol. Chem., 1920, xli, 195. A METHOD FOR THE ESTIMATION OF TOTAL BASE IN URINE. By CYRUS H. FISKE. (From the Biochemical Laboratory, Harvard Medical School, Boston.) (Received for publication, December 26, 1921.) The investigation of diurnal variations in the excretion of bases (with the exception of ammonia) has received practically no attention. While this might formerly have been due to the fact that suitable methods for the estimation of the important individual non-volatile bases (sodium, potassium, calcium, and magnesium) were not available, that is no longer so. The main deterrent appears to be the large amount of work required for the separate determinations of these four constituents in the many samples of urine involved in most short period metabolism experi- ments. When, as may often be the case, the information sought is the total equivalent of all four bases, rather than the amount of each one, the magnitude of the task is out of all proportion to the value of the final result, which includes the accumulated errors of the four individual analyses. Since, in the interpreta- tion of the results of such investigations, the equivalent of the chloride content of the urine must ordinarily be subtracted, and since the remainder may be comparatively small, the micro methods so far described are not sufficiently precise to furnish figures for the sum of the bases that have the requisite degree of accuracy. The method described here is designed to accomplish this. Up to the final step in the analysis, it is merely an adaptation of long used schemes for the separation of these bases from other sub- stances. The urine, after being ashed with sulfuric and nitric acids, is treated with ferric chloride to remove the phosphate, and with ammonium acetate to remove the excess iron (as the basic acetate). The filtrate from these operations contains the bases, which are finally obtained as sulfates, free from inter- fering substances, by ignition with sulfuric acid and then with 55 56 Estimation of Total Base ammonium carbonate. The final residue is analyzed for sulfate by the benzidine method,! and the total quantity of 0.1 N base present calculated from the result of this analysis. In connection with the removal of phosphate and excess iron, the approximate directions commonly given in text-books of analytical chemistry for the adjustment of acidity, while adequate for use in the older gravimetric methods, are not sufficiently precise with the small quantities used in the present method, where more exact neutralization is required to prevent significant amounts of alkaline earth phosphates from being carried down with the iron precipitate. Since the removal of phosphate proceeds more smoothly when the amount to be removed is known beforehand, a preliminary determination of phosphate is assumed in the detailed description following. This information can now be very readily obtained by Bell and Doisy’s colorimetric method? with more than sufficient accuracy for the purpose. As an alternative, ferric chloride may be added until no further precipitate appears, but the presence of a large amount of basic ferric acetate makes the subsequent filtration much slower, and is likely to result in the consumption of more time than would be required for the phosphate determination. Two variations of the method will be described. In the first, which is probably slightly more accurate, the iron precipitate is washed free from the desired bases with hot ammonium acetate solution, the filtrate cooled and diluted to a definite volume, and the analysis completed with an aliquot. This is time-consuming when many determinations must be made, and for that reason a much more rapid modification has been devised. In this case, the volume of the hot solution, after the addition of ferric chloride and ammonium acetate, is adjusted to 10 ec., and then cold water is added to make the total volume 25 cc. (and the temperature, therefore, about 50°C.). The solution can then be mixed, and the filtration completed in about 2 minutes. This procedure is, of course, necessitated by the fact that the basic ferric acetate would redissolve if the solution were allowed to become cold. 1 Fiske, C. H., J. Biol. Chem., 1921, xlvii, 59. 2 Bell, R. D., and Doisy, E. A., J. Biol. Chem., 1920, xliv, 55. - C. H. Fiske 57 The adjustment of volume at a temperature considerably above that of the room is not a serious matter, since a correction can easily be applied. The only real cause for concern lies in the possibility of concentration changes during the filtration, due to the escape of water vapor. By working under uniform conditions, and taking certain precautions, this error can be diminished to less than 0.5 per cent. Removal of Phosphate. First Method.—Measure into a large lipped Pyrex test-tube (200 by 20 mm.) a sample of urine containing an amount of chloride equivalent preferably to between 10 and 25 mg. of NaCl, but not more than 5 mg. of inorganic phosphorus. Add 1 ce. of 4 n sulfuric acid and 0.5 ec. of concentrated nitric acid, and boil down until white fumes appear. If the residue does not soon become colorless after this stage has been reached, cool slightly, add a few drops more of nitric acid, and continue the heating. When the remaining drop of sulfuric acid has become clear and colorless, let cool for a few minutes, and add 10 cc. of water and a drop of a saturated alcoholic solution of methyl red. Neutralize with powdered ammonium carbonate until the color of the indicator just begins to change, and restore the pink color by adding 4 n sulfuric acid, 1 drop at a time. Heat to boiling, and continue adding sulfuric acid in the same manner until 1 drop makes the solution faintly but unmistakably pink. Add a 10.5 per cent solution of ferric chloride (FeCl; 6H2O) in 0.1 N hydrochloric acid in the proportion of 0.1 cc. for each mg. of inorganic phosphorus present in the original sample of urine,? — shake, and run in 1 ce. of a 5 per cent solution of ammonium 3 The ferric chloride solution may be delivered from an ordinary 1 ce. pipette, and measured with sufficient accuracy by counting the drops. The amount prescribed is about 20 per cent in excess, whichshould be enough to take care of any phosphate derived from organic phosphorus compounds in the original urine. To be perfectly safe, the filtrate should be tested for phosphate with the reagents of Bell and Doisy.? 5 ce. of the filtrate, when treated with 1 cc. of the molybdate solution, 1 cc. of hydro- quinone solution, and 2 cc. of the carbonate-sulfite reagent, should give a color distinctly less than that obtained with 5 cc. of a phosphate solution, containing 0.005 mg. of phosphorus, treated in the same way. 58 Estimation of Total Base acetate. Heat again to boiling. Filter on a 9 cm. ashless paper* into a 50 cc. volumetric flask, and wash with five 8 cc. portions of boiling 0.5 per cent ammonium acetate solution. Cool under the tap, make up to the mark, and mix. Second Method.—Ash as described above, but add only 2 ce. of. water to the cooled residue. Transfer to a test-tube marked at 10 and at 25 cc.,® rinsing three times with 2 cc. of water. The neutralization with ammonium carbonate and the precipitation with ferric chloride and ammonium acetate are conducted exactly as described under the first method. The hot solution at this stage should have a volume of 10 or 11 cc. Have ready a dry funnel fitted with a 9 cm. ashless paper,* and a dry 25 cc. Erlen- meyer flask. To the contents of the test-tube, which should be at the boiling point, add cold water (room temperature) to the 25 ec. mark. Immediately close the mouth of the test-tube with a clean, dry rubber stopper, invert two or three times, and filter without delay into the small Erlenmeyer flask, keeping the paper nearly filled as long as possible, and collecting only about 20 cc. of filtrate. The paper should fit the funnel snugly, and the entire filtration must not, in any event, occupy more than 2 minutes.® Stopper the receiving flask at once, and let cool (or cool under the tap if the analysis is to be completed immediately). The filtrate may contain a trace of iron, but this does no harm, since ferric sulfate decomposes on ignition. Larger quantities of iron would interfere mechanically. 4 Nothing but the best quantitative paper will give dependable results. One determination with Schleicher and Schill No. 604 filter paper on 2.5 ec. of 0.1 mM KCl took 0.535 ec. of 0.1 N NaOH for an aliquot amounting to one-fifth of the total. This, calculated to the basis of the entire sample, gives 2.675 cc. of 0.1 N base, instead of 2.500 ce. 5 The tubes used as receivers in the Folin-Wu urea method are suitable (Folin, O., and Wu, H., J. Biol. Chem., 1919, xxxviii, 96). The 25 ec. mark must, of course, be accurate; that at 10 cc. need be only approximate and can be made with a glass pencil if necessary. 610 ce. of n sulfuric acid were heated to boiling, diluted to 25 ce. with cold water exactly as described, and filtered through a funnel plugged with paper so that 3 minutes were required for the filtration of 20 cc. The filtrate, after cooling, proved to be 0.4055 Nn, or, after applying the correc- tion for temperature (1 per cent), 0.4015 N, instead of the theoretical 0.4N. The error incident to evaporation was, therefore, only 0.4 per cent, although the filtration took half again as long as prescribed in the method. C. H. Fiske 59 Ignition and Sulfate Determination. Transfer to a small platinum dish (about 5 cm. in diameter) an amount of the filtrate corresponding with the chloride equiva- lent of about 2.5 mg. of NaCl in the original urine,’ add 1 cc. of 4 n sulfuric acid, and evaporate on the water bath until nearly dry. Place the dish on a metal triangle, and heat, cautiously at first, over a micro burner, gradually raising the flame until fumes have ceased to come off. Let cool, sprinkle over the residue a small amount of powdered ammonium carbonate, and ignite again, finally raising the flame to its maximum and moving the triangle about until each part of the dish has been momentarily subjected to a dull red heat. When the dish has cooled, add 2 ec. of water. Agitate until the residue has dissolved, using a rubber-tipped rod to assist in its solution if necessary. ‘Transfer the contents of the dish to a large lipped Pyrex test-tube (200 by 20 mm.),® and rinse four times with 2 cc. of water. Add 2 cc. of benzidine reagent,® and, 2 minutes later, 4 cc. of 95 per cent acetone. After waiting 5 minutes for the completion of the precipitation, filter through paper pulp in the special filtration tube recently described,! wash three times with 1 cc. of 95 per cent acetone and once with 5 cc. With the aid of about 2 cc. of water and a nichrome wire, transfer the precipitate and mat through the bottom of the filtration tube into the test-tube in which the precipitation took place, and rinse off the wire with a few drops of water. Heat to boiling and add through the filtra- tion tube (still suspended in the mouth of the test-tube) 2 drops of a 0.05 per cent aqueous solution of phenol red, and (from a 7 For example, if the second method has been used with a sample of urine containing 12.5 mg. of NaCl, the proper amount of filtrate is one- fifth of the total, or 5 ce. 8 Large Pyrex test-tubes provided with satisfactory lips are more con- venient than beakers for use in connection with the benzidine sulfate method. Beakers were recommended before! because properly lipped Pyrex tubes had not been obtained at that time, and tubes of thinner glass cannot be relied upon to stand the treatment to which they must be sub- jected. The lipped Pyrex tubes used in this work were obtained from the Macalaster-Bicknell Co., 28 Wendell St., Cambridge 38, Mass. ® Suspend 4 gm. of benzidine in about 150 cc. of water in a 250 ce. vol- umetric flask. Add 50 cc. of n hydrochloric acid, shake until dissolved, and make up to volume. Filter if necessary. 60 Estimation of Total Base micro burette) about 0.3 cc. of 0.1 N sodium hydroxide.!® Rinse down the sides of the filtration tube with about 3 cc. of water from a wash bottle, boil until steam escapes from the mouth of the test-tube, rinse again with about 2 cc. of water, remove the filtration tube, and continue the titration until 0.005 cc. of 0.1 N alkali produces a distinct pink color that does not change on further boiling. The volume of alkali used gives, without further calculation, the amount of 0.1 N non-volatile base in the sample of filtrate analyzed in case the first method has been followed in removing the phosphate. With the second method, a correction of 1 per cent must be subtracted to compensate for the contraction in volume on cooling the filtrate from 50° to room temperature. Since reagents that give a wholly inappreciable blank appear to be easily obtainable, there is no reason for being satisfied with anything less good, for the necessity of correcting each determina- tion for impurities in the reagents not only diminishes the accuracy of the method, but may also considerably complicate the calcula- tion, inasmuch as some of the reagents are used in very variable amounts. In order to test them, all that is required is a ‘‘deter- mination” (by the second method) on the reagents alone (includ- ing 0.5 cc. of the ferric chloride solution). 5 cc. of the filtrate from the iron precipitate should be ignited with sulfuric acid and ammonium carbonate in the manner described, and 2 cc. of benzidine reagent introduced into the platinum dish containing the residue. The absence of any precipitate of benzidine sulfate (after pouring the contents into a test-tube) is a very sensitive test, and shows that no significant amount of interfering impurity is present in the reagents. Analysis of Mixed Salt Solution. After sufficient preliminary work with potassium chloride solutions had been done to determine the most suitable conditions, the method was tested with a solution containing equivalent amounts of sodium, potassium, calcium, and magnesium," with 10 More dilute alkali may be substituted if desired but the amount of alkali that will be required is large enough to permit the use of a 0.1 N solu- tion. 11 Urine ordinarily contains much smaller proportions of alkaline earths than this, and they are the only bases of the four that are likely to cause trouble for specific reasons (by being carried down with the iron precipi- tate). C. H. Fiske 61 and without the addition of phosphate (as NH,H2PO,). The following materials were used. Sodium Chloride—Recrystallized by saturating an aqueous solution with hydrochloric acid gas. Dried, and heated to incipient fusion. Potassium Chloride—Kahlbaum’s (zur Analyse). Calcium Sulfate—Prepared by mixing equal volumes of 20 per cent calcium chloride and 20 per cent sulfuric acid. Filtered, washed, dried, and ignited to constant weight. Magnesium Sulfate—Prepared from Merck’s ‘‘Reagent’? magnesium oxide and sulfuric acid. Recrystallized, dried, and ignited to constant weight. Monoammonium Phosphate.—By recrystallization of a Baker and Adam- son preparation (wrongly labelled (NH:)2HPO;). The final product was free from non-volatile bases. The purity of the sodium and potassium chlorides was further checked by chloride determinations, and that of the calcium and magnesium sul- fates by sulfate determinations. A solution of each of these salts was prepared, equivalent to a 0.1 n solution of base (the calcium sulfate being dissolved in 0.5 n hydrochloric acid), and a mixture made of equal volumes of the four solutions. The results obtained with this mixture by the two methods are shown in Table I. TABLE I. Determination of Total Base in Solution Containing 0.1 mM Na, 0.1m K, 0.05 M Ca, and 0.05 mM Mg. First method Second method. oe Btided. | srttrnte |. 9-1 base- Filtrate Oa ee _ FSS ee eg Oo a ae ee lyzed.* | Found. | Present. | 1¥2¢¢-1 | Found. Beare: Present. ce mg ce ce. ec ce ce cc ce 10 0 5 1.005 | 1.000 3 1.210 | 1.200 | 1.200 10 5 5 1.005 | 1.000 3 12054) e951 12200 1 0.410 | 0.405 | 0.400 5 0 10 1.000 | 1.000 5 1.015 | 1.005 | 1.000 5 = 10 0.995 | 1.000 5 1.015 | 1.005 | 1.000 > 0 5 0.505 | 0.500 5 5. 5 0.505 | 0.500 2a 0 10 0.395 | 0.400 5 0.400 | 0.395 | 0.400 2 5 10 0.400 | 0.400 5 0.405 | 0.400 | 0.400 * Total volume of filtrate, 50 cc. t Total volume of filtrate, 25 cc. t By subtracting1 percent. All figures are given to the nearest 0.005 ce. FAT-SOLUBLE VITAMINE. X. FURTHER OBSERVATIONS ON THE OCCURRENCE OF THE FAT-SOLUBLE VITAMINE WITH YELLOW PLANT PIGMENTS.* By H. STEENBOCK anp MARIANA T. SELL. (From the Department of Agricultural Chemistry, University of Wisconsin, Madison.) (Received for publication, January 6, 1922.) When in 1919 (1) the senior author called attention to the apparent intimate association in the occurrence of the fat-soluble vitamine and yellow pigmentation it was stated that the data would be presented in detail as the evidence secured might war- rant. In the meantime advantage has been taken of this obser- vation to study the nature of the vitamine, assuming that such association was of more than mere physiologic significance. While this was in progress the publication of our data has been delayed. Originally we were led to surmise a possible intimate relation- ship between pigment and fat-soluble vitamine on the premises that the leafy parts of plants, and carrots and sweet potatoes— all containing large amounts of yellow pigment—were also rich in the vitamine. Many fats low in pigment had already been found correspondingly poor in vitamine. Furthermore, tubers and roots such as the potato, dasheen, mangel, sugar beet, and red beet—in contrast to carrots and sweet potatoes—all poor in yellow pigment content, were also found poor in vitamine or even entirely free from it. These conditions were taken as merely suggestive as even to the casual observer many instances would immediately occur which indicate absolute non-conformity with the above con- ditions. Such is especially true with many fats, some of them * Published with the permission of the Director of the Wisconsin Experi- ment Station. 63 THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI, NO. 1 64 Fat-Soluble Vitamine. X being highly pigmented and yet impotent as a source of vitamine. Also in many plant materials we early noted a lack of propor- tionality between vitamine and yellow pigment content. In animal husbandry circles there has always prevailed a vague suspicion that the yellow and white varieties of Indian corn were not entirely equivalent in feeding value. Such was also our inference when we found that the only instances of ophthalmia in our stock colony of rats were observed when we substituted white corn for yellow corn in their rations. Apparently our animals had been kept on a ration dangerously low in fat-soluble vitamine which fact did not, however, become immediately evi- dent until this substitution was made. Carefully planned feed- ing trials with many different varieties of maize have now left no question but that yellow maize as compared with the white varieties is immeasurably superior in fat-soluble vitamine con- tent (2). Peas have been graded variously as to their fat-soluble vitamine content. Chick and Delf (3) found both dry and germinated peas deficient. Coward and Drummond (4) credit them with more or less fat-soluble vitamine activity. McCollum and co- workers (5) obtained good growth on 87.5 per cent of peas but state that the addition of butter fat would have improved the ration for reproduction. We have found it possible to harmonize these points of view as upon analyzing different varieties we (6) found that they differed tremendously in their vitamine content; those richest in vitamine were again found highest in yellow pigment content. Outside of the data reported from the author’s laboratory (7) unfortunately there appear to be none available on the vitamine content of sweet potatoes and carrots. However, as our original observations have now been repeatedly confirmed in successive years on different samples, there appears to be no question but that the ordinary yellow and reddish yellow varieties are abun- dantly supplied with it. In fact our most recent data indicate an occasional occurrence of even larger amounts of vitamine than we had originally surmised to be present. With the association of vitamine with yellow pigment established in the previously mentioned plant materials we have now extended our investi- gations to include the white as well as the pigmented varieties H. Steenbock and M. T. Sell 65 of carrots and sweet potatoes. Our observations here are limited, but so far, the non-pigmented varieties were again found very low in vitamine. Leafy materials have repeatedly been shown to be well sup- plied with the fat-soluble vitamine. We have here also availed ourselves of an opportunity to demonstrate the association of vitamine with yellow pigments by feeding green and etiolated cabbage leaves, leaves respectively rich and poor in yellow pigments. EXPERIMENTAL. All our demonstrations of fat-soluble vitamine activity were made with rats of mixed black and albino ancestry. These were taken at an age of 3 to 43 weeks at the corresponding weights of 40 to 70 gm. and put on a basal ration of salts, casein, dextrin- ized starch, and a source of water-soluble vitamine, with which the unknown to be tested was incorporated. The salts used were respectively Salt Mixture 32, the composi- tion of which has already been given (7), and Salt Mixture 38 which represented a mixture of 3 parts of Salt Mixture 32 with 1 part of Salt Mixture 35 (7). The casein was a high grade of commercial casein purified by washing for a week with two changes of dilute acetic acid daily. It was finally washed with distilled water, then dried and baked in shallow pans at 95° for at least 72 hours. The dextrinized starch was prepared by barely moistening corn- starch with a 0.1 per cent solution of citric acid and then auto- claving it for 2 to 3 hours at 15 pounds steam pressure. After drying at 80 to 85° for at least a week it was ground to a fine powder. The water-soluble vitamine was incorporated with different materials. In some of the experiments we used 40 per cent of white corn. This introduced a superabundance of water-soluble vitamine, but with it were also introduced small amounts of the fat-soluble vitamine or some other growth constituent, because with it present in the ration usually better growth is secured than with water-soluble vitamine from other sources. For compara- tive purposes its use was, however, entirely permissible. In some instances 2 per cent of dried yeast was used as the carrier 66 Fat-Soluble Vitamine. X of the water-soluble vitamine and again in other cases 3 per cent of ether-extracted wheat embryo or a larger amount of the alcohol extract of ether-extracted wheat embryo evaporated on and made up to the original weight of the wheat embryo. This latter preparation we have used repeatedly in our laboratory and for the sake of convenience have given it the name of dextrin wheat embryo. The sweet potatoes were sliced and dried at room temperature in an air current or else, as noted, were dried in an oven at ap- proximately 80 to 85°. As the circulation of air in this oven was very limited, the potatoes were well steamed by their own mois- ture content before desiccation had ensued to any noticeable degree. The carrots were trimmed to remove the crown in order to avoid contamination of the root with the attached leaf stems. They were shredded in a power-mill and dried at room tempera- ture in an air current. Results with Sweet Potatoes. In Chart 1 are shown the growth curves obtained when 15 per cent of oven-dried sweet potatoes furnished the sole source of fat-soluble vitamine. The three varieties on which all our tests were conducted are represented in these trials. The Triumph variety was a root with flesh of a creamy white color, Little Stem Jersey had flesh of a light yellow color, and Nancy Hall flesh of an intense yellow color. As is seen in Chart 1 in Lot 932, failure of growth was com- plete in all individuals on the Triumph samples at the end of 7 weeks. By this time the eyes of Rats 3722 and 3725 were edematous and at times completely closed—symptoms entirely. characteristic in the reaction of the mammal to a deficiency of the fat-soluble vitamine. When this deficiency was removed by the introduction of butter fat, the eyes promptly cleared up and growth was resumed. Individuals in Lots 820 and 821 on the yellow pigmented roots grew at what might be called the normal rate. Though young were produced none were reared, which was to be expected as lactation is far from normal on such rations even when their fat-soluble vitamine content is more than sufficient. Further data are shown in Chart 2. H. Steenbock and M. T. Sell 67 Ratiq et Pqtate 15 dali 7 [aavae aoe ) Seta ea c77 Ai Berit iti POA eg TIAL VIA | Lee Pee ATA] | potest fst | ah | ft SSR uaZaaaee PEE ae he = me | AY + PV) | _ (28 4/2488 5940e _ 9 65478004558 L | Vste Ueto Moe | Veo | | | mH BS | sad ako Behe ie He ‘4 @ ~ in¢ Emp in Ezeh as PS La malls N ae Ps Ly ae eo FRE ° po | ct re A) op @ CuartT 1. 68 Fat-Soluble Vitamine. X As in the case of the experiments shown in Chart 1 complete failure again resulted on the Triumph sweet potatoes which as well as the other varieties were air-dried. This was to be ex- pected as they were fed at a lower level, 5 per cent as compared ae BO1l ue eet Potat Gn. 120 80 40 a ° OF fh OF SLEP LLEV 9LEP GLEr| € | §€ G8EF ISEP OSEP 6LEF, € | & O98F 6S8P SS8P LG8F| % |G 2 *SU104}OG MOT[A *sdoq MOTTO A *SU09JO OFIT A 9LET LLEI SZZT COST ‘ane }UI019g "sdoq OFT MA 6Z6T “OUI M |79L *S]OLIGO 07 a ee ‘Il GTGVL 73 H. Steenbock and M. T. Sell *qUOTJIOX [[B@ SPRY “OUIVIVJIA O[QN[OS-} VF quoloyop of onp viuyeyzydo ue you Ajureyi09 ‘eurnesy ATqeqoid ‘soko Sulyoo] snorrdsns pey Gzzp pur CCCP SPVY «“UOTYIPUOD pooOs UI SzVyY "yooM YET Sulanp ouo pure Y46 94} SulMp ouo ‘sunod JO $104}I]T OM} PLY GISF qey “UOTyIpuod YuUdT[I2xo Ul s}eyy ‘UOT}IPUOD Pood UI S}EVY *YoOoM YIOT OY} BSutanp Sunod pevy 88h FEY “UONyTpuod poos ul szVyY 161 C61 £0 908008 082 rozlogzlore gezieezlogz €16|L9C|0FG 0ZZ|80G|Z8T 89Z|ZSS/8e O82 LEC 0&6 0LZ/096|S66 69Z|/0SE/SEC 022|SSa/S6T 062,082|SS2 S9TSOTISLT/SST|SST OFZ|8ET|8ZS|LTS|STSG OGTILGT|TTZ|L8T\ZST|SFI|SET/STI}Z6 |c9 Z6Z|E87Z|82Z|026| F9Z|8ZS|S0G|L1\ST 1) 99 ——— |§ ——— | ———— SIZ II@ S61 806 ITZ OST|ZST\06 |99 00Z/ZFT/O6 |ZS GST|OST|OOT|$9 Q8T/SZT\28 |OS SLT\GET} 2 OLTOOT| TZ ZST/LET\C6 LLUOFT|S6 LOTISST|S8 ZOT/STT/O8 GPTISTT08 8ZS|S6T/ZS TOOT) SS SETISTTIS6 [EZ |SP OGT/SLTIOFT/OS |SP 20Z|80Z/86T|S8TOSTIO9T/SFIETTIOL |EF OZZT/ETS/SET|LSTISST|LGTIZET/SOTZ9 |SP OIZ/E1LS|ZOST|OGT/00Z|G9T|OFT|OZT/OOT/99 O&Z/OES|ZIZ|E0Z|00T|Z8T|GOT|OET/SS |SPF Pe as eee PU On OL Te PI | SS re PD HE PS Re WE a es P| a ort aceor oes ®o OF "dD "oO Z| ‘SdAvo] JOIIvO pay |LEET *SOARO] MOT[PX |9EET Z 4OIIvd z | SOAvOT JOITV OPI AA [CEST € *SU10}}0q poy |SLET € ‘sdo} pow |6L8T g ° SS SRenP AP on em or ig Vaes oe aL, i ee cn BE act Neat A ee a O Yo 0 o on - “4 4 | se} = cs a ~ “ we ° ”wo oe ° °o n =] fo] °o °o oO oO o wo N @ vv N a a Oa). iS. Oo) 6 Pe, Vie aK On at te) ore OF ee Se CO eo ao CMe Oi” SCR FO ee H. Steenbock and M. T. Sell 75 ours, nevertheless, it is interesting to note that they likewise report a considerable vitamine content in the green material. Delf (9) has also stated that the inner white leaves of cabbage - do not contain “growth-promoting vitamines.’”’ McCollum, Sim- monds, and Pitz (10), Osborne and Mendel (11), and the author and coworkers (12), have all reported the presence of the fat- soluble vitamine though in lesser amounts than in other leafy materials. We believe it safe to assume that the fat-soluble vitamine is generally present in more than the minimal demon- strable amounts in white cabbage and that Coward and Drum- mond failed to demonstrate it because the level at which it was fed was too low to bring out its presence. As Delf’s data have not been published in detail we are not able to explain the results obtained unless it is to be assumed that some varieties contain much more of it than others. A sample of the green and white leaves analyzed for pigment content by the method used with carrots, revealed a tenfold greater pigment content with the green leaves than with the white. CONCLUSIONS. Data are presented which further emphasize the fact that the fat-soluble vitamine often occurs most prominently where there are found the largest amounts of certain yellow pigments. White sweet potatoes and white carrots were found to contain little fat-soluble vitamine which stands in marked contrast to our observations on the yellow pigmented varieties. The tops of white carrot roots, slightly pigmented with chlorophyll and con- taining a small amount of yellow pigment were found richer in fat-soluble vitamine than the bottoms containing only one-half as much pigment. Green cabbage leaves taken from the heart of cabbage plants which failed to ‘“‘head’”’ were found much richer in fat-soluble vitamine than white cabbage leaves in the head. The latter contained only one-tenth as much yellow pigment. The authors wish to express their appreciation to Dr. J. S. Caldwell, Bureau of Plant Industry, for furnishing them with the sweet potatoes used in this investigation. 76 or WN tite 12. Fat-Soluble Vitamine. X BIBLIOGRAPHY. . Steenbock, H., Science, 1919, i, 352. . Steenbock, H., and Boutwell, P. W., J. Biol. Chem., 1920, xli, 81. . Chick, H., and Delf, E. M., Biochem. J., 1919, xiii, 199. . Coward, K. H., and Drummond, J. C., Biochem. J., 1921, xv, 531. . McCollum, E. V., Simmonds, N., and Parsons, H. T., J. Biol. Chem., 1919, xxxvil, 287. . Steenbock, H., Sell, M. T., and Boutwell, P. W., J. Biol. Chem., 1921, xlvil, 303. . Steenbock, H., and Gross, E. G., J. Biol. Chem., 1919, xl, 501. . Coward, K. H., and Drummond, J. C., Biochem. J., 1921, xv, 535. . Delf, E. M., South African J. Sc., 1920, xvii, 121; abstracted in Chem. Abstr, 1921, xv, 1154. . McCollum, E. V., Simmonds, N., and Pitz, W., Am. J. Physiol., 1916, xiv 361% Osborne, T. B., and Mendel, L. B., J. Biol. Chem., 1920, xli, 549. Steenbock, H., and Gross, E. G., J. Biol. Chem., 1920, xli, 149. WATER-SOLUBLE B AND BIOS IN YEAST GROWTH. By ELLIS I. FULMER anv VICTOR E. NELSON. (From the Department of Chemistry, Iowa State College, Ames.) (Received for publication, November 21, 1921.) During the past 2 years there has been considerable discussion relative to water-soluble B as a possible stimulant of yeast growth. Until recently the prevailing opinion was to the effect that the yeast growth stimulant, bios, and water-soluble B are identical. In fact, the stimulative effect upon the growth of yeast of extracts of materials was proposed (1) as a quantitative measure of the water-soluble B content of the materials. In recent articles! by Fulmer, Nelson, and Sherwood (2), experimental data were presented which showed that the method of measuring vitamine content by yeast growth stimulation was of no value for the following reasons: (a) the relative potencies of two materials as yeast growth stimulants cannot be arrived at on the basis of the stimulative effect of the extracts from equal weights of materials; (b) treatment with alkali does not impair the stimulative effect of the alcoholic extracts, a treatment known to destroy water-soluble B; (c) alcoholic extracts of alfalfa and wheat embryo contain sufficient nitrogenous and mineral nutri- ment for the growth of yeast; (d) a medium of known composition has been developed, namely Medium F, which promotes the growth of yeast without vitamine. The addition of alcoholic extracts of wheat embryo or alfalfa does not improve this medium. Yeast has been growing in the above medium for nearly 2 years being subcultured at least 300 times, each time with a dilution of 1:50. In a further communication Nelson, Fulmer, and Cessna (3) presented data showing the synthesis of water-soluble B by yeast. In a recent paper by Eddy, Heft, Stevenson, and Johnson (4) certain exceptions are taken to the work mentioned above and, 1 Reports on which were read at the 1920 spring meeting of the Ameri- can Chemical Society, St. Louis. 4a 78 Vitamine B and Bios in Yeast Growth furthermore, some interpretations have been placed upon that work which we feel to be unjustified. For instance, they state: ‘In this connection they [Fulmer, Nelson, and Sherwood] propose a formula for the culture medium (Medium F) which they believe will pro- duce maximum stimulation to yeast growth and which will not be improved or altered by the addition of extracts of organic substances unless the latter disturb the optimum salt concentrations of the medium.”’ We have not maintained that Medium F will produce maximum stimulation of yeast growth. If such were the case it would imply that Medium F is as good a medium as beer-wort. What we said and implied was that Medium F is the best synthetic medium for the growth of yeast which has been developed up to the present time and that it is the best possible medium which can be made from the constituents employed. Furthermore, we do not maintain that Medium F cannot be improved by means of organic extracts. Since Medium F is not as good a medium as is beer-wort it would necessarily follow that certain organic extracts would improve Medium F. We maintain only that the addition of alcoholic extracts from alfalfa and wheat embryo, containing water-soluble B, will not improve the medium. The same authors go on to say: ‘In brief, they [Fulmer, Nelson, and . Sherwood] imply that ‘bios’ is simply a matter of salt concen- tration.” In a recent communication Funk and Dubin (5) apparently are unable completely to substantiate the work of Fulmer, Nelson, and Sherwood, and agree with the conclusions of Eddy, Heft, Stevenson, and Johnson (4). Moreover, they maintain that Medium F is not as good as Nageli’s solution for the growth of yeast, a conclusion at variance with that of Eddy, Heft, Stevenson, and Johnson (4) who say that “Medium F is superior to Nigeli solution as a basal medium.” Furthermore Funk and Dubin (5) propose a new name for bios, namely vita- mine D, a procedure which we believe to be wholly unnecessary and confusing. The term bios has proved satisfactory and, moreover, the name vitamine D should be reserved for any new vitamine that may subsequently be found necessary for the growth and well being of the animal. There seems to be a tendency to confuse the terms vitamine, bios, and auximone. Vitamines are materials of unknown compo- sition, which are necessary in the diet, other than fats, proteins, ih Mia E. I. Fulmer and V. E. Nelson 79 carbohydrates, and mineral salts, for the growth and well being of the animal. Bios is the term originally applied by Wildiers (6) to designate substances of unknown composition which are neces- sary for the best growth of yeast. Auximone is a term originally used by Bottomley (7) to designate those substances of unknown composition believed to be necessary for the growth of plants. Any interchange of the above terms is in our present state of knowledge unjustified. While yeast will grow so satisfactorily in Medium F that addition of water-soluble B will not improve it, the addition of certain other extracts will improve the medium greatly. Our point was that bios and water-soluble B are not identical. In the paper referred to (Eddy, Heft, Stevenson, and Johnson, 4) they agree with us on the first two of our claims; 7.e., that the relative stimulating potencies of two materials cannot be arrived at on an equal weight basis, and that alkali does not destroy the stimulant; as a result, at least temporarily, they discard the yeast growth stimulation method for the estimation of water- soluble B. : The authors do, however, take definite exception to our state- ment that medium F is not improved by the addition of extracts (alcoholic) of alfalfa and come to the conclusion ‘‘that while Medium F is superior to Nigeli solution as a basal medium it is still capable of stimulation by every concentration of the extract used.” It seemed advisable, in view of the above statement, to check our previous data. Dried alfalfa was extracted for 12 hours in a continuous extractor with 95 per cent alcohol (the method out- lined in our previous papers, 2). Varying concentrations of the extract were added to Medium F which had the following compo- sition: 100 ce. contained 0.188 gm. of ammonium chloride, 0.100 gm. of calcium chloride, 0.100 gm. of dipotassium phos- phate, 0.04 gm. of precipitated calcium carbonate, 0.60 gm. of dextrin, and 10 gm. of cane-sugar. Of course, it is necessary when the dextrin is prepared by heating starch with citric acid to extract the acid with alcohol previous to its use in the medium. The flasks were inoculated with an initial count (when the count = 1, there are 250,000 cells per cubic centimeter) of about four, the yeast being taken from cultures which had been growing in 80 Vitamine B and Bios in Yeast Growth Medium F for 2 years. The cultures were incubated at 30°C. After 48 hours the yeast count was determined. The results are given in Table I, Column 1. By ‘‘dry equivalent” is meant the number of grams of dry alfalfa added as extract per 100 cc. of medium. The addition of alcoholic extract of alfalfa did not improve Medium F, thus verifying our previous statement that not only is water-soluble B not necessary for the growth of yeast but also it is of no advantage in Medium F. A study of the methods used by Eddy, Heft, Stevenson, and Johnson, revealed two points wherein their method differed from that used by us and which might account for their failure to TABLEI. (1) (2) (3) Dry equivalent. Alcoholic extract. Pigeons caen penal 0 214 214 0.10 216 250 0.20 214 260 0.40 221 300 0.60 223 400 0.80 220 325 1.00 206 306 1.20 226 300 1.40 193 246 2.00 184 check our results. First, there is the matter of temperature con- trol. They used temperatures varying between 30-35°C. Medi- um F is optimum for 30°C. and for that temperature only. Secondly, the above authors did not use a 95 per cent alcoholic extract of alfalfa. Their method was as follows: ‘400 gm. of dried alfalfa were repeatedly extracted with distilled water, the filtered extracts combined . . . .” They sometimes pre- pared the extract by ‘‘diluting the water extract with 95 per cent alcohol to 40 per cent of the volume and filtering off the precipi- tated protein complex.” Extract of alfalfa was prepared according to the method out- lined above. While the alcoholic (95 per cent) contained 0.20 gm. of dry material per 100 ec. the water extract contained 0.70 gm. of dissolved material per 100 ce. E. I. Fulmer and V. E. Nelson 81 Varying concentrations of the extract were added to Medium F in the same manner as outlined above for the 95 per cent alcohol extract. The results are tabulated in Table I, Column 3. The addition of the water extract did stimulate the growth of yeast in Medium F. This extract evidently contains, along with water- soluble B and other materials, bios, the yeast growth stimulant. The 95 per cent alcoholic extract contains water-soluble B but not bios. We wish to thank Mr. H. W. Wright for aid in obtaining experimental data. SUMMARY. 1. Data is presented in this paper which verify previous work to the effect that 95 per cent alcoholic extract of alfalfa does not improve Medium F for the growth of yeast. 2. Two explanations and supporting data are given why Eddy, Heft, Stevenson, and Johnson, and also Funk and Dubin, were unable to corroborate our results. BIBLIOGRAPHY. jot . Williams, R. J., J. Biol. Chem., 1919, xxxvili, 465. 2. Fulmer, E. I., Nelson, V. E., and Sherwood, F. F., J. Am. Chem. Soc., 1921, xliii, 186, 191. 3. Nelson, V. E., Fulmer, E. I., and Cessna, R., J. Biol. Chem., 1921, xlvi, rie 4, Eddy, W. H., Heft, H. L., Stevenson, H. C., and Johnson, R., J. Biol. Chem., 1921, xlvii, 249. . Funk, C., and Dubin, H. E., J. Biol. Chem., 1921, xlviii, 437. . Wildiers, E., La Cellule, 1901, xviii, 313. . Bottomley, W.B., Proc. Roy. Soc. London, Series B, 1915-17, Ixxxix, 102, 481. “IQ O1 ~ Orn eae : - me mas oT a 7 F- Wel Ts fi as ¥ . A REPLY TO FULMER, NELSON, AND SHERWOOD CON- CERNING MEDIUM F. By WALTER H. EDDY, H. L. HEFT, ann H. C. STEVENSON. (From the Department of Physiological Chemistry, Teachers College, Columbia University, New York.) (Received for publication, December 21, 1921.) Opportunity having been given us to examine the preceding paper prior to its publication we take this opportunity to present certain comment thereon. In our original paper we were concerned solely with the question of whether the stimulatory action of vitamine B-containing extracts owe their stimulatory effect to the vitamine and whether the yeast test first suggested by Williams and later modified by several workers is a reliable index of the vitamine B content of an extract. In the two papers of Fulmer, Nelson, and Sherwood that were before us at the time! the claims bearing on this point were as follows: (1) Alcohol extracts of alfalfa and wheat embryo do stimulate yeast growth when added to synthetic media of the Williams’ type. (2) Treatment of these extracts by alkali fails to reduce their stimulatory powers. (3) Alcohol extracts of alfalfa and wheat embryo fail to show stimulatory effect when added to a medium devised by the authors and called by them Medium F. The authors say: “From the above data it is evident that we have developed a medium, namely Medium F, composed of known constituents, which is not improved by the additions of vitamine containing extracts.” We reported in our paper substantial agreement with all except one of the claims of the authors (superiority of Medium F over the Williams’ type, the negative effect of alkali upon the stimula- tory extracts, and the impossibility of using the test in its present form as a test for vitamine content). We believed at the time that the authors had used alcohol extracts solely as a means of obtaining a vitamine B extract, after the method devised by 1 Fulmer, E. I., Nelson, V. E. and Sherwood, F. F., J. Am. Chem. Soc., 1921, xliii, 186, 191. 83 84 Medium F McCollum for extraction of that vitamine from the navy bean. We used a water extract for the same purpose, as it had been repeatedly shown by various investigators that 95 per cent alcohol is a relatively poor extractant of vitamine B. We had no idea at the time that water and alcohol extracts were to be considered as qualitatively different for the purpose in mind, viz. to determine the effect of vitamine B on the medium. We, therefore, believed our criticisms of the conclusions were a correct statement in view of the fact that our water extracts did stimulate Medium F. In the present paper the authors repeat our work with water extract and confirm our results. They repeat their own work with alcohol extract and confirm their own results. They then criticize our interpretation, made on the above basis, and advance a new view-point, vzz., that water extracts both bios and vitamine B but alcohol extracts only vitamine B. Herein is a suggestion quite in accord with that of Funk, and forecast by Emmett, namely that the yeast stimulatory substance is something other than vitamine B and in the present paper admittedly not present in Medium F. They prefer to retain for this factor the term bios, Funk to call it vitamine D. Emmett declined to name it, but . suggested it was unlike vitamine B in not being antineuritic. While such a view is quite possibly accurate, it has seemed to us desirable first completely to exclude vitamine B from the factors involved in the yeast growth and evidence to this effect seems at present definitely lacking. We have not yet had the opportunity to fully investigate the relative properties of alcohol and water extracts of alfalfa but we have prepared an alcohol extract of this substance after the manner of Fulmer and associates,? and used this in the concentrations given by the authors and in higher concentrations. In our incubations this time, our incubator was set at exactly 30°C. The results are given in Table I. We continued to use the Funk method of measurement as we believe it to be more accurate as a measure of total growth than their method. The results make us sceptical as to the qualitative differences in alcohol and water extract as we again get stimulation of Medium F and with the alcohol extract. The stimulation is not as great as 2 Fulmer, Nelson, and Sherwood,! 187. Eddy, Heft, and Stevenson 85 with the water extract, which is quite in harmony with the view we had held, that alcohol is a poorer extractant of vitamine B than water. It-is also to be noted that the results obtained with us which showed greatest stimulation are with concentrations much greater than those used by Fulmer and associates, which facts combined with what we believe to be greater accuracy in our measuring, could easily account for their failure to observe stimulation with the alcohol extract. TABLE I. Growths ob- tained by us Growths obtained by us and reported in with an alcohol previous paper when extract of alfalfa water extract of alfalfa was used. Seenba- | inuiactae meentra- | 1000ths of a cc. eee temeed | of yeast cle, | | tions used by us in Funk method by Fulmer, z! uickmnctis Nelson, and oe Sherwood. Walesa nad Duplicate Dry equivalent | Corresponding Dry equiva- Sherwood’s | determinations. expressed concentrations| 4 Varage of lent in 100 ce. k ___soanfter manner | as expressed in fi iets : abe of Fulmer, Table III, bedi cesiag = Nelson, and | Series C of our cae 1 2 & Sherwood. original paper. S _. =. = SE SE en 0.00 0.00 15440) 11 0.00 Control. 15 0.10 0.10 15) to | 15 ‘ 0.20 0.20 AO P15; |, 12 0.40 0.40 13 | 17 | 15 0.60 0.60 15 | 16) 15 0.60 0.015 19 0.80 0.80 Lael lo: Ae: 1.00 1.00 Sp Los 14 1.00 0.025 18 1.20 1.20 16 | 15 | 15 1.40 1.40 TZ | 22°| 19 1.40 0.035 19 2.00 2.00 1G TAG 1.80 0.045 19 2.50 16 | 20 | 18 3.00 0.075 22 5.00 16 | 18 | 17 5.00 0.125 25 We have no desire to develop controversy in this field. Our own results, including those accumulated since the publication of our paper, have only served to increase our distrust of the test as an accurate measure of vitamine B but they have also clearly indicated the need of very careful study to evaluate all the factors concerned in the response of the yeast cell and if these factors can be evaluated we believe that it may result in showing that vitamine B plays a part in the action and in perfecting the test for the detection of that factor. r ts oO Se eee ere: _— ‘ i _ é e © P| f q Prlp + yeti av oh ee F Aa i‘. j + i . - . , , ~~ LS * f THE HYPOBROMITE REACTION ON UREA. By PAUL MENAUL. (From the Oklahoma Agricultural Experiment Station, Stillwater.) (Received for publication, October 4, 1921.) In the determination of urea in the urine of experimental animals a discrepancy in the results was noticed upon using the urease method of Folin and Youngburg! and the hypobromite method of Stehle,? the hypobromite method giving low results. Two published articles,** were brought to my notice by Dr. C. T. Dowell of this Department, as bearing upon the subject. These articles declare the formation of dichloro urea and nitrogen tri- chloride in reactions between chlorine and urea. Only 93 to 96 per cent of the nitrogen being liberated in gaseous form, showing that this is not a stoichiometric reaction, which is also shown by Krogh,* who finds that in the reaction of sodium hypobromite on urea, that part of the nitrogen is set free, and that the rest is con- verted into oxides of nitrogen. With a variation in the concentra- tion of the bromine, there is also a variation in the amount of gas liberated, a more dilute solution of bromine liberating more gas. The volume of nitrogen varied from 86.8 to 96.5 per cent of the nitrogen in the urea. By what law of physical chemistry should such an oxidation reaction be altered by vacuum? The reaction between sodium hypobromite and urea is nearly instantaneous, and it is not apparent to the author why Stehle considers that in his method the reaction takes place zn vacuo, for it is apparent that the reaction: has proceeded for some time before any vacuum can be obtained. In order to solve this question a number of determinations were made using pure urea oxalate. The apparatus and pipette were calibrated with mercury. The hypobromite solution was added 1 Folin, O., and Youngburg, G. F., J. Biol. Chem, 1919, xxxviil, 111. 2 Stehle, R. L., J. Biol. Chem., 1921, xlvii, 13. 3’ Chuttaway, F. D., Chem. News, 1908, xeviii, 285. 4 Dowell, C. T., J. Am. Chem. Soc., 1919, xli, 124. 5 Krogh, M., Z. physiol. Chem., 1913, Ixxxiv, 379. 87 88 Hypobromite Reaction on Urea through the exit tube (of the Van Slyke apparatus) subjected to vacuum, the air expelled, then the urea solution and rinse water were added through the graduated cup, the air dissolved in them having been previously determined. The following is a table of representative results obtained by following the procedure in the original article of Stehle, using also the strength of bromine recommended by Dehn and Krogh.’ TABLE I. Nitrogen in Urea Determined by the Following Methods. Hy pobromite. Folin and Youngburg. Kjeldahl. Theoretical. 6 per cent bromine. mg. iadenen: mg Meoretical mg. mg. 1.000 93.98 1.062 99.81 1.064 1.064 1.016 95.48 1.062 99.81 1.064 1.01 94.93 1.060 - 99.62 1.015 95.39 1.063 99.91 1.020 95.86 1.063 99.91 1.016 95.48 1.062 99.81 0.644 - 96.84 0.662 99.55 0.665 0.665 0.646 97.14 0.663 99.69 0.665 0.643 96.69 0.663 99.69 0.640 96.25 0.663 99.69 10 per cent bromine. 1.00 93.98 1.064 1.02 95.86 1.00 93.98 1.00 93.98 1.01 94.93 0.5 per cent bromine. 1.01 94.95 1.064 1.03 96.8 1.03 96.8 1.02 95.86 1.03 96.8 These figures are in accordance with the findings of Chattaway,° Dowell,* and Krogh,* that there is not a quantitative liberation of nitrogen when urea is acted upon by hypobromite. 6 Dehn, W. M., J. Am. Chem. Soc., 1907, xxix, 1317. NOTE ON THE GASOMETRIC DETERMINATION OF UREA. By RAYMOND L. STEALE. (From the Laboratory of Pharmacology, McGill University, Montreal, Canada.) (Received for publication, December 13, 1921.) This note is occasioned by the preceding communication, for an early knowledge of the contents of which the writer is indebted to the author. Additional analyses have been conducted on pure urea and urea oxalate solutions with the following results: Urea oxalate N. Urea N. Found. Calculated. Found. Calculated. 38.9 38.7 193 195 39.5 38.7 198 195 74.4 74.6 198 195 75.0 74.6 193 196 73.9 74.6 193 196 wouo 74.6 193 196 38.3 38.0 In calculating the results of urea determinations by the gaso- metric method described by the writer? a certain amount of error is involved. The solution left in the apparatus at the end of a determination is a rather concentrated one and hence it is not correct to assume that its vapor tension is the same as that of water. This erroneous assumption was made in the original de- scription of the method because vapor tension data for hypobro- mite solutions were unknown to the author at the time, and the error introduced is not large enough to invalidate the method for 1Menaul, P., J. Biol. Chem., 1922, li, 87. *Stehle, R. L., J. Biol. Chem., 1921, xlvii, 13. 89 90 Gasometric Determination of Urea most purposes. Recently, some data on this point in a paper by Dehn’ have come to the writer’s attention. In order to make them applicable to the method it is necessary to prepare the hypobromite solution as described by Dehn with the omission of the final step; viz., doubling the volume with water. Then by using 1 cc. of the solution to be analyzed, 1 cc. of rinse water, and 2 ec. of the hypobromite solution the vapor tension of the solution in the apparatus should approximate closely the tension called for by Dehn’s table. Dehn’s solution is prepared by dissolving 100 gm. of NaOH in 250 cc. of water, adding 10 cc. of Br to each 100 cc., and doubling the volume with water. The vapor tensions are as follows: Temperature. NaOBr H20 Difference. °C. mm. mm. mm. 4 nF -6.1 5.0 8 1.8 8.0 6.2 12 3.3 10.5 7.3 16 Do 13.6 8.3 20 8.6 17.4 8.8 24 12.8 22.2 9.4 28 18.0 28.1 10.1 32 24.0 35.4 11.4 36 29.5 44.2 14.7 40 36.0 55.0 19.0 These data were utilized in the calculation of the results reported above. ‘The dissolved air contained in this hypobromite solution is negligible. Concerning the work of Chattaway,* Dowell,> and Krogh,‘ all of which Menaul states to show the urea-hypobromite reaction not to be stoichiometric, the following comments are apropos. In the paper by Chattaway an explanation is suggested for the failure of the urea-hypobromite reaction to give quantitative results. When urea and hypochlorous acid react in acid solution dichloro urea is formed. This substance hydrolyses easily according to the equation 3 Dehn, W. M., J. Am. Chem. Soc., 1907, xxix, 1317. 4 Chattaway, F. D., Chem. News, 1908, xeviil, 285. 5 Dowell, C. T., J. Am. Chem. Soc., 1919, xli, 124. 6 Krogh, M., Z. physiol. Chem., 1913, lxxxiv, 379. R. L. Stehle 91 NHC ou co +2H.0=CO +2NH,CI \NHCI \ou The NH.Cl is then assumed to decompose as follows: The end-products then ae react, forming nitrogen and hydrogen chloride, the latter at once combining with the free ammonia and allowing the remaining nitrogen trichloride to escape, as this does not react with ammonium chlo- ride. . . . In presence of alkalies, on the other hand, the reaction between the ammonia and the nitrogen trichloride goes on to completion, since the hydrochloric acid formed in it is at once fixed; no nitrogen tri- chloride therefore is set free, since twice as much ammonia is formed as is required to decompose it.’’ In order to account for the incompleteness of the urea-hypobro- mite reaction Chattaway continues: “The urea, is, without doubt, converted into dichloro or dibromo urea, which is at once hydrolyzed in the manner above described. In presence of the excess of hypochlorite or hypobromite, the mono substituted ammonia formed in the hydrolysis may be further substituted to a greater or less extent, nitrogen being evolved quantitatively only when this takes place under such conditions that the amount of hydrogen attached to nitrogen in the reacting system is always sufficient to react completely with the chlorine attached to nitrogen.” Two facts concerning Chattaway’s paper are these: (1) It is not stated that the reaction necessarily falls short of being quantita- tive; and (2) no data are presented to show that with alkaline hypobromite nitrogen tribromide is formed. Conditions which might favor its formation are suggested but that is all. Indeed from the foregoing quotations one would be inclined to believe that the reaction would be quantitative rather than otherwise. __ The paper by Dowell concerns the action of chlorine upon urea. This is somewhat different than the action of a strongly alkaline hypobromite solution and hardly justifies the conclusion that the low results of the urea-hypobromite reaction as observed in pre- vious methods are due to the formation of nitrogen tribromide. It is worth noting that Dowell found no evidence of the forma- tion of NH;Cl in the hydrolysis of dichloro urea as assumed by 92 Gasometric Determination of Urea Chattaway. Neither did he observe the oxidation of nitrogen to nitric or nitrous acid. This latter fact is interesting in connection with the following comment on the work of Krogh: Krogh* has studied the urea-hypobromite reaction as it is ordinarily carried out and found that when the hypobromite solu- tion is poor in Br some CO (0.62 per cent is the highest value recorded) may be formed but that when the solution is rich in Br little or no CO is evolved and the evolution of N is also less quantitative (13.8 per cent is the greatest deficiency recorded). The low N Krogh believed to be due to the formation of oxides of N. She obtained a qualitative test for nitric acid. In order to learn whether similar results could be obtained in the vacuum apparatus a hypobromite solution similar to that employed by Krogh in obtaining the last mentioned result was employed. The liberation of nitrogen was complete. Indeed the hypobromite solution of Dehn contains more Br than that employed by Krogh yet gives satisfactory results. It appears, therefore, that when conducted in vacuo the reaction does not follow the same course as when conducted at atmospheric pressure. If carrying the reaction out 7m vacuo has no effect on the result one may wonder why Menaul’s results are surprisingly independent of the hypo- bromite concentration instead of in accord with the results of Krogh on this point. THE CARBOHYDRATE CONTENT OF THE SEED OF ASPARAGUS OFFICINALIS L. By W. E. CAKE anp H. H. BARTLETT. (From the Laboratory of Plant Chemistry of the Department of Botany, University of Michigan, Ann Arbor.) (Received for publication, December 27, 1921.) INTRODUCTION. The reserve hemicelluloses of seeds have been investigated especially by Schulze! and his associates, to whose numerous papers we owe most of our present knowledge of these com- pounds. They occur as stored foods in seeds of plants belonging to the most distantly related groups, and offer a large field for biochemical study. It has been found that in most cases the reserve hemicelluloses yield galactose and mannose upon hydroly- sis, and are, therefore, either mixtures of mannan and galactan or else galactomannans. Baker and Pope? found that vegetable ivory, the hemicellulose from seeds of Phytelephas, yielded mannose and about 5 per cent as much fructose, pointing to the existence of a fructomannan. Our results, showing a large percentage of fructose among the cleavage products of asparagus seed, suggest that fructose may have been overlooked in much previous work. If this surmise should not be true, the presence or absence of fructose will afford an important criterion in the further study and classification of the hemicelluloses. In general, it may be said that little is known about the constitution of individual hemicelluloses, since methods have not been devised for differen- tiating them as chemical individuals. In the absence of such methods, complete analytical data, accounting for all cleavage products, may throw light upon their constitution. It is hoped 1Schulze, E., and associates. For list of references see Czapek, F., Biochemie der Pflanzen, Jena, 2nd edition, 1913, i, 420, 655. 2 Baker, J. L., and Pope, T. H., Proc. Chem. Soc., 1900, xvi, 72. 93 94 Seed of Asparagus officinalis L. that the results presented in this paper will be useful along this line. After the removal of the large oil content of asparagus seed, the hemicellulose may be extracted by dilute alkali and precipi- tated either by acidifying or by adding alcohol. The amorphous precipitate may be somewhat purified by redissolving and re- precipitating. After drying, t turns to a dark blue color with strong iodine solution, and may quickly be washed colorless by suspending it in water and decanting two or three times. In the analysis of asparagus seed by acid hydrolysis the poly- saccharide is accounted for as mannose, fructose, glucose, and galactose. The total mannose is only 24.2 per cent, and the other three sugars total 21.9 per cent, suggesting that there is an equivalent of mannose for each equivalent of every other sugar. Since starch was not present in the seeds, nor inulin or any similar compound, and since cellulose was not attacked in the method of hydrolysis used, it is probable that the individual polysaccharide or polysaccharides of the asparagus seed are mixed, each being a condensation product of two or more hexoses. Unless the hemicellulose molecule be excessively large, it is fair to assume that a very small galactose content is derived from a polysaccha- ride quite distinct from that yielding the bulk of the mannose. This polysaccharide might be a galactomannan, a glucogalactan, or a simple galactan—most likely a galactomannan. The bulk of the polysaccharide might conceivably be of a single kind, a glucofructomannan, without requiring the hypothesis of a very large molecule, or it might be a mixture of two or more simpler kinds, as for example a glucomannan with a fructomannan. At any rate, the occurrence of mannose in a practically 1:1 ratio with the total remaining hexoses is distinctly suggestive of the possible manner in which the hemicellulose is constituted. Two former papers touching on the chemistry of asparagus seed have been published—those of Reiss* and of Peters. The latter worker estimated all the hydrolytic sugar of the seed as mannose, failing to detect the other hexoses. He did not report even galactose, as it would be inferred he did from the reference to 3 Reiss, R., Ber. chem. Ges., 1889, xxii, pt. 1, 609. 4 Peters, W., Arch. Pharm., 1902, cexl, 53. W. E. Cake and H. H. Bartlett 95 his paper in Czapek.* The earlier paper of Reiss was the first significant investigation in the field of hemicellulose chemistry, and dealt with asparagus seed only incidentally, as one of the sources of his newly. discovered sugar, ‘‘seminose,’’ afterwards found to be the same as mannose. Material. The seeds of Asparagus officinalis L. used in this work were obtained from the firm of D. M. Ferry and Co. of Detroit, Michi- gan, and were of the variety ‘‘Palmetto.”” They were first separated from most foreign matter by picking them over, and then ground in a large grinding mill until they were fine enough to pass through a 40 mesh sieve. The sample was thoroughly mixed and kept in a sealed jar until needed for use. Preliminary Examination. An analysis was made of the sample so prepared according to the methods of the Association of Official Agricultural Chemists® TABLET. Preliminary Analysis of Asparagus Seed, var. ‘‘Palmetto.’’ per cent LOT SS. ee a 1S SR eT Pee eer ee 10.41 ee ee ye ok. aalsiate vine aie 01a @ ein 0e mye 0[diaje ern 9:0cidjein aie 2.02 2 OLo EPEC E Di og. 2 oes eee 13.59 EIIIRINGRGEET 09253)... 52 ee een ee ene eee ee ees 17.38 yi it pe eS i 0 4.94 Nitrogen-free extract (by difference)....................0-0eeeeeeee 51.66 except that for the determination of nitrogen, used in computing the “crude protein,’ the iodometric method of Willard and Cake? was used. These analyses gave the results shown in Table I. Part of the total nitrogen of the asparagus seeds is contained in the black seed coats. The black pigment in these seed coats seems to be of a melanoid nature, and is being investigated further. The relatively high percentage of ether extract, a fixed oil, is interesting. Work has been about completed on a study of the 5 Czapek, F., Biochemie der Pflanzen, Jena, 2nd edition, 1913, i, 421. 6 Official and tentative methods of analysis of the Association of Official Agricultural Chemists, 1920. 7 Willard, H. H., and Cake, W. E., J. Am. Chem. Soc., 1920, xlii, 2646. THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL, LI, NO. 1 96 Seed of Asparagus officinalis L. physical and chemical properties of this oil, and these results, in some respects at variance with those of Peters,* will be published shortly. Qualitative Study of the Carbohydrate Content. Qualitative experiments were first conducted in order to deter- mine what sugars were produced by acid and enzyme hydrolyses of the asparagus seeds. For this purpose one portion of the seeds was boiled with 2 per cent sulfuric acid, and after neutra- lization with calcium carbonate and purification with lead ace- tate, the osazones of the sugars were prepared by heating with an excess of phenylhydrazine in dilute acetic acid solution. To prepare the sugars resulting from enzyme hydrolysis another portion of the seeds was placed in water to which a little chloro- form had been added, and allowed to undergo autolysis. A portion of this solution after purification with lead acetate was heated with an excess of phenylhydrazine in dilute acetic acid solution. In both cases the osazones were separated by recrystal- lization into ten different fractions, and each fraction proved to contain nothing but phenylglucosazone (M. P. 208-213°). This indicated that no sugars could have been present in the original solution other than glucose, mannose, and fructose, except per- haps in mere traces. For further identification of the sugars present their hydrazones were prepared by adding phenylhydrazine acetate to the cool neutral solutions. There separated at once the white insoluble precipitate of mannose phenylhydrazone, which after recrystal- lization from 60 per cent alcohol, melted at 194-195°. This hydrazone was further identified by heating with an excess of phenylhydrazine in dilute acetic acid solution, when phenylglucos- azone was formed (M. P. 210°). A portion of the purified mannose phenylhydrazone was reconverted into mannose by the formaldehyde method of Ruff and Ollendorff,? as modified by Browne and Tollens.? This mannose was crystallized from 80 per cent alcohol, and the anhydrous crystals so obtained melted at 131.° All attempts to bring any other phenylhydrazones to crystalli- zation in the filtrate from the mannose phenylhydrazone proved ® Ruff, O., and Ollendorff, G., Ber. chem. Ges., 1899, xxxii, 3234. * Browne, C. A., Jr., and Tollens, B., Ber. chem. Ges., 1902, xxxv, 1457. W. E. Cake and H. H. Bartlett 97 futile. This filtrate, however, on heating with an excess of phenylhydrazine in dilute acetic acid solution yielded large quantities of glucose phenylosazone (M. P. 206-207°). Since the mannose had been almost quantitatively removed as the insoluble phenylhydrazone, the remaining sugar could have been only glucose or fructose or both. A filtrate from the precipitation of mannose phenylhydrazone, containing about 25 gm. of the sugars in the form of their phenyl- hydrazones, in 200 ce. of water and 280 cc. of alcohol, was refluxed for 5 hours with 32 gm. of benzaldehyde according to the method of Herzfeld'® and de Witt, for converting the hydrazones back into sugars. After the removal of the benzaldehydrazone, excess benzaldehyde, benzoic acid, and alcohol, the solution was boiled with a little animal charcoal and filtered. In this solution the presence of glucose was established by oxidation with 25 per cent nitric acid according to the method of Gans and Tollens." On neutralization with potassium carbonate and making slightly acid with acetic acid, the characteristic potassium salt of saccharic acid crystallized out. The presence of glucose was further indicated by the optical rotation of the solutions, as shown in Table IV. The presence of fructose was established both in the original solution resulting from the acid hydrolysis of the asparagus seeds, and also in the solution after removal of the mannose, by the Seliwanoff” test, of heating a few drops of each with con- centrated hydrochloric acid and a little resorcinol—a _ bright cherry-red color developing almost ‘mmediately. The diphenyl- amine reaction of Ihl-Pechmann carried out as directed by Jolles'® also gave positive results—a deep blue color being developed in 2 minutes. In addition, the methyl fructosazone was prepared from the solution after the removal of mannose, according to the directions of Neuberg. This osazone melted at 158-159°. 10 Herzfeld, A., Ber. chem. Ges., 1895, xxviii, 442. 11 Gans, R., and Tollens, B., Ann. Chem., 1888, cexlix, 219. 1 Seliwanoff, T., Ber. chem. Ges., 1887, xx, 181. 13 Jolles, A., Ber. pharm. Ges., 1909-10, xix, 484. 14 Neuberg, C., Ber. chem. Ges., 1902, xxxv, 959. 98 Seed of Asparagus officinalis L. Quantitative Examination of the Carbohydrates. The constituents of the carbohydrate content of asparagus seeds were quantitatively determined in the ether-extracted samples according to the following methods: Total Reducing Sugars on Acid Hydrolysis ——The sample was re- fluxed for 23 hours with 3 per cent hydrochloric acid. After neu- tralization the reducing sugars were determined in an aliquot part by means of Fehling’s solution. The precipitation was carried out according to the directions of Munson and Walker,’ and the copper in this precipitate estimated by the iodometric method of Low.‘ Reducing Sugars Resultant from Hydrolysis of Polysaccharides.— The oil-free sample was extracted with 150 cc. of 10 per cent alcohol on a hardened filter paper, the insoluble residue hydrolyzed, and the reducing sugars in the solution so obtained were estimated by Fehling’s solution. Free Sugars.—The difference between the total reducing sugars on acid hydrolysis and the reducing sugars resulting from the acid hydrolysis of insoluble polysaccharides gave the free sugar content. (It became obvious from the analytical results that none of the free sugar was mannose.) Starch.—The oil-free seeds gave no blue color on treatment ‘with iodine, nor did malt diastase affect the ether-extracted sample. Pentosans.—The sample was boiled with 12 per cent hydro- chloric acid and the furfuraldehyde in the distillate estimated as the phloroglucide according to the directions given in the Official Methods. The pentosan results are apt to be a little high since the hexoses on distillation with hydrochloric acid yield small amounts of furfuraldehyde, a fact of especial im- portance in connection with this work, since mannose is especially apt to give rise to furfuraldehyde. Fischer and Hirschberger’’ have shown that this change occurs even when a 5 per cent solu- tion of mannose is heated in a sealed tube at 140° for 4 hours. 15 Munson, L. S., and Walker, P. H., J. Am. Chem. Soc., 1906, xxviii, 663. Walker, P. H., J. Am. Chem. Soc., 1907, xxix, 541. 16 Low, A. H., J. Am. Chem. Soc., 1902, xxiv, 1082. 17 Fischer, E., and Hirschberger, J., Ber. chem. Ges., 1889, xxii, 365. W. E. Cake and H. H. Bartlett 99 Galactans.—The sample was treated with 25 per cent nitric acid and the mucic acid estimated by the method given in the Official Methods.® Mannose.—The oil-free sample was refluxed for 23 hours with 3 per cent hydrochloric acid. After cooling, the solution was neutralized with dilute sodium hydroxide, then 2 cc. acetic acid were added. The solution was next filtered and the residue thoroughly washed with water. The filtrate was concentrated on a water bath to about 30 cc. and then cooled The mannose TABLETI. Determination of the Total Free Sugar and of the Several Classes of Polysaccharides. Non- s bl summable | percentages. per cent per cent Total sugars on acid hydrolysis................. 54.5 ‘Sugars from hydrolysis of polysaccharides...... 50.4 ss 4.1 eh chee cnr cee aie oo Crees ee = 45.4 or cine win tae bate eewssses’ 2.0 oe ee eer 0.4 Polysaccharides yielding glucose, fructose, and TLL LLIt.. 22234 55062 BAR 43.0 Total carbohydrates.................56.5 se. 49.5 in this solution was now determined as mannose phenylhydrazone according to the method of Bourquelot and Hérissey.1® Glucose: Fructose Ratio.—The oil-free sample was hydrolyzed with 3 per cent hydrochloric acid. After neutralization with dilute sodium hydroxide, and making slightly acid with acetic acid, the solution was purified with lead acetate. The solution after filtration was freed from excess lead by hydrogen sulfide. After filtration the hydrogen sulfide was removed by blowing a stream of air through the solution, which was finally heated to boiling. The total sugars were determined in an aliquot portion. The aldoses were determined in another aliquot by the hypoiodite 18 Bourquelot, E., and Hérissey, H., Compt. rend. Acad., 1899, exxix, 339. 100 Seed of Asparagus officinalis L. method of Willstitter and Schudel.!® Since the ratio of mannose to total sugars had already been determined, it was simply a matter TABLE III. Determination of Aldoses and Ketoses by Willstatter’s Hypoiodite Method. Non- s bl percentages, | PeFeentages per cent Aldoses in ‘total sugatses.. «4. <5 «) a2 eee eee « © insoluble polysaccharides........... Wree aldosess is. 22 ae auees «08 ss ease oes 3.0 Wondensed aldoses;o os) 2: 22 nos oe ee eee 39.4 Ketoses:in total suvars....0052-.csees- cere eee << © condensed: SUgZATS ©. ..,..--,2 ee eee eee Free-kcetoses) ..., ncice ter store owies ee le ee 1.0 Coridensed: ketoses?. 5c... °. - 240s 0. se eee 6.3 Totals cce . aA se ena ee See eee 49.7 TABLE IV. Determination of Glucose, Fructose, and Mannose from the Optical Activity. Non- - Ss bl ponaemabes, | pereentages. per cent per cent Mannose:initotalisugats:2.<22uc--e seer 24.2 : “« polysaccharides: ...2:55<5-5eeee ee 24.2 Nrée Mannose 22 fess cade vice ae See eee 0.0 Condensed .mantiose << ¢222 S500 Gan vessel eee 21.8 Glucose m.fotal sugars... 22. --o0 seacsseeeeaees 20.1 re “ polysaccharides, . 0. o)5-7..5>-aee ee 16.8 Bree @lucos@s.o2 iss pos 585 33 php a Ore 3.3 Condensed Chitose 24.0. oc a cieeoe tee hin veel ee 15.1 Fructose in! total guears! S232 «. 2+ soe eo 8.5 rf “* nolysaccharides:: 24s.0& . 2 eae yee ree. fructosene. 5. to cae pcos ae see ee 1.4 CGendensed fructose... ..< - 28 323-6 e020 sced eee 6.4 Dotal3; 20 Beck IN Os ieee 48 .0 of computation to determine the glucose: fructose: mannose ratios. The same process was carried through after the removal of the free sugars. 19 Willstaitter, R., and Schudel, G., Ber. Chem. Ges., 1918, li, 782. W. E. Cake and H. H. Bartlett. 101 Knowing the total sugar content and the mannose content of a solution of glucose, fructose, and mannose, it is possible to calculate the ratio of these three sugars from the optical activity of their solution. Accordingly, determinations were made of the optical activity of a purified solution of the total sugars obtained by acid hydrolysis. The same was done with the solu- tion resulting from acid hydrolysis of the sample after the re- moval of the free sugars. RESULTS. In the determination of the glucose: fructose ratio, 2.24 per cent was subtracted from the sugar content of both solutions TABLE V. General Summary of Analysis of Asparagus Seed, var. ‘‘Palmetto.’’ per cent Sa noe eo So wails crenci eo gyapd Sip) sve 20 /x 6 wisiein vale eG sree 8258 10.41 en ee ois wre a dren cep ne Wcpe ules elec ccewcds 2.02 se earn wis cisiac Se Hes siac.escsbs a cloceceadeae 13.59 I Se Nee he tcc aw cre nde rceceeeswcscen 17.38 eT See. OP eer ST? ek ORS Oe ee ge ae 4.94 EL ITiPET. 22.5. 8 2.02 a ola id ict oe nia wow d's Svein dd de ace sieeabewics 0.42 i os. ow da cin ow ub cine ne nice seucivana cd 0.00 aN ORIN ee ec coc ew dice cw aes cones sacecsaneneece 0.00 ase cee ccs scar vasewncansecedsnacese 3.3 eae eeu hot ee let ee hee cbaccsecebdadessvee 1.4 es oy eet lay os. ficken kde doo Seoo Ree wee 21.8 a So rc ans aad Sl xsd aii’ wig itatale oat oh 15.1 . re a rt ou Umd/ae dedaaied Koos 6.4 ee ci aos ona suse wcte evade dss cee 98.8 to make allowance for the pentoses present. Furthermore, it was assumed that the effect of this small pentose content on the | plane of polarization was zero. In event that all the furfuralde- hyde in the pentosan determination resulted from the mannose present, the glucose and fructose estimations are low—about 95 per cent of what they should be. SUMMARY. 1. A quantitative examination of asparagus seed (Asparagus officinalis L., var. ‘‘Palmetto’’) has shown that the reserve carbohydrate is in the form of hemicelluloses, which give, on hydrolysis, mannose, glucose, fructose, and galactose. 102 Seed of Asparagus officinalis L. 2. The galactose is in such small quantity that it appears likely that it forms part of a different hemicellulose from that which constitutes the bulk of the reserve carbohydrate. 3. The mannose is in a ratio of 1:1 to the total remaining hexoses. The absence in appreciable quantity of carbohydrates having the properties of cellulose, starch, and inulin makes it seem likely that the hemicelluloses are either glucomannans occurring with fructomannans or else glucofructomannans. 4. The proof that fructose is one of the cleavage products derived from the hemicellulose of asparagus seed confirms the single previous report of fructose among the hydrolytic products of a hemicellulose, namely, the report by Baker and Pope? in the case of Phytelephas. A CHEMICAL STUDY OF THE PROTEINS OF THE ADSUKI BEAN, PHASEOLUS ANGULARIS.* By D. BREESE JONES, A. J. FINKS, ann C. E. F. GERSDORFF. (From the Protein Investigation Laboratory, Bureau of Chemistry, United States Department of Agriculture, Washington.) ~- (Received for publication, January 5, 1922. Chemical studies of the proteins of various beans, including the navy,! lima,' mung (1), Chinese velvet (2), Georgia velvet (3), and the jack beans (4), have shown that nearly all of them contain two globulins, one occurring in relatively small amounts and the other constituting the chief protein of the seeds. The former have been designated a-globulins and the latter 8-globulins. The a-globulins were separated by addition of solid ammonium sulfate to the sodium chloride extracts of the meal in an amount sufficient to make the solution from 0.3 to0.4saturated. Small intermediate fractions consisting of a mixture of the two globulins were then removed by further addition of ammonium sulfate to the filtrates from the a-globulins until the solutions were about 0.6 saturated. These intermediate fractions were discarded, and from their filtrates the 6-globulins were precipitated by com- plete saturation with ammonium sulfate. The a-globulins differ chiefly from-the 8-globulins in having less nitrogen and a much higher sulfur content. The correspond- ing globulins obtained from the various beans which have been * A preliminary report of this paper was presented at the sixty-second meeting of the American Chemical Society held in New York City, Septem- ber 6 to 10, 1921 (cf. Science, 1921, liv, 444). The adsuki beans used in the experiments described in this paper were furnished by the Bureau of Plant Industry, Department of Agriculture, and were of the maroon variety. A complete description of the adsuki bean is given by Piper and Morse (Piper, C. V., and Morse, W. J., U. S. Dept. Agric., Bull. 119, 1914). 1 Chemical studies of the navy and lima beans are still in progress, and the results will be published later. 103 104 Proteins of the Adsuki Bean thus far examined, especially those of the Phaseolus genus, bear a close resemblance to one another in both their general proper- ties and chemical composition. Data other than chemical be- havior and composition, however, will be necessary to show whether or not they are identical. Work along this line is now in progress. A similarity in the proteins of these beans has also been shown by a study of their nutritive values as found by feeding experi- ments with albino rats. The proteins of the navy (5) and lima (6) beans were found to lack two factors necessary to make them available for the normal nutrition of the rats: first, a deficiency of cystine, which was corrected by the addition of 0.3 per cent of this amino-acid to a diet adequate in other respects; second, cooking—without which growth could not be obtained. In the case of the adsuki bean (7), however, it was found that while its proteins were deficient in cystine, cooking was not required. This difference in the nutritive behavior of the adsuki bean makes a further chemical study of its proteins of additional interest. The adsuki bean is one of five oriental species of beans which at various times have been introduced into the United States. Its cultivation was originally confined to Japan, Manchuria, | China, and Korea. Next to the soy bean it seems to be the most important legume grown in Japan, and commands a higher price than any other bean. At the Arlington farms of the United States Department of Agriculture, in experimental tests, average yields of about 22 bushels per acre were obtained. The beans used for the experiments described in this paper were identified as Phaseolus’*angularis. Piper and Morse (8) have called attention to the fact that “While the species is clearly and abundantly distinct, it has been con- fused with related species by most botanists. . . . . In most Japanese botanical works the adsuki bean is confused with the mung, and therefore called Phaseolus mungo or Phaseolus radiatus, from both of which it differs greatly.’’ The proteins of the Japanese adsuki bean were first studied in 1897 by Osborne and Campbell (9), who isolated, by dialysis of a 10 per cent sodium chloride extract of the bean, a globulin the properties and composition of which closely resembled those of ee eee eel rrr rrmrmErereeeeeeeeeee in ae) 4 ae ¢ tus, Jones, Finks, and Gersdorff 105 phaseolin, the chief globulin of the navy bean. Later Osborne and Harris (10) made a nitrogen distribution in several prepara- tions of both this globulin and phaseolin, and found that the preparations of the adsuki bean globulin gave uniformly higher results for the basic nitrogen. They also isolated a small amount of a substance from the extract after all of the globulin had been removed by dialysis in water, by further dialysis against alcohol. This protein resembled in its composition the legume- lins obtained from the pea, vetch, lentil, and other legu- minous seeds. The adsuki bean meal used for the preparation of the proteins _described in this paper contained 21.13 per cent of protein (N X 6.25). Preliminary experiments showed that the maximum amount of protein was extracted by thoroughly mixing the meal with aqueous 5 per cent sodium chloride solution in the propor- tion of 4 cc. of solvent to each gram of meal, and allowing the mixture to stand for 40 hours in cold storage at 1-3°C. It was found that if the extraction was allowed to take place at room temperature, a smaller yield of the a-globulin was obtained. This globulin denatures readily, changing into an insoluble form, a tendency which is less marked at a low temperature. In this way 79 per cent of the total protein in the meal was dissolved, or 16.7 per cent, based on the amount of meal used. By applying to the adsuki bean the same methods which have been used in this laboratory to separate the globulins of the mung, navy, and lima beans, we have again been able to isolate two globulins closely resembling the globulins obtained from those beans. Both globulins gave positive tests for tryptophane, and 2.13 per cent of tyrosine was isolated from the $-compound- The difference in composition of these two globulins is shown by the following percentages (Table I) which are the averages of those for several preparations of each protein. Determination of the diamino-acids by the Van Slyke method showed that the $-globulin contained more arginine, and about half as much cystine as the a-globulin, while the percentages of histidine and lysine found were practically the same. Argimne, histidine, and lysine were also determined in the §-globulin by the direct method of Kossel and Patten (11). The values ob- tained by the Van Slyke method were in all cases higher than those 106 Proteins of the Adsuki Bean obtained by the direct method, especially that for lysine. The percentages of the diamino-acids are summarized in Table II. After removal of all of the globulins from a distilled water extract of the bean meal, a small amount of an albumin was ob- tained by heating the dialysate for 2 hours at 70°C. TABLE I. Average Composition of the a- and B-Globulins. a@-Globulin. 8-Globulin. per cent per cent ees ok a See ee etek see oa oe 52.75 53.57 |: ree EP con eo econ 6.97 6.79 LS ee re ne a eee fs 15.64 16.46 Re ne Pee Serer errr em ae 1.21 0.40 TABLE II. Percentages of the Diamino-Acids in the a- and 8-Globulins. a-Globulin. B-Globulin. Ami id. E oe hi Van Slyke Van Slyke ee method. method. ee Batten: per cent per cent per cent Arpimine. i. .2...usseee = -seen ewes 5.45 7.00 5.30 Hlsstidine; 4 20). Seie ne hoon oe clers eee 2.25 2.51 1.76 Dapeines 21.054 Sa oece Seen 8.30 8.41 4.18 Cystine. nd.c.. sieeenk be cesses 1.63 0.86 EXPERIMENTAL. Preliminary Experiments——The beans were ground to a fine meal in a power driven pulverizing mill. Extractions were then made, using small samples of the meal and varying concentra- tions of sodium chloride in distilled water in the proportion of 4 ec. of solvent to each gram of meal. The mixtures were allowed to stand at a low temperature for 40 hours, and nitrogen* was 2 Acknowledgment is due to Mr. S. Phillips of this laboratory for the Kjeldahl nitrogen determinations made in the course of the present investigation. Jones, Finks, and Gersdorff 107 determined in aliquot portions of the extracts. The results (Table III) show that an extraction with 5 per cent sodium chloride solution for a period of 40 hours gave the maximum yield of protein. Precipitation experiments made with small quantities of the clear sodium chloride extract of the meal showed that some of the protein was precipitated at 0.3 of saturation with ammonium sulfate (Preparation I). Other fractions were obtained at 0.45 (Preparation II), and 0.65 of saturation (Preparation III), and a final fraction at 0.7 to complete saturation (Preparation IV). TABLE III. Extraction Experiments.* Protein Solvent. extracted from meal (N X 6.25). per cent I es a ass wo wis 5 oe 2 SL ee eee 7.25 En a ee Bi aan 11.88 1 bi See en winnie aleiske Ns che 8 216 8102's 13.81 ced re SR ee Oo Likes cic Sieyae Ae nes vgieveceis 14.44 vA | Plea Oe SGA SSE Bae eee ene ae eee 14.69 es ry SET RI ers oy s.cnid os ale See faye vic ihn aso es 15.31 Bia sig To. ae od ROSS CCIE ene See ee ee 15.69 3.5 “ vd “2 ) 22) = bd SOSA RGe ee eae te: aaa 16.56 4.0 “ “sf “Laas CaS GOR SBE AEs SERS Se I eter Irn 16.63 5.0 “ ay era 2 rs sins ete e Seles aiw' nie 2.2% 16 6s 16.75 role be So OE SE Se a ee 16.06 10.0 “ =~ MORE Shae oS ORS yy Acie ficiciere. odie -Sisicie,< 2 a0. 15.25 * The extractions were carried out at 1-3°C. for a period of 40 hours each. These various fractions were filtered and well washed with solu- tions of the same salts and concentrations from which they had been precipitated. They were then redissolved, dialyzed, washed, and dried in the usual way. These preparations contained 1.28, 0.95, 0.90, and 0.43 per cent of sulfur, respectively. These results show that two globulins, represented by Preparations I and IV, were present, differing markedly in their sulfur content, while Preparations II and III consisted of mixtures of the two globulins. Preparation of the a-Globulin.—Extractions were made, using from 1 to 9 kilos of meal and 5 per cent sodium chloride solution. 108 Proteins of the Adsuki Bean The meal was thoroughly mixed with the solvent, and the mix- tures allowed to stand for about 40 hours in a cold storage room at 1-3°C. Dry filter paper scraps were then pulped in the mix- tures until a consistency suitable for pressing was obtained. After pressing in muslin bags, the expressed liquors were filtered through a mat of paper pulp on a Buchner funnel. The clear filtrates were then made 0.3 saturated by addition of solid ammo- nium sulfate and the precipitate was allowed to settle by standing for 24 hours. The precipitates were filtered on folded filter papers and washed with large volumes of 5 per cent sodium chloride solution which had been previously made 0.3 saturated with ammonium sulfate. The precipitates thus obtained were found TABLE IV. Averages of Duplicate Analyses of the a-Globulin. Preparation. I II Ill IV MA Average. per cent per cent per cent per cent per cent | per cent | ONS 8 Se eres 52.99 | 52.69] 52.62 | 52.75 | 52.71 | 52.75 Hes y (ea 6.87 7.00 6.85 6.97 6.97 1 er 15.75 | 15.69 | 15.50] 15.60] 15.64] 15.64 ° eee 1.28 1.29 1.12 1.23 1.13 1.21 Scans Fee eee oo 22.81 | 23.46 | 23.76} 23.57 | 23.55 | 23.43 IMGISTIPG. pocr.s 25.24 +10 8.69 5.99 5.83 5.29 6.35 ASE Se aot eae! 6 ea 0.42 0.54 0.37 0.47 to denature rapidly, so that considerable portions would not redissolve in 5 per cent sodium chloride solution. The dissolved portions were dialyzed against chilled running water for 9 days. In the case of Preparations I and V (Table IV), the denatured portions were mixed with the soluble portion and dialyzed to- gether. After dialysis the preparations were washed, and dried with alcohol and ether in the usual way. An average yield of 0.35 per cent of the meal used was obtained. Averages of dupli- cate analyses of the five preparations of the a-globulin are given in Table IV. Moisture and ash were determined in all of the preparations, and the analyses calculated on the ash- and mois- ture-free basis. Jones, Finks, and Gersdorff 109 Preparation of 8-Globulin——The filtrates from the original precipitates of the a-globulin, which were 0.3 saturated with ammonium sulfate, were measured and made 0.65 saturated. The small amounts of precipitates were filtered off and discarded, as preliminary experiments had shown that these fractions con- sisted of a mixture of the a- and $-globulins. From the filtrates of the intermediate fractions the 8-globulin was precipitated by adding ammonium sulfate to saturation. The purification and drying of this material were carried out as described for the a-globulin. The average yield was 2.75 per cent of the meal extracted. The elementary composition of seven preparations is given in Table V. TABLE V. Averages of Duplicate Analyses of the B-Globulin. Preparation. | Aver- I II III IV av; VE VII age. per cent\ per cent|per cent\per cent per cent| per cent|per cent|per cent cent ye aa 54.17| 53.91) 53.05) 53.62 53.24) 53.35) 53.62) 53.57 sd) Dea 6.64, 6.94) 6.89] 6.92) 6.62} 6.74) 6.77| 6.79 eee as on «'s 55 16.38) 16.62) 16.45) 16.57) 16.37) 16.44) 16.37) 16.46 > SL Pee 0.46; 0.41) 0.36; 0.35) 0.41) 0.41) 0.37) 0.40 Wye is ors a cana: - 22.36) 22.12) 23.25) 22.54) 23.36) 23.06} 22.87] 22.78 4 1 | 5.30 4.66] 5.24 1 5. 67 -36) 0.32 0.15) 0.26) 0.21 Oo uu No bo Properties of the Globulins.—The finished preparations are light, dusty powders, the a-globulin ranging from a pale gray to a dark gray color, while the other is of a very light gray to an ivory white. Apparently the former, being the first to precipi- tate from the meal extract carries with it some of the coloring matter derived from the hulls of the beans. The a-globulin was found to coagulate at about 88°C. It gave a faintly positive test for tryptophane by the Hopkins and Cole reagent, and a positive test for tyrosine with Millon’s reagent. The 8-globulin coagulated on heating for 10 minutes at 97°C., and gave a strongly positive test for tryptophane. The precipitation limits with ammonium sulfate have already been noted. 110 Proteins of the Adsuki Bean Analyses of the a- and 8-Globulins by the Van Slyke Method.— Duplicate samples of about 3 gm. of each of the globulins were hydrolyzed by boiling for about 30 hours with 150 cc. of 20 per TABLE VI. Distribution of Nitrogen in the a-Globulin as Determined by the Van Slyke Method.* Sample I, ash- and moisture-free, 2.8343 gm. protein, 0.4422 gm. BoreEer- T Average. Sample 1H “ “ce “ “ 2. 8302 “ I Il gm am JAMIGEAINS- 4s eee ae cee oes 0.0444 | 0.0443 Humin N adsorbed by lime...| 0.0061 | 0.0063 Humin N in ether-amy! alco- hnoliexttact. poe eee eee 0.0005 | 0.0006 OyAiities IN... 2. Gem pees a aoe 0.0055 | 0.0053 Arginime Nc oe ores hoes 0.0499 | 0.0495 Huastidme Ns. eee = 0.0170 | 0.0177 ysine No onesies Sees asp ees 0.0451 | 0.0450 Amino N of filtrate...........] 0.2585 | 0.2596 Non-amino N of filtrate...... 0.0167 | 0.0156 Total N regained...........| 0.4437 | 0.4489 0.4415 “ Preparation. I II per cent per cent 10.05 10.03 1.38 1.43 Ott 0.13 1.23 1.19 11.28 |} 11.22 3.83 4.01 10.20 | 10.19 58.46 | 58.80 3.78 3.53 100.32 | 100.53 * Nitrogen figures corrected for the solubility of the bases. +t Nitrogen content of protein, 15.60 per cent. TABLE VII. Basie Amino-Acids in the a-Globulin. Amino-acid. I per cent Arming: 25405855 d:%i2 8: ae nee 5.47 Histidine! o:20 Bes i ae, Pe a1 IV SING 2 5 48e yee ae oes ee 8.30 CEP SATIO Crore PEAS det heed Costs, neat ae 1.66 II per cent 5.44 2.31 8.29 1.61 per cent 10.04 1.40 - Noon Ne SSESnarn a Average. per cent 5.45 2.29 8.30 1.63 cent hydrochloric acid. The phosphotungstates of the bases were decomposed by the amyl] alcohol-ether method (12). The results of the analyses are given in Tables VI, VII, VIII, IX, and X. Jones, Finks, and Gersdorff wg! TABLE VIII. Distribution of Nitrogen in the 8-Globulin as Determined by the Van Slyke Method.* Sample I, ash- and moisture-free, 2.8557 gm. protein, 0.4732 gm. nitrogen. Sample II, “ ra me — 2.855, “* = 0.4732 “ ee 0.0517 | 0.0515 | 10.92} 10.89} 10.91 Humin N adsorbed by lime...| 0.0039 | 0.0041 0.83 0.86 0.84 Humin N in ether-amy!] alco- UD) Garis 0.0001 | 0.0004 0.02 0.07 0.05 Ot ee 0.0029 | 0.0028 | 0.61 0.59 0.60 PPI DE Deke 2 2'2= sid << 62" 0.0653 | 0.0635 | 13.80) 13.42 | 13.61 0 TL ee 0.0183 | 0.0204 3.87 4.31 4.09 lv 0.0458 | 0.0460 9.66 9.85 9.75 Amino N of filtrate...........| 0.2643 | 0.2589 | 55.85 | 54.71 | 55.28 Non-amino N of filtrate....... 0.0170 | 0.0208 3.60 4.40 | 4.00 Total N regained........ 0.4692 | 0.4690 | 99.16 | 99.10 | 99.13 * Nitrogen corrected for solubility of the bases. + Nitrogen content, 16.57 per cent. TABLE IX. Basic Amino-Acids in the 8-Globulin. Amino-acid. I II Average. per cent per cent per cent loon Ae 7.10 6.90 7.00 ME cia ce di sp ks oes 2231 2.64 2.51 a i: rr 8.33 8.48 8.41 Jo “ge eee ae 0.87 0.84 0.86 TABLE X. Distribution of Nitrogen in the a- and B-Globulins as Calculated from the Van Slyke Analyses in Terms of Per Cent of the Proteins. a-Globulin. 8-Globulin. N pee ee ee eee I II Average. I II Average. per cent per cent per cent per cent per cent | per cent 0) a he aired 1.57 1.57 1.81 1.80 1.81 ee 0.23 0.24 0.24 0.16 0.15 0.15 SG SRS a 4.15 4.15 4.15 4.63 4.67 4.65 Non-basic............ 9.71 9.72 9.71 9.75 9.79 9.77 a 15.66 | 15.68 15.67 16.35 16.41 16.38 12 Proteins of the Adsuki Bean Determination of Tyrosine and the Basic Amino-Acids in the B-Globulin by the Direct Method of Kossel and Patten. Tyrosine.—50 gm. of the @-globulin equivalent to 47.6 gm. of the ash- and moisture-free protein were hydrolyzed by boiling for 48 hours with a mixture of 150 gm. of concentrated sulfuric acid and 300 cc. of water. The solution, freed from sulfuric acid and concentrated until crystals began to form, yielded 1.01 gm. of tyrosine equivalent to 2.13 per cent of the protein. Without recrystallization it was analyzed with the following results: 0.1335 gm. substance required 7.60 cc. of 0.1 N acid. CoH,103N. Calculated. N lar Found. ee ree!)- The filtrates and washings from the tyrosine were used for determinations of the bases according to the method of Kossel and Patten. Histidine.—The solution of the histidine = 500 cc. 50 ce. solution required 16.15 ec. 0.1 N acid = 0.2267 gm. N in 500 ce. = 0.8368 gm. histidine or 1.76 per cent in the protein. The histidine was converted into the dihydrochloride for identification. 0.1230 gm. substance: 0.1534 gm. silver chloride. 0.0792 “ a 0.0915 “ carbon dioxide and 0.0372 gm. water. CsHi1,02N3Cle. Caleulated. Cl 31.14, C 31.58, H 4.86. Found. ‘e - 80:85; “S¢ VSI Dey Geeeeeetan Arginine.—The solution of arginine = 1,000 ce. 50 cc. solution required 27.70 cc. 0.1 N acid = 0.7778 gm. N in 1,000 ce. = 2.4140 gm. arginine + 0.1080 gm. = 2.5220 gm. or 5.30 per cent. The arginine was converted into the copper nitrate double salt for identification. 0.2023 gm. substance: 0.0278 gm. copper oxide. Cy2H2304N sCu(NO3)2 3H20. Calculated. Cu 10.79. Found. “. 1 OLG3= Lysine.—The lysine picrate weighed 5.10 gm. equivalent to 1.99 gm. lysine, or 4.18 per cent. Jones, Finks, and Gersdorff 113 0.1638 gm. substance required 21.65 cc. 0.1 N acid. C.5H1402.N2°C5H307N3. Calculated. N_ 18.67. Found. “18.47. A summary of the percentages of amino-acids in the two globu- lins is given in Table II. The Albumin. A small amount of albumin was also isolated from the adsuki bean. The meal was extracted for 40 hours at 1-3°C., with three times its weight of distilled water, and most of the globulins were removed from the clear, filtered extract by dialysis for 18 days. A further separation of a small amount of globulin was effected by passing into the filtered solution after dialysis a stream of carbon dioxide for 35 hours. After filtering, the clear solution was heated for 2 hours at 70°C., and the coagulated albumin washed and dried in the usual way. The yield was about 0.05 per cent of the original meal. Owing to the small amount of material available, its analysis was limited to that of its sulfur content, which was found to be 2.19 per cent. SUMMARY. The adsuki bean contains about 21.13.per cent of protein (N X 6.25), 16.7 per cent of which is extracted by means of a 5 per cent aqueous sodium chloride solution. By fractional precipitation of the sodium chloride extracts with ammonium sulfate two globulins, designated as the a- and 6-globulins, have been isolated. The former was precipitated by addition of ammonium sulfate in sufficient amount to make the original extract 0.3 saturated. A small fraction consisting of a mixture of the two globulins was separated from the filtrate by increasing the concentration of ammonium sulfate up to 0.65 of saturation. This fraction was discarded, and the 6-globulin precipitated by making the solution completely saturated. A small amount of an albumin was found in distilled water extracts of the bean, after the globulins had been removed. The two globulins differ markedly in their sulfur and nitrogen content and in their nitrogen distribution as determined by the method of Van Slyke. 114 Proteins of the Adsuki Bean SOON of wh BIBLIOGRAPHY. . Johns, C. O., and Waterman, H. C., J. Biol. Chem., 1920, xliv, 303. . Johns, C. O., and Finks, A. J., J. Biol. Chem., 1918, xxxiv, 429. . Johns, C. O., and Waterman, H. C., J. Biol. Chem., 1920, xl, 59. Jones, D. B., and Johns, C. O., J. Biol. Chem., 1916-17, xxviii, 67. . Johns, C. O., and Finks, A. J., J. Biol. Chem., 1920, xli, 379. . Finks, A. J., and Johns, GC. O., Am. J. Physiol., 1921, lvi, 205. . Johns, C. O., and Finks, A. J., Am. J. Physiol., 1921, lvi, 208. . Piper, G. V., and Morse, W. J., U. S. Dept. Agric., Bull. 119, 1914, 3. . Osborne, T. B., and Gampbell, G. F., J. Am. Chem. Soc., 1897, xix, 509. . Osborne, T. B., and Harris, I. F., J. Am. Chem. Soc., 1903, xxv, 323. . Kossel, A., and Patten, A. J., Z. physiol. Chem., 1903, xxxviii, 39. . Van Slyke, D. D., J. Biol. Chem., 1915, xxii, 281. CLINICAL METHOD FOR THE ESTIMATION OF CHLORIDES IN BLOOD. Bry HERMAN FRIEND. (From the Department of Physiology, College of Physicians and Surgeons, Columbia University, New York.) (Received for publication, December 8, 1921.) All clinicians readily appreciate the importance of sodium chlo- ride estimation in affording very valuable data for the diagnosis, treatment, and prognosis of certain diseases and conditions. The value of a reliable and convenient clinical method is, therefore, obvious. An ideal clinical method is one that is easy of manipulation, accurate, rapid, and one that utilizes as little apparatus and rea- gents as possible. The estimation of chlorides may be carried out in several ways: gravimetrically, volumetrically, and by jodimetry. A gravimetric method is perhaps always best in that it is most exact, but for clinical purposes it is unpractical for the reason that it is time-consuming, requires more or less apparatus (bal- ance, etc.), skill, and rather large amounts of material. Iodime- tric methods are very good but irksome and call for too many reagents. The proposed method is very simple, requiring very little glassware and reagents such as an ordinary office laboratory can readily provide. A factor generally neglected in blood chemistry is that for many determinations only the plasma should be analyzed. This was emphasized by Kuttner! in describing his uric acid test. It is particularly true of the chloride content. Some observations recorded in the literature give the chloride content of whole blood, others the content of plasma or serum. As early as 1850 Carl Schmidt? gave figures for the chloride content of both whole blood 1 Kuttner, T., J. Am. Med. Assn., 1915, Ixv, 245; 1916, Ixvi, 1370. 2 Schmidt, C., Charakteristik der Epidemischen Cholera, Leipsie and Mitau, 1850. 115 116 Chlorides in Blood and plasma and found that the chloride content as NaCl is about 0.12 per cent higher in plasma. Recent observations by Van Slyke and Cullen* confirmed by McLean‘ and Myers and Short*® call attention to the change (increase) of NaCl in plasma if plasma remains in contact with the red cells. They interpret this in- crease as due to a loss in bicarbonate. Since, however, the plasma rather than the whole blood bathes the tissues of the body, it seems only logical to study the chloride content of plasma only; unless, however, the plasma is separated immediately there appears as stated, and from my own observations, a gradual increase in the chloride content due to a decreasing CO, tension. The method to be described has been compared with the stand- ard one of McLean and Van Slyke. The tests were performed under similar conditions and these are the results: NaCl per 100 ce. McLean and Van Slyke’s method. Author’s method. mg. mg. 585.0 581.1 650.8 645.4 628.8 625.1 592.3 589.3 606.9 603.5 731.2 728.1 599.6 592.3 A salt solution was also made 100 cc. = 0.509 gm. NaCl accord- ing to Folin’s method. Several blanks were run using this solu- tion with the following results: Solution No. NaCl per 100 ec. mg. 504.5 504.5 504.5 504.5 em Whe 3 Van Slyke, D. D., and Cullen, G. E., J. Biol. Chem., 1917, xxx, 317. 4 McLean, F.C., J. Exp. Med., 1915, xxii, 212. 5 Myers, V. C., and Short, J. J., J. Biol. Chem., 1920, xliv, 47. a Herman Friend 117 The same solution was. again added in 1 and 2 cc. quantities to blood, the NaCl volume of which was known—results were ob- tained as follows: No. 1 No. 2 NaCl per 100 cc.| NaCl per 100 cc. mgm. mgm ME ss esos ss 0.6435 | BA COYOT bifid ns Ad RUE Os she a 0.6727 PAMNNCCS Wr. sie s es 1.1475 oO cle ee cic Be ee 1.1720 PUMMEGC EOI it 2. ot! 1.6525 ZS kw (0 Cog S oe tee eae > 1.6765 Very careful tests were conducted for the detection of interfering constituents such as sulfates, phosphates, and uric acid. These substances were not found. Briefly, the protein of the plasma is precipitated by a neutral precipitant, made up to volume, filtered, and titrated directly against a standard silver nitrate solution. A chart is then con- sulted, avoiding all calculation. Method. Reagents.—The following reagents are used: Aluminum cream, 0.02 n silver nitrate, and 5 per cent solution potassium chromate. Aluminum Cream.—It is best to prepare this and not to rely upon commercial products. To a liter of saturated aluminum alum (c. Pp. not necessary) add concentrated NH,OH until com- plete precipitation. Boil till weakly alkaline to litmus. Let stand to settle. Wash by decantation till supernatant fluid is neutral to phenolphthalein. Adjust to volume of 800 cc. Place in well corked bottle. Procedure takes about 24 hours. This cream is free of chlorides. Silver Nitrate-—This is made by accurately measuring from a burette 20 cc. of 0.10 n solution into a 100 cc. volumetric flask and adding distilled water up to the mark. Kept in a dark bottle, this solution does not deteriorate upon short standing. Technique. Blood must be taken on a fasting stomach. It is important to remember that the plasma must be separated immediately to obtain correct results. Pipette 1 ec. of clear plasma (slight 118 Chlorides in Blood AgNOs used. NaCl per 100 ec. AgNOs used. NaCl per 100 cc ce mg. ce mg. 2.00 292.5 4.00 585.0 2.05 299.8 4.05 592.3 2.10 307.1 4.10 599.6 7a ks 314.4 4.15 606.9 2.20 321.6 4.20 614.2 p75) 329.0 4.25 621.5 2.30 336.3 4.30 628.8 PARAS. 343.8 4.35 636.1 2.40 351.0 4.40 643.5 2.45 358.3 4.45 650.8 2.50 365.6 4.50 658.1 2.99 372.9 4.55 665.4 2.60 380.2 4.60 672.7 2.65 387.5 4.65 680.0 2.70 394.7 4.70 687.3 Zo 402.1 4.75 694.6 2.80 409.5 4.80 702.0 2.85 416.8 4.85 709.3 2.90 424.1 4.90 716.6 2.95 431.4 4.95 723.9 3.00 438.7 5.00 731.2 3.05 445.6 5.05 738.5 3.10 453.3 5. 10 745.8 3.15 460.6 5.15 753.1 3.20 468.0 5.20 760.5 3.25 475.3 5.25 767.8 3.30 482.6 5.30 775.1 3.35 488.9 5.35 782.4 3.40 497 .2 5.40 789.7 3.45 504.5 5.45 797.0 3.50 511.8 5.50 804.3 3.95 519.1 5.55 811.6 3.60 526.5 5.60 819.0 3.65 533.8 5.65 $26.3 3.70 541.1 5.70 833.6 3.75 548.4 oto 840.9 3.80 555.7 5.80 848.2 3.85 563.0 5.85 855.5 3.90 570.3 5.90 862.8 3.95 SV ale 5.95 870.1 6.00 877.5 Computation; 1 ce. 20.0 n AgNO; = 1.170 gm. NaCl. Since the filtrate represents # of 1 cc. of the plasma it is necessary to take $ of the result obtained. Titration X 1.170 X i X 100 = mg. NaCl per 100 ce. Herman Friend 119 hemolysis does not interfere) into a 25 cc. volumetric flask con- taining about 10 cc. water. Add from a graduate 3 cc. of alumi- num cream and make up to volume with water. Shake well, stand 10 minutes. Filter through 5 cm. dry filter paper. 22 cc. of filtrate should be obtained and filtrate should be water-clear; slight color, however, will not interfere. Pipette 20 cc. of this filtrate into a 50 cc. beaker or Erlenmeyer flask. Add 5 drops of the 5 per cent potassium chromate. Titrate until the yellow color is changed to a first tint of dirty brown. Note number of ec. used. Consult chart for reading. For very exact work, deduct 0.05 cc. from reading for the reason that this end-point is actually silver chromate. For clinical use, thisis not necessary. Milk may be done in exactly the same way. The determination of chlorides in blood has as yet found little applicability in clinical medicine. The value of such an estima- tion, however, is not to be denied, particularly where diet restric- tions are to be followed. Generally, such conditions as cardiac diseases, anemia, and malignancy have shown high chloride content, while fevers, diabetes, and pneumonia have shown low figures. ' In conclusion, I wish to thank Dr. H. R. Miller for his interest and ever ready advice. STUDIES ON THE AMINO-ACID NITROGEN CONTENT OF THE BLOOD. By SEIZABURO OKADA anno TOWORU HAYASHI. (From the Medical Clinic, Imperial University of Tokyo, Tokyo, Japan.) (Received for publication, March 26, 1921.) It has been shown by various workers, especially by Van Slyke and Meyer (1), Gyérgy and Zunz (2), and Bock (3) that the amino-acid content of the blood of fasting animals for each species is fairly constant. It was further known that this constancy is maintained with remarkable tenacity except during digestion, when a transient increase occurs. Under special conditions, however, certain deviations from this constant may take place. The purpose of this paper is to study the conditions under which these deviations might develop. Experiments were carried out on animals and on the bloods of various ward cases. To control the results and to get further explanations we have studied also the contents of blood sugar, urea nitrogen, and non-protein nitrogen. The amino-acid nitrogen was determined by the method des- cribed previously by one of us in this Journal (4). The blood sugar content was determined by the Lewis and Benedict (5) method modified by Myers and Bailey (6). The urea nitrogen content of bloods was estimated by the Van Slyke and Cullen (7) method with slight modifications. The non-protein nitrogen was determined by Bang’s micro method (8). Physiological Amino-Acid Nitrogen Content of the Blood. Van Slyke and Meyer, using the alcohol precipitation method, found from 3 to 5 mg. of amino-acid nitrogen per 100 ce. of blood in dogs which had been fasting from 20 to 24 hours. Gyorgy and Zunz, using the same method, and removing urea by treatment with soy bean and subsequent aeration, obtained 4.8 mg. of 121 122 Amino-Acid Nitrogen of Blood amino-acid nitrogen in 100 cc. of blood from dogs fasted for 24 hours. Bock, using the heat-trichloroacetic acid precipitation method, found an average of 7.47 mg. of amino-acid nitrogen per 100 cc. of dog’s blood, 8.43 mg. from pigs, 8.68 mg. from cats, 7.63 mg. from sheep, 6.84 mg. from calves, 6.58 mg. from oxen, and 7.13 mg. from human venous blood. As Okada reported in a previous paper that there was no appreciable difference between the results of the heat-trichloroacetic acid method and the heat-kaolin procedure, we first tried to see whether we could get the average amino-acid nitrogen content in the blood of dogs that Bock found. We always used animals that were fasting from 20 to 48 hours and found an average of 7.3 mg. of amino-acid nitrogen per 100 cc. of blood, coinciding fairly well with the result of Bock. The content of amino-acid nitrogen of the blood of rabbits is somewhat higher than that of dogs, being on the average 9.2 mg. per 100 cc. The results of twenty-eight analyses of dog’s blood are given by the following figures, which are arranged in ascending order. Amino-acid Amino-acid Amino-acid Sample No. aerogens Sample No. nitroged: Sample No. nitrogen. mg. mg. mg. 1 6.33 11 7.03 21 7.66 2 6.34 12 7.05 22 at 3 6.38 13 7.07 23 8.14 4 6.39 14 7.08 24 8.40 5 6.42 15 7.31 25 8.46 6 6.49 16 7.39 26 8.51 7 6.51 17 7.43 27 8.61 8 6.62 18 7.66 28 8.79 9 6.62 19 7.66 oa 10 6.90 20 7.66 Average... 7.31 Most of our experiments on dogs were carried out under anesthe- sia with morphine or other narcotics, so that we examined first whether or not anesthesia tends to change the content of amino- acid nitrogen in the blood. From Table I, it is obvious that anesthesia has no influence upon the amino-acid nitrogen content of the blood, though there was always more or less increase of blood sugar. For rabbit’s blood the following results were obtained. Sample No. Amino-acid Sample No. Amino-acid Sample No: Amino-acid nitrogen. nitrogen. nitrogen. mg. mg. mg. 1 7.22 13 8.74 25 9.92 P 7.40 14 8.74 26 9.97 3 7.60 15 8.97 27 9.97 4 8.11 16 9.04 28 10.05 5 8.21 Lr 9.06 29 10.18 6 8.30 18 9.11 30 10.18 7 8.38 19 9.38 31 10.21 8 8.42 20 9.64 32 10.37 9 8.48 21 9.71 33 10.57 10 8.48 22 9.73 34 10.58 11 8.55 23 9.73 35 10.60 12 8.64 24 9.76 SSS | SSS Average.. 9.20 TABLE I. The Relation between Anesthesia and Amino-Acid Nitrogen Content of the Blood of Dogs. Bef Af ide Bod e : efore or ter Blood |2cid 2i- No. | weight oe anesthesia, | thesia. | 8¥= | Bopp cc. kg. hrs. |per cent) mg. as 5.0 | Chloroform. Before. 1.40 | 6.90 After. 3 | 2.06 | 6.90 2 5.4 | Morphine hydrochloridum Before. 1.62 | 6.57 hypodermically, 0.25 gm. After. 3. | 2.25 | 6.92 3°| 15.2 Morphine hydrochloridum Before. 1.25 | 8.61 hypodermically, 0.3 gm.; After. T (| 1580*) S.6l and ether insufflation. . 3 | 2.00 | 8.42 4 | 17.7 | Ether insufflation. Before. 1.10 | 7.96 After. 1 1.90 | 8.14 ss 1.92 | 8.14 5 |+8.5 | Morphine hydrochloridum “i 0.25 | 1.60 | 7.72 hypodermically, 0.26 gm. 5 4 | 2.08 | 7.7 ne 6 | 2.30 | 7.63 m 8 | 2.00 | 7.83 6 | 11.5 | Morphine hydrochloridum + 0.5.) 1.13.) 8.12 hypodermically, 0.2 gm.; . a PIS Se and injection of 30 cc. of - 6 | 1.18 | 8.12 blood into the peritoneal S 8 | 1.34 | 8.51 cavity. 123 124 Amino-Acid Nitrogen of Blood In Experiment 6 we injected the blood of the same animal into the peritoneal cavity to see whether any change in the amino- acid content of the blood would result, as some of our experiments TABLE II. The Influence of the Removal of the Pancreas on the Sugar and Amino-Acid Nitrogen Content of the Blood of Dogs. No. Body weight. ae kg. gm 1 18.5 0.25 2 8.75 0.2 3 13.0 0.25 4 20.8 0.3 5 11.0 0.11 6 9.5 (Sugar in urine, 3.6 per cent). 7 10.1 (Sugar in urine, 3.3 per cent). Before or after removal of pancreas. Before. After. “ce “cc Before. After. Before. After. Time after removal. oO On One 22 28 Blood sugar. per cent 1.23 1.72 1.15 2.28 2.50 Amino- acid nitro- gen per 0 ce. involve bleeding in the peritoneal cavity. The result shows that there is no influence. The close relationship between the blood sugar content and the functions of endocrine glands such as pancreas, suprarenals, S. Okada and T. Hayashi 125 thyroid, etc., is well known. It seems, therefore, interesting to investigate whether the amino-acid content of the blood has any relation to the functions of such endocrine glands. First we removed the pancreas of dogs and observed the influence. The complete removal of the pancreas was established after each experiment. The removal of the pancreas causes an increase of amino-acid in the blood as well as of blood sugar. Whether this result may be of transient character or permanent will be seen in Table III. The above experiments show us that the increase of amino-acid nitrogen content in the blood is still remarkable 2 days after the thorough removal of the pancreas, while thereafter it diminishes again even to the subnormal amount. The blood sugar shows continually its high rate of percentage, though the excretion of sugar in urine diminishes after a while. When a part of the pan- creas is left under the skin, no increase of amino-acid nitrogen or of sugar occurs. The change in percentage of urea nitrogen and of non-protein nitrogen in the blood is inconstant. Adrenalin was found to be without effect on the amino-acid content of blood. Five rabbits each received hypodermically 1 ce. of 1 : 1,000 adrenalin solution per kilo. The amino-acid nitrogen figures before and 2 to 3 hours aiter were respectively 9.11 and 7.11, 10.05 and 10.05, 8.38 and 8.35, 7.40 and 7.23, 8.74 and 8.74 mg. per 100 cc. of blood. Similarly, negative results were obtained in eight experiments in which pituitrin was injected. It is well known that the thyroid gland is essentially important to metabolism. Bock reported a case of hyperthyroidism showing an increase in blood amino nitrogen over the normal. We exam- ined five cases of hyperthyroidism (Graves’ disease) and found that all of them were of normal amino-acid content. We also extirpated the thyroid glands of four dogs without effect on the blood amino nitrogen. In one case the parathyroids also were removed so that the animal died of tetany 5 days after the opera- tion. In the other cases no tetany occurred. We may therefore conclude that the thyroid gland has no influence upon the amino- acid nitrogen content of the blood. Amino-Acid Nitrogen of Blood 126 *4u90 Jed ZZ ‘outin Ul IvsNng |Z76'6E 60°0S ¥0'€€ 88°09 Fe a FL'0S ’ “qu90 Jod 6] ‘oulin ul Iv3ng |Eeg’se “queo Jod g‘Q] ‘eulin Ul arvsng |¢1'OF 91° LP *qu90 jod Zp ‘oulin ul aesng “But “poo[q jo 09 00T “SHIBUYA Jod ues -O1}1U utayoid -u0ON 6r'0OT | 187 | cE | I ” Gc OT | 86°9 | cbs | 9 ” Geol | 116 | 99'E | @ “LOVFV *poaoul cL OL | 6F'9 | G8'S ‘alojog =| -o1 ATYsno10y} SvotoUeg | G*L SE°ST | O6'TT] G2°E | @ “1945 V *poaour Go'OL | 6€°L | 99'T ‘aLoyog, | -e1 ATYsnos0y} svoroURd | G°CT LOT | 06°S | OO'F | LT ” GO°ST | COL | PSE | ET ” 8S°ST | 648 | OF F| 2 “194FV *poaoul 86ST 12 Pa L ‘alojog =| -oa ATYSnoLoy} svosoued | PIT "U014B19 82°81 | 48°8 | GLE | g |-doer 199FV *poaoul -o1 ATYSNOLOYY sva1oueg 86°SI | 9F-O | PL) 8 ‘I04JV | “UL{S oy} JOpuN 4Jo] SvAK COTEL ee8e9 Wesel ‘a10jog | svorouvd oy} jo waved V | FZ “Bu “bus \yuao wad| shop “BY “poolq eRe anit “09 001 ‘1e3ns peas ‘uo1ye10d0 . “qVysIoM Jed ues peace poold | 10qje | 1ozje IO o10JOgT tied 0) Apog -O1fIU ae oulL T, Bol) -oulury tg | ‘sbog fo svasouvg ay} fo Daoway ay) UO spuaUrsad x iT TIT GTAVL ‘ON 7a 127 S. Okada and T. Hayashi *STZ1U0}LIOd ou Asdoyne Aq ‘yuU90 dod g*y ‘eulin ul avsng [ez FF "quao Jed g°6 ‘eulin ul re3ng ‘sIquopiod 94n9v Jo porq ie ae a ee 08 GP GL’ 9% 0g° 29 0L°F9 GP I9 61°2S 60°8 | FEE Ze'9 | STI o2"¢'| 1¢°2 66'8 | 19°€ 68°9 | cE°% PI'S] FF £0°L | 22% Z9'OT| eg'e LO FET ILS] 99°F GO'L | FF'% ‘IOVJV ‘a10 Jog “10FV ‘OLOJOE “LOUSY ‘d10 JOG “10YV ‘O10 JOT ‘LOVJV "O10 JOET ‘UOoT}V1OdOd.l 104, V *poaoul -o1 ATYSNOLOYY Svo19 Ug ‘UTYS OY} LOpuN Yay SUM svotouvd ayy jo yard vy *poaoul -o1 ATYSNOLOY} SBIIOUB ‘poaou -01 ATYSNOLOYY SvotIUv *poaour -a1 ATYSNoIOY} svarouRg *poaour GL oh PGI Sie ~ol ATYINOLOYY SvaLOUB | OST g € ”? ” ” ‘00q ” ”? "490 ”? ”? ” ” ‘ydag © THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI, NO. lL 128 Amino-Acid Nitrogen of Blood The effect of kidney extirpation on the amino-acid nitrogen content of the blood was also studied. It is well known that when kidneys are extirpated or the ureters are ligated, a pro- nounced increase of urea and non-protein nitrogen in the blood will occur. Bock investigated pathological bloods and found the most pronounced variations from the normal in nephritis. We have confirmed Bock in observations on patients with uremia and TABLE IV. Experiments on the Removal of the Thyroid Gland of Dogs. Amino- Urea acid nitro-| nitrogen sugar. gen per |per 100cc. Body Before or Time Biowl No. Date. Gisht after after extir- extirpation. | pation. 100'co-" | of Blood. is | ko. (" =| .dipees| ponioent /ealenalen rena 1 Aug. 29 12.0 Before. 1.69 8.46 10.49 351 After. 2 1.53 1.80 10225 2 Sept. 14] 16.2 Before. 1.37 8.79 | 11.18 © 16 After. 2 2.00 Tae elas &¢ 21 $ a 2.30 6.72 11.18 C530 s 16 1.70 Gidl 9.79 Oct Uf “6 23 Peel 2 6.61 10.25 3 | Sept. 21] 16.5 | Before. 1.46 | 6.42 | 11.65 < 23 After. 2 2.00 6.19 10.25 “ 28 es i 1.98 5212 11.18 4 Oct. 317) 1220 Before. 1.98 7.66 | 10.25 Nov. 2 After. Pe 152 ieon 10.65 se iby sd 2, a 357 6.71 9.79 Dec. 13 ae 43 0.98 7,65 =| 10-72 1919 Jan. 10 al 71 1.02 (cit 9.79 Feb. 13 US 7 ad 105 b15 7.05 10.72 severe nephritis, and have also found a marked increase of amino- acid nitrogen in the blood of dogs after ligating the ureters or extirpating the kidneys. Pilocarpin, as is well known, is an excitant for the secretion of the salivary and lachrymal glands, the mucous glands of the mouth, throat, nose, and deeper respiratory passages, the gastric secretory glands, the pancreas, and the intestinal glands. It also S. Okada and T. Hayashi 129 excites the suprarenal bodies and causes their increased intestinal secretion and influences the nervous control of glycogen in the liver. The amylase and the maltase in the blood will copiously increase as in the case of the ligature of the pancreatic duct. It seems reasonable, therefore, to try to find out whether such a drug also has any influence upon the amino-acid content of the blood. The result of the hypodermical injections of pilocarpin hydrochloricum to five dogs tells us that it causes a definite increase of the amino-acid content in the blood (35 to 65 per cent). TABLE V. Experiments on the Removal of Kidneys or the Ligature of Ureters on Dogs. : £3 | 82 | #5 + 5 : Fs a0 | of 4 . Before or | 5 ZA a5 | 3m No Date 3 Operation. after 2 5) 85 Bias log = operation. | ‘3 - n Seal" sas 52 2 as| 3 |28| #8 | 483 Bt 4 ~~ a a oo 5 oO ° 1918 kg. days pac mg. mg. mg. 1 | May 24/ 9.8} Ligature of | Before. ‘1'1.91] 6.42] 23.77 * 20 both ureters. | After. 2 | 2.10)16.44/131.09 2 | June 2 | 17.7; Both ureters) Before. 1.74) 7.08} 11.65) 42.24 5 eee: | cut off. After. 2 | 1.57|18.17|°67.07|157.98 ge) 3 | 1.92/17.82/181.74/226.09 3 “« 24 | 21.2) Bothureters | Before. 1.37| 6.62] 12.58] 56.89 | 25 cut off. After. 1} | 1.26/11.71) 44.12)100.65 4 | July 1| 9.5) Extirpation | Before. 1.38) 7.43] 24.32) 70.11 yaar of both After. 2 | 1.50)15.31/115.57/224.61 kidneys. We have determined the blood amino nitrogen in a large number of ward cases, with consistently abnormal results only in advanced nephritis and leucemia. Since the results in leucemia appear to be new, we give them in Table VII. From Table VII, it will be seen that the amino-acid nitrogen content of the blood increases or decreases in a fairly constant proportion to the number of white corpuscles. X-ray treatment decreases the amino-acid nitrogen content as well as the number of white corpuscles, while it increases the red corpuscles in the 130 Amino-Acid Nitrogen of Blood blood. Therefore, the red blood cells do not seem so essential a factor for the amino-acid content in the blood as the white blood cells. To ascertain more exactly the relationship between white cells and amino-acids we divided the blood into fractions TABLE VI. Experiments of the Hypodermic Injections of Pilocarpin Hydrochloricum to Dogs. : . Bef Ti bef i i No. Body weight. P loegan pa or Be Blood sugar. ea mnjected. injection. injection. per 100 ce. kg. gm. hrs. per cent mg. 1 13.9 0.14 Before. 2.82 (20D After. 2 3.75 10.51 7 4 1.90 8.21 2 etl 0.08 Before. 2 1.66 Fe iley oh 1.96 coon After. 2 2.00 11.96 ef 3 13'S Ln 27, 3 17.8 0.18 Before. 1 2.50 6.38 2.50 6.54 After. 2, Sie 70: 8.83 | se 4 1.53 7.38 oe ed 1.56 8.58 4 9.2 0.095 Before. 1 2.98 8.13 a: S207 6.83 After. 1 4.00 11.06 % 2 4.29 10.09 ee 3 2.50 11.39 5 1s 0.075 Before. 1 1.60 7.60 x feae 7.60 After. 1 Laws 10.03 se 2 1.51 10.64 4 3 1.41 11.01 and determined the amino-acid content in each fraction. The results show that the white blood cells contain six or seven times as much amino-acid as the plasma. The red and white blood cells cannot be separated perfectly from each other except by means of high speed centrifuge, so that the relatively high content of S. Okada and T. Hayashi t31 amino-acid in the red blood cells seems mainly due to the mixing of white blood cells. TABLE VII. Amino-Acid Nitrogen and Cor puscles in the Blood by Leucemia. : y Amino-acid| No. of Nev orred oe Diagnosis. a io pe a mg. c. mm. c.mm., a Chronic myeloid leucemia......... 11.27 | 212,600 31 Subacute lymphoid “ ......... 6.11 4,000 | 1,040,000 32 Chronic myeloid leucemia......... 12.86 | 244,500 | 4,780,000 Same after x-ray treatment....... 8.60 35,310 | 5,032,000 51 Chronic myeloid leucemia........| 17.81 | 286,100 | 3,520,000 Same after x-ray treatment....... 6.68 13,800 | 4,185,000 52 Chronic myeloid leucemia......... 29.22 | 582,400 | 2,840,000 Same after x-ray treatment....... 19.24 | 399,000 | 2,688,000 5d Chronic myeloid leucemia........} 13.14 | 205,800 | 3,520,000 56 Acute lymphoid “ih Lirias inte 5.60 1,000 886,000 58 Subacute myeloid “ ........ 8.02 64,000 852,480 59 Chronic myeloid a eR Re 12.12 | 132,600 | 3,766,400 61 Acute lymphoid wy Pesos 14.16 | 507,200 | 3,500,000 TABLE VIII. Amino-Acid Nitrogen Content in the Fractions of the Blood of Leucemia. ' Per 100 cc. of blood. Sample No. Whole blood. Red corpuscles. | White corpuscles. Plasma. 51 32.88 5.18 55 17.40 37.57 4.80 Bock has shown that red blood cells of birds are exceptionally rich in amino-acids. These cells, like human leucocytes, are characterized by the fact that they contain nuclei. Unfortunately 132 Amino-Acid Nitrogen of Blood we are not able to separate nuclei from protoplasm without changing the amino-acid content in them, so that the conclusion may not be definitely established that it is nucleus which contains amino-acid in exceptional concentration in the human white corpuscles and in the bird red blood cells. A suggestive fact is observed, however, in the analysis of birds’ eggs. Here the yolk corresponds to the nucleus, the white to the protoplasm. We found that the yolk of eggs of domestic fowls contained 32.79 mg. of amino-acid nitrogen per 100 cc., while the white contained only 6.49 mg. SUMMARY. 1. The amino-acid nitrogen content of the blood of twenty-eight fasting dogs per 100 cc. was found to vary from 6.33 to 8.79 mg.; that of rabbits is somewhat higher, from 7.22 to 10.60 mg. 2. Anesthesia does not influence the amino-acid nitrogen content of the blood during the course of 8 hours. 3. The thorough removal of the pancreas causes a transient increase of the amino-acid nitrogen in the blood; the increase begins within a few hours and lasts for at least 2 days after extirpation, after which it diminishes again even to a subnormal figure. When a part of the pancreas is left under the skin in connection with the blood vessels, the blood amino nitrogen remains normal. 4. Hypodermic injection of neither adrenalin nor pituitrin influences the amino-acid nitrogen content of the blood. 5. Neither the removal of the thyroid gland nor hyperthyroidism influences the amino-acid nitrogen content of the blood. 6. The ligature of both ureters or the extirpation of both kidneys causes a remarkable increase of the amino-acid nitrogen, in parallelism with the increase of urea and non-protein nitrogen in the blood. 7. Hypodermic injection of pilocarpin causes an increase of the amino-acid nitrogen content in the blood. 8. The amino-acid nitrogen content in the blood is increased in leucemia. The increase parallels the number of white corpuscles. 9. The analysis in fractions of the blood of leucemia shows that the white blood cells contain six or seven times as much amino-acid nitrogen as the plasma. It is suggestive that the nuclei S. Okada and T. Hayashi 133 of the white corpuscles are the essential part for containing such a high rate of amino-acid, though the conclusion is not definitely established. It is a pleasure, as well as a duty, to acknowledge our indebted- ness to Professor R. Inada for his constant inspiration and courtesy during this work. BIBLIOGRAPHY. . Van Slyke, D. D., and Meyer, G. M., J. Biol. Chem., 1912, xii, 399. . Gyorgy, P.,and Zunz, E., J. Biol. Chem., 1915, xxi, 511. . Bock, J. C., J. Biol. Chem., 1917, xxix, 191. . Okada, S., J. Biol. Chem., 1918, xxxiii, 325. . Lewis, R. C., and Benedict, S. R., J. Biol. Chem., 1915, xx, 61. . Myers, V. C., and Bailey, C. V., J. Biol. Chem., 1916, xxiv, 147. . Van Slyke, D. D., and Cullen, G. E., J. Biol. Chem., 1914, xix, 219. Bang, I., and Larsson, K. O., Biochem. Z., 1913, li, 193. CONDO PON Ee oe a ff sf Su “ Tice ant Paks oy » ‘ + tis tes G f af Fea: Cl q THE HYDROGEN ION CONCENTRATION OF THE INTESTINAL CONTENTS. By SEIZABURO OKADA anp MINORU ARAI. (From the Medical Clinic, Imperial University of Tokyo, Tokyo, Japan.) (Received for publication, March 26, 1921.) The intestinal juice has been regarded as distinctly alkaline. Auerbach and Pick (1) examined the intestinal juice of dogs and found the hydrogen ion concentration to be 0.5 to 5.01078, but, allowing for errors in the method, they estimated that the real reaction of the intestinal juice was 2 * 10-8. This reaction coincides with that of pancreatic juice estimated by them and with that of bile from the liver determined by one of us (2). It also corresponds fairly well to the optimal reaction of trypsin estimated by Meyer (3) and Michaelis and Davidsohn (4), to that of erepsin shown by Rona and Arnheim (5), and to that of the lipase of pancreatic juice found by Davidsohn (6). McClendon (7) determined the human adult duodenal contents taken with the duodenal tube and found most cases to be very near to 2 X 10-8. He also examined the duodenal contents of sucklings and found them to be 4.0 « 10-* to 1.5 & 1077, an average of 8 x 10-+. He reports that when the stomach contents of the sucklings are strongly acid, the duodenal contents are less acid, but when,.the former are weakly acid, then the latter are more acid. McClendon, Shedlov, and Thomson (8) examined the ileal contents of seven puppies from 4 to 46 days after birth and found the hydrogen ion concentration to be from 2 * 10-° to 1.8 X 10-7. McClendon, Shedlov, and Karpman (9) examined the contents of the ileum of four dogs from 232 to 43 hours after a mixed diet of cooked food, in two cases of which magnesium sulfate was also administered, and found the hydrogen ion con- centration to be 2.5 X 10-* to 2.5 x 10-7. Long and Fenger (10), using the duodenal tube, observed a variation in the adult duodenum from pH 3.80 to 7.81. Myers and McClendon (11), 135 H+ Concentration of Intestinal Contents 136 zo'9 = Hd ‘goint yovuloyg Lo? Ey ‘sort youuroyg o3'9 = Ha ‘aoint yowuloyg ert = Hd ‘gornt yovuloyg *SHIVUIOI s—OT X L'T s—-Ol X ZF s-Ol X FS s-OI X T'T s-Ol X Z'E .-OL X PT .-0OI X 9°S s-OT X LE geOL on Te +H Ha ”? ” ” ”? “OUlTSATV "[RAYQNO N "1B9]O pue MoTjeA ploy "IGOTO puv MoTeA ploy "IVOTO pue MoT[IA poy "IvOTO PU NstMole x *IBOTO puv YSIMOTIAA “prqing peVyM -OUIOS PUG MOT[O A “Pratny ‘PION PIPES 9 ” ” “OUlTeALY ”? ” ? ‘“IvOpD puve MOT[OA ‘reded snw4t] 0} UOTPOBAY ‘ping 944 Jo Aqyeng nin § wD a = S *[eour Jaqyye url, ”» ” “SUIYST *[eoul 480} SBOg-p[BVAnT “SUTYSV,T ‘d99 [ puv ded 9011 Jo sdno ¢ “BUIYSET “atu JO 09 00Z ‘SUTYStyT *[BOW “‘eqsueg *ASIINO[T *pioo yeurds ey} jo osvosiqy "BOLIYSVS TVITAYOV ‘QSVOSIP ,SOAVID "SI4TVVOLOUR pts alfa: | "BIIVOTYLL) ‘AydorysAp repno -SNUI OAISSOISOIg *sIsousBICT “squayvg wpidsopyy fo sjuajuog youapong ay) fo uoynmuau0g Uuoyl wabouph yy 9y J, ‘LT OTaVL ‘ON o[dureg 137 S. Okada and M. Arai s-OT X 1-OL X s-OT X s—-O1 X o—OT X o-OT X 6° we G 0° ¢° 9° s-OT X &'E OI X TT s-Ol X ZL s-O1 X FT ~ *[BaAp No Ny ‘ploy ”? “OULLRALV ”? ” ” “IvOT[D PUB MOT[OX *prqin} 979 “HT BCs, ” “AVOTO puv MOLVA “1O]O pue YSIMOT]O A ” ” ” “prqiny pus MOT[OA *prqany AT -010A0S PUB MOTIOX “prqiny A] “ploy j-O10A0S puv MOTO X ” "OUlBALV “PPV ” ” “OUr]eyLY “prqany STFU] B Pus MOTIO A *prqany yey “oulos ,, ” ‘prqing puw MOTI A “‘prqany O79 “HT 8s, ” ‘IVO[D ,, i “AUS puv MOT[OA fe ‘sao 7 pue Yp[rar jo *99 YOZ] *‘s350 7 puB YLT Jo *99 YOST] ” “SULYST | “sido 7 puB Y[LUT JO °99 YOT ‘sddo 7 pUw SITU Jo “99 00% "S390 Z PUB Y[TUL JO 99 OST ‘s390 7 pu YUH JO “99 OZ “BUIySV AT “ZOIp [BNS/) "S30 Z PUB Y[TUL JO °09 0OZ "S330 7, PUL I[TU JO “09 0OZ ” “Surljseyy ‘yavoy OY4 JO OSVOSIP ALVINATCA |Z7Z |L|GT “SISUIULOJSO[ADUY IZ | LPT ‘Snyypour sopoqeiq, |pP9 | OleL “SI[VSIOP SOGBT, |GE | LZT gg |S IIT “rowny [VIqoIaDH |6z |L|OT 138 H+ Concentration of Intestinal Contents applying the same method to one of themselves, found the reaction of the duodenum between 3 and 4 hours after various meals to be from pH 3.20 to 7.82. Michaelis and Mendelssohn (12) found the reaction of feces in the adult was 1 X 10-8, in the suckling 1 X 10-* tol X 10-7; while Howe and Hawk (13) found a hydrogen ion concentration of 0.15 to 9.8 & 10-8 in the adult. We examined electrometrically the duodenal contents of various hospital patients, removed with an Einhorn duodenal tube under TABLE II. The Hydrogen Ion Concentration of the Intestinal Contents of Dogs. $| 5 | 5 2| 3 ven |5,| Pete | eae re aE} 3 # Es a\ala Et... kg. hrs 1 | 6.6} 9 | 1 Ib. meat.) 3 | Last part of | Alkaline. |6.78)1.7 X 1077 ileum. PING 7l gat tae 1 | Last part of 71763 Xi? duodenum. 315.51 to |1% “ |2 | Last part of a 7.68|2.1 X 1078 duodenum. y #. °° | 2 | ast pare ot) eee 6.59|2.6 xX 1077 ileum. ey AR ee Ue ee = 2 Last part of cf 615) Sek Oe AO duodenum. 1“ “ |2 | Last part of| Alkaline. |7.83)1.5 X 1078 ileum. 5 | Pup of 25 days, whose stomach was filled with coagulated milk; the pH of which was 3.82, the pH of the ileum was 5.92. various conditions, and the duodenal and ileal contents of dogs taken out directly. The results obtained will be seen in Tables Iand Il. > In fasting, the duodenal contents are mainly composed of a mixture of bile and intestinal and pancreatic juices. That both the latter juices are alkaline has been described above. The bile from the liver is also alkaline, but this is mainly poured into the duodenum by the time of digestion and only the bile from the gall bladder will be occasionally poured into it. The reaction of the S. Okada and M. Arai 139 bile from the gall bladder is variable; it may be acid, neutral, or alkaline. The hydrogen ion concentration was found by Okada (2) to be from 4.7 X 10-*to3.4 X 10-8. The reaction of the duodenal contents may not logically surpass the limit of the reactions of these fluids. Of the eleven cases examined by us for the duodenal contents when fasting, we found eight cases to be alkaline and three cases acid, the hydrogen ion concentration of which was from 2.6 X 10-7 to 1.3 X 10-%. By the time of digestion the acid chyme may be poured into the duodenum and the secretion of the digesting juices may be copious and in pathological cases the action of microorganisms must also be accounted for, so that the relation is somewhat confused. Of the twelve cases examined we found that seven cases were acid, four cases alkaline, and one case was neutral, the hydrogen ion concentration being from 1.6 * 10-* to 1.1 * 10-%. The tests for free hydrochloric acid were always negative. No special relation seems to exist between the acidity of the stomach and the reaction of the duodenal contents. The reaction of the duodenal or the ileal contents examined in four dogs was acid or alkaline. On the reaction of ileal contents, the action of the microorganism seems especially to dominate. One puppy was examined and the ileal contents were found to be slightly acid. BIBLIOGRAPHY. 1. Auerbach, F., and Pick, H., Arb. k. Gsundhtsamte, 1912-13, xliii, 155. 2. Okada, S., J. Physiol., 1915-16, 1, 114. 3. Meyer, K., Biochem. Z., 1911, xxxii, 274. 4. Michaelis, L., and Davidsohn, H., Biochem. Z., 1911, xxxvi, 280. 5. Rona, P., and Arnheim, F., Biochem. Z., 1913, lvii, 84. 6. Davidsohn, H., Biochem. Z., 1912, xlv, 284. 7. McClendon, J. F., Am. J. Physiol., 1915, xxxviii, 180. 8. McClendon, J. F., Shedlov, A., and Thomson, W., J. Biol. Chem., 1917, Xxxi, 269. 9. McClendon, J. F., Shedlov, A., and Karpman, B., J. Biol. Chem., 1918, xxxiv, l. 10. Long, J. H., and Fenger, F., J. Am. Chem. Soc., 1917, xxxix, 1278. 11. Myers, F. J., and McClendon, J. ¥., J. Biol. Chem., 1920, xli, 187. 12. Michaelis, L., and Mendelssohn, A., in Michaelis, L., Die Wasserstoffion- enkonzentration, Berlin, 1914, 112. 13. Howe, P. E., and Hawk P. B., J. Biol. Chem., 1912, xi, 129. bbatd leg, ah a wie aa ‘ int = ae oS izol FO ve it a rity AO aa quite Ron Oe hes an rows ry) mG noid v4 a abit , woot OA) Dee sf an ih od? I¢6 ine teste, + 1-01, falewA th ay it “4 fy Ay ik" 128 Dae oie THE INFLUENCE OF PUTREFACTION PRODUCTS ON CELLULAR METABOLISM. II. ON THE INFLUENCE OF PHENYLACETIC AND PHENYL- PROPIONIC ACIDS ON THE DISTRIBUTION OF NITROGEN IN THE URINE. By YOSHIZUMI HIJIKATA. (From the Medico-Chemical Institute, Kyoto Imperial University, Kyoto.) (Received for publication, January 31, 1921.) The investigations of H. and E. Salkowski! have established the fact that in the putrefaction of proteins, aside from other products, two are formed which are homologous with benzoic acid, namely, phenylacetic and phenylpropionic acids; and that these acids when introduced into mammals, reappear in the urine conjugated with glycocoll, the phenylacetic as phenaceturic? and the phenylpro- pionic, having been oxidized to benzoic, as hippuric acid. It is possible, of course, that putrefaction of protein material in the intestinal canal may give rise to the acids investigated and thus account for the occurrence of phenaceturic and hippuric acids in normal urine. Unfortunately no attempts have thus far been successful in establishing directly the presence of either phenyi- acetic or phenylpropionic acid in the intestinal contents or excre- 1 Salkowski, E., Z. physiol. Chem., 1878-79, ii, 420; 1883-84, viii, 149; 1885, ix, 229, 491. Salkowski, E., and Salkowski, H., Z. physiol. Chem., 1882-83, vii, 161, 169; Ber. chem. Ges., 1879, xii, 107. 2 Recently, Thierfelder and Sherwin reported that, after administering phenylacetic acid, there appeared in the urine of man, not phenylaceturic (as in the case of dogs and rabbits) or phenylacetornithuric acid (as with birds) but phenacetylglutamic acid which could be isolated, partly com- bined with urea and partly uncombined (Thierfelder, H., and Sherwin, C. P., Ber. chem. Ges., 1914, xlvii, 2630-34; Jahresb. Thierchem., 1914, xliv, 1058; Z. physiol. Chem., 1915, xciv, 1). 141 142 Cellular Metabolism ment,’ possibly, owing to the difficulty of isolating small quantities of these acids from such mixtures. Several studies have been made of the influence of the acids on the excretion of nitrogen in the urine. E. Salkowski* has shown that, on feeding phenylacetic acid to rabbits, the nitrogen excreted in the urine is increased. Desgrez and Guende’ have reported that phenylpropionic acid causes a decrease in the nitrogen in the urine when fed to guinea pigs; the decrease being due to a smaller excre- tion of urea. This leads to the conclusion that two homologous aromatic acids, which probably arise from the putrefaction of protein material in the intestine and which can combine with the same nitrogen-containing conjugate, glycocoll, are playing quite opposing réles in the excretion of nitrogen in the urine. If we assume with Salkowski and others* that the introduction of benzoic acid causes a marked increase in the decomposition of tissue pro- teins, the same effect must also be ascribed to phenylpropionic acid, since it is oxidized in the animal body to benzoic acid. Con- trary to Desgrez and Guende’s statement, therefore, introduction of the latter substance should cause an increase in thenitrogen excreted in the urine just as in the case of phenylacetic acid. The assumption is justified by the results of the following experi- ments which were carried out in the manner described in an earlier > paper.” A number of the rabbits used were maintained in N bal- ance by feeding “‘tofukara,’’® while others were fasted. The total nitrogen was determined by the Kjeldahl method, the urea by means of urease, the ammonia by the method of Kriiger, Reich, and Schittenhelm, and the amino-acids by Van Slyke’s method. 3 Tappeiner has shown the presence of phenylpropionic acid in the paunches of ruminants fed with hay; but it is uncertain whether the acid existed preformed in the hay or was formed in the alimentary canal from protein (Tappeiner, H., Z. Biol., 1886, xxii, 236-40). 4 Salkowski, E., Z. physiol. Chem., 1888, xii, 222. 5 Desgrez and Guende, Compt. rend. Soc. biol., 1906, lviii, 526. 6 Salkowski, E., Z. physiol. Chem., 1877-78, i, 1; Virchows Arch. path. Anat., 1888, exiii, 394. Virchow, Z. physiol. Chem., 1877-78, i, 78. Kum- agawa, M., Virchows Arch. path. Anat., 1888, exiii, 134. 7 Hijikata, Y., Acta scholae medicinalis universitatis imperialis in Kyoto, iv. 8 Tofukara, a waste product obtained in the preparation of bean cheese, was pressed daily and a weighed quantity soaked in a fixed amount of water. - = Y. Hijikata 143 Experiments with Phenylacetic Acid. Experiment 1.—A 1,900 gm. rabbit, in N balance, received on Oct. 30, 0.5 gm. of the acid? mixed with his food and on the following day, 1.0 gm. In both administrations the acid was neutralized with sodium carbonate solution. On both days in which the acid was given in the food, the excre- tion of amino-acids increased strikingly, while that of the total nitrogen was unchanged, causing a considerable increase in the proportion of the former to the latter. The excretion of ammonia was unaltered. A small decrease in the ratio of urea to total nitrogen is mainly due to an insignificant decrease in the absolute quantity of urea. The fact that there was no increase in the total nitrogen excreted, was to be expected since the amount of acid administered was too small to cause any decomposition of tissue proteins. Experiment 2.—A rabbit, weighing 2,420 gm. and in N balance, received on both Nov. 13 and 14, by means of the stomach tube, 1.2 gm. of phenyl- acetic acid, dissolved in sodium carbonate solution. It is apparent that on both days in which the acid was given, not only the output of amino-acids, but also those of total nitrogen, urea, and ammonia were increased, in spite of a considerable inter- ference with the appetite. While the relative quantity of urea was very slightly decreased, the relative quantities of ammonia and amino-acids were more than doubled. The experiment was ended when the animal refused to take any more food. Experiment 3.—The rabbit weighed 2,830 gm. Until fasting was begun, it received tofukara; then 35 cc. of water once daily through the stomach tube. On both the 4th and 5th days of fasting, it received 1.0 gm. of phenyl- acetic acid, neutralized with 35 cc. of sodium carbonate solution through the stomach tube. The total nitrogen excretion decreased during fasting until the acid was administered. There was then an increase for 2 days ° The phenylacetic acid was prepared by Totani, according to Mann; and was analytically pure. It was identical with that used in his experi- ments reported in his famous work, “Uber das Verhalten der Phenylessig- siure im organismus des Huhns.’’ I take this occasion to express to him my thanks for his courtesy in supplying it (Totani, G., Z. physiol. Chem., 1910, Ixvili, 75; Mann, W., Ber. chem. Ges., 1881, xiv, 1645). Cellular Metabolism 144 EE ee "U0}v9 SUIYJON ‘u0489 YOU WOI4VI s,Aep sty} jo $ ‘oqn} yovmo4s Yysno1y} prow o1yeov,Auoyd “Ws ZT ‘u9}89 JOU UOI4RI s,Aep siq} jo ¢ ‘aqn} yovuro}s Ysnoi1y} plow oiyeowyAueyd ‘wis ZT ” *qoIp OY} Ul prov o1yooeTAUoYd “us G°0 ”? ” ” “SY IVUIOY “syIVULOY ” » OT NTOA zo'z | 0°18 | ze'0 | Z'OT | 80°88 | O'1g8‘Z | ¢°S0%‘S | TZO'T | Sze | 088s 19°2 | 62 | 2¢°0 | gor | 12°28 | 2°9T9°% | $'698'S | TOT | SFT | OZF'S zit | 9°22 | 220 | ¢'r | 28°88 | P'F6L‘T | Z'0Z0'S | GIO'T | S6T | OFFS zit | $ez | sto | te | 29°88 | ¢°SF8‘T | 2°F80'S | OTOL | OFZ | OFFS sL1 | sez | 200 | pz | g¢'88 | 6'Sg8‘T | 6960's | 0ZO'T | 00% | OZr*% srt | zez| I1'0| az | 21°88 | F6cL‘T | P'196‘T | STO'T | 002 | Oars AE Pte| tun VP] ple 2) Aloe ae | ou 00 | "wb ‘N plov-oulury | “N Braowuy “N very, *N 1F70.L gubede fees an “II WIAVL 90't | t'zt | tro | 2°t | 02°28 | SOTF‘T | F°809‘T | OZO'T | OST | O86‘T G10 | T'Z1. | 60°0 | ST | SP'68 | 2°8Eh‘T | F'809'T | 8TO'T | SOT 086 ‘T e120 | 2:11 | 80°0 | ST | 89°28 | &°SOF‘T | 2°Z09°T | 8TO°T | O9T 086 ‘T 96°0 | 6' FT | 80'0 | &°L | 22°88] 6 28e'T | 6° 2gS°T | OZO'T | SST OF6‘T ore | #o¢ | OL'0 | OL | 2G°F8 | O'FSE‘T | 2°ZO9'T | TZO°T | OFT | 086'T eo'z | ore | 110 | 2'T | 09°S8 | 9°608‘T | 6'6zS‘T | OZO'T | OST | 0Z6'T 73°90 | L'F1 | O10 | 2:1 | £8°68 | Z:009‘T | 0'029‘T | OZO'T | SOT | OT6‘T 92°0 | 9°11 | o1'0 | st | 92°88 | O'8ce‘T | 6°GZS‘T | OZO'T | O9T | 006°T 13°0 | 6°zE | 11:0 | 2°L | #688 | G’OTF‘T | 6°S8G'T | 6TO'T | GOT | 006‘T 03'0 | gz | g0°0 | zt | 2°88 | err‘ | 1°26S‘T | OZO'T | OLT | 006°T ARNG one WEE 06 Ate ae | Sa vo | wo ‘N plov-oulmry | ‘N viuomury “"N BoIQ *N [279.L Spee ae ae al ” BE 3» €1 ”? CL» LT OL => 6 “AON 8T6r o7eq OoonnAnNn ost ie ° 7, 8I6T ra | ee eee ee) si Ne eee ‘I OUTaVL 145 Y. Hijikata G80 | 0°42 | T9°0 | 8'9T | 92°48 | 2°246‘% | 6 E82‘ | TOD 4 86. S0etee Gio1 | 0865) G60. 1°29 -28°se.) Sven" T. horees's x POST UBB) MORZaeu ae o8'I | $' bE | 320) 8 F | 16°28] Z°S90'T | 2° F68'T » |980'T|0L jloze'z| 9 ” »” ” 99°% | © 99 | S80 | 0'6 | IS°F8 | S’ZgI‘zZ | 0'6Rgg‘Z <2 [| SSRORL 4 2B. ORR ot ‘aqn} yowwMoys ysnosyy prow oreouyAuoyd “w3 O'T | $9 | S16 | 28:0 | 9°¢ | 6o'zs | 2°:6Fs‘T | L'eIg‘T » | SPOT] 2h lore’ | F 18°0 | O'IT | $2°0 | FE | 90°28 | T'2ZZE‘T | O'zee'T » | SOT | Sr jose’s]| & SLT. | PSl| £60 | S P= | 90-88.) 62207 T-vee'T » | ShO TL] Sh joso'z| z €L°T | 1°02 | S20 | SF | 90°68 | T'z6S‘T | L'zgZ‘T ‘ploy | Ss0'T | G2 jozz‘z} TI ‘Burpooy Jo Avp ys] | ¥8'0 | G'9z | 12'0 | 2°9 | 82°26 | FGe6‘z | e-goT‘e ‘OuTTBATV | OZO'T | Sez jogs‘z jtaeal Oe |ilied| em | ALE] com | ow | uo “APABIS eee “FU SIOM | Burysey ‘S| IVULOYY ‘N plov-ourury | “N vruourury ‘N Borg, ‘N [870], | ‘worzovoyy oytoadg oe oa Apog | jo ug en ee 8 eee ae ee eee ‘hog yoo aqny, yoouorgy ybnowy) uM 4 poz}USIG ‘9 EE ‘TI ATAV.L 146 Cellular Metabolism followed by a decrease. The outputs of urea, ammonia, and ami- no-acids showed increases; the urea absolutely and the others both relatively and absolutely. Experiment 4.—The rabbit, weighing 3,130 gm., was fed tofukara until the beginning of the experiment, after which it received neither food nor water. On both the 4th and 5th fasting days, it received 0.5 gm. of phenyl- acetic acid, twice daily (10 a.m. and 5 p.m.). The acid, neutralized with 10 cc. of sodium carbonate solution, was injected subcutaneously. The results agree with those obtained in the preceding experi- ments in which the animal received the substance per os. Experiments with Phenylpropionic Acid. Experiment 5.—A rabbit, weighing 2,550 gm., was maintained in N bal- ance. It was given phenylpropionic acid?® during two periods of 2 days each. The acid was neutralized with sodium carbonate solution and intro- duced through the stomach tube. On both Dec. 15 and 16 the animal received 1.0 gm. of the acid. On both the 21st and 22nd of Dec., 1.5 gm. were administered. The total nitrogen and the ammonia excreted were practically unchanged when the acid was introduced. The output of am- monia increased strikingly on both occasions, both absolutely and relatively. In the first case, urea decreased relatively and in the. second both absolutely and relatively. The amino-acids were greatly increased. Experiment 6.—A rabbit, weighing 2,560 gm., in N balance, was treated in the same manner as the preceding one but with larger doses. On Dec. 23 and 24, it received 2.5 gm. daily of neutralized acid and on the 28th and 29th, 3.5 gm. daily. On the first day in which the acid was administered, the total nitrogen excreted was increased. On the second day, it was nor- mal. There was, also, an absolute increase in the output of urea on the first day. Relatively, however, it was decreased on both days. Both absolutely and relatively the ammoniaand amino-acid excretions were increased on both days on which acid was given. During the second period of feeding the acids, similar changes were observed but of greater degree. 10 The acid was prepared by the reduction of cinnamic acid with sodium amalgam as described by Erlenmeyer and Alexejeff (Erlenmeyer and Alexe- jeff, Ann. Chem., 1862, exxi, 375. 147 Y. Hijikata SPT | 8°98 | 020 | O'S GET |eSree, | OO | 67s c6'T | 6 L2E | FLO | 8G GO DSS SAO: |e. S,. ” ” ” GO" |) SOL te.0 |e.0 ‘Ay[worutrepodAy peyoolur ‘UOIZN]OS BPOS “90 OT paapos -sip ‘plow o1joov,Aueyd “us G0 (urd ¢ pus ‘ue OT) BOIM, | OS'S | 9°98 | LTO | IF G81 |9°S8 | 710192 16° 1) G-86 | OT.0 | % 2 L381 | 862) 120 | 7 ‘Burpooy Jo Avp ysv'T | €8°L | GSE | 910 | 6'E Hi] com |fseee| A “Sy IBUIY "N plov-ourmy | "NV vruourury 99°06 | 9° $92°% G0'68 | O'8FI‘S 9°68 | T°892‘T Go'98 | O'FIO'S FL°S8 | 6°99GS €6'F8 | 0°S20°Z 86°68 | 8° Se9'T 0°28 | F 1e8‘T IP'98 | 2 S2z8‘T 69°16 | L°L8%‘% Md'ieh| 0m “N Borg ? ” ? ”? ” ‘ploy “OUITeALV 1¥0'T | 09 jose'z | 6 140'T | #9 jOTh's| 8 crO'T | 09 joos‘'z| 2 OFO'T | 92 joss‘z}| 9 4401 | G8 |069°2 | ¢ 9F0'T | 69 |008‘c| F ctor | 1 joss‘z|] & ZrO | 8h j0s6's| Z Geol | 09 joro‘s | T S101 | Sse josr‘s Says) “wb “N 187O.L UOTJOVAT “APLIABIS jo og1oods |aunjo, ‘AI WTA VL “PY SIOM |" Surjsey Apog | jo Aeq Cellular Metabolism 148 69°0 | S'ST | OL'0 | O'S | OL'06 | T°ZS2‘T | 9° FFE'T | STO'T | OFZ | OLG'T | FS >» €2'0 | L'FL | 11°0 | 2% | 8G°06 | G9FZ‘T | 8'226‘T | OTOL | OFZ | 09G'S | Ee » » » ” Oc’? | 2°78 | ZT°0 | T'S | Tr'98 | 2°2429°T | 68ST | SIO'L | O22 | 09ST 1 ce “aqn} YovVUoys ysno1yy pros oruordoadjAuoyd ‘wS gt | 90°F | FSS | $20 | SF | Fe'L8 | LZ TLL‘T | 9°820°S | 6IO'T | O1Z | OLGZ| TS » 92°0 | 2° | 62'0 | 9°S | F016 | S'S9L‘T | O'6E6'T | 9TO'T | OZZ | O9G°Z/02 » 6L°0 | &°ST | 92°0 | O'S | 16°68 | O'SFL‘T | O'686‘T | ZIO'T | OTS | O9S'S | 6I » 92°0 | 6 FL | 42°0 | 8S | 62°68 | STFL‘T | Z:0S6‘T | ZTO'L | OZ | O2G°Z | SI » @)°0'1 8:eT | 12:0 | Sh | TL06.| O SE8-1.| SPOS | .SIOT |, O12 |. OLE Z2F : » ” ” Te | S:8h | 26.0 | O27, | 80° 28.1 21022 18°0 | 8'8T | ST'0 | €°S | GOOG | T'SS6‘T | F'FLT‘S | STO'T | OFZ | OGG'Z) 22 OC CLO TM) ILO | SS | B8'68:| 2962 TL.) S000. Sul LO Ta Ose Oe et yy 1Z°0 | TST | O10 | 3S | GL'06 | S°HE'T | L'OFT‘S | STO'T | OFZ | 099‘ | 02 99d Roe nts) aes Meteobengctl cy il Age acts ce ote “out ‘9 | ‘wd si6t ‘ounn *BYIBUIOY *N plov-ourmy | ‘N Bruouumy *N vory, "N [BIOL ombedes wuteon ie *oyBq ‘IA GTA VL Cellular Metabolism 150 i 660 | OPE 06°0 | 1 TE 60°T | Te ” ” ” OI’ 2 | 6° 8T% ‘aqn} YoORulojs YSnoryy prow oruordordy;Aueyd “ws Q'g | 98'S | 6 TPZ 080 OL T | 2 FG Of T | 88% Set | F CE ZT | LP€ “SuIpooy JO Avp ysuT | 6O'T | POE N 20797 |. quad sad ou *SyIBUI0YY *N plov-ourury Gs'0 | FTE | 96°06 | O'TEe‘s | F'829°E 3» | Gaur es 98°0 | 8°21 | SI'16 | P'LPI'S | L'OPP'S ”» 1h0'T | 16 6b'0 | O'FT | 92°06 | ¥'699% | 8'9FS‘S » ORO" T ee gcd | 8°91 | 00°28 | F'2z9'Z | 2°280'S » | @80'T "SOF LS"| OS eS | OSLE'S 2168's » | PSO'T | SOT €1'0 | 8% | 98°06 | O'I66‘T | 2 161'% Pr all (iy 30 iy Ue ed > eT'0'| SS | F208 | 7 S46 T | O'el2 cS » | e0 TC) Pe Z1'0|8'% | €8°68 | T'20T'2 | S'ahe 2 Pele BOOM DA eee TO | Oe” | CG6s | G'ese's | e"SIS's ‘ploVv | OO'T | S2T F101 9°P | 82Z'16 | g'gg0‘e | 8'8zE's | OUNTVATV | LZTO'T | SEE ae (leas “Bu fae 9 thas “Bu “Bue 99 : e : alae ak ‘APIAGIB ‘oun N vruowwy N Bory) N [®9O.L WOTpOVOY aytoodg ana ‘fipwg eqny, yoouoiy ybnowy? 1270 44. P27]U8U *9D SS TIA WTA VL 080‘2 OSL‘S 08%‘z 028'% 0gc‘s 089‘% OFL‘S OFS‘% 06'S OzI‘s “ub “QU SIOM | Furgsey Apog OD Hu jo Aud 151 1.70 ‘AUp 4xou pvop punoy | 96'T | 2°eT | 910 | TT | FO'6L | L'eg¢ ¢ 002 » | 90D | 02. 086 fb las 70'% | G'LZT | 0€'0 | 9% | O8'8Z | G°Ez9 F'Lg8 s | OF0'L | #2 — 1068'L | -9 ” 9 % €6°8 | LT'6IT] 08'0 | 6S | 22°08 | 6'O90‘T | S'aze'T 8 |) BGO" Te TOS. OZ0REar 2 “AT [BOTUL -lapoddy poqoofur ‘WoTyn] 8 -OS Bpos ‘00 0% ul PoeATOSstp a prow oruordoadjAuoyd “wi = O'T (urd g puv wv OT) OOIMT, | 682 | F'6OT| ST'O | 2'% | 24°18 | 6 eIs’‘L | S'6ZF'T » | 880'L | 06 o9l's| F an eT eSL | Tee S'S | Te’ ee ie" 226 9 FFO'T x (O80 T, | SO% Re al ¢ . vG'T | O'IT | 28'0 | 8'°Z | 08°98 | F892 P S88 » | 820°T (00 joze's! Z > Zo". OrE ee) | O'S: Te" ee: t Tee G'S16 ‘ploy | GzO'T | OL Osr's] T ‘surpooy JO Avp ysv'T | 98°0 | OZ | 80°0 | 2:1 | $¢°26 | E'868'T | E'1¢0‘z | ourpexty | o1o'L | 09% 019'% hat “Du Bal ah ft hceg eae 14 Ou Du ‘29 “Ud “eis : breil ; ‘ Wee: e F yy | Apavaa | PFT custom gery YAVULOYT, N plovroutury N viuourury N vou) N [®10, WOTPIVOY oytoodg jo Apog |jo Auq ouUIn[OA st et ee 8 a ee eee TIA WTa VL Cellular Metabolism 152 ” ” ” * AT [BOTUL -ropoddy poyoofur ‘uory -N[OS BPOS "09 GT UL poATOssIp ‘prow oruordord;Aueyd “wa ZT (Curd g puv ‘u's OT) 90147, ‘Burpooy Jo Aup 4su'T *BYIBVULOY 98'°T | T & cO'T | 9°&% LE'T | 0°26 66°6 | P'8TZ 98°6 | GIST 98°0 | IFT PL E16 oT PIT | 6°02 iat isBale ager om ‘N plow-ourmy 122) - o SOO OON MOAN 810 610 LT'0 610 02°0 N 70207) quad wad Ou N oD OwH19 on = ON oD ‘N Bruowmury 18°62 | 9 97°98 | 0° 98°28 | $°36% 00°06 | GPP 6P'°6 | 9° SLL N 1°70} 5 qua sad ou "Ni Bor, 0°869‘T 0°029‘T 9° ZL6‘T P'20S'S ‘N 1890.1 “XI UTaAV »”» ” ” “ploy “OUPVALV “MOTJOVOY “APAVIT oywodg ‘oun jo ouInjoA “QySIOM | DurySVy Apog in O rr 0 jo Aq wee Y. Hijikata 153 Experiment 7.—The rabbit, weighing 3,120 gm., was fed with tofukara for a time and then fasted. During the experiment it received 35 cc. of water daily. On both the 5th and 6th days, 3.0 gm. of the acid, in 35 ce. of sodium carbonate solution, were introduced through the stomach tube. The total excretion of nitrogen decreased regularly until the acid was administered. It then increased, with the exception of the seventh day, until death. The urea, also, rose absolutely beginning with the fifth day, but on the 2 days in which the acid was given it decreased relatively. The ammonia excretion paral- leled that of the total nitrogen. There was a great increase in the quantity of amino-acids both absolutely and relatively. Experiment 8.—A fasting rabbit, weighing 2,610 gm., received, on the 4th and 5th days, two subcutaneous injections daily of 1.0 gm. each of phenylpropionic acid in 20 cc. of sodium carbonate solution. On both days in which the acid was given, the total nitrogen, as well as the urea and amino-acids, was increased. Relatively, however, the urea decreased. On the second day, the absolute quantity of ammonia was increased but there was relatively no change. ‘The increase in the amino-acid excretion was both abso- lute and relative. Ezperiment 9.—The rabbit weighed 3,210 gm. On the 4th and 5th days of fasting, it received, twice daily, 1.2 gm. of phenylpropionic acid. The acid, neutralized with 15 cc. of sodium carbonate solution, was injected subcutaneously. The result of the experiment was to confirm the foregoing ones. DISCUSSION. The experiments reported show that both substances under investigation have the same effect on the distribution of nitrogen in the urine. The effect of the phenylpropionic acid is perhaps smaller. The total nitrogen, as well as the ammonia, shows no essential change with asmall dose. An insignificant decrease is noted in the - output of urea. No simultaneous decreases in the excretion of total nitrogen and urea, such as those reported by Desgrez and Guende, in the case of phenylpropionic acid, were observed. The amino-acids in the urine showed a very considerable increase, not 154 Cellular Metabolism only absolutely but relatively. Although no attempt was made to isolate glycocoll, it seems quite probable that the increase may have been largely due to this substance in the form of a conjugate. The fact that the increase in the amino-acids is accompanied by a decrease in urea suggests the possibility that the amino-acids may arise from urea or its antecedent. The results may be of importance in clarifying our conception of the formation of glyco- coll in the animal body, no experimental reduction of the urea in the urine having been observed heretofore, on introducing ben- zoic acid or analogous substances. When the dose was increased the total nitrogen excretion was, without exception considerably augmented, thus confirming the observations of Salkowski on phenylacetic acid. The increases in the absolute and relative quantities of ammonia and amino-acids are remarkably great. The urea also rises but not in proportion to the total nitrogen. The increase in the ammonia, even though the acids were neutralized, favors the view that the simultaneous increases in the quantities of total nitrogen and urea excreted may have been due to a pathological decomposition of tissue proteins. These changes were observed not only with rabbits maintained in N balance but with fasting animals. ON THE CLEAVAGE PRODUCTS OF THE CRYSTALLINE LENS. By YOSHIZUMI HIJIKATA. (From the Medico-Chemical Institute, Kyoto Imperial University, Kyoto.) (Received for publication, January 31, 1921.) Since the study of the hydrolysis of the proteins has shown that they are composed of various combinations of certain substances, which, for given bodies, are united in characteristic proportions, it is desirable that as many protein substances as possible be exam- ined in regard to their cleavage products. While Mérner' has found that the protein substances, obtained from the lens of the Ox, occupy a special position among the proteins, no detailed investigation of the cleavage products of the lens has been reported. Under these circumstances, it is worth while to report the following results of an investigation of the cleavage products of the crystal- line lens of the ox as a contribution to the knowledge of the chemi- cal composition of the lens tissue. EXPERIMENTAL PART. The material for the investigation was obtained from a nearby slaughter house and was always perfectly fresh. The lenses were freed from all foreign tissue and used immediately. Water and Ash Content. The water content of the fresh material was determined by drying to constant weight at about 105°C. 1 Morner, C. T., Z. physiol. Chem., 1894, xviii, 64. 155 156 Cleavage Products of Crystalline Lens Experiment No. Fresh substance. Substance lost. gm. gm. per cent 1 2.2627 1.4352 63.46 2 2.2393 1.5113 64.06 3 2.1787 1.3872 63.67 4 2.3805 1.5546 65.36 5 2.3090 1.4832 64.23 6 2.6449 1.5929 64.04 7 2.6509 1.6294 65.25 ASVOTA BO oi. 15.50 Sioloaio mio she bel geioeisieieicle See eR Ee ee 64.30 Experiment No. Fresh substance. Ash. gm. gm. per cent 1 2.3690 0.0187 0.79 2 2.4954 0.0192 0.77 3 2.5219 0.0184 0.73 AVETAEC Sie co 00's hem cle tare es ein are oe ne 0.76 Hydrolysis of the Lens by Hydrochloric Acid and Fractionation of. the Esters of the Monoamino-Acids. 750 gm. of fresh lens were dissolved in 3 liters of concentrated hydrochloric acid by warming on the water bath, and the solution was boiled under a reflux condenser for 8 hours. A negative biuret reaction indicated that the hydrolysis was complete. After separating 2.7 gm. of dark brown material by filtration, the filtrate was concentrated to a syrup under diminished pressure. The residue was taken up in 1 liter of absolute alcohol, the solution saturated with dry hydrogen chloride and boiled. ‘The alcohol was then removed under diminished pressure and the esterification repeated three times. No glycocoll ester hydrochloride could be separated. The esters in the residue were liberated by treatment with so- dium hydroxide and extracted with ether in the presence of potas- sium carbonate. After drying 12 hours over sodium sulfate, the ether was distilled and the esters were fractionated. Y. Hijikata 157 Fraction. Bath. pine te of Pressure. Amount. mT i aero Ud gm I Water Up to 60 12 46.7 II “s 60-100 12 69.6 ERE Oil. 100 0.5 4.0 IV “ 100-175 0.5 17.9 The residue weighed 97.2 gm. Each of the first three fractions were boiled, under a reflux con- denser, with 10 volumes of water until the reaction was no longer alkaline, and then evaporated to dryness under diminished pres- sure. In order to remove any proline the residues were boiled with absolute alcohol and the extract united. The proline was isolated in the usual manner as the copper salt. The active and racemic salts were separated by means of boiling absolute alcohol. The yields were 5.6 gm. of active, and 1.8 gm. of inactive copper salts, dried at 120°C. 0.1827 gm. of air-dry inactive copper salt lost 0.0196 gm. at 120°C. Calculated for CioHieOs N2Cu.2H20. Found. per cent per cent oo) a ae 10.99 10.37 0.1631 gm. of dry inactive copper salt gave 0.0437 gm. CuO. Calculated for CioHisOsN2Cu. Found. per cent per cent 0 A OS 21.81 21.40 The free active proline melted at 211°C. An aqueous solution of the amino-acid was strongly levorotatory. 0.0992 gm. of substance neutralized 8.3 cc. of 0.1 N H2SO3. Calculated for CsHsO2N. Found. per cent per cent oe oy ARE EC aa oe on 12.17 11.72 Fraction I.—The material remaining after the extraction of the proline was suspended in absolute alcohol and dry hydrogen chloride introduced. No glycocoll was found. The solution was freed from hydrochloric acid by boiling with yellow lead oxide; the lead removed by treatment with hydrogen sulfide; and alanine and leucine were isolated as their copper salts, of which 3.2 and 0.2 gm., respectively, were found. 158 Cleavage Products of Crystalline Lens 0.1138 gm. of alanine copper salt gave 0.0378 gm. of CuO. Calculated for CeHi204N2Cu. Found. per cent : per cent OTUs aise bre so Sire Athoyeinia's Galea ee Dee eee 26.50 26.56 The alanine hydrochloride was dextrorotatory. 0.1092 gm. of substance neutralized 11.8 cc. of 0.1 x H2SO.. Calculated for CzH702N. Found. per cent per cent IN cavanavanalaretay evel atest oe Yate ois thane Glow are Se eee 15.73 15.38 Fractions II and III.—These fractions, after removal of the proline, were united and converted into copper salts. In order to separate the valine and isoleucine the mixture was exhaustively extracted with methyl alcohol according to Ehrlich and Wendel.? The soluble fraction was decomposed and the amino-acids were racemized by heating with aqueous barium hydroxide at 180°C. in an autoclave. The copper salts were again formed, and extracted with 96 per cent ethyl alcohol. On recrystallization of the soluble fraction, 0.2 gm. of copper salt was obtained. The analytical figures, unfortunately, did not check well with those of the copper salt of isoleucine. The yields of the copper salts of valine, leucine, and alanine were 3.2 gm., 17.2 gm., and 13.5 gm., respectively. There was a further yield of 3.7 gm. of free leucine. Leucine. 0.1372 gm. of copper salt gave 0.033 gm. of CuO. Calculated for CyzH2sOs.Ne2Cu. Found. per cent per cent Cua. .ckkisdes doe ek eans deeeekaaerteeee 19.65 19.56 0.1653 gm. of leucine gave 14.6 cc. of moist nitrogen at 5°C. and 758 mm. Calculated for CceHi1202N. Found. per cent per cent Ni esedee da ERR SE REL Bes See ene oe eee 10.69 . 10.77 The optical rotation of 0.6369 gm. was measured in 20 per cent hydrochloric acid. The total weight of the solution was 17.3998 gm. and its specific gravity 1.1. Ina2dm. tube at 20°C. it rotated sodium light + 1.27°. lol?” = + 15.77° 2 Ehrlich, F., and Wendel, A., Biochem. Z., 1908, viii, 399. Y. Hijikata 159 Alanine. 0.1795 gm. of copper salt gave 0.0595 gm. of CuO. Calculated for CsHi20:N2Cu. Found. per cent per cent OU. ado lea) gE eee eee 26.50 26.42 Valine. 0.2090 gm. of copper salt gave 0.0560 gm. of CuO. Calculated for Ci0H2001N2Cu Found. per cent per cent “DU. fit 230 3 A eee eae 21.51 21.41 Fraction IV.—This fraction was worked up in the usual manner. After the phenylalanine ester had been extracted with ether, the insoluble residue was saponified by barium hydroxide and the glutamic acid sepa- rated as hydrochloride. The mother liquor was freed from hydrochloric acid and divided into two equal parts for independent estimations of the aspartic acid and serine. $-naphthaline sulfoserine could not be isolated in a perfectly pure state. 2.2 gm. of aspartic acid were recovered, equiva- lent to 3.5 gm. in the entire fraction. The yield of phenylalanine was 5.9 gm. The phenylalanine, liberated by ammonia and crystallized from hot water, decomposed at 278°C. and gave the following analytical figures: 0.1158 gm. of substance neutralized 7.0 cc. of 0.1 N H2SOx.. Calculated for CsHu02N Found. per cent per cent SS | SSS ee 8.49 8.47 The hydrochloride, which melted at 183°C., was also analyzed. 0.1023 gm. of substance contained 0.0177 gm. of chlorine. Calculated for CsH102N.HCl. Found. per cent per cent oS I 17.59 17.31 The aspartic acid gave the following figures: 0.1079 gm. of substance gave 9.5 cc. of moist nitrogen at 5°C. and 763 mm. Caleulated for CsH70OwN. Found. per cent per cent (ae eg OE Se ee ee eee 10.53 10.76 THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI, NO. | 160 Cleavage Products of Crystalline Lens Estimation of Tyrosine. 146 gm. of fresh lens were boiled with dilute sulfuric acid under a reflux condenser and the sulfuric acid was then quantitatively removed with barium hydroxide. The precipitated barium sul- fate was repeatedly boiled with water until the Millon reaction was negative and the united filtrates were concentrated until crystallization began. After standing in the cold the tyrosine was filtered off and the concentration and crystallization were repeated until the filtrate was free from tyrosine. The fractions were then united and crystallized from hot water after treatment with bone- black. The yield of purified tyrosine was 2.3 gm. It melted at 293°C. 0.1141 gm. of substance neutralized 6 ee. of 0.1 N H2SO.. Calculated for CsHuO3N. Found. per cent per cent 1 AE ee al acre aa Gorctors hin cic 7.74 7.37 A solution of the amino-acid in hydrochloric acid was levoro- tatory. Estimation of Glutamic Acid. The glutamic acid was isolated as hydrochloride. For this pur- pose, 323 gm. of fresh lens were hydrolyzed by boiling for 8 hours with 4 volumes of concentrated hydrochloric acid. The solution was then bone-blacked, concentrated under diminished pressure, and saturated with hydrogen chloride. After seeding it was al- lowed to stand at 0°C. When the crystallization was complete, the glutamic acid was filtered off, and the concentration and satu- ration repeated as long as a further crop separated. The fractions were then united and, after treatment with animal charcoal, reprecipitated by hydrochloric acid, repeating, as before as long as more crystals could be obtained. The total yield of hydrochloride was 21.6 gm. It melted at 193°C. and gave the fol- lowing figures on analysis: 0.1108 gm. of substance neutralized 5.8 ce. of 0.1 N H.SO.. 0.1121 gm. of substance contained 0.0215 gm. of chlorine. Calculated for CsHsOsN.HCI. Found. per cent per cent NIN Sb sroce biaps ate aratena tee date sore atoha ro) oiatce ches to aye Roe eee 7.63 7.39 LS Pr OA ee fe SE ae A bec 19.31 19.18 Y. Hijikata 161 Detection of Tryptophane. 90 gm. of fresh lens were incubated for 10 days in 300 cc. of water, saturated with chloroform, to which sodium carbonate and pancreatin (Merck) had been added. The tryptophane reaction was then positive. Estimation of the Hexone and Purine Bases. For this purpose, 1,025 gm. of fresh lens were heated under a reflux condenser on a Babo hot air funnel, according to Kossel and Kutscher,? with 3 parts by weight of sulfuric acid and 6 vol- umes of water until the Biuret test was negative. The solution was then freed from sulfuric acid by means of barium hydroxide and the filtrate and wash waters were united and concentrated. 5 per cent of sulfuric acid was then introduced and the hexone and purine bases were precipitated as tungstates. The tungstates were decomposed by barium hydroxide and the barium was re- moved as carbonate. The purine bases were precipitated from the resulting solution, and made weakly acid with dilute nitric acid by adding an excess of silver nitrate. Purine Bases.—The silver precipitate was suspended in ammo- nium hydroxide and allowed to stand for 24 hours. The insoluble fraction was then decomposed by means of dilute hydrochloric acid, filtered, and evaporated to dryness. Only a very small amount of material, not enough for identification, was obtained when the residue, dissolved in water, was treated with an excess of ammonia. The filtrate was freed from ammonia, weakly acidified with di- lute hydrochloric acid, and treated with a saturated solution of sodium picrate. A crystalline precipitate was produced, which, after recrystallization from hot water, formed needles. The ma- _ terial melted at 279°C., the melting point of adenine picrate. The - yield was too small for further identification. The filtrate from the adenine picrate, after removal of the picric acid, gave no xanthine or hypoxanthine on treatment with ammon- iacal silver solution. Hexone Bases.—From the filtrate from the purine silver salts, the arginine and histidine were precipitated by silver nitrate and * Kossel, A., and Kutscher, F., Z. physiol. Chem., 1900, xxxi, 165. 162 Cleavage Products of Crystalline Lens barium hydroxide and filtered off. The histidine, after removing the silver and barium, was reprecipitated by mercuric chloride in the presence of carbonic acid. The mercury was then removed and the histidine again passed through the silver-barium combination. After removal of the silver and barium, the solution was concentrated. One-fifth was converted into the picrolonate and the remainder into the hydro- chloride. Owing to an accident the yield of picrolonate could not be determined. The hydrochloride recovered weighed 6.7 gm., indicating a total yield of 8.4 gm. It melted at 231°C. and gave the following analytical figures: 0.0945 gm. of substance contained 0.0291 gm. of chlorine. Calculated for CsH902N3.HCl. Found. per cent per cent GIS. OE Po SAA eee eee 31.14 30.79 The picrolonate blackened at 255° and melted at 265°C. 0.0652 gm. of substance gave 12.7 cc. of moist nitrogen at 12.5°C. and 759 mm. Calculated for CeH 02N3.(CioHsOsN4)2. Found. per cent per cent IN. 2s os eh 504 oh want 6 4.eie sO ae omens 22.56 22.92 The filtrate from the histidine precipitate was completely freed from mercuric chloride by treatment with hydrogen sulfide and silver carbonate. The silver was then removed by hydrogen sul- fide and the arginine precipitated as picrate. The yield was 27.5 gm. It melted at 205°C. 0.0712 gm. of substance gave 15.4 cc. of moist nitrogen at 22.5°C. and 760 mm. Calculated for CcsH1402Ns.CeH307Ns. Found. per cent per cent IN eoihicigd tian Seoic aR aa son cuits oer aneenele 24.32 24.34 Silver and barium were removed from the filtrate from the his- tidine and arginine, and the solution was treated with phospho- tungstic acid. The precipitates, after decomposition by barium Y. Hijikata 163 hydroxide and removal of the barium, were converted into hydro- chlorides and the solution was evaporated to dryness. The choline in the residue was extracted with alcohol and precipitated from its alcoholic solution by mercuric chloride. The mercury compound was decomposed by treatment, in water, with hydrogen sulfide and the choline, after evaporation to dryness, dissolved in absolute alcohol, and precipitated with alcoholic chloroplatinic acid. Although an abundant flaky precipitate was formed, no perfectly pure choline chloroplatinate could be isolated. One-fifth of the hydrochlorides, insoluble in alcohol, was con- verted into the picrate, and the mother liquor, after removal of picrie acid, precipitated by phosphotungstic acid. The filtrate was then worked up for lysine picrate. The lysine was also recovered from the alcoholic filtrate from the mercury salt of choline. The solution was treated with mercuric chloride and barium hydroxide, and the lysine, after removal of mercury and barium, isolated as picrate. The mother liquors were again worked up in the manner described above. The yield of lysine picrate was 5.8 gm. indicating 14.5 gm. were present. The picrate melted at 252°C. and gave the following figures on anaylsis: 0.1036 gm. of substance gave 17.2 cc. of moist nitrogen at 25.5°C. and 761 mm. Calculated for CeH1s02Ne2.CsH307Ns. Found. per cent per cent (a AOD ES A OIA 18.67 18.40 SUMMARY. 1. The hydrolysis of the lens of the ox yielded thirteen cleavage products: alanine, valine, leucine, aspartic acid, glutamic acid, lysine, arginine, phenylalanine, tyrosine, proline, tryptophane, his- tidine, and adenine. 2. Calculated to ash- and water-free substance the following quantities of amino-acids were found: 164 Cleavage Products of Crystalline Lens TABLE I. Amino-acid. Amount found. per cent PA AAT EE gous corsa yaiorho net ob lerersrenes ee era eae ee 4.7 NERA TN ys mo. seh cas vescssctray Saget Sis a © ate egetanec 810 eke eRe ae 1.0 WSUGCINE) 0 oc odo See cye es on Meine oe Tee eee 6.8 ABPATHIC ACIG.. 0.2.2 S ok see ee cas boa et a oes Cee 1.4 Glutamice acid): sek A, i ee eee 15.5 LYSINE os, i oO > cE a 5 = ze 6 S 3 Aa Sul ta = 3 re ae ee: s | 8 a] 6 e |Als |e le Dog 24. | ce. | om | om | | Bre | Bae | Ba | June 6) 13*| 0.12 | 0.75 6.25 0.25710.25| After 2 hrs. ether anes- | | thesia. Dog 25 June 16 | | 0.08 | Blood taken just before anesthesia. June 16 | 42*| 0.39 | 2.87 '7.36) 0.147 0.12 Blood taken after 2 hrs. | ether anesthesia. Dog 26. Weight, 8.86 kilos. June 13 | flit “© 14 [570 | 5.47 12.90 12.35) Three injections of adren- | alin (0.04 mg. per kilo) Pra at 3 hour intervals. “ 15 |625t/ 5.941/16 .067/2.70) 0.11) Four injections of adren- | alin, as above. 1G 70 oh IEC oU 2 16 0.09 “ 17 |565 | 6.83 {11.94 |1.75) 0.09) Etherized, 10 to 12 a.m.; 4 to5 p.m. “ 18 |280¢| 4.571) 8.521/1.86) 0.107 Carotid blood taken just | ] before anesthesia. jams 35 EY | '0.1270.12 After 2 hrs. ether anes- | | | thesia. * Volume of urine secreted during the 2 hour anesthesia period. 7 Carotid artery. ¢ 24 hour volume incomplete. Ys Pui Nash: Jr. TABLE I—Continued. Urine. s| § Date. B S 5 sS s = > rs) n S|) 3 | 3/2 a = ~ A Dog 27. ce gm. gm June 28 29 |195 | 2.35 |13.54 |5.76 “30 (188 | 5.73 |17.24 |3.01 “ 30 | 34* 0.47 | 3.84 8.17 Dog 29. | 500 | 9.40 |36.24 aa Om or “7 1625 {14.48 |41.68 |2.87 “8 |555 {13.43 |39.52 |2.94 “8 | 16*,| 0.23 | 0.86 a. Dog 30. Oct. 10 “11 [240 | 6.95 [24.28 |3.49 12 |320 | 8.79 |29.35 |3.34 “ 13 {310 | 8.78 |28.24 |3.21 “14 |300 | 7.08 |20.64 |2.91 “14 | 34* 0.39 | 1.77 |4.54 Blood sugar. Femoral artery. Jugular vein. Renal vein. 177 Remarks. Weight, 6.18 kilos. per per per cent cent | cent 0.14 “Fi 0.21 0.17 Fed 61 gm. lean, cooked meat. Fed 100 gm. lean, cooked meat. 100 cc. water by stomach tube at 1 p.m. Blood taken after anesthesia from 2 to 4 p.m. Weight, 9.70 kilos. 0.13! | a 26 “ae 26 | | Fed 100 gm. lean, cooked ee <4 if4 “ “ 100 cc. water by stomach tube at 1 p.m. Blood taken after ether anes- thesia from 1.50 to 3.50 p.m. Weight, 6.80 kilos. 0.13 0.08 0.08 0.08 0.07 Fed 100 gm. lean, cooked meat. Fed 50 gm. lean, cooked meat (all animal would eat). “ “ “ “ “ “ce Animal refused meat. 400 cc. water by stomach tube at 1 p.m. Blood taken after ether anes- thesia from 2.40 to 4.40 p.m. TABLE I—Concluded. Urine. Date. 5 2 S _ i en So 3 =] > ‘= n | 3/3 /2 a a = 2) Dog 31. ce. gm. gm. Oct. 24 “ 25 |815 | 8.78 |27.00 |3.07 26 |940 |10.45 |30.19 |2.89 27 \850 | 9.84 |30.48 |3.09 28 |810 | 8.37 |27.44 |3.27 28 | 34*| 0.43 | 3.30 7.67 Nov. 1 |335 | 8.14 |30.74 (14). 2 455 |10.00 |32.96 3 |640 |12.96 |41.64 4 1785 |12.23 |40.46 |: 4| 60*| 0.79 | 6.24 Blood sugar. > . ~ Ab g 2 a 5 = = ) = 2 5 > & Renal vein. Kidney in Phlorhizin Diabetes Remarks. Weight, 8.78 kilos. per cent 0.13 0.10 0.09 per cent per cent Fed 150 gm. lean, cooked meat. “ “ “ “ “c “ Fed 110 gm. lean, cooked meat (all animal would eat). Fed 150 gm. lean, cooked meat. 0.13 |0.12) 500 cc. water by stomach tube at1.25p.m. Blood taken after ether anes- thesia from 2.45 to 4.45 p.m. Weight, 10.02 kilos. 0.11 Fed 150 gm. lean, cooked ~ meat. Fed 65 gm. lean, cooked meat (all animal would eat). Fed 150 gm. lean, cooked meat. “ “ “ “ce “cc “ 0.17 |0.14| 300 cc. water by stomach tube at1.05p.m. Blood taken after ether anes- thesia from 2.30 to 4.30 p.m. Dog 26, as shown in Table I, received on the 2nd and 3rd days of the experimental period subcutaneous injections of adrenalin Again, on the day preceding the conclusion of the experi- fee Wash, or: 179 ment, the dog was subjected to two periods of anesthesia. This regimen was designed to deglycogenize the animal and prevent hyperglycemia during the final anesthesia, but led to such impair- ment of renal function that it was not employed in the later dogs. On 2 days this dog voided outside of the cage; hence the 24 hour urines corresponding to these days were incomplete. The last five dogs of the series received warm water by stomach tube a short time before they were anesthetized. With Dogs 31 and 32 the urine secreted during the 2 hour period of anesthesia was measured at 30 minute intervals, in order to determine whether the effect of the anesthetic was con- stant or progressive. ‘The following results were found: Dog 31. Dog 32. cc ce. PRIM IESESITINUGH 0 2 ke ee ee ceeds 11 13 Second 30 eee ee ae 6 14 Third 30 Sener PO ee SEE ret 1s eb eloctats 8 18 Fourth 30 SERRE rcs int aca cig cial sia atsu 3's st 9 15 DISCUSSION. Inspection of Table I reveals in the last six cases a typical picture of phlorhizin glycuresis, accompanied by a definite hypo- glycemia. In several instances following the injection of phlo- rhizin the blood sugar has fallen until it is little more than one-half the normal value. Following ether anesthesia we have observed in every case except Dog 30 an increase in blood sugar. This observation is in general accord with that of other workers (15). The increase in the case of Dog 29 was more than 200 per cent above the value just prior to anesthesia, and in general was above the normal value. The degree of hyperglycemia is determined both by kidney activity and the apparent glycogen reserve of the animal. Thus Dog 26, which had received a series of injections of adrenalin during the fore period, secreted only 2 cc. of urine under anes- thesia, yet the percentage of blood sugar increased only from 0.10 to 0.12. Dog 27 secreted 34 cc. of urine containing 3.8 gm. of sugar during the ether period, yet the blood sugar increased 180 Kidney in Phlorhizin Diabetes from 0.10 to 0.21 per cent; in this case the high D:N ratio of 8.17 is indicative of an unusually high glycogen reserve. In striking contrast is Dog 30, which also secreted 34 ce. of urine during anesthesia, but only 1.7 gm. of sugar. Here the blood sugar showed no change, and the relatively low D:N ratio of 4.54 bears out the probability of a dep!eted glycogen reserve prior to anesthesia. With regard to this point we find in our notes, under date of October 12, a memorandum that the weather had suddenly turned cold, and the dog, lying in a metal cage in an unheated room, was constantly shivering. It is generally accepted that shivering will deglycogenize a phlorhizinized dog. A comparison of the sugar concentration in arterial and renal venous blood shows that in no case have we found more sugar in the renal blood. On the contrary, in five of the eight dogs we have found appreciable differences in favor of the arterial blood sugar. It will be noted, furthermore, that in the three cases where we found no change in the blood sugar content the amounts of sugar excreted during the anesthesia period were the smallest obtained in any of the experiments (less than 1 gm.). The greatest observed loss in blood sugar in passing through the kidney was 0.04 per cent, Dog 27, and in this case the sugar excreted during the 2 hours was 3.84 gm. The average excess of sugar in the arterial blood of the five positive cases was 0.022 per cent. It is significant that there is a direct proportion between the sugar excreted in the urine and the loss from the blood. Whether the loss from the blood is sufficient to account for the sugar recovered in the urine is a question which obviously involves other and uncertain factors. Not only must the average volume flow of blood through the kidney be known, but it must be assumed that this average rate obtained at the moment of drawing the renal blood. If the rate of blood flow through the kidneys is 150 cc. per minute per 100 gm. of kidney tissue (16), and the average weight of kidney tissue in the dogs of our experiment was 60 gm., the average volume of blood passing through the two kidneys during the anesthesia period was 10,800 cc. With an average loss in arterial blood sugar of 0.022 per cent, we could thus account for 2.38 gm. of excreted sugar. The total sugar excreted during the anesthesia period by the five dogs upon which Ps Nask.. Jr. 181 the calculation is based was 18.02 gm., an average of 3.6 gm. per dog. In other words the amount calculated is 66 per cent of the amount found. In the case of Dog 27, the actual sugar excreted was 3.84 gm., as compared with a calculated value of 4.32g¢m. We believe, in view of the several indeterminate factors involved, that the correspondence between actual and calculated excreted sugar is close enough to warrant the conclusion that the phlorhizinized kidney is not concerned in a function of specific sugar production. CONCLUSION. In phlorhizinized dogs the renal venous blood shows a lower concentration of sugar than the general arterial blood, provided the kidney function has not been abolished or seriously impaired. Failure to control the effect upon the kidney of general anesthe- tics, and inherent errors in the methods employed for blood collec- tion and sugar estimation, account for contrary results previously reported. It appears that the kidneys, under the influence of phlorhizin, do not acquire a specific sugar-producing function. Our results confirm an increased permeability of the renal epithelium. BIBLIOGRAPHY. 1. von Mering, I., Verhandl. Kong. inn. Med., 1886, v, 185; Z. klin. Med., 1888, xiv, 405; 1889, xvi, 431. 2. Levene, P. A., J. Physiol., 1894-95, xvii, 259. 3. Zuntz, N., Sex: Physiol., 1895, 570. 4. Biedl, A., ee Kolisch, R., Verhandl. Kong. inn. Med., 1900, xviii, 573. 5. bavy, F'. W., Brodie, T. G., and Siau, R. L., J. Physiol, 1903, xxix, 467. 6. Lépine, R., and Boulud, Ganiue. rend. Acad., 1904, exxxix, 497. 7. Lépine, R., Compt. rend. Soc. biol., 1910, lxviii, 448. 8. Erlandsen, A., Biochem. Z., 1910, xxili, 329. 9. Coolen, F., Arch. Pharm., 1895, i, 267. 10. Underhill, F. P., J. Biol. Chem., 1912-13, xiii, 15. 11. Reilly, F. H., Nolan, F. W., and Lusk, G., Am. J. Physiol., 1898, i, 395. 12. Epstein, A. A., Reiss, J., and Branower, J., J. Biol. Chem., 1916, xxvi, 25. MacNider, W. deB., J. Med., Research, 1913, xxviii, 403; J. Pharm. and Exp. Therap., 1921, xvii, 289. And many others. 13. Benedict, S. R., J. Biol. Chem., 1918, xxxiv, 203. 14. Ringer, A. I., J. Exp. Med., 1910, xii, 105. 15. Epstein, A. A., and Aschner, P. W., J. Biol. Chem., 1916, xxv, 151. Ross, E. L., and McGuigan, H., J. Biol. Chem., 1915, xxii, 407. San- sum, W. D., and Woodyatt, R. T., J. Biol. Chem., 1915, xx, p. xxix; xxi, 1. And many others. 16. Burton-Opitz, R., and Edwards, D. J., Am. J. Physiol., 1917, xliii, 408. Gs ro Me ipeey ¢.4y i? ~ ©) (prey i io MRD eye fT : 7 ie tt res | ll ium La ‘) re os yif (ties ¥ i/ a6 ‘ 4 #)a) = . tie 2 » (ASO Sepa) or ee me! Fl SING ne wees Shay | NOTE ON THE AMMONIA CONTENT OF BLOOD. By THOMAS P. NASH, Jr., anp STANLEY R. BENEDICT. (From the Department of Chemistry, Cornell University Medical College, New York City.) (Received for publication, January 4, 1922.) In a personal communication, Dr. T. Addis, of Stanford Uni- versity Medical School, has suggested that the increased ammonia which we find (1) in blood of the renal vein may be due to return of previously excreted ammonia from a kidney which has ceased to function normally under the experimental conditions necessary in drawing the blood. Addis and Shevky (2) have reported an increase of urea in renal venous blood of rabbits at a time when no urine was being secreted. The suggestion made by Dr. Addis seemed to us of sufficient interest to warrant a presentation of some considerations which we believe conclusively show that the increased ammonia content of the blood of the renal vein cannot be explained upon such a basis. Addis and Shevky found more urea in the blood of the renal vein only when urine secretion had ceased after clamping and ligating the renal vessels of the opposite side. When blood was taken quickly the renal venous blood showed, usually, a loss of urea. In our animals, except those whose urine secretion was controlled over a definite period (see below), the renal blood was taken as quickly as anesthesia could be effected and the operation performed; the kidney on only one side was exposed, and this disturbed as little as possible; and blood was taken from the vein without even a temporary stoppage of the renal circulation. We have found, without exception, more ammonia in the renal vein than in other blood, and we obtained this finding regardless of whether the blood was taken early or late in anesthesia. Such results are not to be expected if one is dealing with an effect of the - 183 184 Ammonia Content of Blood anesthesia upon urine secretion. That we were not dealing with such an effect is definitely proved by the following considerations: In those animals from which blood was taken late in anesthesia the urine secretion was controlled. Thus, we reported a dog which received intravenously during 2 hours of ether anesthesia 196 cc. of 2 per cent sodium bicarbonate solution. At the end of this period (during which a total of 81 cc. of urine was taken by catheterization at 30 minute intervals), the renal venous blood now contained 0.18 mg., and the femoral arterial blood 0.07 mg., of ammonia nitrogen per 100 ce. Dogs 24, 25, and 27 were diabetic from phlorhizin injection. From each of these dogs blood was taken after a 2 hour ether anesthesia during which period the urine was collected by catheter, TABLE I. Urine gemne 2 he s. ether Blood at end of 2 hrs. ether anesthesia. Dog No. Carotid. Renal vein = = Total Volume. | Total N. sugar. | NH:-N NHs-N per 100 ce: Sugar per 100 ee. Sugar ce. gm. gm. mq. per cent mg. per cent 24 13 O12 0.75 0.14 0.25 0.25 0.25 75) 42 0.39 2.87 0.13 0.14 0.18 0.12 27 34 0.47 3.84 0.05 0.21 0.14 0.17 its volume noted, and its sugar content determined. The same sample of each -blood was analyzed for ammonia and sugar. The blood sugar values are included in a series recently reported (3); they were not given in our original paper dealing with the ammonia content of the blood because they were intended as a separate investigation, and in only the three dogs did our problems overlap. The comparative data for these dogs are summarized in Table I. It will be seen that Dog 24 shows no change in the sugar con- centration of the renal blood, but a marked increase in the ammonia content. The other two dogs (which were secreting urine more freely) show a definite decrease in the sugar content of the blood after traversing the kidney, while the ammonia content of the same blood sample is increased (in one case to T. P. Nash, Jr., and 8. R. Benedict 185 nearly 300 per cent of its former value). These facts further substantiate our conclusion! that ammonia is formed by the kidney. BIBLIOGRAPHY. 1. Nash, T. P., Jr., and Benedict, S. R., J. Biol. Chem., 1921, xlviii, 463. 2. Addis, T., and Shevky, A. E., Am. J. Physiol., 1917, xliii, 363. Geengen. t. Po, Jr., J. Biol. Chem., 1922, li, 171. THE DETERMINATION OF URIC ACID IN BLOOD. By STANLEY R. BENEDICT. (From the Department of Chemistry, Cornell University Medical College, New York City.) (Received for publication, January 17, 1922. The determination of uric acid has been usually regarded as the least accurate and the most tedious and exacting of any of the analyses commonly made upon the blood. Certainly the heat coagulation method followed by concentration and precipitation with silver magnesia mixture as proposed by Folin and Denis (1) and modified by the present writer (2) was a laborious process, and required relatively large amounts of blood and careful ana- lytical work for satisfactory results. Folinand Wu (3) haverecently suggested a very interesting procedure which provides for direct precipitation of uric acid from highly diluted blood filtrates. After decomposing the precipitate with sodium chloride in hydro- chloric acid solution the uric acid is determined practically as formerly, but instead of using 10 to 20 cc. of blood for a determina- tion, as in the older process, Folin and Wu made use of an equiva- lent of only 2 cc. of blood for the development of color in their method. The depth of color thus obtained in most bloods is exceedingly weak. To the mind of the present writer it is questionable whether the great advance made by Folin and Wu in precipitating the uric acid directly from the dilute blood filtrate is not more than offset by having the final solutions so weak in color that it is questionable whether they can be com- pared accurately in a colorimeter by most analysts. The observation of the writer several years ago that cyanide caused an increase of the color given by uric acid and phospho- tungstic acid in alkaline solution offered the possibility of greatly increasing the color obtainable from a given quantity of uric acid. With a large excess of cyanide (several cc. of a 5 per cent solution of potassium or sodium cyanide) the reaction was found to become 187 188 Uric Acid Determination exceedingly delicate, but advantage could not be taken of this fact for analytical purposes for two reasons. With the large excess of cyanide the color reaction was found not to be closely enough proportional to the quantity of uric acid present. Furthermore, turbidity always developed in such solutions, which was a very objectionable feature in analytical work, as the precipitate would continue to form slowly over a considerable period of time. The results obtained in this connection did, however, lead us to keep in mind the possibility of developing a more satisfactory technique for uric acid determination, based upon the use of large quanti- ties of cyanide. It seemed possible that the reaction might be thus made more specific for uric acid, so that the determination could be made directly upon the blood filtrate, without preliminary precipitation of the uric acid. Experiments along this line have been carried on from time to time for the past few years, but the results obtained have hitherto been unsatisfactory for one reason or another. Last September experiments were begun in which the possible application of arsenic tungstic acid in uric acid de- termination was studied. Results obtained in this connection have proved very satisfactory, and the purpose of the present paper is to describe new uric acid reagents based upon the use of arsenic and phosphoarsenic tungstic acids, and to present a method for uric acid determination in blood which can be applied directly to small quantities of the Folin-Wu blood filtrate. Arsenic tungstic acids have been described by Kehrmann (4), Fremery (5), and Gibbs (6). In the present work it was soon found that boiling sodium tungstate and arsenic acid together in solution would give a reagent which could be used for uric acid determination, but the condensation of the two acids under such conditions was very slow. Long boiling was required, and the solutions were apt to become quite strongly colored during the continued heating. We, therefore, added hydrochloric acid as a condensing agent. Using this acid, it was found that arsenic and tungstic acids will react very promptly in hot solution, so that the reaction is complete within 20 minutes or less, and the solution colors even less than do the phosphotungstic acids. Our very first experiments showed that reagents prepared in this way are definitely superior to the older phosphotungstic acid reagents, even when used in exactly the same way. Under such S. R. Benedict 189 conditions they develop about 20 per cent more color than do the phosphotungstie acid reagents, and are very much less apt to yield turbidity. It was found using 10 gm. of sodium tung- state, and condensing the tungstic acid derived from this by means of hydrochloric acid with 1, 2, 3, 4, 5, 6, 7, 8, 9, and 10 gm. of arsenic pentoxide and making to a final volume of 100 cc., that any quantity of the arsenic pentoxide above about 1.8 gm., yields a reagent with a color-yielding power equal to the solution obtained when any of the larger quantities of arsenic acid are used. We shall describe the arsenic tungstic acid reagent which we first employed, though we no longer use it. It may be of service to those who wish to continue to use the old precipitation method for uric acid determination, since it is more easily prepared than is the old phosphoric acid reagent, and is remarkably free from any tendency to formation of turbidity. This reagent is pre- pared as follows: 100 gm. of sodium tungstate (J. T. Baker’s c. P. is satisfactory, or the ‘‘Primos”’ brand may be used if ob- tainable) and 30 gm. of pure arsenic acid (As:O;) are placed in a liter flask and about 700 cc. of water added. 50 cc. of concen- trated hydrochloric acid are then added and the mixture is boiled for about 20 minutes. It is then cooled and diluted to 1 liter. This reagent may be employed exactly as is the phosphotung- stic acid reagent of Folin and Denis, and is to be recommended in place of that reagent to anyone who may prefer to continue using the old method. We found, however, that with the new arsenic tungstic acid reagent it is possible to use an excess of sodium cyanide solution, and to dispense with the use of sodium carbonate. Under such conditions a much deeper color develops for a given quantity of uric acid than can be obtained under the old condition. The color obtainable with the new reagent under the new con- dition was so great that it seemed probable that the new reaction would be much more specific for uric acid than was the old. Hence we were led to try the reaction directly upon blood filtrates and also compared the relative color given by a typical ‘‘inter- fering substance” in the new process. The results obtained here fell just short of our hopes. Resorcinol, used in minute quanti- ties, would give about one-fourth to one-third as much color as 190 Uric Acid Determination an equal weight of uric acid,’ and when the reaction was applied directly to blood filtrates, a considerable proportion of bloods yielded distinctly high figures by the direct method, while others gave figures which agreed closely with those obtained by the pre- cipitation method of Folin and Wu employed as described later in this paper. As the results obtained in this connection so nearly fulfilled our expectations we were led to investigate the possibilities further before abandoning the hope of obtaining a reagent which could be applied directly to blood filtrates. We therefore tried, among other things, combining tungstic acid with both arsenic and phosphoric acids. Curiously enough a reagent was thus obtained which is certainly more specific for uric acid than either arsenic tungstic or phosphoric tungstic acid alone, and as we shall see below, we believe that the new reagent fulfills reasonable demands for direct determination of uric acid on the Folin-Wu filtrates from human blood. The reagent used in the method described below is prepared as follows: 100 gm. of pure sodium tungstate are placed in a liter flask and dissolved in about 600 cc. of water. 50 gm. of pure arsenic pentoxide are now added, followed by 25 ec. of 85 per cent phosphoric acid and 20 cc. of concentrated hydrochloric acid. The mixture is boiled for 20 minutes, cooled, and diluted to 1 liter. The reagent appears to keep indefinitely. When used as described below this arsenic phosphotungstic acid reagent yields nearly seven times as much color from a given weight of uric acid as does the old phosphoric acid reagent as formerly employed. The new reagent is scarcely affected by a typical polyphenol such as resorcinol in the presence of uric acid. If five times as much resorcinol as uric acid be present in a solu- tion, the quantity of color obtained is increased by only about 15 per cent, and there is no influence on the shade or quality of color produced. With fifty times as much resorcinol as uric acid, there is about 66 per cent increase in the color obtained. It is 1 It is not correct to generalize regarding the amount of color given by resorcinol and uric acid with the old (or new) reagent, since under the conditions employed for uric acid determinations the color given by resorcinol is only very roughly proportional to the quantity of the sub- stance present. S. R. Benedict 191 interesting to note that if no uric acid be present, resorcinol yields considerably more color than when even very minute amounts of uric acid are contained in the solution. This is probably best explained by assuming that the uric acid reacts first with the color-yielding substance in the reagent. Since the substances in blood, other than uric acid which give color with the old reagent are unknown (though it is commonly assumed that they are probably polyphenols), it is obvious that the true specificity of a reagent for uric acid determination in blood can be determined only by comparing the results obtained when the reaction is applied directly to blood filtrates with those obtained by the use of some accepted precipitation method. We have made such comparison in some detail for human blood, and the findings will be discussed below. The technique of the new method for uric acid determination in human blood is as follows. Standard Solutions—The ‘color obtained in the new method from a given quantity of uric acid is so intense that the standard solutions employed have a concentration of uric acid considerably below the solubility of uric acid in pure water. For this reason we are able to employ as standard, solutions of uric acid which are strongly acid with hydrochloric acid. We shall report later upon the keeping quality of these solutions. At present we prepare them fresh once in 2 weeks by appropriate dilution of the phosphate standard solution described by Benedict’ and Hitchcock in a previous paper. Unless kept in an excessively warm room, the phosphate standard may be relied upon to keep about 2 months. It is desirable to keep on hand two standard solutions, one of which contains 0.01 mg. of uric acid per cc., while the second contains 0.02 mg. of uric acid in 5 cc. of solution. The second standard is the one commonly employed, but the first may oc- casionally be of service, and is valuable in instances where it is desired to use the Folin-Wu procedure for comparison of results by the old and ‘new procedures. For the preparation of the first standard, 25 cc. of the phosphate standard solution (containing 5 mg. of uric acid) are measured into a 500 ec. volumetric flask, and the flask is about half filled with distilled water. 25 cc. of dilute hydrochloric acid (1 volume of concentrated acid diluted to 10 volumes with water) are added, and the solution is diluted THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI, NO. 1 192 Uric Acid Determination to 500 ec. This solution contains 0.01 mg. of uric acid in 1 ce. For preparation of the second standard (the one which is most frequently employed) the procedure is the same, except that instead of starting with 25 ce. of the phosphate solution, 10 ce. are employed and diluted after acidification exactly as for the other standard. It should be remembered that these standard solutions should be freshly prepared once in 2 weeks. In connection with the use of the standard solution it should be pointed out that in the new method, as with most colorimetric methods, results are most accurate when standard and unknown correspond closely in depth of color. Using 0.02 mg. of uric acid as standard, results are satisfactorily exact when the unknown reads between 10 and 24 mm., when the standard is set at 15 mm. This represents a quite satisfactory range for bloods, as one stand- ard is applicable for bloods containing from about 2.5 to 6 mg. of uric acid per 100 cc. With 2 mg. of uric acid results may be about 10 per cent too high: 7.e., we might obtain 2.1 or 2.2 mg. of uric acid per 100 cc. instead of the true value of 2 mg. Absolute results, where desired, can, of course, be obtained for bloods on the outer limit of any standard by a repetition of the determination, using a closer standard. Increased accuracy cannot be obtained by chang- ing the dilution of either the standard or unknown after the reaction is completed. The determination is so simple, and the quantity of material required so small, that it is no hardship to repeat deter- minations on the few bloods which fall outside the limits of 2 to 6 mg. per 100 cc. In repeating the determination on such bloods it is desirable to correct by using more or less of the blood filtrate, as indicated by the determination, with corresponding change in addition of water to make a total volume of 10 cc. before addi- tion of cyanide and reagent as described below. It is safer in such cases to repeat with 5 ce. of filtrate, using 0.01 mg. of uric acid in the standard. Where more than 5 ee. of blood filtrate are employed the reading should be made promptly, in order to avoid the development of turbidity. It should be stated that the method described below yields a slight trace of color in a blank determination. To those who have been in the habit of reading the Folin-Wu uric acid solution this color may appear a serious matter. In the present method, where the color developed is relatively so much more intense, the S. R. Benedict 193 color from a blank is negligible, inside the limits above mentioned, since it is present in both standard and unknown. The blood is precipitated with tungstic acid as described by Folin and Wu (3), the blood being allowed to stand at least 10 to 20 minutes after adding the tungstate and sulfuric acid, before filtration. This tends to insure complete protein precipitation. The use of excess of acid in the precipitation is to be avoided. 5 ce. of the water-clear filtrate (representing 0.5 cc. of blood) are transferred to a test-tube? and 5 cc. of water are added. The standard solution, containing 0.02 mg. of uric acid (prepared as described above), is placed in another tube and the volume like- wise made up to 10 cc. To both standard and unknown are added 4 cc. of 5 per cent sodium cyanide solution’ containing 2 ec. of concentrated ammonia per liter. To each tube is then added 1 cc. of the arsenic phosphoric acid tungstic acid reagent. The contents of each tube should be mixed by one inversion immediately after addition of the reagent, and placed immediately in boiling water, where the tubes should be left for 3 minutes after immersion of the last tube, but the time elapsing between immersion of the first and last tubes should not exceed 1 minute. There will be no difficulty here in getting the tubes all into the hot water within 1 minute, unless it is attempted to run more than about five unknown solutions in one series. A 3 minute sand-glass is very convenient in connection with the heating. After the 3 to 4 minute heating the tubes are removed and placed in a large beaker of cold water for 3 minutes and read in a colorimeter against the standard as soon as may be convenient. Long standing before reading may lead to devel- opment of turbidity. It is best to read the solution within ? As will be seen from the description of the method the test-tubes used need not be graduated, since no dilution is ordinarily made prior to reading in the colorimeter. The test-tubes employed should be of uniform diameter (18 to 20 mm.). * On account of the high toxicity of cyanide solutions they should never be handled in pipettes, but should always be measured from a burette. The reagent is also best measured from a burette. 4Some cyanide solutions were found to yield more color in blank determinations than did others. Experiments on this point showed that it is desirable to have a trace of ammonia present in the cyanide solution. It is desirable to prepare the cyanide solution fresh once in 2 months. 194 Uric Acid Determination 5 minutes after removing from the cold water. Where this is done we have never encountered turbidity. Where a large number of bloods are to be analyzed it is best to run not more than four at a time with one standard. This provides for greater uniformity in handling, does not cool the bath down too much, and makes it easy to finish reading before any turbidity may develop.* A few words of explanation concerning the adoption of heating in connection with the uric acid determination may be desirable. We adopted the immersion in hot water for a few minutes after a very careful study had demonstrated that it is not possible to obtain conditions where the reaction goes to even approximate completion in less than 15 minutes at room temperature. Such standing may lead to development of turbidity, and we have found it much better to adopt the short period of heating, which is certainly no hardship. Prolonged heating is to be avoided, since the color will fade under such conditions. There is no difficulty whatever if the tubes are placed in water which is within 10° of boiling and left there for 3 to 4 minutes, and then immersed in cold water as described above. Calculation.—Employing the standard solution containing 0.02 mg. of uric acid and using 5 cc. of the 1:10 blood filtrate, the calculation for the uric acid content of the original blood is as follows: S R x 4 = mg. of uric acid per 100 ce. of original blood in which S represents the height of the standard solution in milli- meters, and FR the reading of the unknown solution. If instead of using 5 cc. of blood filtrate in the determination, 2.5 or 10 cc. are employed, the final figure is multiplied or divided by 2 accordingly. 5 Excess of potassium salts in blood filtrates is undesirable in both the uric acid and the creatinine determinations. It would probably be advan- tageous for these determinations to employ sodium oxalate instead of the potassium salt to prevent coagulation. There is no apparent advantage in the almost universal habit of using potassium oxalate as an anticoagulant for blood. S. R. Benedict 195 Discussion of the Results by the New Method. ‘ In Table I are given the results of determination of uric acid by the new method upon fifty samples of human blood. We have also included in the table figures for the uric acid on the same blood samples obtained by a slightly modified Folin-Wu proce- dure, together with figures for the non-protein nitrogen. Suff- cient blood was not available for comparison by the older silver magnesium precipitation method, but Folin and Wu have stated that their method essentially duplicates the results by the older procedure. In using the Folin-Wu procedure our technique has been identical with that described by these writers up through the decomposition of the precipitate by means of 10 per cent sodium chloride in 0.1 Nn hydrochloric acid. From this point our procedure differed from that described by Folin and Wu, because these writers advocated the use of a sulfite-containing standard, which has not appealed to us. The use of the sulfite standard, together with the dilution employed by Folin and Wu led to such weak final colors that we were unable to read them. We shall describe our exact technique from the point of precipi- tation of the uric acid, as it may prove useful to others in con- nection with checking up results by the new method upon any particular samples of blood. In using the Folin-Wu technique we have invariably employed 20 cc. of filtrate for the precipitation, except where the blood uric acid exceeded 6 mg. per 100 cc. In such cases 10 cc. of the filtrate were used. The measured volume of filtrate, contained in a centrifuge tube, is precipitated by means of the silver lactate- lactic acid solution’ as described by Folin and Wu. After cen- — 6 These bloods were obtained from the pathological laboratories of the Roosevelt Hospital through the courtesty of Dr. Wm. G. Lyle, Director. _ We are also indebted in this connection to Mr. Harry Osterberg, whose _ figures for non-protein nitrogen on these bloods we have included in the _ table. _ ‘Silver lactate purchased in the market is apt to contain considerable _ amounts of reduced silver, and the product is quite expensive, and often _ difficult to obtain. We have prepared our silver lactate-lactic acid solu- _ tion starting from silver nitrate, as follows: Dissolve 22.5 gm. of silver nitrate in about 500 cc. of water in a liter stoppered cylinder. Add an _ excess (about 60 cc.) of approximately 5 n sodium or potassium hydroxide 196 Uric Acid Determination trifugation the residue in the tube is very thoroughly stirred with the acid sodium chloride solution, using 1 cc. where 10 cc. of blood filtrate were precipitated, and 2 cc. where 20 cc. were used. A volume of water is now added to the centrifuge tube® which will bring the total volume of the solution to 10 cc. where the quantity of uric acid expected does not exceed 0.08 mg., or to 12 ce. where a quantity between 0.08 and 0.12 mg. is expected. (These figures were selected because they represent the maximal concentration obtainable where we employed the standard solu- tion containing 0.01 mg. of uric acid per ec.) The contents of the centrifuge tube are again thoroughly stirred, and after cen- trifugation are poured as completely as possible into a clean dry test-tube. In two other tubes quantities of the standard solution previously described (containing 0.01 mg. of uric acid per cc.) are measured to give a total uric acid content in one tube of 0.05 mg. and in the second of 0.08 mg. of uric acid. If high uric acid is expected it is desirable to have a third standard containing a total of 0.1 mg. of uric acid. 2 ee. of the acid sodium chloride solution are added to each of the tubes, and the volume made to 10 cc. for the weaker standards, and left at 12 ec. for the stronger standard. To each standard and unknown tube are now added 3 drops of 5 per cent sodium cyanide solution, and 1 ce. of the Folin- Denis uric acid reagent, followed by 2 ec. of 20 per cent sodium carbonate solution. The contents of the tubes are mixed by a single inversion, and read promptly in the colorimeter, setting the standard at a height of 20 mm. Should any of the unknown solution, then add water to about 1 liter and shake. Allow to stand for a few minutes and decant the supernatant fluid. Add distilled water to about 1 liter, shake, and pour off the supernatant fluid after a moment or two. Repeat this washing by decantation until the wash water fails to react alkaline to litmus paper. The total washing process need not take longer than 10 minutes. After the last decantation make up to a volume of about 200 ce. with distilled water, and add 35 ee. of lactic acid (sp. gr. 1.2). Shake thoroughly and dilute to 500 ec. Filter through a dry folded filter, returning the first portions until the filtrate appears as clear and colorless as distilled water. Filter the entire solution through this filter, and pre- serve in an amber, glass-stoppered bottle. 8 In the description it is assumed that either the 10 or 20 ce. of filtrate has been precipitated in one centrifuge tube. Where two tubes have been employed one-half the indicated quantity of chloride solution and of water should be added to each of the two tubes. 2 S. R. Benedict 197 tubes having a dilution of 10 cc. match the 0.1 mg. standard more closely than any of the weaker standards, this comparison may be made after adding 2 cc. of water to the unknown tube. In employing the Folin-Wu procedure it should be remembered - that unknown and standard solutions must match quite closely - for satisfactory results. We shall now discuss the comparative figures obtained by the two methods. For facilitating this comparison we have tabulated in Column 4 the difference by the two methods in milligrams of uric acid in 100 ce. of blood, using the Folin-Wu figures as stand- ard. Thus where the new method gives results higher than the old, figures in Column 4 are preceded by a plus, and where lower, by a minus sign. Even a casual inspection of Table I reveals that the new method tends to give higher results than does the Folin-Wu procedure, at least on those bloods which contain less than 50 mg. of non- protein nitrogen per 100 cc. Thus out of the total of fifty bloods, we find that in thirty-one, or 62 per cent, the new method yields higher figures than the other. This fact in itself could not, however, lead us to question the accuracy of the new method, Since in the new procedure we have eliminated precipitation, decomposition, and transfer of the final solution, which would theoretically tend to cause slight losses. For a proper considera- tion of the differences by the two methods it is, therefore, more correct to take account only of differences which exceed, let us say, 0.5 mg. per 100 cc. of blood, since duplicates by the same method will frequently vary by this amount, and plus differences up to the 0.5 figure can properly be credited in favor of the new method. Out of the total of fifty bloods we find that there are twenty, or 40 per cent, in which the new method exceeds the old by more than 0.5 mg. per 100 cc., eight (16 per cent) in which the figures lie between 1 and 1.5 mg., and none in which the new method exceeds the old by as much as 2 mg. of uric acid per 100 ec. of blood. We shall return later to a discussion of the higher figures by the new method. From a general standpoint it would appear that if there are no cases where the figures by the new method exceed those by the old by as much as 2 mg. per 100 cc., that very probably the ques- tion is a theoretical rather than a practical one. But a very 198 Uric Acid Determination TABLE I. Results Obtained by the New Method Compared with Those Obtained by a Slightly Modified Folin-Wu Procedure, on the Same Samples of Human Blood. Per 100 cc. of blood. Sample No. F Difference, : New method. aide he Folin-Wu figures i; beh 7) mg. mg. mg. mg it 4.4 5.2 —0.8 103 2 2.5 2.6 —0.1 33 3 3.6 2.5 Ved 38 4 3.2 3.0 +0.2 41 5 3.2 ae 60 6 3.4 5.0 —1.6 33 2G 6.6 6.6 69 : 4.0 3.2 +0.8 30 9 ie 4.8 +0.4 33 10 3.5 3.2 +0.3 30 11 13.6 15.7 —2.1 197 2.4 2.0 +0.4 28 13 4.0 2.4 +1.6 33 14 2.8 2.3 +0.5 29 2 2.6 1.8 +0.8 36 16 4.2 eal iy 204 _ 2.6 2.4 +0.2 40 18 4.0 4.4 —0.4 45 +) 2.4 2.3 +0.1 46 ” 4.1 3.4 +0.7 39 21 0.8 1.2 =14 27 ee 2.7 2.1 +0.6 41 23 2.8 2.0 +0.8 35 ee 3.0 2.0 +1.0 36 25 6.0 6.6 —0.6 105 26 5.6 6.0 —0.4 201 aa 3.7 3.4 +0.3 45 28 4.3 3.0 +1.3 41 2 4.2 4.0 +0.2 37 30 3.6 4,2 —0.6 43 eo 2.9 2.7 40.2 37 Fe 3.6 2.7 +0.9 38 eS 4.6 3.8 +0.8 49 34 B68 2.9 = Oed) 36 vs 3.8 2.9 +0.9 28 36 3.6 2.5 +1.1 44 S. R. Benedict 199 TABLE I—Concluded. Per 100 ec. of blood. Sample No. Sa ee Difference, x : ee) retin. | | Eee eon mg. mg. mg. mg 37 an 2.4 +0.3 28 38 9.5 10.0 —0.5 260 39 3.0 3.4 —0.4 29 40 4.6 3.5 +1.1 43 41 4.5 4.0 +0.5 57 42 3.0 3.0 30 43 4.0 5.0 —1.0 50 44 4,2 3.5 +0.7 36 45 2 al 1.5 +1.7 27 46 10.9 17 —0.8 277 47 4.7 4.5 +0.2 29 48 3.2 3.4 —0.2 36 49 2.9 2.3 +0.6 30 50 5.2 3.8 +1.4 45 crucial question arises at this point. If the tendency to higher results by the new method is due to an interfering substance, and this substance tends to increase in the blood in conditions of kidney inefficiency, then we would have to conclude that the new method could not be of service for determining uric acid. This point seems adequately covered by the results on. the bloods which showed 60 mg., or more, of non-protein nitrogen per 100 ec. (Table I; Nos. 1, 5, 7, 11, 16, 25, 26, 38, 41, and 46). These bloods cover a wide range of nitrogen retention as indicated by figures from 60 to 277 mg. of non-protein nitrogen per 100 cc. There is excellent agreement by the two methods in five out of the ten bloods. In three of the remaining five, the new method gives lower results than the old, ranging from 0.4 to 2.1 mg. per 100 ce. of blood. In only one case of a blood with marked accumu- lation of non-protein nitrogen does the new method yield an appreciably higher figure than the old; vzz., 1.1 mg. per 100 ce. in Blood 16. We believe that there is definite basis for assuming that the lower figures which the new method gives in the majority of the retention cases are the more correct. Theoretically the lower 200 Uric Acid Determination figures should be regarded as correct, and it can be shown that ‘there is experimental evidence in favor of this view.° The silver lactate precipitation of uric acid from dilute blood filtrates as suggested by Folin and Wuis a remarkable example of accomplishing the seemingly impossible, and is not at alla simple ordinary precipitation as might be inferred from reading a descrip- tion of the method. Folin and Wu provide for precipitating uric acid from the blood filtrate where the actual concentration of the uric acid may be even less than 1 mg. in a liter, or 1 part in 1,000,000. Even the most insoluble salts known, such as silver iodide cannot be simply precipitated from such great dilution. Furthermore, if one adds the silver lactate solution to pure uric acid solution of five or even ten times the -concentration existing in normal blood filtrates there is no visible precipitation. Yet the fact remains that the uric acid is actually removed from the blood filtrate by the silver lactate precipitation. Hence we must infer that the uric acid (probably as the silver salt) is adsorbed by, or in some way dragged down with, other substances simultaneously precipitated. Probably chloride, oxalate, and tungstate (all of which are usually present in the Folin-Wu blood filtrate) con- tribute to this removal of uric acid from these dilute solutions. It is, however, doubtful upon theoretical grounds whether a precipitation accomplished in this way can be highly specific for any one compound. It is easy to demonstrate that in the case of blood filtrates other substances are carried down with the uric acid, and what is important for our present discussion, that most of the color-yielding substances (in the uric acid reaction) are so precipitated. Here we encounter findings which will impress many readers as not only difficult of interpretation, but as prob- ably involving false assumptions or incorrect work. Yet the present writer has worked on this question in considerable detail and can offer only the conclusions given below, together with the facts upon which they are based. It is notable that Folin and Wu, instead of dissolving their precipitated uric acid in cyanide, and making the determination directly upon this solution as they do in the case of the urine, ® Uric acid added directly to blood filtrates is recovered very satis- factorily in the new method, indicating that there is nothing in blood filtrates which inhibits the development of the color due to uric acid. S. R. Benedict 201 take the extra steps of decomposing the precipitate with a chloride solution and centrifuging off the silver chloride. No explanation is offered by Folin and Wu as to why this apparently unnecessary procedure is adopted in the case of the blood. But if we attempt to get along without it, we immediately get into trouble. If the ‘silver urate” precipitate is dissolved in enough cyanide to readily bring it into solution (about 10 drops of 5 per cent solution) and an equal quantity of cyanide added to the standard solution (employing, of course, the old phosphotungstic acid reagent for the final color production) we commonly obtain results which are from 150 to 300 per cent too high. The high results under such conditions might be due to one of three factors: (1) The silver present in the solution where the precipitate has been directly dissolved might be respons ble for the high result; or (2) other color-yielding substances may have been precipitated from the blood which dissolve unaltered in the cyanide, but which are decomposed by treatment with acid sodium chloride so that they no longer yield color with the uric acid reagent; or, (3) other color-yielding substances may have been precipitated from the blood which pass into solution when the whole precipitate is directly dissolved, but which remain in the precipitate when treatment with acid sodium chloride is employed. It can be proved beyond a reasonable doubt that neither of the first two suppositions can explain the high results. If the same silver lactate solution employed in the precipitation is added to a standard solution of uric acid it has practically no effect upon the reading. If a few drops (1.5 drops for each cc. of solution) of 10 per cent sodium tungstate solution (the one em- ployed in the original precipitation of the blood) are added to a standard solution of pure uric acid of about the concentration found in blood filtrates, and this solution is then treated with the silver lactate, the precipitate centrifuged off, and the procedure concluded as for the blood, it will be found that the uric acid taken is recovered almost exactly whether the precipitate be decomposed by the acid chloride solution, or whether it be dissolved directly in cyanide. These results seem to demonstrate quite conclusively that the high results where cyanide is employed directly on blood filtrates 202 Uric Acid Determination are not due to the presence of the silver. The demonstration is of importance and shows that the high figures where cyanide is employed directly, are not due to finely divided metallic silver (which would, of course, reduce the phosphotungstic acid re- agents). That metallic silver plays no part in this connection is also indicated by the light color of all the precipitates obtained with the blood. The merest traces of metallic silver in a mixture~ will darken it perceptibly, and this effect is absent from the pre- cipitates with properly prepared and kept silver lactate solution. The second possibility above suggested, wz. that the high re- sults may be due to a compound which is destroyed by the acid solution, is shown not to be correct by the fact that if the pre- cipitate from blood is stirred with acid sodium chloride and then the entire precipitate dissolved in cyanide the results are still two or three times too high. By exclusion then, we arrive at the conclusion that the silver lactate precipitates other compounds from blood filtrates which give the uric acid reaction with the old reagent. This conclusion is brought practically to the point of certainty by noting that the figures obtained by dissolving the “‘silver urate” directly in cyanide increase with increase of total interfering substances in the blood. Indeed for many bloods it seems that practically all the color- yielding substance present in the blood, uric acid, and non-uric acid, is precipitated by the silver lactate solution. Illustrative results in this connection for a few bloods are given in Table II. In order to show the much greater specificity of the new process for uric acid determination, figures obtained by applying the new process directly to the same blood filtrates, are also included. It seems clear from the results reported in Table II that there is a direct relationship between the total color-yielding compounds in the blood, and the color obtained where the ‘‘silver urate” is dissolved directly in cyanide. We must, therefore, conclude that while these substances may be largely precipitated by silver lactate in lactic acid, Folin and Wu found a means of decomposi- tion for the precipitate which effects a remarkable separation between the uric acid and the other color-yielding compounds. It seems quite inexplicable to the present writer how, upon theo- retical grounds, this separation can be effected, but the facts seem to admit of no other explanation. It seems reasonable, S. R. Benedict 203 however, to assume that the tendency to higher results by the Folin-Wu method for the bloods high in non-protein nitrogen is best explained upon the assumption that some of the precipi- tated non-uric acid color-yielding material may be set free during the decomposition of precipitates containing excessive quantities of this material. Certainly it appears that the lower results for such bloods by the new method are probably more nearly the correct ones. The tendency to somewhat higher figures by the new method for ordinary bloods as exemplified in Table I can scarcely be definitely explained at present. It is possible that the excess TABLE II. Showing the results obtained with human bloods when the “‘silver urate”’ precipitate is dissolved directly in cyanide solution. Comparative figures are given by the modified Folin-Wu decomposition method, and by the old and the new reagents applied directly to the blood filtrate. Per 100 cc. of blood. Sample No. Folin-Wu_ | ‘Silver urate’’| Old reagent | New process | regular decom- | dissolved applied applied Non-protein position of directly in directly to directly to | nitrogen. precipitate. cyanide. blood filtrate. | blood filtrate. i mg. my. mg. mg. mg. 1 1.5 5.2 10.1 2.3 43 2 10.2 27.4 28.1 9.5 281 3 2.4 6.5 7.6 3.4 38 t 7A | 5.6 7.8 2.9 37 5 2.5 5.0 7.4 3.1 39 6 15.7 19.1 20.5 13.6 197 figures represent an effect of, some interfering substance which shows no tendency to increase with kidney inefficiency. This explanation is unlikely, especially since the difference between the two methods tends to disappear or is reversed as soon as the uric acid by the Folin-Wu method approaches a figure of 3.5 to 4.0 mg. The only evidence that we have that the figures by the new method represent essentially only uric acid in the bloods showing the discrepancy as well as in the others, is the fact that where the determinations are made by letting the filtrates stand in the cold, so that the maximal color develops slowly, it is found that where standard and unknown contain approximately the same 204 Uric Acid Determination amounts of the reacting substance as measured by the final color, results are practically identical whether the readings be made early or late in the process of color development. This similarity in rate of color development in standard and unknown lends sup- port to the view that the figures represent uric acid only. Con- versely, since the exact conditions of salt content, acidity, etc., and other factors which may influence the adsorption precipita- tion of the silver urate in the blood filtrates cannot be readily duplicated upon solutions totally free from uric acid for tests upon complete recovery, there is some question as to whether the small increment of uric acid shown in the new direct method may not be lost in some way in the precipitation and decomposition method. Another possible explanation which suggests itself concerning these differences is that in these bloods we may be dealing with the presence of some second form of uric acid—combined or al- tered in some way so that it reacts more strongly in the new method than in the old. We investigated this possibility, and soon be- came convinced that it deserves no consideration. When we applied the new process to the solution obtained by decomposing the silver precipitate as in the Folin-Wu method we invariably obtained figures identical with or lower than where the old reagent ~ and procedure were employed. In no instance where the new and the old processes have been applied actually to portions of the same solution, whether the blood filtrate directly, the solu- tion after sodium chloride decomposition, or the solution obtained by dissolving the silver precipitate directly in cyanide, have we obtained higher figures by the new process. They have usually been appreciably lower. Hence it seems that there can be no question of a second form of uric acid which reacts differently with the two procedures in these bloods. It seems that an impartial study of Table I should lead to the adoption of the new procedure for general work.'° The method 10 Tf one prefers to continue using the Folin-Wu precipitation and de- composition method, the new process can, of course, be applied to the solution thus obtained. 5 or 10 ce. of filtrate are precipitated and de- composed as in the Folin-Wu method, the volume being made to 10 ce. with water before the final stirring and centrifugation. The clear solu- tion obtained after centrifuging is poured as completely as possible into a test-tube, and the process concluded as in the regular new method. S. R. Benedict 205 is more expeditious and requires the use of less blood than does any other procedure. It is simple and gives a satisfactory depth of color and is far safer in the hands of many analysts. The differences by the two methods are in many instances negligible, and in only one or two instances out of the fifty analyses reported, would acceptance of the new figures suggest a different interpretation as regards uric acid abnormality or retention. In one instance, Blood 13, the new method gives a figure above the accepted normal, while in another, No. 43, the interpretation would be reversed if the new figures are accepted. In the latter case the figure by the new process is almost certainly more cor- rect. The difference by the two methods chiefly affects the bloods where the Folin-Wu method gives 2.5 mg. or below. Thus we find that out of the fifty bloods, the Folin-Wu method gives a figure of 2.5 mg. or below, per 100 ce. for fifteen bloods, while only five bloods give similarly low figures by the new process. Where the new process is used it is probable that the commonly accepted figures should be increased by about 0.5 mg. per 100 cc., so that the usual normal figures would probably be about 3.0 mg., while no evidence of definite uric acid retention would be evident before a figure of 4.0 mg. or over per 100 cc. was reached. The question as to whether the figure 4.0 mg. per 100 cc. of blood represents a true pathological retention must remain open, just as it is for quite similar figures by the old methods. Such border-line questions can be answered only by extensive data. A few words covering the quantity of blood filtrate necessary or desirable to carry out the new procedure may be of value. As the method is above recommended it will be noted that we finally heat a rather large volume (15 cc.) of a very dilute solution. This solution may contain as low a concentration of uric acid as 1 part in 1,500,000 parts of solution or even less, and accurate results can still be obtained. It may seem that it would be better to carry out the reaction with more concentrated solutions, or with smaller volumes. We have found that higher concentra- tions are not so desirable, as turbidity may occur, and the results as a whole are not so uniform. It is, of course, perfectly possible to cut everything in half in the process as above described, using 2.5 cc. of filtrate, and having an ultimate volume of 7.5 cc. dur- ing the heating. This is permissible where circumstances may 206 Uric Acid Determination require it, but the writer believes that the larger volumes are commonly measured with greater accuracy, and since 0.5 ce. of blood is certainly not excessive for a uric acid determination, it seems better to recommend starting with that equivalent unless there are reasons to the contrary. There are, of course, certain instances where it may be desirable or necessary to work with very minute quantities of blood, such as can be obtained readily by puncture. In the case of infants, or where repeated examinations on the same individual are to be made, the use of blood obtained by puncture may be essential. For this reason a modification of the new procedure is here de- scribed which might perhaps be properly called an “ultra-micro”’ method, since it can be used to determine as little as 0.002 mg. of uric acid, and yields results very comparable with those obtained by the procedure described earlier in this paper. The modifica- tion requires the use of only 0.2 cc. of blood (0.1 ce. of filtrate), and is carried out as follows: 0.2 ec. of blood is carefully pipetted ar a narrow pointed cen- trifuge tube. 1.4 cc. of water are added and the mixture is stirred with a fine glass rod. 0.2 ec. of 10 per cent sodium tung- state solution is added, followed by 0.2 ce. of 0.75 N sulfuric acid, and the mixture is thoroughly stirred and allowed to stand for 10 minutes. The tube is then centrifuged at high speed for a few moments and 1 ce. of the clear supernatant fluid is pipetted into a long narrow (1 cm.) test-tube. 1.8 cc. of 2.8 per cent sodium cyanide solution (containing 1.5 cc. of concentrated ammonia per liter) are then added, followed by 4 drops (0.2 ec.) of the arsenic phosphotungstic acid reagent. The mixture is then treated as. is the unknown in the regular process, and can be compared with the 0.02 mg. regular standard solution. The 3 cc. of colored solution are readily read in the narrow form cup of the Bock-Benedict colorimeter up to a height of about 24 mm. In using this modification it is, of course, essential that cali- brated pipettes be employed, and that all the measurements be made with great care. Under such conditions we have obtained surprisingly close duplicates between the two modifications. Of ten bloods analyzed by both methods the maximal difference was 0.6 mg. of uric acid per 100 cc., and in six of the samples the variation did not exceed half of this figure. S. R. Benedict 207 In conclusion one or two points of general interest in connec- tion with the new process may be mentioned. The increased specificity of the new procedure as regards uric acid lies probably chiefly in the reagent employed and only partly in the use of cyanide instead of carbonate for development of the alkalinity. A corresponding, though not so great, increase in color is obtained with the old phosphotungstic acid reagent through the use of large quantities of cyanide, but the reaction thus obtained is not very specific for uric acid. The cyanide acts in a dual capacity in causing the increased color. It has an effect upon the reac- tion, probably through combination with the uric acid as exempli- fied in the older procedures, and it is superior to carbonate as an alkali, because it causes a much slower decomposition of the reagent than does the stronger alkali, and hence gives more time for the reaction at the critical point when the linkage of the complex tungstic compound is being broken. That this view is correct is rendered probable by a fact observed by the writer some years ago that the use of borax in hot solution would markedly increase the color yield in the uric acid determination. The results were not reported because the color developed was not closely proportional to the quantity of uric acid present. In this instance the speci- ficity of the reaction was not appreciably increased, although the color yield for a given quantity of uric acid was greater than in the carbonate process. It is to be noted that the new process described in the present paper is now recommended only for application to the Folin-Wu filtrates from human blood. Studies of the new process under certain other conditions of precipitation, and for other bloods, are being carried on. We shall shortly report upon the application of the new pro- cedure to the determination of uric acid in urine. BIBLIOGRAPHY. . Folin, O., and Denis, W., J. Biol. Chem., 1912-13, xiii, 469; 1913, xiv, 29. . Benedict, 8. R., J. Biol. Chem., 1915, xx, 629. Folin, O., and Wu, H., J. Biol. Chem., 1919, xxxviii, 98. Kehrmann, F., Ann. Chem., 1888, cexlv, 45. . Fremery, M., Ber. chem. Ges., 1884, xvii, 296. . Gibbs, W., Am. Chem. J., 1885-86, vii, 313. Do on THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI, NO. 1 ee a ee et iy | ioe ale arate a’ a j ; A COLORIMETRIC METHOD FOR THE DETERMINATION OF SUGARS IN NORMAL HUMAN URINE. By OTTO FOLIN anp HILDING BERGLUND.* (From the Biochemical Laboratory of the Harvard Medical School, Boston.) (Received for publication, December 31, 1921.) The colorimetric principle underlying the method of Folin and Wu! for the determination of sugar in blood should prove equally useful for the determination of sugar in normal urines provided that a suitable process could be found for removing substances which can interfere. The practicability of the process would necessarily depend on the preliminary treatment required for the removal of creatinine, uric acid, and other materials which might have considerable reducing power. - The process finally adopted is extremely simple, and from the very extensive use which we have made of it during the past year we have become convinced that it meets all practical re- quirements. The preliminary treatment consists only of shaking the urine with ‘‘Lloyd’s alkaloidal reagent,’? a concentrated fullers’ earth. This reagent removes most of the coloring matters together with the uric acid, creatine, and the creatinine, yet, unlike all or most effective charcoals, does not take away the sugar. | It is not necessary that every trace of creatinine should be removed for relatively considerable amounts have absolutely no effect on the sugar method of Folin and Wu. That this is the case was shown in connection with the application of the method to blood. * Work conducted under a grant from the Swedish Society for Medical Research. 1 Folin, O., and Wu, H., J. Biol. Chem., 1920, xli, 367. ? The authors wish to express their appreciation of the courtesy and generosity with which J. U. Lloyd (Cincinnati, Ohio) has supplied them with all the ‘‘alkaloidal reagent’’ required for this work. 209 210 Sugars in Normal Human Urine The process is as follows: To 5 ec. of urine add 5 ec. tenth normal sulfuric acid and 10 cc. of water. Add 1.5 gm. of Lloyd’s reagent and shake gently for 2 minutes. Filter. 2 ec. of the filtrate are the usual amount used for the sugar determination. The above mentioned dilutions are for concentrated urines. With more dilute ones, one takes 10 or 15 ee. and reduces the amount of water taken. The shaking with Lloyd’s reagent should not be continued longer than 2 minutes because the reagent is gradually dissolved by the acid and because longer shaking does not take out any more. The dissolved aluminate from the reagent does not disturb the determination at any stage. The colorimetric determination of the sugar in the filtrate is made in exactly the same manner as in the case of blood filtrates. For the determination of the total sugar we hydrolyze as follows: To 10 cc. of the filtrate obtained after shaking with Lloyd’s reagent add 1 cc. of 10 per cent hydrochloric acid and heat in boiling water for 75 minutes. This heating should be done in test-tubes graduated at 20 cc.—for purposes of subsequent dilution. After hydrolysis, cool thoroughly and neutralize with normal sodium hydroxide. Phenolphthalein may be used as indicator, if desired, but is not necessary as the cloud produced from the material dissolved out of Lloyd’s reagent furnishes an adequate indicator of the degree of neutrality required. Add the alkali until the cloud so formed does not disappear on shaking. Dilute the neutralized hydrolysate to the 20 ec. mark. Then add a small pinch of Lloyd’s reagent and invert half a dozen times. This is for the purpose of removing most of the coloring matter formed during the hydrolysis. 2 cc. of this more dilute filtrate are usually a suitable amount to take for this determina- tion also. The standard sugar solutions to be used are the same as for the blood; namely, such as contain 1 and 2 mg. of glucose per 10 cc. A few remarks concerning such standard sugar solutions may appropriately be introduced here. It is a common ex- perience that whereas moderately concentrated sugar solutions can easily be preserved the dilute ones often deteriorate. We suspect that the cause for this deterioration is alkali given off from the containers rather than destruction by microorganisms. O. Folin and H. Berglund 211 At all events, we have found that the dilute as well as the con- centrated solutions keep perfectly in 0.3 per cent of benzoic acid and we, therefore, make the original stock solution containing 1 per cent of glucose by means of 0.3 per cent benzoic acid solution. The same benzoic acid solution is then used, instead of water, for the preparation of the dilute standard solutions. SOME NEW OBSERVATIONS AND INTERPRETATIONS WITH REFERENCE TO TRANSPORTATION, RETEN- TION, AND EXCRETION OF CARBOHYDRATES. By OTTO FOLIN ann HILDING BERGLUND. (From the Biochemical Laboratory of the Harvard Medical School, Boston.) (Received for publication, December 31, 1921.) CONTENTS. Review of the more important literature with some comments......... 213 Sun MEIEIEVE OC IMIVESIIPAGION. coc. cs ee ee tes ee cleceeseceees 227 Experiments with glucose, maltose, dextrin, and starch.............. 229 Penman wre trmetone.. 0.2.6). eee cee 241 Experiments with galactose and lactose..................0000 cece eee 247 Nature and origin of the sugar of normal urine...................... 254 On the nature of the glucose threshold, together with some observa- 0 LeU oC ee 256 On emotional hyperglycemia and glycosuria.......................... 262 On subfasting blood sugar levels (hypoglycemia)..................... 265 Distribution and character of the blood sugar....................... 270 SUE 272 Review of the More Important Literature with Some Comments. Many investigations have been reported representing the simul- taneous determination of certain waste products in blood and in urine, the general purpose of which has been to elucidate how the concentration of these products in the blood determines the rate at which they are eliminated through the kidneys. A fairly large number of analogous investigations have been made with “for- eign” substances and also with certain food products, particularly glucose. Glucose has heretofore been almost the only food mate- rial so investigated, and for many obvious practical reasons. The field as a whole is an interesting one, for it has to do not only with the activity of the kidneys but with other equally important pro- blems, notably the power of the tissues to abstract from the blood with a high degree of speed and completeness many products such 213 214 Carbohydrates as uric acid, creatine, amino-acids, and probably many other substances, besides the water-soluble food materials. In the present paper we deal mainly with the rise and fall of sugar in blood and urine following the intake of glucose and other carbohydrates. There is a very large literature both old and new on this subject, yet both as to the essential facts and still more as to the interpre- tations the modern literature is almost as conflicting as the old. Compare for example the opinion of Shaffer! (1921) on the sugar of normal urine with that of Benedict? published in 1918. The one virtually denies the presence of glucose in normal urines, the other almost implies that all are more or less on the verge of be- coming really diabetic. When it comes to such concepts as limits of sugar assimilation, sugar tolerance, and the renal threshold for sugar, a careful sifting of literature is greatly needed. Our original purpose was to include such a sifting of the litera- ture in this paper, but it soon became apparent that to do so would take up too much space, and we have confined ourselves to brief summaries of only the more important investigations. The existence of a relationship between the concentration of sugar in the blood and the appearance of sugar in the urine was clearly recognized by Claude Bernard,? and was brought out in connection with the experimental glycosuria produced in dogs by his sugar puncture as well as by the administration of curare. Claude Bernard even gives definite figures (224 to 260 mg. per 100 cc. of arterial blood) for the blood sugar concentration below which no glycosuria occurs; and in another connection he observes that the blood sugar may sometimes rise to 300 mg. without caus- ing the ‘“‘healthy’”’ kidneys to eliminate any sugar. The latter figure is still cited in some text-books.* Claude Bernard had thus clearly in mind the renal threshold concept for sugar, although he did not use the term. Claude Bernard did not specifically deny the presence of minute traces of sugar in all normal urines; he simply omitted to express a conviction on that problem. 1 Shaffer, P. A., and Hartmann, A. F., J. Biol. Chem., 1920-21, xlv, 365. 2 Benedict, S. R., Osterberg, E., and Neuwirth, I., J. Biol. Chem., 1918, XXXxiv, 217. 3 Bernard, C., Legons sur la diabéte et la glyeogénése animale, Paris, 1877. 4Cushny, A. R., The secretion of the urine, London, 1917. ; O. Folin and H. Berglund 215 Pavy,® though originally a pupil of Claude Bernard, came to entirely different conclusions concerning the conditions and rela- tionships existing between the sugar of blood and the sugar in the urine. Pavy held that the kidneys are completely permeable for glucose, that normal urine always contains traces of sugar, and that the reason why it does not contain more is the fact that the blood contains only similar insignificant traces of free sugar. The major part of the so called blood sugar Pavy regarded as a post- mortem phenomenon, for in life the absorbed sugar is held in com- bination with protein material in non-dialyzable form. It is easy enough to comprehend why Pavy never could give up the concept of some chemical glucose-protein combination in blood as the mechanism by which the escape of all but insignifi- cant traces of sugar into the urine are prevented. While he never succeeded in proving the correctness of his theory he nevertheless did prove by his discovery of the glycoproteins that sugar-protein combinations do exist to a limited extent. Definite proof that the Pavy-Lépine interpretation is incorrect, that the blood sugar exists in free, diffusible form, was finally furnished through the compen- sation dialysis experiments of Michaelis and Rona.® The reten- tion of a normal amount of blood sugar as well as the failure of such retention at a given concentration, above the sugar threshold, must therefore presumably be a feature of the kidney activity alone, and this is indeed the conclusion arrived at by Hamburger and Brinkman in their highly interesting work on the excretion of sugar urines by the glomeruli of frog’s kidneys, Hamburger’s investigations are interesting not only in connection with the problem of why the kidneys eliminate glucose when the glucose of the plasma exceeds a certain concentration (the threshold), but the results are also extremely interesting as illustrations of the selective activity of kidneys, or of phenomena simulating such activity. By varying the calcium and the sodium bicarbonate content of the modified Ringer’s solutions with which the frog’s kidneys were perfused the activity of the kidneys to added glucose could be varied from complete permeability to complete impermeability for glucose in concentrations corresponding to those found in frog plasma. In response to an excess of sugar, 0.2 to 0.3 per cent, the impermeability of the kidneys broke down. eS EE ® Pavy, F. W., On carbohydrate metabolism, London, 1906; J. Physiol., 1899, xxiv, 479; 1900-01, xxvi, 282; Lancet, 1908, 1499, 1577, 1727. ®° Michaelis, L., and Rona, P., Biochem. Z., 1908, xiv, 476. 216 Carbohydrates The conditions resulting in the complete retention of added glucose had no such effect on other sugars. When mixtures of glucose and levulose or glucose and lactose were used in suitable concentrations, the levulose or the lactose was passed quantitatively into the “‘urine,’’ while glucose remained in the perfusing fluid. Similar results representing varying degrees of separation were obtained with a number of other sugars. One must necessarily hesitate about the extent to which one is willing to accept the theoretical discussion given by Hamburger, but the data are at least interesting and seemingly represent several years of patient research.? A characteristic feature of the more ‘recent investigations has been the establishment of the necessary data with a greater degree of precision than was possible by the methods of Pavy or any of the older investigators. Jacobsen® in Copenhagen was probably the first who on the basis of a modern method for the determination of blood sugar (Bang’s) attempted to find the blood sugar level at which positive (clinical) tests for sugar are obtained in the urine. He gave 100 gm. of glucose to each of fourteen normal persons (either before, or 2 to 3 hours after breakfast) and deter- mined the blood sugar every 15 minutes and tested the correspond- ing urines for sugar, mostly by Almén’s test. In eight of the four- teen subjects he obtained positive alimentary glycosuria, and he found that the glycosuria began when the blood sugar rose above 160 to 170 mg. per 100 cc. of whole blood. The remarkable fea- ture of Jacobsen’s results is that he should have obtained clinically positive sugar in the urine and such high levels of glucose in the blood in so large a proportion of his subjects after the taking of only 100 gm. of glucose. It is perhaps worth noting that all of the subjects responding with positive alimentary glycosuria took the glucose 2 to 3 hours after breakfast, whereas the common experience indicates that the sugar tolerance is lowest before breakfast (Naunyn’). Quite recently, Hagedorn,’ also from Den- 7 Hamburger, H. J., and Brinkman, R., Biochem. Z., 1918, Ixxxviii, 97; 1919, xeiv, 131; Proc. K. Acad. Wetensch. Amsterdam, 1918, xxi, 548. Ham- burger, H. J., and Alons, C. L., Biochem. Z., 1919, xciv, 129. Hamburger, H. J., Proc. K. Akad. Wetensch., Amsterdam, 1919, xxii, 351, 360. Brit. Med. J., 1919, 267. 8 Jacobsen, A. T. B., Biochem. Z., 1913, lvi, 471; Blodsukkerindholdet hos normale og ved Diabetes mellitus, Copenhagen, 1917. 9 Naunyn, B., Der Diabetes melitus, Vienna, 1906. 10 Hagedorn, Acta med. Scandinav., 1921, liii, 672. bens O. Folin and H. Berglund 217 - mark, seems likewise to have met with very little difficulty in obtaining alimentary glycosuria after giving from 63 to 120 gm. of glucose; only in Hagedorn’s cases the sugar was taken before breakfast. In commenting on Jacobsen’s results Hamman and Hirschman" in their paper on the same subject suggest that the persons giving positive sugar tests for the urine must have had a subnormal tolerance for sugar. Yet this is scarcely a satisfactory explana- tion in view of the large proportion of positive results and in view of the fact that the blood sugar reached such high levels before the appearance of sugar in the urine. A more plausible explana- tion occurred to one of us (B.) who has had experience with Bang’s drop method of blood analysis. The physical pain pro- duced by repeated needle pricks in the fingers is distinctly greater than that experienced in having blood taken from a vein, and a certain amount of superimposed psychical hyperglycemia may have been sufficient to produce the positive results obtained. Jacobsen, it may be noted, obtained as high levels in the blood sugar by the giving of 160 gm. of bread as by giving of 100 gm. of glucose, whereas according to most other investigators starch feeding does not produce nearly so high a level of hyperglycemia. The obstacle encountered in trying to determine the renal thres- hold for sugar in normal persons lies in the fact that the hyper- glycemia just adequate for the production of glycosuria, if pro- duced at all, lasts for so short a time that one very seldom suc- ceeds in collecting samples of urine representing the desired period. It is, therefore, not surprising to find that the available valid data on this point are very few. The majority of the figures so far reported indicate that the glucose threshold for strictly normal persons most frequently lies between 160 and 180 mg. per 100 ee. of whole blood—and somewhat higher, about 190 mg., as the upper limit for the plasma. Threshold values lying above 180 for whole blood, or 190 for plasma, have not yet been found for normal persons. The figures cited, 160 to 180 mg., are usually referred to as the normal glucose threshold level, but it would perhaps be more correct to call them the highest normal level for the glucose thresh- old, because lower levels are frequently obtained. Thus Hamman "™ Hamman, L., and Hirschman, I. I., Arch. Int. Med., 1917, xx, 763. 218 Carbohydrates and Hirschman gave 100 gm. of glucose to each of six normal per- sons, and in none of these could they find a glucose level above 150 mg. per 100 cc. of whole blood (Lewis-Benedict method), yet two of the six showed sugar in the urine. In several other persons not normal these authors obtained threshold values of 170 to 180 mg.; and by the subcutaneous injection of 0.6 to 1 mg. of epinephrine they succeeded, on the basis of an intake of only 65 gm. of glucose, in obtaining such gradual increases of the sugar level in the blood that accurate threshold values could be obtained in two cases. In these two cases the threshold was again found to be 170 to 180 mg. Goto and Kuno” have recently published an extensive series of threshold experiments on 53 Japanese subjects using plasma instead of whole blood and employing Benedict’s method (Myers- Bailey modification) for the sugar determination. The sugar, 100 gm. of glucose + 250 cc. of water, was given on an empty stomach in the morning. On the basis of a critical study of their essential data we would make the following summary of their results. Twenty-two of the subjects showed sugar in the urine. In eight of these the sugar threshold (for plasma) was between 180 and 190 mg., in five between 160 and 172, and in eight the threshold fell between 128 and 155 mg. In one case the value for the threshold fell as low as 120 mg., without there being any reason to suspect the existence of diabetes. The exact records of the blood sugar level at which glucose begins to appear in the urine are far from numerous, yet it is clear that while fairly definite figures can be assigned to the upper limit of the sugar threshold the lower levels have practically no definite limitation, but gradually descend until one comes to the ‘‘cyclic glycosuria of the renal type’ (Faber and Norgaard) where there is a transient mild glycosuria after every meal, and finally to the chronic renal glycosuria, where the sugar threshold is so low that it is more or less difficult or even impossible to cause the level of the blood sugar to sink below the threshold. There are two other important points to be noted in connection with the sugar threshold in normal persons, important because they emphasize the uncertainty which must necessarily be at- 12 Goto and Kuno, Arch. Int. Med., 1921, xxvii, 224. 13 Faber, K., and Norgaard, A., eid med. Scandinav., 1920-21, liv, 289. O. Folin and H. Berglund 219 tached to any definite figure supposed to represent the threshold concentration. The first of these is the fact that the threshold determined on the basis of a descending curve of sugar concentra- tion in the blood may be entirely different from the threshold derived from an ascending curve. Goto and Kuno, for example, report one experiment where the sugar elimination began at 150 my. of glucose per 100 cc. of plasma, but continued until the plasma sugar sank to 60 mg.; and two others where the elimination began at about 168 mg. and continued until plasma levels of 77 and 99, respectively, were reached. The other important point is one recently elucidated and emphasized by Henriques and Ege.* The level of the blood sugar is not the same in venous and in arterial blood, particularly when there is hyperglycemia due to ingestion or injection of glucose. It is, therefore, precisely under conditions demanded for threshold determinations that such differences imply uncertainties as to the meaning of any exact figures given for the threshold. The differences are by no means uniform, but are quite insignificant when the level is distinctly below the thres- hold. That any such differences should be ascertainable shows how much more accurate are the modern methods for the deter- mination of the blood sugar than were the methods by means of which the earlier investigators tried to prove the retention of sugar by the liver. We do not feel justified in omitting from this discussion of the literature on the blood sugar threshold the contributions which have come from Ambard® and his pupils. These French investi- gators have introduced conclusions, explanations, and generaliza- tions which will undoubtedly figure in the literature for some time to come—yet we have been unable to satisfy ourselves that they are based on sound experimental data or that the new concepts advanced represent much else than empirical, fanciful, and mis- leading speculations. According to Ambard, the characteristic feature of the sugar threshold consists of its constant endeavor to adjust itself to the fluctuations in the blood sugar level, and the threshold itself is, therefore, constantly changing, up or down, each change being a 14 Henriques, V., and Ege, R., Biochem. Z., 1921, exix, 121. 18 Ambard, L., Physiologie normale et pathologique des reins, Paris, 2nd Edition, 1920. Chabanier, Arch. d’Urol. de Necker, 1919, ii, 1. 220 Carbohydrates response to the stimulus of a change in the blood sugar concen- tration. In actual figures Chabanier thus finds that the threshold may correspond to any level between 107 and 520 mg. of glucose per 100 ec. of blood. Ambard assumes that the “kidney secretion constant, 0.07,” a constant derived from an empirically selected urea concentration in urine of 2.5 per cent, is equally applicable for all other substances which find their way into the urine by virtue of the secretory activity of the kidneys—provided that the concentrations of these substances in the urine are recalculated to concentrations which are isotonic with 2.5 per cent urea solution. It will be no small task to prove the validity of such a remarkable assumption. In their attempts to explain the sugar elimination in terms of the Ambard hypothesis Chabanier and his associates endeavored to eliminate or annul the sugar threshold by means of increasing doses of phlorhizin until further increases of the phlorhizin pro- duced no more increase in the sugar elimination. On the basis of this experimental condition they endeavored to determine whether the sugar threshold alone had been eliminated or whether at the same time there had been a general change in their so called secre- tory constant, 0.07. This constant had not changed for urea, and because of no change in the urea elimination they concluded . that the secretory constant for glucose must also remain unaf- fected. It is not at all strange that a considerable degree of similarity should be found between the elimination of urea and the elimination of sugar in a completely diabetic animal to whom the sugar is a useless, indeed one might say, a foreign substance. Any other foreign innocuous water-soluble substance would prob- ably behave in the same way. But to conclude from this that the reason for it all is a certain empirically selected “secretory constant”’ is fanciful speculation, for the kidneys are by no means the only organ involved in the retention or elimination even of foreign materials. The absorption, retention, and release of those substances by the tissues in general play an important part. But Ambard and Chabanier now make use of their secretory con- stant as if it represented an established fundamental fact, and proceed to calculate on the basis of it the sugar threshold prevail- ing in other glycosurias than that of complete phlorhizin diabetes, and the figures obtained for the threshold vary between 107 and O. Folin and H. Berglund 221 520 mg. In the light of the experimentally observed thresholds found by other investigators we consider these calculated thres- hold values as mathematical proof of the non-existence of the secretory constant. And indeed, to us at least, it seems infinitely more probable that the secretory constant, rather than that un- known something called the glucose threshold, should change normally merely in response to changing concentrations of sugar in the blood. It is at least important for the present to keep sharply apart all directly determined blood sugar threshold values and all so called threshold values obtained by calculation by means of Ambard’s formula. A review of the sugar threshold in diabetes does not come with- in the scope of this paper. While there may be a more or less general tendency toward a gradual increase of the blood sugar con- centration at which glycosuria begins, it seems nevertheless to be a fact that many diabetics show the normal, and some even the subnormal, thresholds encountered in non-diabetic individuals (Faber). It should be important to determine, if possible, whether the high threshold values found in diabetes are an inevitable ultimate feature of the abnormal carbohydrate metabolism or an indirect result depending on a general deterioration of the kidneys. In connection with this problem it is suggestive that abnormally high blood sugar thresholds have been found in ne- phritis—associated with retention of nitrogenous waste products. In one such case Hamman and Hirschman found a threshold value of about 205, Mason" found a value of over 217, and Williams and Humphreys!’ obtained one threshold value at over 250 mg. (by Benedict’s method) and two others where the value found exceeded 200 mg. The increased level of the sugar threshold is not the only abnormality of carbohydrate utilization encountered in these cases—the high blood sugar values so frequently occur- ring together with non-protein nitrogen values of 150 mg. or more indicate that some other factor than a decrease in the efficiency of kidneys is involved. A clear understanding of the fluctuations of the blood sugar and the existence of a renal threshold for glucose would seem to be indispensable in any attempt to interpret such concepts as 16 Mason, EH. H., Arch. Int. Med., 1918, xxi, 216. 17 Williams, J. R., and Humphreys, E. M., Arch. Int. Med., 1919, xxiii, 537. 222 Carbohydrates the limits of sugar assimilation or that of carbohydrate tolerance, yet most of the literature on these two subjects is based merely on qualitative tests for sugar in the urine. The term ‘‘carbohydrate tolerance’ was introduced in con- nection with diabetes. It represented originally the amount of carbohydrate which a diabetic patient could take in the course of 24 hours without showing sugar in the 24 hour urine. In recent years the term ‘‘carbohydrate” in this connection has come to include the carbohydrates which are produced within the organ- ism from protein materials. By the “‘limit of sugar assimila- tion’’ a term introduced by Hofmeister,'* is understood the amount of sugar which can be taken in one dose without any demon- strable loss of sugar with the urine. This concept has also had clinical applications; in a number of diseases other than diabetes it was found that the limit of sugar assimilation was materially lower than in normal individuals. The test has usually been made in the form introduced by Bloch'*—100 gm. of glucose taken after breakfast. The test was extensively used, particularly in Germany 20 to 30 years ago, but the results obtained were far from uniform either in normal persons, or in others, and the clinical value of the test has long been considered rather small.*° In its original form, and as generally understood, alimentary glycosuria implied the false assumption that the taking of a larger dose of sugar than just sufficient to produce some glycosuria meant that all the surplus sugar taken would pass into the urine. The erroneous character of this assumption was clearly pointed out, in 1895, by Linossier and Rogue,*! who introduced the term “coefficient of utilization”’ to express the fact that the more sugar that is taken the more is retained and utilized by non-diabetic individuals. The same point was expressed by Allen” (1913) in the even more emphatic terms of his “‘ paradoxical law of dextrose,” according to which there is absolutely no limit (‘‘except death”’) to the amount of sugar which the non-diabetic individual can utilize. The most exact data concerning the utilization of dex- 18 Hofmeister, F., Arch. exp. Path. u. Pharmacol., 1888-89, xxv, 240. 19 Bloch, G., Z. klin. Med., 1893, xxii, 525. 20 Kraus, F., and Ludwig, H., Wien. klin. Woch., 1891, iv, 855. 21 Linossier, G., and Rogue, G., Arch. med. exp., 1895, vii, 228. 22 Allen, F. M., Glycosuria and diabetes, Cambridge, 1913. O. Folin and H. Berglund 223 trose have been furnished by Woodyatt* and his associates. By continuous intravenous injection of glucose these investigators found that 0.85 gm. per kilo body weight per hour could be utilized without loss of sugar with the urine. By continuous injection of twice as much sugar, 1.7 gm. per kilo per hour, about 10 per cent is lost with the urine; and with about four times as much, 3.6 gm. of glucose per kilo per hour, the loss is still only 35 to 40 per cent. Further increases in the amount of sugar injected are not ac- companied by any increase in the percentage of sugar lost through the kidneys. The rate of glucose absorption from the digestive tract, ac- cording to Woodyatt, never exceeds 1.8 gm. per kilo per hour, from which one would conclude that a transient loss of about 10 per cent of absorbed glucose is the maximum that could be ob- tained on taking glucose by mouth, and the rate of loss would necessarily be of very short duration even when very large amounts of dextrose, several hundred grams, are taken. The retention of glucose by the non-diabetic individual might well be interpreted as a most striking illustration of the applica- bility of the mass law to metabolism processes. But, as in most biological processes, all the factors involved are not clear or known. The renal threshold for blood sugar and the mass law cannot yet be combined into a harmonious whole as an explanation of the facts in the case. According to the mass law alone there should be no sharp break anywhere in the relation between the amount of sugar absorbed and the amount excreted by the kidneys, yet in the renal threshold concept we have an expression of the fact that a very sharp break does occur at a certain point of blood sugar concentration. Theoretically speaking it should be and is incorrect to divide urines, as clinicians always have done, into sugar urines and sugar- free urines, but there is no general agreement as to how much of an error or how important an error is represented by such a classi- fication. The most extensive investigations on this point yet published are undoubtedly those of S. R. Benedict and associates. Before taking up the conclusions reached by Benedict concern- *3 Woodyatt, R. T., Sansum, W. D., and Wilder, R. M., J. Am. Med. Assn., 1915, Ixv, 2067. Woodyatt, R. T., The Harvey Lectures, 1915-16, xi, 326. 224 Carbohydrates ing carbohydrate metabolism, particularly sugar retention versus losses of sugar, we venture to make the following summary of his experimental findings. The basis for the work is the colorimetric determination of sugar in normal urines after all disturbing substances have been removed by the mercuric nitrate treatment of Patein and Dufau.4% The sugar elimination, by 2 hour periods, under the influence of dif- ferent diets was then determined. The subjects were two normal persons (assistants). The older person (O., age 51) eliminated on mixed diets 996 mg. per 24 hours; on a diet containing very little carbohydrates the sugar output sank to 775 mg., and in response to high carbohydrate diet the sugar output rose to 1,479 mg. The corresponding figures obtained from the younger subject were respectively 639, 500, and 1,528 mg. of sugar for 24 hours. These variations in sugar elimination did not immediately follow changes in the carbohydrate content of the food, as in the case in diabetic individuals. It took 3 days of high carbohydrate feed- ing, for example, to bring the excreted sugar up to 1,528 mg. in the younger person. The urines were usually collected and analyzed, as stated, in 2 hour periods. The lowest values, 36 and 19 mg. per hour, were obtained between 6 and 8 in the morning, though equally low figures sometimes occurred between 3 and 7 in the afternoon. Every meal was followed by a marked temporary increase in the sugar output; but this increase was not materially dependent on the carbohydrate content of the meal, for the mixed diet occa- sionally produced a larger increase than the carbohydrate-rich diet. The effects noted on giving definite amounts of carbohydrates in the form of sugar also merit attention. The giving of 40 gm. of glucose before breakfast was followed by an increase of the sugar elimination amounting to 50 mg. per hour (for 2 hours), while 60 gm. produced an increase of 254 mg. per hour in the case of the older subject. In the younger person 50 gm. of glucose had no effect, and 60 gm. produced an increase of only 7 to 13 mg. per hour. 20 gm. of glucose before breakfast had no effect on the older person, but the same together with the breakfast produced 24 Patein and Dufau, J. pharm. et. chim., 1902, xv, 221. Benedict, 8. R., and Osterberg, E., J. Biol. Chem. 1918, xxxiv, 195. O. Folin and H. Berglund 225 an increase of 129 and 137 mg., respectively, above the increase obtained from the breakfast alone. 20 gm. of cane-sugar, whether with or without breakfast, had no effect on the sugar elimination. Similar but much smaller increases were obtained from the younger person. Finally on taking 25 gm. of glucose with each meal the 24 hour sugar elimination rose from 935 mg. to 2,350 mg. the Ist day and 2,190 mg. the 2nd day. At this point the authors felt that the sugar eliminations were reaching dangerous proportions, and they did not feel justified in continuing the experiment. Yet in the case of the younger person 25 gm. of pure glucose added to each meal produced no increase whatever in the 24 hour sugar elimination. In many of the experiments the sugar in the urine was determined before and after fermentation. The investiga- tions did not include determinations of the blood sugar. It is interesting to note that Benedict in his discussion of the work uses the term carbohydrate tolerance exactly as is done in writings on diabetes. In order to get away from the misleading clinical term glycosuria he introduces the term ‘“‘glycuresis’ to express the increase of the sugar elimination above that of the control periods. ; The most important point that Benedict manifestly means to convey is that the quantitative increases in the sugar elimination, measured in milligrams and encountered in supposedly normal persons, is qualitatively or in origin of the same nature as is the clinical elimination of sugar in diabetes. In fact, he distinctly suggests that the difference between the normal and the diabetic individual may be wholly a quantitative one, just as is that be- tween the mild and severe diabetic. And he suggests that if the total sugar elimination amounts to more than 1.5 gm. per day, the diet should be altered until this figure, or a lower one, is reached as the upper limit. As in frank diabetes so here with apparently normal subjects does Benedict lean toward the view that the cause of glycuresis is probably associated with some failure or deficiency or exhaustion on the part of the pancreas. While Benedict’s experimental data constitute a slender sup- port for his rather startling conclusions, those data are interesting in that they furnish exact figures for the quantities of sugar occur- ring in the urines of normal persons and for the fluctuations in that Sugar occurring at different times of the day and in response to the intake of food. 226 Carbohydrates The experiments of Woodyatt and of Benedict represent oppo- site extremes with regard to the levels of sugar intake used as a basis for the study of the retention and the loss of glucose. Before taking up the consideration of our own experimental material there is one other point in the preceding literature which needs to be briefly reviewed. It is generally acknowledged that for quantitative studies of the kidney function it is the concentra- tion of a given substance in the plasma rather than in the whole blood which should give the more correct information. But the use of whole blood instead of plasma is so much simpler from an analytical standpoint as to be quite justifiable unless it can be shown that results so obtained are misleading. Among the modern investigators on the sugar threshold only Goto and Kuno have used the plasma. At present the general concensus of opinion seems to be that the distribution of the blood sugar between the plasma and corpuscles is, on the whole, so nearly equal in the case of human blood that no serious error is involved in the use of whole blood for the analysis. Quite recently, however, the correctness of this view has been flatly contradicted by Falta and Richter-Quittner,” and they have published a series of analyses which would seem to show that not only the sugar, but also the chlorides, and even the whole. of the non-protein nitrogen, including the urea, occur exclusively in the plasma. The analytical findings are accompanied by far reaching theoretical explanations and deductions, but these can well be left out of consideration until the analytical data have been verified. The essential point in Falta’s experimental work according to his own interpretation lies in the use of hirudin for the prevention of the clotting. Other anticlotting reagents, such as citrate or oxalate, are said to be useless, because they render the blood cells permeable to sugar, urea, etc., and thus hide the fact that in life the blood cells are quite impermeable to these products. By means of this explanation Falta and his coworkers have at once disposed of the contrary earlier results of other investigators, and have at the same time increased the difficulties of showing that their own results are not correct. It is very difficult at the pres- 28 Falta, W., and Richter-Quittner, M., Biochem. Z., 1919, c, 148; 1921, exiv, 145. O. Folin and H. Berglund 227 ent time to obtain hirudin. Several Danish investigators,”> Ege, Hagedorn, Gad-Andresen, and Warburg have, however, been in a position to repeat the work, have done so, and all find that the remarkable analytical results reported by Falta and his assistants cannot be verified in their laboratories. For the present, therefore, the findings of the above mentioned Austrian investigators might be left without serious consideration, but the subject is important and we have in connection with our own work included the study of the distribution of sugar, amino- acids, urea, etc., between the plasma and the corpuscles. Plan and Scope of Investigation. Our own experiments with carbohydrates were begun for the purpose of determining the relationship which might be supposed to exist between the elimination of sugar and the concentration of sugar in the blood. In order to throw additional light on the character of the carbohydrates present in urine and blood we have determined the sugar both before and after hydrolysis. This procedure appeared to us particularly important when the ingested carbohydrate is more complex than glucose. Even when the ingested carbohydrate consists of a simple hexose the hydrolysis might furnish some supplementary information as to the mooted questions of how much of the blood sugar, or the sugar in non- diabetic urines, can be spoken of as glucose. The conflicting results obtained by means of fermentation tests have not proved particularly illuminating on this point, and it has seemed to us doubtful how dependable fermentation results can be in connection with sugars so small in amount as those present in the urine. Indeed, it seems to us doubtful whether the presence or absence of glucose in normal urine can be settled by chemical tests applied to the urine alone, and we hoped to solve that problem on the basis of indirect evidence. In nearly all of our experiments the blood sugar has been deter- mined in plasma as well as in whole blood, and was also determined both before and after hydrolysis. The volume per cent of the corpuscles was determined by the hematocrit method of 26 Ege, R., Biochem. Z., 1920, evii, 246. Hagedorn, H. C., Biochem. Z., evii, 248. Gad-Andresen, K. L., Biochem. Z., evii, 250. Warburg, E. J., Biochem. Z., evii, 252. 228 Carbohydrates Hedin. While theoretical objections can be raised against the absolute values obtained by this method, it certainly does give consistent figures. The sugar content of the corpuscles has been obtained by calculation. The change in the amount of sugar found in whole blood, plasma, and corpuscles after hydrolysis is shown in our tables by the plus or minus numbers accompanying the figures obtained before hydrolysis. Much attention was given to the quick separation of some per- fectly clear plasma. The blood was collected directly from the needle inserted in the cubital vein, in carefully paraffined centri- fuge tubes, and the separation was finished in from 2 to 4 minutes after the tube was filled with the blood. 4 minutes may be given as the regular time, but this time was not used until we had thor- oughly satisfied ourselves that it gave the same results as the shorter period. No oxalate or anything else, except the paraffin, was used to prevent clotting during the centrifuging. Clotting has at no time entered into the results, but minute amounts of potassium oxalate, about 15 mg. per 10 ec. of blood or plasma, were added before diluting with water. The “post mortem” diffusion between plasma and corpuscles, to which Falta ascribes positive findings for sugar in the corpuscles, is, we believe, abso- . lutely excluded in our experiments. Within 15 to 20 seconds after the blood was drawn a part of it was beginning to rotate in a powerful centrifuge, and in less than 8 minutes both the whole blood and the plasma had been precipitated by the sodium tung- state method of Folin and Wu.27 We, therefore, cannot entertain the slightest doubt about the fact that our analyses reveal the true distribution of the sugar between the plasma and the corpuscles. No attempt was made to prescribe any uniform time interval for the passing of the urine, but the time of each urination was noted, and the amount of sugar and urine obtained was calculated as so much per hour. . *7 Folin, O., and Wu, H., J. Biol. Chem., 1919, xxxviii, 81. For the precipitation of the plasma proteins we use only half a volume of 10 per cent sodium tungstate and of 2 normal sulfuric acid. Careful prelimin- ary investigations had shown that these amounts precipitated all of the protein and gave filtrates suitable for all the determinations included in the system of blood analysis of Folin and Wu. a aE ee ze O. Folin and H. Berglund 229 With reference to the analytical methods used, there is need for but little discussion. The blood sugar was determined by the method of Folin and Wu,?8 both before and after hydrolysis. For the hydrolysis 2 ec. of blood or plasma filtrate were acidified with 3 drops of 10 per cent hydrochloric acid, and the mixture (in the blood sugar tube) was heated in boiling water for 75 minutes. The added acid was then neutralized with 3 drops of a sodium hydroxide solution which was equivalent to the hydrochloric acid. The method used for the sugar determinations in urine is given in the preceding paper. In experimental work of the kind involved here it was inevitable that many experiments should prove disappointing in the sense of not yielding results of the kind expected, or indeed any information of visible consequence. In the interest of space we have omitted many such; we are, in fact, giving only a selected series of results, but we have not omitted any experiments merely in order to diminish apparent contradictions. Experiments with Glucose, Maltose, Dextrin, and Starch. The marked increase in the sugar of normal urine which comes after every carbohydrate meal, and to which Benedict has given the name glycuresis, would seem to indicate that there is some glucose excretion below the blood sugar threshold. In order to discover the extent to which this is the case we have made a large number of feeding experiments with pure glucose. Pure sugars, particularly glucose, should be more effective than starches in raising the level of the blood sugar, and in our experiments with this sugar we hoped also to get further data on the threshold. From the very beginning the results obtained have been quite different from what a careful perusal of the literature had led us to expect. In Tables I to V are given the essential data of five typical experi- ments with not less than 200 gm. of glucose. The sugar was taken on a fasting stomach in the forenoon after a morning sample of the blood and urine had been obtained. The fluctuations of the sugar in the blood and urine were followed for several hours. The sugar of the urine is given in milligrams per hour, but the 28 Folin, O., and Wu, H., J. Biol. Chem., 1920, xli, 367. 230 Carbohydrates volume of the urine expressed in cubic centimeters per hour is also given so that the percentage of sugar can be calculated if desired. The percentage is, however, of little importance exceptin relation to the question as to whether the urines would give positive qualita- TABLE I.* Subject D-n. Age 22 years. Weight 75 kilos. 200 gm. glucose taken. Result: Maximum subthreshold hyperglycemia but no glycuresis. Time. | Blood. Urine. | Remarks. ~ Sugar per 100 cc. l Nol= Basar | 2, + ij < at Whole | Cor- puscles. ‘per ian | pane blood. | Plasma. puscles. hour. a.m. mg. mg. mg. atl? ce. mg. 10.12 | 250 ec. water taken. Urine voided. 11.55 |105+0 |107—9 |102+12) 43 p.m. 12.00 50 | 21+3 12.05 200 gm. glucose taken in 830 cc. | | water. 12.50 |152+4 |172—4 |122417| 41 1.30 111 | 18+2 1.50 |121—6 |127—13/112+4 41 2.30 35 | 238+4 3.00 |186—21/143—13)127—34| 41 4.00 29 | 28+5 4.35 105—3 |105—7 |105+-2 42 ! 5.30 27 | 21+2 6.05 | 95+2 101—4 | 88+9 6.50 | | 76 | 224-4 | 7.00 Dinner. 9.25 _ 111 | 65+21; Note glycuresis | after meal! . . ¥ . *Small, plus or minus, figures show increase or decrease after hydrolysis. tive tests for sugar. The sugar excretion per hour is quite inde- pendent of the volume of the urine. 5 to 7 hours after the ingestion of the glucose, each subject was allowed to eat an ordinary mixed evening meal. A record of what he ate was kept, but these details were found to be of no impor- 4 QO. Folin and H. Berglund 2a1 tance and are therefore omitted. The tables are arranged in the order of the highest blood sugar found, the highest coming first. From these tables it will be seen that the highest blood sugar * level attained was 152 mg. per 100 cc. of whole blood and 172 mg. TABLE II.* Subject T. W. W-r. Age 24 years. Weight 69 kilos. 200 gm. glucose taken. Result: Maximum subthreshold hyperglycemia but no glycuresis. Time. Blood. Urine. Remarks. Sugar per 100 cc. | Vol- Cor- ume puscles.| per Sugar ner Fasting. Urinated and took 200 ce. water at 9 a.m. — Plasma. | ed hour. hour. a.m mg. mq ated ce mg. 10.30 | 93+-11) 9643 | 89422) 41 | 10.46 238 | 20+6 10.50 | Took 200 gm. glu- 11.25 |188+4 |166— 24 100-442 43 | | cose in 600 ee. water. 11.46 100 | 33+4 p.m. 12.20} 90+15 12.58 125 | 35+8 1.30 | 54+9 | 58+4 | 47+19| 38 1.58 25 | 32411) 3.35 |100+2 |106+6 | 90-4) 38 4.12 104 | 30+6 5.10 142 | 23+5 5.10 Dinner. 7.05 | 92+2 |102—12) 78+22| 40 Note glycuresis after meal! vice 366 | 43+6 10.00 52 | 32+8 a.m. | 6.00 ee March 9. ee aed show inercana or decrease after hydrolysis. plus or minus, figures show increase or decrease after hydrolysis. in the plasma. We have ten other similar experiments with as many different normal persons (students), but the average of these is about the same as for the five given, and none reached as high a figure for the blood sugar as that shown in Table I. 232 Carbohydrates In not a single normal case, therefore, have we been able to obtain a blood sugar high enough to give the alimentary glyco- suria which comes when the glucose threshold is reached. Further discussion of this somewhat unexpected result will be given in a - TABLE III.* Subject P. A. Ch-r. Age 24 years. Weight 60 kilos. 215 gm. glucose taken. Result: Maximum subthreshold hyperglycemia but no glycuresis. Time. Blood. Urine. ; Remarks. maps | | cop | ume | SSE) Beto Croat and, © | BRO | Plasma. | cis, [P| our. | BOUr pm rivige a.m. mg. mg. mg. ae ce mg. 10.20 | 90+9 | 96+4 | 81+17 10.29 66 22 10.33 | | Took 215 gm. glu- cose in 500 ce. water. 11.20 |138+14/154+3 |116+-28 11.33 140 26 p.m. 12.17 | 9442 |110—5 | 72412) 43 12.39 51 23 1.45 | 87—5 |100—0 | 70—12 1.53 22 19 4.45] 85+5 | 92—8 | 77421 5.00 31 19 els Dinner. 7.40 |125—2 |182—14116+14 8.03 | 23 57 | Note glycuresis after meal! 10.35 38 38 * Small, plus or minus, figures show increase or decrease after hydrolysis. subsequent section, but we would have the reader recall that A. E. Taylor®® some years ago also met with uniform failure in attempting to produce alimentary glycosuria in normal persons (students) by feeding pure glucose. Taylor evidently was not 29 Taylor, A. E., and Hulton, F., J. Biol. Chem., 1916, xxv, 173. O. Folin and H. Berglund 233 aware of the transient character of the hyperglycemia produced by sugar, for he examined only 24 hour urines. The giving of 200 gm. of pure glucose not only failed to produce the glycosuria which accompanies the blood sugar threshold, but the tables show that the hyperglycemias which we did obtain are TABLE IV. Subject J. D. C-d. Age 22 years. Weight 61 kilos. 200 gm. glucose taken. Result: Maximum subthreshold hyperglycemia but no glycuresis. Time. Blood. Urine. Remarks. og _ eel = Les shag gr parece at 9.00 a.m. Whole ae om puscles. eal hour. ook 200 cc. water. a.m. mg. mg. mg. P baeie ce. mg. 10.35 90 94 85 | 43 10.43 252 21 10.45 Took 200 gm. glucose in 5990 cc. water. 11.30} 118} 130| 102| 438 11.40 61 21° p.m. 12.22} 102) 101 104) 41 1.15 98 26 1.45 94; 102 83 | 41 2.50 52 20 4.30 78 81 75 | 43 5.10 27 17 5.30 Dinner. 7.30} 134] 142] 123 7.40 25 33 a.m. 12.45 32 38 wholly without effect on those lower levels of sugar excretion comprised in the term glycuresis. This is manifestly important. Since the sugar of normal urine is quite independent of the level of the blood sugar it must be considered exceedingly doubtful whether the sugar of normal urine is glucose, quite independently of whether it does or does not ferment. But if that sugar is not 234 Carbohydrates glucose it is difficult or impossible to assume that it has any con- nection with such problems as carbohydrate tolerance, limits of glucose ‘assimilation, or with any ‘‘diabetic tendency” or “pre- diabetic stage.’”” The absence of any unmistakable glycuresis following the ingestion of 200 gm. of glucose is also important in that zt proves the concept comprised in the term glucose threshold to TABLE V.* Subject H. B-d. Age34years. Weight 90kilos. 200 gm. glucose taken. Result: Maximum subthreshold hyperglycemia but no glycuresis. Time. Blood. Urine. | Remarks. Ss ec. Tale ei 2 ane per ae Mane wees ere Urinated 2 | Ligse pees eee | cles pe | hour. | yt) Pm ie? vol. p.m. mq. per ce mq | cent 1.25) | t 39 | 2047 2.05 } 84+10) 93+3 | 71420) 41 3.00 | 30) 17-42 3.05 Took 200 gm. glucose in 1,000 ce. water 3.40 |116+11/129—2 | 97+30) 42 | 4.00 | 44 | 19-4 5.00| 70+5 | 70+1 | 70+10 2 | | 5.30 80 | 23+0 6.30) 57-3 | | 7.00 | 30} 2145 9.00 | | Protein-fat meal. a.m. 12.45 | | 32 26+6 | No glycuresis. 9.00 | 30 | 23+4 | Following morning. | | * Small, plus or minus, figures show increase or decrease after hydrolysis. be not something only approximately true; the concept 1s absolutely correct, however uncertain the exact figures given for the threshold may be. Hyperglycemia definitely below the threshold does not normally produce the slightest leakage of glucose through the kidneys, and normally not a trace of absorbed and circulating glucose ts lost. The striking difference between pure glucose and ordinary mixed meals in their effects on the sugar excretion is shown in each of the rik hid ©. Folin and H. Berglund 235 five experiments recorded in Tables I to V. The definite unmis- takable glycuresis described by Benedict is obtained after every carbohydrate meal, but not after meals containing no carbohy- drate (Table V). By hydrolysis the sugar of the urine is increased. The sugar of urine is, therefore, made up, at least in part, of di- or poly- saccharides, for we have satisfied ourselves that the increase after hydrolysis cannot be explained on the basis of glucuronic acids. The conditions under which glycuresis is obtained and the increase after hydrolysis suggested the possibility that the sugar of normal urine is due to a little premature absorption of incom- pletely digested carbohydrate material, such as maltose and dextrins. Two experiments were made with maltose, but only one is recorded here (Table VI). The maltose was a Kahlbaum prep- aration but was not absolutely pure. It contained enough dextrin to give a faint but unmistakable reaction with iodine. From the figures given in Table VI it will be seen that 200 gm. of maltose have had extremely little effect on the level of the blood sugar, and like glucose has no effect on the sugar of the urine. The urine passed at 1.07 p.m., 62 minutes after the maltose ingestion, evidently contained a trace of ‘“‘dextrin.”” It may seem rather extraordinary that any evidence of dextrin in the urine should be obtainable from the small amount of such polysaccha- rides present in the maltose, but the significance of the phenom- enon is clear enough from the results which we have obtained from our feeding experiments with large quantities of dextrin. Three experiments were made with dextrin with the subject H. B-d. The results obtained from these experiments seemed to furnish an adequate explanation of the origin of the sugar in normal urine and determined the trend of many subsequent experi- ments, but we discovered after a while that our original interpre- tation was erroneous, and the significance of the dextrin results was considerably diminished. We omit, therefore, one of the experiments in which 200 gm. of dextrin were taken. Table VII gives the results obtained from 350 gm. of dextrin. The dextrin was taken in two doses about 2 hours apart. The subject suffered no discomfort from the large amount of dextrin taken and enjoyed his evening meal as well as the meals taken 236 Carbohydrates during the subsequent days. There was no reason to doubt that the dextrin was digested and absorbed very much as an equivalent quantity of other carbohydrate food. From all this dextrin we obtained no increase in the level of the blood sugar and no increase TABLE VI.* Subject H. B-d. Age34years. Weight90kilos. 200 gm. maltose taken. Result: No glycuresis. Time. Blood. | Urine. Remarks. a Sugar per 100 cc. Vol- mi ey 9, faa Oar: and i ors ie pee 21. . }puscles.} e t 10. .m. essay Plasma. em res ae hour c =; a.m mq. | mg. | mg. ee cc. mg. 11.45 | 90—0) 90—4/90+5 | 41 p.m. 12.00 | 20 |20+7 12.05 Took 200 gm. maltose (Kahlbaum) in 600 ce. water. 12.30 | 80+4)/108—6/36+20; 39 1.07 70 |25+16 1.20} 78+7 1.41 42 |21+7 2.22) 69+8) 73—0/63+20) 38 2.45 26 |22+6 4.05 | 75+7| 78—5|70+26) 39 4.33 24 |22+6 21 |19+4 | Dinner at 6.45 p. m. 26 |39+11) Note glycuresis after a.m. | meal! 10.12 51 |43+9 | Second day. p.m. 12.04 24 |22+2) Fasting. * Small, plus or minus, figures show increase or decrease after hydrolysis. of the preformed sugar of the urine. The urine contained, how- ever, an abundance of dextrin. This was our second dextrin experiment, and its main purpose was to see how long the dex- trin excretion continued, for we had found in the first experiment that the excretion was still rising at the end of 10 hours. And, as may be seen from the table, the dextrin excretion is at its height O. Folin and H. Berglund 237 in the morning of the day following the dextrin ingestion. The elimination of dextrin continued for 4 days. There can, there- fore, scarcely be any doubt about the interpretation. Some dextrin or dextrin-like product had been absorbed from the diges- TABLE VII.* Illustrating the Slow Excretion of Absorbed Dezirin. Subject H. B-d. Age 34 years. Weight 90 kilos. Time. : Blood. | Urine. Remarks. Besar pe 100 eo. Vol- Sugar | Fasting. Drank 200 cc. water Jan. 22 Cor- | ume | t 11.45 a.m. 1921 ae inca, eee puscles. ae ‘oa, and fain at 12.40 p.m. Pe cone mu, mg Sst ae mig: 12.50 20 2414 | Preceding urine passed 9 a.m. 3.10 |101+5109—5 89+20) 38 3.35 24 |234+2 | Took 200 gm. dextrin in 800 cc. water. 4.18 |101+0,100—3102+4 | 41 4.52 195 \25+54 | Marked diuresis. 4.55 | Took 150 gm. dextrin in 500 cc. water. 5.20 460 \28+-137| Diuresis. 5.35 | 87—2| 88+2) 86-9 | 39 5.50 580 '34+170| Diuresis. 7.37 67 |29+176 7.50| 83+2| 95—1| 6647) 41 8.52 42 \30+173 a.m. ; 9.00 30 |35+185| Second day. 10.00 28 \44150 | Third day. 9.00 32 |38+28 | Fourth day. eee ee eee * Small, plus or minus, figures show increase or decrease of sugar after hydrolysis. tive tract which after absorption behaved entirely like a foreign unusable product which slowly found its way into the urine. The absorbed material did not stay in the blood, for if it had been there our sugar determinations after hydrolysis would have revealed its presence. It must, therefore, have been absorbed by the tissues and was later gradually released and eliminated. 238 Carbohydrates In order to be absolutely sure that the long continued dextrin excretion was not due to some unusual delay in the absorption from the intestine, another experiment was made with a single dose of 200 gm. of dextrin followed 2 hours afterwards by a large dose (3 tablespoons) of castor oil. The bowel was thus thoroughly cleaned out in the course of about 5 hours. The course of the dextrin elimination is shown in Table VIII. TABLE VIIL.* 2 Showing that Slow *‘Dextrin’’ Excretion Does Not Depend on Slow Absorption. Subject H. B-d. Age 34 years. Weight 90 kilos. Time. Saeioa Th eee Remarks. a.m. ce. mg. 11.23 32 | 2549 Fasting. 11.32 Took 200 gm. of dextrin in p.m. 600 cc. water. 12.18 48 27+51 1.20 37 28 +150 1-25 Took 45 gm. of castor oil. 2.24 50 | 81+201 3.44 179 40+ 250 4.13 93 | 27+-156 5.01 34 | 244133 7.12 33 | 25+118 a.m. | ey) 28 | 22+75 Second day. 9.50 18 | 17+38 11225 15 | 174-24 Fasted throughout the experi- | ment. * Small plus figures show increase after hydrolysis. The experiments with dextrin furnished the clue to the origin of sugar in normal urine, but our assumption that the dextrin excretion was due to the fact that we had overtaxed the digestive mechanism and that this was the reason for the absorption of unchanged dextrin was not correct. At least we do not now think that premature absorption of digestible dextrin has anything to do with it. We would recall in this connection the fact that we could detect the presence of some dextrin-like substance in one of the urines from the maltose experiment. O. Folin and H. Berglund 239 Our main reason for believing that the polysaccharide excretion after taking dextrin is due to the presence of some denatured, indigestible, and unusable carbohydrate rather than to the escape of real dextrin is based on the results obtained with pure starch. The commercial preparation of starch differs from the processes used in the preparation of dextrin and sugars in that practically TABLE IX.* Showing Absence of Glycuresis after 175 Gm. Pure Raw Potato Starch. Subject H. B-d. Age 34 years. Weight 90 kilos. Time. Blood. | Urine. Remarks. : s 100 ce. Vol- — eres i, | oe ae | Hee | eee quniated 2 7) | | S a = a b a.m. wpa Plasma. | Pn ragoaasia hain | hour / i em fom fm | m fica = | we 11.20) 93—4 | 98—13 85-10) 39 11.30 | | 30 | 30+10 11.55- Took 175 gm. raw 12.25 | potato starch in | 600 ce. water. p.m. 1.00 130 | 25+9 1.12} 83—5 | 96—26) 63+-27 39 1.23 625 | 33414 2.00 | 70+9 | 983-3 | 37+27 41 222 | 58 | 21+7 3.25 | 6849 | 81-22) 47458) 39 | | 3.57 | 33 | 21+10 4.44) 80+1 | 79—9 | 81+17 aD or iw) s +- co 28 * Small, plus or minus, figures show increase or decrease after hydrolysis. no heat and no vigorous chemical reagents are used in the manu- facture of starch. There is thus good reason to believe that starch is free from decomposition products and free from soluble impurities of a carbohydrate nature. Whether the carbohydrate elimination after taking dextrin is or is not due to impurities is of minor importance, for it is, after all, on the basis of pure starch and the sugars, maltose and dex- THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI, NO. 1. 240 Carbohydrates trose, that one can definitely determine whether the normal sugar excretion after meals is a feature of our main carbohydrate metab- olism or only a side issue due to the presence in our food of prod- ucts which from the standpoint of carbohydrate metabolism can be regarded as impurities. Two experiments with starch are recorded in Tables IX and X. The starch (175 gm.) was first taken in a raw, uncooked, condition. 25 gm. of starch were first converted into a thin paste by means of hot water and into this was stirred the remaining starch (150 gm.). The mixture was eaten with a spoon and was very difficult to take. TABLE X.* Showing Absence of Glycuresis after 200 Gm. Pure Dextrin-Starch. Subject H. B-d. Age 34 years. Weight 90 kilos. Time. Urine. Remarks, Nov. 28, 1921. Bie Beso Fasting. Urinated at 9.39 a.m. a.m. ce. mg. 10.15 33 23+-11 10.40-10.50 Took 200 gm. dextrin-starch in 900 cc. water. 11.30 100 25+11 p.m. 1227 137 29+-4 1.40 37 24+-10 3.36 35 22+9 * Small plus figures show increase after hydrolysis. The analytical results given in Table [IX show that from raw (potato) starch one does not obtain the glycuresis which is so prominent after ordinary mixed carbohydrate meals. The blood sugar level did not rise appreciably, as was to be expected from the low levels which we have obtained from the sugars and from dextrin. The experiment with raw starch was not deemed conclusive. While there was nothing to indicate that the starch was not com- pletely digested and assimilated, the digestive process may have been so slow that the intermediate digestion products were im- mediately and completely converted into maltose and dextrose, thus eliminating the chance for any premature absorption of a Cian pag IN NR = ome eee eee O. Folin and H. Berglund 241 intermediate products. We, accordingly, in the next experiment (Table X) first boiled 200 gm. of the starch with 900 cc. of water and then treated the paste at 70°C. with a little freshly prepared aqueous extract of malt. This diastase treatment was very mild, barely enough to soften the mixture so that it could be taken with- out the use of a spoon, but considerable dextrin was, of course, produced. This cooked and partially digested starch failed to produce glycuresis and also failed to yield any dextrin in the urine. Before giving a more detailed discussion of the origin and character of the sugar of normal urine it is necessary to present some additional feeding experiments with sugars, for these have a more positive bearing on the problem than did the experiments with glucose and maltose. Experiments with Fructose. The feeding of fructose should give substantially the same results as those obtained with glucose, if it is true that fructose is converted into glucose in the liver and normally never gets beyond that organ. The assumption that fructose does not escape the liver is based on the fact that neither normal nor diabetic urines contain fructose even when large quantities of fructose or cane- sugar are taken. The fructose test for a much reduced efficiency on the part of the liver rests on the same assumption. The assimi- lation limit for fructose is also believed to be substantially the same as for glucose. In Table XI are given the essential data on blood and urine obtained after taking 200 gm. of fructose. As far as the whole blood is concerned it will be observed that its sugar content has not increased. From the plasma sugar values we find, however, that there is here a slight increase (8 mg.), an increase which is hidden in the whole blood analysis by a nearly compensating de- crease in the sugar of the corpuscles. From the practically unchanged level of the blood sugar one should, of course, not expect to find any increase of the sugar in the urine, especially since none was obtained when the blood sugar did rise, as in the glucose experiments. A material glycuresis lasting for several hours was nevertheless obtained, and the negative results obtained with Seliwanoff’s 242 Carbohydrates reaction showed that the sugar excreted could not be fructose. The urine passed at 5.43 p.m., 4 hours after the fructose ingestion, contained sugar equivalent to 0.25 per cent of hexose, and we could not have failed to get a positive reaction for levulose if even a small fraction of the sugar had been fructose. These data seemed to us beyond possibility of plausible explanation in terms of metabolism processes. TABLE XI.* Showing that 200 Gm. of Levulose Produces no Hyperglycemia and Does Give Rise to Some Unknown ‘‘Sugar”’ in the Urine. Subject H. B-d. Age34years. Weight 90 kilos. Time. Blood. Urine. Remarks. Sugar per 100 cc. | Vol- 1921. Cor sen a! Partie Tene voided y scles ol . ° Mead Plasma. eee ose fear hour. > sete p.m. mg. mg. mq. Bical cc. mg. 1.10 100 102 97 38 1.30 27 | 26+6 1.40 Took 200 gm. levulose in 900 cc. water. 2.10 102 110 90 38 2.40 63 40+1 2.50 101 103 99 39 4.05 36 42+2 4.25 98 99 96 36 5.43 22 | 67+0 5.45 94 98 86 36 8.10 93 95 90 8.35 36 | 40+3 * Small plus figures show increase after hydrolysis. It occurred to us that the true explanation might be the presence of reducing decomposition products of levulose. The sugar used was a very high grade preparation of Kahlbaum’s. Like all levulose preparations it had some color, and its concentrated solutions were yellow. Fructose is so unstable that it is very questionable whether any obtainable is entirely free from decom- position products. A very little of such reducing impurities (0.1 O. Folin and H. Berglund 243 per cent) would be adequate to account for the sugar found in the urine. The prolonged character of the ‘‘sugar’’ excretion pointed distinctly to some foreign unusable impurity as the cause of the glycuresis. Subsequent experiments made on this point have only served to confirm our conviction that we were here dealing with artificial products which have nothing to do with carbohydrate metabo- lism. How much such factors may have figured in all past feeding experiments with fructose one cannot tell. In the course of our investigation on the subject we heated 200 gm. of our pure levu- lose (dissolved in 300 ce. of water) at 145°C. for 1 hour. The levulose was so extensively destroyed that the sweet taste of this concentrated solution entirely disappeared. The resulting solu- tion was considerably darker than before the heating, but by no means black. The solution was diluted to a volume of 500 ce. and its reducing power was determined. This was still equiva- lent to about 70 per cent of the levulose taken. In view of the unknown and doubtful character of this solution we hesitated about having anyone take the whole. Our fructose subject, H. B-d., took a small sample and observed no unpleasant symp- tom of any kind. A little diarrhea followed, but as a meal had been taken in between, it was not ascribed to the levulose decom- position products. The following day H. B-d. therefore took enough of the solution to correspond to about 150 gm. of the levu- lose. In about 1 hour a most violent diarrhea began and lasted for several hours. Considerable nausea was also experienced, but no headache or other intoxication symptoms. We deem it im- portant to call attention to the vigorous laxative effects of decom- posed levulose solutions. Maple sugar, molasses, and certain candies doubtless owe their laxative effects to similar decomposi- tion products of sugar. Because of the unexpected and dramatic outcome of this experi- ment we unfortunately neglected to pursue the original purpose for which the experiment was made as long as we should have done. Several consecutive samples of urine were, however, collected, and the analytical results obtained from these certainly confirm our original conjecture that the sugar in the fructose urines was de- rived from decomposition products of fructose. The sugar elimi- nation, expressed in milligrams per hour, is shown in Table NII. 244 Carbohydrates At a later time H. B-d. took 200 gm. of Schering’s fructose. This sugar was in the form of a powder and looked white enough, but in solution it was much darker than similar solutions made from Kahlbaum’s levulose. The diarrhea which followed the taking of this product was almost as intense as that obtained from the levulose which had been decomposed in the autoclave. The main purpose of this experiment was to determine whether it is really true that fructose is completely retained and converted into glycogen in the liver. Since fructose has nowhere near the same effect as has glucose in raising the level of the blood sugar, TABLE XII.* Glycuresis after Decomposed Fructose. Subject H. B-d. Age34years. Weight 90 kilos. Urine volume Urine sugar Time. per hour. per hour. Remarks. a.m. mg. 10.45 28+9 Fasting. 10.45-11. 150 gm. decomposed fructose in p.m. 450 cc. water. 12.14 94+8 Diarrhea. 1.37 143+ 26 $ 2.49 126+39 ‘s 3.45 128+12 * Small plus figures show increase after hydrolysis. one is forced to assume that the transformation of fructose into glycogen within the liver is accomplished much more easily, and, therefore, more completely than is the manufacture of the same kind of glycogen from glucose. From the standpoint of chemical dynamics as well as from the standpoint of biological adaptation such a conclusion does not seem plausible. Levulose is fortunately a sugar for which we have a reasonably dependable and extremely sensitive test in Seliwanoff’s reaction. By means of preliminary experiments we satisfied ourselves that we could detect levulose with perfect certainty in blood plasma if as much as 5 mg. per 100 cc. were present. The proteins of blood plasma can be precipitated by the sodium tungstate process without any preliminary dilution, and a filtrate can thus be obtained, the dilution of which is only double that of the original plasma. A O. Folin and H. Berglund 245 strictly fresh Seliwanoff’s reagent containing 3 mg. of resorcinol per ec. of concentrated hydrochloric acid was used, and.2 ce. of this were added to 4 ce. of the blood filtrate—all for the purpose of eliminating avoidable dilutions. The blood for this test was taken 20 minutes after the ingestion of the levulose. The control blood was absolutely free from any substance giving the levulose reaction, and the test for levulose in the plasma representing a 20 minute absorption of levulose was unmistakably positive. On the basis of these tests we affirm that levulose up to at least 5 mg. per 100 cc. of plasma may be present in blood at the height of active levulose absorption from the digestive tract. The blood sugar in the plasma was 98 mg. before and 112 mg. 20 minutes after the ingestion. At least one-third of the increase was, there- fore, represented by levulose. The urine, as in other levulose experiments, contained enough sugar (reducing substances) to give positive clinical tests, yet we could not satisfy ourselves that any levulose was present. And a doubtful Seliwanoff’s test applied to urine must necessarily be called negative. Our own interpretation of the fate of absorbed fructose is as follows: The liver retains fructose as well as every other usable sugar to a greater extent in proportion to its weight than do the general tissues such as the muscles. But such retentions by the liver are never even approximately quantitative, and a large fraction of absorbed sugars, possibly the greater part, gets by this organ. Other tissues, such as the muscles, take up sugar from the plasma of arterial blood, and it is this general absorption which prevents excessive accumulations of sugar in the blood. But tissue sugar like blood sugar is normally and predominantly glu- cose, partly because the major part of our carbohydrate food is made up of glucose, partly because all other usable sugars are gradually converted into this essential sugar. The tissues being relatively well stored with glucose and empty of other sugars, such as fructose, may well be able to absorb these other sugars from the blood so nearly completely that the venous blood used for our analyses shows only traces. The glycogen formation may or may not begin immediately, and at all events need not be the immediate cause for the rapid disappearance of levulose from the blood. Levulose happens to be an excellent glycogen former, but that this is not the immediate cause of its disappearance from the 246 Carbohydrates blood is strongly indicated by the results which we have obtained with a much poorer glycogen former—galactose. Before taking up the discussion of our galactose and lactose experiments, we would call attention to one other important con- clusion which we think can be drawn from our finding of levulose in the blood. Strictly speaking, there is only one substance for which there has been any reason to assume a renal threshold. The renal threshold for glucose is an abundantly established fact, and no substantial evidence for the existence of analogous thresholds for other materials has as yet been produced. Other alleged thresholds are based mostly on speculations. It would seem, how- ever, that with reference to levulose we may have something corre- sponding to the threshold for glucose. Unless such a threshold for levulose came into play it is difficult to see why there should not have been levulose in the urine corresponding to the period of . its proved presence in the blood plasma. The levulose should have been far more abundant than the reducing impurities which so readily found their way into the urine. The existence of a levu- lose threshold similar to that for glucose would serve to explain the occasional (though very rare) occurrence of fructosuria. The elimination of fructose in such cases would correspond to the elimination of glucose in renal glycosuria, and is caused by a lower- ing of the threshold for fructose. Fructosuria on the basis of this explanation would in most cases be a comparatively mild disorder, whereas without this explanation the elimination of fructose would have to mean a profound disturbance in the metabolism of carbo- hydrates, more or less analogous to diabetes mellitus. Our facts and interpretations concerning the absorption and transportation of fructose suggest the possibility of making more extensive use of fructose in diabetes than has been possible in the past. It is impossible to tell how far past failures in the use of fructose may have been due to impurities, for they could doubt- less bring about very serious upsets in the general condition of the patient. There is room for the suspicion that the early favorable impressions obtained as to the effects of fructose were based on the use of the purest obtainable brands, and that later, from economic considerations, cheaper brands loaded with impurities came into use. Continuous use of such would necessarily bring trouble. O. Folin and H. Berglund 247 The tissues of the diabetic person should contain much higher concentrations of glucose than tissues of normal persons, and this is the immediate reason why the blood sugar rises so high after the intake of glucose in any form. The high concentration of glucose in the tissues would probably have little or no effect on their absorption of fructose. From the giving of fructose we should, therefore, get a much higher concentration of total sugar (glucose plus fructose) in the tissues without any material increase in the sugar of the blood. Because of this higher concentration an increased utilization might well take place. Continuous or excessive use of fructose would, of course, defeat the purpose of the treatment, because gradually this would become equivalent to the loading of all the tissues with glucose, and, because of the large bulk of the tissues, an excessive amount of glucose would pour into the blood and be eliminated with the urine. Such a series of results has, however, no bearing on what would happen from small doses of pure fructose given at carefully regulated intervals. For under such conditions there is not only a larger concentration of sugar in the tissues in relation to the level of the blood sugar, but there is also the possibility that nascent glucose is more easily utilized than ordinary, preformed glucose. At least it seems safe to say that until the effects of pure fructose on diabetics have been carefully investigated from this standpoint the subject has not been exhausted. Experiments with Galactose and Lactose. All investigators who have made experiments with galactose are agreed that the assimilation limits for this sugar are entirely different from the assimilation limits for glucose and fructose. 20 to 40 gm. are usually given as the largest amounts that can be taken at once without resulting in glycosuria. The limits of assimilation found by any particular investigation would depend to a considerable extent on the analytical methods used in proving the presence or absence of glycosuria, or rather of galacto- suria. Our own investigations based on quantitative sugar determina- tions and planned to reveal the origin of the sugars found in normal utine have, of course, shown glycuresis at much lower levels of galactose (and lactose) ingestions than those based on qualita- 248 Carbohydrates tive tests for sugar. Our data bearing on this point are given in Table XIII. The finding of minute increases in the sugar of the urine after small doses of galactose has been to us a matter of minor signifi- cance in this particular case. It is the level of the blood sugar which from our point of view was a matter of fundamental im- portance. Galactose differs from levulose in that the former TABLE XIII.* Showing the Glycuresis, in Milligrams per Hour, Due to Galactose and Lactose. All except the second experiment with 30 gm. galactose on Subject H. B-d. and that one on McC-n., the renal glycosuria subject. Pure galactose. Pure lactose. Time after ee ingestion. 30 30 10 gm em. | em. 100 gm. 10 gm. | 30gm. | 45 gm. | 50gm.} 200gm. Before. |24+11} 22] 13 23+13 |26+8 |21+11)22+10/18+13) 19+8 lhour. |40+6 | 395 | 158 70+9 |61+-15 68+ 23/444 -14/111+ 23 13 hours. |43+9 | 153 538+46 |33+-10/53+-20/80+-31 252+-148 2 C 284 | & 29+-12/42+-21 24 = 2248] 23) 37 23++8 40+ 16\63-+- 24 3 5 24 |3,020—540 25+15 26+ 19)266+ 130 =) 17 4 a 21+-7 5 3 452—12 27+-14 175+99 7 a 33-20 76+61 8 ? 25+-17 21+9 * Small, plus or minus, figures show increase or decrease after hydrolysis. is assimilated very imperfectly, whereas the assimilation limits for the latter are so high that they can scarcely be exceeded. There is probably no difference of opinion about the cause of these widely separated assimilation limits. Fructose is on a par with glucose as a glycogen former; galactose is not. But if the glycogen for- mation in the liver were the chief factor by which the accumula- tion of absorbed sugar in the blood is prevented, then galactose should be immensely more effective even than glucose in raising the level of the sugar in the blood. 200 gm. of glucose will not yield a trace of sugar in the urine, whereas 100 gm. of galactose may yield as much as 10 gm. of excreted sugar. ©. Folin and H. Berglund 249 Since a sugar excretion amounting to as much as 10 gm. has been obtained from 100 gm. of galactose it did not seem worth while to take any more, especially as we know from glucose experi- ments, that 100 gm. produce just as high a hyperglycemia as does twice that quantity. The results obtained are given in Table XIV. TABLE XIv.* Showing that 100 Gm. Galactose Produces no Hyperglycemia but Does Give Rise to a Marked Galactosuria. Subject H. B-d. Age 34years. Weight 90 kilos. Time. Blood. Urine. Remarks. Sugar per 100 cc. Vol- 0 SS ae eg eed ieee Oa al er ee ee Plasma. Baie. our. p.m mg mg. | mg Bee, mg. 12.15 93 91 96 41 12.24 33 23+13 12:27 Took 100 gm. pure galactose in 300 12.35} 100; 101 98 | 43 cc. of water. 127.52 89 89 89 | 42 1.15 33 70+9 | Qualitative sugar test positive. 1.23 87 88 86 | 42 2.00 32 | 538+46 | Qualitative sugar test positive. 3.28 64 |3,020—540) Qualitative sugar test positive. 5.24 36 | 452—12 | Qualitative sugar test positive. 6.50 31 33-+-20 | After 6.50 p. m. free water intake. 8.11 119 25+-17 Total extra sugar excreted 5,685 mg. * Small, plus or minus, figures show increase or decrease after hydrolysis. The change in the level of the blood sugar is just about the same as that obtained from levulose. The transient maximum increase is 10 mg. per 100 cc. of plasma, yet in the course of 6 hours the sugar excretion amounted to 5.7 gm. 250 Carbohydrates If to these data we add the observation that 10 gm. of galactose are sufficient to produce unmistakable glycuresis it is difficult to see how glycogen formation in the liver can possibly be the factor which prevented accumulation of sugar in the blood. On the basis of our explanation—absorption of sugars by the tissues—the observed facts are freed from inexplicable contradictions. Nor- mally there is neither fructose nor galactose in the tissues, and within this enormous empty reservoir the galactose disappears as readily as does fructose and more readily than glucose. With respect to absorption and distribution galactose and fructose are alike. As far as excretion is concerned, the difference between the two is not even of a quantitative character. One is, the other is not excreted. Such a qualitative difference in the matter of excretion cannot very well depend on quantitative differences in the speed with which the two sugars are transformed into glycogen. The sugar excretion observed after very small doses of galactose clearly points to the absence of a renal threshold for galactose. In the absence of such a threshold galactose will continue to be ex- creted as long as there is any galactose in the tissues, for a little of it will constantly leak back into the blood. The excretion would be like that of absorbed dextrin except for the fact that even galactose is gradually converted into glycogen thus diminishing and finally exhausting the supply of galactose. In a later section (page 270) we call attention to the diminution of the blood sugar obtained after hydrolysis. A similar diminution has in cer- tain cases been found in the urine also. An extraordinary diminution was found in the galactose experiment. We have no explanation to offer, but are certain that the finding does not depend on analytical errors. In connection with galactose we venture to make the same sort of a suggestion concerning its possible value in diabetes that we did for fructose (page 246). The peculiar possible merit of the galactose in this connection is that it is only very gradually con- verted into dextrose and glycogen. In this very slow transition ‘we might have within the tissues a source for a relatively prolonged supply of nascent glucose. And if only 10 gm. were given at a time, the loss through the urine of unchanged galactose would be insignificant, notwithstanding the fact that it appears to be without a threshold. eal O. Folin and H. Berglund 251 Almost all authorities assign a much higher limit of assimilation for milk-sugar than for galactose, 100 to 150 gm. for the former, against 20 to 40 gm. for the latter. It is certainly true that much more sugar is excreted from 100 gm. of galactose than from 200 gm. of lactose, but with respect to the assimilation limits based on quantitative sugar determinations and glycuresis there is scarcely any difference between the two. 10 gm. of milk-sugar, corre- sponding approximately to the amount taken with a single glass of milk, is sufficient to produce a temporary, but unmistakable increase in the sugar of the urine. The interesting fact about the sugar excretion after lactose ingestions is that lactose (as well as galactose) is present in the urine unless the amount taken is very small—less than 30 gm. This is instructive, for it indicates that there 1s no mechanism in the digestive tract for preventing the absorption of soluble but incompletely digested carbohydrate materials. If such are not absorbed it is only because the activity of the endocellular enzymes is capable of accom- plishing the hydrolysis during the transit of soluble carbohydrates through the mucous membrane of the intestine. In the case of lactose the endocellular hydrolytic activity is easily exceeded, and this disaccharide is absorbed. Lactose, as such, is supposed to be unusable, and it is customary to refer to the occasional occurrence of milk-sugar in the urine of pregnant women as proof of how com- pletely unassimilable milk-sugar really is. We do not think that the question of the assimilability of milk- sugar can yet be considered as definitely settled. The old experi- ments based on the injection of known quantities of milk-sugar into the blood and finding the whole of it in the urine can no longer be considered conclusive, because under such conditions the sugar in the urine would be increased by virtue of superimposed glyco- suria. Our own experiments have not furnished conclusive evi- dence, but the fact that the excretion of milk-sugar comes to a definite end in the course of a few hours suggests that a part of the absorbed lactose may be utilized. The question of whether or not some lactose is utilized seemed to us unimportant in com- parison with the problem of why milk-sugar taken by mouth results in less total sugar excretion than is obtained from the galactose corresponding to the same quantity of milk-sugar. 100 gm. of galactose will give twice as much sugar in the urine as 200 252 Carbohydrates gm. of lactose. Our attempts to solve this problem have yielded extraordinarily interesting results, but our investigation is not yet completed, and the statements and interpretations here given we explicitly consider only as a preliminary communication. When the hydrolysis of the milk-sugar during absorption from the intes- tinal tract is quantitative, as it presumably is in normal infants, TABLE XV.* Showing that 200 Gm. Lactose Produces no Hyperglycemia and Does Give Rise to Lactosuria and Galactosuria. Subject H. B-d. Age 34 years. Weight 90 kilos. Time. Blood. Urine. Remarks. Feb, 6 | Sugar per 100.ce. cor, wine Sue ae go: = a Viele Plas aM) Need hour. voided at 9.30 a.m. a.m. mg. mg. mg. ee ce. mg. 11.15 | 9444) 96—8/90+24) 38 11.37 26 | 24+6 11.40 Took 200 gm. lactose p.m. in 750 cc. water. 12.18 |101+3)107—9/92+21) 38 1.00 135 |160+52 1.23 | 88+6| 97+4/72+10| 36 2.00 56 |290+-46 3.08 | 64+5| 62—0/68+15} 36 4.00 29 |256+58 5.05 | 88+5) 90—3/71+19 5.50 21 |108+34) Dinner at 6.30 p. m. 8.10 | 883+-7| 86+1/78+16 9.00 25 | 66+34 11.05 29 | 50+10 a.m. 11.00 34 | 21+5 | Second day. * Small, plus or minus, figures show increase or decrease after hydrolysis. then the giving of milk-sugar is the same as the giving of equal quantities of galactose and glucose. The difference between the adult and the nursing infant with reference to the power of split- ting lactose can be eliminated by giving pure glucose and pure galactose—equal amounts of each. ' O. Folin and H. Berglund 253 Table XV gives a picture of the sugar in blood and urine as obtained from 200 gm. of lactose. Table XVI gives the corre- sponding data obtained from 100 gm. of galactose plus 100 gm. of glucose. ‘The sugar excretion is less than one-fifth as great from the mixture of the two sugars as from the milk-sugar and is less TABLE XVI. Showing that Galactose (100 Gm.) Is Almost Completely Retained if Ingested together with Glucose (100 Gm.). Subject H. B-d. Age 34years. Weight 90 kilos. Time. Blood. Urine. Remarks. Sugar per 100 cc. ae yet — fh en Br per asting. Urinated at 9.45 a.m. 1921 Whole Plas Cor- puscles.| per hour. blood. ma. yuscles. Our. vol. p.m mg ™). mg per cent ce mg. 12.00 94 97 90} 41 12.12 55 19 12.15 Took 100 gm. glucose and 100 gm. galac- ; tose in 600 cc. water. 12.30} 116| 113| 120] 39 12.51 79 72 87 | 44 1.05 79 84 72} 43 1.25 250 50 1.35 82 85 78 | 42 2.35 30 | 102 4.13 30 | 137 | After 4.13 p. m. free water intake. 5.08 224 71 6.21 54 24 Total extra sugar excreted 371 mg. than one-tenth as great as the excretion obtained from 100 gm. of galactose when taken alone. The figures are 0.37, 2.8, and 5.7 gm., respectively. The results given show that the extent to which galactose is retained and utilized by the human organism depends on the quan- tity of available glucose. The bearing of this finding on all past literature on the assimilation of galactose is clear. 254 Carbohydrates The discovery is also interesting in that it seems to supply a simple and clear explanation of why milk-sugar should be advan- tageous for the young notwithstanding the fact that it has seemed to be the least useful sugar for adults. Galactose ts needed, evidently in abundance, for the building of nerve tissue; and the | surplus galactose is easily utilized, because of the presence of the glucose. Possible practical applications of the knowledge thus acquired suggest themselves. Less milk-sugar and some pure glucose (or maltose) might be better than milk-sugar alone for infants whose rate of growth is subnormal or whose urine contains sugar. Galac- tose rather than milk might be advantageous for nursing mothers whose milk is low in lactose. One of us (H. B-d.) is particularly interested in these feeding problems and will report on them later. Nature and Origin of the Sugar of Normal Urine. From the experimental data given in the preceding sections there can be little or no room for doubt as to the origin of the sugars of normal urine, including the glycuresis discovered by S. R. Benedict. These sugars must consist of a considerable variety of reducing carbohydrates other than dextrose and fruc- tose. A small portion may be derived from the sugar of milk. In the main they are derived from foreign, partly or wholly unus- able, carbohydrate materials present in grains, vegetables, and fruits. Artificial decomposition products due to the heat used in cooking, canning, and baking probably contribute a considerable share. Baking and toasting should produce decomposition prod- ucts similar to those encountered in dextrin. And wherever fructose or fructose compounds (cane-sugar, raffinose) are present, laxative decomposition products of fructose may be expected. All sweet preserves and many canned vegetables should contain such decomposition products. Pentoses and pentose compounds are doubtless represented, at least when fruit is eaten. The escape of such miscellaneous unusable carbohydrate ma- terial into the urine has no connection with the main carbohydrate metabolism, and, therefore, none with the various problems of diabetes, except with reference to the correct interpretation of weak but positive tests for sugar in the urine. These foreign glycuresis products may sometimes be abundant enough to cause lain O. Folin and H. Berglund 255 confusion or to give positive clinical tests for sugar. It is clear that one cannot expect to find any sugar reagent which will en- tirely discriminate between such foreign sugars and traces of real glucose. Excessively sensitive copper reagents, such as the copper-glycerol reagent described by Folin*® (1915), will show sugar in every urine. TABLE XVII. 24 Hour Amounts (Milligrams) of Sugar in Normal Urine. Subject H. B-d. Age 34years. Weight 90 kilos. . Day. | Before hydrolysis. Wraee Diet. 1 980 738 Meat, eggs, 50 gm. glucose. 2 611 159 = Sage ee o 3 714 103 “ stage Ge | Was % 4 756 73 “ “ 50 ce “cc 5 652 201 “ “ 50 “ “cc 6 543 101 “ “ 200 “ “ * § 585 85 “ “ee 50 “ “cc 8 624 52 “ “ 50 “ “cc 9 1, 064 910 Mixed; including starches. 10 1, 360 770 # 9 rs 11 1,090 200 = “5 z 12 1, 650 485 2 13 1,170 450 - = Py 14 885 815 5s si 53 While the absorption and excretion of foreign carbohydrates is to be sharply separated from the main carbohydrate metabolism, it does not necessarily follow that the precursors of the practically inescapable sugar of normal urine can be dismissed as of no impor- tance. There is no reason to doubt that these products are absorbed from the blood by the tissues as readily as galactose and lactose and “‘dextrin.”” And the constant harboring of such for- eign products in all the tissues of the body may very well be the cause of at least some minor disorders. As long as the etiology of so many skin and joint troubles remains unknown one cannot entirely dismiss the fact that the human organism is constantly charged with impurities of vegetable origin. %® Folin, O., J. Biol. Chem., 1915, xxii, 327. 256 Carbohydrates A survey of the various cooked and uncooked materials used as food can easily be made on the basis of feeding experiments ac- companied by determination of the sugar of the urine before and after hydrolysis. Such a survey will show which products are particularly rich in soluble unassimilable carbohydrate materials. Fresh cider, for example (500 cc.), yields a little reducing sugar and considerable polysaccharides. The sugars of normal urine cannot be entirely eliminated from the urine by abstaining from carbohydrate food, so that in addi- tion to the precursors already referred to one must also ascribe a part to the endogenous metabolism. That this must be so is shown by the experiment continued for 14 days, recorded in Table XVII. During 8 days the subject, H. B-d., ate only meat, eggs, and butter, together with black tea and 50 gm. a day of pure glucose. The latter was added to exclude acidosis, because the fuel value of the food taken was probably not adequate for the ~ daily energy requirements. The increase in the sugar excretion when the subject returned to ordinary mixed food was prompt and large. On the Nature of the Glucose Threshold together with Some Observa- tions on Renal Glycosuria. . The term alimentary glycosuria has been used to express the fact that some apparently normal persons eliminate moderate quantities of sugar after the ingestion of cane-sugar, glucose, or even after starch. The term alimentary glycosuria is a misnomer. The sugar is excreted either because the level of the blood sugar has risen above the normal threshold, or because the threshold itself is below the normal, as in renal glycosuria. There can be no doubt as to the existence of some mechanism by which the excre- tion of glucose is normally absolutely prevented. If we here venture to advance something approaching a new explanation we realize that it may be wholly wrong, but we believe that it should be at least partly correct. Each of us independently has in former researches seen and emphasized the avidity with which tissues may absorb materials from the blood.*!* And throughout this 31 Folin, O., and Denis, W., Protein metabolism from the standpoint of blood and tissue analysis, J. Biol. Chem., 1912, xi, 87. 82 Berglund, H., Studier oever koksaltomsaettningens pysiologi och patologi, Stockholm, Norstedt, 1920. O. Folin and H. Berglund 257 investigation we have agreed that it is absorption by the tissues rather than glycogen formation which prevents excessive accumu- lations of the sugars in the blood. With reference to glucose we assume that the tissues always contain at least as high a concen- tration of free sugar as the blood plasma and probably more. The glycogen formation need not begin until the tissues have begun to possess a much higher concentration than that present in fasting. Much absorbed sugar can thus be distributed without any large increases of the sugar in the blood. But the kidneys receive their quota of sugar, just as do the other tissues, and this increase of sugar does not involve the slightest degree of strain. The strain comes only when the holding capacity for free sugar is reached and when the glycogen formation must come into play to keep the sugar concentration within normal limits. The speed of glycogen formation is of a much lower order than is the earlier process of merely absorbing the sugar from the blood. At this stage, therefore, the sugar backs up in the blood and the holding capacity of some tissues including the kidneys is exceeded. As a result of the strain thus produced the kidneys are finally compelled to make use of a more efficient process than the glycogen forma- tion for reducing the sugar concentration in the kidney cells, and the elimination of sugar suddenly begins. Thata real local strain has preceded the escape of the sugar is indicated by the fact that the sugar excretion once begun does not stop as soon as the blood sugar has fallen below the threshold, but, in fact, continues until the level of the blood sugar has gone away down, even to sub- fasting values (hypoglycemia). It is as if there had sprung a leak for sugar, a leak which cannot be immediately repaired. Yet the total amount of sugar eliminated need not be very large, except in clinical or experimental diabetes, because other tissues continue to absorb sugar, as well as to make glycogen. We cannot here develop this interpretation in more detail be- cause of its many ramifications. The essential point is that it makes clear why increasing concentrations of sugar in the blood below the threshold values involve no strain and no elimination of glucose. From our interpretation it also follows that excretion of glucose, as in emotional glycosuria and renal glycosuria, does not represent a finely adjusted normal process analogous to the excre- tion of sodium chloride or of waste products. It is also in harmony 258 Carbohydrates with the belief in the importance of not permitting diabetic patients to excrete any sugar at all, if it possibly can be prevented. From our interpretation of the renal threshold for glucose the study of renal glycosuria receives added importance. That inter-. TABLE XVIII.” Showing that in Renal (as in Emotional) Glycosuria the Fall of the Blood Sugar Does not Immediately Stop the Excretion of Sugar. Subject McC-n. Age 22 years. Weight 65 kilos. Time. Blood. Urine. ; Remarks. Sugar per 100 cc. Vol- IA pF side) | evens eee beatae severed |p Core ume Fasting. Urinated at 8.56 Gas puscles.| per a.m. and took 200 cc. water. ree Plasma. puscles. es ae! a.m mg mg mg cr ce mg 10.15 |100—2 | 98—4 |1014+2 | 45 10.21 127 14 10.25 Took 200 gm. glu- cose in 500 cc. water. 11.05 |118+-3 |188—15}) 82+26| 44 THERA 145 | 526 p.m. 12.15 | 96+11|107-+7 | 84+16] 46 aA 97 | 365 1.47 |104—2 |101+-2 |107-—7 | 44 1.58 159 97 3.40 | 61—3 | 54-2 | 70—5 | 44 3.48 178 34 6.45 29 14 | 7.00 p. m. Dinner. 9.03 |178— 13/188 — 28|166-++5 9.08 48 | 200 11.05 37 86 a.m. 7.05 41 20 | Second day. * Small, plus or minus, figures show increase or decrease after hydrolysis. pretation raises, for example, the question whether other tissues than the kidneys are involved. Do these also have a smaller capacity for free sugar than normal tissues, or is this smaller capacity confined to the kidneys? We only raise the question, we are not now in a position to answer it. O. Folin and H. Berglund 259 Renal glycosuria is by no means uncommon; most observations on ‘alimentary glycosuria” represent nothing else. From a class of 100 students one can usually find at least one, and often two or more, who find sugar in their own urine when learning to make the tests for sugar. McC-n. was such a student. He found sugar in the urine after every meal, as well as after 50: gm. of cane-sugar, maltose, or glucose. In Table XVIII are given the figures for blood and urine obtained from this subject after the ingestion of 200 gm. of glucose. The highest observed level of the blood sugar was only 138 mg. per 100 cc. of plasma and 113 mg. per 100 cc. of whole blood. In less than 2 hours the blood sugar had fallen to 96 mg. per 100 cc. of whole blood. Such a typical fall in the level of the blood sugar may be taken as positively excluding dia- betes—a view confirmed in the case of this subject by Dr. E. P. Joslin, who at the request of the young man’s father (himself a physician) later examined him from the standpoint of possible diabetes. The sugar in the urine rose quickly to about 0.4 per cent and remained excessive for over 3 hours notwithstanding an extreme fall in the level of the blood sugar. The evening meal following the glucose produced in this case an excessive rise in the level of the blood sugar, 178 mg. as against only 113 mg. obtained from 200 gm. of glucose. It is a fairly well established fact that the effect on the blood sugar obtained from taking glucose is scarcely, if at all, dependent on the quantity of sugar taken. The maximum level, coming 20 to 45 minutes after the ingestion, may be as great from 50 gm. as from 200 gm. Variable uncontrollable factors, which one must loosely refer to as the condition of the subject, play more of a role than the quantity of sugar taken. From Table XIX it will be seen that McC-n. gave a maximum blood sugar of 139 mg. from only 30 gm. of glucose, as against only 113 mg. previously obtained from the taking of 200 gm. of glucose. Some uncertainty is involved in this comparison because the maximum blood sugar level lasts only for a few minutes and is probably missed more often than found, but the conclusion that 30 gm. of glucose has in this case produced at least as high a blood sugar as did 200 gm. is justified by the figures obtained. The table is given because it seems to show how small a quantity of glucose may produce a temporary glycosuria in cases of this 260 Carbohydrates kind (renal glycosuria), and to emphasize how misleading it might be to interpret the results as indicating a “break down of the carbohydrate metabolism.’”’ When the blood sugar falls to the fasting level, or lower, in 2 to 3 hours after the taking of glucose, there is no reason to believe that the pages to assimilate sugar is impaired. TABLE XIX. Showing that in Our Case of Renal Glycosuria Only 30 Gm. of Glucose Have Yielded Both Glycuresis and Glycosuria. Subject McC-n. Age 22 years. Weight 65 kilos. Time. Blood. Urine. Remarks. Sugar per 100 cc. Vol- June Car Sane ’ - So Stel Ue | Sl F. . Urinated at 9.23 a.m. Cor- |Puseles. per B scei asting rinated at 9.23 a.m Dood, Plasma. puscles ho wig Fo! ie 7: i sae Ce Tee 11.29 97 | 102 92 | 44 11.30 40 15 11.31 Took 30 gm. glucose in 200 ec. water. 11.54] 139] 143] 185] 48 11.55 48 27 p.m. 12.92) 11S |} 123 99 | 41 12.13 57 94 | Qualitative test posi- tive. 12.36 75 re 72 12.37 50 40 1.08 45 14 As a control on the results obtained with 30 gm. of glucose in a case of renal glycosuria, we give the experiment recorded in Table XX. The subject, D-n., had previously been used in an experiment with 200 gm. of glucose (Table I). In this case the | sugar level in the blood rose only to 111 mg. and no glycuresis was obtained. D-n., who at this time was giving more or less assis- tance in the work, had accordingly become much interested, and had, of course, lost every trace of fear. aE O. Folin and H. Berglund 261 From our experiments with fructose and galactose we arrived at - the conclusion that in the human organism there is a threshold for fructose, but none for galactose. On the basis of this interpreta- tion McC-n. should not show any excessive (abnormal) excretion of galactose, and if he eliminated fructose it would indieate that the threshold for this sugar is intimately associated with the thresh- old for glucose. TABLE XX.* Control Experiment with 30 Gm. of Glucose (Compare Table XIX). Subject D-n. Age 22 years. Weight 75 kilos. Time. Blood. Urine. Remarks. Sugar per 100 ce. Vol- Sugar June Cor- ume y + 1921 Whole ee a puscles per ae ; Fasting. Urinated at 8 a.m. bl puscles : a.m mg mg. mg ae ce. mg. 11.10 | 94—5 94-15, 94+10) 41 11.12 28 | 17+3 11.14 Took 30 gm. glucose in 300 ce. water. 11.30} 93—14| 94—16) 91-11) 40 11.32 21 17+1 11.47 |111—17|114—23}106—6 | 40 11.52 21 17+0 p.m 12.05 |101—15/103—17| 97—10, 39 12.15 18 | 19+0 12.28 | 97—19|102—21| 90—17 12.31 22 | 22+1 12.52 | 89—12) 92—22) 86+2 12.57 18 19-++2) * Small, plus or minus, figures show increase or decrease after hydrolysis. The urine passed 1 hour after taking 50 gm. of our purest levu- lose (Kahlbaum’s) gave with Seliwanoff’s reagent exactly the same color as the urine passed immediately before the taking of the sugar. We are, therefore, certain that no fructose excretion took place. This negative result does not, of course, prove anything concerning the fructose threshold, because we do not know how much of it passed through the liver without being changed into 262 Carbohydrates glucose. The urine gave also negative results for glucose (quali- tative tests), whereas positive qualitative tests were always obtained after 50 gm. of cane-sugar. It is, therefore, not probable that transformation into glucose was the reason why no fructose - appeared in the urine. The quantitative determination gave scarcely any higher figures than would be accounted for by the impurities of the levulose. The galactose experiment was made as follows: H. B-d., serving as control, and McC-n., the renal glycosuria subject, took on a fasting stomach 30 gm. of galactose in 200 ce. of water. In3 hours the control subject eliminated 474 mg. of sugar after deducting the normal fasting amount (obtained by analysis of the urine passed before taking the galactose). The renal glycosuria subject eliminated only 404 mg. The lowered threshold for glucose is, therefore, without influence on the excretion of galactose, as should be the case on the basis of our conclusion that there is normally no threshold for galactose. The figures obtained are included in Table XITI. On Emotional Hyperglycemia and Glycosuria. One striking feature of all our experiments including those with glucose, maltose, dextrin, and starch, is that the blood sugar has not risen to the high levels which practically all other investigators have obtained from similar experiments. It is evident that these other investigators have introduced some factor favoring excessive hyperglycemia which we have accidentally avoided. We believe that this added factor is nothing else than the emotional state of the subjects. Our subjects, aside from H. B-d., were healthy medical students who offered to serve and who knew and learned from each other that no pain or danger was involved. It seems to us distinctly doubtful whether an equally favorable state of mind can often be obtained in the case of patients who stand in awe of the ‘“‘doctor’’ and are unused to laboratory surroundings. We also believe that a relatively insignificant emotional disturb- ance has a greatly increased effect on the process of sugar mobili- zation when there is an active absorption of sugar from the diges- tive tract. It is quite true that those who have taken the blood from the fingers should get somewhat higher values for the blood sugar because if they prick deeply enough they will get uncontami- O. Folin and H. Berglund 263 nated arterial blood, but that this does not account for more than a part-of the difference between their results and ours is shown by the fact that they so often get unmistakable glycosuria. Cammidge, Forsyth, and Howard® have recently published a paper on the factors controlling the normal concentration of sugar in the blood, the results of which show, according to our interpretation, what misleading results may be obtained because of emotional sugar mobilization superimposed on the other experi- mental conditions. These authors took the blood from the roots of the finger nails. They have obtained substantially the same degree of hyperglycemia (140 to 180 mg.) from any kind of food except fat, as they obtained from sugar or other carbohydrates. They have accordingly reached the conclusion that the absorption of sugar from the digestive tract has nothing special to do with the obtained hyperglycemias. Our experiments with egg white and with gelatin, dating from February, 1921, have yielded results entirely different from those reported by Cammidge, Forsyth, and Howard, and we are, there- fore, forced to conclude that these authors were really dealing with emotional glycosuria. In one respect our results agree with theirs, namely in the fact that fat ingestion may lead to hypogly- cemia—and it is in connection with hypoglycemia (next section, page 265) that we give the details of our findings. Even under our conditions of working we have not entirely escaped the superimposed effects of emotional glycosuria. One experiment illustrating the same deserves to be recorded (Table XXI). The subject, a medical student, S-g., never had any sugar in his urine (qualitative tests) either before or after the day of the experiment. When the preliminary sample of blood was taken he grew pale and very faint. When the second sample was taken he fainted completely. The plasma of this blood sample contained 210 mg. of sugar, whereas the first contained only 105 mg. In the course of the next hour the sugar content sank to 111 mg., and neither at this time nor later did the taking of the blood disturb him. The figures for the sugar of the urine may be cited as one more illustration of the fact that when the glucose threshold is once 33 Cammidge, P. J., Forsyth, J. A. C., and Howard, H. A. H., Brit. Med. J., 1921, 586. 264 Carbohydrates overcome by the rising tide of the blood sugar, the normal thresh- old is not restored for several hours. In this case the excretion of glucose evidently continued till the plasma sugar had fallen below 80 mg. TABLE XXI.* Showing Effects of Glucose Ingestion Combined with Emotional Disturbance. Glycosuria Once Begun Does Not Stop at the Threshold. Subject Mr. S-g. Age 29 years. Weight 74 kilos. Time. Blood. Urine. Remarks. Sugar per 100 cc. woe Vol- oe pus- | ume ean ee bet at 9.10 1921. er er hour. a.m. and too. cc. water Whole Cor- | cles. |,? ig 5 e Blood. Plasma. paseles hour , vol a.m mg mg mg. per ce mg 10.15 | 90—1 |105—18) 69-425} 40 Took 200 gm. glu- cose and 100 ce. casein digestion (4 gm. N) in 650 cc. water. 11.22 |189—31/210—47|160—10| 42 11.40 55| 1,260—310 12.15 |108—6 |111—20)105+12) 42 526) 1,273—173 1.37 | 88—5 | 81—8 | 970 | 40 1.41 162 43+-7 4.18| 77—7 | 74—7 | 82—7 | 40 4.26 60 22+8 5.45 | 96—9 |101—14| 87+0 6.00 21; 20+6 6.30 Dinner. 9.10 57|1,235—45 * Small, plus or minus, figures indicate increase or decrease after hydrolysis. In the evening, after dinner, the blood sugar must again have exceeded the threshold, for the urine passed at 9.10 p.m. contained over 2 per cent of sugar. In this respect the result resembles that obtained from McC-n., the renal glycosuria subject. We are under finden O. Folin and H. Berglund 265 the impression that excessive hyperglycemias are more easily obtained in the evenings than during the early forenoon hours, but have made no experiments bearing directly on this problem. In order to prevent future misinterpretation of Table XXI we must specifically mention that S-g., like several other subjects, got several grams of mixed amino-acid nitrogen from casein to- gether with the glucose. This variation was introduced in order to see whether nitrogenous materials could contribute anything to the raising of the sugar level in the blood, but as no such effect was found it has not seemed worth while to dwell upon this detail. On Subfasting Blood Sugar Levels (Hypoglycemia). In the preceding section we have referred to the important réle which we ascribe to the mental state of a subject as a factor in the production of excessive hyperglycemias during active sugar absorption. Normally the general tissues are abundantly capable of taking up all the sugar which can be delivered by way of the intestinal absorption, but the tissue absorption from the blood may be interfered with by many factors which have no influence on absorption from the intestine. The rise in the blood sugar has heretofore received a dispro- portionate share of attention. The fall of the blood sugar below the fasting values has been almost entirely neglected, yet such a fall we believe to be fully as normal and almost as regular a con- sequence of a carbohydrate meal as the rise. MacLean seems to be the only one who has given any special attention to this phase of carbohydrate transportation, or who has attempted to interpret the same. The intial rise in the blood sugar is due, according to MacLean to the dormant glycogen-forming function of the liver, which does not begin to act until the blood sugar concentration begins to approach the threshold of 160 to 170 mg. The initial rise of the blood sugar is, therefore, due to this lag in the glycogen-forming process. Once started the glycogen-forming process gains momen- tum, overtakes and passes the speed of the sugar absorption from the intestine, thereby producing the subsequent fall in the blood sugar to concentrations far below fasting values. In the main this interpretation seems intelligible and plausible, except that it as- sumes a greater lack of sensitiveness or adjustibility, to varying 266 Carbohydrates requirements, on the part of the sugar-stormg mechanism than might be reasonably expected. We find it difficult to suppose that the glycogen formation in the liver should not begin until the level of the sugar in the general circulation is approaching the threshold value. The sugar concentration in the portal vein must TABLE XXII.* Showing that Protein (Raw Egg White) Does Not Change the Level of the Blood Sugar and Does Not Cause Glycuresis. Subject H. B-d. Age 34 years. Weight 90 kilos. Time. Blood. Urine. Remarks. Sugar per 100 ce. Vol- Jan. a a5 Se PGT © ae |e ee oae: pe pees Fasting. Urinated at 9.15 a.m lsh Plasma. te cles. hour. hour. vol. p.m. mg. mg. mg. per ce. m4. cent 12.55 | 108—3/114—7 | 99+3 | 40 1.35 22 | 30+5 2.15- Took 1,000 gm. raw 2.25 egg white. 3.03 | 102—2/106—S | 96+6 | 39 3.25 115 | 2747 4.05 | 101—9 4.35 282 | 28+6 5.35| 98—4) 98—4 | 98—4 | 39 6.05 71 | 26+5 7.45 | 101—7| 97—3 |108—14| 35 8.35 58 | 24+-6 9.35 | 100—6|105—12) 9243 | 38 10.25 49 | 21+-4 * Small, plus or minus, figures denote increase or decrease in sugar after hydrolysis. reach this and higher values comparatively quickly, and the initia- tion of glycogen formation in the liver comes presumably from the sugar concentration in the portal blood. MacLean and de Wesselow,* like most other investigators, have obtained materially higher levels of the blood sugar in response to various carbohydrate intakes than we get, and it must be in 34 MacLean, H., and de Wesselow, O. L. V., Quart. J. Med., 1921, xiv, 103. O. Folin and H. Berglund 267 part because of this fact that they assume that such levels are reached before the glycogen formation comes into active play to check the sugar accumulation. The fall of the blood sugar, according to our interpretation, is not due to any unsuitable adjustment or excessive momentum of the glycogen-forming process. That fall comes when there is- TABLE XXIII.* Showing that Protein (Gelatin) Does Not Raise the Level of the Blood Sugar. Subject H. B-d. Age34years. Weight 90 kilos. Time. Blood. Urine. Remarks. Sugar per 100 ce. Vol- ae 2 > i aoe nny en eg Fasting. Urinated at 9 a.m. Blood. | Plasma. | a acetes, our, | Hour p.m mg mg. mg ee cc mg. 1.55 | 97-4} 98—7| 96+1 1] 41 2.20 27 33 Took 135 gm. gel- atin in 900 ee. nig hot water. No ; other food taken all day. 4.00 | 98-8 | 93+3) 93—32) 42 4.37 39 38 4.52 | 8441] 90-2) 754+5) 41 6.03 39 44 6.10 | 92-6 | 93-9] 91-1 | 39 7.15 ; ff 70 10.34 82 52 10.55 | 98—7 | 102—5| 92—10) 39 a.m. 9.00 30 31 | Second day. * Small, plus or minus, figures show increase or decrease after hydrolysis. reason to believe that all tissues are well supplied with available food material. It is an index of the fact that the need for sugar transportation from some tissues to others has temporarily fallen very low or ceased altogether. It is for this reason that the lowest levels and the longest periods of low levels are usually obtained after very large ingestions of sugars. 268 Carbohydrates This interpretation is very elementary and perhaps incomplete, but it does agree with facts which cannot very well be explained on the basis of excessive glycogen formation, as, for example, the hypoglycemia obtained after the ingestion of gelatin and olive oil. Tables XXII, XXIII, and XXIV give the results obtained with 1,000 cc. of egg white, 135 gm. of gelatin, and 200 gm. of olive TABLE XXIV.* Showing Early Hypoglycemia after 200 Gm. Olive Oil. Subject H. B-d. Age34years. Weight 90 kilos. Time. Blood. Urine. Remarks. s 100 ec. Heb 1, ese as ne = eee pean — Fasting. aunts : > : t 11.40 a.m. epee | Plaame.| Sots || et oa eed p.m mq mg. mg eer cc mg. 2.15 | 96—10} 100—8) 89—13) 37 3.10 48 | 2449 3.15 Took 200 gm. olive oil and 250 ce. water. 4.20} 70+1 72—8| 68+12) 4] 4.45 232 | 27+8 5.53 | 88 95 77 40 6.05 60 | 20+11 7.30} 92—1 93+=0) 91-4 | 38 8.07 39 224-4 9.45 | 87-1 90—2) 83+1 37 10.17 42 | 22+5 * Small, plus or minus, figures show increase or decrease after hydrolysis. oil. The prompt and pronounced fall in the blood sugar (from 100 to 72 mg.) in the plasma after the olive oil is particularly remarkable, coming as it does within 1 hour after the ingestion. The fall in the blood sugar after the gelatin is small, too small perhaps, to permit drawing any definite conclusion. The result is further obscured by the fact that it gave rise to a glycuresis. It is difficult to tell what impurities may be present in commercial gelatin preparations. At all events there is certainly no rise of the blood sugar either after gelatin or after egg white. — O. Folin and H. Berglund ~ 269 TABLE XXvV. Showing that Hypoglycemia Is No Less Frequent than Hyperglycemia after Glucose Ingestions. Whole blood-plasma sugar per 100 cc. Subject. orang Remarks, : Before.* Highest.* |Lowest.* gm. mg. mg. | mq. T. W. W-r. .| 200 93-96 | 138-166 54-58 McC-n...... 200 | 100-98 | 113-138 }61-54 ws... 200 84-93 | 116-129 57 L. A. C-n. ..| 200 86-96 | 132-143 66-51 Took 15 gm. urea with the sugar. J. 0: ee....| 200 95-105 | 168-203 66-74 | Renal glycosuria and | emotional hyper- | glycemia. Mr. Ft... . 150+150} 89-92 | 109-113 67 | 2 hour intervals be- tween the sugar por- tions. Took 15 gm. urea with the first one. Mr. S-g..... 200 90-105 | 189-210 77-74 | Took 100 ce. casein | digestion with the sugar. Emotional | hyperglycemia. McC-n...... 30 97-102 | 139-143 75-77 The lowest figure 1 hour 5 minutes after | ingestion. J. M. F-r....| 120 94-93 | 131-139 77-83 | Took 250 ce. casein digestion (10 gm. N) with the sugar. J.D. C-d...| 200 | 90-94 | 118-130 [78-81 | P. A. Ch-r. 215 | 90-96 | 138-154 |85-92 | _ 2 eee 30 | 9494 | 111-114 89-92 | The lowest figure | 1 hour 38 minutes | after ingestion. iy... J-n-% 200 | 105-107 | 152-172 95-101, * The first number in this column represents whole blood, the second plasma. In Tables XXV and XXVI are given summaries of the fluctua- tions encountered in the sugar of whole blood at various periods after taking different kinds of food. In explanation of the many extremely low minimum values found it should be stated that most 270 Carbohydrates of the subjects availed themselves of our invitation to rest on a couch during the greater part of the day. This resting condition was unfortunately not imposed on all the subjects. It was neces- sarily omitted in the many experiments with H. B-d., one of the authors of this paper. TABLE XXVI. Showing Effects of Other Food than Glucose on Hyper- and Hypoglycemia. Whole blood-plasma sugar Subject. Food intake. per Mike Remarks. Before.* | Highest.* | Lowest.* mg. mg. mg. H. B-d....| 200 gm. lactose. 94-96 | 101-107 | 64-62 Mr. C-c-a.| 200 “ maltdse. 88-93 | 116-122 | 67-71 H. B-d....| 175 “ raw starch. | 93-98 | 83-96 | 68-81 H. B-d....| 200 “ maltose. 90-90 80-108 | 69-73 He Bat 48200, olive) oul: 96-100} No rise.| 70-72 H. B-d....}| 100 “ galactose + 100 gm. glu- cose. 94-97 | 116-113 | 79-72 H. B-d....} Mixed lunch. 95-90 | No rise.| 77-80 | Fasting. O;, Faniee = es 100-103' “ “ | 86-83 < H. B-d....| 350 gm. dextrin. 101-109} “ - ™ | (33-35 H. B-d....| 135 “ gelatin. 97-98 | 660) SU SO H. B-d 100 “ galactose. 93-91 | 100-101 | 87-88 H. B-d....| 200 “ fructose. 100-102) 102-110 | 93-95 H. B-d....| 200 “ dextrin. 93-96 | 115-110 | 98-94 | Dextrin taken 3 hours after breakfast. H. B-d. | 1,000 gm. egg white.|108-114| No rise.| 98-98 * The first figure in each column represents whole blood, the second plasma. Distribution and Character of the Blood Sugar. From the many blood sugar figures recorded in this paper for whole blood and for plasma, there is, we believe, no escape from the conclusion that the corpuscles in the circulating blood are permeable to glucose. We are inclined to think that they are not only permeable, but that they actually can take up more sugar from the plasma than can be explained on the basis of diffusion pees OR TT | O. Folin and H. Berglund 271 alone. We do not venture to stress the possible active absorption of sugar by the corpuscles and absorption analogous to that postu- lated for tissues in general; the conditions are too complicated to justify any dogmatic statement. But if we consider the sugar concentration in plasma and corpuscles in relation to the water content of each, we find that in every case the sugar concentration in the water of the corpuscles is greater, sometimes fully twice as great, as the sugar concentration in the water of the plasma. On the basis of 90 per cent of water in the plasma and 50 per cent in the corpuscles, a concentration of 90 mg. in each would corre- spond to concentrations of 100 and 180 mg., respectively. Without making any allowance for different water contents we see from the tables that there is no very uniform correspondence between the sugar of the plasma and of the corpuscles. In the fasting condition (before breakfast in the morning) the sugar contents of blood and corpuscles are so nearly together that one gets very nearly the same sugar value on whole blood as on plasma. The latter tends to give slightly higher figures, but the differences are not such as to justify the demand that plasma must be used in clinical work for sugar determinations. During periods of active sugar absorption accompanied by considerable increases in the sugar concentration there is a rela- tively greater increase in the plasma than in the corpuscles, but this difference is naturally of short duration because the hyper- glycemia itself does not last very long. What the situation is in hyperglycemia of diabetes we do not know. The records in the literature bearing on this point can scarcely be accepted because no adequate precaution has been taken to separate the plasma under conditions precluding subsequent changes in the distribu- tion of the sugar. The lack of a decided uniformity in the relationship between the sugar of plasma and corpuscles is not to be wondered at. The ~ incoming sugar can reach the corpuscles only by first passing through the plasma. In addition there are other factors to be considered. In the plasma there is no catabolism and no poly- merization of sugar, whereas both these processes must be assumed to occur in the corpuscles—in the absence of binding evidence to the contrary. If we are justified in assuming that the sugar in reality is of a higher concentration in the water of the corpuscles THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI, NO. 1 272 Carbohydrates than in the water of the plasma, we have the added possibility that the corpuscles at times may lose sugar to the plasma. From the sugar obtained after hydrolysis it has become clear to us that the subject under discussion, the distribution of sugar between plasma and corpuscles, is further obscured by the uncer- tainty as to the nature of a part of the blood sugar. There are present in the blood, and, we believe, also in the urine, reducing substances which disappear during the hydrolysis (heating on a water bath with 1 per cent hydrochloric acid). These substances are most prominent in the plasma, as is indicated by the fact that the plasma filtrates very frequently yield less ‘‘sugar” after than before hydrolysis. The whole blood filtrates may show either an increase or a decrease in the reducing value as a result of the hydrolysis. It is evident, therefore, that as far as the corpuscles are concerned, the predominant effect of hydrolysis is an increase of the sugar, and this increase is usually sufficient to more than counterbalance the decrease obtained from the plasma. As there is no reason to doubt the presence in the corpuscles of the same substances which in the plasma cause a diminution of the sugar, the increasing effects obtained from hydrolysis of the corpuscle extracts is probably in reality larger than it appears to be. The increases produced by hydrolysis derived chiefly from the corpuscles may be due wholly, or in part, to glycogen. This must perhaps be admitted though we have not been able to prove it. Oyster glycogen added to whole blood immediately before precipi- tating with sodium tungstate and sulfuric acid is precipitated so completely that not a trace goes into the filtrate. It is conceivable that the glycogen of human blood corpuscles is a different glycogen, but it is perhaps more probable that the extra sugar obtained by our method is derived from some inter- mediate carbohydrate. SUMMARY AND, CONCLUSIONS. 1. In the absence of emotional complications or a subnormal renal threshold (renal glycosuria) the ingestion of pure glucose (up to 200 gm.), does not raise the level of the blood sugar above the threshold in normal persons, and no glycosuria is obtained. 2. Fructose, galactose, or lactose, as well as dextrin or starch, are much less effective than glucose in raising the level of the blood sugar. O. Folin and H. Berglund 273 3. Absorption of sugars from the blood by the tissues, rather than glycogen formation, is believed to be the immediate reason why the sugar fails to accumulate in the blood. 4, A renal threshold analogous to that for glucose is believed to exist for fructose, but not for galactose or lactose. 5. The retention and utilization of galactose depends on the amount of available glucose. 6. Hypoglycemia (subfasting blood sugar level) is probably quite as normal a consequence of carbohydrate ingestion as hyper- glycemia, but comes later. 7. Hypoglycemia is probably a reflection or index of a decreased need for sugar transportation from one set of tissues to another. This condition is obtained when there is an abundance of available carbohydrate material in all the tissues. A general abundance of other suitable food than glucose, notably fat (olive oil) may, there- fore, also produce hypoglycemia. Hypoglycemia (in venous blood) can occur even during a prolonged moderate sugar absorp- tion from the intestine, because the absorbed sugar may get by the liver but does not get into the venous blood. 8. Definite ‘‘glycuresis’’ (Benedict) is obtained after every ordinary carbohydrate meal. 9. Glycuresis is independent of the level of the blood sugar, and is not normally obtained from the ingestion of pure glucose, maltose, dextrin, or starch. The transition from below to above _ the renal threshold appears to be so sharp and decisive that there is practically no intermediate stage represented by glycuresis. 10. Glycuresis represents the absorption and excretion of (1) foreign unusable carbohydrate materials present in grains, vege- tables, and fruits, and (2) decomposition products due to cooking, canning, and baking of such food. 11. The sugar of normal urine consists, therefore, of a motley variety of carbohydrate products and carbohydrate derivatives including di- and polysaccharides. 12. The blood sugar is distributed in somewhat varying pro- portions between plasma and corpuscles. In fasting the distri- bution is nearly equal between the two. 13. The plasma sugar is usually diminished by hydrolysis, while the sugar of the corpuscles is usually increased. The cor- puscles probably contain polysaccharides—possibly polysac- charides other than glycogen. = - ‘Ne Tia ; cn : Ripsk rs 6 : et! Sp. i rare ‘ i ‘ ‘ met: oad eet eel “VT? Lenityit: tr sor 7 sch “itieh CFR citiagt (OMNI Oa eae i) 1 Views Lowy olantive J yeaa te: eth ee eT cei) iy, ite ‘oth ty : 1 Rereyge fy . ee ore ‘ neil pt Tha. = x sp) E i, o> peel ” iy ie ral ad } . oe a) : a} On ee THE PREPARATION OF INULIN, WITH SPECIAL REFER- ENCE TO ARTICHOKE TUBERS AS A SOURCE.* By J. J. WILLAMAN. (From the Division of Agricultural Biochemistry, Minnesota Agricultural Experiment Station, St. Paul.) (Received for publication, December 6, 1921.) From time to time the suggestion is brought forward that the tubers of the Jerusalem artichoke, or girasole, would be an excellent source of commercial fructose, either as a syrup or as erystals.} The advantages offered are the enormous acre yields of the tubers, and the greater sweetness of fructose over other sugars. Stein proposed to separate the inulin from the expressed juice, and then hydrolyze it to fructose. Cockerell’s syrup was made by hydrolyz- ing the juice directly. Hudson suggests the separation of fruc- tose as the calcium compound; this would be more economical when used in connection with the second method mentioned above. Other writers have not suggested a definite procedure. It thus appears that information is needed on the probable merits of the two methods. In making some preparations of inulin from girasoles, the writer had certain experiences and discovered certain facts that might prove of value in this connection, and also in connection with the preparation of pure inulin. They are recorded in the present paper. Methods of Preparing Inulin. There are three general methods of separating inulin from con- taminating substances: (a) direct crystallization from water, * Published with the approval of the Director as Paper No. 295, Journal Series, Minnesota Agricultural Experiment Station. 1 Stein, S., Int. Sugar J., 1908, x, 218. Daniel, A., British Patent No. 109, 813, 1917; in Chem. Abstr., 1918, xii, 156. Hudson, C.S., J. Ind. and Eng. Chem., 1918, x, 176. Cockerell, T. D. A., Mo. Bull. State Comm. Hort., Colorado, 1919, viii, No.5. Willaman, J. J., Science, 1920, lii, 351. 275 276 Preparation of Inulin (b) precipitation by 60 per cent alcohol, and (c) precipitation as the barium compound in excess of Ba(OH)s, and subsequently decomposing with CO. All three methods involve a preliminary clarification of the juice, usually with lead acetate. Method (a) was first used by Kiliani,? and then by many other investiga- tors with various modifications. Alcoholic precipitation was first used by Dragendorff,? and has frequently been used in connection with the third method, which was instituted by Tanret.4 Dean® tested many modifications of these methods, and found that the best procedure is to clarify with lead acetate, add Ba(OH): to excess, then methyl alcohol to 30 per cent. The Ba precipitate is decomposed with COs, and the filtrate evaporated to dryness. This material is then successively extracted with boiling 84, 70, and 60 per cent alcohol. The final residue is true inulin, and is purified by a further precipitation with 60 per cent alcohol. Irvine and Steele used recrystallization from water, followed by prolonged washing in cold water, and dehydration with a series of alco- hol and of ether solutions. The most essential object to be at- tained in any method is the separation of the so called true inulin from the inulides, or inuloids, such as the pseudoinulin, helianthenin, inulenin, and synanthrin of Tanret. These bodies form a more or less progressive series with inulin, differing indis- tinctly in solubilities and in specific rotations. They are all primarily anhydrides of fructose, although Tanret’ believes that they are characterized by glucose components in different amounts. Thus, from the standpoint of preparing pure inulin, the problem is to segregate the latter from the inulides. But from the stand- point of preparing fructose from girasole tubers, the problem is to separate the whole group from the other constituents of the juice. In experimenting with the preparation of inulin, various steps in the existing methods were subjected to scrutiny. Without 2 Kiliani, H., Ann»Chem. Pharm., 1880, cev, 145. 3 Dragendorff, Material zu einer Monographie des Inulins, St. Peters- burg, 1870; quoted in Abderhalden, E., Handbuch der biochemischen Arbeitsmethoden, Berlin, 1911, ii, 185. 4 Tanret, C., Compt. rend. Acad., 1898, exvi, 514. 5 Dean, A. L., Am. Chem. J., 1904, xxxil, 69. 6 Irvine, J. C., and Steele, E.S., J. Chem. Soc., 1920, exvii, 1474. Irvine, J. C., and Soutar, C. W., J. Chem. Soc., 1920, exviii, 1489. 7 Tanret, C., Bull. Soc. chim., 1893, ix, 200, 227, 622. Se ttm { ; ; | J. J. Willaman aia giving all the details of the work, the points on which any concrete evidence was obtained will be briefly discussed. Professor G. R. McDole, of the Idaho Agricultural Experiment Station, very kindly furnished the artichokes. The variety was not definitely known, but was probably Helianthus tuberosus albus. The tubers were worked up during the winter, some being used as late as February. Extraction of the Juice——Since inulin is extremely insoluble in water at freezing temperatures, it was thought that by freezing the tubers and expressing the juice immediately on thawing considerable impurities might be removed. Accordingly, a sack of tubers, 2,000 gm., was left outdoors for 24 hours at about -25°C., ground, thawed at 0-5°C., and pressed in a hand-press. 530 cc. of juice, of 1.14 specific gravity, were obtained. The residue was extracted with boiling water in the usual way, and both juice and extract were worked up for inulin. The extract gave a lighter colored preparation than usual, and of lower yield, indicating it to be somewhat purer inulin. The juice gave a crop of apparently good inulin; but when the mass was diluted with ice water and centrifuged (see below), the crystals almost com- pletely dissolved. A second lot of tubers was treated in the same way. The first crop of inulin from the extract gave [a], = — 34.4°, whereas that from the juice even after two crystallizations from water, gave {a}; = —30.9°. Thus it is apparent that the above freezing procedure gives a better preparation of inulin, but that it is wasteful for the prep- aration of fructose from the total inulin bodies. Some further experience with the process is required before it can be unquali- fiedly recommended. Clarification of the Juice-—The usual procedure is to add lead acetate, filter, and delead with either hydrogen sulfide or a sulfate. It was found that the inulin bodies have great peptizing power; as soon as a slight excess of HS is present, colloidal lead sulfide is formed, and it is very difficult to destroy. Lead sulfate is often difficult to filter. Sodium carbonate and ammonium oxalate were tried as deleading agents, and the latter found to be superior to any of the others. Daniel,! in his patent covering the manufacture of inulin, specifies clarification with alkali, using any one of seven common 278 Preparation of Inulin hydroxides. Cockerell’s syrup’ was clarified with Ca(OH)s, COs, and SO,. Alkalinity with Ca(OH). was tried on an extract of girasole tubers. A quantity of precipitate was formed, but it could not be filtered. After centrifuging, it was acidified with phosphoric acid, and again filtered. This gave a liquor almost devoid of color; in fact, it was the best decolorization obtained in any of the experiments. Other trials gave similar results. Crys- tailization of inulin was not as good as from a solution clarified with lead, there apparently being some gummy material present that impeded crystallization. Because of these results, lead acetate followed by ammonium oxalate, was always used in clarification. ‘The lime and phosphate method is not good for the preparation of inulin, but it shows promise for a process in- volving hydrolysis of the clarified juice directly into sugar. The Purification of Inulin.—The separation of inulin by means of its barium compound is a long process; at least it is much longer than crystallization from water or precipitation by alcohol. It was decided to make a study of the relative merits of the latter two methods, comparing the yields and quality of products ob- tained by crystallizing from an aqueous syrup, and by precipitating from a medium containing 60 per cent of alcohol, this being the concentration generally agreed upon by workers as differentiating inulin from its congeners. Two portions of a sample of inulin, [a]; = —34.4°, were dis- solved in water to a 12.5 per cent strength. One was allowed to crystallize over night at 0-5°C., and was then filtered, washed with ice water, and then with 20, 50, 80, and 95 per cent alcohol, and finally with ether. It was then dried in an oven at 95-100°C. The other solution was made to a 60 per cent aleohol content and let stand over night. It was then filtered, washed with 60, 80, and 95 per cent alcohol and ether, and dried. The following data were obtained: Crystallization from: | Yield. lal per cent Wiater::: .ccRugeekics floes. cee eee 80 —37.1° 60:per: cent alcohol :.4305.33..3 0042 hee ee eee 94 —36.5° 8 Personal communication. os J. J. Willaman 279 The data in Table I were selected from the results of several experiments, which were not designed primarily to show the solu- bility of inulin in alcohol. : Table I shows that inulin has an appreciable solubility in 60 per cent alcohol, and that hence the yield of inulin from alcoholic precipitation is greater the more concentrated the solution of inulin that is used. A more thorough comparison of crystallization in water and in alcohol was next made. A composite sample of inulin, made up of many small lots, and totaling about 100 gm., was washed by grinding in a ball-mill with 65 per cent alcohol, filtering, dissolv- ing in boiling water, filtering with kieselguhr, and evaporating to a thin syrup (about 1.13 specific gravity). This was divided into TABLE I. tai” ws tts f Suceueason of a of = inulin in the solution | po oovery of inulin. inulin in the - ted b 60 alcohol preparation used. | Precipitated by a per cent per cent per cent —34.4° 2.0 38 0.46 —34.4° 2.5 59 0.40 —36.5° 12.5 94 0.13 —36.3° 25 to 27 0.50 —31.4° 25 to 27 0.35 —33.2° 25 to 27 0.51 —33.3° 25 to 27 0.64 two portions. One portion was used for successive crystalliza- tions from water, the other for successive crystallizations (pre- cipitations) from 60 per cent alcohol. The one for the aqueous process was kept at 0-5°C. for 24 hours, at which time it had formed a solid mass of crystals. It was stirred with an equal volume of ice water, filtered, washed with ice water, then with 20, 50, 80, and 95 per cent alcohol and then ether. It was dried to constant weight at 100°C., and its specific rotation determined. It was then redissolved in hot water, filtered with washing, con- centrated to a content of 25 to 27 per cent inulin, and crystallized as before. The filtrate and washings from each of these lots of crystals were concentrated to a content of 12 to 16 per cent inulin, made to 60 per cent alcohol by means of 95 per cent, and then treated in the same way as the other crystallizations from alcohol. 280 Preparation of Inulin That portion of the original inulin syrup designed for successive precipitations by alcohol, after standing 24 hours at 0-5°C., was filtered cold, washed with cold 60, 80, and 95 per cent alcohol and then with ether, and dried as in the case of the aqueous samples. Each sample was redissolved in hot water, filtered with washing, Specific rotations O-Crystallized from waler e- ” « 60% alcohol @-Filtrate from © 60% Yields in percentage Crystallization Fig. 1. Chart showing the specific rotation and yield of crystals of inulin prepared by successive crystallizations from water and from 60 per cent alcohol. concentrated to a content of 25 to 27 per cent inulin, and again made to 60 per cent alcohol. Fig. 1 shows the results of these successive attempts at puri- fication. Both the specific rotation and the yield are given in each case. J. J. Willaman 281 It will be seen that the rotations of both the aqueous and the alcoholic samples increase during three recrystallizations, then decrease markedly on the fourth, then increase to a maximum, and finally undergo another decrease. The filtrate from the aqueous preparations does not show this phenomenon, but exhibits an almost continuous increase in rotation. It shoyld be remem- bered that these latter preparations represent the inulin which is left in solution in the mother liquor, and which is precipitable by 60 per cent alcohol. They always show a considerably lower rotation than the parent samples, but this difference becomes less and less during successive recrystallizations. Unfortunately the last sample was lost; and the recrystallization had to be dis- continued at the seventh one because the amount of material had become too smail to work with. Dean reports 'the same ex- perience in trying to purify inulin to a constant rotation. When his preparations, crystallized from water, reached values of 38 to 40, a further crystallizing always brought about a reduction in this value. In the present work, the inulin from the aqueous medium always had a higher rotation than the corresponding preparation from alcohol. The yields were higher from alcohol, however. The above data constitute rather convincing evidence that inulin is not a single substance, but a mixture; and a somewhat unstable mixture at that. The successive crystallizations amount- ed to a fractionation, with the leaving behind in the mother liquor of those molecules. of lesser specific rotation. When 60 per cent alcohol is the medium, the separation is less marked, due probably to the fact that the differential solubility of the two groups is greater in water than in alcohol. This is further shown by the greater yields from alcohol. Dean came to this same conclusion as regards the identity of inulin. He says, “May we not, in a purely empirical way, call the carbohydrate, or carbo- hydrate mixture, which is readily precipitated in cold alcohol of 60 per cent strength, and hasa specific rotation of la]p = —33° to —40° inulin; and the undetermined mixture of lower rotatory power and greater solubility, the levulin mixture?’’ He postulates the existence of various aggregates of molecules, more or less loosely combined; the greater the aggregate, the 282 Preparation of Inulin greater is its specific rotation, and the lower is its solubility in water. This theory explains fairly well the behavior of inulin on successive crystallizations, either from water or from alcohol. The Girasole as a Source of Inulin.—Tanret reports an analysis of the extract of girasole tubers which showed that of the total carbohydrates, inulin was 20 per cent; pseudoinulin, 0.3; inu- lenin, 9; helianthenin, 6; synanthrin, 50; sucrose, 12; and glucose and fructose, 3. This would indicate that these tubers are a poor source of true inulin. The writer’s experience has shown this to be the case; the inulides greatly predominate over the inulin in all the lots of tubers examined. After precipitating a given extract with 60 per cent.alcohol, no appreciable further precipitate was obtained until the concentration of alcohol reached about 84 per cent, when a voluminous coagulum was obtained. On the average, the yield of inulin of specific rotation of —33° or over was about 2 to 5 per cent of the weight of the tubers, whereas the total inulin bodies must amount to three or four times this. The records in the literature go to show that dahlia tubers are a better source of true inulin, . CONCLUSIONS. 1. On the basis of the experience outlined in this paper, the writer recommends the following procedure for the preparation of inulin from artichoke tubers: Grind the washed tubers as fine as possible, and put into boiling water containing calcium car- bonate. For each kilo of tubers use 1,300 cc. of water and 30 gm. CaCO;. Boil 15 to 20 minutes, extract the juice with a press, reboil with 1,000 cc. of water and 10 gm. CaCOs, extract, and combine the extracts. Clarify with lead acetate, avoiding a large excess. Centrifuge, or filter, remove the lead with amnio- nium oxalate, and centrifuge again. The clear liquor may here be treated with decolorizing carbon, although this is usually not necessary. It is then evaporated under vacuum to a content of 40 to 60 per cent of solids. This syrup is allowed to cool slowly, then it is kept at 0-5°C. for several hours, thoroughly stirred with an equal volume of ice water, and centrifuged. The crystals are redissolved in about 3 volumes of water, filtered hot, concentrated to about twice the volume of the original crystals, and allowed to crystallize in the cold as before. They are again J. J. Willaman 283 stirred with ice water, filtered on paper or on silk bolting cloth with suction, keeping everything as cold as possible. The crys- tals are washed with cold water, then with 20, 50, 80, and 95 per cent alcohol and ether, and dried in an oven at 100°C. The specific rotation of the preparation should now be at least —33°. A third crystallization may be performed, although it is useless to attempt to obtain a higher rotation than —38° or —39°. The washing method of purification of Irvine and Steele® may also be used, although the writer has not had experience with it. 2. Girasole tubers are not a satisfactory material for the pre- paration of true inulin, but are admirably adapted for investiga- tions of the whole group of inulin substances. 3. For the manufacture of a fructose syrup from these tubers, a preliminary separation of the inulin bodies and their subsequent hydrolysis to fructose is not feasible, because only a small propor- tion of the inulin bodies present will crystallize from water. In all probability the best method for making a syrup will be along the lines of the following procedure: (a) the extraction of the juice by diffusion; (b) the clarification of the juice by means of lime, phosphoric acid, and decolorizing carbon; (c) the acid hy- drolysis of all the inulin bodies; (d) the separation of fructose as calcium fructosate; and (e) the decomposition of the calcium fructosate and evaporation of the fructose to a syrup. 4, Dean’s hypothesis that inulin is a group of substances with large, loosely bound molecules, and not a single substance, is supported by the evidence in the present paper. AE vont likeodlie dy oF AR « airy den. ee i gnllemals “gh, Al a i } ‘Pe . art, la ovettt i FSO A aed \ eb iho wa) 0 es Litt Pelle ’ - - tty ea vy OY re Ay a? oe nalts ee cE iw ity) (eden oe ie any) oO oa Bite z Me aul Te HOPE ! a yt ’ hen weit Oat : Wales ednieo mina Ahriiss lh re rasadaand im toy Allene ae {0G pa fio © ah oh aiaeias s. chtnoRae ti nj dt ~ i i! Ps "y mlvenrgg te . meres. Per 00) aie oie mich utes: (MAT. { _ ne 4 ‘ ENG ay | as CL dim ' THE UNSATURATED FATTY ACIDS OF LIVER LECITHIN. By P. A. LEVENE anv H. 8S. SIMMS. (From the Laboratories of The Rockefeller Institute for Medical Research.) (Received for publication, January 3, 1922.) In previous communications on liver lecithin results were reported showing that the unsaturated fatty acids obtained on hydrogenation gave stearic and arachidic acids. It was further reported that the iodine number of the mixed unsaturated fatty -acids was 154, and that the mixed unsaturated acids on bromina- tion formed a bromide, which on the basis of its bromine content and of its melting point seemed to be the octobromarachidic acid. From these findings it was concluded that liver lecithin con- tained more than one unsaturated fatty acid, and that the indi- vidual acids differed in the degree of their unsaturation. It then seemed suggestive that the acid of higher unsaturation was the tetra unsaturated arachidic acid. Since there was no accurate knowledge either as to the number of the unsaturated acids present in the mixture or as to the degree of their unsaturation, there was no indication as to the ratio of the individual acids. In order to complete our knowledge of the unsaturated fatty acids, it became necessary to find: (1) the number of the unsaturated fatty acids present in the liver lecithin; (2) the exact nature of the acids of higher unsaturation; and (3) the ratio of the individual fatty acids. In the course of the present work it was found that liver lecithin yields only two unsaturated fatty acids, oleic and arachidonic! acids, and of these oleic acid predominates. 1 Since the tetra unsaturated arachidic acid is called arachidonic acid by Lewkowitsch (Lewkowitsch, J., Chemical technology and analysis of oils, fats and waxes, London, 4th edition, 1909, i, 211) this name will be adopted by the present writers. : 285 286 Fatty Acids of Liver Lecithin In connection with the finding of arachidonic acid as a con- stituent of liver lecithin, it is interesting to recall the work of Hartley,? done under the direction of Leathes. On oxidation of the unsaturated acids from liver fat, including phosphatides, this author isolated an unsaturated fatty acid which was oxidized to a tetrahydroxyarachidic acid. Furthermore, on bromination of the same fatty acids Hartley obtained a bromide, having a bromine content only 0.7 per cent higher than that required by theory for octobromarachidic acid. The present writers,? and Levene and Ingvaldsen,* ees pre- viously reported that on hydrogenation of the unsaturated acids of liver lecithin, arachidic acid was isolated. They have also obtained from liver lecithin a bromide, which on analysis gave a bromine value about 1 per cent lower than that required by theory for the octobromarachidic acid. In the course of the present work. an octobromarachidic acid was isolated which gave correct analyt- ical values. This acid was then reconverted into a tetra unsatu- rated arachidic acid, which on hydrogenation was converted into arachidic acid. From the product of bromination of the unsatu- rated fatty acids, fractions were obtained which analytically resembled the hexa- and tetrabromides. On purification of these fractions only the octobromide was isolated in pure form. Hence present day evidence does not permit us to assume the presence of more than one poly unsaturated fatty acid in lecithin. On the other hand, from the more soluble fraction of the bromination product a substance was obtained which yielded unsaturated fatty acids with an iodine number of 107. This is not far removed from that required by oleic acid which has an iodine number, of 90. From this fraction on hydrogenation pure stearic acid was isolated. Assuming that only two unsaturated acids, arachidonic and oleic acids, are present in the lecithin fraction, and taking into consideration the fact that the iodine number of the mixed un- saturated acids of this lecithin is 196, it becomes easy to calculate the ratio of the two acids from the equation 90% + 335y = 196 (x + y). It follows that the ratio is approximately 1.3 parts of oleic acid to 1 part of arachidonic acid. ? Hartley, P., J. Physiol., 1908-09, xxxviii, 353. 5 Levene, P..A., and Simms, H.S., J. Biol. Chem., 1921, xlviii, 185. 4 Levene, P. A., and Ingvaldsen, T., J. Biol. Chem., 1920, xliii, 359. nay | P. A. Levene and H. 8. Simms 287 The lecithin referred to above was obtained from the acetone extract of the liver. From the ethereal/éxtract there was ob- tained a lecithin of lower unsaturation. The iodine number of the unsaturated acids was 136. This corresponds to 4.3 parts of oleic acid to 1 part of arachidonic acid. Bearing in mind the molecular weight estimation of dihydro- lecithin reported in the previous communication, namely 810, one is justified in concluding that the liver contains several leci- thins and that the oleyl lecithins predominate. EXPERIMENTAL. Preparation of Material. Beef livers in 100 lb. lots were minced, dried, and extracted, first with acetone, then with ether, and last with alcohol. These extracts were treated by the methods described in an earlier paper. A cadmium chloride salt of pure lecithin was obtained which was free from amino nitrogen. Analysis 5. 1.0 gm. substance was hydrolyzed with HCl, neutralized, concentrated, and made up to 15 cc. 2 ce. of this solution: no N» (Van Slyke). Sa “ 1.95 ce. of 0.1 Nn HC] (Kjeldahl). Amino N Total N X 100 = 0 per cent amino nitrogen. The Unsaturated Fatty Acids. 765 gm. of the above mentioned material were hydrolyzed with 10 per cent HCl for 15 hours. The fatty acids which separated were dissolved in ether and washed with water until free from mineral acid. The ether solution was dried and concentrated to about 1.5 liters. This solution was allowed to stand at 0°C. over night. A white precipitate settled out. This was filtered off and recrystallized twice from ether and once from dry acetone. [t analyzed as follows: Analysis 15. 0.1019 gm. substance: 0.1184 gm. H.O and 0.2834 gm. CO». CisH3602. Calculated. C 75.93, H 12.76. Found. “75.84, “13.00. THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI, NO. 1 288 Fatty Acids of Liver Lecithin The ether mother liquors from which this stearic acid had been separated were then concentrated to dryness under diminished pressure. The residue was dissolved in methyl alcohol, treated with lead acetate solution, and made alkaline with ammonium hydroxide. On cooling this solution to—5°C. the lead salts of the fatty acids settled out. These were filtered off and extracted repeatedly with ether until ether would no longer extract material giving a precipitate with hydrochloric acid. The combined ether extracts of the lead salts were treated with HCl to convert the salts into free fatty acids. The ether solution was then washed with water until free from mineral acid and dried with sodium sulfate. The ether was removed by distilla- tion under diminished pressure. The residue of unsaturated fatty acids weighed 160 gm. The iodine number was obtained. 0.2281 gm. of substance absorbed 0.453 gm. of iodine by the Wijs method. Average molecular weight 290. Calculated (two double bonds). Iodine number 175. Found. a Fo OG. These unsaturated acids were dissolved in glacial acetic acid and brominated with a 25 per cent solution of bromine in glacial acetic acid. The temperature was kept as low as possible with- out freezing the acetic acid. The bromine solution was added from a special vacuum jacket burette constructed for the purpose. This consisted of two tubes of Pyrex glass sealed together at both ends, with a stop- cock attached. The lower end of the inner tube consisted of spiral capillary tubing to take up the thermal expansion. The space between the outer and inner tubes was evacuated and sealed. Bromine solution in this burette could be kept cool longer than in an ordinary burette. The acids were brominated in 5 to 10 gm. lots and allowed to stand over night with a slight excess of bromine. A yellow pre- cipitate formed which was filtered off. The precipitate (Fraction A) was expected to contain the higher bromides and the acetic acid mother liquor (Fraction B), the lower bromides. P. A. Levene and H. S. Simms 289 Fraction A, the Fraction Insoluble in Glacial Acetic Acid. Identification of Arachidonic Acid—The precipitate was ex- tracted repeatedly with ether until it no longer contained any ether-soluble material. This residue analyzed as follows: Analysis 17. 0.1066 gm. substance: 0.0356 gm. H.O and 0.0998 gm. COs. 0.1996 “ a :0.3166 gm. AgBr. C29H3202Brs. Calculated. GC 25.438, H 3.42, Br 67.72. Found. pp aeae, ce fo, °° Gea In melting point determination this material darkened, melted, and decomposed at 245°C. (A slightly less pure sample of this material previously obtained contracted at 240°C. and decom- posed at 244°C. while it began to darken below 200°C.) 20 gm. of the above material were obtained. 10 gm. of this were suspended in dry methyl alcohol and reduced with zine dust and dry hydrochloric acid gas. The methyl alcohol was filtered, diluted with water, and shaken repeatedly with gasoline. The gasoline solution was then shaken with water, dried, and evaporated to dryness under diminished pressure. The iodine number of the resulting material was obtained. 0.0492 gm. substance absorbed 0.150 gm. iodine by the Wijs method. C2oH3202. Calculated. Iodine number 335. Found. S05 This unsaturated material was reduced with hydrogen in al- coholic solution by the method of Paal. The solution, after being filtered, was concentrated and cooled to —5°C. The reduced acid separated out. This was recrystallized three times from dry acetone. It analyzed as follows: Analysis 34. 0.1034 gm. substance: 0.1176 gm. H.O and 0.2912 gm. COs. Co pH 4.02. Calculated. C 76.85, et? 91° Found. * fa.88;) “12:72. _ Identification of Oleic Acid—It was stated above that the octobromarachidic acid was extracted repeatedly with ether to remove the lower bromides. The ether mother liquors from this extraction were concentrated to 500 cc. and cooled to 0°C. A white precipitate settled out. This was separated by filtration and will be referred to as Fraction I. 290 Fatty Acids of Liver Lecithin The ethereal solution was evaporated to dryness under dimin- ished pressure and cooled to 0°C. Two oily layersformed. These were separated and dissolved separately in methyl alcohol. Each on cooling formed a sticky sediment. These were separated. The one from the upper layer will be referred to as Fraction IT, that from the lower as Fraction ITI. The methyl alcohol mother liquors from Fractions II and III © were then combined and reduced with zine dust and hydrochloric acid as described above. The resulting material weighed 30 gm. It gave an iodine value as follows: 0.2776 gm. substance absorbed 0.296 gm. iodine by the Wijs method. CisH3,02. Calculated. Iodine number 90. ' Found. = ”* AGG 10 gm. of this material were reduced with hydrogen by Paal’s method and recrystallized twice from dry acetone. In order to decompose any methyl ester which might have formed during the zine reduction the material was saponified with NaOH in methyl alcohol solution. The soap was precipitated in acetone, dried, and converted into free acid with hydrochloric acid in the presence of ether. The ether solution, freed from inorganic material by shaking with water, was dried and evaporated. After three more recrystallizations from dry acetone the acid analyzed as follows: Analysis 33. 0.0994 gm. substance: 0.1154 gm. H,O and 0.2776 gm. CO2. Cy1gH310>. Calculated. G 75.93, H 12 1Gs Found. 7645, “2 Oo: Melting point of stearic acid 70-71°C. Found. 68°C. Attempt to Isolate Other Unsaturated Acids than Arachidonic and Oleic Acids.—Fraction I, referred to above, was a grayish brown dry powder, difficulty soluble in ether. For purification it was extracted with hot methyl alcohol and filtered hot. The insoluble material analyzed as follows: Analysis 21. 0.1062 gm. substance: 0.0374 gm. H,O and 0.1030 gm. CO». 0.2041 <<‘ ef : 0.3030 gm. AgBr. C,sHgoO2Bre. Calculated. C 28.49, H 3.99, Br 63.26. CooH3202Brs. ss “< 25.43, .“ 3.42) "sGeened Found, “26.46, “ 4.04, “ 65.26. P. A. Levene and H. S. Simms 291 The alcohol extracted only a small portion of the material. When this extract was cooled to 20°C. a precipitate formed which was too small for further treatment. On cooling the mother liquor to — 5°C. asecond precipitate formed. This was separated and analyzed. Analysis 22. 0.1062 gm. substance: 0.0408 gm. H,O and 0.1108 gm. CO, 0.2026 “ 4 :0.3018 gm. AgBr. C,3H3.02Brs. Calculated. C 28.49, H 3.99, Br 63.26. Found. ae 42. ee ae. In a melting point determination this material softened at 130°C., darkening up to 170°C. when it partially melted and partially decomposed. At 240°C. it became liquid and decom- posed in the manner of octobromarachidic acid. Since the analysis of the material approached that of a hexa- bromide, the substance was extracted six additional times with a large volume of methyl alcohol. In a melting point determina- tion the residue became black at 200°C. and decomposed at 230°C. The combined extracts on cooling to 0°C. formed a pre- cipitate which analyzed as follows: Analysis 31. 0.2016 gm. substance: 0.3146 gm. AgBr. CisH3002Bre. Calculated. Br 63.26. C2oH3202Brs. se & Gial2: Found. “ 66.41. In a melting point determination this material softened at 140°C., darkening with rise of temperature, and melting with decomposition at 243°C. The methyl alcohol mother liquors were concentrated to 100 ee. and cooled to 0°C. A second precipitate formed which melted as follows. It softened at 140°C., partially decomposed at 180°C., and melted with decomposition at 243°C. Thus each fraction obtained from Fraction I on purification ap- proached the character of ‘the octobromide, hence it is justifiable to assume that this fraction consisted mainly of the octobromide. It was next attempted to isolate a tetrabromide from the material referred to as Fractions II and III (page 290). Each was dissolved in excess of ether and treated with 5 parts of gasoline. Precipitates were formed in each and were separated. These had a sticky consistency which indicated that they were not pure. 292 Fatty Acids of Liver Lecithin They were obtained in a quantity too small for analysis. The combined gasoline mother liquors were concentrated to small volume and again treated with gasoline. A precipitate was formed which analyzed as follows: Analysis 25. 0.1003 gm. substance: 0.0408 gm. H.O and 0.1118 gm. CO2. 0.1972. “ se : 0.2806 gm. AgBr. CisH;.0.Br,. Calculated. C 36.01, H 5.38, Br 53.28. C1sH3002Bre. a “< 28.49, “ 3.99, “ 63.26. Found. “<°30.39, 4.55, “Glee: The mother liquors from this precipitate were again concen- trated, dissolved in dry methyl alcohol, and reduced with zinc dust and hydrochloric acid. The reduced material had an iodine number as follows: 0.1990 gm. substance absorbed 0.264 gm. iodine by the Wijs method. Found. lJodine number 1238. The reduced material was dissolved in glacial acetic acid and rebrominated. This bromine will be referred to as Fraction IV. The acetic acid solution was concentrated to dryness under diminished pressure. The residue was dissolved in hot absolute alcohol and cooled to —5°C. A precipitate formed which was again dissolved in hot absolute alcohol and reprecipitated by cooling to —5°C. After reprecipitating eight times from absolute alcohol and once from methyl alcohol, a melting point determina- tion was made. It softened at 150°C. and melted with decom- position at 240°C. Apparently this is an impure octobromide. The mother liquors were concentrated and cooled to — 5°C. A precipitate was formed which, after several reprecipitations from absolute alcohol and from methyl alcohol, melted as follows: At 130°C. it softened and became black. At 230°C. it melted with decomposition. This also appears to be an impure octobro- mide. In order to ascertain whether a tetrabromide was present in the mother liquors from the purification of the last substance they were concentrated to dryness under diminished pressure and dis- solved in gasoline. Since no precipitate formed even on cooling to —5°C., it was concluded that no tetrabromide was present. ie P. A. Levene and H. §S. Simms 293 The three precipitates obtained from Fractions II and III were combined and also reduced with zinc dust and hydrochloric acid. The reduced material gave an iodine number as follows: 0.2815 gm. substance absorbed 0.488 gm. iodine by the Wijs method. Found. Iodine number 180. This material was also rebrominated in glacial acetic acid solution. It will be referred to as Fraction V. The bromide solution was treated in the same way as Fraction IV. It yielded similar results. Purification of the insoluble material gave an octobromide. Fraction B, the Fraction Soluble in Acetic Acid. This consists of the acetic acid mother liquors from which Fraction A was filtered. The liquors were concentrated under diminished pressure and treated with gasoline. An oily layer settled out. This was separated, dissolved in methyl] alcohol, and reduced with zine and hydrochloric acid. 20 gm. of reduced material were formed having an iodine number as follows: 0.1974 gm. substance absorbed 0.356 gm. iodine by the Wijs method. Found. Iodine number 180. The acids were rebrominated in glacial acetic acid. These bromides (Fraction VI) were treated in the same manner as Fractions IV and V and yielded similar material. Had there been any tetrabromides they would have been found in Fractions EV; V. and: VI. Lecithin Containing a Smaller Percentage of Highly Unsaturated Acids. A sample of lecithin was obtained from the alcohol-insoluble portion of the ether extracts of liver. This was purified by dissolving in acetic acid and treating with alcohol. The solu- tion after being separated from the precipitate which formed was concentrated under diminished pressure, emulsified with water, and precipitated with acetone. This precipitate was dissolved in alcohol and cooled to 0°C. A precipitate formed. The solution was separated, concentrated under diminished pressure, dissolved in ether, and precipitated with acetone. 294 Fatty Acids of Liver Lecithin This precipitate was converted into the cadmium chloride salt and purified once by the ether crystallization method. The purified compound was converted into free lecithin by means of ammonia in methyl alcohol solution. The lecithin was again converted into the cadmium chloride salt and again purified by the ether crystallization process, using a large volume of ether. © This material was free from amino nitrogen. About 1.0 gm. of material was dissolved in 15 cc. of glacial acetic acid. 2 cc. of this solution: no N (Van Slyke). cea “« 1 1.35 ee. 0.1 N HCl (Kjeldahl). This cadmium chloride salt of lecithin was then hydrolyzed with 10 per cent HCl. The iodine number was determined on the fatty acids after they had been freed from mineral matter and dried. 0.2120 gm. substance absorbed 0.256 gm. iodine by the Wijs method. Found. Iodine number 65. These acids were then converted into the lead salts and extracted with ether. The ether-soluble lead salts were converted into free acids with HCl. Their iodine number was determined. 0.2672 gm. substance absorbed 0.363 gm. iodine by the Wijs method. Found. Iodine number 136. _ a i CARBONIC ACID AND BICARBONATE IN URINE. By JAMES L. GAMBLE. (From the Department of Pediatrics, the Johns Hopkins University, Baltimore. ) (Received for publication, January 3, 1922. INTRODUCTION. In the course of some studies of acid excretion undertaken in this laboratory it became desirable to learn something if possible, of the factors which determine the elimination of carbonic acid in urine. Although carbonic acid may be regarded as only inci- dentally a urinary acid, the bulk of its production leaving the body by way of the lungs, it is, nevertheless, regularly present in urine, both free and bound as bicarbonate. Appraisement of the extent to which it may enter the urine is, therefore, necessary from the point of view of an understanding of the factors controlling the reaction of the urine, and is also required in quantitative studies of the acid-base content of urine. To explain the presence of HCO; in urine several conjectures are permissible: (a) that HCO; enters the urine from the blood plasma at a regulated rate, 7.e. is secreted by the kidney; (6) that BHCO;! entering the urine reacts with acid phosphate and liber- ates H»CO; to such an extent as will be determined by the phos- phate concentration; and (c) that H.CO; diffuses readily through renal epithelium with the result that the concentration of H,CO; in the urine is determined by that obtaining in the blood plasma. Considering the inappreciable part of the total HsCO; production conveyed from the body in the urine, its volatile character, and the fact that its concentration in the plasma is regulated by the re- spiratory mechanism, an active control of its elimination by the kidney seems unlikely. That the renal epithelium is no barrier . 1B represents a univalent atom of base. This convenience in notation is taken from Van Slyke (1). 295 296 Carbonic Acid and Bicarbonate in Urine to its obedience to the gas laws is a much easier assumption. An equilibrium between the tension of CO, in urine and in plasma does not, of course, preclude the derivation of HCO; from bicarbonate in the urine (see (b) above). The data presented in this paper were obtained with the purpose of testing the hypothesis of a direct relationship between HeCO3 concentration in urine to that obtaining in blood plasma. If the COz tension of urine is determined by the COs, tension of plasma, then both the H:CO; and BHCO; concentrations in urine are definitely indicated. H,COs3 will remain a nearly constant value, the variation of its concentration in-urine being no greater than in plasma, regardless of variation of the reaction of urine. The con- centration of H»COs; remaining stationary, or nearly so, the BHCO; concentration must then vary exactly inversely with the urinary pH.2 In other words BHCO; concentration will become a linear function of pH. Urine will thus always contain at a given pH approximately the same amount of bicarbonate. ‘The size of the BHCO; concentration at a given pH will be in such proportion to the value for the stationary H.CO; concentration as is determined by the it ratio which must obtain at thispH. The range of 3 variation of BHCO; concentration at a given pH should be rela- tively no greater than the slight variation (under normal circum- stances of metabolism) of the H.CO; concentration in plasma. The total amount of carbonic acid (H2zCO; + BHCOs) entering the urine will, owing to the variation of BHCOs; with pH, be deter- mined by the reaction of the urine, with the result that the total carbonic acid elimination in urine will decrease with increase in urinary acidity. If the above outlined consequences of a direct relationship between the CO, tension of the plasma and the CO: tension of the urine actually obtain they should be easily demon- strable by measuring in a series of urine specimens, the concen- trations of H»CO3, BHCO;3, and pH. The finding regularly in urine of the values for H»CO; and BHCO; postulated above would 2 The pH of solutions containing carbonic acid and bicarbonate being determined by the ratio of the concentrations of these substances, it follows that if one of the three factors in this equilibrium, in this instance (H:CO;), be immobilized, the other two, (BHCOs;) and pH, must neces- sarily vary exactly inversely. Judy Garfible 297 constitute satisfactory evidence for regarding carbonic acid elimi- nation in urine as a consequence of the CQO, tension in blood plasma. Such measurements obtained from a series of 55 urine specimens are presented below. Methods. Collection of Urine.—It was found to be a matter of essential importance that the urine be collected in such a way as to avoid as nearly as possible loss of CO2. The free carbonic acid tends to leave the urine rapidly after voiding. In consequence of the loss of HCO; there occurs a diminution also of BHCO; with shift of base to acid phosphate. Measurement of the H,CO; and BHCO; content of 24 hour urine collections necessarily gives low and irregular values.* It was decided to use in this study only freshly voided specimens. In order to minimize loss of CO, the voidings were collected in large test-tubes fitted with long stemmed funnels which delivered the urine in the bottom part of the tube. The tubes were nearly filled with urine and then immediately stoppered. Measurement of pH.—The pH of the specimen was determined colori- metrically reading to the nearest first decimal numeral of the logarithmic notation scale. Undiluted urine was used and comparison with the stan- dard carried out by the “comparator method,’ of Walpole, i.e. a tube containing urine was placed behind the standard and a tube containing distilled water behind the urine sample receiving the indicator. Phos- phate standards were used, the accuracy of which had been tested by potentiometer readings. Measurement of Total CO2.—Total CO: was determined in the Van Slyke apparatus for measuring the CO, content of plasma. The sample was taken from the bottom portion of the contents of the collecting tube and was introduced under 1 per cent NH.OH solution into the cup at the top of the burette. The sample was measured between marks on the pipette so that the upper part of the urine column, from which occurs the greatest loss of CO2, was left in the pipette. In order to obtain a satisfactory burette reading it was necessary to use samples of various sizes; viz., 5, 2, 1,and0.25cce. It was found that the amount to be taken was indicated by the pH of the urine. A single extraction was made and the burette reading 3 The only extensive study of the CO: content of urine which has been reported is that,of Denis and Minot (2). Their specimens were 24 hour collections which were kept in well stoppered containers. There was, however, apparently no precaution taken to prevent an initial loss of CO. Estimation of H,CO; and BHCO; from their data, which include also a measurement of pH, gives results which are widely variable as compared with the values presented in this paper. It is believed that the discrepancy must be due to loss of CO2 from their specimens during voiding and while the urine was being transferred to the container. ° 298 Carbonic Acid and Biearbonae in Urine corrected for CO: remaining in samples after a single extraction and for the air content of the samples. The factors for correction for unextracted CO, were calculated from the ratio of the burette capacity (50 cc.) to the volume of the samples plus the 1.5 ce. of reagents. The factors used for the amounts given above were 1.15, 1.075, 1.05, and 1.02, respectively. Their accuracy was tested by a series of double extractions. The value for air content of urine plus reagents was directly determined in a series of specimens by absorbing the CO: after extraction with 5 per cent KOH. The total air content, depending on the size of sample used, was found to be 0.11, 0.06, 0.04, or 0.02 cc. After multiplying the burette reading by the factor for unextracted CO:, the appropriate value for total air con- 273 tent was subtracted and a third correction for temperature {| ————— : Bae then applied. Calculation of “‘Free’’ and ‘‘Bound’’ CO:.—Fig. 1 represents graphically H,CO the ( : :) ratio over the range of urinary pH. The curve was obtained BHCO; BHCO,* H.CO; percentage of the total CO, which is free at various degrees of pH. These values were read from the curve (Fig. 1) and were used in estimating the H,CO; and BHCO; content of the urine specimens after measurement of pH and total CO, content. Accuracy of Measurements.—The accuracy of the measurements of free and bound CO; in urine depends in the first instance on the prevention of loss of CO. from the specimen while it is being collected. The collection error with the technique used cannot, unfortunately, be directly deter- mined. The necessity for the precautions taken and the size of the possible error in estimating H,CO; without them is indicated by the data in Table II, in which the results obtained by analysis of urine voided into a beaker are compared with those obtained from a portion of the same voiding collected in the manner described. The analysis of the urine voided into the beaker was carried out immediately, 7. e. the sample for analysis was introduced into the burette within a few moments after voiding the speci- men. The large error in the H,CO; measurement due to loss of CO: is apparent. In the case of the specimens in which the .phosphate con- centration is low, this error is due not only to the direct loss of CO2 but to a consequent alkaline shift of reaction. Decrease in pH greatly reduces the calculated value for HCO; (see Table I). ° from the equation pH = 6.1 + log In Table I is given the ‘Hasselbalch’s equation. For derivation of this equation see Van Slyke (1). Owing to the wide variation of BHCO; concentration in urine the extent to which it dissociates does not remain a constant. This equation with its fixed value for pK, will, therefore, measure the Hx,CO;: BHCO; ratio with only an approx®’mate accuracy. J. L. Gamble 299 It should be admitted that the technique used probably did not prevent a very slight loss of CO2. The values obtained in the most acid urine specimens in which the total CO: is a very small value are possibly some- what low for this reason. pH 40 42 44 4b 48 50 52 54 5b 58 60 62 MH bb 68 7072 TH 76 78 80 TTT. i AEROREREOTOP AONE HNNURUODOAAYAUENIR ONSATARNATSY ANTS TSATOHRGPAHRNITE 10 Fig. 1. Gurve indicating per cent of total CO, which is free or bound at various degrees of hydrogen ion concentration. A much more serious obstacle to an accurate estimation of H:CO; and BHCO; is encountered in the wide range of the total CO. content of urine. This range may be indicated as approximately 1 to 100 between pH 5.0 and pH 8.0. pH readings to the first decimal place give 30 points over 300 Carbonic Acid and Bicarbonate in Urine this range. The pH reading on which depends the partition of the total CO. into free and bound CO; is, therefore, not closely enough determined to provide more than an approximate accuracy in the estimation of H»COs; and BHCO;. This difficulty is increased over the middle part of the pH TABLE I. Percentage of Total COz as H2CO,at Various Degrees of Hydrogen Ion Concentration. | Free CO2 as Free CO: as Free CO> as pH per cent pH per cent pH per cent of total COe. of total COz. of total COs. 5.0 93 6.0 55 ao 12.0 eel 91 6.1 50 (fea! 9.5 ane 89 Oe 45 da? 8.0 ae 86 6.3 40 des 6.5 5.4 83 6.4 35 7.4 5.0 DAD 80 6.5 30 TE 4.0 5.6 75 6.6 25 The(t) 3.0 Sead 70 6.7 20 Th cath Dat 5.8 65 6.8 17 7.8 2.0 5.9 60 6.9 14 7.9 1.5 TABLE II. Showing Error in Estimation of H2CO3; Due to Loss of COz from the Urine Specimen. Specimen. Method of collectien. pH ss H2COs HPO, oh vol. ee gm. mol 1 Test-tube. jee 0.32 4.3 0.028 Beaker. 52 0.28 3.5 : 3 Test-tube. 6.3 0.70 bee 0.018 ‘a Beaker. 6.3 0.53 E's 3 Test-tube. 6.6 0.98 4.7 0.002 Beaker. 6.8 0.90 2.9 4 Test-tube. Call 0.60 Dee 0.008 Beaker. Wad 0.56 By: : H.CO; range by the rapidly shifting character‘of the ratio. Inspec- BHCO; sion of the curve, Fig. 1, and the values in Table I will make these points clear. ————— J. L. Gamble 301 Evidence Indicating Direct Relationship of the Elimination of Carbonic Acid in Urine to the H2CO3 Concentration an Plasma. As indicated above, measurements of the H,CO; and BHCO3; content of urine carried out in the manner described are to be regarded as of only an approximate accuracy. Collected, how- ever, from a fairly large series of urine specimens over a wide range of pH they serve satisfactorily the purpose of this study which is a statistical rather than direct attempt to learn whether or not there is a direct relationship of CO, elimination in the urine to the CO, tension of plasma. In estimating the significance from this point of view of the values for H,CO; and BHCO; found in urine, it should be remembered that since the HCO; concentration in plasma varies normally to the extent of about + 15 per cent of its average value, we may expect the same degree of variation in the values for HeCO; and BHCO; in urine prescribed by the assump- tion that they are determined by the CO, tension of plasma. The values for total, free, and bound carbonic acid found in urine over a range of pH extending from 5.0 to 8.0 are given in Table III. Inspection of these measurements shows very clearly that the H.CO; concentration in all of these specimens is a fairly constant value and bears no relation to pH. The average of the measurements of H,CO; concentration is 4.2 volumes per cent COs. Variation from this average is + 25 per cent. In contrast with this fairly steady value for H.CO3, BHCO; concentration varies over an enormous range and, moreover, is directly related to pH, the thousandfold increase in pH from 8.0 to 5.0 being accompanied by an approximately thousandfold diminution of BHCQO;. In consequence of this decrease in bicarbonate with increase in acidity of the urine there is also a rapid fall in the total amount of car- bonic acid (free + bound) entering the urine. This large decrease in the total CO, content of urine*which accompanies increase in ‘pH has been shown by Denis and Minot (2). In order to demon- strate that the regularity of the relationship of the total CO, con- tent of urine to pH is independent of variation in the phosphate concentration the data in Table IV are presented. These findings will excellently suit the hypothesis that carbonic acid elimination in urine is simply a consequence of the CO» tension of plasma provided it can be shown that the nearly 302 Carbonic Acid and Bicarbonate in Urine TABLE III. Measurements of the Total Bound and Free Carbonic Acid Content, and pH of 55 Urine Specimens. Subject.| pH | Total. | BHCOs| HCOs Sue | pH | Total. | BHCOs| HsCOs vol. vol. vol. vol. vol. vol. per cent | per cent | per cent per cent | per cent |per cent CO2 COz CO2 CO2 CO2 CO2 R. 8.0 430* | 426 4.3 R 6.6 16.6] 12.4] 4.2 G. 6.6 16.2] 12.1] 4.1 R. 7.9 |~350* | 345 5.3 G. 6.6 16.0 | 12.0] 4.0 G 6.6 15.9] 11.9} 4.0 G. 7.8 222* | 218 4.4 6.5 Od H, 6.4 12.9 8.4 |° 4.5 G. 7.6 174* | 169 5.2 G. 6.4 12.9 8.4] 4.5 7.5 G. 6.3 12.1 7.3 | 4.8 G. 6.3 11.8 7.0} 4.8 Ge 7.4 94* 89 4.7 G. 6.3 1sO.09-47 G. 7.4 88 74 4.4 W. 6.2 10.8 3.9 | 4.9 Es (sa) id 8" |) te a) ee 8. 6.2 Ss 5.3 | 4.4 G. 7.3 | 64.8 60.6 | 4.2 Gi: 7.9 (1) GO. 25 BH.St) \B22 H. 6.1 9.5 4.7| 4.8 G. 7.2 | 56.5 52.0 | 4.5 C. 6.0 8.5 “Bol eek SE S: 2 | 51.4 47.3 | 4.1 R. 5.9 7.3 2.9) 4.4 T. T2428 W430 44.3 | 4.7 te 5.9 6.5 2.6/| 3.9 G.t |) Lass 44.2) 4.6 G. 7.1 | 45.2 40.9 | 4.3 G. 5.8 6.3 2.2| 4.1 G. they 438.0 38.9 | 4.1 G. 5.7 5.9 Poo) 2.4 G. 7.0 | 39.4 oA. | 4.7 G 5.7 5.3 LO So. 1 G. 7.0 | 38.37 | 33.7 | 4.6 G. 7.0 | 35.5 31.2 | 4.3 G. 5.6 5.6 1.4) 4.2 G. 7.0 | 35.3 31.1 | 4.23 G. 5.6 5.4 1.4] 4.0 G. 6.9 | 36.77 | 31.6) &1 K. 5.4 4.4 i | o.7 ee 6.9 | 30.1 25.9} 4.2 G. 6.9° | 27.3 23.5 | 3.8 FT: 5.3 4.4 0.6} 3.8 B. 5.3 4.0 0.6| 3.4 G. 6.8 | 25.37 | 21.0] 4.3 G. 5.3 4.0 0.6 | 3.4 G, 6.8 | 21.7 18.0 |. oat * After ingestion of NaHCOs3. + After ingestion of Na2HPOs,. J. L. Gamble 303 TABLE I1I—Concluded. Subject.| pH | Total. | BHCO:! H:Cos| SY | pH | Total. | BHCOs| H:COs | ject. tol. tol. tol. : | vol. ; vol. , vol. ; CO. | COs | Co: "COR | OnA Oke F. 5.2 4.4 0.5-| 3.9 of 6.7 | 19.8 15.8 | 4.0 WY 5.2 3.7 0.4] 3.3 oaemy 119-8 | 15.8| 4.0 R 5.0 5 Ey A 0.2 3.5 D 5.0 3.5 0.2| 3.3 K 5.0 3.5 022.) 3:3 constant H.CO; concentration in urine is of an appropriate size in relation to the same value in plasma. The basis of the CO2 tension of the body fluids is the CO2 tension maintained in the alveolar air by the respiratory mechanism. The average COs tension in alveolar air is given as 45 mm. Dividing 45 by 760 this value becomes 5.9 volumes per cent COs. Multiplying 5.9 by 0.54 (the absorption coefficient of CO. in blood serum given by TABLE IV. Molecular Conceniration of—HCO; and—HPO, in Urine in Relation to pH. Subject. pH — HCO; | — HPO: | Subject. | pH — HCO: | — HPO: . = | | PD R. 7.9 0.157 0.016 ee dy Oe 0.009 0.017 G. 6 0.078 0.006 C. 6.0 0.004 0.028 ‘EA Gh 0.022 0.018 G. 5.3 0.002 0.017 G. 6.9 0.016 0.040 B. 5.3 0.002 0.040 G. 6.7 0.009 0.002 i ae 5.3 0.002 0.045 Bohr*), 3.2 volumes per cent CO» are indicated as the average HCO; concentration in blood plasma. The higher concentration (4.2 volumes per cent CO.) found in urine may be explained by accepting it as evidence of a higher CO, tension in renal tissue than obtains in plasma. Urine during its passage through the tubules of the kidney would quite certainly acquire the CO tension of renal tissue. A few direct measurements of the CO, tension of urine have been recorded. Fredericq (3), using a micro 5 Cited from Van Slyke (1). 304 Carbonic Acid and Bicarbonate in Urine tonometer found that the CO, tension in urine, collected without exposure to air, from dogs was in average (6 observations) 11 per cent of an atmosphere. Strassburg (4), a much earlier worker, gives as the result of three measurements 9 per cent of an atmos- phere as the CO: tension in the urine of dogs. Assuming that the absorption coefficient of CO, in urine does not appreciably differ from that in plasma‘ these values indicate 5.9 and 4.9 volumes’per cent CO. as H.CO;. They are thus still larger in relation to the usual concentration of H,CO; in plasma than the average value found in this study, and may be taken to indicate that the CO, tension in renal tissue is higher than in plasma. There is thus the probability that the H,CO; concentration of urine as it enters the bladder is determined by a CO; tension which is somewhat higher than obtainsin plasma. Presumably, however, this higher tension to which the urine is exposed in the kidney is directly proportional to the CO, tension of plasma and we may, therefore, correctly regard the elimination of CO, in the urine as a diffusion process controlled by the tension of the gas in the plasma. Relation of Carbonic Acid to the Reaction of Urine. Accepting 4.2 volumes per cent CO2as an approximate estimation of the average concentration of HCO; in urine, the average BHCO; content of urine at a given pH may be readily calculated by reference to the ee ratio obtaining at this pH. The estimat- ed average bicarbonate content of urine at frequent points in the range of urinary pH, expressed as volume per cent COs, and also as cc. of 0.1 N base and as gm. of NaHCO; per liter, is given in Table V. It will be noted that the values expressed as volume per cent CO, agree approximately with the direct measurements of BHCO; (Table III). It is desired to call attention by means of this table to the increasingly huge amounts of bicarbonate which are required to force the reaction of the urine in the direction of alkalinity. Fig. 2 is presented as further illustration of this point. Evidently the maintenance in the urine of a constant H,CO; con- 6 The solubility coefficient in water at body temperature is given by Bohr as 0.555 (cited from Van Slyke, 1). COs is thus only slightly less soluble in blood serum than in water. In urine it is doubtless not less soluble than in serum nor more soluble than in water. RA ee rs Re J. L. Gamble 305 centration offers resistance to an alkaline shift of reaction which becomes more and more effective as the amount of base entering the urine increases. For example at pH 7.4 the urine contains 3 gm. of bicarbonate per liter. In the absence of HCO; this large concentration of bicarbonate would force down the pH of the urine to 7.9 or 8.0. The small amount of acid phosphate in urine at pH 7.4 could only slightly check the shift. With the relatively small HCO; concentration present in urine about 15 gm. of bicar- bonate per liter are required to force the reaction to pH 8.0. It should be noted that the maintenance in the urine of a con- stant H.CO; concentration will offer effective resistance to depres- sion of pH due to base carried into the urine as alkaline phosphate. The events which will transpire following the arrival in urine of alkaline phosphate, HCO; being maintained, may be breifly stated as follows: H2CO; will deprive BsHPO, of some of its base to form BHCO; thus releasing a certain amount of BH:PO;. This proc- TABLE V. Estimated Bicarbonate Content of Urine, Assuming 4.2 volumes per cent CO. Unbound. pH BHCO: pH BHCO: Ll. t gm, ie gm. oh “ie cc.0.1N Joshi me bor cc.0.1N Babee 7.8 206 920 7.70 6.4 (ies 35 0.29 7.6 136 600 5.10 6.2 Sea! 23 0.19 7.4 80 360 3.00 6.0 3.4 15 0.13 12 48 216 1.80 5.8 2a 10 0.08 7.0 31 140 P45 5.6 1.4 6 0.05 6.8 20.5 92 Oia 5.4 0.9 4 0.03 6.6 12..6 56 0.47 nge 0.5 fe 0.02 Z fm BH.PO, H.CO3 ess will go on until the ————* and ———— ratios stand in equilib- 8 B.HPO, BHCO, rium, and the basis of this equilibrium, it should be noted, will be the ‘“‘normal’’ H.CO; concentration. Finally it should be pointed out that the reaction of the urine of a pH below 6.8 or 7.0 becomes largely a function of the A BH2PO, B,HPO, ratio rather than of the ratio which as Henderson has 306 Carbonic Acid and Bicarbonate in Urine pointed out determines the reaction of urine over the usual range of pH in urine. This being the case, a large alkaline shift of reaction is likely to occur after voiding, in specimens of a pH oH HCO, Se St 5o 58 60 OL b4 bb 66 1072 TH I 78 - 3 -0 BHCO, -10 -20 “40 a 2 S| -60 + = ro] oO -30 ¢ a eC. -100 w = > -1n0-S -140 -(60 \ -180 ~100 Fig. 2. Diagram of the HCO; and BHCO; concentrations in urine at various degrees of hydrogen ion concentration constructed on the basis of a constant H.CO; concentration of 4.2 volumes per cent COz. between 7.0 and 8.0, unless the specimens are collected with pre- cautions against loss of CO. This point has already been indi- cated in Table II, and is of practical importance in any study of acid excretion depending on a measurement of pH. J. L. Gamble 307 Regmaficance of the Usual Acidity of Urine from the Standpoint of Economy in Base Elimination. In the process of conveyance into the urine of the acid end- products of metabolism it is necessary that base be used with a regulated economy. Henderson (5) has pointed out the large economy of base in the elimination of phosphoric acid due to the fact that the pH of urine is usually much greater than that of blood plasma. The manner of this saving of base can be briefly described. Phosphoric acid is bound both in the body fluids and in urine as acid phosphate and as di-basic phosphate. In the body fluids at pH 7.4 the ratio of BH2PO.: B2xHPO; is 1:4, whereas in urine of average reaction, pH 6.0, the ratio is more than re- versed, becoming 9:1. From these ratios it. may be calculated that the base bound by phosphoric acid while being conveyed for excretion is 1.8 times (in terms of normality) the molecular con- centration of HPO, whereas in urine at pH 6.0 the base equivalent is 1.1. The elimination of phosphoric acid in urine of pH 6.0 is thus accomplished with a saving of approximately 40 per cent of the base bound by HPO, in the body fluids.. There is also a slight economy of base in the elimination of the organic acids normally present in urine due to the fact that a small part of their total concentration is unbound at the usual reaction of urine. In the presence of ketosis, with a large elimination of organic acids, this base economy although remaining small in relation to the total organic acid excretion, may equal in size the economy obtained in the excretion of phosphoric acid. Or —— THE CENCO HYVAC PUMP For Obtaining a Vacuum of 0.001 mm. is recommended for exhausting anaerobic jars, suction filtration of serums, cultures and culture media, and for all other laboratory processes requiring reduced air pressure. No. 10999 The Cenco Hyvac Pump consists of two stages, mounted vertically at a slight angle to each other, with the motors mounted on a common shaft. The horizontal shaft passes through a stuffing box which can be tightened up as necessary to prevent leakage. The shaft carries a 7-inch pulley, grooved for a 5/16-inch round belt which provides the proper speed of operation when connected to the usual] fractional horse-power constant speed motor. For the convenience of those who have motors available, the pump is listed unmounted, or mounted on a wooden base of the proper size. For those who wish a complete self-contained outfit, it is furnished mounted with the proper motor and belt on a single wooden base, 9x18 inches, complete with snap switch and 5-foot connecting cord, with separable attachment plug for connection to any lamp socket. The inlet tube ends in a hose-end for the convenient attachment of heavy rubber vacuum tubing. To prevent sucking back of oil into the exhausted vessel when the pump is stopped, an automatic trap, which effectively cuts off the flow of oil, is provided. Directions and a supply of special high vacuum oil are furnished with each pump. 10996. Pump, Cenco Hyvac, unmounted........ a aieysiete tins obtain (s,s eromane $50.00 10996A. Pump, Cenco Hyvac, mounted on hardwood base with belt, with- I? nO he i AIP arnt Be, os wierctpre ase or talats ... $60.00 10999. Pump, Cenco Hyvac, mounted on hardwood base, with motor No. A B C D AVC, D.C. Cee cena adit an.» 110 220 110 220 ee es ai aews x. $90.00 $91.00 $89.00 $90.00 For further information write for Bulletin 92K CENTRAL, SCIENTIFIC COMPANY LADCRATORY CEN UPPLIEX tur ical: 460 E.Ohio St., oer * Chicagis, epee USAA. The Mark af Purity RARE SUGARS Public epee ye Work Pfanstiehl Sugars because of their purity are used by the leading investigators and are the standard for bacteriology. 1-Arabinose i i titol La eer ai Raffinose (hydrate) d-Galactose Rhamnose (hydrate) d- Sorbitol (Sorbite) d-Glucose (anhydrous) Starch d-Glucose for Intravenous {free from Ery- oe Pisco we Soluble) ycero Glycogen F pee Inositol (Inosite) ylose Eoutest EXTREMELY RARE Lactose < d-Levulose from Inulin SUGAR d-Levulose from Invert A qonite (in research) Sugar d-Arabinose sear 1-Arabitol Maltose (hydrate) d-Mannitol (Mannite) d-Mannose Melezitose (Glucose-free) Cellose (Cellobiose) Dextrin (Bacteriological) Galactonic Lactone b-d-Galactose The following are available: Perseito] (a-d-Mannohep- a-Galaheptitol a-d-Galaheptonic Lactone a-d-Galaheptose a-d-Glucoheptonic Lactone a-d-Glucoheptose Glucosamine Hydro- cloride (Chitosamine) beta-d-Glucose d-Gulonic Acid Lactone a-d-Guloheptose Lactositol (Lactosite) a-d-MMannoheptonic Lactone a-d-Mannoheptose d-Mannoketoheptose Melibiose a-Methyl Glucose Quercitol (Acorn Sugar) d-Ribose d-Glucose for intravenous injection Carried by leading Distributors Write for our list of Amino-Acids, Reagents for Blood Chemistry, and Analyzed Laboratory Chemicals Specify Standard of Pfanstiehl SPECIAL CHEMICALS COMPANY DEVOTED TO THE CHEMICAL INDEPENDENCE OF AMERICA HIGHLAND Park. ILL. POLARS Our Booklet No. 1026 (S) on Polari- scopes, Refractometers and Spectro- scopes, available. recently printed, is now Kindly write for same. A well assorted stock of instruments, described in above booklet is at our disposal to permit prompt delivery. If you have to solve problems which require special instruments, kindly favor us with your wishes; we pride ourselves on being able to make suitable recommendations, since our staff is practically trained. Our manufacturing department is equipped to render repairs of any and all scientific instruments in an expert manner. 60 East Cy =aaD 10th St. Hay Fever Memoranda EARLY SPRING TYPE Patients whose hay fever develops in late March, April, or begin- ning of May, should be tested with pollens of early flowering trees as:—poplar, birch, maple, willow, walnut and oak, which polli- nate in the order named. LATE SPRING TYPE ‘ Patients whose hay fever develops in the latter part of May, dur- ing June or early July, should be tested with the pollens of sweet vernal grass, June grass, orchard grass, timothy and red top. The one giving the major reaction should be selected for treatment to the group. The unrelated rose pollinates simultaneously and is the primary or secondary cause in an occasional case—hence, should be included in tests especially where direct exposure exists. The same is true of dandelion, daisy and in some sections alfalfa. ARLCO-POLLEN EXTRACTS For Cutaneous Tests and Treatment cover early and late spring, also summer and autumn. Literature and List of Pollens on Request THE ARLINGTON CHEMICAL COMPANY Yonkers, New York The Endocrines, Digestive Ferments, Catgut Ligatures, Etc. HE ARMOUR LABORATORY is maintained for the purpose of handling the glands, membranes and other raw materials sup- plied by our abattoirs in immense quantities, from which important therapeutic agents are extracted and fabricated. Among the products that the physicians and surgeons use daily are: LABORATORY \ PRODUCTS 4 Mo ce |} ot Leg Suprarenalin Solution 1: 1000; Suprarenalin Ointment 1: 1000; Pepsin, U. S. P.; Pancreatin, U.S. P.; and other preparations of the Digestive Ferments that are used in stomachic and intes- Corpus Luteum; Suprarenals» U.S. P.; Parathyroids; Pitui- tary, Whole Gland; Pituitary, Anterior; Pituitary, Posterior; and other glandular substances in po. and tabs. Pituitary Liquid in 1 c.c. and 3 cc. tinal disorders and as vehicles for : ampoules. nauseating drugs. A We also make Sterile Surgical Catgut Ligatures, plain and chromic, boilable; and Iodized Ligatures, non-boilable. {The Armour ligatures are made from Lambs’ gut, selected especially for surgical purposes and sterilized at opportune stages in such manner as to preclude the possibility of contamination in the finished strings. We are headquarters for the Organotherapeutic Agents and are always glad to coorerate with the medical profession ARMOUR 480 COMPANY CHICAGO, U.S. A. BOVIE HYDROGEN ION OUTFIT A convenient arrangement of Hydrogen Calomel Cell with special support as used by the author in connection with the Bovie Potentiometer for the investigation of biological and other fluids available only in small quantities . 4900 Bovie Hydrogen ion Outfit, with diagrammatic illustration of connections The Bovie Potenticmeter (See Journal of Medical Research, 33: 295, 1915) is a semi-automatic recording potenticmeter by means of which H ion concentrations are directly plotted as a function of some measurable, variable quantity with which the H ion concentration is changing. The variable quantity may, for example, be amount of titrating solution added or elapsed time. Records are made on a special cross- section paper by means of a marking device automatically moved along the H ion scale as the Potentiometer is balanced. This cross-section paper has millimeter divi- sions along the horizontal scale and fourteen sections of logarithmic divisicns along the vertical scale, the latter reading H ion concentrations directly. | It measures and records both the actual (effective) and the total (titratable) re- actions of solutions, and plots a curve of actual reaction as amounts of acids or bases are added to the solution. The curves obtained shcw the changes in H ion concentra- tion of a solution with any other variable, such as time, growth of organisms, or the addition of acids or alkalies. 4900. Hydrogen ion Outfit, consisting of Bovie Potentiometer, Bovie Hydrogen Calomel Cell, with special support and two dry cells. Code With resistance for galvanometer lamp for operation on 110 volt Word CIRCUIO. ooo sod oe cence cveccccuectbeces ct Dre: ee 283.00 Ezvlf 4890. Bovie Potentiometer, only... ... .c.c.i.1 v DHDDDDDNDNDDDDNDNDMDMD GO OD) = rT DODADDDDHDODDDNDNDAH DH ny ) “1110}01 FE HODDDDHDHNHNHNDDNHDMOMDMMO HO és 701d 5S HHH HHH HHH HHH HH HHH SH DOOD BDAADDAAADAAAGaDS om ‘ayerpAyoqied SE BHBDBADABAGDGBABDAAGS oF > Re ee ee ee ee a ee ee ee Oe | SBHHAMHMORDAOHAMHMNON oO I oe oe ee ee = S. ys Q ~s Soy 2s SSS eae a yee * & oe o fo SSS 318 | 19° Z—| LP O\LS0 960 Goh |I88 |P'S8 |6°6 |O'ST | FL » 1°18 |0L°0 |24°0 | 02—|0'98 46° 1—| L8°0/99°0 |I6'T g¢'g 88 E8166 'ST|SI » 93° S—| Gr OLF'0 |68°% 8h I88 |P'€8 66 |9'ST |ZT » ZO'S—| SS O|TL'0 |02'°% 08°6 (09'% |I88 |F'€8 |6°6 (OST IIT ,, €°ze |89'0 [0L°0 | SZ—|T'S8 |6°8S |88°% |6T0'O\G8z'O|S8' Z—| LP O}FS'O \2e°% €9'8 ehh 188 F866 ST /OL » 8'TS.|89°0 |69°0 | ZI—|e" 28 LL’ 3—| Sb'0IZ9'0 |26'0 ISc'6 |GE'F [88 |F'e8 166 'STI6 » ZI'S—| 29°OZ8'0 |TS°% Ge'T [ce'Iz02'e |I88 |F'e8 |6'6 OST |S » ¢'T |89°0 |89°0 | PI—|2'88 26° P—| OF OLS'T (00'S S6'T [0S'8zl0G'9 |I88 |F'E8 166 [STIL » 0'TE |89°0 |69°0 | ST—|2'°6e 9G¢°E—| LG°OS'T 00'S [G°s |PI'ZZ/PT'S |T88 |P'E8 |6°6 9'ST 19 » [A 9°88 |0Z TT/SS0'0|Z8z 0|62'F—| G8°O88'T |4h'°S |1Z'S [ST PEGCO' TIO TE S01 6E |9°6 r Av 6°18 |0L°0 |89°0 | 6 —|h' Th Zh O—| 8O'IiFr'T [G88 [82° |S9°2900'ZTIzZ ‘162819 SOTe'S |b 5 L'1e |12°0 |69°0 | OL—|8"OF |6'98 [OSE |Es0'Olsez O|g0' O+] 9E"TIZe'T [OG "2TI86'S OZ GOSS OTIIZZ‘Te'zel9'sOoTe's | » LTE |0L°0 |69°0 | TI—|8'6E |2°Sh |Te'T 670" 0/94 O/G2' T+] OF TIFH'T [29'S [94'S [80 GSES 'FIIZZ ‘TI6ZET|9'EOT!E'S |S » 8'1e |€2°0 |TZ°0 | 6I—|6'98 82 I—| 69° TOP'T \S8°ZI|IP's 0c F998 ZTIZZ ‘T6'Ze19' soTe's | 1 ABI A 16° 0—| OS 'TIS8'0 |eL'F [96'S [OG SSS" ZT|IZZ ‘TI6 LZET|9 SOT!e’s | 08 “Ady IZ [020 |0L'0 | II—|8'OF |0' LF 080°0 62 1—| OF T1660 |68°S [96'S [02 0S9E OTSTES ‘TIT'66 | 6 |S'T | 62 » 0'2E [FLO |82'0 | ST—[G'6E |F'OS |TZ°T [920° O|Z6T 'O]66' Z—| ZET\PL'T [68'S [92°C PP 1990 STGTE ‘TIT'66 |2°F6 |S'T | 8% » 9'TS |9°0 |99°0 | IIT—|T ‘OP £6 O—| 92° T1060 |Z8'F |86°% [OF GOO OTST ‘TII'66 |Z'F6 [8'T | 26» AI [2°18 |69°0 |22°0 | 8IT—|0°ze 10° Z+| 02'T/28'0 \9%°% |10°% [88° 2z/00 STSTE ‘TI 66 |Z'F6 |S'T | 9% “Ady Le \TA:0 (62:0 | St—07 28 86° Z—| TO'T/G9'0 |68'T |242°0 |42°8 |eh OT/OTO ‘TIS’S8 |6°9P |2°0 |GZ » CL S—| ZI'TG6'0 [06'T |I8'T [99 PZGZ'ST/OTO ‘TIS"88 |6°9F |L°0 | Fo » LL: G—| €0'Teb'T \e8°% |88°T [09 SzjZz TOTO ‘Tie'88 6°9F |L°0 |Z» €°eE |29'0 |89°0 | FZ—|L' FE |6'OF 020 O|SZI 0/86 S—| 8608S 'O (00'S |9E'T [06 STEP ETOLO ‘TESS G'9F L°0 |Z» [II |I'¥e |F9°O |89°0 | 42—|T "FE |9'99 020 0/991 ‘O/#0' F—| GO'TZ8'0 [S8°% |6z°0 (00H [PS IT9TO ‘Te ’8s 6'9F \4°0 | 12 “Ady 4Z° 0—| 08°0/80'0 |s¢"T 92'S [80'S 690 ‘198 |8'Sh 6°6T |0Z eh b—| Sh O10 620 BRO) be Die lee ele ae II 89° ¢— L1'0 |96'0 [2z'0 |Gz'T jgo'g | 0 | O | O | O {SI ady G°0E |89°0 |69°0 | &I—|6'°8E ZG I—| SSOIT'T PS°9 [ST OF OS94°S |269 ‘TIP IST|S 9% |9°Gh POE |89°O |0L°0 | FI—|€8E 00° T—| FS°0Z80 |FS'9 [SI'T |0S SF's |269 ‘TIP TST|S°9% [9° GF 908 |12°0 |z2'°0 | ZI—|2'68 |2'FP |FI'F [S20 O\Z1E'O|22 O—| GF O90 \2a°9 |PS'0 \0E Sh 96°F 669 ‘TIP TST|S 9% |9 CF G08 |h4°0 |hL°0 | OL—|€ OF LE 0—| Sh°0\8S'°0 9F'E 06°68\19'F |Z69 ‘TIF TST|S°9z% |9° CF P'0S OF'% |0€0 O/E0E'O/9G T—| 99° O9'0 OFF 9Z°T (06 6F08'S |Z69 ‘TIF TST|G°9% |9° ZF 8°0E |29°0 |0L°0 | AT—|€ "LE 9T' O—| 88°0\9S'°0 |2E°S OL 8E0F'F [269 ‘T|F TST|S° 9% |9° SF G0 |99°0 |69°0 | 06—|6'SE 18° 0-—| 1F°009°0 96°% 03 SS|IT'S |Z69 ‘T\F TST|¢"9% |9° ZF [ITA|O'TE |29°0 |69°0 | Go—|o'SE Sh I—| so O\SS°0 \22°S 00°S8iz2°S |Z69 ‘TIF TST|S°9% |9 SF €°TE |04°0 |22°0 | d2—1€ SE ¥G'0—| 990/670 ZI'c GbE |10Z ‘T|TGOT|Z°0€ |2°F1 oT 169°0 102°0 | 9I—|8° 2 8¢°0—| 6h 00S "0 |249°% 6F'S |10Z ‘TIT 6OT|2°08 |L° FT z'1e |0L°0 |22'0 [88—|¢ 08 [0° 6F |0&°S [920° O/00E ‘0/80 0-F| 9F'O\SS"0 |26°E €8°P |10G ‘T/T 601/208 2° FT IIA el 0+] OF O16S'0 88"E 82°F |10Z ‘TIL 601/208 2° FT quao by ee dad ub “Ub “Ub “ub “wb | ‘wh | “wb “mb | “Ub “Ud Ud “ub “700 sa] = vA bg oo 4 Q se > mn A a] i ra o Q ea ca] ie] tg Q 8 g. o| 8 | Be a | & + | $3 re ma 2a haa Met al a Ee | & S et 3 y pe) ESE a TEIEEL ia [elf] ge ele lee) ei eje|e| |elé pe Ne Gee Mie Bie he Ble Bie Se ese are Bt Bieler ie Cal ecein | see ater sliae Bae BS 3 © ce = e S oo 4 + © o sh : ) a SE lice lea geet gl oe faa |e lo peal tee TIRE iotnee aoe f & 2 pO, 9 Be 8 Se |) 1S 5 ¢ = 2 o e iotet | ee ir i ef | 8 B Bl g g : cP a he s ? : ce » © 8 ® a - : @ Ofc URES ie ae 8 B z OS ra =a o i 4 ‘uIsTpoqeyou A10;BITdSexT *poolg ‘ouTIN JO uottoeds NOY FZ ‘OyBPUL POO ————————— Oe ves OZ iy oan 8S ” cern cise 0G » 6T Avy SI'S: 3 aes OT ” GI AVN S oD Té61 ‘978q “papnjIUoQ—lT ATAV.L ‘UdSOIPIU [BOO] OJ OPVUI SVM DIUBMOT[V OU DdUL[LG UoSOI}IU OY} JO UOT}B/NI/BO oY4 UT "TX polog YYIM popiodsed YSnoyye KX polog 07 ssuojoq ‘alojotoy} ‘puve Z oune jo ByVp AIVUTIN OY} YIIA poqgordioqzur oq Jsnur g ouNL Jo Vyep ALOyvaTdsor oy} ‘ojdwexd IoJ ‘pottod Surpodoid oyy 07 sduojoq AT[vos potted Mou B JO Suru10Ul JsIy oY} UO 480} WAST[OGLJoW BV YvYZ puw Ap 4vY} IO} popsOdeI poo] 9Yy4 Aq pooUoNPUr yOu o1OA\ 8404 OY} IVY} potoqurouIo OG Ysnul 4I WAST[OGuyot AIOZVIIdSoL [VSVq OY} UO POoO}J JO doUSNYUT OY} OUTULOJOp 04 so[qvy OY} BuIsN uy ‘posn o10M Avp snorAoid oy} Jo SIsSATVUIIN WoIy vYEVp 91 QuOTyOND ATOZvATdSoOI UTOYOI1d-UOU OY} SUIye[No]vo Uy ‘uadorjyiu U19}01d-UOU pooTq puU¥ ‘oUIUIYBvOID pool ‘Born poolq ‘sopr1o;yo vousv]d oy} ‘oulIn 94} JO Soplo;Yyo oy4 puv ‘ouraryvorto oy} ‘UOT}VIYUSDTIOD UOT UdssoOIpAY oYy ‘AZIpPIOV o[qvyvs41y OY} ‘SUOTJVUTTUeXO OUTIN AIvUIPIO ‘ooUL] eq 1048M : 90Rds JO YOR] LO] 9[Gvy OY} UI pojyw1OdiOoUl U99q JOU OABY YN POUIV}GO o1OA BUIMOT[OJ OY} ‘OAOG’ poyL[Nge, VYep OY} OF UOIJIPpw UT OPSFUL FBF OF JO UOTZBNGB, oY} UI opeUr USSG SLY SOUVMOT[V SITY} ‘SOD9J OY} UL PUNO] SUA POO} OY} UI 4vJ OY} Jo 490 Tod g Jo OdvIOAR UL SY ‘savok 6Z = O8B {'uId GTZ = ysoyo f'ud G'z7g = (Buryyis) yYSroy {*wd O'ZGT = (Surpuvys) VYSoP] yl Meee 69°0 (69°S | 09°F [900 ‘TIT'F6 |2°6T |€°SI | ZT» IZ'0 |€2°0 | 91-6" LE |9°9F |G2°% [$20 0/0SZO|ZF' T—| 8h 0ZL°0 S6°< | 29°F [900 ‘TIT F6 |2°6T EST | IT» cr I—| 8h '008'0 €6'T 8L°9 09°F |900 ‘TIT'F6 |2°6T |E°ZT | OT», ¢9°0 |49°0 | 8I—|6 9& 62° T—| FF 01680 |T8°2 80°6 F6'F |900 ‘TT F6 |2°6T [€°2T 16» 68° S—| OS OFE'T OS'S 18°0 08 9TFPG'S 900 ‘T|\T'F6 |Z°6T [EZ |S» 89°0 |89°0 | HI—|9'8E G8’ Z—| 9S OSS'T |ZO'F OTT |26°ST/00'9 |900 ‘TIT'F6 |4°6T [EST | L » PI'S—) 9S O9'T |ZL'S |6T'S 88"Esl6s"S |900 ‘T/T F6 |2°6T [EST }9 » ZI S—| P9'OO8'T \8Z°L \24h'S |6L° 221429 [900 ‘TIL'F6 |2°6T |€°2T |e» c9°0 |29°0 | SI—|6°2E 19° 3— 6L°T OF OLT9'S [OL S892°S |900 ‘TIT'F6 |Z°6T [EST |b €24'0 |02'0 | $ —|2°GP be 9—| GL°O29'T (69'S ZE°T (06 FHSS'S \222 (0°C9 [OFT [9 E |e Sune 69'°0 |89°0 | b —|8°GF |8"Ez [08S [6G0°O|S8z'O|ZE° T+] 00TH '% |FS"S \PS"E |00'S9OF'ST|TZO TIE 9ZT/S"FOT|S'S |S » 120 |69°0 | T —|t FF [8°92 |20°S |090°0/0FZ'O/ZE° I-++| Zh T98"T ZS ET/c9'S 09 6SOF 'ST/TZ9 ‘Tig 9ZT/8'FOT|S*S | 1 unr 99°0 |99°0 | $ —|L°GF HE O+) ST TSO'T |G2'OT/89'S OS FIEF OT|TZO ‘TIS 9ZT|S"FOTIS'S | IE » ¥9°0 |89°0 | €I[—|0°6& ZO’ Z+| 201290 |\Z0'F GE'S ES" ESS2 FI[TZ9 ‘Tie 9ZTI8 FOTIS"s | OF AVIN 88° S—| 28°0'28'°0 06'S |6E'T OT OS 82'ST/999 |S°6E |0'OF [OEE | 66 » €8'S—| 62°0\89'0 |22°0 [28'T 22°24 68'S | O 0 0 0 |86 » 99°0 |29°O | IL—|€'6& [4 Ob |V9'T SEO°O/LTE “OSh &— 91°049'0 |19°0 \88'T 0S'9 |sh's | O 0 0 0 | 22 AB 321 322 Metabolism of Diabetes During this time an average daily excretion of 37 gm. of glucose and 10.74 gm. of nitrogen occurred. The D:N ratio fluctuated somewhat, averaging 1.52. The blood sugar was 0.294 per cent on admission and a slight acidosis was indicated by the ammonia nitrogen excretion which exceeded 1 gm. and by total acetone bodies of over 3 gm. in the 24 hour specimen of urine. The carbon dioxide-combining power of the plasma was 47.0 volumes per cent onadmission. A daily administration of 15 gm. of sodium bicarbonate was commenced Apr. 11, but discontinued Apr. 20. Thereafter no drugs were tised. The average basal metabolic rate for the period was —12 per cent, the average respiratory quotient was 0.68, and the non-protein respiratory quotient, 0.65. The morning of Apr. 7, the very questionable non-protein respiratory quotient of 0.61 was obtained. The utilization of glucose from the carbohydrate of the diet and from protein metabolism averaged 24.6 gm. for the first 6 days, but after that, fell to an average of 21.2 gm. for the rest of the period. Period II.—Apr. 18 and 19 the patient was fasted. Apr. 20 she again received the food given in Period I. The beneficial effect of the 2 fast days confirmed the general belief in the value of an occasional fast day in the management of diabetes. Apr. 19, the urine was free from sugar. High Protein, High Fat Diets with Development of a D:N Ratio 3.65; 14 Days, Periods III, IV, and V. Period III.—From Apr. 21 to 25, the patient was placed on a protein-fat diet of meat, butter, eggs, and soy bean bread, containing less than 1 gm. of carbohydrate, 46.9 gm. of protein, and 88.3 gm. of fat. This was con- tinued 5 days without much appreciable effect. The carbon dioxide-com- bining power of the plasma remained normal. The D:N ratio did not exceed 1.88. The basal metabolic rate fell to —27 per cent Apr. 21, the result probably of the 2 fast days, and had risen to —18 per cent by Apr. 25. Period IV.—From Apr. 26 to 29, the protein allowance was increased by approximately 50 gm. to twice the previous amount, the fat of the diet remaining about the same as before, and the carbohydrate not exceeding 2gm. The effect of this change was evident immediately. The urine of the first day showed a D:N ratio of 2.01. Apr. 27, 28, and 29, the ratios were 2.98, 2.75, and 2.96. Acidosis, however, was still slight and the patient felt entirely comfortable. The total acetone bodies in the urine of Apr. 29, were 2.39 gm., that is, not higher than the acetone excretion during the preliminary period; the carbon dioxide-combining power of the plasma remained normal, 47.0 volumes per cent, and the ammonia nitrogen excretion averaged less than 1 gm. The pernicious influence of this added protein on sugar utilization is very striking. Unless we assume a flooding out of stored glycogen, which seems unlikely in view of the long preceding period of very low carbo- hydrate intake, or unless nitrogen retention had occurred, for which there is no evidence, the rising D:N ratios indicate a sharp drop in sugar toler- Wilder, Boothby, and Beeler 323 ance. This break in tolerance cannot be attributed to increased acidosis or ketogenesis since neither of these increased over that of Period I. Another significant result of raising the protein may be the elevation that takes place in the basal metabolic rate. This increases from an average of —20 per cent for three determinations between Apr. 22 and 26 to an average of —14 per cent for four determinations between Apr. 27 and 30. This is likewise unexplained by acidosis and seems to be attributable to a prolonged cumulative effect of the specific dynamic action of the extra protein (25, 26, a). Period V.—From Apr. 30 to May 4, the fat in the diet was increased by 39 gm., and the protein by 10 gm., the carbohydrate quota remaining practically the same. The result of this was a rapid loss of the remaining ability to utilize sugar and the excretion of glucose in amounts equal to the maximal theoretic yield of glucose from the protein metabolized. The D:N ratios obtained in the 24 hour collections of urine of May 1, 2, 3, and 4 were 3.41, 3.76, 3.98, and 3.78, or an average of 3.73. The non-protein respiratory quotients of May 2, 3, and 4 were 0.70, 0.71, and 0.70. The rate of excretion of acetone bodies increased, the blood acetone rose to 0.049 May 2, and 0.053 May 3, and the carbon dioxide-combining power of the plasma fell to 36.9 volumes per cent May 3. The percentage of fat in the blood increased very rapidly. The basal metabolic rate rose only slightly, from the average of —14 per cent of the previous period to an average of —10 per cent for this period. In spite of the increased acidosis and a D:N ratio of 3.73, the elevation of the basal metabolic rate was not, therefore, as great as that caused by the increase of about 50 gm. of protein in the preceding period. The morning of May 5, the patient complained of fatigue and was drowsy and it was considered advisable to change the diet. Low Protein, Relatively High Fat Diets with Recovery of Former Carbohydrate Tolerance; 14 Days, Periods VI and VII. Period VI.—Beginning with the noon meal of May 5, and continuing until May 14, the diet was made similar to the therapeutic diets em- ployed by Newburgh and Marsh (29). The daily allowance was 15.6 gm. of carbohydrate, 9.9 gm. of protein, and 83.4 gm. of fat, a total of 881 calories. According to the Du Bois (13, 19) formula the basal calorie requirement for this patient, 158 cm. in height and weighing 31 kilos, would be, 1,070 calories, provided the basal metabolic rate was normal. This, however, fell within a few days after the new diet was instituted and at a basal metabolic rate of —20 per cent, 856 calories suffice for the basal energy exchange. As the patient was kept in bed, the new diet provided, therefore, nearly enough calories to avoid calling on endogenous food sup- plies. The therapeutic results of the diet on this completely diabetic subject with increasing acidosis and impending coma are very satisfactory, as the daily figures in Table I reveal. Within 7 days the urine became sugar-free and ketogenesis was practically controlled. The ferric chloride 324 Metabolism of Diabetes test of the urine was negative within 2 days and the rate of ammonia excretion fell almost equally rapidly. The carbon dioxide-combining power of the plasma reaccumulated to 58.9 volumes per cent by the 5th day and the blood fat was appreciably reduced. The postabsorptive blood sugar value, however, remained elevated. The nitrogen excretion fell off rapidly reaching 3.55 gm. on the 9th day. The figures for the respiratory metabolism reveal a decrease in the level of the basal metabolic rate to —20 per cent on May 13th. The respiratory quotients remained unchanged and their failure to mirror improvement in carbohydrate utilization may be accounted for in a slight measure by glycogen storage, but in a large measure by the fact that the tests were made in the morning before breakfast, too long after the last feeding to be affected by the small amount of carbohydrate ingested. The rapid control of ketogenesis and acidosis, which the new diet accomplished, is fair evidence that sugar was actually burning. Judged by data from urinalysis the patient’s sugar tolerance at the end of the period exceeded that of the preliminary period. The average daily sugar utilization for the 3 days, May 12, 13, and 14, equaled 30 gm. Period VII.—Between May 15 and 18 the protein allowance was in- creased to 30.7 gm. and the fat allowance to 109.1 gm. Traces of sugar reappeared in the urine, but otherwise the condition of the patient re- mained practically unchanged. The glucose utilization for these 4 days averaged 29.7 gm. The increase in protein to 30.7 gm., which is nearly 1 gm. for each kilo of the patient’s body weight, had no such cumulative effect on the basal metabolic level as was observed in the higher protein periods, nor did the increase in fat to 109.1 gm. alter the basal metabolic rate which continued —20 per cent. Low Protein Diet with Fat Sufficient to Make It Isocaloric with the High Protein, High Fat Diet of Period V, and Carbohydrate Sufficient to Yield Available Glucose Equal to That of Period V; 8 Days. Period VIII.—From May 19 to 26, the diet contained 42.6 gm. of carbo- hydrate, 26.5 gm. of protein, and 151.4 gm. of fat. Our purpose was to compare the effect of this high fat, low protein combination with that of the former high protein diet of Period V (Apr. 30 to May 4, inclusive). The two diets are practically isocaloric, the former yielding 1,721 calories and thelatter 1,692. Bothdietsshould theoretically yieldnearly the same amount of glucose. In the former, this was derived from 3.3 gm. of carbohydrate and 58 per cent of 103.6 gm. of protein; the latter contained only 26.5 gm. of protein which could yield only 15.4 gm. of glucose, hence 42.6 gm. of carbohydrate were added toit. The two diets are practically comparable in energy value and available glucose, but different in protein. They may also differ slightly in ketogenic properties by Woodyatt’s (45) assumption that 46 per cent by weight of the protein of the diet should be calculated as the equivalent of a like weight of oleic acid in ketogenic properties. Wilder, Boothby, and Beeler 325 The diet of Period V contained the equivalent of 171.7 gm. of theoretic oleic acid (F. A. = 0.46 P. + 0.9 F.) while that of Period VIII has only 148.5 gm. On this diet, as was to be expected, sugar reappeared promptly in the urine. The amount increased gradually, but did not exceed 50.4 gm. The D:N ratio was 1.26 on the 4th day, 1.13 on the 7th day, and 1.35 on the 8th day. The excretion of acetone bodies increased, but did not exceed 6.54 gm. The ammonia nitrogen excretion remained below 1 gm. a day until the 8th day when it reached 1.11 gm. The carbon dioxide-combining power of the plasma was 50.4 volumes per cent on the 4th day, and 44.7 on the 6th day. The blood acetone value of the 6th day was 0.025 gm. The basal metabolic rate increased from an average of —20 to —12 per cent May 23 to 27. The non-protein respiratory quotients continued to hover close to 0.70 in tests made 14 hours after a meal. Quotients in tests conducted 3, 1, 2, and 3 hours after meals showed a peculiar behavior, to be discussed under respiratory quotients. ' Subjectively, the patient continued free from complaints. Pulse and respiration were normal and weight remained constant. It was evident that the diet was tolerated much better than the former isocaloric diet of high protein; while there was a general tendency toward a lower sugar tolerance this was by no means so marked as it had been on the high protein diet, and even on the 8th day there seems to have been a utilization of 13.2 gm. of glucose from protein. It is impossible to know whether to attribute such effects as were ob- served to the high fat value of the diet or to the high sugar value. The latter was well above the patient’s tolerance of 30 gm. of glucose and it is a generally well established principle in diabetic therapy that the feeding of carbohydrate in excess of sugar tolerance will depress such tolerance as exists. Had the high fat diet been given with a smaller amount of carbohydrate it is possible that it would have been tolerated even better than it was. Period IX.—2 fast days were instituted at this time, May 27 and 28, and on May 29 the food contained 33 gm. of carbohydrate, 40 gm. of protein, and 39.5 gm. of fat. The purpose of this fast period was to correct the moder- ate acidosis which had been provoked by the preceding diet and to restore the patient to the favorable condition of the beginning of Period VIII so that the effect of the diet of the following period could be compared with that of Period VIII. The benefit of this fast was less striking than that produced by the 2 fast days which followed Period I, suggesting that the diet of Period VIII had produced not only an immediate deleterious effect, but one from which recovery was slow. Whether to attribute this to the fat or to the carbohydrate is questionable, the observation is like those made by Allen (2) on partially depancreatized dogs fed on fat. 326 Metabolism of Diabetes High Protein, High Fat Diet with Ketogenic Value Nearly equal to That of Period VIII; D:N Ratio 3.65; 4 Days. Period X.—It appeared desirable to compare the effect of the diet of Period VIII with that of a diet which would be exactly comparable to it in ketogenic constituents as calculated by Woodyatt’s (45) formula and yet contain a larger proportion of protein. Accordingly a diet was given from May 30 to June 2 which contained 3.8 gm. of carbohydrate, 104.8 gm. of protein) and 126.3 gm. of fat. Inserting these values into the formula (F. A. = 0.46 P. + 0.9 F.) it will be seen that this diet provides the equiva- lent of 161.8 gm. of oleic acid, which is close enough to the ketogenic value of the diet of Period VIII to satisfy the requirements. The new diet also contained practically the same amount of available glucose (3.8 gm. as carbohydrate and 58 per cent of 104.8 gm. of protein) and differed very little in its amount of available energy. This second high protein, high fat diet was started May 30 and con- tinued for 4 days. The D:N ratio of the urine of May 29 was 1.39. The data from urinalysis reveal a sugar utilization of 27.7 gm. on this date. The ammonia nitrogen excretion was 0.82 gm. and acetone excretion below 3 gm. The D:N ratio of the urine of the first 24 hours of the new high protein diet jumped to 3.39 and May 31, June 1, and June 2, it was that of complete diabetes, 3.68, 3.62 and 3.84, an average of 3.63 for the 4 days. Accelerated ketogenesis and acidosis developed more slowly than the drop in sugar tolerance. The total acetone bodies of the urine of the Ist day were 4.02 gm.; on the 2nd day they were 10.75 gm., on the 3rd day 13.52 gm. The excretion of ammonia nitrogen was only 0.62 gm. on the Ist day rising to 1.08 gm. on the 2nd day, to 1.86 gm. on the 3rd day, and to 2.44 gm. on the 4th day. The total acetone bodies of the blood rose to 0.060 and 0.059 gm. on the last 2 mornings. On the 3rd morning the carbon dioxide-com- bining power of the plasma was 26.8 volumes per cent; on the 4th morning it was 23.8. By the 3rd day the breath smelled strongly of acetone and a slight hyperpnea was noticeable. On the 4th day slight nausea was complained of and it was decided that the period should be discontinued. The respiratory metabolism behaved as it had on the high protein diet of Period V. The basal metabolic rate rose from —13 per cent the morning of May 30, the beginning of the period, to —5 per cent May 31 and —1 per cent June 1. The average rate for the period was —4 per cent. The experimental basal respiratory quotients for the 4 days were 0.66, 0.69, 0.68, and 0.70 and the corresponding calculated non-protein respiratory quotients were 0.66, 0.71, 0.69, and 0.73 (average 0.70). The effect of meals on these quotients is referred to in the discussion of the respiratory quotient. The effect of the protein on sugar tolerance and ketogenesis appeared to be out of proportion to its ketogenic properties as calculated by Woodyatt’s (45) formula. The effect is certainly much more marked than was that of the nearly isocaloric high fat diet of Period VIII which it was designed to equal in ketogenic properties. r Wilder, Boothby, and Beeler ond Low Protein, Relatively High Fat Diet with Recovery of Former Tolerance; 10 Day Period. Period XI.—The morning of June 3 a test breakfast of 25 gm. of pure glucose was given (to be discussed under specific dynamic action of foods), but at noon of this day the patient was placed on a diet similar to the New- burgh (29) type of diet of Period VII. It differed merely in detail, being designed to provide 0.66 gm. of protein for each kilo of body weight, fat in sufficient amount for adequate calories, and carbohydrate equal to, but not exceeding, the patient’s tolerance as previously determined. Account was taken of a theoretically desirable ketogenic antiketogenic ratio between the various constituents of the food. This diet was determined as follows. On the basis of the Du Bois standards, a woman aged 29 years, weighing 31.3 kilos, 158 cm. in height, and with a normal metabo- lic rate would require 1,070 calories to meet her basal requirement. The protein allowance was fixed at 20.7 gm., 0.66 gm. for each kilo of weight. This amount of protein may theoretically yield 12 gm. of glucose (0.58 X 20.7). Assuming the glucose tolerance to be 25 gm., the patient should be able to burn additional carbo- hydrate to the amount of 13 gm. Fat was required to supply the necessary calories, but to avoid ketogenesis a definite ratio must not be exceeded between the fat of the diet and the available glu- cose. On the basis of other studies it seems possible that a diet which is constructed so as to provide not more than 1 gm. of pro- tein for each kilo of body weight and calories not exceeding main- tenance requirements may, without being ketogenic, contain fat in an amount such that one-half a molecule of glucose will be available to accompany the metabolism of every molecule of fatty acid. A diet of 13 gm. of carbohydrate, 20.7 gm. of pro- tein, and 99.5 gm. of fat would give 1,064 calories, an energy value sufficient for the basal energy exchange of the patient. Since a basal metabolic rate of —15 to —20 per cent was to be antici- pated, such a diet would be a maintenance diet without extra allowance for the specific dynamic action of food or for the slight exertion the patient was to be allowed to make. It was hoped that the diet would clear up the glycosuria, control the ketogenesis and acidosis, and maintain the patient in nitrogen and calorie equilibrium. The subsequent course justified our expectations. The diet actually given, on recalculation contained 12.3 gm. of carbohy- THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL, LI, NO. 2 . 328 Metabolism of Diabetes drate, 19.7 gm. of protein, and 94.1 gm. of fat, totalling 1,006 calories. Sugar disappeared from the urine in 8 days and the ferric chloride reaction was negative on the 6th day. The carbon dioxide-combining power of the plasma reached 46.6 volumes per cent on the 9th day and 54.1 volumes per cent on the 11th day. The data from urinalysis revealed a glucose utilization of 29 gm. on each of the last 2 days before dismissal. A satisfactory nitro- gen balance was not obtained; a daily negative balance of 1.5 gm. occurred on each of the last 2 days. It seemed possible, however, that with time this would adjust itself, and in any case, the patient was unable to remain longer in the hospital. She was dismissed, therefore, on this diet, weighing 31.7 kilos. She left Rochester and returned to her home in a near-by town reporting to us weekly by telephone. For 3 months she maintained her weight and the urine remained free from sugar and with a negative ferric chloride reaction. Sugar then reappeared in the urine and the patient becoming discouraged discontinued her diet and died in coma soon after. In view of the low tolerance, the very narrow dietary margin of safety, the type of the disease, and its duration, it is unlikely that disaster could have been prevented for long under any circumstances. The Respiratory Quotient. The average respiratory quotients for the various dietetic periods are as follows: Period. | SAURRST | Respiratory quotient. | Nompressiesesptratory I 8 0.68 0.65 II 1 0.68 0. 64 Ill 3 0.71 0.67 IV 0.70 0.70 V 3 0.69 0.70 VI 5 0.70 0.68 VII 4 0.71 0.69 Vill 7 0.70 0.69 IX 1 0.68 0.64 xs 4 0.68 0.70 XI 4 0.69 0.67 Average..... 44 , 0.693 0.679 a e onh-4 Wilder, Boothby, and Beeler 329 Excepting some of the respiratory quotients obtained in Periods I and III, we feel quite sure of the technical accuracy of the figures. The fact that they vary only slightly above and below the quotient for fat (0.707) indicates, we believe, that at the time of their determination no carbohydrate and only a small amount of protein was being burned. The variation from 0.71 was prob- ably due to the following factors: (a) temporary alteration in the character of the respirations so common in all respiration experiments; (b) variation in the carbon dioxide-combining power of the blood depending on whether the acidosis is increasing or diminishing at the time of the test; and (c) utilization of oxygen and carbon dioxide in the formation of acetone bodies. As pointed out by Lusk (24) it is impossible to make correct allowances for these factors which influence the respiratory quotient and, there- fore, we have not attempted such calculation. In consequence of the possible magnitude of the first of these factors we believe that the occasional respiratory quotients below 0.71 cannot be taken to indicate the formation of sugar from fat, particularly, since such deviation is neither marked nor constant; furthermore, the other data of this study are contrary to any such assumption. In this connection the five experiments performed under postabsorp- tive standard conditions May 27, give the range of variation caused by these factors the most important of which, in this instance was probably the unconscious alteration in the character of the respira- tion: 0.67, 0.69, 0.68, 0.71, and 0.72. The effect of fright is reflected in the high respiratory quotient, 0.74, of May 23 which without question was caused by slightly forced breathing following the dropping of a tray of dishes beside the patient’s door while the test wasin progress. Our patient was most willing and cooperated in every way; we fear that possibly she tried overhard to be quiet and breathe normally and asa result at times during the collection of the expired air slightly underventilated her lungs which resulted in decreased elimination of carbon dioxide and a consequent lower- ing of the respiratory quotient. The non-protein respiratory quotient averages, calculated by Lusk’s (24) method, are only slightly different from those found experimentally. It is doubtful if they are a more exact approxi- mation of the average 24 hour combustion quotients for fat and carbohydrate because they are calculated on the assumption that 330 Metabolism of Diabetes the relation between the protein, fat, and carbohydrate metabo- lism remains constant for the entire day; this assumption is prob- ably not justified. Likewise, Shaffer (85) has shown that no more accurate results are obtained by a method of calculation which attempts to allow separately for both the increase in the respiratory quotient from the glucose quota of the protein and the oppasite ketogenic effect of protein; he advises, therefore, the use of the total respiratory quotient for the analysis of experimental data. Shaffer has constructed a very interesting and valuable table by which the ratio of the ketogenic to antiketogenic mole- cules in the metabolic mixture may be determined from the re- spiratory quotient and also the percentage of calories obtained from glucose and fatty acid; the latter is an elaboration of Lusk’s (26, a) modification of the table of Zuntz and Schumburg. This method, however, is not strictly applicable to respiratory quotients obtained as ours, for short basal morning respiratory periods with the patient taking food subsequent to the test, and should prob- ably be reserved for respiration experiments conducted in a respiration calorimeter for periods of 24 hours when the combined influence of rest, work, and food can be obtained in the day’s average and where the effect of temporary alterations in the character of the respiration do not play a part in the elimination of carbon dioxide. z In spite of these theoretic objections to the submission of the data on Bessie B. to an analysis by Shaffer’s calculations, 1t must be admitted that the results are noticeably consistent therewith except in regard to the intensity of the acidosis that would be expected if it depended primarily on the magnitude of the keto- genic antiketogenic ratio as calculated by his method from the respiratory quotients. The Specific Dynamic Action of Food and the Behavior of the Respiratory Quotient Following the Ingestion of Food. Nine experiments were made in order to study the specific dynamic action of food and its effect on the respiratory quotient. The results are listed in Table II. They can be conveniently divided into three series corresponding to diets of Periods VIII, X, and XI. Each experiment consisted of a series of five determina- tions of the metabolic rate and respiratory quotient. The first "MOIyVAIOSqo Jo Avp oY} IOJ OIA Ni: 9Y} OF Al[PSO[9 o10UuL puodsas109 [JIM [BOUL 480} oY} JO WOIJSoSuL oY} 194jv SUOT}vAdosqo YuONbosqns 94} ‘IOAOMOY /O17VL N? Cd pu yorp s,Anp snotaotd oy} JO FYSH OY} UI poroprsuoo oq ysniu ‘oTOJoLoy} ‘pus Avp yey} IO} popso090I1 pooy oy} Aq poouonpyur jou o10M 84909 jeseq oy} YVY} poroquioUer oq YSNUL 4! OJVI O1TOGeIOUL OY} UO POO JO oOUONYUI oY} oULULToJOp OF So[qey oy} SaIsn uy y cal oOo me - 19'0 19°0 99°0 v9°0 cL'0 69°0 02°0 690 69°0 ‘guoryonb 1oyeardsayy Ps a a 96+ Zia Gia ¢ + 6 + let OI+ | "°°" ‘aqea O1foqeqgoyyf 02°0 19°0 e9°0 cg" 120 19°0 02°0 02°0 ILO ‘quoryonb A10;yeatdsoy | sinoy z ine 9— Het | OG a G+] ptt} b+] OT+ | °° + 09es orfoquzayy! . uM = 19°0 02°0 g9°0 €9°0 89°0 89°0 02°0 01:0 (02°0 “guoryonb ArToyeardsoy , al | ag e+ a 9I+ = et aa | as 8+ [coger orfoqeyo{ ” rm 89°0 69°0 690 19°0 69°0 99°0 02°0 69°0 69°0 ‘quorjonb A10ywardsoy | inoy § = Z 0 i+] 8+ ST— 9+) 24] P+] PIF [777 oper oroqugeyy/ ; 9} Ba — OTJOGeJOUI JO SUOTJRALASGO 104] ate) a = I'? (osoonjs) ¢z| e'T Gar ‘poojou | Z'FT CFI CFI GPL rorsesesssseauB ‘oyerpAqoqie fo) PTE l'2P lov ‘yuourtsed G'os G'Os C'0S GOS . iis) =) ere) >: si gt oa, ORY 7H 4) ‘4B iq faa) e"9 6 FE 6° Fe -x9 [oIjUOD | Q's 9'8 8° gree | ates Sao te 15> s08 “eI a [4sBpyVolg JSa] o a €1°0 0L°0 89°0 69°0 19°0 69°0 0L°0 cL'0 p10 nereesssssess-quargonb £107811ds9y = ) hes a ame ire |) ta eI= id Soe A ae OI— [°° qua wad ‘ayed o1foqeqzoul jeseg — ce I PS € oo & 88° T Set ty PS 0 = “*(uorywadosqo Jo kvp) 4, N?d = 8° zoe = gore cI Ea > pew — 921 |'*'**'(A8p Surposard) o1je1 N Pq TT oune g ounr gounr | young 1g Av 9g ABW | Sc AUIN | PAC | £2 AUN 1261 pottv<) “(X poHeg) IT sorog “(IITA Polo) J sonog ITT 8ef10g (gecoeeV) “a 1ssog] 332 Metabolism of Diabetes determination immediately preceded a test meal; the other four followed 4, 1, 2, and 3 hours after the meal. The first series of experiments, five in number, the last of which was a control without food, were made May 23, 24, 25, 26, and 27, during Period VIII when the diet of the day preceding each observation consisted of 26.5 gm. of protein, 151.4 gm. of fat, and 42.6 gm. of carbohydrate. During this time the D:N ratio varied as follows: 1.26, 0, 0.54, 1.13, and 1.35; and, May 23, when glucose utilization was greatest, the calculated combustion of glucose from all sources did not exceed 20 gm. Therefore, these experiments, with the exception of those on the 2nd day, were made on an organism which presumably was not utilizing carbo- hydrate as such and was metabolizing only a part of the sugar derived from protein. The test meal given, protein 8.8 gm., fat 50.5 gm., carbohydrate 14.2 gm., was one-third of the total daily food allowance. In view of the pure glucose experiment of June 3, " it is unlikely that the carbohydrate quota of this meal played any material part in the reaction produced; unfortunately no determinations on the effect of protein or fat alone were obtained and it is, therefore, impossible to assign exactly the part each of these played in the reaction. A rough estimate of the relative rdéle of protein and fat can, however, be gathered from the experiments of the second series carried out in Period X in which the protein in the test meal was increased to 34.9 gm. and the fat decreased to 42.1 gm. The second series of experiments was conducted in Period X, June 1, 2,and 3. The diet of the day preceding each observation consisted of 104.8 gm. of protein, 126.3 gm. of fat, and 3.8 gm. of carbohydrate. From the D:N ratios of 3.68, 3.62, and 3.84 it is assumed that the patient was completely diabetic and burning neither the carbohydrate of the food nor the sugar derived from the protein. The test meal in two of these experiments was pro- tein 34.9 gm., fat 42.1 gm., and carbohydrate 1.3 gm. In the third experiment 25 gm. of glucose were given alone. The third series consisted of one experiment. This was made June 11 in Period XI after the patient had been several days on a diet of 19.7 gm. of protein, 94.1 gm. of fat, and 12.3 gm. of carbo- hydrate and after the urine had become sugar-free and the acquired sugar tolerance was approximately 30 gm. It is assumed there- — Wilder, Boothby, and Beeler 333 fore, that not only the protein quota of sugar but also all the ingested carbohydrate was being burned. The test meal con- sisted of 6.5 gm. of protein, 31.4 gm. of fat, and 4.1 gm. of carbo- hydrate, which is not far from two-thirds of the test meal of the first series. The variation of successive tests during the forenoon without food is shown in the control experiment of the first series, May 27. The basal metabolic rates observed were —11, —15, —10, —8, and —12 per cent, an average of —11.2 per cent with an average variation of 1.8 and an extreme variation of 3.8 from the mean. The respiratory quotients were 0.67, 0.69, 0.68, 0.71, and 0.72, an average of 0.694 with an average variation of 0.018 and an extreme variation from the mean of 0.025. It is not unlikely that there was a slight underventilation of the lungs in the first three experiments witha corresponding compensation during the last two. As may be seen in Table II there is a slight difference in the metabolic rate following the ingestion of food in each of the four experiments of the first series; however, it seems best to discuss the average results for these experiments. The basal rate before the test meal averaged —12 per cent; 4 hour after the meal +8 per cent; 1 hour after, +5 per cent; 2 hours after, +10 per cent; and 3 hours after, +6 per cent. Therefore, in the first series of experiments, as a result of the specific dynamic action of 8.8 gm. of protein, 50.5 gm. of fat, and 14.2 gm. of carbohydrate, the metabolic rate was raised above the basal level 20, 17, 22, and 18 points 3, 1, 2, and 3 hours, respectively, after ingestion; the rise was prompt and remained at essentially the same increased level for at least 3 hours. The respiratory quotients of the postprandial observations showed no elevation whatsoever from those of the preprandial, completely failing to mirror the effect of the combustion of 14.2 gm. of carbohydrate. Indeed in the second and fourth experiments of the series the respiratory quotients are actually depressed im- mediately after the meal, rising again to the preprandial level only after the 3rd hour. The basal respiratory quotient (0.74) of the first experiment may be discarded for the reason that the patient was startled during the test by the dropping of a tray. In the first two experiments of the second series the normal level of the basal metabolic rate averages —3 percent. Thetest meal con- 334 Metabolism of Diabetes sisted of 34.9 ¢ ». of protein, 42.1 gm. of fat, and 1.3 gm. of carbohy- drate. 4 hour after its ingestion the metabolic rate increased to an average of +11, after 1 hour to +11, after 2 hours to + 22, and after 3 hours to +20 percent. The specific dynamic action of the food, therefore, increased the metabolic level 14, 14, 25, and 23 points above the basal level 4, 1, 2, and 3 hours, respectively, after the food was taken. This increase was superimposed upon a base line which was 9 points higher than that of the first series of experiments. Comparison with the first series shows that the effect was less rapid but slightly higher with an apparent tendency to be more prolonged. It is possible that this difference in action. may be due to the larger proportion of protein in the test meal. The behavior of the respiratory quotients observed in the second series was surprising. No -postprandial elevation was anticipated because the D:N ratio justified the assumption that glucose combustion during the period was minimal, but no such’ depression of the quotient as occurred was expected. In both experiments this is unmistakable; there was a drop from prepran- dial levels of 0.69 and 0.68 to 0.61 and 0.62, 4 hour after the meal, remaining as low as 0.64 and 0.66 after 3 hours. In the light of the marked reaction observed in these two experiments the slight postprandial depression of the quotients in two of the experiments of the first series assumes significance as does that. observed in the single experiment of the third series. The third experiment of the second series-in which the test meal consisted of 25 gm. of glucose was conducted primarily as a control of the completeness of the loss of sugar tolerance. The D:N ratios of the previous 3 days had been close enough to the theoretic ratio of 3.65 to justify the assumption that the patient was a total diabetic. The effect of the sugar on the rate of metabolism was so slight that interpretation is difficult. An elevation of 5 and 8 points was noticeable 4 and 1 hour after ingestion, but this was not much greater than the fluctuations of the rates recorded May 27 when no food was given. The respiratory quotients remained the same or were slightly decreased after the ingestion of glucose and are, therefore, evidence against its immediate combustion. The D:N ratio for this day, 1.32, suggests, it is true, that some glucose was burned. This ratio, however, reflects the metabolism of the entire 24 hours and —. Oe Wilder, Boothby, and Beeler 335 a more probable explanation is that resumption of glucose oxida- tion occurred at some time later in the day as the result of the beneficial effect of the lower protein intake. It may also indicate the retention of some of the ingested sugar to resupply depleted glycogen stores. The single experiment of the third series conducted with a test meal of 6.5 gm. of protein, 31.4 gm. of fat, and 4.1 gm. of carbo- hydrate revealed the same phenomena observed after the larger test meals of the first and second series. The metabolism rose 14, 17, 11, and 9 points $, 1, 2, and 3 hours after ingestion of food, the curve being less high and less prolonged than those of the two former series. This smaller reaction is probably explained by the smaller size of the test meal. The respiratory quotients obtained in this experiment exhibit the same behavior noted in two of the experiments of the first series and unmistakably in the two experiments of the second series, namely, a postprandial depression. This is slight and would be insignificant were it not for the previous observation. We will not attempt, at this time, the interpretation of these postprandial depressions of respiratory quotients although we believe they are significant because they occurred in five of six good experiments, or six times in seven, if the doubtful basal quotient in the first series is included. It may be noted, however, that the depressions of the quotients are so much greater after high protein meals as to suggest that they may represent a protein effect. ‘ The Basal Metabolic Rate. During Period I of the observations the basal metabolic rate averaged for eight determinations —12 per cent and the daily variations were slight. ‘The protein in the diet was 48.8 gm. and the patient’s glucose tolerance was limited as shown by an average D:N ratio of 1.52. In Period III the basal metabolic rate was 8 points lower than in Period I. The reason for this lower rate is not quite clear. Period III was preceded by a 2 day fast with resulting decrease in the amount of glucose eliminated in the urine, a lower D:N ratio, and increased sugar tolerance; the lower basal metabolic rate seems primarily dependent on the general benefit obtained 336 Metabolism of Diabetes by the fast, as indicated by the rate of —27 on April 21 (end cf Period II); however, the fact that the main alteration in the diet of Period III was a decrease in the carbohydrate intake, must not be overlooked as a factor. Although there is legitimate doubt as to the cause of the fall in the basal metabolic rate in Period III there is no question with regard to the interpretation of the marked and sudden rise from an average of —20 per cent of Period III to the average of —14 per cent found in Period IV. This rise definitely followed the increase of approximately 50 gm. of protein, making a total intake of 94.2 gm., 3.0 gm. for each kilo of body weight. Acidosis de- veloped subsequently to the rise in the basal metabolic rate and did not materially exceed that of Period I so that acidosis did not seem to be the cause. The average for Period IV of —14 per cent was materially lowered by the observation made April 30 which is so far below the average for the period that the accuracy of the determination may be questioned. The increase in the basal metabolic rate in Period V over that of Period IV is slight especially if the determination of April 30 of —19 per cent is excluded from the average of the latter. There- fore, the increase of approximately 40 gm. of fat, the protein and carbohydrate intake remaining practically the same, increased the basal metabolic level only 2 (or 4) points from —12 (or —14) to —10 per cent. The D:N ratio rose to an average of 3.73 in this period. Acidosis was simultaneously increased. From an exami- nation of the data for this and the preceding period the conclusion seems warranted that the increase in the basal metabolic rate was not caused at least in major part, either by the fat in the diet or by the acidosis that developed later, but was mainly, if not entirely, caused by a cumulative specific dynamic action of protein (25, 26, a). The low protein and low carbohydrate with relatively high fat, Newburgh type of diet of Period VI, allowed the basal metabolic rate to fall quite rapidly to —20 per cent (May 13) from the previous high level of —10 per cent. This diet contained approxi- mately 10 gm. of protein, 83 gm. of fat, and 16 gm. of carbohydrate which totalled only 881 calories. The food ingested exceeded the basal requirement (856 calories), but since it did not allow a sufficient excess to balance the 10 per cent needed for the specific dynamic action of the food and 10 per cent for the movements in a le . Vilder, Boothby, and Beeler 337 bed, it was below the maintenance requirement at this level. Therefore, on this diet the patient had in all probability to draw on her own reserves for approximately 147 calories, or about 16 gm. of endogenous fat daily; however, as there was a negative nitrogen balance of 2 to 3 gm. half of this deficit was probably made up by the combustion of endogenous protein. The acidosis decreased, but apparently more slowly than the basal metabolic rate and does not, therefore, appear to have been a factor in producing the drop in metabolism observed. In the production of the decreased basal metabolic rate we have, then, two factors, the elimination of the cumulative specific dynamic action of a high protein diet, and the reduction of the calorie intake to a subnormal level of maintenance. The second of these factors was eliminated in Period VII when the diet was increased to 30.7 gm. of protein, 109.1 gm. of fat, and 14.7 gm. of carbohydrate totalling 1,200 calories which was within 84 calories of the patient’s normal daily requirement (normal basal requirement of 1,070 calories plus 10 per cent for specific dynamic action and 10 per cent for movements) and was 172 calories in excess of her needs calculated on a basal metabolic rate of —20 per cent. The average basal metabolic rate remained unaltered at —20 per cent despite this increase in the diet so that the sudden reduction in the basal metabolic rate in Period VI seems to be owing primarily to the elimination of the protein which had temporarily raised it above the low level which had become cus- tomary for this patient. The diet of Period VIII was constructed to make it practically isocaloric with the high protein, high fat diet of Period V and totalled 1,692 calories. This diet was 408 calories in excess of the normal daily needs of this patient at rest which was 1,284 calories (1,070 plus 10 per cent for specific dynamic action and 10 per cent for muscular movement). As a result of this new diet the basal metabolic rate rose 6 points from an average of —20 to —14 per cent. It is difficult.to determine the specific cause of this rise as there was an increase of 42.3 gm. in the amount of fat and also an increase of 27.9 gm. in the amount of carbohydrate in the diet. The carbohydrate as such, and that from protein, therefore, was 58.0 gm. [42.6 + (26.5 X 0.58) = 58.0] which was about twice -the patient’s best sugar tolerance. 338 Metabolism of Diabetes These two factors caused a moderate break in the sugar tolerance. The rise in the basal metabolic rate in this instance was obviously not owing to the protein, and might or might not have occurred from an increase in the fat if the carbohydrate intake had re- mained constant. The increase in acidosis was slight and did not precede the rise in the basal metabolic rate sufficiently for us to attribute the cause to the acidosis. While it is impossible, there- fore, to draw conclusions with regard to the exact reason for the effect of this diet, it seems probable that its high calorie value was in part responsible. This is in agreement with the arguments of Allen against the Newburgh diets and these experiments support the validity of his objections when such diets exceed a certain optimal calorie value. The fast of 2 days in Period IX caused a rapid improvement in the acidosis and a decrease in sugar elimination. The basal metabolic rate, however, did not fall as it did after the previous fast (Period II) nor was the improvement in the sugar tolerance sufficient materially to decrease the D:N ratio. Some factor or accumulation of several factors in the diet of Period VIII appar- ently produced not only an immediate elevating effect in the rate, but one from which the recovery was slow. The diet of Period X was essentially the same as that of Period V, namely, high protein and high fat. The previous low protein, high fat diet of Period VIII may have produced a more susceptible status, as we have pointed out, but in any case the reaction to the diet of Period X was so marked and so rapid contrasted to that of Period VIII that no doubt can exist as to the harmfulness of excessive protein. The D:N ratio of total diabetes resulted al- most at once, and paralleling this fall in sugar tolerance the basal metabolic level rose to an average of —4 per cent. This is 9 points above the level determined May 30, and from 12 to 16 points higher than that of the periods in which the patient seemed to be at the metabolic level best suited for her condition (Periods III, VI, VII, and XI). In Period XI there was a big reduction in protein from 105 to 20 gm. and in fat from 126 to 94 gm. with an increase of carbo- hydrate to 12.3 gm. Therefore, the good effect of this diet must be credited in great part to the reduction in protein; however, if contrasted with the results produced by the diet of Period VIII a i is ee es = hn Wilder, Boothby, and Beeler 339 it is further evident that the lower fat and a total carbohydrate intake slightly less than her tolerance [12.3 + (19.7 * 0.58) = 23.7] also was beneficial. Throughout these experiments the basal aieiabobe rates varied inversely with the general condition of the patient. During Periods III, VI, and VII when the rates were lowest the glucose utilization was best and acidosis was either controlled or decreas- ing. When the basal metabolic rates rose as in Periods IV, V, and X sugar tolerance diminished and acidosis increased. This parallelism between the degree of depression of the basal meta- bolic level and the rise of the D:N ratio is shown graphically in Chart 1. Respiratory Metabolism of Thirty-One Patients with Diabetes. In 1910 and 1912, Benedict and Joslin (6, 7, 8) published exten- sive studies of the basal metabolism in diabetes from which they concluded that the basal metabolic rate was increased in severe diabetes. Allen and Du Bois (3), however, have shown that the normals which were used by Benedict and Joslin averaged 8.6 per cent below the Du Bois normal standards and that the nineteen patients studied by Benedict and Joslin were only 2 per cent above the Du Bois normal, although 11 per cent above the normal controls selected. Allen and Du Bois also pointed out that of the twenty-six carefully investigated cases of diabetes reported up to that time, the basal metabolic rates were normal in thirteen, increased in nine, and decreased in four. We have recenily investigated the basal metabolism of thirty- one patients with diabetes. A brief summary of the data obtained is given in Table III. The basal rates were found to be within 10 per cent of the Du Bois normal standards in 10 out of 12 un- complicated cases without emaciation; the other two had de- creased rates. Body weights of 20 per cent or more below the standard weight in medico-actuarial mortality tables (28) are interpreted as indicating prolonged subnormal diets; the patients, with uncomplicated diabetes but with these low body weights, have, with only one exception, shown depressions of their basal meta- bolic rates of from —11 to —32 per cent, the degree of depression being possibly roughly proportional to the degree of loss of weight below the standard. Two emaciated patients, Cases A350440 and Metabolism of Diabetes 340 206 OT IP |206 OT 0g ZF (206 OT PP |206 OT Ph 206 OT ze 6z¢ ‘T | 09 GPa ‘T | 09 16g ‘T | 09 98¢‘T | 09 cs £°¢ (206 FI ee ra) 0S ce “UB ‘mb \yuaa sad / ‘Soll |, RIB. ‘oun “18D ees Meee ay} jo N ‘sAep ¢ snotAoid ‘yoIp oF BIDAY 02h 0°29 0'vS _ «pre -puryg “WYSIOM “poyeNn v LST ¢°69T 0'S89T 9° 2ST “Mba 94310 0g— 66— L6— 0Z— ce ig 41— Ge Lu | oa Z1— 6G /4| eh— Gis cE — 86— quaa sad *O7B1 oT[O -qeyjour |sor1oyeo [este 6°68 |S os T‘Or | &T » VIP | at “ydag Gp | FT 9°9F | FI» eo ee. (El = Gs Ogn | . it Fa ward SS a OS i i ea hires Me ? THE INTERACTIONS OF OXYGEN, ACID, AND CO, IN BLOOD. By A. V. HILL. (From the Physiological Laboratory, Manchester University, Manchester, England.) (Received for publication, December 31, 1921.) In a recent paper (1) I have shown that the effect of acid or CO: upon the oxygen absorption curve of blood can be deduced quantitatively from the hypothesis that hemoglobin is a weak acid, dissociated electrolytically to a very slight degree, according to the equation, B (Sb), ==" A= (b)’,, while oxyhemoglobin is a stronger acid, dissociated to a greater degree, according to the equation, H(HbO2), = H’ + (Hb0:)’, This hypothesis leads to a relation between the K of ‘“Hill’s equation” and the ratio of the hydrogen ion concentration (Cy) to the basic ion concentration (Cg) inside the corpuscle Cy. where k and a are constants. It can be shown that a oa is much B greater than unity, so that the equation can be written, without sensible error, Taking logarithms, log K = pH + constant, so that the relation between log K and pH is a straight line inclined at 45° to the axes. The linearity of this relation has been established experi- mentally by Barcroft (2) and by Donegan and Parsons (38), 359 THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI, NO. 2 360 Oxygen, Acid, and CO, in Blood and it is noticeable that their curves are all inclined at ap- proximately 45° to the axes; small divergences, however, do occur, and these are probably to be explained as follows. In the calculation of K for comparison with pH it has been assumed that n = 2.5. In any particular case this assumption may not be exact. In determining K the practice adopted by Barcroft has been to make an observation of the percentage saturation, at any oxygen pressure adjusted to give about half saturation. y 100—y Assuming y = 50 the equation =Kz" becomes 1 = 1 Mn n (say n’) be assumed, an incorrect value of K (say K’) will be calculated from it, the relation between the assumed and the true values being, Kz", from which log x = log K. Thusif anincorrect value of 1 1 - log K = — log K’ . n n Thus the true value of K will be given by the equation n nr log K.= — loge kK’ = ——lagike BD 7 oe ae In the cases, therefore, in which the relation between log K and PH is not inclined exactly at 45° to the axes it is obvious that the inclination can be adjusted to 45° by a suitable small: alteration in the assumed value of n. There is further justification for the hypothesis. Many investigations have recently been made of the relation between the pH of blood and its CO: tension; if this relation be expressed in terms of Cy, instead of pH, it will be | found that pract‘cally all the published curves become straight lines; the relation between Cg and pCO, therefore is linear, or 1 : : nearly linear. Thus K being a linear function of Cy must also be a linear function of pCO.. Thus theory confirms the empirical relations established independently by Henderson (4) and Adair (5). It should be noticed that the assumption as to the acid nature of hemoglobin differs from that commonly made, which is! that 1 See Henderson (4), p. 408. A. V. Hill 361 the unit? of hemoglobin containing 1 atom of iron is an acid HHb, the degree of electrolytic dissociation of which changes when it combines with 6xygen. The assumption made here is that the molecule (Hb),, occurring in the formula which leads to Hill’s equation of the dissociation curve, v7z., (Hb), + nO2 = (HbO2)» behaves as a mono-basic acid, dissociating according to the equation, H (Hb), — H’ + (Hb)’, In this respect the dissociation of H (Hb), is analogous to that of HsCOs, the first dissociable hydrogen ion alone of the complex molecule is involved in the dissociation, at any rate over the range of Cy possible in blood. Thus we do not necessarily assume that hemoglobin is not (H Hb),, but that over the range con- sidered only the first hydrogen of the poly-basic acid is dissociable. It is possible that we should regard the acid hemoglobin rather as the acid sodium (or potassium) salt H Na,_; Hb,, but until we know more of the nature of the agtions of bases on the hemo- globin molecule it would be idle to speculate further. In any case the assumptions (1) of the mono-basic properties of (Hb),, and (2) of the change wrought in the degree of electrolytic dis- sociation by the combination with Os, are sufficient to account quantitatively for the effects of acids and CO: on the dissocia- tion curves of blood. We c-me now to the effects of oxygenation of blood upon its absorption curve for CO... The CO: absorption curves of oxygenated and reduced blood are known to be different (Chris- tiansen, Douglas, and Haldane, 6) and this fact has naturally been credited to the greater electrolytic dissociation of oxyhe- moglobin. Henderson (4) has attempted quantitatively to con- nect the effect of O2 on the CO: absorption curve, with the effect of CO, (and acid) on the O2 absorption curve, as he admits, how- ever, on'y with partial success. There can be no doubt that, thermodynamically, the two sets of facts must be related, and should be deducible the one from the other. The difficulty in 2 It is not justifiable to call this unit a molecule, unless it be shown that the unit (Hb),, which appears to act as a molecule, is a fiction. 362 Oxygen, Acid, and CO, in Blood my opinion has arisen from the assumption that hemoglobin in blood is ionized according to the formula H Hb = H’ + Hb’, and not H (Hb), = H’ + (Hb). With the lattér assumption the quantitative connection between the two sets of phenomena is at once clear. At any considerable CO. pressure reduced hemo- globin in blood, being an exceedingly weak acid, is present as undissociated H(Hb),, while oxyhemoglobin, being a much stronger acid, is present as the salt B (Hb O.),,, holding onto the base B very firmly and denying it to the CO2. Thus in combin- ing with n molecules of O2 the hemoglobin has seized on 1 mole- cule of base, and has deprived the blood, therefore, of the ability to combine with 1 molecule of CO. As Henderson very rightly points out? it is necessary to consider the “base transferred from carbonic acid to haemoglobin when fully reduced blood is fully saturated with oxygen zsohydrically,” 7.e. at a constant hydrogen ion concentration, and he gives a figure from which it is seen that, in the blood of J. 8. H., as the Cy increases the extra COz taken up by reduced blood reaches a maximum of about 8.8 volumes per cent. Parsons (7) has investigated this subject by direct experi- ment, and from Fig. 4 of his paper,‘ it is seen that as the Cy in- creases the extra CQ, taken up by his reduced blood, in excess of that taken up by his oxygenated blood at the same Cy, reaches a steady value of about 9 volumes per cent, thus entirely confirming Henderson’s calculations with the blood of J. S. H. Assuming that the blood of J. 8. H. had the normal Oy» capacity.of 18.5 vol- umes per cent, we see then that the taking up of 18.5 volumes of O, has deprived the blood of base sufficient to take up 8.8 volumes 1S:h.. of COs. Thus n should be equal to ——; 2.e., to 2.1. Recent 8.8 experiments by Brown and myself in this laboratory make it probable that the usually accepted value of n, viz. 2.5, is rather too high; that at any rate in the blood of sheep, oxen, or pigs, a value of n between 2.2 and 2.3 more exactly fits the facts. The agreement between this revised value and that calculated above from the Oz: COs ratio is very good, and appears a striking con- firmation of the general validity of the hypothesis used to explain ’ Henderson (4), p. 407. 4 Parsons (7), p. 452. AL Wer El 363 the actions of acid and CO: upon the oxygen absorption curves of blood. We see, therefore, that the hypothesis leading to the equation a K2" YO + Ka") the effects of acid and CO. upon the O, absorption curve of blood, and in correlating these with the effects of oxygen upon the ab- sorption of CO:. It is clearly more than a mere mathematical chance that the equation is of such general application. It was deduced originally (8) from the hypothesis that salts affected the degree of “aggregation” of the colloid hemoglobin molecule. A precise theory, however, is not necessary; any factor which causes the hemoglobin in blood to exert an osmotic press- bears the strain imposed upon it in explaining 1 Pi ee ure equal to — of the pressure of the oxygen with which it aie combines must necessarily lead to an equation of this type. Expermients, hitherto unpublished, made by Miss Atkinson in this laboratory, have shown that the osmotic pressure of hemo- hes 3 : 1 : globin in blood is about 59 of the pressure of the O2. combined with it. Thus far, therefore, the equation is necessary from the standpoint of thermodynamics; it remains to determine the ac- tual chemical mechanism by which the osmotic pressure is lowered by the presence of salts. This mechanism is not yet clear and one can scarcely assent to an ‘“‘aggregation” theory in its crude form; it is difficult to believe that a mere physical clumping to- gether of the simple hemoglobin molecules, caused by a surface effect of salt on colloid, could have the extensive and exact re- sults observed. I am inclined to the view that the electrolytic dissociation discussed in this paper is connected with these “‘os- motic” changes. If the real hemoglobin molecule be a complex weak poly-basic acid (like boric or phosphoric acid) its electrolytic dissociation, and the size of its molecule, may be largely influenced (according to the laws of mass action) by the presence even of small concentrations of basie or of hydrogen ions. My friend Mr. W. E. L. Brown prepared the attached note on the analo- gous case of boric acid, in which he shows that hydrolysis may largely affect the molecular weight of the boron-containing com- 364 Oxygen, Acid, and CO, in Blood plex, may in fact cause a chemical ‘‘aggregation” or “ polymeriza- tion’’ of its molecules. In pursuing the subject, therefore, of the electrolytic dissociations of hemoglobin we are probably taking the shortest road towards a knowledge of the causes of those varia- tions of osmotic pressure, which can be shown experimentally to exist, and whose existence necessarily involves the observed changes in the oxygen absorption curves. The Effects of Hydrolysis on the Osmotic Pressure of Salts of Weak Acids. By W. E. L. Brown. If hemoglobin exists in the blood as the sodium or potassium salt of a weak acid it is reasonable to suppose that it must under- go a certain amount of hydrolysis, as do the salts of inorganic weak acids with strong bases; e.g., borates, cyanides, carbonates, etc. ; It is stated by Mellor (9) that in a concentrated solution borax, Na2B,07, is hydrolyzed as follows: NazB.07 o 3H,0 = 2NaBO;z 2H;BO; It is also stated that sodium metaborate, NaBOkz, is converted into borax by CO, as follows: 4NaBO, — CO2 — Na.CO; — Na2B.0;7 If the hydrolysis of the first equation proceeds completely from left to right the total osmotic pressure of the boron-containing compounds or ions will be zncreased in the ratio 4:1, while if the second reaction proceeds completely from left to right the total osmotic pressure of its boron-containing compounds or ions will be reduced in the ratio of 1:4. If either reaction goes on only partially the osmotic pressure of the boron-containing compounds will be increased, or reduced, in some intermediate ratio. Thus a fairly extensive change in the osmotic pressure per atom of boron can be caused, either by hydrolysis or by combination with COs, and these reactions provide an example of a chemical mechanism possibly analogous to the case of hemoglo- bin. It is conceivable that a similar change of osmotic pressure might be brought about in hemoglobin by varying concentra- A. V. Hill 365 tions of Na’ or H’, and that the mechanism of such a change might be explained by the assumption that hemoglobin is a salt of a weak acid, hydrolyzed in a manner similar to borax. SUMMARY. The S-shape of the O, absorption curve of blood can be ex- plained by the hypothesis that the osmotic pressure of hemoglo- 1 bin in blood is only a part of the pressure of the O2 with which it combines, where n has a value usually of rather less than 2.5. This reduction of the osmotic pressure of hemoglobin is probably analogous to similar effects occurring with weak acids, e.g. boric acid, in the presence of varying concentrations of hydrogen and basic ions, and described under the general heading of ‘‘hydroly- sis.” The effects of acid and CO, upon the Os: absorption curve of blood can be deduced from the hypothesis that the complex hemoglobin molecules (Hb),, and (HbO:), are like H2COs, elec- trolytically dissociated in the forms H (Hb), = H’ + (Hb)’n H (HbO2), = H’ + (HbO2)’n the degree of dissociation of the latter being far greater than of the former. The same hypothesis accounts quantitatively for the extra CO, taken up by reduced in excess of that by oxygenated blood. At high CO, pressures, and at a given Cy, the taking up of a given V volume V of QO: results in a diminished CO2 absorption equal to 7, BIBLIOGRAPHY. 1. Hill, A. V., Biochem. J., 1921, xv, 577. . Barcroft, J., Camis, M., Mathison, C. G., Roberts, F., and Ryffel, J. H., Phil. Tr. Roy. Soc. London, Series B., 1914-15, eevi, 71. , . Donegan, J. F., and Parsons, T. R., J. Physiol., 1918-19, lii, 325. . Henderson, L. J., J. Biol. Chem., 1920, xli, 403. . Adair, G.S., J. Physiol., 1921, lv, p. xvi. . Christiansen, J., Douglas, C. G., and Haldane, J. S., J. Physiol., 1914, xlviii, 244. . Parsons, T. R., J. Physiol., 1917, li, 452. . Hill, A. V., J. Physiol., 1910, xl, p. iv. - Mellor, J. W., Modern inorganic chemistry, London, 1918, 627, 628. bo > oe co oon or =~ i cy) bra eat ravi leah ray ents h lf ai sam oaat i bok eek > ‘ 4 \ , cae a 5s + aera 2 “ [ J } bien + hse theta 's “3 Ts Fe ity! a meek ‘ ‘ 7, ‘ # : } ii ; M> o 1 - Be * “30 j - i - Ae , a 4 A MICRO METHOD FOR THE ESTIMATION OF AM- MONIA IN BLOOD AND IN ORGANIC FLUIDS. By K. L. GAD-ANDRESEN. (From the Laboratory of Zoophysiology, University of Copenhagen, Copenhagen, Denmark.) ‘ (Received for publication, November 2, 1921.) In an earlier publication! dealing with a micro method for the estimation of urea in blood, I also described a micro method for the estimation of ammonia in blood and in organic secretions. For the determination of the ammonia two analyses are made; in the first the total urea plus ammonia is determined; and in the second the urea alone after the ammonia has been removed by evaporating the blood in vacuum. The difference between the two analyses gives the ammonia content, ~ In the urea method itself the blood proteins are precipitated with 0.01 N acetic acid to which sodium acetate is added as a buffer. The precipitate is filtered off and the urea in the filtrate is estimated by means of its decomposition with hypobromite. The nitrogen produced is measured in Krogh’s micro respirometer and calculated with the aid of Krogh’s formula in terms of ¢.mm. When this figure is multiplied by 1.256 a correction for the weight of nitrogen is obtained in mg. and can, if desired, be found in terms of urea. For the purpose of this determination the micro respirometer is provided with ellipsoid bottles of 10 to 15 ce. capacity, fitted with ground glass stoppers on which are sealed small containers for the hypobromite. The determinations re- guire from 0.10 to 0.15 ce. of blood and the error is +0.5 mg. per 100 ce. Since the value for the ammonia by this method is obtained as the difference between two such estimations, both subject to error, the possible error on the ammonia estimation itself is relatively high: 1 Gad-Andresen, K. L., Biochem. Z., 1919, xcix 367 368 Estimation of Ammonia in Blood As it is well known from the work of Henriques and Christian- sen,? the amount of ammonia in blood is very small and is prob- ably constant at about the value 0.25 mg. per 100 cc., a figure of about the same magnitude as the error in the determination. - Estimations by this method, therefore, give positive results only when the ammonia content is abnormally high. The method was worked out at the time because I felt that my urea method would be more accurate if the urea and ammonia were analyzed separately. But when Henriques and Christiansen’s work ap- peared, it seemed that further study in the interest of the accuracy of the urea method was unnecessary; the ammonia correction of 0.25 mg. per 100 ce. could simply be subtracted from the total found in order to get the true urea value. A series of ammonia estimations were, however, made in part on human and in part on ox blood. The highest value found was 0.8 mg. per 100 ec. but as a rule none was found, indicating the amount present to be less than 0.5 mg. per 100 ce.—in good agreement with the results of Henriques and Christiansen with their macro method. If 0.1 to 0.15 cc. of blood be used the method can indicate only, whether the ammonia content is about normal or not. Greater accuracy might be obtained with larger amounts of blood but in its present form the method can only be used with small amounts of blood. Were larger quantities of blood used, correspondingly larger quantities of fluid would be required for coagulating the blood and washing the precipitate, the size of containers must needs be increased, and hence the change in the manometer would be reduced. The requirements for an exact method are evidently that the amounts used for analysis be large and the value for the container as small as possible. The method to be described meets these requirements. The amount of blood used is 1 cc. and the volume of the container is reduced to about 10 cc. The ammonia is estimated directly. After addition of borate (Sérensen*), the blood is evaporated to dryness in the apparatus presently to be described. The ammonia 2 Henriques, V., and Christiansen, E., Biochem. Z., 1917, Ixxviil, 165; Ixxx, 297. 3S6rensen, S. P. L., Emzymstudien. Mitteilungen vom Carlsberg Laboratorium, 1909, viii, pt. 1. Kk. L. Gad-Andresen 369 volatilized is driven by means of an air current over into the bottle of the micro respiration apparatus which contains weak sulfuric acid. It is thereupon estimated in the same way as in the urea method described above, the ammonia being decomposed by hypobromite, and the volume of nitrogen is determined. The evaporating apparatus is very simple. It consists of a glass tube 25 em. X 10 mm. provided with a bulb, as shown in Fig. 1, in which the blood and borate are mixed. The air current which is introduced at the other end through a small bent glass tube and rubber stopper first passes through a wash bottle con- taining concentrated sulfuric acid, which not only frees it from ammonia but also dries it and thus greatly increases the rate at which the blood is evaporated. The other end of the evaporating apparatus is drawn out into a narrow tube and bent as shown, so that it will dip under the surface of the sulfuric acid in the analysis bottle. During evaporation the apparatus is placed in a thermostat at about 25°. A temperature exceeding 30° must be avoided because, as it is well known, urea is then converted into ammonia. A hypobromite solution of constant composition must be used, controlled, and corrected for by a Kjeldahl determination, as Krogh?’ * showed for urea determinations. . Her formula is used for 4 Krogh, M., Z. physiol. Chem., 1913, lxxxiv, 379. ’ Krogh, M., Deutsch. Arch. klin. Med., 1916, exx, 272 370 aereatien of Ammonia in Blood this purpose, the solution containing 1 cc. of bromine to 100 ce. of 2 n sodium hydroxide. The ratio between the nitrogen as determined by a Kjeldahl estimation and as found by the present method gives the factor for. correction. In the present instance the factor was found as follows: An ammonium sulfate solution containing 0.4715 gm. per 100 ce. of water yielded 0.101 gm. of nitrogen. A series of estimations with the micro respirometer on the same solution gave an average value of 0.0919 gm. 0.10 of nitrogen per 100 cc. and the correction was, therefore, 0.0919 ~ 1.09. The estimations with the micro respirometer were made under precisely the same conditions as those for blood. 0.1 cc. of the ammonium sulfate solution was measured into the analysis bottle in which had been placed 0.5 ce. of 0.2 n sulfuric acid and 1.9 ec. of water, making in all 2.5 cc. of fluid of the same acidity as under the conditions of blood ammonia analysis. The procedure was precisely as described in the urea method. The ammonia estimations on blood are done as follows: 1 ce. of blood is accurately measured out and placed in the bulb of the apparatus, into which ‘has been previously put 0.1 ec. of borate (9 ec. of borate + 1 cc. of NaOH). The stopper is inserted and the tube rotated so that blood and borax are thoroughly mixed. The apparatus is then tilted so that the blood flows towards the end of the tube where the air current enters. Care must be taken not to get any of the blood into the narrow connecting tube be- cause when the air current is opened it may then be carried over into the acid and ruin the determination. The blood is distrib- uted over the entire surface of the tube and the latter is then placed in the water bath and connected up with the bottle of the micro respiration apparatus, in which has been placed 0.5 cc. of 0.2~ H.SO;. The air current is cautiously opened and regulated so as to prevent the possibility of the material splashing out of the bottle. Evaporation to dryness is complete in about 30 minutes and the ammonia is estimated in the usual way in the - micro respiration apparatus. In order to facilitate the calculation of the evolved quantity of nitrogen it is to be noted that the same quantity of fluid is to be used in the analysis reservoir each time. This can be made so that the reservoir is tared and then, when the analysis is finished, weighs up with distilled water to a certain weight; for instance, Kk. L. Gad-Andresen 371 2.5 gm. This weighing must be made as correct as possible and must not differ at the utmost more than 0.10 gm. If care is taken that the volume of the analysis reservoir is not altered, the calcula- tion of the analysis result is very simple, as the c.mm. of evolved nitrogen is multiplied by the proportion: 1.256 X 1.09 < 10? AR SB A in which V signifies the quantity of blood used for the deter- mination. Analysis of ammonia both in normal blood and in the blood to which definite amounts of ammonia, as ammonium sulfate, have been added, are given in Table I. TABLE I. Estimations on Blood to Which Ammonia Has Been Added, Calculated as Ammonia Nitrogen per 100 Ce. Double estimation directly on blood. Blood. | Ammonia nitrogen per 100 cc. mg 1 0.41 — 0.44 2 0.51 — 0.46 3 | 0.41 — 0.38 = | 0.39 — 0.39 Blood 4 added 3.43 mg. of ammonia nitrogen per 100 cc. Ammonia nitrogen per 100 cc. Blood. Found. Average. Calculated. mg. mg. mg. 8 3.86 — 3.84 3.85 3.82 Blood 4 added 5.36 mg. of ammonia nitrogen per 100 cc. Ammonia nitrogen per 100 cc. Found. Average. Calculated. mg. : mg. mg. 5.78 — 5.81 5.79 ay A 4 Estimation of Ammonia in Blood The results show an average error of 0.03 mg. per 100 cce., but an error of 0.05 mg. may be taken as the limit. Secretions as well as blood may be analyzed by this method, but it cannot be used for tissues, because, as I® have shown in an earlier paper, there is a rapid conversion of urea into ammonia after death. It may be prevented if the tissue is at once placed in alcohol cooled to— 20°, but the employment of the alcohol will pre-— vent an exact estimation with the micro respirometer. I am obliged to Professor A. Krogh, the chief of the Zoophysio- logical Laboratory of Copenhagen for his kind attention to my work. 6 Gad-Andresen, K. L., J. Biol. Chem., 1919, xxxix, 267. eo. ee A MICRO UREASE METHOD FOR THE ESTIMATION OF UREA IN BLOOD, SECRETIONS, AND TISSUES. By K. L. GAD-ANDKESEN. (From the Laboratory of Zoophysiology of The University of Copenhagen, Copenhagen, Denmark.) (Received for publication, November 2, 1921.) In the previous communication I have described a micro method for the estimation of ammonia in blood and in organic secretions. Utilizing the same principle, a micro urease method has been worked out for the estimation of urea in blood, tissues, and organic secretions. The procedure is exactly the same as that described under the ammonia method, except that in this case the urea is first converted into ammonia by means of urease anhydride, after which borate is added and the blood evaporated to dryness. The same correction is applied to the nitrogen found as in the ammonia method (that is, multiplication by the factor 1.09), and the calculation of results is quite the same. The procedure estimates at the same time the preformed ammonia, but as this is very small and almost constantly 0.25 mg. per 100 cc., this figure may safely be used for subtraction to obtain the true urea value. If desired, however, or where the blood is abnormal, the ammonia may be determined separately and the correction found directly. As previously reported! I follow the procedure of Cullen and Van Slyke? in their modified macro urease method, using a solution of urease anhydride instead of the aqueous extract of the soy bean originally employed by Marshall.2 The employment of urease anhydride is rendered safer if the time required to split the approximate amount of urea is previously determined. Otherwise one runs the risk that other nitrogen-containing substances present may be acted upon and 1 Gad-Andresen, K. L., J. Biol. Chem., 1922, li, 367. 2 Cullen, G. E., and Van Slyke, D. D. ‘Te Biol. Chem., 1914, xix, 211. 3 Marshall, E. K., Jr., J. Biol. Chem., 1913, xiv, 283. 373 374 A Micro Urease Method yield ammonia, as is the case when blood stands 12 hours with an aqueous extract of soy bean. For this standardization it is convenient to use 0.1 ec. of a 0.1 per cent urea solution in water together with 0.2 cc. of a 5 per cent urease anhydride solution which also contains 0.6 per cent primary potassium phosphate (KH.2PO,). In this way an extra measurement is avoided and besides the urease anhydride solution is more stable when it contains primary potassium phosphate. The urease anhydride is placed in the bulb of the apparatus described in the ammonia method, the urea is added to it, and after the apparatus has been connected it is rotated in such a way that the two solutions are mixed. It is then immersed in a water bath at 38-40°C. for about 20 minutes, after which time 0.01 ce. of borate (9 ce. of borate + 1 ce. of sodium hydroxide) is added, and the tube is rotated in such a way that the solutions are well mixed. The procedure is then the same as described under the ammonia method except that the evaporation is carried out with the apparatus immersed in the water bath and is completed in about 15 minutes when the air current is properly regulated. The ammonia is then estimated in the micro respirometer in the usual way. In order to make the calculation of the volume of nitrogen evolved easier, the same amount of fluid should be employed each time in the bottle, by previously taring the latter. The urease which I have used splits the urea up in the course of 20 minutes. The number of c.mm. of nitrogen found is converted into mg. of nitrogen per 100 cc. of solution by multiplying by: 1.256 X 1.09 X 10? 0.10 X 10° In this formula the figure 0.10 represents the amount of urea solution taken. Urea estimations on blood are made as follows: From 0.05 to 0.1 cc. of blood is measured in a micro pipette and at once inserted into the bulb of the apparatus, in which 0.2 ec. of the urease solution has been placed, previously standardized, as already described with respect to the time needed for the complete con- version of the urea. The apparatus is placed in the water bath at 38-40°C., and after the necessary time has elapsed 0.01 cc. of ee + Samy oie K. L. Gad-Andresen 375 borate is added and the evaporation carried out exactly as has been described. Urea estimations in secretions are carried out in exactly the same way. Because of the relatively large urea content, it is necessary to dilute urine for estimation in the ratio 1:50 and take 0.1 cc. for an estimation. It is, of course, necessary to estimate the ammonia content of the urine as well, since this is not constant as in blood and organic secretions, and individual corrections must be made. For this ammonia estimation 0.05 to 0.10 cc. of urine is used instead of 1 cc. as in the case of blood and other secretions. For tissue estimations I use 0.05 to 0.1 gm. of tissue, weighed into a tared weighing bottle, and which is macerated in a small TABLE I. Urea nitrogen per 100 ce. Blood. | Found. Average. | pet Liquid analyzed. | mg. 0.1 per cent urea solution 1 47.2 — 47.0 — 46.7 | 47. ABST |S. 2 15.6 — 15.9 — 16.1| 15.9 Ox blood. 63.1 — 62.8 62.9 62.6 | Blood 2 plus 0.10 gm. urea | per 100 ce. 3 ta — 7.7 7 iets Frog blood. : is — 8.0 7.8 Frog bile. 5 7.7: — 7.5 7.6 Frog muscle. agate mortar before it is put into the apparatus. With such small quantities, grinding with sand is unnecessary and the latter if used will later cause difficulty in the estimation. The crushing one can facilitate by cutting the tissue into very small] bits, and making it as fine as possible with the aid of the mortar. By means of a small card and a glass spatula it can be transferred into the bulb of the apparatus. The mortar is rinsed twice with 0.25 ce. of water, which is transferred to the bulb with a micro pipette. Finally, 0.2 ec. of the-urease anhydride is added. Care must be taken that the tissue is completely covered with the fluid so that the urease can act on the urea present and on this account more water must occasionally be added. Because of the time required for the urea to diffuse from the tissue the reaction with the urease is slower than in the case of THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI, NO. 2 376 A Micro Urease Method blood. When the tissue has been finely divided one can count on an extra 10 minutes being sufficient for the purpose. In Table I some results are given which will indicate the accuracy of the method. As the results show, the average error is 0.4 mg. per 100 ce. The greatest error in individual determinations is 0.5 mg. per 100 em. which indicates that the micro urease method as here given is more accurate than the macro urease method in which there is an error of 1 to 2 mg. per 100 ce. SUMMARY. A micro urease method is given for the estimation of urea in blood, secretions, and tissue. The method gives directly the content of urea plus ammonia, but the ammonia is allowed for by a correction of 0.25 mg. per 100 ce. or can, if desired, be sepa- rately determined. ; The limit of error is 0.5 mg. per 100 ce. I am obliged to Professor A. Krogh, the chief of the Laboratory of Zoophysiology in Copenhagen for his kind assistance and advice during my researches. A SYSTEM OF BLOOD ANALYSIS. SUPPLEMENT III. A NEW COLORIMETRIC METHOD FOR THE DETERMINATION OF THE AMINO-ACID NITROGEN IN BLOOD. By OTTO FOLIN. WITH THE ASSISTANCE OF HSIEN Wu. (From the Biochemical Laboratory, Harvard Medical School, Boston.) (Received for publication, January 30, 1922.) INTRODUCTION. Next to the urea nitrogen the largest portion of the non-protein nitrogen in blood is represented by the nitrogen of the amino- acids.- It is manifestly undesirable to try to obtain figures for this large portion by ‘‘calculation” as is frequently done. The undetermined nitrogen obtained by calculation contains not only amino-acids, but contains also nitrogenous products approaching the polypeptides in composition and character. These products must receive further study and it is partly with the idea of paving the way for the investigations of these unknown products that I have endeavored to find a simple direct method for the determi- nation of the amino-acids in the tungstic acid blood filtrates. The amino-acid work described in this paper was begun in the fall of 1919 and a preliminary report was made at the annual meeting of the American Society of Biological Chemists the following December. I gradually became convinced, however, that the method as we then used it was not sufficiently reliable to merit publication and it is only now after a great deal of further study that I venture to describe and recommend the method. The work was begun as a joint investigation by myself and Dr. Wu; but he was not able to remain until it was finished. My immediate and main object was to find a method for the determination of the amino-acid nitrogen in blood, but I have, Sil ' 378 Blood Analysis of course, also had in mind the determination of amino-acids in urine; and since the preliminary critical work involved the test- ing of all available amino-acids present in proteins a suitable method if obtained would necessarily also be suitable for the determination of such amino-acids in many other materials such as gastric contents, milk, meat extracts, protein hydrolysis mix- tures of various kinds, and innumerable food and medicinal prep- arations, representing various stages of decomposed (hydrolyzed) protein materials. In this connection I must express my great obligation to Drs. H. D. Dakin, C. O. Johns, J. H. Koessler, R. B. Gibson, V. C. Myers, D. W. Wilson, J. C. Bock, and 8. P. L. Sérensen for valu- able gifts of pure aaah preparations. While looking through the literature for hints as to stint kind of a substance might be able to yield the required color reaction I came upon an old observation of Wurster’s to the effect that ordinary o-quinone gives a color reaction with amino-acids and with proteins. This quinone was soon found to be unsuitable, but the observation served as a basis for the investigation of all o-quinones which we could obtain; and through the courtesy of Pro- fessor E. P. Kohler we were able to test out a considerable number. None seemed to be as promising, however, as 6-naphthoquinone- sulfonic acid, the sodium salt of which happened to be among our own stock of chemicals. $-naphthoquinone-sulfonic acid has long been known to give a bright red precipitate with aniline. It has also been used as a color reagent for the quantitative estima- tion of indole with which it gives a blue color; but the fact that it also gives a very intense and decidedly stable color with amino- acids appears to have so far escaped discovery. It may be men- tioned that the reaction with aniline has been shown to be due to the sulfonic acid group, whereas the reaction with amino-acids is presumably due to the o-quinone group, since ordinary o-quinone and many of its derivatives show a tendency to give such reactions. The great majority of the quinones are quite unsuitable, either because of their own intense color or because of their lack of suitable solubility. The fact that the quinones have so much color of their own would seem a priori almost to destroy their usefulness for quantitative work, especially in connection with the analysis of blood filtrates, 10 cc. of which contain less than 0.1 mg. of Otto Folin 379 amino-acid nitrogen. This aspect of the problem has indeed been a great stumbling block and it is only by a happy combina- tion of circumstances that I have been able completely to over- come the difficulty. A colorimetric estimation of amino-acids has been sought for by others, notably by Harding who made use of “ninhydrin” originally introduced by Abderhalden as a test for pregnancy. This reagent is difficult to procure. According to Harding it, moreover, fails in the presence of urea and thus could not be made available for the investigation of blood or urine except on the basis of more or less complicated and uncertain isolation processes. $-naphthoquinone-sulfonic acid on the other hand does not give a color with any of the main nitrogenous waste products, except ammonia, and ammonia is very easily removed. Urea, uric acid, creatinine, creatine, or hippuric acid give no color. As ‘already indicated indole does give a blue color. The color ob- tained with amino-acids is red and, moreover, indole, at least in reasonable amounts, does not give a color with 6-naphthoquinone- sulfonic acid under the conditions used for the production of the color reaction with amino-acids. The blue indole reaction is ob- tained only in the presence of strong alkalies (sodium hydroxide). Theoretically it is, of course, rather unfortunate that 6-naphtho- quinone-sulfonic acid should give a reaction with ammonia. Theoretical considerations are of minor importance, however, in a problem of this sort, because that problem can probably never be solved on the basis of any perfectly specific reaction. The amino-acids are too numerous and too different in chemical constitution-for that. It seems to me that the investigators of the ninhydrin reaction have rather overemphasized the fact that ammonium salts give the reaction. The ammonia in blood is certainly so insignificant as to be of no consequence in con- nection with the determination of amino-acid nitrogen. In the case of urine the situation is different, and there are doubtless other materials where the reaction with ammonia (and amines) cannot be left out of consideration. 380 Blood Analysis Discussion of the Reagents Used for the Determination of the Amino- Acid Nitrogen in Blood. Tl. Standard Amino-Acid Solution.—The standard solution of amino-acid used in blood analysis should contain 0.07 mg. of nitrogen per cc. Where there is a possibility of one’s wanting to make other amino-acid determinations it is more convenient to make a stock solution containing 0.1 mg. per ce. The solu- tion is made with 0.1 nN hydrochloric acid, and 0.2 per cent of sodium benzoate. All the amino-acids (except perhaps trypto- phane) seem to keep indefinitely. From the stock solution con- taining 0.1 mg. of nitrogen per cc. the blood standard is made by diluting 70 ce. with 0.1 n hydrochloric acid to a volume of 100 cc. The question of which amino-acid to use in the preparation of the standard solution depends, of course, on which can be obtained in pure condition. Any one of the following amino-acids can be used: glycine, glutamic acid, leucine, phenylalanine, tyrosine. To this list I might add aspartic acid and cystine. Cystine has the advantage that it can be prepared so easily in strictly pure condition. The color obtained from cystine is, however, a trifle more yellow than the color obtained from the other amino-acids mentioned and in addition it is rather less dependable than the others with respect to the intensity of the color. I have, in fact, had much trouble about adjusting the conditions so as to get the full value from cystine. If the conditions are not right it will be too weak by from 5 to 7 oreven 10 per cent. The objec- tion to aspartic acid is that it is the slowest of all the amino-acids in relation to the development of the color. For miscellaneous work at least, this is a distinct drawback. Glutamic acid also tends to be rather slow and for this reason is the least desirable of the amino-acids mentioned as serviceable for the preparation of standard solutions. Personally I now use glycine exclusively as the standard and my only reason for not recommending this amino-acid alone is that the American supply of the pure article is rather uncertain. This objection is perhaps now not so impor- - tant as it was a short time ago. It is scarcely safe to depend on nitrogen determinations to check up the purity of glycine because the impurities in it contain nitrogen. Glycine, as well as a num- ber of other amino-acids, can easily be recrystallized by dissolving af Otto Folin 381 in water and precipitating by the addition of 0.5 to 1 volume of alcohol. 2. Special Sodium Carbonate Solution——The color reaction be- tween amino-acids and 6-naphthoquinone-sulfonic acid takes place very slowly in neutral solutions. The stronger the alkalin- ity, up to a certain point, the more rapidly does the color develop. The different amino-acids do not show the same acceleration, however, to definite increases in alkalinity, and the chromophoric reagent (the quinone) is more rapidly destroyed when the alka- linity is increased, giving rise to deep colored decomposition products. The finding of the right degree of alkalinity has, therefore, been rather difficult. Sodium hydroxide, magnesium hydroxide, mixtures of trisodium and disodium phosphate, diso- dium phosphate alone, sodium pyrophosphate, borax, sodium bicarbonate, and sodium carbonate—all in various amounts— have been tested out. Particular attention was given to borax, after it had been found that it has a special solvent effect on the quinone and also protects it from spontaneous decomposition, but it was not possible to obtain so good results with borax as could be obtained with other alkalies. The required carbonate solution is made as follows: 50 cc. of approximately saturated solution are diluted to a volume of 500 ec. The strength of the resulting solution is determined by titrat- ing 20 ce. of 0.1 N hydrochloric acid with the carbonate and with methyl red as indicator. On the basis of the titration value thus obtained the carbonate solution is diluted so that 8.5 cc. are equivalent to 20 cc. of 0.1 Nacid. The carbonate solution is about 1 per cent. The titration method given serves as a guide for duplicating the carbonate solution which I am using The correct degree of alkalinity is obtained when 1 cc. of this sodium carbonate solution is added to 1 cc. of amino-acid solution which at the same time is a 0.1 N solution of hydrochloric acid. The alkalinity is, therefore, represented by a mixture of carbonate and bicarbonate. A drop of phenolphthalein solution should always be used, is in fact indispensable when working with amino- acid solutions of unknown and variable acidity. It is desirable that the alkalinity in the different solutions, the standard and the unknowns, should be approximately the same, but there is no need of trying to make them exactly equal. 382 Blood Analysis 3. Fresh 0.5 Per Cent Solution of the Sodium Salt of 8-Naphtho- quinone-Sulfonic Acid.—For the preparation of this quinone in pure form see page 386. Enough of this compound for several thousand amino-acid determinations can be made in the course of two morn- ings. In solution $-naphthoquinone-sulfonic acid is gradually decomposed and the solution becomes visibly darker in the course of a few hours, particularly if it is not kept in the dark. For this reason only freshly prepared solutions should be used. For practical work I find it convenient to charge a series of clean specimen tubes with 100 or with 500 mg. of the dry quinone in roughly powdered form. The samples need not be weighed any more accurately than can be done on a small torsion balance, because a variation either way of 5 per cent makes no difference. Transfer 100 mg. of the quinone to a small flask, add 20 ce. of water, and shake. Complete solution is obtained almost at once. For miscellaneous amino-acid determinations when 0.1 mg. of nitrogen is the standard, 3 cc. of the reagent are taken; for 5 cc. of blood filtrate only 1 ce. 4. Special Acetic Acid-Acetate Solution.—Dilute 100 ce. of 50 per cent acetic acid with an equal volume of 5 per cent sodium ace- tate solution. The presence of the sodium acetate in this solu- tion serves two purposes. It serves to increase unmistakably the color of the quinone-amino-acid derivative, and it retards very much the onset of turbidity due to the liberation of sulfur from the added sodium thiosulfate. Both of these results are due to the weakened acidity of the acetic acid. 5. A 4 Per Cent Solution of Sodium Thiosulfate (Na2S8,03-5H.0). —This solution is used to destroy the surplus quinone remain- ing after the full color obtainable from the amino-acids has developed. It destroys the surplus color. of the quinone and under the conditions prescribed has no effect on the colored qui- none-amino-acid derivative, at least during the first 1 or 2 hours. Nor do the colored solutions become turbid from liberated sulfur within the first 2 hour period. Before I had discovered how to destroy the surplus reagent with thiosulfate the proportionality of the color obtained with differ- ent amounts of the same amino-acid was not at all satisfactory and the amount of reagent which could be used was limited by the fact that with every increase the proportionality became poorer. _* Otto Folin 383 The proportionality was also unduly variable throughout the earlier stages of the investigation because of the influence of light on the spontaneous decompositions. The color from 0.066 mg. of amino-acid nitrogen would give a reading of 25.5 to 27 mm. instead of 30 mm., the theoretical value, when the standard rep- resenting 0.1 mg. was set at 20 mm. Since learning how to destroy the surplus quinone I have used as much as 50 mg. for 0.1 mg. of nitrogen, but as nothing is gained by so doing I have returned to the amount found adequate in the earlier work; namely, 15 mg. of quinone for 0.1 mg. of nitrogen. For the destruction of the quinone eight times as much of sodium thio- sulfate (Na.S:03-5H.O) should be used. Description of the Determination. As the number of practical determinations applicable to the tungstic acid filtrate increases, it becomes increasingly important not to use more of the filtrate for any one determination than is necessary for reliable results. 5 cc. of the filtrate are adequate for the amino-acid determination, but if the filtrate is abundant 10 cc. make the process perhaps a little more convenient. Transfer to a test-tube (capacity 30 to 35 cc.) 1 ce. of the standard acid glycine solution representing 0.07 mg. of nitrogen and add 3 ce. of water. To another similar test-tube add 5 cc. of the blood filtrate. Add 1 drop of 0.25 per cent phenolphtha- lein solution to each. Add 1 ce. of the 1 per cent sodium car- bonate solution to the standard and then add carefully, drop by drop, enough of the sodium carbonate solution to the blood fil- trate until it has approximately the same pink color as the standard (3 or 4small drops are usually required). Add another 5 ce. of water to the standard; the volume of the standard is to be twice that of the blood filtrate. Then prepare a fresh 0.5 per cent solution of the sodium salt of 8-naph- thoquinone-sulfonic acid; add 2 cc. of this solution to the standard and 1 ce. to the blood filtrate. Shake a little to make the solu- tions uniform and set them aside in a completely dark cupboard and leave them there till the following day; that is, for 19 to 30 hours. At the end of the time specified add the acetic acid-acetate solution—2 cc. to the standard and 1 cc. to the blood filtrate. 384 Blood Analysis After the acetic acid has been added (never before) add the thio- sulfate solution—2 cc. to the standard and 1 cc. to the blood fil- trate. Finally add with a “blood pipette” 14 cc. of water to the standard giving a volume of 30 cc. and add 7 cc. of water to the blood filtrate (final volume 15 ec.). Mix and make the color comparison setting the standard at 20 mm. For this work it is more convenient to use test-tubes graduated at 15 and at 30cc. In order not to multiply unnecessarily the kinds of graduated test-tubes needed in blood analysis I now use exclusively the test-tubes, graduated at 25 cc., which are used for the blood urea determination. A second graduation is made at 15 cc. for the blood filtrates and the test- tube containing the standard is first diluted to the 25 cc. mark and an ~ additional 5 cc. of water is then added before the final mixing. 20 divided by the colorimetric reading in mm., times 7; or 140 divided by the colorimetric reading, gives the amino-acid nitro- gen in milligrams per 100 ce. of blood (5 ec. of the blood filtrate corresponds to 0.5 ce. of blood). If 10 cc. of blood filtrate can be spared for the amino-acid de- termination it is more convenient to make use of test-tubes grad- uated only at 25 cc. Transfer 1 ec. of the standard amino-acid solution containing 1 cc. of 0.1 N acid and 0.07 mg. of nitrogen to one such test-tube and to it add 8 cc. of water. To another such test-tube add 10 cc. of blood filtrate. Add 1 drop of phenolphthalein to each. Add 1 ce. of the special 1 per cent sodium carbonate to the stand- ard, and to the blood filtrate add the carbonate solution, drop by drop, until the color obtained matches approximately that of the standard. Add 2 cc. of afreshly prepared 0.5 per cent 6-naphtho- quinone solution to each, mix, and set aside over night in a perfectly dark place. The next day add first 2 cc. of the acetic acid-acetate solution and then 2 ce. of the 4 per cent thio- sulfate solution to each test-tube. Dilute to the 25 cc. mark, mix, and make the color comparison. 140 divided by the colorimetric reading obtained gives, as before, the amino-acid nitrogen in milligrams per 100 ce. of blood. The second process is more satisfactory than the first in the case of bloods which are unusually low in amino-acid nitrogen, be- cause in such cases, when the unknown is much weaker than the standard the match in color is not very good, the weak one being Otto Folin 385 more yellow. This difference in shade is not due to something other than amino-acids in blood, for the same phenomenon is encountered when working with unequal, but unusually weak solution of pure amino-acids. It may become necessary with the first method to use a second standard containing 0.05 mg. of nitrogen, as I do in connection with the determination of the amino-acids in blood plasma. The merits and the shortcomings of the method are adequately shown by the figures recorded in Table I (except in the case TABLE I. Showing the Colorimetric Readings (in Mm.) of Different Amino-Acids with the Color from Glycine as the Standard. 15 mg. reagent. 25 mg. reagent. Amino-acid. (a [ihr yeo rss lbs: | dt breast dvhrsealeal hrs: (Lo si ee 20 20 20 20 20 20 SUN ee a es 30.7 | 30 30.3 | 30.7 | 30.7 | 29.7 (3 | De be a ee 18.4 | 18 18.3 | 18.1 |.18.5 | 18 PARMEIe APM 2. oxen). s. 2. ~| 30 24 | 20.3 | 26 21.5 | 20 CEA MIO tenis a erie sss») 22.6 | 20.3 | 20 20.7 | 20.2 | 20 Ret eS oe one ene ee 25.2 | 24 20 2a0 \pZom le eeoad RR tees cesses 20.5 | 20.6 | 20.6 | 21.7 | 21.2 | 20 Poy IAIEINIE oss es es 20 19.4 | 20 20 20 19.6 co) 2 De 5 Sr ; 20 20 20.3 | 20.3 | 20.3 | 19.6 Wreme tw). 20:6... 2... Pot a Boy e19i6-1:20" 4: 20. (80 TG) ee PIE le20RSu\ 20 22e55 lee 19.6 Jo) Se 17 15 16 17 15 16 Tryptophane (0.5 N).............. Peo 16-5) | A 17 16.5 | 16.5 of arginine) and there is little need for added discussion. Glycine corresponding to 0.1 mg. of N was used as the standard, and another sample of glycine representing 0.066 mg. was introduced to show the proportionality of the reaction. All the other amino-acid solutions were made to contain the same amount of nitrogen as the standard, except in the case of histidine, which reacts with one-third of its nitrogen, and tryptophane, which has been taken to react with one-half. The values obtained for arginine are decidedly unsatisfactory. Because of the peculiarly low values obtained with arginine I was long in doubt about the purity of the 386 Blood Analysis material, but substantially the same values were found for all the arginine preparations including a fine sample of arginine carbonate obtained from Professor S. P. L. Sérensen in Copenhagen and several samples of arginine phosphotungstate prepared by my- self (from protamine). In all of these the color obtained cor- responds only to about 16 per cent of the total arginine nitrogen. The reason why arginine behaves in such a peculiar manner with this amino-acid reagent must be left for further investigations. These color comparisons were made as follows: The amino- acids were dissolved in 0.1 N hydrochloric acid and were so made as to contain 0.1 mg. of nitrogen (or active nitrogen) per cc. To 1 ce. of each solution in a graduated test-tube were added 1 drop of phenolphthalein solution, 1 cc. of the 1 per cent sodium carbonate solution, 7 ec. (or 5 cc.) of water, and 3 ce. (or 5 cc.) of fresh 0.5 per cent amino-acid reagent. The test-tubes were then placed in a dark closet. At the end’ of 1, 3, or 2} hours the contents were acidified by the addition of 1 cc. of special acetic acid-acetate solution. 3 ce. (or 5 ec.) of the thiosulfate solution were then added; the solutions were made up to volume (25 ce.); and the color comparisons were made (within 1 hour). Detailed Description of Preparation of the Amino-Acid Reagent. The process described below for the preparation of strictly pure 8-naphthoquinone-sulfonic acid sodium salt is the outcome of a great many trial experiments. No effort has been spared to make _ the preparation simple as well as reliable. From the list of needed chemicals and from the different steps in the process enumerated, the preparation may seem a rather formidable undertaking but in actual practice it will be found that the amount of work in- volved is not very large. The description has purposely been made so explicit that a person with limited chemical experience cannot go astray, except by not following the directions. I would warn against introducing variations or modifications of any kind, for I have tried a great number of plausible short cuts and variations, only to find that they had no merit. The following chemicals are needed for the preparation of 75 to 90 gm. of the pure reagent: oo Otto Folin 387 Cold 10 per cent sulfuric acid 1,000 ec. Concentrated nitric acid 100 ce. 4 hydrochloric acid 500 ce. 10 per cent sodium hydroxide solution 300 cc. epee “chloride solution 2,600 ec. Bromine 1 ce. Sodium nitrite 50 gm. “nitrate 100 gm. «sulfite 50 gm. “bisulfite 100 gm. Borax 400 gm. Resublimed 8-naphthol 100 gm. Alcohol about 2,000 ec. Ether about 200 ce. Ice 1,000 gm. 1. Transfer 100 gm. of 6-naphthol to a liter beaker; add 300 ec. of 10 per cent sodium hydroxide solution and stir with a glass rod until complete solution is obtained (10 to 15 minutes). 2. Transfer 50 to 55 gm. of sodium nitrite to a 4 liter beaker; add 600 cc. of water and shake until solution is obtained (3 to 5 minutes). 3. Pour the alkaline 6-naphthol solution into the 4 liter beaker holding the nitrite solution, and rinse with about 100 cc. of water. 4, Add 800 gm. of crushed ice to the naphthol-nitrite mixture. 5. Fill a 200 ce. cylinder with cold dilute (10 per cent) sulfuric acid and pour it slowly down one side of the beaker, while stirring vigorously and continuously with a heavy glass rod. Continue the stirring for 1 to 2 minutes after all the acid in the cylinder has been added. Then fill the cylinder again and add this in the same way. Repeat this addition of dilute sulfuric acid, 200 ce. at a time, until 800 ec. have been added. The additions should be continued until the mixture in the beaker gives a distinct and permanent acid reaction with Congo red paper (time 15 to 20 minutes). . A yellow precipitate begins to form with the first addition of acid and increases in quantity until the whole mixture becomes a semisolid paste. The precipitate’ will have a slight greenish tint; if it is distinctly green the conditions are not right and a less good yield is obtained. Let stand for 1 hour after the last of the acid has been added. It is important not to omit this detail because without such a period of standing much unchanged 388 Blood Analysis B-naphthol remains in the mixture and is encountered when try- ing to dissolve the precipitate in sulfites. Longer standing does no harm, but is superfluous. 6. Filter through a 20 em. Buchner funnel with moderate suction and wash with about 1,500 cc. of cold water. 7. Transfer the precipitate (nitroso-G-naphthol) by blowing, to a large evaporating dish and sprinkle over it 100 gm. of sodium bisulfite and 50 gm. of sodium sulfite; stir with a spoon (glass or enameled ware). An extremely soluble bisulfite addition product is formed and the mixture becomes liquid. Filter imme- diately on a Buchner funnel (diameter 12 fo 15 em.) through a double layer of good (quantitative) filter paper, from the small amount of black residue and wash with a little water. 8. Transfer the filtrate and washings at once (to avoid excessive darkening) to a 5 liter flask or wide mouth (colored) bottle, con- taining 2,000 cc. of water and 500 ce. of concentrated hydrochloric acid. Cover with a funnel and one or two watch-glasses and let stand in a dark closet for about 36 hours (24 hours is not quite enough). The whole flask becomes filled with a network of white needles which carry down a little dark and pink matter as impuri- ties. The greater the exposure to light the more of the dark decomposition products will be formed. Filter on a 20 em. Buchner funnel with moderate suction and wash with about 2 liters of cold water. 9. Blow the precipitate (1-amino-2-naphthol-4-sulfonic aN to a large filter paper and from there transfer it to a large beaker (3 to 4 liter capacity). Sprinkle over the precipitate 100 gm. of sodium nitrate. Dilute 100 ec. of concentrated nitric acid with 350 ec. of water and pour the whole of this lukewarm dilute acid into the beaker. Reaction: begins immediately and nitric oxide fumes begin to come off. Leave without stirring for 10 minutes while the greater part of the reaction takes place. Then stir thoroughly for a few minutes and let stand for another 20 to 30 minutes. If no visible reaction takes place on adding the nitric acid the ~ cause is probably to be found in the presence of sodium carbonate in the sodium nitrates used. Even samples of sodium nitrate which are labeled ‘‘The Standard of Purity”? may contain consid- erable amounts of carbonates. If no reaction takes place add a little (1 to 5 ec.) of concentrated nitric acid. Pec ee = _ Otto Folin 389 At the end of about half an hour filter on a Buchner funnel (diameter 15 cm.) and wash with about 1,000 cc. of 10 per cent sodium chloride solution. The light brown product on the funnel is the desired sodium salt of 1-2-4-naphthoquinone-sulfonie acid. But it is not pure. It contains dark-colored decomposition products, and also traces of ammonia. Considerable ammonia is formed (from the amino group) during the oxidation with the nitric acid, but because of the large amount of sodium nitrate present only traces of the ammonia are carried down as the ammonium salt of the quinone. By recrystallization under the conditions described in the next section (Section 10) the product is freed from all disturbing impurities. 10. Transfer the moist precipitate to a large porcelain dish. Add 200 gm. of powdered borax and 450 cc. of water. Mix with a pestle until all but a few flakes of the quinone have dissolved. Filter through a good quality (quantitative) filter paper on a 10 em. Buchner funnel from the surplus borax and a little undissolved black residue. Because of the latter the filtration is apt to be rather slow and it is better not to apply too strong a suction. ~ Wash with 100 to 150 ce. of water. While the filtration is proceeding, transfer 850 ec. of 95 per cent alcohol and 150 cc. of concentrated hydrochloric acid to a Florence flask. Cover the mouth of the flask with a beaker and cool under running water. Transfer the quinone-borax filtrate to a 4 liter beaker. Add a few drops of liquid bromine to the cooled acid-alcohol. Shake until the bromine is dissolved and then pour the resulting straw- yellow solution into the quinone-borax mixture and stir quickly and vigorously for a few moments so as to secure complete mixing. Let stand for 5 minutes. The quinone has all come down at the end of this time. Filter on a Buchner funnel (diameter 15 cc.) and wash with 700 to 800 cc. of 10 per cent solution of sodium chloride. This one recrystallization is adequate as far as the color of the product is concerned, but it still contains some ammonia. The recrystallization must, therefore, be repeated once more. This final recrystallization is conducted in the same way, except with the difference that the washing of the quinone on the Buchner Ms 390 Blood Analysis funnel with sodium chloride is omitted, and for it is substituted the washing, first with 300 to 400 ce. of alcohol and finally, with about 200 cc. of ether. 75 to 90 gm. of perfectly pure quinone is thus obtained. The purification process described above is the outcome of a large number of experiments directed toward obtaining this quinone free from colored decomposition products. The com- pound cannot be recrystallized from water alone, because each time it is dissolved there is some decomposition and when precipi- tation is made (usually by the addition of sodium chloride) the decomposition products adhere to the compound and it thus be- comes darker than it was before. Before I discovered the protecting and dissolving action of borax on 8-naphthoquinone-sulfonic acid I had prepared it in pure form by the fol- lowing much more expensive and wasteful process. To 100 gm. of powdered quinone in a 2 liter beaker were added: first, 1,000 cc. of bromine water; and then, with stirring, 700 cc. of boiling water. The warm solution was at once Fiiiotad through a large folded filter into a 4 liter flask. To the filtrate was added 2,500 cc. of 95 per cent alcohol followed immedi- ately by 150 gm. of powdered sodium chloride. The mixture was shaken hard under running water until cold (10 minutes) and was filtered and washed with alcohol and ether. A second similar recrystallization (but omitting the first flieetion) gave a perfectly pure product. As this process required the use of 6 liters of alcohol and involved considerable losses of quinone the borax process described above is manifestly much superior. The following tests for the purity of the sodium salt of 1-2-4- naphthoquinone-sulfonic acid will be found useful. Color.—Prepare a fresh 1 per cent solution of the quinone in water and compare its color with that of a 0.5 N solution of potas- sium bichromate with the latter set at 20 mm. in the colorimeter. The quinone solution will read 26 to 27 mm. Colored Decomposition Products.—Transfer 2 cc. of the fresh 1 per cent quinone solution to a test-tube. Dilute to a volume of 25 cc. Add: first, 1 ec. of 50 per cent acetic acid; and then, 1 ce. of 15 per cent sodium thiosulfate solution. The solution will bleach in the course of a few seconds so completely that it is only by looking through the length of the tube that a faint yellow shade is visible. pwr. Otto Folin 391 Ammonia.—Transfer 10 cc. of the solution to a small flask. Add about 2 gm. of permutit and shake gently for 3 to 4 minutes. Decant and wash with distilled water four or five times, until all of the yellow color is gone. Then add to the permutit powder a few drops of 10 per cent sodium hydroxide solution, 5 ec. of water, and 5 cc. of Nessler’s solution. No color is produced. an Gas), <4 vet j ty oh 7 cna a hy on ee ae Vetere hey iw hevhietagh i S110 oom MN WT bch at a s nid A COLORIMETRIC DETERMINATION OF THE AMINO- ACID NITROGEN IN NORMAL URINE. By OTTO FOLIN. (From the Biochemical Laboratory, Harvard Medical School, Boston.) (Received for publication, January 30, 1922.) For the colorimetric determination of the amino-acid nitrogen in normal urine by the method described in the preceding paper! it is only necessary to first remove the ammonia. The removal of the ammonia is easily accomplished by means of permutit. -Urine differs from blood filtrates in that the concen- tration of the ammonia to be removed varies within very wide limits and the amino-acid concentration is also subject to very large variations. It is, therefore, not always easy to tell just how much urine should be taken. The amino-acid excretion is sub- stantially independent of the volume of the urine and, since the normal excretion is usually between 4 and 12 mg. per hour, one can practically always get suitable amounts for the color reaction. The process is as follows: Dilute from 5 to 25 cc. of urine to a volume of 25 cc. ina 50 ce. Erlenmeyer flask. Add 2 to 3 gm. of permutit? and agitate very gently, but continuously for 5 minutes. Decant the supernatant urine into another 50 ce. flask. Again add 2 to 3 gm. of permutit, and shake as before for 5 minutes. By this double extraction with permutit every trace of ammonia is removed. Decant the supernatant urine into a flask or test-tube. It may be a little turbid, but this fact does not interfere with the determination. To test-tubes graduated at 25 cc. add 1, 2, and 3 cc., respectively, of a standard glycocoll solution in 0.1 N hydrochloric acid plus 1 Folin, O., J. Biol. Chem., 1922, li, 377. ? The special permutit to be used is that prepared by the Permutit Company of New York. The product, I believe, is now obtained from the leading American dealers in chemicals. It is, of course, essential to know that the product is active and does take ammonia. 393 394 Amino-Acid Nitrogen in Normal Urine 0.2 per cent of sodium benzoate. This standard solution should contain 0.1 mg. of glycocoll nitrogen per cc. To these tubes add 1, 2, and 3 cc., respectively, of the special 1 per cent sodium carbonate solution described in the preceding paper! (1 ce. of sodium carbonate for each cc. of 0.1 N hydrochloric acid present). Dilute the contents of each test-tube to a volume of 10 cc. Transfer 5 ce. of the ammonia-free (usually diluted) urine to another test-tube graduated at 25 ec. Add 1 ce. of 0.1 N hydro- chloric acid and 1 ce. of the 1 per cent sodium carbonate solution. Dilute to 10 cc. Dissolve 250 mg. of the amino-acid reagént in 50 ec. of water, and add 5 ec. of this solution to each standard and to the unknown urine. Mix and set in a dark place over night. It is often advisable to take out the test-tubes and inspect them after they have stood 10 to 15 minutes. If the test-tube containing the urine appears much darker than the darkest standard, as may happen, especially in connection with experiments planned to produce excessive amino-acid excretion, then it is, of course, necessary to start another sample of the urine, taking only 1, 2, or 3 cc. and treat- ing it in the same way as the first sample, not omitting to provide for a final volume of 15 ce. The following day the standard and the unknown or unknowns are first acidified by the addition of 1 ec. of the special 25 per cent acetic acid-acetate solution. To each are then added 5 ce. of the 4 per cent sodium thiosulfate solution. The contents of all the tubes are diluted to a volume of 25 cc. and, after mixing, the color of the unknown is read against that of the standard having most nearly the same intensity of color. For the calculation it is, of course, essential to know which standard is used, and the actual volume of undiluted urine taken for the, determination. (HIGCING meee =< THE RETENTION AND DISTRIBUTION OF AMINO-ACIDS WITH ESPECIAL REFERENCE TO THE UREA FORMATION. By OTTO FOLIN anp HILDING BERGLUND. (From the Biochemical Laboratory, Harvard Medical School, and the Medica! Service of the Peter Bent Brigham Hospital, Boston.) (Received for publication, January 30, 1922. CONTENTS. Ee er es 395 Critical and experimental examination of the hypothesis that deamina- tion and urea formation are localized in the liver................ 396 On amino-acid excretion as a test for liver function.................. 405 Significance of the normal excretion of amino-acids.................. 408 Distribution of the normal non-protein nitrogen, including an hypoth- esis as to the character of the undetermined nitrogen............ 417 INTRODUCTION. In connection with our recently described investigations on the transportation, retention, and excretion of carbohydrates, we have made a very large number of analyses of blood and urine bearing on the simultaneous fluctuations of nitrogenous constituents such as amino-acids, urea, and creatine. Since one of the questions involved was that raised by Falta concerning the permeability of human blood corpuscles it seemed well worth while to settle this question for the nitrogenous products as well as for sugar— a settlement which we believed that our method for separating plasma enabled us to furnish. Our amino-acid determinations were made by the new colorimetric methods described in the two preceding papers.! . The extent to which the amino-acid excretion is influenced by fluctuations in the amino-acid concentration in the blood represents a practically untouched problem. While one cannot expect to find any really constant relationship between the concentration 1 Folin, O., J. Biol. Chem., 1922, li, 377, 393. 395 THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI, NO. 2 396 Amino-Acids with Reference to Urea in the blood and the concentration in the urine or the hourly excretion, it is important to have data on the normal simultaneous variations in blood and urine, because these data can at least serve as a guide and check by which to judge alleged pathological deviations. At present it is for,example impossible to tell what significance, if any, can properly be ascribed to the “excessive”’ amino-acid excretion which is obtained after giving glycocoll or gelatin to patients in whom the deamination power of the liver is believed to be reduced. We shall discuss this test for the liver function in connection with some of our experiments. Critical and Experimental Examination of the Hypothesis that Deamination and Urea Formation are Localized in the Liver. The localized character of the deamination process by which the absorbed amino-acids lose their nitrogen and yield urea was definitely disproved by Folin and Denis? in a series of articles published in 1912. By determinations of the urea and the non- protein nitrogen in the blood during the absorption of single pure amino-acids, glycine, alanine, aspartic acid, and asparagine they proved that deamination and urea formation did not occur, at least to any material extent, while the amino-acids passed through the walls of the intestine or through the liver. On the other hand, they did obtain increases in the non-protein nitrogen of the blood, and also of the tissues, in amounts sufficient to account for approximately the whole of the absorbed nitrogen. They obtained substantially the same results whether they used a single amino-acid, or pancreatic digestion mixtures which could have recombined into protein material during the absorption. These results definitely excluded both deamination and protein regeneration as features of amino-acid absorption and were inter- preted as excluding equally definitely the doctrine that the urea formation is localized in the liver. The experimental results of Folin and Denis were abundantly confirmed the following year (1913) by Van Slyke,’ and as his 2 Folin, O., and Denis, W., J. Biol. Chem., 1912, xi, 87; 1912, xii, 141. 3 Van Slyke, D. D., and Meyer, G. M., J. Biol. Chem., 1912, xii, 399; 1913-14, xvi, 197, 213. O. Folin and H. Berglund 397 results were obtained by his new gasometric method for the deter- mination of amino-acids he naturally regarded them as even more conclusive. In one fundamentally important point Van Slyke’s conclusions differed from the conclusions drawn by Folin and Denis. He revived once more the old doctrine that the formation of urea from the nitrogen of amino-acids is a special function of the liver. The experimental results which Van Slyke interpreted as evi- dence in favor of the liver as a special ‘locus’ for the urea forma- tion are very important. Unless the interpretation adopted by Van Slyke is correct those results tend to prove quite the opposite, namely that the liver does not even transform the nitrogen of the amino-acids which it does take up into urea, but soon lets go of those amino-acids for distribution to the other tissues. The experimental results in question are as follows: The liver takes up amino-acids more rapidly and extensively than the muscles just as might be expected from its enormous blood supply. Within 3 hours after the injection of the amino- acids into the blood the surplus stored in the liver has already left it, whereas the amino-acids absorbed by the muscles have then scarcely begun to diminish. Is this extraordinary disappear- ance of the amino-acids from the liver due to urea formation in that organ or is it due to the same cause by which it at first re- ceived an excessive proportion of the amino-acids, namely the abundant supply of blood, of blood from which the muscles and other tissues have in the meantime abstracted much of the amino- acids? In our opinion the last named process is the more plausible explanation of Van Slyke’s interesting data. The vitality of the old concept that the liver representsaspecial- ized locus for the urea formation is such that it has received over and over again the benefit of the doubts, in attempts to explain data which could possibly be construed as having any bearing on the origin of urea. The worthlessness of all the older evidence purporting to show that the urea was formed more or less ex- clusively in the liver was clearly brought out by Miinzer and Winterberg* in 1893. The theory was again revived as a corollary to the hypothesis that-the amino-acids lose their nitrogen during 4 Miinzer, E., and Winterberg, H., Arch. exp. Path. u. Pharmakol., 1893- 94, xxxiii, 164. 398 Amino-Acids with Reference to Urea the process of absorption; and almost immediately after this hypothesis, with its corollary, had been destroyed by the experi- ments of Folin and Denis the theory of the predominant urea forma- tion in the liver was again revived by Van Slyke. Van Slyke does not claim that his results prove that urea is formed mainly in the liver, but in adopting such urea formation as the most plausible explanation of his experimental results, he has again given a new lease of life to this old and popular theory. In Van Slyke’s interpretation the old theory is formulated with precision and in terms of the modern knowledge of the chem- istry and the metabolism of the proteins. The surplus of amino- acids stored in the liver, in consequence of the excessive absorption, is speedily deaminized and the nitrogen removed in the form of urea, and, later, the amino-acids stored in the muscles and other tissues are likewise transferred to the liver and there subjected to the same process. As we cannot accept this interpretation we purpose to discuss the available data bearing on the subject. The experimental conditions of Van Slyke as well as those of Folin and Denis led to excessive temporary storage of amino-acids both in the liver and in the muscles. It is possible that in con- sequence of such excessive retentions the muscles received more material than they could utilize, and that the surplus under such conditions is transferred to the liver as soon as this organ has relieved itself of its own excessive supply and is capable of receiv- ing the new influx. It seems to us at least equally plausible, however, that a process substantially the reverse of that thus described by Van Slyke takes place under such conditions. The liver, partly by its location, partly because of its enormous blood supply, acts as a buffer or safety valve in relation to a large and rapid influx of amino-acids from the digestive tract or as a result of amino-acid injections. The absorption by other tissues, especially the muscles, is not so rapid, because all the capillaries through which the muscles receive materials are not equally and simultaneously open (Krogh), hence all the muscles are not filled equally fast; but for the same reason the muscular tissues, taken as a whole, can continue to take up materials for a consid- erable time. The muscles thus continue to abstract amino-acids from the blood, but the loss in the latter is made good by corres- 5 Van Slyke, D. D., The Harvey Lectures, 1915-16, xi, 146. O. Folin and H. Berglund 399 ponding acquisitions from the liver, until after 3 hours, according to Van Slyke’s results, the liver has been freed from its surplus. For evidence on the problem one must turn to the data avail- able on the urea production. Van Slyke’s hypothesis requires immediate and rapid urea formation while the liver is loaded, practically to full capacity, with amino-acids. At the end of 3 hours when the liver has only its normal amount and the muscles are still nearly at their maximum content of amino-acids the urea formation should certainly be reduced to a very small fraction of what would take place while the liver is full. Van Slyke recog- nizes the importance of immediate and extensive urea formation for the maintenance of his theory and in his Harvey lecture,* given 3 years after the publication of his paper on the locus of the urea formation, he refers to experiments purporting to show that the urea formation begins the very moment that amino-acids begin to reach the liver. These experiments are lacking in the precision necessary to prove so fine a point. They consisted of heavy meat feeding to dogs and finding that the urea content of the blood had already begun to rise, when the gastric contents were beginning to enter the duodenum, the exact moment of which was determined. by means of x-rays. But Folin and Lyman® showed (1912) that absorption from the stomach takes place and leads to an unmis- takable increase of the urea in the blood. Such absorption must manifestly have occurred in Van Slyke’s meat feeding experi- ments, not only on the basis of the digestion products formed in the stomach, but also on the basis of the ammonia and urea present in the meat. More definite data on the urea formation associated with the amino-acid absorption are contained in the papers of Folin and Denis. Take for example Experiment 3 of the first paper. After a 6 minute absorption period following the injection of glycocoll into the intestine the non-protein nitrogen of the blood had risen from 30 mg., the original value, to 34 mg., and the urea nitrogen had risen only 1 mg.—from 18 to 19; and during the following 12 minutes the non-protein nitrogen rose from 34 to 47 mg. while the urea nitrogen increased only from 19 to 22 mg. The very 6 Folin, O., and Lyman, H., J. Biol. Chem., 1912, xii, 259; 1912-13, xiii, 389. 400 Amino-Acids with Reference to Urea large increase in the non-protein nitrogen in this blood showed that the liver had been flooded with glycocoll and that the muscles also had been offered a more abundant supply than they were able to take up. Yet the urea nitrogen content of the blood. was increased by only 3 mg., and this notwithstanding the fact that the exit of urea through the kidneys was prevented by ligatures. The validity of these results and other similar ones have been accepted by Van Slyke and others, including Abderhalden, as showing that deamination of amino-acids does not take place while they pass through the walls of the intestine. But no one has attempted to explain why they do not also show that no material deamination and urea formation occurs while they pass through the liver. Whether the given, relatively small, increase in the urea nitrogen of the blood was due to the liver, the muscles, or both, is of course debatable, but the figures certainly show that the muscles had had an ainda opportunity to take part in the urea formation that did occur. To prove that the liver is even the predominant place for deamin- ation and urea formation it is necessary to show that during the early stages of absorption of amino-acids the increase in the urea content of the blood precedes the increase in its content of amino-acids. If the increase in the amino-acid content precedes that of the urea as in the experiment cited this does not conclu- sively prove that the liver is not the special seat of the forma- tion, because the urea-forming process may be very much slower than the absorption process under such unusual conditions. But such a condition certainly leaves intact the view that the liver has not developed any special powers of deamination. In order to obtain more convincing evidence as to whether the liver has the function ascribed to it by Van Slyke it is necesary to provide for a more moderate speed of absorption than prevailed in the experiments of Van Slyke or of Folin and Denis. And on the basis of the further refinements in the analytical technique which has been obtained since 1912, it is now possible to obtain unmistakable results under much more normal conditions. As all are now agreed that the greater part of the protein is absorbed as simple amino-acids and that the amino-acids, in so far as they pass the liver, are absorbed by all the other tissues,’ the whole problem has really narrowed down to the one question of whether O. Folin and H. Berglund - 401 the absorbed nitrogen does or does not pass out of the liver in the form of urea. In Table I we give the results obtained from one substantial meal taken, without any preceding breakfast, at 11.45 a.m. to 12m. We have two experiments of this sort made at the same time on two different persons, but, as the results are substantially identical, only one is recorded. The meal consisted of three boiled eggs, a pint of milk, together with bread, butter, and coffee. From the figures given it will be seen that from such a perfectly ordinary mixed meal there was obtained a definite, unmistakable increase in the amino-acid content of the blood. Within 14 hours the amino-acid content of the plasma rose from 5.2 to 5.8 mg. per 100 cc. In the course of the next hour it rose to 6.4 mg., and 53 hours after the ingestion it had again sunk to very near the fasting value. The figures cited will naturally raise the question as to whether the amino-acid method is really capable of finding so small a difference as only 1 mg. of amino-acid nitrogen per 100 cc. The difference amounts to 20 per cent, and we are certain that the figures, relative to each other, are correct. The amino-acid determinations were all started and finished together. It will further be noted that the amino-acid excretion with the wrine confirms the values obtained for the blood. As we have other . different experiments it is not necessary to dwell too much on this one. It is given mostly for the purpose of showing that an ordinary meal, not excessively rich in protein, is accompanied ‘by a slowly increasing concentration of the amino-acids in the blood, an important fact since we know that accompanying that increase there must be a large increase in the amino-acid content of the muscles. Table II gives the figures for blood and urine obtained after taking 135 gm. of (Knox’s) gelatin, dissolved in 900 cc. of water. In taking so large a dose of protein we have, of course, exceeded the strictly normal conditions, for man, but the results obtained are also correspondingly striking. It should be explained that the subject, H.B-d., had eaten freely, and probably rather too liberally, for 1 or 2 days before and that the fasting values for the blood at the beginning of the experiment were, therefore, unusually high. 402 Amino-Acids with Reference to Urea TABLE I. Subject H. B-d. Age 34 years. Weight 90 kilos. Time. Blood per 100 cc. Urine per hour. Remarks. Z J Mar. 20, 1921. z a z oars page a a Be a Be 4)| @ologe ta) 2 aaa 4)p)4 |e |e hae a.m. mg.| mg.| mg.| mg. | cc. | mg.|mg.| mg. Blood. 6.2)10.4)11.8)28.4 11.42 Plasma. 5.2)11.4| 5.7/22.3) 80\7.2/431/571 Corpuscles.| 8.0} 8.6)/22.6|39.2 11.45 a.m. : Lunch: 3 eggs, to 450 ce. milk, 12200, eer: bread, but- ter, and cof- fee. p.m Blood. 6.6)11.0 1.23 Plasma. §.8/11.8) 5.2/22.8 Corpuscles.| 8.0) 9.6 2-20 51/9 .3|470/572 Blood. 7.2)10.8)12.4/30.4 2.33 Plasma. 6.4/12.0) 6.0/24.4 Corpuscles.| 8.6) 8.7|23.7|41.1 3.47 43/8 .9)392/491 Blood. 7.4)11.0/15. 2/33 .6 4.08 Plasma. 6.4/13.2) 6.4/26.0 Corpuscles.| 9.2) 7.1/31.0/47.2 5.00 64/9 .8|520)644 Blood. 6.4)/12.1]13.5/32.0 5.42 Plasma. 5.613.4) 6.0/25.0 Corpuscles.| 7.8) 9.7|27.0/44.5 6.10 | a oe 572 O. Folin and H. Berglund 403 TABLE Il. Subject H. B-d. Age 34 years. Weight 90 kilos. Time. Blood per 100 ce. Urine per hour. Remarks Zz 2 E 3 ‘ Feb. Z 2 Z Fasting. Uri- ra = zr |3 ! niendagiaatieasi. 2/2/4212 /e|3] 24] 2 2lelelsal/2l2l¢| 3 Stee fe le pa) SPs p.m. mg.| mg. mg mg ce mg. mg mg Blood. 6.7/15.2/17 5/39 .4 2.25 | Plasma. 5.5 17.3 6229.0 27 |10.0| 403) 463) 2.55 to 3.20 Corpuscles. | 8.412.2)33.8)54.4 p.m. Took 135 gm. gel- atin and 900 cc. water. Blood. 7 .6)14.2/17 639.4 Fasted during 4.00 | Plasma. 9.416.0 8.2:33.6 the whole Corpuscles. | 5.1/11.7/30.6,47.4 experiment. 4.37 39 | 9.3) 459) 518 Blood. 10.7/15.2)18.9/44.8 4.52 | Plasma. ("1:93 Shee ee Corpuscles. |10.2)16.6/25.652.4 6.03 39 43.0, 540) 651 Blood. 11.7/18.9]17.9]48.5 6.10 | Plasma. 7.2'20.213.6.41.0 Corpuscles. |13.5|16.9/29.8 60.2 q-10 77 (96.0) 962/1,158 10.34 82 |63.0/1,095)1 ,327 Blood. 7.0/24.618.9}50.5 10.55 | Plasma. 6.8 25.011.4.43.2) Corpuscles. | 7.4/24.030.4161.8 a.m. 9.00 | 30 — ig 783, ! 404 Amino-Acids with Reference 6 Urea In this case the amino-acid content of the plasma increased 100 per cent in 2 hours, rising from 5.5 to 11 mg. Gelatin is believed to be the most easily digested protein material and our results are in harmony with that view, because within 3 hours the amino content of the plasma had already fallen from the high point, 11 mg., to 7.2 mg. It is by no means certain, however, that gela- tin taken in large doses does not in part escape complete digestion. Our figures for the undetermined (rest) nitrogen are quite high enough to permit the conclusion that some materials of a peptide character were absorbed. The most interesting facts about this experiment are the fluc- tuations of the amino-acids and the urea. In our opinion they definitely prove that the liver has no specialized function in connection with deamination and subsequent urea formation. Here we have a 100 per cent increase in the amino-acid nitrogen of the blood without any increase in the urea content of the same, accompanied in fact by a slight diminution in the concentration of the urea. The urea formation clearly begins to gather force only after the incoming tide of the amino-acids has begun to subside. During the first 2 hours while the amino-acid nitrogen rose from 5.5 to 11 mg. the urea nitrogen sank from 17.3 to 14.2 mg. Then during the next hour and a quarter, while the amino- acid nitrogen sank to 7.2 mg., the urea nitrogen rose from 14.2 to 20.2 mg. From that point on there are only further reduc- tions in the level of the amino-acids, but the concentration of the urea nitrogen increases still further, and 8 hours after the gela- tin ingestion it stands at 25 mg. in the plasma. The figures for the urea nitrogen or the total nitrogen of the urine tell the same story as the figures for the blood. During the first 2 hour period, while the amino-acid accumulation in the blood was approaching its maximum, the urea nitrogen excretion (per hour) rose only from 403 to 459 mg. At the end of about 3 hours, a period including the first real beginning of active urea accumulation in the blood, the urea nitrogen of the urine is still only 540 mg. per hour. At the end of this period the urea excre- tion jumps in a single hour from a rate of 540 mg. of urea nitrogen to one of 962 mg. per hour. And during the following 3 hours, beginning about 4 hours after the gelatin ingestion, urea nitrogen is excreted at the rate of 1,095 mg. per hour. O. Folin and H. Berglund 405 While we, for the sake of conciseness, have concentrated the discussion of the alleged localized urea formation on the gelatin experiment, the reader is asked to examine Tables I to VI. They all show that the predominant increase in the non-protein nitrogen of the plasma during the early stages of absorption is associated mainly with an increase in the amino-acid nitrogen—and the later stages with increases in the urea nitrogen. As a check on the results note also that in Table VIII, where we are dealing with ‘ carbohydrate absorption, no such increases are to be found. It is perhaps also worth while to point out that the results have been obtained with normal human subjects, and under conditions which are as nearly normal as it is possible to make them. As far as we can see there is now nothing more that need be said concerning the old tradition that one special function of the — liver is to split off amino nitrogen and convert it into urea. We _ therefore refrain from referring to the mass of other evidence showing that the urea production continues when the liver is virtually eliminated, damaged, or destroyed. On Amino-Acid Excretion as a Test for Liver Function. A few brief comments on the so called liver function tests in cirrhosis of the liver may not be superfluous. These tests consist usually of giving 25 gm. of glycocoll or 50 gm. of gelatin, and then determining the amino-acids, by the formol titration, and the total nitrogen in the 24 hour urine. It is held that in advanced cirrhosis the percentage of nitrogen excreted as amino-acids is distinctly higher than the figures obtained under similar conditions from normal persons. From the different tables included in this paper it will be seen that increases in the amino-acid excretion following the intake of nitrogenous food like gelatin or glycocoll does not last for more than a few hours. If the intake has been large the excretion will outlast the high level in the blood, presumably because the kidneys, like other tissues, have received excessive amounts. Even under these circumstances the excretion comes to an end long before the 24 hour period is over. 24 hour periods are, therefore, unsuitable for the test, especially if stress is laid on the percentage of the total nitrogen contained in the amino-acid fraction, for the urea excretion comes in later and obscures the 406 Amino-Acids with Reference to Urea result. It might be thought that our findings concerning the special deamination power of the liver has completely removed the scientific foundation upon which the liver function test in question rested. As a test for a non-existent liver function the test has no justi- fication, but we are not prepared to say that the test in suitably modified form may not have some value. A substantial scientific foundation for the test is to be found in the fact, discovered by Van Slyke, that the normal liver does take up an excessive charge of amino-acids when the blood passing through it is rich in the same. It serves, as we have expressed it, as a safety valve or buffer, thus materially diminishing the concentration in the general circulation, and thereby also the elimination with the ‘urine. But in cirrhosis the liver cells gradually disappear until practically nothing or very little is left. As the liver diminishes, its activity as a buffer is reduced and the consequence might well be an abnormally large excretion with the urine, especially during the first 4 to 6 hours. But the test is, of course, not a functional test. It is only a test of the mass of the remaining liver cells. In Table III we give the analytical data for blood and urine following the intake of 25 gm. of glycocoll in 600 cc. of water. The amount is perhaps rather small for a person weighing 90 kilos but we believe that the figures obtained may, nevertheless, serve as a normal basis of comparison for the making of similar experiments on the subject suffering from cirrhosis of the liver, on the basis of 4 to 6 hour periods. It might be pointed out that for the amino-acid determinations in the blood the plasma figures are necessarily somewhat more instructive than those obtained with whole blood, because the latter include the more variable values obtained from the corpuscles. There is no justification for insisting on using the plasma, however; to do so requires the use of more blood, is more laborious, and does not involve the finding of anything which would escape detection by means of whole blood analysis. From the figures for the blood taken at 11.26 a.m. it will be seen that at the end of only 20 minutes absorption there is an unmistakable increase in the amino-acid content of the plasma (and whole blood). At the end of 2 hours the amino-acid nitro- O. Folin and H. Berglund 407 TABLE III. Subject H. B-d. Age 34 years. Weight 90 kilos. Time. Blood per 100 cc. Urine per hour. Remarks. Z 2 F . 3 oe ° Jan. 11, Ei rh = 3s Fasting. Urinated 1922. = 4 3 |6 at 9.00 a.m. & | ae & Z 2h) |e is ar B= 3 » | 3 ime ee, 3 a Bape ud | Sel ot of a de ae eh | ee) | ty | elt ee a.m. mg.| mg.| mg.| mg.| cc. | mg.| mg.| mg. Blood. 6.05/10 .3|13.7|/30.0 11.03 | Plasma. 4.8 |13.0| 7.0/24.8) 28 | 5.8) 452) 474 Corpuscles. |8.1 | 6.0/24.3/38.4 11.06 Took 25 gm. glycocoll and 600 ce. water. Blood. 6.7 |11.3]12.4/30.4 11.26 | Plasma. 5.8 |13.9) 6.3/26.0 Corpuscles. |8.3 | 6.7|23.4)38.4 p.m. Blood. 7.85/12 .7|11 .8)32.4 12.08 | Plasma. 8.3 |14.7| 6.5/29.5 Corpuscles. |7.1 | 9.0/21.7|37.8 ess! 49 115.8} 585} 641 Blood. 8.0 }11.8]11.2/31.0 1.12} Plasma. [7.8 |12.5) 9.8/30.1 Corpuscles. |8.4 |10.4/13.6)/32.4 1.50 60 |26.4| 642] 754 Blood. 6.35/12.5/15.1/34.0 4.07 | Plasma. 5.4 13.9) 6.9/26.2 Corpuscles. |8.0 |10.0/30.0)/48.0 478) 543 4.34 34 |13.7 gen has risen from 4.8 to 7.8 mg., an increase of over 60 per cent. So far as the urea production is concerned it is not possible to draw definite conclusions on the basis of the small change found 408 Amino-Acids with Reference to Urea here—and often encountered. It must be remembered that the urea concentration in the blood is a resultant, partly of what the kidneys are doing and partly of the forces at work in the muscles. Folin and Denis always found the urea content of the muscles higher than in the blood, and no one has ever found less in the muscles than in the blood though Gad-Andresen’ has recently found the same values for plasma and-for muscles. For the urea production more definite results are, of course, obtainable from the urine. It has been suggested (Folin and Denis’) that theore- tically one might expect to obtain a more rapid production of urea from a given nitrogen absorption when the nitrogen is ob- tained from a single amino-acid, than when it is obtained from a mixture of many. This suggestion is based on the hypothesis that the tissues probably tend to maintain amino-acid mixtures of more or less constant composition. It is from this point of view at least suggestive that the nitrogen elimination from the ingested glycocoll has risen from 474 to 641 mg. per hour, 13 hours after the ingestion.. At 1.50 p.m., not quite 3 hours after the intake of the glycocoll, the maximum nitrogen elimination, 754 mg. per hour, is reached. The hourly amino-acid excretion has risen from 5.8 to 26.4 mg. in about 3 hours, and it is still at a rather high level, 13.7 mg. per hour, at the end of 4 to 5 hours. Significance of the Normal Excretion of Amino-Acids. One of the questions which we have attempted to elucidate in this investigation is that of the origin and significance of the amino-acid nitrogen found in all normal human urines. It is generally taken for granted that the excretion of amino-acids represents loss of material which the organism could use to ad- vantage, if the loss could be prevented. From a physiological standpoint it seems rather remarkable that the animal organism should be able to retain every trace of absorbed fat or glucose, yet is constantly losing amino-acids. It occurred to us that the amino-acids of normal urine, like the carbohydrates of the same, might for the most part, at least represent denatured or more or 7 Gad-Andresen, K. L., Biochem. Z., 1921, exvi, 266. . 8 Folin, O., and Denis, W., J. Biol. Chem., 1912, xii, 141. O. Folin and H. Berglund 409 less foreign nitrogenous materials, rather than the ordinary amino- acids which we know as the main constituents of food proteins. The increased amino-acid excretion obtained from the ingestion of 25 gm. of glycocoll (Table III) shows, however, that the amino- acids found in urine do indeed represent losses of ordinary usable amino-acids. To get further light on this question we have studied the effects of casein. We chose casein because it is a protein in which there should be no unusable nitrogenous materials. Three of our experiments are worth recording. The first one, Table IV, is interesting not only because it illustrates the transient character which the amino-acid excretion may sometimes exhibit, but also because it proves, as clearly as the gelatin experiment, that urea formation does not accompany the passage of amino-acids through the liver. : . The subject, Mr. G. Ph. R-s., received 12 gm. of nitrogen in the form of predigested casein, as well as 4 gm. of creatine. The creatine was included for a separate study and is mentioned here only to make the record accurate. Nearly 3 hours after the ingestion of the casein the subject received 100 gm. of pure glu- cose; this was also for a different purpose and need not be further referred to here. Mention must also be made of the fact that this subject, a student, was accustomed to eat heavily of meat products. His average 24 hour nitrogen was over 20 gm. and he was. later instructed to reduce his consumption of protein. The high level of his nitrogen intake is also shown by the high fasting values for his blood urea and non-protein nitrogen. In 45 minutes the amino-acid nitrogen of the plasma rose from 5.3 to 9.2 mg.—an increase of 73 per cent, and is accompanied by practically no increase in the urea nitrogen of the plasma. The urea elimination, as it happens, fell from 806 mg. of nitrogen to 592 mg. per hour during the same period of maximum amino-acid absorption. At the end of 2% hours the deamination is extremely active as shown by the 27 mg. of urea nitrogen in the plasma and a urea nitrogen elimination of 1,140 mg. per hour. The elimination of amino-acid nitrogen rose from the high fasting value of 9.6 mg. per hour to 19.5 mg. The next two experiments (Tables V and VI) should be con- sidered together, because the two subjects received substantially 410 Amino-Acids with Reference to Urea TABLE IV. Subject Mr. G. Ph. R-s. Age 23 years. Weight 90 kilos. Time. Blood per 100 ce. Urine per hour. Remarks, iz a Z °° 2 a 5 F > ris 11 3 ay = , dated Gsd a oh, a ete ead cee eines fee 213 |e | oe | Soe 21S |e eer | = es a.m mg.| mg.| mg.| mg.| cc mg mg. mg Blood. 6 .8|20.0)16.2/43.0 11.00 | Plasma. 5 .3/22 .0| 8.2/35.5) 72 | 9.6} 806} 922 Corpuscles. | 7.4)17.2/28.8/53.4 11.05 Took 300 ce. predigested casein (12 gm. N) and 4 gm. crys- talline cre- atine. Blood. 9 .6/21.0)18.7|49.3 11.48 | Plasma. 9 2/23 .0|10.0/42.2 Corpuscles. |10.1/18.2/30.7|59.0 p.m. 12.04 51 | 8.0| 592) 784 Blood. 9.35/23 .0)17 .6|50.0 1.15 | Plasma. 9.0 |27 .0|12.3)/48.3 Corpuscles. |9.8 |17.5/25.1/52.4 1.30 88 |19.5|1,140)1,445} Took 100 gm. glucose at 1.45 p.m. Blood. 6.5 |25.0)15.0/46.5 3.05 | Plasma. 5.6 |25.0/10.0/40.6 Corpuscles. |7.8 |25.0/22.2)55.0 3.30 80 |13.1/1,115]1,379 O. Folin and H. Berglund 411 TABLE I1V—Concluded. Time. Blood per 100 ce. Urine per hour. Zi A April, A g z Remarks. 1921. = io = ; ; pen ee he | 8 ite Mee Pee a jeer | a | eal flejeljelel|s|s|é p.m. mg.| mg.| mg. mg.| cc. | m mg. mg Blood. 6.2 |22.0]14.8 143.0 4.45 | Plasma. 4.6 |25.6) 9.2 |39.4 Corpuscles. |8.3 |17.2/22.2 |47.7 5.00 55 | 8.8] 846) 942 Blood. 5.9 |23.5/11.6 |41.0 ‘ 5.35 | Plasma. 4.65/25 .0| 9.75|39.4 Corpuscles. |7.6 |/21.4/14.0 |43.0 6.00 60 | 8.9 976)1,060 the same amount of casein (60 gm.), only in one case the casein was given in predigested form, and in the other it was taken in the form of a suspension. The taking of casein in the latter form proved unexpectedly difficult and, for the benefit of others who may wish to take or administer casein in a similar manner, we may say that it is almost hopeless to try to take much of it, as long as there remains dry casein dust floating on top of the suspension. The two sets of figures recorded in Tables V and VI are not nearly so different as might have been expected. The predigested casein was, however, taken together with 120 gm. of glucose and the glucose may have. modified, in a measure both the speed of absorption and the subsequent distribution of the amino-acids. From the predigested casein the amino-acid nitrogen rose in the plasma from 4.9 to 7.4 mg. The corresponding figures from the experiment with casein suspension are 5.05 and 7.6 mg. The amino-acid excretions rose from 4.9 to 9.3 mg. per hour in the former case and from 6.3 to 8.2 mg. in the latter. The amino- acid excretion is evidently independent of diuresis. The nitrogen 412 Amino-Acids with Reference to Urea TABLE V. Subject J. M. F-r. Age 22 years. Weight 72 kilos. Time. Blood per 100 cc. Urine per hour. Remarks. Z a Apr. 12, a = Z. Fasting. Urinated 1921. 3 g 3 } at 9.03 a.m. @ | Be) eel Saab ae a lee | ae | Si eae | ta ae a{/plealels|<|Pp] eo a.m. mg.| mg.| mg.| mg.| cc. | mg.| mg.| mg Blood. 6.6/12.5/12 .9/32.0 10.52 | Plasma. 4.9/12.5) 9.2/26.6) 36 | 4.9] 371) 469 Corpuscles. | 9.0)12.5/18.3/39.8 10.55 Tookaeo0sscc. predigested casein (10 gm. N), 120 gm. glucose, and 500 ce. water. Blood. 8.1)13.9)16 4/38 .4 11.32 | Plasma. 7.4)14.1) 7.3/28.8 Corpuscles. | 9.1|13.7/29.4|52.2 11.53 44 | 5.0) 464! 576 p.m. Blood. 7 4/13 .2)17 .8/38.4 12.55 | Plasma. 6 .3/14.3/10 .4/31.0 Corpuscles. | 9.1)11.5/29.4/50.0 RS Be 43 | 6.9} 530) 662 Blood. 5.9)14.7/12.7/33.3 1.53 | Plasma. 5.9)14.5) 9.6)30.0 Corpuscles. | 5.9)15.1)17.5/38.5 2.52 41 | 9.3) 564) 734 Blood. 6.0)14.1)13.5)33.6 3.47 | Plasma. 4614.4) 9.3/28.3 Corpuscles. | 8.4 ipa apes li ot et O. Folin and H. Berglund 413 TABLE V—Concluded. Time. Blood per 100 ce. Urine per hour. Sat dui lade 2 | A> Zz 3 Zz Remarks. ATS. =a 2 = ete (seiil a3 = ac AN iad) RG ee ules qafeleleale|4|/5)e p.m. mg.| mg.| mg fant cc mg.| mg.| me 4.25 46 | 8.3) 566) 730 Blood. 6. 2/15 .0) 014. 3/35. 5 5.40 | Plasma. 4. lie: 711. ae: 9 Corpuscles. | 9.9/13.9/19. 4/43.2 5.48 | 39 | 6.3) 519) 610 excretion would seem at first glance to be more rapid from the undissolved than from the predigested casein, but the extraordinary diuresis ‘obtained in the undissolved casein experiment may have influenced the nitrogen excretion to a material extent. From all the experiments given, but especially those with casein and glycocoll, it is quite clear that losses of ordinary usable amino-acids is a normal occurrence, and that these losses are increased during and immediately following increases in the amino-acid content of the blood. There is nothing corresponding to the sharp glucose threshold at which the eliminations begin. In fact the retention and excretion phenomena associated with the amino-acids are more like those which we have found for galactose,’ which has no threshold, than for anything else so far described. In the absence of a threshold or other sharp mecha- nism for the retention of the absorbed amino-acids one would. expect to find a continuous excretion, a continuous loss, with the urine. And that is exactly what we do find. Under such. con- ditions there should also be an increase in these losses, whenever there is an influx of amino-acids from the digestive tract sufficiently ‘rapid to materially raise the amino-acid content of the blood. This happens after every fairly protein-rich meal. 9 Folin, O., and Berglund, H., J. Biol. Chem., 1922, li, 213. \ 414 Amino-Acids with Reference to Urea TABLE VI. Subject H. B-d. Age 34 years. Weight 90 kilos. Time. Blood per 100 cc. Urine per hour. Remarks. Zz, = Zz z r=) ie Pia. : ae. = Fj a a eae Br | ee |e ema eee Boy ge | ee ele a | eee che ba et et) Sit p.m. mg.| mg.| mg.| mg.| ce mg.| mg.| mg. Blood. 6.25|12.4)17 .3/36.0 12.25 | Plasma. 5.05|13.0| 8.6/26.6) 23) 6.3) 360) 445 Corpuscles. |8.1 |11.6/31.3/51.0 12.30 Took 60 gm. to Hammarsten’s 115 casein and 600 cc. water. 1e52 190) 5.7| 606) 670 Blood. 7.4 |14.2/12.4/34.0 2.03 | Plasma. 7.25/13.6| 8.6/29.5 Corpuscles. |7.65|15.0)18.3/41.0 2.18 462] 7.1] 650} 712 Blood. 8 .15]15.2/12.6|/36.0 3.00 | Plasma. 7.6 |15.6| 7.2/30.4 Corpuscles. |9.1 |14.5|21.4/45.0 3.46 57) 8.2) 635} 650 Blood. 7.05/15. 2/13 .7/36.0 4.55 | Plasma. 6.15/16.6) 8.2/31.0 Corpuscles. |8.5 |13.0)22.5)44.0 5.43” 311 5 5\ Olam Ambard and his associates have endeavored to expand the con- cept of a definite constant relationship between the concentra- tion in the blood and the excretion, which was first propounded for urea, into a general law governing the excretion of all products. O. Folin and H. Berglund 415 Since the amino-acid excretion is not controlled by anything corresponding to the glucose threshold this law of Ambard’s ought to apply to amino-acids. A candid examination of our tables will show, however, that nothing suggesting the existence of a constant, in the sense of Ambard, is to be found. Increased accumulations in the blood will be followed by increased excre- tions, but it is perfectly clear that it is not only the level in the blood plasma, but also the extent to which the tissues, including the kidneys, have been charged with amino-acids, which deter- mine the rate of the excretion. TABLE VII. Normal Minimum, Maximum, and Average Blood Content of Non-Protein Nitrogenous Products (From 12 Young Men). Whole Blood Corpuscles per 100 cc. Plasma per 100 ce. per 100 cc. ae ae ire ag a4 ~ > — a |! jaa Cam ee ‘ eS hs = I ee & | 3 Zz 3/214 See ee oe = =} Q =I 2, Eo} f=] en 5 = s AS) ‘g zg mt ¥ ay See ese eee || 8 pS). | ays a eae a 3 3 3 om 8 Eo} oS Spal 3 so 3 Boll os ad |o ) £ | .& a | 6 5 a | 6 ejatlelae mg.| mg.| mg.| mg. | ce. | mg.| mg.| mg. 10.3)13.9|30.0 Remarks. Fasting. Urinated at 10.12 a.m. 5.8 4.3/10.6| 9.3/24.2| 50'5.7/387|486) Took 200 gm. 7.8 .8| 9.9/20.0/37.7 111/2.9|327|356 35/3 .8|362/480 4.9)10.0/11.1/26.0 3.6/10.4) 5.2/19.2 6.8) 9.4/19.6/35.8 29|4 .9|376/485 5.2] 8.5)12.7/26.4 3.7| 9.2) 5.3/18.2 7.3) 7.5/23.0/37.8 27|4.3|276/384 5.0} 8.8)13.0/26.8 3.6) 9.6) 5.2:18.4 6.9] 7.7/23.8/38.4 glucose and 830 ec. water. at noon. 76|3.7|382/453] Dinner at 7.00 111)/6.2|348/462 416 p.m. O. Folin and H. Berglund 417 taken, together with similar figures for urea, undetermined nitro- gen, and total non-protein nitrogen, are given in Table VII. This table also gives corresponding minimum values obtained under the influence of carbohydrate intake, not accompanied by any intake of protein (for 1 day). In Table VIII are given data showing in detail the effect of carbohydrate intake on the composition of blood and urine. 200 gm. of pure glucose in 830 cc. of water were taken at noon, immediately after taking a sample of the subject’s blood. As was to be expected from the absence of nitrogen intake and from the protein-sparing action of the sugar we meet here with small, but definite reductions in the different nitrogenous constituents in the blood. The amino-acid nitrogen sinks from 4.3 mg. per 100 cc. of plasma to 3.6 mg.; the urea nitrogen from 10.6 to 9.6 mg.; the undetermined nitrogen from 9.3 to 5.2 mg.; and the total non-protein nitrogen from 24.2 to 18.4 mg. Distribution of the Normal Non-Protein Nitrogen. In an earlier part of this paper we alluded to Falta’s assertion that the blood corpuscles are free from non-protein nitrogenous constituents. We believe that the many simultaneous analyses of plasma and whole blood reported in this paper, and they repre- sent only a small fraction of the total number made, prove beyond possibility of doubt that for the nitrogenous constituents human blood corpuscles are as permeable as we have previously shown them to be permeable for glucose. So far as urea, the chief single nitrogenous compound, is con- cerned, our data point not only to permeability, but also to the production of urea in the corpuscles. It is not worth while to try to prove definitely this point, because there are so many variable factors involved, the effects of which cannot be accurately estimated, at least by the kind of experiments which we have made. It is according to our experience by no means uncommon to find, particularly after heavy protein intakes, samples of blood in which the urea concentration is higher in the corpuscles than in the plasma, even without making any allowance for the fact that the corpuscles contain much less water. This cer- tainly suggests, but, of course, does not prove that the higher concentration is due to local production. 418 Amino-Acids with Reference to Urea In all our tables will be found figures for the undetermined, or rest, nitrogen. These figures are a little too high since they in- clude the nitrogen of the uric acid, the creatinine, and the “crea- tine.” It will be seen that this unknown fraction is from two to five times as abundant in the corpuscles as in the plasma. . The nature, origin, and significance of the nitrogenous materials repre- sented by this nitrogen must necessarily be left for future investi- gations to unravel. We venture, however, to advance one surmise. It is a well known fact that different blood protein precipitants give filtrates which, with the same blood, yield different amounts of non-protein nitrogen. These differences must necessarily be due for the most part to how much of the unknown materials are taken by the precipitant. If these unknowns are of the nature of proteins it may also be expected that the amino-acid nitrogen found will be influenced by the presence of these nitrogenous sub- stances and the materially higher amino-acid nitrogen found in the corpuscles than in the plasma may in part be due to the rich- ness of the corpuscles in the undetermined nitrogen. The extraordinarily high values (20 mg.) obtained for the amino-acid nitrogen of the blood of birds and the fact that such blood is also enormously rich in undetermined nitrogen stand in need of some explanation. It is on the basis of these facts that. we venture to suggest that the undetermined nitrogenous materials are in part made up of histones. It was from the corpuscles of birds that Kossel obtained histones and it is difficult to see how one can fail to obtain varying quantities of histones when working with the blood of birds. It is precisely with such bloods that the different protein precipitants give the greatest differences in the amounts of non-protein nitrogen obtained. The tungstic acid method for example gives 3 to 6 mg. less than the trichloroacetic acid or the m-phosphoric acid method. Histones must be more abundant in bird’s blood than in the blood of man or other mammals, or they would doubtless long since have been found in such bloods also; but it seems to us extremely probable that the undetermined nitrogen of human blood and other bloods, for which the corpuscles are largely re- sponsible, is in very large part due to the presence of histones similar to those which have been obtained from the corpuscles’ of birds. NOTE ON THE NECESSITY OF CHECKING UP THE QUALITY OF SODIUM TUNGSTATE USED IN THE SYSTEM OF BLOOD ANALYSIS. By OTTO FOLIN. (From the Biochemical Laboratory, Harvard Medical School, Boston.) (Received for publication, January 30, 1922.) Because of the many inguiries received concerning different brands of sodium tungstate for use in precipitating the blood proteins it seems suitable to publish this brief note on the subject. The statement made in the first paper! of ‘‘A system of blood analysis” that we obtained our sodium tungstate from the Primos Chemical Company was not intended to convey the impression that this company had been requested or had undertaken to supply the chemical according to our specifications. We simply found that their product was satisfactory and mentioned the fact. Being manufacturers, we thought that their brand would continue to be available. The firm has since ceased to make the product. At that time we supposed that the only difficulty that might be expected in connection with sodium tungstate would be the presence of variable large excesses of sodium carbonate, and directions were given for titrating the alkali with 0.1 N acid, using phenolphthalein as indicator. We suggested that no tungstate should be used which contained more than 0.5 per cent of sodium carbonate. The amount of 0.1 Nn acid required for the titration of 10 cc. of 10 per cent tungstate solution should not exceed 0.4 ec. As a matter of experience, up to date, the sodium tungstates which show an alkaline reaction have come within this figure. I have, however, since encountered sodium tungstates which show no alkaline reaction to phenolphthalein, though they are alkaline to litmus paper. These tungstates may be perfectly pure in the sense that they contain nothing but tungstates, yet 1Folin, O., and Wu, H., J. Biol. Chem., 1919, xxxviii, 86. 419 / 420 . Sodium Tungstate in Blood Analysis they will not yield a good uric acid reagent and they will give erroneous figures for blood analyses. I deem it particularly important to call attention to this variety of sodium tungstate because almost any one might fail to note that they are unsuitable, since they will give perfectly clear blood filtrates. The titratable acidity of the filtrates will be excessive, however; the total nitrogen will be somewhat too low, and the uric acid and creatinine will be more or less completely removed from the filtrate. These acid tungstates are usually so slightly soluble that one cannot readily make a 10 per cent solution in cold water, whereas the true sodium tungstate is soluble in 2 parts of water. The acid tungstates need not be discarded. They can be made perfectly serviceable in the following manner: Prepare 100 cc. of a 10 per cent solution in water, using heat if necessary. Cool. Titrate 10 cc. of the solution with n sodium hydroxide to a faint, but permanent, pink reaction, using phenol- phthalein as indicator. The pink color should persist for at least 3 minutes after the last addition of alkali. The reason for this requirement is that the complex acid tungstates are only gradually and slowly decomposed by the slight excess of alkali added during. the titration. . In making subsequent 10 per cent solutions of the tungstate add the amount of alkali indicated by the titration. The essential point in this treatment is, of course, that the complex acid tungstates are easily converted into the simple salt, Naz:W0.'2H20, by the addition of the required amount of alkali. With the 10 per cent sodium tungstate so prepared and correct, 0.66 N sulfuric acid, we never fail to obtain blood filtrates suitable for. subsequent analysis unless excessive amounts of potassium oxalate have been added to prevent clotting, and it is only because of such excess of oxalate that the addition of a little more sulfuric acid is occasionally required. The preliminary conversion of the complex tungstates into the simple one by the addition of a suitable amount of alkali is also important for the preparation of the uric acid reagent and the phenol reagent. The more complex tungstic acids are seemingly not decomposed in acid solution; they, therefore, fail to combine with the phosphoric acid and the reagent will be weak. COLORIMETRIC METHODS FOR THE SEPARATE DETER- MINATION OF TYROSINE, TRYPTOPHANE, AND CYSTINE IN PROTEINS. ' By OTTO FOLIN anp JOSEPH M. LOONEY. (From the Biochemical Laboratory, Harvard Medical School, Boston.) (Received for publication, February 8, 1922.) CONTENTS. ag alc BA UI a 421 Separation and estimation of tyrosine and tryptophane in mixtures of Ee RS ee Og ee oo en ee 423 Colorimetric determination of cystine................-..----------2-: 427 Applications of the colorimetric methods for tyrosine, tryptophane, ene Ge POLE Wateridis..... 2... 2. kee ee eee 429 Amounts of tyrosine, tryptophane, and cystine in different proteins... 431 INTRODUCTION. The colorimetric method for the estimation of tyrosine in proteins proposed by Folin and Denis! in 1912 has been shown to yield results which in many cases must be too high, because tryptophane and its decomposition products give the same blue reaction with the phenol reagent as does tyrosine. The method has been adversely criticized by a number of different investiga- tors. Much of this criticism is manifestly based on inexperience in the field of colorimetry; but the error due to tryptophane, an error pointed out by Abderhalden,? tentatively admitted by Folin and Denis,’ and later verified by Gortner and Holm‘ as well as by Thomas? has left us with a wide gap of uncertainty as to the true tyrosine content of many protein materials. 1 Folin, O., and Denis, W., J. Biol. Chem., 1912, xii, 239. 2 Abderhalden, E., and Fuchs, D., Z. physiol. Chem., 1913, lxxxiii, 468. ’ Folin, O., and Denis, W., J. Biol. Chem., 1913, xiv, 457. 4 Gortner, R. A., and Holm, G. E., J. Am. Chem. Soc., 1920, lxii, 78. 5 Thomas, P., Bull. Soc. chim. biol., 1921, iii, 198. 421 422 Tyrosine, Tryptophane, and Cystine We shall not here discuss the extraordinary criticisms which some of the authors cited have expended upon the method. Gortner and Holm have tabulated nine different reasons why tyrosine cannot be estimated by means of the phenol reagent,® and, incidentally, have proved to their own satisfaction that the colorimetric sugar method of Folin and Wu must be worthless. It is, of course, possible that definite sources of error in the colorimetric tyrosine method other than that due to tryptophane may be discovered soon or later. But a colorimetric analytical method yielding maximum figures must be considered very useful, at least as a check on the extremely uncertain, usually minimum, values obtained by direct isolation processes; and when a source of error in such a colorimetric method is found the discovery would seem to call for revision rather than for unqualified destruction. The investigation here described was begun for the purpose of removing such errors as were found to be due to tryptophane. As the work progressed it soon became evident that the color due to tryptophane and its decomposition derivatives is quantitative for the amount of tryptophane present. The research was accordingly broadened so as to include the separate determina- tion of both tyrosine and tryptophane. Several years ago the observation was made in this laboratory (by T. Zucker) that cystine, though by itself substantially un- affected by either the phenol or the uric acid reagent, does give an intense blue color with the uric acid reagent—in the presence of sodium sulfite. The reaction is evidently due to the fact discovered by Heffter? that cystine is reduced to cysteine by the sulfite. The reaction with the uric acid reagent has not pre- viously been developed into a quantitative method, but we have now used it for that purpose and believe that the process is worth incorporating in this paper as a quantitative colorimetric method for the determination of cystine. 5 Boil 15 gm. of molybdic oxide and 10 gm. of sodium hydroxide in 200 cc. of water until the solution no longer smells of ammonia. Then add 100 gm. of sodium tungstate, 50 cc. of 85 per cent phosphoric acid, 100 ce. of concentrated hydrochloric acid, and enough water to bring the volume up to 800 cc. Boil the mixture, for 10 hours, in a 1,500 ce. flask, fitted with a reflux condenser, and then remove the condenser and add a few drops of bromine to decolorize the solution. Boil off the excess of bromine, cool and filter the solution, and make up to a liter. 7 Hefiter, A., Chem. Zentr., 1907, ii, 822. O. Folin’and J. M. Looney 423 In this paper are accordingly described colorimetric methods for the determination of (a) tyrosine, (b) tryptophane, and (c) cystine—in protein materials. Separation and Estimation of Tyrosine and Tryptophane in Mixtures of Both. As we happen to have on hand at the present time a consider- able number of amino-acids in pure condition it was thought hest to apply the phenol reaction to them although there was no reason to believe that any except tyrosine and tryptophane would give the reaction. No color was obtained with 10 mg. of each of the following amino-acids: arginine, lysine, aspartic acid, alanine, phenylalanine, histidine, glutamic acid, proline, leucine, valine, glycine, asparagine, cystine, isoleucine, isoserine. Of the seven- teen amino-acids tested only tyrosine and tryptophane gave a color. In this connection we would call attention to one erroneous statement found in the first paper on the uric acid and phenol reagents by Folin and Denis, an error emphasized by Gortner and Holm. Indole and its derivatives were said to give no reaction with the phenol reagent whereas in point of fact it does give an unmistakable though comparatively weak reaction. Indole gives no reaction with the uric acid reagent. The fact that the phenol reagent does give a reaction with indole has scarcely any bearing on the determination of tyrosine in proteins, though it is doubtless of some importance in other connections. In order to determine the quantitative relationship in the color produced by tyrosine and by tryptophane the following experiments were made. Experiment 1.—From stock solutions of each containing 1 mg. per cc., 2 mg. of the tryptophane and 1 mg. of tyrosine were transferred to sepa- rate 100 cc. volumetric flasks. 20 cc. of saturated sodium carbonate were added and enough water to give a volume of 50 cc. 2 cc. of the phenol reagent were added to each and after 30 minutes the flasks were filled to the mark and the color comparison was made. The tyrosine was used as a standard and was set at20mm. The tryptophane solution gave a reading 17.2 corresponding to 1.16 mg. of tyrosine. According to this result tryp- tophane gives 58 per cent of tle color obtained from an equal weight of tyrosine. 424 Tyrosine, Tryptophane; and Cystine Experiment 2.—100 mg. of tryptophane were boiled for 24 hours with 25 ec. of 20 per cent hydrochloric acid in a small Kjeldahl flask connected with a Hopkins condenser. The solution was cooled, diluted to 100 ec., and 2 cc., equivalent to 2 mg. of the tryptophane, were assayed as before, using 1 mg. of tyrosine as the standard. The color value was the same as with the original tryptophane. Boiling with acid does, therefore, not remove the disturbing effects of tryptophane in connection with tyrosine determinations. From the preceding two experiments as well as from unre- corded repetitions of the same it is clear that in order to deter- mine tyrosine in protein digestion mixtures containing trypto- phane the latter must first be removed. The quantitative, or substantially quantitative, removal of tryptophane from such digestion mixtures can be accomplished by means of the excellent mercuric sulfate reagent of Hopkins and Cole. But the usefulness of the process in connection with such small scale work as is involved in colorimetric analysis could be determined only by a systematic investigation. In addition to the question of how completely minute amounts of tyrosine and tryptophane could be separated it was recognized that the intro- duction-of mercury carried with it the same sort of complications which were encountered from the silver salts in the analogous work on uric acid. It was found, however, that sodium cyanide is Just as useful for the removal of mercury ions as for the removal of silver, and the necessary check work was thus much simplified. Without going into a detailed description of the 26 different preliminary experiments which were made to determine the conditions for the quantitative separation of 1 mg. of tyrosine from 1 mg. of tryptophane the results may be summarized as follows: (a) If the acidity of the final solution represents less than 3.5 per cent of sulfuric acid there is danger of loss of tyrosine, due to its precipitation by (2 per cent) mercuric sulfate. (b) The tryptophane is quantitatively precipitated within 2 hours by 2 per cent mercuric sulfate, when the acidity lies between 3.5 and 7.5 per cent of sulfuric acid. After having ascertained the conditions necessary for the separation and colorimetric estimation of tyrosine and trypto- phane a series of analyses of various mixtures of the two was made in the following manner: —_— O. Folin and J. M. Looney 425 Standard solutions of tyrosine and tryptophane in 5 per cent sulfuric acid were prepared—each solution containing 1 mg. of amino-acid per cc. By means of accurate 5 cc. burettes graduated in 0.02 of a ce. (“sugar burettes’’) definite amounts of the amino-acid solutions were measured into 15 cc. centrifuge tubes which had been graduated at a volume of 10 ce. 2 cc. of the mercuric sulfate solution (containing 10 per cent of mercuric sulfate and 5 per cent of sulfuric acid) were added and the mixture was at once diluted with 5 per cent sulfuric acid to the 10 cc. mark. A rubber stopper was inserted and the solution was shaken vigorously to insure thorough mixing. The tubes were allowed to stand for 2 hours and were then centrifuged. The clear supernatant liquid containing the tyrosine was poured into a clean dry test-tube and set aside for the tyrosine determination. The centrifuge tube containing the mercuric tryptophane sediment was then filled to the 10 cc. mark with 5 per cent sulfuric acid, its own rubber stopper was again inserted and the mixture was shaken. After removing the stopper and centrifuging, the supernatant liquid was carefully poured out and drained for half a minute. For the subsequent colorimetric determinations of tyrosine and of tryptophane the process is as follows: Tyrosine.—Transfer 5 cc. (one-half) of the tyrosine-containing liquid to a 100 ec. volumetric flask and into another similar flask introduce 1 cc. of the standard sulfuric acid-tyrosine solution containing 1 mg. of tyrosine. To the latter add also 1 cc. of the acid mercuric-sulfate solution and 3 ce. of 5 per cent sulfuric acid. Then add to each flask about 30 cc. of water, 20 cc. of saturated sodium carbonate solution, and 4 cc. of 5 per cent sodium cyanide solution, in the order named. It is best to shake for a moment after the addition of the carbonate, before adding the cyanide. The yellow mercuric carbonate thrown down by the sodium carbonate promptly dissolves on adding the cyanide and shaking. Add 2 ce. of the phenol reagent, mix, let stand for 10, or better for 30, minutes, and make the color comparison in the usual manner, setting the standard at 20 mm. 20 times 2 divided by the reading of the unknown gives the tyrosine found, in milligrams. 426 Tyrosine, Tryptophane, and Cystine Tryptophane.—The determination of the tryptophane in the mercury-tryptophane precipitate is not quite so simple as is the determination of the tyrosine in the supernatant solution. The precipitate dissolves readily and completely in an excess of sodium cyanide, but only under certain rather definite conditions. Moreover, the color obtained from 1 mg. of tryptophane in the form of the mereury compound is considerably deeper than the color obtained from a milligram of tryptophane not previously combined with mercury—even though the same amount of mercuric sulfate and sodium cyanide have been added before the addition of the phenol reagent. It is accordingly necessary to subject the tryptophane standard to the same treatment as the unknown; namely, to precipitate it with mercuric sulfate. This precipitation should, therefore, be started at the same time as the precipitation is made in the unknown or before. Since it makes no difference whether the precipitation is allowed to con- tinue for 2 hours or for several days, a series of precipitated standards can, of course, be kept on hand in well stopped centrifuge tubes. To the unknown mercury precipitate and to a similarly pre- cipitated and centrifuged standard, containing 1 mg. of trypto- phane, add 10 cc. of water, insert the rubber stopper, and shake so as to secure a uniform suspension. Without any unnecessary delay, that is within 2 or 3 minutes, add 4 ce. of 5 per cent sodium cyanide to each, insert the rubber stoppers, and mix. Complete solution occurs at once; whereas a fine, more or less gelatinous residue, containing tryptophane is left undissolved, if the pre- liminary shaking of the sediment is omitted. Rinse the standard and the unknown into 100 cc. volumetric flasks, keeping the volumes approximately equal (50 cc.). Add first 20 ce. of sodium carbonate and finally, with shaking, 2 ce. of the phenol reagent. Let stand for not less than 10, preferably for 30, minutes, dilute to volume, and make the color comparison. When the standard is set at 20 mm., 20 divided by the reading of the unknown in mm., gives, in milligrams, the amount of tryptophane present. From the analytical figures recorded in Table I it will be seen that the process described above for the separate determination of tyrosine and tryptophane in various mixtures of the two yields substantially correct values. a |. O. Folin and J. M. Looney 427 TABLE I. Illustrating Estimations of Tyrosine and Tryptophane in Mixtures. Amount taken. | Amount recovered. Error. Experiment — Tyrosine. ees Tyrosine. ed “ i‘ Tyrosine. ae . mg. mg. mg. mg. per cent per cent 1 1.80 5.00 1.82 +1.1 2 1.60 0.00 1.60 0.0 8 2.40 3.00 2.38 —0.9 A 2.40 0.00 2.38 —0.9 5 1.80 0.85 1.80 0.83 0.0 —2.3 6 2.20 1.20 2.18 1.16 —0.9 —3.3 7 2.10 0.90 2.08 0.89 —1.0 —1.0 8 2.00 1.00 1.94 0.98 —3.0 —2.0 9 1.90 0.95 1.90 0.93 0.0 —2.0 10 1.50 1.50 1.54 1.43 +2.6 —4.6 11 1.80 1.45 1.88 1.39 +4.4 —4.1 12 2.20 1.30 2.24 1.27 +1.8 —2.3 13 2.60 0.85 2.54 0.88 —2.3° | +3.5 14 3.00 0.90 3.08 0.92 +3.6 +2.2 15 1.50 1.45 1.54 1.41 +2.6 —2.7 16 1.80 1.35 1.84 1.30 +2.2 —3.7 17 2.20 0.75 2.16 0.79 —1.8 +5.2 18 2.60 0.85 2.56 0.86 —1.5 Seibel! 19 3.00 0.90 312 0.90 +4.0 0.0 20 0.00 1.30 0.00 1.29 —0.8 7A 1.80 0.95 0.97 +2.0 22 0.00 0.80 0.81 allo 23 2.00 1.15 1.16 +0.9 24 0.00 1.20 1.18 +1.7 Colorimetric Determination of Cystine. Since the colorimetric determination of cystine is based on the use of the uric acid reagent of Folin and Denis and since neither tryptophane, nor tyrosine, nor any other known amino-acid except cystine gives a reaction with this reagent the process for the determination of cystine in amino-acid mixtures is so simple as to require very little preliminary discussion. As previously mentioned, Heffter found that cystine is reduced vo cysteine by (10 per cent) sodium sulfite. Heffter made some use of this reaction in tests for cystine in tissue extracts by means THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI, NO. 2 428 Tyrosine, Tryptophane, and Cystine of nitroprusside; but, so far as we know, no attempt has been made to build up a quantitative method on the basis of these reactions; nor has the reducing effects of sulfites on cystine been ~ worked out on a quantitative basis. The figures recorded below (Table II) show that the quantita- tive or maximum conversion of cystine into cysteine is obtained only in the presence of a very large excess of sodium sulfite. The experiments were made as follows: To 3 mg. of cystine in 100 ce. volumetric flasks were added 20 ce. of saturated sodium carbonate and varying amounts of freshly prepared 20 per cent solution of sodium sulfite. After 5 minutes, 2 cc. of the uric acid reagent were added; the solutions were diluted to volume and the TABLE II. Showing the Amount of Sodium Sulfite Required for the Transformation of Cystine to Cysteine. Amount of cystine found. 20 per cent NazSOs3 solution. Colorimeter reading. Theoretical value = 3 me. ce. mm. 1 39.0 1.53 2, Soe 1.86 3 29.3 2.05 4 28.5 2.10 5 25.8 2.301 10 (Standard) 20.0 3.00 20 20.0 3.00 color comparison was made, taking the one to which 10 ce. of sulfite had been added as the standard. The amount of color obtained from different amounts of cystine is strictly proportionate to the quantities taken within the limits usually governing such color comparisons; that is, when neither the standard exceeds the unknown, nor the unknown the standard, by more than 50 per cent. In point of fact the proportionality is in this case dependable over still wider limits, but it is safer to keep within those mentioned. The large amount of sulfite used for the reduction of the cystine has the same stabilizing effect on the color obtained as when sulfite is used with uric acid. This stabilizing effect is, however, not nearly so great as that obtained with sodium cyanide (and O. Folin and J. M. Looney 429 uric acid). It is therefore essential that the color should be developed substantially simultaneously in the standard and the unknown. This rule is so frequently necessary in colorimetry that it may well be called one of its fundamental principles. If properly followed it makes no difference whether the color com- parison, in cystine determinations, is made at the end of 10 minutes or after 1 hour, or even longer. Sodium cyanide can scarcely be said to serve any useful purpose in connection with the colorimetric method for cystine in protein materials. If added after the blue color has once been developed it reduces somewhat the rate of the fading, but that fading is any way so slow as to render the addition of the cyanide quite superfluous. Moreover, if the cyanide is added in the usual way, that is before the introduction of the uric acid reagent it almost completely inhibits the development of the color. This inhibiting effect is obtained whether the cyanide is added before or after the addition of the sulfite. Applications of the Colorimetric Methods for Tyrosine, Trypto- phane, and Cystine to Protein Materials. It would be highly desirable if tyrosine, tryptophane, and cystine could all be determined in a common hydrolysis mixture. It is doubtful, however, whether the most accurate results are obtainable under such conditions. For the cystine determina- tion it is essential that the hydrolysis should be made with acid, since the cystine is completely destroyed by boiling with alkalies. On the other hand it seems to be generally agreed that the trypto- phane survives better under alkali than under acid hydrolysis. The ‘“humin” formation in acid hydrolysis, which Gortner and Holm have shown to involve more or less loss of tryptophane, constitutes an added reason against this form of hydrolysis for tryptophane determinations, although we are not yet quite sure to what extent tryptophane really loses its reactivity with the phenol reagent by virtue of such humin formation. Informa- tion on this point will doubtless come out of the promised further investigations by Gortner and Holm. The tyrosine can be determined on the basis of either acid or alkali hydrolysis. For the determination of tyrosine and trypto- phane we have used the following process: 430 Tyrosine, Tryptophane, and Cystine The purified protein is dried for 48 hours in a vacuum desiccator over sulfuric acid. 1 gm. of the dried material is introduced into a 300 ce. long necked Kjeldahl flask. After adding 3.5 gm. of erystallized barium hydroxide and 25 ce. of distilled water the neck of the flask is closed by a Hopkins condenser and the mixture is gently boiled over a micro burner for 40 to 48 hours. 30 cc. of 20 per cent sulfuric acid are then added and the flask is heated in boiling water for 30 to 60 minutes to drive off any hyrogen sulfide which may have formed. The cooled mixture is transferred to a 100 ce. volumetric flask, diluted to the mark, thoroughly mixed, and filtered through a dry filter into a dry flask. The amount of amino-acid-containing filtrate obtained is abundant for any reasonable number of tyrosine and tryptophane determinations, since 8 ce. of filtrate are about the maximum amount required for one pair of determinations. The filtrate has also about the right degree of acidity for the separation of tyrosine and tryptophane. Transfer from 1 to 8 cc. to a centrifuge tube graduated at 10 ce. Add 2 ec. of a solution containing 10 per cent of mercuric sulfate and 5 per cent of sulfuric acid (Hopkins and Cole reagent). Dilute with 5 per cent sulfuric acid to the 10 ec. mark. Insert a rubber stopper and give a few vigorous shakes. Let stand for 2 hours and then centrifuge. The supernatant solution con- taining the tyrosine is decanted from the sediment which contains the tryptophane and these two fractions are then assayed colori- metrically as already described for pure mixtures of tryptophane and tyrosine (page 423). The procedure described below has been found satisfactory for the determination of cystine: From 1 to 5 gm. of the dry protein and 25 ce. of 20 per cent sulfuric acid are transferred to a 300 cc. Kjeldahl flask, fitted with a Hopkins condenser. The mixture is boiled gently over a micro burner for 12 hours, after which it is cooled, diluted to 100 cc., and thoroughly mixed. From 1 to 10 ce. of the solution are transferred to a 100 cc. volumetric flask and to it are added first 20 cc. of saturated sodium carbonate solution and then 10 ce. of 20 per cent sodium sulfite solution. The mixture is well shaken and set aside while the standard cystine solutions are prepared. The standard cystine solution is made to contain 5 per cent sulfuric acid and 1 mg. of cystine perce. This solution keeps indefinitely. QO. Folin and J. M. Looney 431 Two standards are prepared containing 1 and 3 cc., respectively, of cystine solution or 1 and 3 mg. of cystine. Add to each 20 cc. of saturated sodium carbonate solution and 10 cc. of 20 per cent sulfite and let stand for 5 minutes. 3 cc. of the uric acid reagent of Folin and Denis are then added (with shaking) to each standard and to the unknown digestion mixture. The three flasks are allowed to stand for 10 minutes; the contents are then diluted to the 100 cc. mark and the color comparison between the unknown and the standard nearest it in color is made in the usual manner. There is no need for any undue hurry in the making of the color comparison, for the slow fading which takes place is exactly the same in the unknown as in the case of the standard. Amounts of Tyrosine, Tryptophane, and Cystine in Different Proteins. We have purposely omitted in this paper to enter upon any discussions or criticisms of other methods previously described and used for the determination of tyrosine, tryptophane, and cystine in protein materials. In Tables III, IV, and V giving our analytical figures are also given corresponding figures taken from the literature and representing other methods. Concerning these figures we would make the following brief comments: 1. Tyrosine-—Our colorimetric tyrosine figures are lower than the figures obtained by Folin and Denis. The difference is in the main due to the tryptophane though certain inconsistencies are evident—as in the case of edestin and gliadin. These inconsis- tencies are presumably due to different degrees of purity in the material used. The main point about the tyrosine values, obtained by the colorimetric methods, is that they remain extraordinarily high in comparison with those obtained by gravimetric methods. Since ‘‘complete” analyses of protein materials always leave a very wide margin of unknown amino-acids we may for the present be permitted to assume that a part of unknown material con- sists of undetermined tyrosine. 2. Tryptophane——In connection with the irvuhaptians values we would call attention:to the last figure in Table IV showing that when 2 per cent of pure tryptophane was added to zein 2.01 per cent was found. Zein alone gave nothing. i yrosvne Content. Acid Ba(OH):} Folin ; Gravi- < a malar 2 d Protein. hy droly y oP y ae raeerioe Investigator. per cent | per cent | per cent |per cent Casein (Hammar- BieN ins. baste: 5.86 | 5.32] 6.5 | 4.5 | Fischer, E., Z. physiol. Chem., 1901, xxxiii, le «i | Rea Ae 5.76 | 5.52 | 6.0 | 2.9 | Abderhalden, E., and Voitinovici, A., Z. physiol. Chem., 1907, lii, 349. Zein (yellow corn).| 5.6 5.52 |, .5:5,.| 3.6 | Osbemnes pies ang Clapp, 8. H., Am. J. Phystol., 1907-08, xx, 477. Hornish eee ee ee 5.28 | 6.5 | 4.6 | Fischer, E., and Dér- pinghaus, T., Z. phys- iol. Chem., 1902, Xxxvl, 462. Glutenin (wheat)..| 4.7 4.5 5.8 | 4.25 | Osborne, 17Ee 6. and Clapp, S. H., Am. J. Physiol., 1906-07, xvii, 231. Abderhalden, E., and Samuely, F., Z. phys- tol. Chem., 1905, xliv, 276. Serum globulin 6.7 Abderhalden, E., Z. (OGM ie L hah ds phystol. Chem., 1905, xliv, Lf Ovalbumin........ 4.2 5.0 | 1.77 | Osborne, T. B., Jones, D. B., and Leaven- worth, C. S., Am. J. Phystiol., 1909, xxiv, 252. Levene, P. A., and Van Slyke, D. D., Biochem. Z., 1908, xili, 440. bo SS Gliadin (wheat)...}| 3.5 3.4 3.3 bo or (Se) bo or Witte peptone.....| 5.87 | 5.86 Edestin (hemp SECC) ett Str ae a.40 | .5276)), 20). bo bo = Abderhalden, E., Z. physiol. Chem., 1902- 03, xxxvii, 499. Rubrinee eee 6.4 6.5 3.3 | Levene, P. A., and Van Slyke, D. D., Biochem. Z., 1908, xiii, 440. POU SoS g ae 2 7.36 4.9.'| Hopkins) )asgne Ae) Savory, H., J. Phys- iol., 1911, xlii, 189. Gélatinn Ee... Trace. |Trace. |Trace. 0 432 —_— es Se O. Folin and J. M. Looney 433 Protein. ad per cent Casein (Hammar- SUGIT) Pe ah Oa Ae? WhoOliscce.c0. 80. 1.45 Zein (yellow corn).. 0 jr itinik . oe 1.438 Glutenin (wheat) ..| 1.68. Gliadin (wheat)....| 1.14 Serum globulin (Ci) Gap eens 2.28 Ovalbuntn...-.... 1.23 Witte peptone ..... 3.03 Edestin (hemp REGO eer ese. 5... 1.40 IPMid it toe ao ee ee 2.90 Bence-Jones pro- on eee) ee 1.67 Geliiers 20. berer2: 0 Zein plus 2 per cent tryptophane..... 2.01 Fasal.10 TABLE IV. Tryptophane Content. Gravi- metric. Investigator. per cent|per cent 0.65 | 1.5 | Hopkins, F. G., and Cole, S2 Wa di: Physio!., 1901-02, xxvii, 418. 0.17 1.0 | Abderhalden, E., and Samuely, F., Z. physiol. Chem.., 1905, xliv, 276. 0.38 1.8 | Neuberg, C., and Popowsky, N., Biochem. Z., 1907, il, 357. Hopkins, F. G., and Savory, VH.))' Jt Physiol., 1911, xlii, 189. In the case of tryptophane it is to be noted that others have attempted to determine it by colorimetric methods, the latest work being that of Fiirth and Noble." The figures which these 8 Fiirth, O., and Lieben, F., Biochem. Z., 1920, cix, 124. ® Herzfeld, E., Biochem. Z., 1913, lvi, 258. 0 Fasal, H., Biochem. Z., 1912, xliv, 392. 11 Fiirth, O., and Nobel, E., Biochem. Z., 1920, cix, 103. 434 Tyrosine, Tryptophane, and Cystine authors have obtained are so very much higher than those given by our method that either their procedure or ours manifestly must be fallacious. Firth and Noble admit errors amounting to 27 per cent when known amounts of pure tryptophane are added to gelatin. We do not think that any experienced investigator could obtain anything approaching such errors by our method. TABLE V. Cystine Content. Colori- | Gravi- Protein. ptEiG Lange Investigator. per cent|per cent Boris ee ee eee 6.67 | 6.8 | Moérner, K. A. H., Z. physiol. Chem., 1901-02, xxxiv, 207. WOO Ue tect ceitan 7.8 | 7.3 | Abderhalden, E., and Voitinovici, A., Z. physiol. Chem., 1907, lii, 348. RAN es ee ee roe 3.5 | 1.2 | Mérner, K. Au By. physiol. Chem., 1901-02, xxxiv, 207. Glutenin (wheat)........ 1.80 | 0.02 | Osborne, T. B., and Clapp, S. H.., Am. J. Physiol., 1906-07, xvii, 231. Gliadin (wheat)......... 2-a2) 1) OL e s Zein (yellow corn)...... 0.5 Casein (Hammarsten)...| 0.25 | 0.1 | Mérner, K. A. H., Z. physiol. Chem., 1901-02, xxxiv, 207. PMeshin < 25 As S550 208 0.75 | 0.3 | Abderhalden, E., Z. physiol. Chem., 1902-03, xxxvii, 499. Clava. 72s tote hie bo 0.2 Witte peptone, .2.. .. <5. 2.00 Human hare nate i. ele 16.50 |14.0 | Buchtala, H., Z. physiol. Chem.., 1907, lii, 474. 3. Cystine.—It is interesting to note that the gravimetric determination of cystine in the keratins, which are rich in cystine, has been accomplished with quite a fair degree of accuracy. The cystine figures recorded for other proteins, as might have been expected, are much less dependable. FREE AND BOUND WATER IN THE BLOOD. Br BENJAMIN S. NEUHAUSEN. (From the Laboratory of the Department of Physiology, the Johns Hopkins University, Baltimore.) (Received for publication, February 9, 1922.) To explain certain pathologic conditions and even normal func- tions of the body, assumptions have been made that the body water is partly bound and partly free. From a chemical stand- point the body water could be bound in two ways—by hydration of the ions and by imbibition by proteins. Many facts from electrical endosmosis, transference experi- ments, and the relation between conductivity and viscosity, all point to hydration of the ions. When, however, one begins to determine the relative hydration of the different ions, the results obtained differ with the method used. Thus Washburn and his coworkers (1) from transference and viscosity data assign to H ion, K ion, and Cl ion hydration values of 0, 1.3, and 5.4 mole- cules of H,O, respectively. On the other hand, Bjerrum (2) on the basis of the lowering of the freezing point and electromotive force measurements, calculates the respective values for these ions as 8 H.O for H ion, 2 H.O for Cl ion, and 0 H.O for K ion. On the basis of the theory of von Hevesy (3) one would expect maximum hydration forthe Hion. The effects of pressure on the degree of ionization and the phenomena of electrostriction lead to the idea that very probably all the solvent molecules are attracted by electrostatic forces to the ions, the attractive force being some function of the distance. This view has been recently advanced by Akerlof (4) to explain his remarkable results on the effect of salts on the activity of H ion as well as by Keller (5) in his discussion of the significance of the isoelectric point. While conceivably the attraction between the solvent molecules and the ions might vary, any appreciable change in the ratio of bound to free water would not be caused, as the body water has but a low salt concentration. 435 436 $ Water in the Blood The great factor, however, in binding water would be the imbi- bition by the colloids. It is well known that lyophilic colloids such as the blood proteins will imbibe large quantities of water, and that as the quantity imbibed increases, the pressure necessary to cause partial desorption becomes less. Highly dispersed particles, which present adsorbing surfaces, are subject to great compressing forces, as has been shown, especially by Williams (6). A change in the quantity or character of the surface active constituents in the blood would entail a change in the adsorption by the colloids and a consequent change in the compressing force upon them. Since the proteins constitute about 8 per cent of the blood serum, a marked change in the quantity of water imbibed would be of import and on such a basis a theory of bound and free water could be founded. It has, however, been found by Hof- meister (7) and others that in the case of the swelling of gelatin in salt solutions the concentration of these salts in the imbibed water and in the solution is practically equal. Should this like- wise be the case with the body proteins, the quantity of water imbibed would have no effect on the colligative properties of blood serum; as a consequence, bound or free water could not be detected by the use of these properties. Burgarszky and Tangl (8) determined the freezing point lowering of dog’s serum and found values between — 0.550 to — 0.639, corresponding to concentrations of 0.297 to 0.354 molar. An average of 55 determinations by different investigators is given by Hamburger (9) as — 0.571 corresponding to 0.308 mols. To note how near these values approached those at 373°C., determinations of the lowering of the vapor pressure of blood serum at about 373°C., compared to that: of pure water at the same temperature, were, therefore, made. Alveolar air was bubbled through the serum, and air through the pure water, and then in each case through sulfuric acid. For the same volume of air passed through, the difference in the quantity of water ab- sorbed by the sulfuric acid through which the air from the pure water passed, and that absorbed by the sulfuric acid through — which the alveolar air from the serum passed was proportional to the difference in vapor pressure. This value divided by the weight absorbed by the sulfuric acid through which the air from the pure water passed gave the relative lowering in vapor pressure. B. S. Neuhausen 437 The value of this obtained as an average of six experiments, in which no value differed by more than 5 per cent from the mean was aoa If serum be considered a dilute solution, the following relation holds: aes Bs In this relation dP equals the difference in vapor pressure between the pure solvent and solution; P the vapor pressure of the sol- vent; n the number of mols of solute; and N the number ‘of mols of solvent. If we substitute for — its value aoe and for N the number of mols of water in 1,000 gm., 7.e. 55.55 mols, we obtain for ‘‘n”’ a value of 0.384 mols. There are two complete analyses of dog’s serum given by Abderhalden (10). If the averages of the two values given are recalculated in molecules per 1,000 gm. of water, and following Burgarszky and Tangl’s assumptions that the NaCl is about 84 per cent dissociated and that the balance of Na, K, and other cations may be considered as equivalent to an equal concentration of Na,CO3 with a degree of dissociation of about 69 per cent, one obtains a molar or- (as Hamburger has termed it) ‘‘osmotic concentration” of 0.3298 mols per 1,000 gm. of water. Con- sidering the unavoidable errors in a dynamic determination of lowering of vapor pressure the agreement is good.! The fair agreement, moreover, between the calculated molar concentration from the analyses, from the lowering in freezing point, and from the lowering in vapor pressure, supports the con- clusion that water imbibed by proteins, as in the case of gelatin, carries practically the same concentration of salts, so that any theory of bound and free water cannot be confirmed or disproved by any physicochemical method based on colligative properties. The author desires to thank Professor E. K. Marshall, Jr. for the samples of blood serum. 1 Tt should be borne in mind that Abderhalden’s results do not contain any figures for COz, likewise the values from the freezing point determina- tions are probably too low because of the loss of some of the COy. 438 Water in the Blood OONA ow wh 10. BIBLIOGRAPHY. . Washburn, E. W., J. Am. Chem. Soc., 1909, xxxi, 322. Washburn, E. W., and Williams, G. Y., J. Am. Chem. Soc., 1913, xxxv, 750. Wash- burn, E. W., and Millard, E. B., J. Am. Chem. Soc., 1915, xxxvii, 694. . Bjerrum, N., Z. anorg. Chem., 1920, cix, 275. - yon Hevesy, G., Kolloid. Z., 1917, xxi, 129. . Akerléf, G., Z. physik. Chem., 1921, xeviii, 260. Keller, R., Biochem. Z., 1921, exv, 134. . Williams, A. M., Proc. Roy. Soc. London, Series A, 1920, xeviii, 223. . Hofmeister, F., Arch. exp. Path. u. Pharmakol., 1891, xxviii, 210. . Burgarszky, 8., and Tangl, F., Arch. ges. Physiol., 1898, lxxii, 531. . Hamburger, H. J., Osmotische Druck und Ionenlehre in den Medi- cinischen Wissenschaften. Zugleich Lehrbuch physikalisch-chemis- cher Methoden, Wiesbaden, 1902-04, i, 459. Abderhalden, E., Z. physiol..Chem., 1898, xxv, 106. THE PROTEINS OF THE TOMATO SEED, SOLANUM ESCULENTUM.* By CARL O. JOHNS anp CHARLES E. F. GERSDORFF. (From the Protein Investigation Laboratory, Bureau of Chemistry, United States Department of Agriculture, Washington.) (Received for publication, February 8, 1922. According to Rabak (1), the total quantity of tomato seed waste which accumulates annually in the United States as a by-product of tomato pulping plants varies with the pack of the particular season and with the variety of tomato used. From figures determined by several experiments, American grown tomatoes contain on an average 1.13 per cent of dry waste of which 46.1 per cent is seeds. Shrader and Rabak (2) found ‘‘that a total of 2,063 tons of seed is available as the output of all the pulping plants in the eastern and middle western tomato belts.’? Most of this seed has gone to waste, some being dumped into rivers and some allowed to rot near the packing plants. A little has been tried as fertilizer. ' The seed and press-cake used for the experiments described in this paper were obtained from different sources. Total nitrogen determinations were made, the results of which are given in Table I. The meal used in preparing the proteins was obtained by grinding seeds of high germinating quality and removing most of the oil by extracting with U.S.P. ether. Thus prepared the meal contained an average of 37.28 per cent of protein (N X 6.25). Two globulins, designated a and 8, have been isolated from the saline extracts and separated by fractional precipitation with ammonium sulfate. According to the elementary analyses, * A preliminary report of this paper was presented at the 15th annual meeting of the Society of Biological Chemists held in Chicago, December 28 to 30, 1920 (cf. Johns, C. O., and Gersdorff, C. E. F., J. Biol. Chem. 1921, xlvi, p. xxvi). 439 440) . Proteins of the Tomato Seed TABLET. Analyses of Seed and Press-Cake. Description. Total N. at ee ee per cent per cent Press-cake* (protein coagulated by heat of pressing)..| 5.78 36.13 ne + (seeds heated to 60°C. before pressing)....| 5.90 36.88 Seedst (unheated, oil extracted by ether)............. 5.96 30.23 5.83 36.42 Seeds§ (highest germinating quality, oil extracted by CUED) s/scrias otter aeenade ties OC ee ee eee 6.08 37.91 AVOTARE: .. . Fees <7 acu eprnte < rycen tak ne a ee 5.91 36.91 * This material was furnished by the Oil, Fat and Wax Laboratory, Bureau of Chemistry, U. 8. Department of Agriculture, Washington. + This material was furnished by the Bureau of Plant Industry, U. S. Department of Agriculture, Washington. t These seeds were obtained in the open market. § These seeds were furnished by the Office of Congressional Seed Distri- bution, U. S. Department of Agriculture, Washington. ‘ these globulins exhibit marked differences in certain of their constituents. The average composition of the two globulins, calculated on a moisture- and ash-free basis, is as follows: C H N Oo per cent per cent per cent per cent per cent * Pt AC} 610 rs ae ae 52.29 6.89 18 .34 1.16 21.32 B- erie oat Niele Oa 6.84 16.02 0.81 25.12 It has been found that the a-globulin coagulates on heating for 10 minutes at 74°C., while the 6-globulin in the same period coagulates at 96°C. Both globulins give qualitative tests for tryptophane and tyrosine. In the distribution of the basic amino-acids as determined by the Van Slyke method, a marked difference is shown in the content of histidine, the a-globulin being low in this amino-acid, while it is unusually high in the 8 compound. The content of arginine and lysine in both globulinsis high. In the distribution of the nitrogen calculated on the basis of the weight of the protein, from the analyses made by the Van Slyke method, the greatest difference appears to be in the content of the non-basic nitrogen. C. O. Johns and C. E. F. Gersdorff 441 Feeding experiments on the nutritive value of the seeds after the removal of the oil, which proved to be of excellent table quality (3), have been made by Finks and Johns (4), who found that the proteins were nutritionally adequate, normal growth of albino rats being obtained. Our chemical study confirms these conclusions. Tests failed to disclose the presence of albumin and glutelin. EXPERIMENTAL. Experiments on Press-Cake. Our first experiments were made on a press-cake which con- tained 36.13 per cent of protein (N X 6.25). It was found that the heat developed during the pressing had rendered the pro- teins insoluble in the ordinary saline solvents. A number of preparations of protein were made, however, by extracting the finely ground press-cake with 0.5 per cent solution of sodium hydroxide. These preparations were uniformly alike in composi- tion, but differed from the preparations obtained from the saline extracts of the ether-extracted meal by having much lower sulfur and nitrogen contents and higher carbon. These lower sulfur and nitrogen results may have been due to the action of the alkali used as the solvent, and the higher content of carbon to the ad- mixture of carbohydrates. These preparations were a light buff in color. Total Globulins Extracted by 0.5 Per Cent Sodium Hydroxide Solution.—Extractions were made, using 10 cc. of solvent to each gram of meal. The quantity of meal used for a preparation ranged from 25 to 500 gm. Extractions ranged from 48 to 96 hours. Duration of the extractions with these time limits had no appreciable effect on yield or analysis, 23.63 per cent of pro- tein being extracted (N xX 6.25), corresponding to a yield of 65.40 per cent based on the total protein in the meal. The clear extracts obtained after being filtered through paper pulp were acidified with 1 per cent acetic acid, the acid being added slowly until the flocculent precipitates coagulated suffi- ciently to settle. The precipitates were collected on folded filter paper and washed several times with distilled water. They were then redissolved in a 0.2 per cent sodium hydroxide solution, 442 . Proteins of the Tomato Seed filtered, precipitated as before, and washed with distilled water until the washings gave no tests for acidity. They were then washed with absolute alcohol four times and once with absolute ether and dried in a vacuum oven at 110°C. The yield averaged 11.2 per cent. The preparations were then exposed to the air for 72 hours to enable them to acquire a moisture equilibrium, so as to facilitate the weighing of samples for analysis. The average analyses of five preparations thus obtained showed the following composition, calculated on a moisture- and ash-free basis: Cc H N Ss per cent per cent per cent per cent 54.12 6.88 15.06 0.84 Our subsequent experiments were made on seeds specially selected for their high germinating quality, this being an assurance that they had never been subjected to a temperature that would coagulate the proteins. Experiments on Seeds. Preliminary Experiments.—Extraction experiments were made. with different concentrations of sodium chloride in water, ex- tracting for 1 hour in each experiment. Nitrogen! determina- tions were made on aliquot portions of the filtered extracts and percentages of protein extracted calculated from these results. The data are given in Table II. Time extraction experiments were next made with the solvent which extracted the maximum amount of protein. These experiments show that a 4 per cent sodium chloride solution, in 1 hour, extracts the maximum amount of protein—21.34 per cent (N X 6.25). These results are given in Table III. Test for Glutelinn—Exhaustive extraction experiments were made using a 4 per cent sodium chloride solution. Three extrac- tions were made with this solvent, at the rate of 8 cc. for the 1 Acknowledgment is due to Mr. S. Phillips of this laboratory for the Kjeldahl nitrogen determinations made in the course of the present inves- tigation. C. O. Johns and C. E. F. Gersdorff 443 first and third, and 10 cc. for the second, for each gram of meal, extracting for periods of 1 hour each. The extracts were filtered and measured. Nitrogen determinations were made on aliquot portions of the extracts measured, allowance being made in case of the third extract for the residual extract in the meal. After the third extract had been removed, the residue was washed once with sodium chloride solution and then several times with dis- TABLE II. Preliminary Extraction Experiments. Solvent: 8 ce. per gram of meal, extracted 1 hour at room temperature. Protein Solvent. N extracted. extracted (N X 6.25). per cent per cent Te ie 0.80 5.02 Sodium chloride: MRSC INIIE fois oo ot cla icleic.- baie n\e = os s"nlajaiaiois sie. 1.63 10.18 Ells oy Saeed ie ES ee i rae 2.02 12.64 Spr IONE IUEE RS Sie cise oie cs 36s oo Saleiaje ee wis c 2.86 17.90 G0) 0. SR a es Soa oo ee 3.42 21.34 a ee MESS OE ii gee 20.92 oe SEL 5 area wet seksy0i isi @syv Ine edits 3.15 19.66 eS oe oo oii co-o a acs ars Sv dies 0's ob a a we 3.01 18.81 REI are cg icc cides ce vasveasenses 2.76 We27 OUI MIEMEE S OL ec oes ic seb ose weed cdeees 2.12 16.99 TABLE III. Time Extraction Experiments. Solvent: 4 per cent sodium chloride solution, 8 cc. per gram of meal. SPANHEUCRERACHOO GROUT Sot acre oleic - 18.28 |18.42 |18.35 |18.50 |18.57 {18.54 |18.24 |18.44 |18.34 eS ee 6? | L.16 We li/a te leslig 1.06 | 1.06 eg 20.87 20.96 21.28 Moisture..... 5.42 7.14 6.32 fo ee ee 0.29 0.02 0.18 Preparation IV. Preparation V. Preparation VI Ceeieside is. : 52.18 |52.23 |52.21 151.82 [51.92 |51.87 |52.47 |52.51 152.49 Bethke sua-'s 6.78 | 6.96 | 6.87 | 6.84 | 6.83 | 6.84 | 6.86 | 7.11 | 6.99 |, 18.21 {18.18 /18.20 |18.38 |18.27 |18.33 |18.63 |18.48 [18.56 eich ee 12165) 21,16 1ea8: | Es BTS ACLS RO ree tesiter av 21.56 21.78 20.81 Moisture..... 5.61 6.85 5.50 y C.| ee ee 0.17 0.05 0.28 Preparation VII. Preparation VIII. eee aett ere one: 51.78 |51.97 |51.88 |52.30 |52.31 [52.31 52.29 Be echt Se 6.84 | 6.90 | 6.87 | 6.94 | 6.74 | 6.84 6.89 i i ee 18.26 |18.10 |18.18 |18.27 |18.23 |18.25 18.34 ee ee LAG) 4.40 1.28 | 1.28 1.16 RS eee 21.97 21.32 21.32 Moisture..... 5.43 5.83 /. (| ee Oa 0.39 0.16 * Calculated on moisture- and ash-free basis. saturation with ammonium sulfate after the removal of the middle fraction at 0.4 of saturation, there is practically no separa- tion of protein. The middle fraction with a sulfur content of 0.89 448 Proteins of the Tomato Seed per cent is evidently a mixture containing merely traces of the a-globulin. Qualitative Tests—By the Hopkins and Cole method, the a-globulin gave a strong positive test for tryptophane, the color ring developing immediately. The 6-globulin, however, gave but a slight test, a faintly colored ring developing in about 10 minutes. TABLE VI. Analyses of the B-Globulin.* Preparation I. Preparation II. Preparation III. Aver- * Aver- Aver- I TEAt eer ieee In jee I Eb ae ee per cent | per cent | per cent | per cent | per cent | per cent | per cent | per cent| per cent Gas Ae 51.43 151.48 151.46 [51.07 (51.16 [51.12 |51.25 151.02 |51.14 1S RR ease | 6.79 | 6.86 | 6.83 | 6.77 | 6.98 | 6.88 | 6.71 | 6.72 | 6.72 INE ass ROE tee 16.02 |15.97 |16.00 |16.03 |16.04 |16.04 |16.01 |16.05 |16.03 See 0.80 | 0.80 | 0.80 | 0.79 | 0.80 0.79 | 0.79 Oe ee cree 24.91 25.16 waa Moisture..... 6.59 7.16 9.28 AGE Gama 1.99 1.62 0.13 Preparation IV. Preparation V. ee = CRS ES! ae! 51.21 (51.07 (51.14 [51.21 (51.19 |51.20 521 15a me pe eee | 6.87 | 6.83 | 6.85 | 6.92 | 6.96 | 6.94 6.84 ING eee: oe 16.04 |16.08 |16.06 |15.97 |15.96 |15.97 16.02 Sisfeiinte BAe 0.82 | 0.82 0.82 | 0.82 0.81 OF AeA 20 13 25.06 25.12 Moisture..... 6.41 6.28 Ashen es nes 0.30 0.29 * Calculated on moisture- and ash-free basis. By Millon’s test, the presence of tyrosine in both globulins was disclosed. Distribution of the Nitrogen in the a- and B-Globulins as Deter- mined by the Van Slyke Method.—Duplicate samples weighing 3.0 gm. each were dissolved in 100 cc. of 20 per cent hydrochloric acid and hydrolyzed by boiling the solutions under condensers for 30 hours. The phosphotungstates of the bases were decom- C. O. Johns and C. E. F. Gersdorff 449 posed by the amyl alcohol-ether method (6). The results of these analyses are given in Tables VII, VIII, IX, X,and XI. The TABLE VII. Distribution of Nitrogen in the a-Globulin as Determined by the Van Slyke Method.* Sample I. Ash- and moisture-free, 2.7852 gm. protein, 0.5164 gm. nitrogen. j Sample II. Ash- and moisture-free, 2.7852 gm. protein, 0.5164 gm. nitrogen. f I II I hope gm gm. per cent | per cent | per cent Pemeigicmigie rte. ee Aa. ee ee 0.0597 0.0601) 11.56) 11.64) 11.60 Humin N adsorbed by lime.............. 0.0033/0.0032 0.64) 0.62) 0.63 “ _N in ether-amy] alcohol extract. .|0.0080,0.0083) 1.55) 1.61) 1.58 0 Se ee 0.004110.0042| 0.79} 0.81) 0.80 NO re eo cig ons ope res 0.1253/0.1250) 24.27) 24.21) 24.24 LPL tb Loe 0.0077/0 0098) 1.49} 1.90) 1.70 Se ee ee 0.0267/0.0255 5.17| 4.94) 5.05 smasayrOr HICEALE. 65. 43s ons ote 0.2772/0.2756 53.68] 53.37| 53.52 Non-amino N of Ss sae ee pe 1269)|(21- 76lRde 73 EE ES | SSS OO Oe moter -resaimed...... . 22.25.2222... (0.5207/0.5208 100 .84)100.86 100.85 * Total N regained corrected for solubility of the bases (7). 7 Nitrogen content of protein, 18.54 per cent. TABLE VIII. Basic Amino-Acids in a-Globulin. . I II Average. per cent per cent per cent Wo EE ee i 1.26 1.29 1.28 Oe ee eee 13.98 13.95 13.97 A se wt, i 1.02 1.30 1.16 Me a se 5.00 4.78 4.89 OES SS Present. figures for cystine are undoubtedly low, since this amino-acid is slowly decomposed in boiling hydrochloric acid. TABLE IX. Distribution of Nitrogen in the B-Globulin as Determined by the Van Slyke Method.* Sample I. Ash- and moisture-free, 2,7366 gm. protein, 0.4389 gm. nitrogen. fT Sample II. Ash- and moisture-free, 2.7366 gm. protein, 0.4389 gm. nitrogen.7j I II I ol dg gm. gm. \percent| per cent| per cent IATAIAGUN 5. 3 bens Aact -oeie)-k Doe eee pe 0.0510)0.0516) 11.62) 11.75) 11.69 Humin N adsorbed by lime..:........... 0.0071 0.0073} 1.62) 1.66) 1.64 “Nin ether-amy] alcohol extract. ./0.0013,0.0012} 0.30) 0.27) 0.28 Cystine Wot) otk.n Aen ee \0.0036'0.0037; 0.82; 0.84) 0.83 21.44) 21.28) 21.36 0.0299; 6.04 6.81) 6.43 Axginine Bn deretongeuourwds so egcen ere HistidinedN 4.8420 eee oe ern ee 0.0265 Bigsinie Nisin. bid esrencape eee cheered deter 10.0340/0.0326| 7.75) 7.43) 7.59 Arnivic WOCRIAT | ee. Be : 2052/0. 2028] 46.75| 46.20) 46.47 Non-ammo! Noof- filtrate: 5525.22.22... ... 0. a 0185) 4.69} 4.22) 4.46 Total N regained...............-. - a (0. 4434 0.4410 101 .03'100.46 100.75 * Total N regained corrected for solubility of the bases (7). + Nitrogen content of protein, 16.04 per cent. TABLE X. Basic Amino-Acids in B-Globulin. I If Average. per cent ee per cent Cystine=. ...+.Jeivek seh at ay ae 1.13 1.16 1.14 ATPiNINE..,...2652-522- = wee eee 10.69 10.61 10.65 HigtiG mie 7. bac sae peice ngs eee oe 3.57 4.03 a eee Liyeities gcc. open wea ene pees 6.48 6.21 6.35 Tryptophane::....: 2 svrecer oe Present TABLE XI. Distribution of Nitrogen in the a- and B-Globulins as Calculated from the Van Slyke Analyses in Terms of Per Cent of the Proteins. | a@-Globulin. 8-Globulin. N Ss I n |) 4 Me eis a per cent | per cent | per cent | per cent| per cent| per cent ANNIE Pas Fee Tee ee me ee 2.14°|°2.16°| 2.15 [2286 4-89 es Gn ee ss a ee ee eee ae 0.41 | 0.41 | 0.41 | 0.31 | 0.31 | 0.31 BASIC TAS ithe soles biceeee ee 5.88 | 5.91 | 5.90 | 5.78 | 5.83 | 5.81 INION-DASIC:.5 4.0 -raen ee eee 10.26 |10.22 |10.24 | 8.25 | 8.09 | 8.17 POLARS dune toc tenis nace 18.69 18.70 |18.70 |16.20 |16.12 |16.16 ————— C. O. Johns and C. E. F. Gersdorff 451 SUMMARY. Analyses of five samples of tomato seed and press-cake show an average content of 36.91 per cent of protein (N X 6.25). The total globulins were extracted from the press-cake by extrac- tion with 0.5 per cent sodium hydroxide. Preparations obtained from these extractions gave uniform results on analysis. From seed of high germinating quality two globulins, a and 8, have been isolated and analyzed. These globulins, coagulable by heating for 10 minutes at 74 and 96°C., are respectively precipitated from their saline solutions by 0.3 of saturation in the case of the a-globulin, ‘and by saturation with ammonium sulfate in the ease of the 6-globulin. They differ in their solubility, the a- globulin being easily denatured, while the 8 compound is highly soluble in mere traces of aqueous sodium chloride. The elemen- tary analyses of the two globulins show the following percentage differences in composition: C N 8 per cent ; per cent per cent cl 0 52.29 18.34 1.16 B- 4, a 51-21 16.02 0.81 The sulfur content of the globulins is in the ratio of 3:2. The analyses by the Van Slyke method show that the nutri- tionally essential basic amino-acids are well represented in these globulins. It is of interest to note that these globulins are high both in arginine and lysine. The §-globulin contains also an unusually large percentage of histidine, but in the a-globulin the histidine content is low. Both globulins respond to the qualita- tive tests for tryptophane and tyrosine. In the distribution of the nitrogen based on the weight of the proteins, as calculated from the Van Slyke analyses, the most marked difference is found in the figures for the non-basic nitrogen. Tests failed to disclose the presence of albumin and glutelin. 452 Proteins of the Tomato Seed BIBLIOGRAPHY. Rabak, F., U. S. Dept. Agric., Bull. 632, 1917. . Shrader, J. H., and Rabak, F., U. S. Dept. Agric., Bull. 927, 1921. Bailey, H. S., and Burnett, T. B., Science, 1914, xxxix, 953. Finks, A. J., and Johns, C. O., Am. J. Physiol., 1921, lvi, 404. Johns, C. O., and Waterman, H. C., J. Biol. Chem., 1920, xliv, 306. . Van Slyke, D. D., J. Biol. Chem., 1915, xxii, 281. . Van Slyke, D. D., J. Biol. Chem., 1911-12, x, 15. NO OR wD A COLORIMETRIC METHOD FOR THE DETERMINA- TION OF HOMOGENTISIC ACID IN URINE. By A. P. BRIGGS. (From the Laboratory of Biological Chemistry, Washington University School of Medicine, St. Louis.) (Received for publication, January 27, 1922.) When o-phosphates are added to ammonium molybdate in acid solution, the well known ammonium phosphomolybdate is formed, which corresponds to the 1:24 type of phosphomolybdiec acid described by Wu.! Bell and Doisy? have found this complex to be especially susceptible to reduction by hydroquinone; a deep green coloration is produced in acid solution which is changed to an intense blue when made alkaline; the excess ammonium molyb- date is not reduced. On a basis of those reactions they have devised methods for the colorimetric determination of phosphorus in blood and urine. While carrying out the tests for homogentisic acid in an alcap- ton urine the probability arose that this p-diphenol would give the same color with phosphomolybdic acid as hydroquinone, and such was found to be the case. Subsequently the homogentisic acid was isolated, its color ratio to hydroquinone determined, and then hydroquinone used as the standard in the quantitative determination of homogentisic acid. The method is very simple: 1 or 2 cc. of the aleapton urine are diluted to about 15 ec. in a 25 cc. volumetric flask, 2 cc. of the molybdate solution and 2 cc. of the phosphate solution are added, and the mixture is diluted with water to the mark. Into another flask containing equal amounts of the phosphate and molybdate solutions an appropriate amount of hydroquinone standard is added and diluted with water up to the mark. The flasks are inverted a few times so that the contents are well mixed, and the comparison is made in the colorimeter after 5 minutes. 1Wu, H., J. Biol. Chem., 1920, xliii, 189. 2 Bell, R. D., and Doisy, E. A., J. Biol. Chem., 1920, xliv, 55. 453 454 Homogentisic Acid in Urine Solutions Used. 1. Phosphate solution contains 1 per cent KH2PO,. 2. Molybdate solution contains 5 per cent ammonium molybdate in5Nn H.S04. 3. Hydroquinone standard contains 1 mg. of hydroquinone per ce. The average of several comparisons gave 1 mg. of hydroqui- none equal to 0.79 mg. of homogentisic acid. It will be observed that this is not a molecular ratio. Various substances which might be expected to interfere with the colorimetric method, such as uric acid and phenols, give relatively little color. This was determined by adding to a series of flasks, each containing 1 mg. of homogentisic acid, 1 mg. of the substance to be tested. Phosphate and molybdate solutions were then added as in the determination and the colors obtained compared with that of 1 mg. of homogentisic acid alone. The following substances gave an error of about 1 per cent: phenol, cresol, resorcinol, and uric acid; pyrocatechol freed from traces of hydroquinone gave an error of about 3 per cent. Sulfides give a very intense blue-green; if present, however, they can be re- moved by precipitation with AgeSO,. Albumin, if present, would interfere by forming a cloud; it may be removed by trichloro- acetic acid. Sulfides and albumin may be conveniently removed together by treating the urine with equal volumes of CCl;COOH (10 per cent) and AgeSO, (0.5 per cent) ;this mixture is stirred well, solid NaCl to make about 2 per cent is added, the mixture stirred again, and then centrifuged. 3 cc. of the supernatant fluid are taken for 1 ec. of urine in the determination. Determinations of homogentisic acid made by the colorimetric method and by that of Wolkow and Baumann’ gave results which checked within about 10 per cent. The method of Wolkow and Baumann is described as very tedious by all who have used it, so the colorimetric method offers a more convenient means for studying the metabolic anomaly of alcaptonuria. 3 Wolkow, M., and Baumann, E., Z. physiol. Chem., 1891, xv, 260. THE RELATION OF PHOTOSYNTHESIS TO THE PRO- DUCTION OF VITAMINE A IN PLANTS.* By J. WALTER WILSON. (From the Biological Laboratory, Brown University, Providence.) (Received for publication, January 25, 1922.) That the synthesis of vitamine A takes place only in plants is generally believed. The green parts of plants appear to be among the richest sources of the vitamine, whereas seeds, in general, contain only traces of it (1-5). It seems apparent, therefore, that the plant must synthesize the vitamine A found so abundantly in its leaves. Since some plant syntheses are dependent upon solar energy it appeared desirable to find out whether or not the production of vitamine A is dependent upon photosynthesis. Recently Coward and Drummond (5) concluded that the amount of vitamine A does not increase in germination and that it is only increased in those plant tissues in which photosynthesis is going on. The following experiments have led me to a different con- clusion; v7z., that photosynthesis is not necessary for the produc- tion of vitamine A in plants. EXPERIMENTAL. Wheat seeds were sprouted in wooden trays on moist paper, one lot in the dark room and another in the sunlight. When the sprouts were from 2 to 3 inches high, they were cut off close to the seeds, dried in a current of air at about 60°C. (Osborne and Mendel, 2), and ground into fine powders. These were fed to young white rats in a diet deficient in vitamine A, 5 per cent of the powdered sprouts taking the place of an equivalent weight of corn-starch (Chart 1). * From a thesis presented in partial fulfillment of requirements for the degree of Doctor of Philosophy, Brown University, 1921. 455 456 Production of Vitamine A. in Plants GES Go PC Lee LL LL En A eee Beebe SY po A | ERR 4S ARR eee ase oe HERES ARR ARERR eee aa Ata ARR nERS la | | sguecdaen tL ALL A ty Ae ge a ae A ALS BESD AREER ARBRE ORS ABSP ST ef Ae Rf re — 1 Y A AN BY OF RO EW A AM rs PPEte -siee #MuEE aun @had Benet Ttebees = Fi { i Ae ee So Te Cuart 1. Showing that photosynthesis is not necessary for the produc- tion of vitamine A in plants. Rats K1, K2, and K3 were fed the basal diet with 5 per cent of dried etiolated wheat sprouts replacing an equivalent amount of corn-starch. Rats S1 and S2 were fed the basal diet with 5 per cent of dried green wheat sprouts replacing an equivalent amount of corn-starch. The basal diet was made up as follows: gm. Casein. 3.9). 5) A022, De eee 20 Cornestarch: 24).00..(0.-5 SIA ee ee eee 50 Dear oe iscccie . avesisie trae s' a(dave siaiete ate cere hen 1 Eee 20 Salt MIKE WTE! «soe, Feo ode caldjepieapei rit ; ‘ IO OF. ean eee ee oie inte eh MUFLINT DNS te.3 ofa ata ds ic 4 fone soidvyeest! off eddie med uliteds ah hebiee ¢ ooh osilnate'a > in yresbe pack? acura vit \isatow dtm a) » oF bem? loniy efeenogaed stones wt foie senor njots tis cele toulj oils fiive Bihev oie. donodaes siti zeigen at “yr bees ; eoraly etoiae ine oie eeyes lt! per THTNITIOD Sy IAN i ‘to suit ito: Ree ree Rees » box sees dio Pewh Uateleiioae sHtitididay of! jomit doodle » Ob eile ba my ‘THI (4 of ore eipeekh) > ehie PRET cee Ain nef't etedngg sstiticigt fal saneliinons ab atortibn inomnhilt tebe wma ita we dint baw uittirtelaies adl: dos heron aya iat , scottertes oat eben rit 188 . ( a ¢ a i> ; e = pe x - he = ¢ ; = bd ; Ske Ls ee ah gin aD * — , i 7 = i THE INTERACTION BETWEEN BLOOD SERUM AND TIS- SUE EXTRACT IN THE COAGULATION OF THE BLOOD. II. ACOMPARISON BETWEEN THE EFFECTS OF THE STROMA OF ERYTHROCYTES AND OF TISSUE EXTRACTS, UNHEATED AND HEATED, ON THE COAGULATION OF THE BLOOD, AND ON THE MECHANISM OF THE INTERACTION OF THESE SUBSTANCES WITH BLOOD SERUM. By LEO LOEB, MOYER 8. FLEISHER, anp LUCIUS TUTTLE. (From the Department of Comparative Pathology, Washington University School of Medicine, St. Louis.) (Received for publication, January 21, 1922. Curve of Tissue Extract with Hirudin Plasma and Fluoride Plasma. If we add to tubes containing 1 cc. of HP1 variable quantities of TE, we do not find the coagulation time to vary in an inverse ratio proportional to the quantity added, but instead irregular curves are found. In the majority of cases a medium quantity of TE represents the optimum; increasing the quantity above the optimum caused a decrease in activity or leaves the latter unchanged. Below the optimum the effect is weaker. In some cases, however, the largest quantities used (1 ec. of TE) were found optimal. The optimal quantity varies with the same ex- tract, if different plasmas are used, and with the same plasma if different extracts are used. When after standing the extract loses in strength, and the optimal quantity loses, for instance, about 100 per cent, large quantities may be found inactive. In some cases we observed with extract, as with red blood corpuscles, that with a homologous plasma the optimum is lower than with a heterologous plasma. With FIPI we find a similar condition, but with FIP1 1 cc. is more often optimal or near the optimal point than with HPI; but even with FIP] smaller quantities may sometimes be as good or better than larger quantities. 485 486 Coagulation of Blood. II Specificity of Tissue Extract. In former experiments in which the tissue extracts were allowed to act directly on the unaltered blood or plasma, a general (class) specific adaptation between tissue coagulins and plasma was shown to exist.! Under certain conditions this specific adapta- tion can also be demonstrated if we use hirudin and fluoride plasma instead of fresh blood or plasma. In all cases it is neces- sary to make cross-tests in order to eliminate variations in the absolute strength of extracts; it is furthermore necessary to use extracts and plasma in various proportions. With HPI it is much more readily possible to demonstrate the specific adap- tation than with FIPI, because with HPl the extracts alone usually cause coagulation of the plasma, while with FIPI, especially GFIPI, extracts alone are in some cases quite ineffec- tive and thus it is impossible to prove the specific adaptation of extract and plasma. In such cases it can, however, some- times be shown to exist by combining extract with serum. The homologous extract in combination with serum proves then in cross-tests relatively more active than heterologous combinations. In other cases the addition of serum to extract may, however, obscure the specific relationship between extract and plasma. This is due to the fact that in the mixture of serum and extract inhibiting substances are produced which counteract the tissue coagulins. The Effect of Heating on Tissue Extract. Tissue extract was heated to 56-57° or to 60° generally for a period of 30 minutes, or in some cases for 15 minutes. The effect of DKE heated to 56° for 30 minutes on DHP! or DPPI depended in some cases on the quantity of DKE used. 0.3 ce. of DKE (added to 1 cc. of HPl) was usually rendered more active through heating, while 1 ec. of DKE remained either unchanged or was even weakened. The improvement in 0.3 ec. of DKE was in some cases slight, but in other cases considerable. Some- times an improvement was produced through heating without a definite change in the character of the quantity curve. When- 1 Loeb, L., Montreal Med. J., 1903, xxxii, 507; Med. News, 1903, Ixxxiii, 212; Virchows Arch. path. Anat., 1904, clxxvi, 10. L. Loeb, M. S. Fleisher, and L. Tuttle 487 ever the smaller quantity was more improved through heating than the larger quantity, the quantity curve was altered through heating and sometimes the relative effects of large and small quantities were reversed through heating. In an experiment with GME 1 cc. was improved through heat- ing, but not 0.3 cc. if we tested the coagulating power of the extract with DHP]. With GHP1 no improvement was produced through heating GME, but in this case also the larger quantity became, through heating, relatively better than the smaller quantity. With DLE (tested with DHPI) 0.3 cc. remained un- changed through heating, but 1 cc. was slightly weakened. While with unheated extracts DLE was more effective than DKE, after heating the relation was reversed, DKE being more effec- tive than DLE. We may explain these facts if we assume that in the extract at least two substances are present, one causing acceleration when added to plasma, the other inhibition. In different extracts these two substances are present in different proportions; in DKE there is a relatively larger quantity of inhibiting substance than in DLE. The relative preponderance of the two substances may also vary in different quantities. In the case of DKE the inhibiting substance happened to be relatively more potent in 0.3 than in 1 ce.; while in GME the reverse relation happened to hold good. Heating the extract to 56° for 30 minutes injures the inhibit- ing substance to a higher degree than the coagulating substance and thus under certain conditions the coagulating power of the extract may be improved through heating. But heating injures not only the inhibiting substance but also the coagulating sub- stance and thus in certain cases heating may fail to produce an improvement or may even cause a deterioration. Heating DKE to 60° for 30 (or even 15) minutes weakens its coagulating power if tested with DHPI, quite definitely. Larger quantities were in our experiments more affected than smaller quantities. Heating DKE to 60° likewise weakens its coagulat- ing effect on DPP]. We may assume. that in heating to 60° the injury to the coagulating factor predominates over the weakening of the inhibiting substance. 488 Coagulation of Blood. II We carried out also a few experiments in which we compared the action of heated and unheated extract on DFIPIl. Heating the DKE to 60° weakened it very much; heating it to 56° weak- ened it in two cases while in one case it produced a very slight improvement. We may assume that with FIP! as with HPI heating injures both the inhibiting and coagulating substances, but with FIPI the inhibiting substance is of less importance than with HP] and PPI, and that, therefore, the weakening of the coagulating substance caused through heating counts in this case more than with HPI and PPI, or at least both the inhibiting and coagulating substance balance each other more nearly in the case of FIPI. In combination with serum the effect of heating on extract cor- responds to the effect of heating on the uncombined extract if we use DHPI! or DPPI to test the coagulating power. As we stated the coagulating effect of extract on DHPI or DPP! is usually in- creased through heating. Correspondingly we find that the inhibiting effect of S on DKE is usually diminished or entirely removed through heating the extract to 56° for 30 minutes; or instead of an inhibition we may even find a slight acceleration of the mixture as compared with the extract alone provided the mixture of extract and serum is added to the plasma at once. In the majority of these experiments 0.3 cc. of extract was mixed with 0.7 cc. of serum and the mixture added at once to 1 ce. of HPI or PPI. In case the heating does not improve the coagulat- ing effect of the uncombined TE, the inhibiting effect of serum in combination with such heated extract, if added at once to the plasma, is not diminished. This occurred, for instance, in our experiment with DLE. If the mixture of heated extract (56°) and serum is allowed to stand 6 minutes before being added to the plasma, an inhibiting effect of the serum is still noticeable, but the loss in the mixture of DKE + § on standing is much reduced through the heating of the extract. While with the unheated DKE the mixture usually loses many times in efficiency, the loss of the mixture of heated extract and serum was much smaller and varied in our experiments between 150 and 300 per cent. The time curve in the mixture acting on DHP1! or DPPI has, therefore, become not unlike that of the mixture of unheated E + S when acting on L. Loeb, M. S. Fleisher, and L. Tuttle 489 DFIPI]. A great part of the inhibiting effect on standing has been eliminated through heating. Quantitative variations in the diminution of the inhibiting effect through heating do, how- ever, occur and in one case the diminution in the loss of the mix- ture after standing was less than usual. In this case DPPI had been used. In cases in which the inhibiting substance in the extract plays a smaller part as in a mixture of DLE + DS, the time curve is not very much changed through the heating; the effect of the heating on the time curve is likewise diminished, if we decrease the amount of serum added to the mixture, be- cause a constituent of the serum is necessary for the production of the inhibiting effect. The more the serum is diminished in experiments with HPl or PPI, the less becomes the inhibiting effect, and the less significant becomes the effect of heating on the extract. If we test with DHPI] or DPPI an extract-serum mixture in which the extract has been heated to 60° for half an hour, the mixture resembles in its action much more a mixture of unheated extract and serum. This holds good if the mixture is added at once to the plasma as well as when it is added after a previous standing for a period of 6 minutes. In other cases the effect of the mixture of DKE + DS, in which the DKE had previously been heated to 60°, is intermediate between that of a mixture of unheated DKE+DS and of DKE+DsS in which the DKE had been heated to 56°, but usually it is nearer the former. We may interpret these facts by assuming that through heating to 60° the coagulating component of the TE is injured to approxi- mately the same extent as the inhibiting substance or only slightly less; and that thus through heating to 60° the balance between coagulating and inhibiting substance is’ always less favorably affected than through heating to 56°. The gain shown through the heating to 56° is, therefore, again lost or at least diminished through heating the extract to 60°. If we test the effect of the mixture of 56° heated DKE+DS with DFIP1 instead of with DHPI] and DPPI, and if we add the mixture to the plasma at once, we usually find some improvement as compared to the effect of the mixture of unheated DKE+DS on DFIPI1; in some cases no change in the efficiency of the mixture takes place through the heating. 490 Coagulation of Blood. II The time curve of the 56° heated DKE+DS with DFIPI is usually similar to that of the unheated DKE+DS§; there is, therefore, usually some loss after 6 minutes standing which is, however, considerably smaller than in the case of DHPI or DPPI. But inasmuch as there is usually some gain, if the mixture of 56° heated DKE+DS is added at once to the DFIPI, there is also, after 6 minutes standing, usually some absolute gain in the coagu- lating effect of the 56° heated DKE+DS mixture as compared to the mixture of unheated DKE+DS. Even 56° heated DLE may be better in combination with DS as compared to unheated DLE in combination with D§, although with DHPI, and if used uncombined with DFIPI, heating did not improve DLE in the experiment in which we used it. A mixture of 60° DKE+ DS added to DFIPI had less coagulating power than a mixture of unheated DKE+DS. The injury to the coagulating substance caused through heating DKE to 60° is more evident in the case of DFIPI than in the case of DHPIl and DPPI. In the case of the latter the fate of the inhibiting substance is of relatively greater importance. The Effect of Heated Extract on Hirudin and Fluoride Plasma. Mixture. Coagulation time. hra. min 8€¢. lee. DHPI+ 1 = ce. DKE High 2D 1 “c “cc + 0.3 “ “ ‘ 14 30 1 bases paced cat | “ heated (56°) DKE 5” 40 1 a3 “ - 0.3 “ “ (56°) “cc 5 30 ites 6 Ea « GME Sys 1 “ec “ -- 0.3 “cc “ 1 57 dete “« +1 =, “ heated (56°) GME pa Me 1 “ << — 0.3 “ce “ (56°) “ 5 9 et pe Mixture. 1 ec. DHP!] + 0.7 “ DKE “ 1 + 0.7 1 ec. DHP! + 0.7 cc “ DKE “ 1 +0.7 “ 1 ec. DHP! + 0.7 ce “ce “ce 1 = ah GME 1-ec. DHP! + 0.7 ce . DS + 0.3 ec. DKE “« +0.3 “ heated (56°) . DS + 0.3 ce. DKE “ . GS + 0.3 ec. GME “ +0.3 “ heated (56°) . GS + 0.3 cc. GME + 0.3 “ heated (56°) Coagulation time. min. sec, Coagulated next morning. 6 40 Partly coagu- lated next morning. 21 10 Trace coagu- lated next morning. Not coagulated next morning. 1 “ “ _ 0.7 “cc “ _ 0.3 “ heated (56°) “ “ “ GME 1 ec. DFIP] + 0.7 ce. DS + 0.3 ce. DKE Next morning co- agulated. a = +67 “ “ -+0:3 “ heated (56°) 31 15 DKE Mixture. Coagulation time. hrs. min, Sec. ec. DHP]+1 = cc. DKE 1 25 “ “ + 0.3 “ “c 5 15 S “+1 £4 heated (56°) DKE 1 35 cc “ _ 0.3 “ “cc (56°) “ 3, “ “ - 1 “ce ““ (60°) “ 12 “c «- -- 0.3 “ “ (60°) “ 8 15 “- DFIPI+1 “ DKE 12 55 <4 5 AES oa | ita “ 1 12 Partly coagu- lated; next morning partly coagulated. fh “+1 £4“ heated (56°) DKE 14.. .55 “ “ce “= 0.3 “c “ (56°) “cc 29 45 Al most coagulated. ° ies oy a ee 30 45 Most- ly coagulated. * SOG ES * Saintes) = Trace coagulated next morning. 491 THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI, NO. 2 492 Coagulation of Blood. II Time pects Mixture. Coagulation time. ing of D+S. min. hrs. min sec 0 | lec. DHP! + 0.7 ce. DS + 0.3 ec. DKE 30 =630 Almost coagu- lated. ‘Disa BB less eo Ol oe Oa an, Next morning. Trace coagu- lated. 0 |41 G4 iat ON eo Me HE O63, § heated aby) 2.) 3D DKE 6 | lec. DHP] + 0.7 cc. DS + 0.3 cc. heated (56°) G 15 DKE 0 | lec. DHP]1+ 0.7 cc. DS + 0.3 ce. heated (60°) 16 40 DKE 6 | 1 ec. DHP! + 0.7 cc. DS + 0.3 ce. heated (60°) 2 9 Trace. DKE ' Next morning partly coagu- lated. 0 | lec. DFIP!+ 0.7 “* “ +.0.3 ec. DKE Be 6 1 “ “ eh) “6 Se rT “ -6 OM « +0.7 “ “ +0.3 ec. heated (56°) 2 5 DKE 6 | lee. DFIPI1+ 0.7 cc. DS + 0.3 ce. heated (56°) 4 5 DKE 0 | lee. DFIP] + 0.7 ce. DS + 0.3 ce. heated (60°) 5 40 DKE 6 | 1 ec. DFIP1 + 0.7 ce. DS + 0.3 cc. heated (60°) tite, 50 DKE The Effect of Unheated and Heated Stroma of Erythrocytes with or Without the Addition of Serum. A suspension of the stroma of red corpuscles was obtained in the following manner: The corpuscles of defibrinated blood were separated through centrifugation from the serum and washed four times with 0.85 per cent NaCl solution. The corpuscles of 100 cc. of blood were then suspended in 76 cc. of salt solution and an equal amount of distilled water was added. After standing the stroma was separated from the hemoglobin through centrif- ugation. To the suspension of stroma so much NaCl was added that a 0.85 per cent solution resulted. L. Loeb, M. 8S. Fleisher, and L. Tuttle 493 We found the curves of TE with hirudin, peptone, and fluoride plasma representing the coagulation time as a function of the amount of TE to be irregular. The same holds good in the case of stroma of erythrocytes (Cp). The character of the curve depends to some extent upon the species whose corpuscles were used; and even the corpuscles obtained from different individuals may differ in some respects from each other. It also depends upon species and individual variations in the plasma. There is an optimum for each Cp preparation below and above which it is less active. With 1 cc. of DHPI| the optimum is in many cases between 0.4 and 0.7 cc. of RCp, but it may be as high as 1 cc. With DCp and CCp it is often between 0.2 and 0.4 cc. The curve may be flat or show a sharp turning point. With RCp it is on the whole flatter than with DCp or CCp. The latter two may, for instance, show a sharp turning point between 0.7 and 1 cc. With the little active turtle corpuscles larger quanti- ties are best. Also GCp are absolutely less active than mamma- lian Cp. Cp may cause coagulation not only of the HP! and PPI but also of FIPI, but less actively than serum; in this respect they again resemble TE. The effect depends in this case also on the kind of plasma used. With DFIP1 DCp may be better in smaller quantities; with RFIPI both large and small quantities may be equally active; and with GFIPI larger quantities may be better. Also with different samples of DHPI the optimum of DCp and RCp may vary and both may vary in the opposite direction. In several cases the optimum varied with different plasmas in such a way that with the same kind of Cp the optimum was lower with the homologous than with the heterologous plasma. Through cross-tests between Cp and plasma it is not only possi- ble to establish a relative class specificity, as in the case of TE, but also a specific species adaptation between plasma and cor- puscles.2 Dog, cat, rabbit, and goose corpuscle act relatively best each with its own plasma. While, therefore, GCp are ab- solutely little active, relatively they show the greatest coagulating power in combination with GPIl. We compared the action of Cp on DHPI and DPPI.2 Cp 2 Loeb, L., and Fleisher, M. 8., Biochem. Z., 1910, xxviii, 169. . 3 While the tests made with DPP] were numerous and concordant with each other, the specimens of PP] used in our experiments had all been obtained from one individual dog. It would be desirable to make addi- tional tests with the PPI! from different individuals. 494 Coagulation of Blood. II was found to be much less effective towards DPP] than towards DHP]. DTE, on the other hand, was about as active with DPPI as with DHPI; often it was more active with DHP1 and only once was it found to be slightly less active. Also in combination with serum Cp was relatively inactive towards DPPI as compared with DHPI. TE + S, on the other hand, is at least as active towards DPP] as against DHPI. DS and RS, unheated and heated, behaved alike towards both kinds of plasma. A combination of S with Cp, if added at once to DPPIl, may cause as much acceleration of the coagulation as corresponds to a summation of the individual effects of Cp and S on DPPI. With S alone DHPI coagulated somewhat better than DPPI. This greater responsiveness to Cp of DHPIl as compared with ~ DPPI1 may also influence the optimum of Cp. In one experi- ment, for instance, with DPP1 1 cc. of Cp was better than 0.3 ec.; while with DHPI, on the contrary, 0.3 cc. was better than 1 cc. It often holds good that the more inactive the Cp is, the higher relatively is their optimum. As we have seen, a mixture of S and TH, if added at once to HPI or PPI, is usually less active than TE alone; the mixture added at once to FIP] usually shows activation. After standing of the mixture, there is a loss in efficiency with all three kinds of plasmas, but it is much more marked with HPI and PPI than with FIPl. While Cp behaves essentially like TE, there is a marked difference between the behavior of Cp + S on the one hand of TE + S on the other hand towards the three kinds of plasma. A mixture of Cp + S, added at once to HPI, is usually as active as Cp alone; or sometimes it is slightly more active. In a few cases it was even markedly more active, and in a few other cases there was just an indication of a loss as compared with the Cp alone. After 6 to 10 minutes standing, the mixture has in most cases gained in coagulative power, even in cases in which if added at once it had not been more active than the Cp alone. This improvement may be slight or marked. There are | cases in which there is no improvement, or even a slight loss in the coagulative power of the mixture on standing. The time curves of the mixture vary in different cases and the effect de- L. Loeb, M. S. Fleisher, and L. Tuttle 495 pends on the time when the mixture has been added to the HPI. The gain in power through standing may extend over a period of several hours, though the optimum is usually reached at an earlier period; this optimum is followed by a loss which is, how- ever, very gradual. The shortening of the coagulation time with the standing of the mixture may be progressive up to a period of 30 minutes. In other cases the optimum is reached after 5 to 10 minutes standing, and after 30 minutes standing there is again some decrease which has become greater after 3 hours standing. In one case there was no improvement in the mixture, when added at once; after 6 minutes standing there was a slight improvement _ and after 10 minutes there was again a slight loss. The improve- ment in the mixture may already be very marked if it is added at once, while after 10 minutes standing there may still be some improvement, although less than when the mixture has been added at once. In some cases there was no improvement in the mixture, either when added at once, or after 6 minutes standing. With DFIP1 the improvement caused through mixing Cp and § as compared with Cp or S alone, is usually greater than with DHP! if the mixture is added at once to the plasma. The same effect was found in the mixture of S + TE. In some cases, how- ever, there was no activation through the mixing of S and DKE, when adding the mixture at once to DFIPI. After 6 tol0 minutes ‘standing of Cp + S, there may be great improvement over the mixture added at once; this applies to cases in which the mixture, if added at once, did produce an acceleration as well as to other cases in which it did not produce an acceleration. In other cases, however, there was not found any improvement or even a slight deterioration after 6 minutes standing. If we compare the mixture of the same kind of Cp and § with HPI and FIPI in a particular case, we find corresponding time curves with both kinds of plasma and the optimum is situated at similar points of the curve, although the absolute quantities of acceleration differ, the improvement being usually greater with DFIPI. If, on the other hand, we compare the mixtures of Cp and § in their action on HPI and FIPI of different species, we may find differences in the curves. With DHPl and RHPI the curves may differ; with DS there may be after 6 minutes standing more loss with DCp than with RCp. Those kinds of Cp which alone 496 Coagulation of Blood. II are most active with HPI usually accelerate also most effectively in combination with serum. Effect of Heated Cp.—While tissue extract heated to 56° is often more effective in causing coagulation than unheated ex- tract, such an improvement is not found after heating Cp to 56° for 30 seconds. Effect of Heated Cp on DHPI.—There is uniformly a loss in the coagulating power of the Cp when uncombined with serum. This loss usually varies between 50 and 300 per cent, but it may be much greater (1,200 to 2,000 per cent). In cases in which the unheated Cp is more active in the smaller quantity (0.3 cc.), heating may in some cases cause a reversal and make the larger quantity (1 ce.) more active than the smaller one without, however, producing an absolute improvement. The same reversal we observed in the case of heated TE. Inasmuch as even unheated Cp is relatively inactive with DFIPI, heating renders the uncombined Cp quite inactive with FIPl. A mixture of S + heated DCp acts on HPI as we might expect from a knowledge of the effect of a mixture of S + un- heated Cp on HPI] and of the effect of heating on the uncombined Cp. The combination of S + heated Cp, if added to HPI at once, is to the same extent less active than the mixture of S8 + unheated Cp as heated Cp alone is less active than unheated Cp. Or the loss caused by heating may appear even somewhat greater in combination than uncombined. Thus if the combina- tion of unheated corpuscles with serum, added at once to the HPI, neither accelerates nor inhibits as compared with the action of unheated Cp alone, then the use of heated Cp may show a slight loss in combination as compared to the action of heated Cp alone. After 6 minutes standing S + heated Cp may be more effective than the same mixture added at once, as we found to be the case with the mixture of serum and unheated Cp; but the mixture with the heated Cp is distinctly less effective than S + unheated Cp after 6 minutes standing. Relatively it may gain as much as S + unheated Cp after 6 minutes standing or the gain may be somewhat less. On the whole the more the mixture of S + unheated Cp surpasses unheated Cp alone in effectiveness, if added to the HPI at once or after 6 minutes standing the more the mixture of S + heated Cp surpasses the heated Cp alone. There is, therefore, a parallelism between the behavior of the L. Loeb, M. S. Fleisher, and L. Tuttle 497 heated and unheated Cp with HPI, but the heated Cp is always distinctly less effective. With DPP! heated Cp alone as well as in combination with S may be quite inactive. This corresponds to the low degree of efficiency of Cp towards DPPI. While with DFIP1 heated Cp alone may be inactive, in combina- tion with serum it may be active, but again less so than unheated Cp in combination with serum. Both unheated and heated . Cp may gain approximately in a similar proportion after standing with serum for 6 minutes; though the gain of the heated Cp may be relatively slightly less than that of the unheated Cp. Abso- lutely the mixture of heated Cp + § is less active than the mix- ture with unheated Cp. If we apply in the case of the Cp the same reasoriing which we used in the case of TE, we may conclude from the fact that heat- ing the corpuscles causes only loss and never a gain in coagula- tive power, that in the erythrocytes the inhibiting substance which we found in TE is’absent or at least present in a much smaller quantity than in TE. This conclusion is corroborated by the fact that on standing with serum the Cp gains in coagu- lative power under conditions in which a combination of TE + S loses, and that the loss of a combination of S + Cp is much more gradual and much slower than the loss in a combination of S+ TE. The gain usually is noticeable for a relatively long period of time in the combination of S + Cp. We may there- fore conclude that the accelerating substance forms mainly in combination of Cp with S and that the inhibiting substance either does not form at all or forms in a much smaller quantity. On the other hand, we again find corroborative evidence for our conclusion that heating tissue coagulin to 56° injures it to some extent; this comes out much more clearly with Cp because here the destruction of an inhibiting effect does not obscure the in- jurious effect of moderate heat on the coagulating substance. We may furthermore conclude that inasmuch as the irregular curves are found in the case of uncombined Cp, as well as of uncombined TE in their action on HPI, although in the former the inhibiting substance is present to a much smaller extent than in the TE, these irregular curves cannot be due, or at least cannot be entirely due, to the same inhibiting substance which combines with the serum. ' The Action of Cp on Plasma. Mixture. Coagulation time. min. sec 1 cc. DHP! + 0.3 ce. DCp 5 50 fees) 4-038. Seip 4 aes Ey t=) di yeaa Next morning, no coagulation. ‘han ee + 0.3 “ DCp 54 es +0.3 “ CCp 5 30 | 7 EOE AS EDS 20 5 Time stand: Mixture. Beygianen S+Cp min. min. sec 0 1 cc. wie + 0.7 cc. DS + 0.3 ce. DCp 3 5 ques “Ei 7 “ “ 3s “ “ 1 35 10 10) «oa oe 1 50 15 Di} jot occ eOet. yet Deo eaeal 2 50 30 1 “ “ == 0.7 * “ + 0.3 “ “ 5 30 hrs. 3 1 “ “ se ay | ce EE =. 0.3 “ “ 10 min. 0 Litton dae 0.8 i tree eta 2 35 10 1 “ “ ag “6 ae O30 “ 2 5 30 1 « “ = D.7 6 66 4. 0.3 * “ 3 0 1“ DFIPI+0.7 “ “ 40.3 “ DCp 1 45 5 i “ + 0.7 “c “ + 0.3 “ “ 35 10 1 “ “ a 0.7 * “c “1.3 “ “ 25 CW 15 1“ “ = Wy re =I Ose “ 28 30 1 “ “ + 0.7 “ “ + 0.3 “ “ 1 20 hrs. 3 1 “ « + 0.7 “ “ + 0 3 it3 “cc 4 50 min. 0 ee COE DFE ME Se a 45 10 fue “ ae On7m Serpe ire “ “ 35 30 ( Iees “ Sag “ “ “fies “ “ 55 O. iat ee: DIE eset POD alata DS 3 10 6 1 “ == 0.3 “ “ Beye Taek “ 2 20 O}1“ “ +0.3 “ heated (56°) DCp+0.7¢ec. DS} 8 15 6 1 “cc “cc + 0.3 “ “cc (56°) “cc +0.7 “ “ 6 35 Ptes sinekig ams 4 10 1 Spode heb yiia DOp 13 20 oj1“ “ +0.7 “ DS+0.3 cc. DCp 3 20 2 1 “ce “ a 0.7 “< “< + 0.3 “<< “ 2 10 1 “ “ + 0.7 “ “c = 0.3 “ “ 1 40 PERO TOE gg ECB 6 15 O Ap Des “« +0.7 “ DS + 0.3 ec. CCp 2 30 9 1 “ “ ae e-/ “ “ INES “ “ Pes 10 1 “ “ fe 0.7 “ “c ay 1p BG: “ 50 30 1 “ “ SE ey 9.3 ¢ “ 45 498 L. Loeb, M. 8. Fleisher, and L. Tuttle 499 Mixture. Coagulation time. hrs. min. sec. 1 ce. DHP! + 0.4 ce. DCp 32 10 1 “ “ — 0.3 “ “ 23 30 1 “* “ a 0.2 “ “ 20 40 i :°6 « + 0.4 “ heated (56°) DCp 1 23 15 A a Fet-Ocal < < (567) 3S 42 Beginning coagula- tion. 5 16 Coagulated. T:. a + 0.2 “ “ (565) 31 Beginning coagula- tion. 5 16 Coagulated. Time of Coagula- stand- Mixture. tion ing of rs time. $+Cp. min. min. 8c. 0 1 cc. DHP!I + 0.3 cc. DCp + 0.7 cc. DS 9 45 6 1 % “ iS “ aig) G7 “ “ 9 55 0 Dy ff “« +0.3 “ heated (56°) DCp + 0.7 cc. DS 38 2 1 “ “ -- 0.3 “ “ (56°) “ — Oz “ “ 40 6 1 “ “ + 0.3 “<< “ (56°) “cc + 0.7 “ “ 40 On the Mechanism of Inhibition. We carried out a few additional experiments in order to deter- mine more definitely the mechanism underlying the inhibiting effect or the loss in coagulating power which occurs when the mixture of serum,and extract is allowed to stand for a short period of time. If instead of letting a mixture of 0.7 cc. of DS + 0.5 cc. of DKE stand at room temperature, we keep it on ice for 10 minutes and then add the mixture to 1 cc. of RHPI, the inhibition is very much diminished. The mixture added to the HP! at once caused coagulation in 64 minutes; after 10 minutes - standing at room temperature, the coagulation was prevented indefinitely. But after 10 minutes standing on ice, the coagula- tion time was 23 minutes. Some inhibiting substance had been formed, but it was much less than at room temperature. The 500 Coagulation of Blood. II prolongation of the coagulation time, after the mixture had been standing on ice, was not entirely due to the formation or action of an inhibiting substance, but partly at least to the lowering of the temperature in the mixture which persisted for some time after the plasma had been added. From this experiment we may conclude that very probably a chemical reaction takes place between the inhibiting constituents of the serum and of the tissue extract and that this reaction is retarded by the lowering of the temperature. If a combination of DKE + DS after standing for a short time no longer causes coagulation of RHPI, the addition of 0.5 cc. of new DKE to the mixture of DHPI, DKE, and § causes coagulation, which occurs either promptly or with a slight delay. We may, therefore, conclude that in the DKE + DS on standing no marked excess of inhibiting substance had been produced such as would have, been able to neutralize an additional amount of Es. If instead of combining DIE with DS we use with the DKE an amount of DFIP1 equal to the DS in the usual mixture, and allow the mixture to stand for 7 minutes before adding it to the HPI, a retardation in the coagulation of the DHPI occurs which, in our experiment, amounted to 100 per cent and which was prob- ably due to the inhibiting effect of the fluoride as such, but in contradistinction to the mixture of DKE + S§S, this inhibition does not increase through standing for 7 minutes of the mixture of DFIP1 + DKE. In this case the interaction between TE and the DFIP1 led to the production of coagulating substance which counteracted the inhibiting substance which may perhaps have formed in a mixture of DKE + DFIPI! on standing. In another experiment we determined whether it is possible to withdraw from:the serum the inhibiting substance by shaking it with the residue of finely divided kidney tissue which had previously been used for the extraction of tissue coagulins. 20 cc. of DS were shaken with this residue of kidney tissue for 1 hour. Such serum when separated from the kidney particles — and allowed to stand with fresh DKE during the usual time proved still very inhibiting. The inhibiting substance had, therefore, not been removed from the serum through this procedure. L. Loeb, M. S. Fleisher, and L. Tuttle 501 SUMMARY AND DISCUSSION. 1. With hirudin as well as with fluoride plasma the curves representing the coagulation time as a function of the quantity of tissue extract added are irregular, a medium quantity usually being optimal. The specific adaptation of the tissue coagulins to the plasma can be demonstrated not only with unaltered blood or plasma, but also with hirudin plasma and in certain cases with fluoride plasma. The inhibiting effect of serum may in some cases obscure the specific adaptation of the tissue coagulins in a mixture of extract and serum. 2. Heating extract to 56° very often, produces an increase in its coagulating power, if the extract is added to the plasma in certain proportions. This improvement in all probability is due to a decrease in the inhibiting power caused by a moderate de- gree of heat which is greater than the injury to the coagulating substance caused by heating. Those extracts are especially benefited through heating in which the inhibiting substance prevails in the unheated extract. Heating the extract to 60° injures its coagulating power more than its inhibiting power if it is used uncombined with plasma. In combination with serum an extract heated to 56°, if tested with hirudin or peptone plasma, also shows a great diminution in inhibiting effect which becomes especially distinct if the mixture of serum and heated extract are allowed to stand for a few minutes. The use of sera and ex- tracts under conditions which relatively increase the importance of the inhibiting effect of the tissue extract and serum increase part passu the beneficial effect of the heating of the extract. A mixture of serum and of extract heated to 60° tends to become similar to a mixture of serum and unheated extract. With fluor- ide plasma a mixture of serum and extract likewise becomes more efficient through heating the extract to 56°. Heating the extract to 60° diminishes the coagulating power of the mixture as compared with the mixture of serum and unheated extract. The differences between the effects of the mixtures on hirudin and fluoride plasma can be explained if we assume that the inhibiting substance formed is of relatively more importance with the hirudin and peptone plasma and the coagulating substance with the fluoride plasma. 3. The curves representing the coagulating power of the stroma of erythrocytes as a function of quantity are similar to those of 502 Coagulation of Blood. II tissue extract; the curves of different species differ. It is possible not only to demonstrate a relative class specific adaptation, but even a species specific adaptation between corpuscles and plasma. While organ tissue extract and serum are about equally active towards DHPI and DPPI, corpuscles were found much less active with DPPI than with DHPI. In contradistinction to what we found in the case of mixtures of tissue extract and serum, mixtures of Cp and § usually show an acceleration, if added at once to hirudin or fluoride plasma; on standing there is in most cases a gain, and only very gradually a decrease occurs in the coagulating power of the mixture. There is, therefore, a marked difference between the behavior of such mixtures and mixtures of TE and 8S. These differences can be explained if we assume that the inhibiting substance which com- bines with a component of the serum is lacking in erythrocytes, or at least much weaker than in the tissue extract. Thus the opposite reaction which leads to an acceleration preponderates for some time. Heating corpuscle stroma merely injures the coagulating substance, the tissue coagulin. It does not lead to an improvement as heating of the tissue extract does, because in Cp the inhibiting component is absent or at best present only in small amounts. Our conclusion is confirmed that even a moder- ate heating injures the tissue coagulin and that the apparent lack of effect of heating is due to the simultaneous injury to an inhibiting substance. Through combination with serum, heated stroma of corpuscles improves relatively to the same extent or somewhat less than unheated corpuscles. As in the case of the unheated corpuscles, this improvement is still more marked after the heated corpuscles have been standing for some time. The coagulating. power of the heated mixture is smaller than that of the mixture with the unheated corpuscles. Inasmuch as in the corpuscles the inhibiting substance is much less prominent than in tissue extract, and yet the irregular curves of the uncombined corpuscles in their action on plasma resemble those of the uncombined tissue extract, these irregular curves cannot be altogether due to the same inhibiting mechanism which acts in combination with serum. 4. The evidence points to the conclusion that a chemical com- bination forms between a constituent of the tissue extract and the L. Loeb, M. S. Fleisher, and L. Tuttle 503 serum which leads to the inhibition in the coagulation of the plasma, and that a lowering of the temperature increases the reaction time. ‘Two processes are taking place simultaneously, if blood serum and extract are mixed. (a) Interactions take place which lead to the formation of a substance accelerating the coagulation of the blood. This is shown by the optimum in the time curve of hirudin and fluoride plasma, if stroma of erythrocytes is combined with serum, and in the optimum of the time curve of fluoride plasma if tissue extract is used in combination with serum. (b) An inhibiting substance is formed through the combination of a constituent of the serum and of the tissue extract. Both these constituents are injured by a moderate degree of heat. In different blood sera the quantity and character of this inhibiting substance vary. It is usually present in large quantities in blood serum of the dog. The quantity and presumably also the char- acter vary apparently in the tissue extracts in different species, and there are also variations in different tissues of the same species. In the mammalian red blood corpuscles, this inhibiting substance is found, if at all, only in very small quantities, while in the kid- - ney and probably also in the spleen of dog it usually is found in large quantities. Preliminary experiments which ought to be extended seem to indicate that there are also variations in the strength of the inhibiting substance in other organs of the same species. We find accordingly, if one kind of blood serum is com- bined and incubated with various kinds of extracts that the results vary and that certain combinations lead to the production of an inhibiting substance to a much higher degree than others. Ac- cordingly we find that those sera and extracts which are most inhibiting lose absolutely more of their inhibiting power through a moderate heating than sera or extracts which have less of the inhibiting substance. We find, furthermore, that through a graded moderate heating of the tissue extract the coagulating power of the tissue extract alone, as well as of the combination of serum and tissue extract, can be increased. This is due to the fact that the inhibiting sub- stance is slightly more sensitive to moderate heat than the coagu- lating substance; the latter is, however, also distinctly heat labile. Heating the serum diminishes both the quantity of the inhibit- ing and coagulating substances and usually an increase in the 504 Coagulation of Blood. II coagulating power of the serum cannot be produced through graded moderate heating of the serum, although in a few experiments we have seen indications that possibly a slight improvement may be procurable under certain conditions. Through graded heating it is possible to separate the tissue coagulin from the inhibiting substance. We may interpret these facts by assuming that both in the serum and the tissue extract there are substances present which unite to form the inhibiting substance. The substances in the uncombined serum and extract are apparently only the precursors of the active inhibiting substance.* As to the mode of action of this inhibiting substance, whether in some way it combines with the newly formed thrombin and thus inactivates the latter, or whether it acts in other ways, must be left undetermined at present. ; 5. In our earlier experiments we concluded that the tissue coagulins are specifically adapted to a substance in the blood plasma, while in the serum no evidence of a specific adaptation of the coagulating substance can be found.! Subsequently when we observed the formation of an inhibiting substance in the mix- ture of serum and extract, we pointed out that the presence of the inhibiting substance might possibly obscure the specific adaptation of both the coagulating and inhibiting substances in the serum and that there are indications which indeed suggest such a specific adaptation of both the accelerating and inhibiting substance.®> Our present results are quite compatible with such an interpretation, although further experiments are necessary in this direction. 6. Our experiments show that while the character of the sub- stances admixed to the tissue coagulins varies greatly in differ- ent extracts and in the erythrocytes, the specific adaptation of the tissue coagulins is not altered thereby. This fact makes extremely improbable the assumption that the specific adapta- tion of the tissue coagulins is due to the admixture of other substances. 4It would be very desirable to extend the study of the effect of graded heat to other extracts; we intend to carry out these and other experiments as soon as possible. 5 Loeb, L., Beitr. chem. Physiol. u. Path., 1904, v, 534; 1907, ix, 185. ‘L. Loeb, M. 8. Fleisher, and L. Tuttle 505 We may draw the further conclusion that such a specific adapta- tion would not be compatible with the opinion of several investi- gators according to which the tissue coagulins are of a lipoid character. From all we know at the present time such a definite specificity is only found in substances of a protein nature, and accordingly we have assumed that there must.be present in the tissue coagulins a protein component; but inasmuch as the studies of Bordet and Delange,® Howell,’ and Zak® undoubtedly proved a coagulating effect of lipoid substances, we formed the opinion that possibly the combination of a protein and lipoid substance might constitute the tissue coagulins. The recent studies of Mills’ likewise point to this conclusion. 7. While our experiments prove that an interaction between a constituent of tissue and blood serum leads to the formation of a substance accelerating the coagulation of the blood, and in so far confirms the conceptions of Morawitz!® and others, the possibility is not thereby excluded that in addition there may be a direct interaction between the tissue coagulins and the coagul- able substance of the plasma. In invertebrate blood our studies have made such an interaction very probable and there undoubted- ly exist very marked analogies between the coagulation of the invertebrate and vertebrate blood.!! Furthermore, certain facts in the coagulation of vertebrate blood likewise suggest such an interpretation. It is therefore possible. that we have not only to deal with different kinds of active components in cells, such as the protein and lipoid constituents, but even with different proc- esses whereby the tissues accelerate the coagulation of the blood. 8. It is well known that coagulating substances are not only found in the blood serum and tissue extracts, but also in other material. Thus we found previously a very strong coagulating substance in cultures of the staphylococcus which is destroyed 6 Bordet, J., and Delange, L., Arch. exp. Path. u. Pharmakol., 1913, Ixxi, 293. 7 Howell, W. H., Am. J. Physiol., 1912-13, xxxi, 1. 8 Zak, E., Arch. exp. Path. wu. Pharmakol., 1912, lxx, 27. 9 Mills, C. A., J. Biol. Chem., 1921, xlvi, 135. 10 Morawitz, P., Deutsch. Arch. klin. Med., 1903-04, Ixxix, 1; Bettr. chem. Physiol. u. Path., 1904, v, 133. 11 Loeb, L., Biochem. Centr., 1907, vi, 829. 506 Coagulation of Blood. II by heat.!2 It would be of interest to determine whether inhibit- ing substances in the blood serum and tissue extract are able to interact also with these substances. In regard to the inhibiting component of the serum, we have made it very apparent that it disappears from the blood as a re- sult of phosphorous poisoning.’ In the tissues we found previously fibrinolytic substances; whether or not they have some relation to the inhibiting component of the tissue extract is uncertain. CONCLUSIONS. The tissues contain constituents which can combine with a substance in the blood serum and thus lead to the production of a substance inhibiting the coagulation of the blood. This sub- stance differs in its action on fluoride plasma on the one hand, and on hirudin and peptone plasma on the other. The quantity of these substances present in various kinds of cells varies greatly and they seem to be either absent or present only in small quantities in the erythrocytes. ‘These substances on the presence of which depends the inhibiting effect of the blood serum can be separated from the tissue coagulins through graded heating; the tissue coagulins, as well as the inhibiting component of the tissues, are injured through moderate heat. Both the inhibiting component in the blood serum and that in the tissues seem to be similarly affected by heat. The specific adaptation of the tissue coagulins remains con- stant, however much the inhibiting substances in the cells may vary in quantity. This characteristic resides, therefore, in all probability in the tissue coagulins proper and is not dependent on admixtures. The specific adaptation of the tissue coagulins makes it very probable that a protein component is concerned in the coagulat- ing effect of the tissue extracts. 12 Loeb, L., J. Med. Research, 1903-04, x, 407. 13 Fleisher, M. S., and Loeb, L., J. Biol. Chem., 1915, xxi, 477. THE UNSATURATED FATTY ACIDS OF EGG LECITHIN. By P. A. LEVENE anv IDA P. ROLF. (From the Laboratories of The Rockefeller Institute for Medical Research.) (Received for publication, February 6, 1922.) The work on liver lecithins by Levene and Ingvaldsen' and by Levene and Simms,? has shown the existence of lecithins which are distinguished from one another by the nature of their unsatu- rated fatty acids. There were isolated from liver lecithin, two unsaturated fatty acids, oleic and arachidonic, as well as two satu- rated acids, palmitic and stearic. Hence on the basis of the present findings it is permissible to accept the existence in the liver of four lecithins. Further work may lead to the discovery of an even greater number. The work on the liver lecithins suggested the need of a reinves- tigation of the lecithins of other organs with a view of finding in them also lecithins containing acids of a higher order of un- saturation than that of oleic acid. ; The present communication contains the report of the work on egg lecithin. This “lecithin” also was found to contain more than one unsaturated fatty acid. Three acids were isolated; namely, oleic, linolic, and arachidonic. Oleic acid was identified by its iodine number and by its hydrogenation product, stearic acid. Linoleic acid was isolated in small quantities only and | was identified as its tetrabromide. Arachidonic acid was identi- fied by its octabromide and by its hydrogenation product, ara- chidic acid. Comparing the mixed lecithins of liver and those of the egg yolk, one is struck by the difference in the proportions of the individual forms in them. The liver lecithin contains a very large proportion of the forms with highly unsaturated fatty acids, whereas in the egg yolk the proportion of the latter is small. 1 Levene, P. A., and Ingvaldsen, T., J. Biol. Chem., 1920, xliii, 359. 2 Levene, P. A., and Simms, H. S., J. Biol. Chem., 1921, xlviii, 185. 507 THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. LI, NO. 2 508 Fatty Acids of Egg Lecithin Thus, the iodine numbers of the cadmium chloride salts from liver lecithin varied between 59 and 84, whereas those from egg leci- thins varied between 30 and 54. It is noteworthy that individual samples of egg “lecithin” differed considerably from one another. An effort is being made to develop a method for fractionation of individual lecithins. EXPERIMENTAL. A. Isolation of the Unsaturated Acids. The source of the unsaturated acids was “lecithin cadmium chloride” free from amino-containing impurities. These cad- mium chloride salts were prepared from both the acetone and ether extracts of egg powder and were purified by the methods described in previous papers.’ Decided variation was observed among individual samples obtained either from different lots of egg powder or by different methods of extraction. However, it was not possible to establish a definite relationship between the composition of the material and these variable factors. The fatty acids were liberated by hydrolyzing the cadmium chloride salt for 6 hours with a 10 per cent solution of hydrochloric acid. On cooling, the mixed fatty acids separated as a solid cake. They were then dissolved in ether and thoroughly washed with water. The iodine numbers (method of Wijs) of several such whole fatty acids, are given below. These acids were derived from salts which were obtained from different lots of egg powder, as well as by varying the methods of " extraction. No. 123 (from an ‘‘acetone extract’). 0.3114 gm. substance absorbed 0.2839 gm. iodine corresponding to an iodine number of 91. No. 117 (from an ‘‘ethereal extract’’). 0.3465 gm. substance absorbed 0.2839 gm. iodine corresponding to an iodine number of 82. No. 301 (from an “‘ethereal extract’’). 0.2798 gm. substance absorbed 0.2102 gm. iodine corresponding to an iodine number of 75. After drying the ethereal solution and concentrating the solvent, a crude separation of the saturated from the unsaturated acids was effected by crystallization from acetone. The mother liquor 3 Levene, P. A., and Rolf, I. P., J. Biol. Chem., 1921, xlvi, 195. P. A. Levene and I. P. Rolf 509 contains all the unsaturated, with a slight admixture of saturated acids. A more complete separation was accomplished through the insolubility of the barium and lead salts of the saturated acids in ether. B. Separation of Acids of Varying Degrees of Unsaturation. Two methods were used in an attempt to separate the constit- uent acids. The first procedure was based upon the greater solubility of the barium salts of the more highly unsaturated fatty acids in a benzene-alcohol mixture. For this purpose Ba(OH). was added to the mixed acids in a methyl alcoholic solution, until the solution was just alkaline to phenolphthalein. The mother liquor was decanted from the gummy precipitate, and concentrated to dryness under diminished pressure. The residue, combined with the original precipitate, was dissolved in a mixture of warm benzene and 5 per cent by volume of 95 per cent ethyl alcohol. To this, after cooling, an additional equal volume of ether was added. On standing over night in the ice box, barium oleate, together with the barium salts of any unsaturated acids present, was precipitated and was easily removed by filtration. C. Purification of Oleic Acid. The precipitated barium salts were decomposed with hydro- chloric acid, and the liberated acids converted into their lead salts. The solubility of the lead salts of the unsaturated acids in cold ether, permitted the removal by filtration of the last traces of saturated acids. The unsaturated acid, obtained from its lead salt by treatment with hydrochloric acid, was a light yellow liquid which had an iodine value of 89, corresponding to oleic acid. No. 307. 0.2578 gm. substance absorbed 0.2283 gm. iodine. CisH34O2. Calculated. Iodine value 90. Found. h6 a5o. After hydrogenation by Paal’s method the resulting saturated acid (No. 307) was analyzed and determinations of the molecular weight and melting point were made. ‘ Farnsteiner, K., Z. Untersuch. Nahrungs- u. Genussmittel., 1899, ii, 1. 510 Fatty Acids of Egg Lecithin All samples of saturated fatty acids reported in this paper were dried by fusion on an electric hot-plate, and the material used for combustion was remelted under diminished pressure at the temperature of xylene vapor, until constant weight was ob- tained. The bromo acids were dried under diminished pressure at the temperature of boiling chloroform. The melting points as recorded, are corrected, and were taken at such a rate that the time per degree rise in temperature was 6 seconds. The molecular weights were calculated by the titra- tion of approximately 1 gm. of acid, dissolved in 10 cc. of toluene and 25 ce. of methyl alcohol (neutral to phenolphthalein) with 0.5 n NaOH, using phenolphthalein as an indicator. No. 307. 0.1000 gm. substance: 0.1176 gm. H.O and 0.2786 gm. COs. 0.8465 “ i required for neutralization 6.00 ce. 0.5 n NaOH. CisHsg02. Calculated. C 75.98, H 12.72. (Molecular weight = 284. Melting point = 70-71°C.) Found. C 75.97, H 13.14. (Molecular weight = 282. Melting point = 71°C.) D. Acids of a Higher Degree of Unsaturation. The barium salts soluble in the mixture of benzene-alcohol- ether, as described under Section B, were decomposed with hy- drochloric acid. The liberated acid, a dark brown liquid, had an iodine value of 165. 0.2055 gm. substance absorbed 0.3392 gm. iodine. This acid after reduction by Paal’s method gave the following analytical data. No. 309. 0.1008 gm. substance: 0.1192 gm. H.O and 0.2832 gm. CO. icvoco ~~ required for neutralization 11.70 ce. 0.5 n NaOH. CisH3s02. Calculated. C 75.98, H 12.72. (Molecular weight = 284. Melting point = 70-71°C.) GooH4902. Calculated. C 76.95, H 12.81. (Molecular weight = 312. Melting point = 75-77°C.) Found. C 76.61, H 13.21. (Molecular weight = 299. Melting point = 66-67°C.) P. A. Levene and I. P. Rolf 511 E. Fractionation of the Methyl Esters. For the fractional distillation of the methyl esters the crude unsaturated acids were converted into their lead salts. The ether-soluble fraction was decomposed with hydrochloric acid and the free acids were esterified. These methyl esters were then saturated with hydrogen by Paal’s method and again esterified. The recrystallized esters were distilled at a pressure of 3 mm. into four fractions, each of which was then saponified and the free acid purified by conversion through the lead salt. The molec- ular weights, melting points, and analyses of the respective acids are recorded below. No. 312. 0.1002 gm. substance: 0.1162 gm. H.O and 0.2750 gm. COs». 1.01382 “ ef required for neutralization 7.18 ce. 0.5 n NaOH. No. 314. 0.1012 gm. substance: 0.1182 gm. H,O and 0.2820 gm. COsz. 1@i6 “ + required for neutralization 7.22 ce. 0.5 n NaOH. No. 315. 0.1000 gm. substance: 0.1144 gm. H2O and 0.2762 gm. COz. No. 316. 0.0970 gm. substance: 0.1172 gm. H,O and 0.2698 gm. COs. 0.7858 “ ry required for neutralization 5.15 ce. 0.5 n NaOH. CisHz;,02. Calculated. C 75.98, H 12.76. (Molecular weight = 284. Melting point = 70-71°C.) C2oHs902. Calculated. C 76.95, H 12.98. (Molecular weight = 312. Melting point = 75-77°C.) Boiling point Analysis of acid. | Weight Molecu- No. oe “ ea of cnt anid 3 mm. Cc H a. acid. st OF gm. "Ge Found. 312 178-182 | 74.88 | 12.98 ee 282 70-71 a 314 ts —tee Wo.98. |-13.07 | 5.3 2 70-71 S 315 177-185 | 75.31 | 12.80} 4.6 62-63 ; ad 316 184-195 | 77.03 | 18.30 | 2.5 308 75 F. Bromine Addition Products of the More Highly Unsaturated Acids. > The acids, whose barium salts were soluble in benzene-alcohol- ether, were dissolved in 10 parts of glacial acetic acid. To the cooled solution a 10 per cent solution of bromine in glacial ace- 512 Fatty Acids of Egg Lecithin tic acid was added very gradually. The bromination was accom- panied by a characteristic color change and the appearance of a very finely divided amorphous precipitate. After standing over night the latter was separated by centrifugalization and extracted with warm ether until further extracts were colorless. On dry- ing, the material darkened slightly. When heated in an open tube it browned gradually above 215°, and contracted at 250°. It neither melted nor did it decompose with gas evolution. In a closed tube, this material after recrystallization from glacial acetic acid, contracted at 250°. It darkened decidedly above this point, and sintered very definitely at 255°. By the hydrolysis of 380 gm. of a lecithin cadmium chloride (No. 123) whose iodine value was 54, mixed acids were obtained which had an iodine value of 91. From these, 60 gm. of acids, whose barium salts were soluble in a benzene-alcohol mixture, were isolated. These acids on bromination yielded 4.5 gm. of an octabromide insoluble in ether, but very slightly soluble in hot glacial acetic acid. Analyses of two products obtained in this manner are given below: No. 336. 0.0932 gm. substance: 0.3000 gm. H2O and 0.0864 gm. CO:. 0.2068 “‘ te 0.3304 “ AgBr. No. 304. 0.1024 gm. substance: 0.0326 gm. H2O and 0.0942 gm. CO:. No. 362. 0.1846 “ a 0.2920 “ AgBr. CooH3202Brs. Calculated. C 25.48, H 3.42, Br 67.72. No. 3836. Found. shy Sy’. amas | a No. 304. s “95.08; “ 3.56, °° iGicae The bromination liquor, after the removal of the octabromide was concentrated to dryness under diminished pressure. The residual syrup was dissolved in ether and washed with sodium thiosulfate. From the concentrated ethereal solution, aggre- gates of spear-like needles were deposited. From the bromina- tion of the acids of No. 123, details of which were given above, about 8 gm. of this crude material were obtained. This material was readily soluble in ether and absolute alcohol, but was insolu- ble in gasoline. After repeated recrystallization from various solvents, two samples, obtained respectively from an “ether” and an ‘“‘acetone” extract, gave the following analyses, molecular weights, and melting points. P. A. Levene and I. P. Rolf 513 No. 355. 0.1034 gm. substance: 0.0512 gm. H.O and 0.1368 gm. COz. 0.2012 <“ se used for Carius determination: 0.2596 gm. AgBr. 0.8401 “ * requires for neutralization 13.70 cc. 0.1 N NaOH, corresponding to a molecular weight of 617. In an open tube it softened at 115° and melted at 117-119°C. No. 307. 0.1012 gm. substance: 0.0476 gm. H.2O and 0.1340 gm. CO:. 0.2010 “ as 0.2526 “ AgBr. 0.5412 “ + required for neutralization 9.00 ce. 0.1 N NaOH, corresponding to a molecular weight of 601. Melted in an open tube at 116-117°C. C,sH3202Br;. Calculated. C 36.01, H 5.38, Br 53.28. (Molecular weight = 600.) No. 355. Found. C 36.08, H 5.54, Br 54.22. (Molecular weight 617. Melting point = 115-116°C.) No. 307. Found. C.36.10, H 5.22, Br 53.48. (Molecular weight = 601. Melting point = 115-116°C.) Tetrabromostearic acid was prepared by brominating the acids of cottonseed oil. This material when heated, contracted at 114° and melted at 115-116°C. When mixed with these tetra- bromo acids from lecithin, no depression of its melting point was apparent. INDEX TO VOLUME LI. CID, amino-, nitrogen content of the blood, studies on, 121 ——, ——, —— in blood, new colori- metric method for the determi- nation, 377 —_, —, —— — normal urine, colorimetric determination, 393 ——, earbon dioxide, and oxygen in blood, interactions, 359 ——, carbonic, and bicarbonate in urine, 295 ——, homogentisic, in urine, colori- metric method for the deter- mination, 453 ——.,, uric, in blood, determination, 187 Acids, amino-, in cow’s milk, 165 ——, ——, retention and distribu- tion, with especial reference to the urea formation, 395 —., phenylacetic and phenylpro- pionic, on the distribution of nitrogen in the urine, influence, 141 ——, unsaturated fatty, lecithin, 507 285 Adsuki bean, Phaseolus angularis, chemical study of the proteins, 103 Amino-acid nitrogen content of the blood, studies on, 121 —— —— in blood, new colorimetric method for the determination, 377 —— —— —— normal urine, colori- metric determination, 393 Amino-acids in cow’s milk, 165 ——, retention and distribution, with especial reference to the urea formation, 395 of egg liver lecithin, Ammonia content of blood, note on, 183 —— in blood and in organic fluids, micro method for the estima- tion, 367 Analyses, separate, of the corpus- cles and of the plasma, 21 Analysis, blood, a system, 377 ——, , note on the necessity of checking up the quality of sodium tungstate used in the system, 419 Aral, Mrnorvu. See Oxapa and Aral, 135 Artichoke tubers, preparation of inulin with special reference to, as a source, 275 Asparagus officinalis L., carbohy- drate content of the seed, 93 BARTLETT, H.H. See Caxe and BaRTLETT, 93 Base, total, in urine, method for the estimation, 55 Bean, adsuki, Phaseolus angularis, chemical study of the proteins, 103 ——, soy, the coconut, and cotton- seed meal, nitrogen distribu- tion of proteins extracted by 0.2 per cent sodium hydroxide solution from, 17 BEELER, Carou. See WILDER, Bootusy, and BEELER, 311 BENEDICT, STANLEY R. The deter- mination of uric acid in blood, 187 See Nasu and BENeEpict, 183 BERGLUND, Hitping. See Four and BERGLUND, 209, 213, 395 Bicarbonate and carbonic acid in urine, 295 515 516 Biological test, delicate, for cal- cium-depositing substances, 41 Bios and water-soluble B in yeast growth, 77 Blood, amino-acid nitrogen in, new colorimetric method for the determination, 377 —, ammonia content of, note on, 183 analysis, sodium tungstate used in the system of, note on the necessity of checking up the quality, 419 —— ——,, system, 377 —., chlorides in, clinical method for the estimation, 115 , coagulation, comparison between the effects of the stroma of erythrocytes and of tissue extracts, unheated and heated, on, and on the mechan- ism of the interaction of these substances with blood serum, 485 3 , interaction between blood serum and tissue extract in, 461, 485 ——, determination of uric acid in, 187 ——, free and bound water in, 435 —, interactions of oxygen, acid, and carbon dioxide in, 359 — and organic fluids, ammonia in, micro method for the esti- mation of, 367 , secretions, and tissues, micro urease method for the estima- tion of urea in, 373 serum, mechanism of the inter- action of stroma of erythrocytes and of tissue extracts with, and comparison between the effects of these substances, unheated and heated, on the coagulation of the blood, 485 —— —— and tissue extract, inter- action between, in the coagula- tion of the blood, 461, 485 Index Blood, studies on the amino-acid nitrogen content, 121 BoorTusy, WALTER WILDER, BootueBy, BEELER, 311 Briaes, =” « ae iy ite BIND “, MAY 2 4 66 QP The Journal of biological 501 chemistry ae CEB & Medical Serials PLEASE DO NOT REMOVE CARDS OR SLIPS FROM THIS POCKET UNIVERSITY OF TORONTO LIBRARY ogee mies ~ ae to) fe: ed wip ste: i Tig tsahe 5 ie eer eos $ Sigh mpboe wre * Weta ees foe etet eee te 0 eh dale tot Prtae new, steaks eh ue tet: de toe" ecw 92 ode peener aes, Her? / Ree onbye - Le ye eatin ae Sep omeee 5 reas! Oe ed - Dipckelene = trends Imre Rims * eo7e> whe oe Dy Seer bet bap ab ceene he bel weed ehnt, A ab perk Pee it yeeros fis ret arora g eu an sta PS cas ie Die) Fr web ow =~) > rb My benee Phd poh bat RP ees eee ete A=) 7 BA Pete ale « teers oR ah 4 7) bent Hera rae: Sats rshee x eats Oe + br ems Srecie desteten nhs -- > Fee er as AE St =e bere Bt Thevprin hie > mPa Dt e774) reheated ia brde ie eae meet «yh Sh Fre g a bebeey ’ fees Cis beach d 7 areas Ceprcsteh Rael rp ahacy Ad ‘a Edens % res Tee Pea te Le Earl oe yi PA pom ¢ orl radia b St Saleh pers te? ph int Sines pA sree © ans bag toe m7 2 heuetam arts = $e Sat i be oo ee Ay 4 ni » nde cargepeee eeepee es a satis ae ee epee Raped STnh atr esi asin ¢ o-4: sogbeaeaee : atag STewtnraipe: ; Hey betes) ; G Eerie S A é on aia bite oe abs toes epee