Nov., 19511 DETERMISATIOS OF CARBOX DIOXIDE 623 The Determination of Amino-Acids Colorimetrically the Ninhydrin Reaction BY A. M. SMITH AND A. H. AGIZA The coloured compound produced in the reaction between amino-acids and ninhydrin can be used as a quantitative measure of the amino-acid decomposed. The depth of colour is not the same for all the common amino-acids, but the results are reproducible to within 3 per cent. and the curve is linear for all the acids examined in the range 5 to 25 pg of amino nitrogen. Hence the reaction provides a convenient method of determining the sub-micro quantities of amino-acid obtained from a drop of a protein hydrolysate fractionated by two-dimensional chromatography on paper. IN the reaction between amino-acids and ninhydrin (triketohydrindene hydrate), each of the products, aldehyde, ammonia, carbon dioxide and the coloured compound, can provide a convenient measure of the amount of decomposed amino-acid.The determination of a volatile aldehyde can be carried out,' but the method is, of course, applicable only to certain acids; ammonia may be detemr~ined,~~~ but apart from the fact that several amino-acids do not give the theoretical amount, the results are not as accurate as those obtained by determining the carbon dioxide. The liberation of carbon dioxide is highly specific and its determination is sufficiently precise for assessing micro amounts of all free a-amino-acids and a number of closely related c o m p ~ u n d s . ~ ~ ~ ~ ~ ~ ~ Nevertheless, it would be convenient in many circumstances if amino-acids could be measured in terms of the coloured compound formed in the reaction.Ruhemanns was the first to examine the nature of the blue or purple compound and considered that it was formed by the condensation of the reduced ninhydrin with the ammonia liberated from the amino-acid and with the excess of ninhydrin. Harding and MacLeanO devised a quantitative technique that gave accurate results for several amino-acids in the range 5 to 50 pg of nitrogen per ml. Later they and Warneford showedlotll that the colour reaction was not specific for a-amino-acids but also occurred with ammonium salts, peptides and certain classes of amines. These compounds, however, are not as reactive as amino- acids, which alone are sensitive, in the absence of a base, when the concentration of nitrogen is less than 0.1 mg per ml of solution.The question of specificity does not arise in the deter- mination of individual amino-acids separated and identified by a chromatographic procedure. Wieland and Wirth,12 for example, have used the method to determine aspartic and glutamic acids separated in a column of aluminium oxide, and Moore and Stein13J* have fractionated the amino-acids in synthetic mixtures and in protein hydrolysates on a starch column and measured their concentrations in the eluate. The mechanism of the reaction has received a good deal of attention, and recent spectro- photometric studies16,16 of ninhydrin and various derivatives have indicated that the purple colour is due to the anion of diketohydrindamine-diketohydrindylidene, the red and blue624 SMITH AKD AGIZA: THE DETERMISATIOX OF AMISO-ACIDS [Vol.7ti components being respectively the mono- and di-valent anions of indanone-enediol. Prolinck and hydroxyproline, from which ammonia is not produced by reaction with ninhydrin, arc exceptional in giving a yellow or red compound.17 The depth of colour is not constant for equivalent amounts of different amino-acids, so the method is not accurate for the determination of the total amino nitrogen in a mixture, but this does not present any difficulty with individual acids, since a standard curve can btb prepared for each and, with proper precautions, the results are reproducible without difficulty and the curve is linear over the range 5 to 25 pg of nitrogen. Differences in shade of colour are probably due to the condensation of the aldehydes produced from different acids with the 1 :3-diketohydrindamine to give the orange - brown dyes (1 :3-diketo-2-arylidene-hydrind- amines) described by Ruhemanns and referred to more recently by Atkinson, Stuart and Stuckey.18 In the course of the present investigation, it was observed that the colour was altered in both depth and shade when traces of aldehydes were added to mixtures of glycine or alanine or ammonium salts with ninhydrin.Moore and Stein13 were of the opinion that difficulties in obtaining a good correlation between amount of amino-acid and depth of colour were caused by oxidation changes and consequently they took precautions to exclude oxygen from the reactants and also added a reducing agent, first hydrindantin, which is formed on the reduction of ninhydrin, and later stannous chloride.In their comprehensive study of the conditions of the reaction they found that the absorption spectrum rose to a maximum value at 570 mp (440 mp for proline and hydroxyproline), and that maximum colour development occurred at pH 5 for all acids except tryptophan (pH 6) after 5 to 20 minutes’ heating at 100°C. These authors were especially interested in the recovery of the individual amino-acids in successive fractions of the eluate from a starch column, and they used three different solvents. They added the amino-acid solution to a mixed solution of ninhydrin in methyl Cellosolve and buffered stannous chloride in a photometer tube, raised the temperature to 100°C for 20 minutes and then shook the solution with a diluent of water and n-propano1 before measuring the colour absorption.In the present work several modifications were made in this procedure. For example, the amount of ninhydrin used by Moore and Stein, about 20 mg for amounts of amino nitrogen of the order of 3 pg, seemed to be an unnecessarily large excess; less than a tenth of that quantity was found to be adequate. I t was also found more convenient to dissolve the ninhydrin in the buffer solution and keep it separate from the stannous chloride than to keep a mixture of the two reagents under nitrogen. The method finally adopted is described below. REAGENTS- Citrate bufer, PH &Dissolve 21.008 g of citric acid (C,H,O,.H,O) in 200 ml of distilled water, add 200 ml of N sodium hydroxide and dilute to 500 ml; add 15 to 20 ml of n-butanol to prevent fungal growth. Ninhydrin solution A-Dissolve 550 mg of Iiinhydrin in 100 ml of citrate buffer.Ninhydrin solution B-Dissolve 100 mg of ninhydrin in 100 ml of n-butanol, saturated This solution is for spraying on to the paper chromatogram. Stannous chloride-Dissolve 0-5 g of stannous chloride in 250 ml of citrate buffer and This solution must be renewed METHOD with water. add a layer of n-butanol to the surface to reduce oxidation. at intervals of about 5 days. PROCEDURE FOR PURE AMINO-ACIDS- Place 1 to 2 ml of the solution of the amino-acid, containing 5 to 20 pg of amino nitrogen, in a test tube graduated at 5, 10 and 15 ml and neutralise if necessary with sodium hydroxide, using phenolphthalein as indicator. Add 1 ml of the buffer solution, or 2 ml if it has bee11 necessary to neutralise the solution, and then add 1 or 2 ml of ninhydrin solution A , according to the amount of amino-acid taken.Place the test tube in boiling water and add 1 ml of the stannous chloride solution. Continue heating for 15 minutes and then remove the tube from the boiling water and allow it to cool in the dark for 10 minutes. Make the solution UP to a volume of 10 ml with a saturated solution of sodium chloride. Add between 4 and 5 ml of n-butanol to the tube, stopper it tightly, shake and set aside for 5 minutes. By this treatment the colour is extracted from the aqueous phase and appears A red colour is first produced, but this changes to blue.Yov., 19511 COLORIMETRICALL\i BY THE SISHYDRIS REACTIOS 625 in the supernatant alcohol layer.Adjust the volume of the aqueous phase to 10 ml by adding water saturated with butanol, and adjust the volume of the upper layer to 5 ml by addition of butanol. With a pipette, transfer 3 to 4 ml of the clear butanol solution into a colorimeter tube and compare the purple colour with that of a blank prepared under the same conditions. 0 5 10 15 Amino nitrogen, p g 20 25 Fig. 1. Relationship between colour and amino nitrogen An Evans Electroselenium colorimeter was found to be satisfactory, with yellow filter No. 626 (maximum transmission at a wavelength of 570 mp), for all acids except proline and hydroxy- proline, which gave a yellow - red colour and for which violet filter No.621 (maximum transmission at 460 mp) was used. With cystine it was necessary to double the amount of ninhydrin in order to obtain reproducible results and a depth of colour similar to that from the other amino-acids. Cysteine yielded only about 15 per cent. of the colour given by an equivalent amount of leucine, an observation previously recorded by Moore and Stein.13 This may be due to the presence of the HS- group, and the following modification in the method was found to over- come the effect. A drop of phenolphthalein was added to the solution of cysteine (or cystine) in the tube and then enough sodium hydroxide to give a pink colour. One drop of dilute bromine solution was then added and this formed sufficient hypobromite to effect an oxidation to cysteic acid.The normal procedure for the other amino-acids was then followed, and the development of the colour was of the same order.626 SMITH A S D AGIZA: THE DETERMISATIOS OF AMISO-ACIDS [l'ol. TG RESULTS For each of twe?ty-three amino-acids, a series of concentrations ranging from about 4 to 20 pg of amino nitrogen was examined by the above method and the results were repro- ducible to between 1 and 3 per cent. of the mean of five determinations for all except glutamic acid (5 per cent.) and tryptophan and tyrosine (4 per cent.). A linear relationship between colour and quantity of acid was found over this range, and five typical curves are shown in Fig. 1. In common with other i n v e ~ t i g a t o r s , ~ l ~ ~ ~ ~ ~ ~ ~ ~ we found that equivalent amounts of different amino-acids did not give the same depth of colour.The relative values interpolated on each curve at 15pg of amino nitrogen are shown in Table I. On the same scale, proline and hydroxyproline gave values of 58 and 57 respectively, but these figures are not comparable with the others as the colours were different. TABLE I RELATIVE DEPTHS OF COLOUR FROM AMINO-ACIDS Relative depth of colour Amino-acids 100 glycine, leucine, isoleucine, norleucine 96 serine 93 phenylalanine, cysteine 90 lysine 89 glutamic acid, tyrosine. tryptophan 88 arginine, histidine. methionine 86 aspartic acid, ornithine 85 alanine 81 threonine, valine, itorvaline 79 cystine DISCUSSION OF RESULTS- The above results differ substantially from the corresponding values reported by Moore and Stein,13 although the analytical procedure was similar except in certain respects previously noted.For example, the figures given here are much lower for alanine, arginine, glutamic acid, histidine, lysine, methionine, threonine and valine, and higher for phenylalanine, tryptophan, cysteine and cystine. Moore and Stein used various solvents to elute the amino- acids from a starch column, added a much larger proportion of ninhydrin to develop the colour and determined the depth of colour with a spectrophotometer a t 570 mp. Fowdenlg has recently reported the results of a study of the reaction with amixed reagent of equal volumes of a 4 per cent. solution of ninhydrin in methyl Cellosolve and a 0-16 per cent. solution of stannous chloride, saturated with hydrindantin and stored under nitrogen.He took 20 to 30 mg of ninhydrin for quantities of acid containing 1 to 10 pg of a-amino nitrogen, heated for 25 minutes, made up the solution to 10 or 25ml with water and acetone, and estimated the colour with a spectrophotometer at 570mp. His method, therefore, closely resembled that of Moore and Stein, and his list of relative depths of colour is similar to theirs except that his values are higher for aspartic acid, glutamic acid, tyrosine, histidine, phenylalanine, serine and valine. The reason for the discrepancies in the three sets of results is not clear, apart from those for cysteine and cystine, because when the leucines are taken as a standard, the present values are in close agreement with those of Moore and Stein in respect of glycine, serine, aspartic acid and tyrosine, and with those of Fowden in respect of glycine, serine, phenyl- alanine and tryptophan.Obviously the method may not give a reliable result for a mixture of amino-acids, but it can nevertheless give reproducible results for individual acids, and is sufficiently sensitive to be suitable for the determination of the very small amounts of amino- acids obtained in the fractionation of a protein hydrolysate by partition chromatography on filter-paper. APPLICATION TO PROTEIN HYDROLYSATES About 170 samples of protein extracted from various species of grassland herbage were subjected to acid hydrolysis and the individual amino-acids were separated by two-dimensional chromatography on Whatman No.54 filter-paper by means of the solvents (a) phenol - 0.2 N formic acid, in an atmosphere of ammonia, and ( b ) n-butanol- glacial acetic acid - glycerol - water (in the ratio 83 : 2.5 : 2.5 : 12) in an atmosphere of coal gas and hydrocyanic acid.Sov., 19511 COLORIMETRICALLY BY THE SI.; HYDRIS REACTIOS 627 After drying at 80" C, the paper was lightly sprayed with ninhydrin solution B and again dried at 80" C for two minutes. The colour developed by the ninhydrin marked the positions of the individual acids, which could be identified from their RF values and from their positions with respect to proline, whose characteristic yellow or yellowish-red spot was a useful guide. Fifteen of the acids included in Table I (and also proline) were identified in this way.One millilitre of the buffer solution was added to each tube and then 1 to 2ml of ninhydrin solution A. The subsequent procedure was exactly the same as that described on p. 624 for pure amino- acids. This method differs from that of Naftalin,20 who developed the colour at pH 7, at which pH tyrosine and cystine are insoluble, and extracted it with 75 per cent. acetone. For various reasons, cysteine, cystine, histidine, hydroxyproline, methionine and tryptophan were not determined on these chromatograms. The sulphur compounds were sometimes difficult to identify on chromatograms of protein hydrolysates, histidine was rather insensitive to development with ninhydrin, and hydroxyproline occurred only in extremely small amounts. The tryptophan could not be determined in an acid hydrolysate and was determined separately after hydrolysis of about 150 mg of the protein in 1 ml of 5 N sodium hydroxide in a sealed tube at 110" C for 90 minutes.The hydrolysate was brought to an acidity of 1.2 N with hydrochloric acid, filtered, and an aliquot was diazotised and then treated with S-( 1-naphthy1)-ethylenediamine &hydrochloride. The red - purple colour was salted out, extracted with n-butanol and measured in the colorimeter with green filter No. 404 (maximum transmission at 530 mp). The curve was linear over the range 0-05 to 0.20 mg of tryptophan, corresponding to 3.5 to 14 pg of a-amino nitrogen. This method differed from that described by Eckert21 only in some minor details adopted for convenience in routine determinations, and the results were not affected by the other compounds likely to occur in a grass-protein hydrolysate.Tyrosine, for example, had no effect on the colour unless it was added in considerable excess. Although this method of determining the individual amino-acids in a mixture does not attain the remarkable standard achieved by Moore and Stein1* with a starch column, it is rapid and convenient for routine analysis and sufficiently accurate for a comparison of the relative amounts of amino-acids present in different hydrolysates. The percentage recoveries of the individual acids from a synthetic mixture were as follows: arginine and lysine, 91; serine and valine, 93; aspartic acid, 94; alanine and tyrosine, 96; glutamic acid, 07; threonine, 99; tryptophan, 100; glycine, 101; the leucines and phenylalanine, 103; proline, 105.In an analysis of a mixture of ten of these amino-acids, separated from two-dimensional chromato- grams, Fowdenls obtained a similar range of figures (93 to 102) although the individual values differ markedly in five instances. The individual spots were cut out and placed under water in test tubes. 1. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 3 -. REFERENCES Virtancn, A. I., and Rautancn, X., Biochm. J . , 1947, 41, 101. Schlenker, F. S., Anal. Chem., 1947, 19, 471. Sobel, A. E., Hirschman, A., and Besman, L., J . Biol. CJieiw., 1945, 161, 99. Mason, M. F., Biochem. J . , 1938, 32, 719. Van Slyke, D. D., Dillon, R. T., MacFadyen, D. -4., and Hamilton, P., J . BioZ. Sci., 1941, 141, 637. Van Slyke, D. D., MacFadyen, D. -4., and Hamilton, P . , Zbid., 1941, 141, 671. Smith, A. M,, and Agiza, A. H.. Analyst, 1951, 76, 619. Ruhemann, J., J . Chenz. SOC., 1911, 99, 792, 1310, 1486. Harding, V. J., and MacLean. R. M., J . Biol. Client., 1915, 20, 217. -- , Zbid., 1916, 25, 337. Harding, V. J., and Warneford, F . H. S . , Zbid., 1916, 25, 319. Wieland, T., and Wirth, L., Bey. chem. Ges., 1943, 76B, 823. Moore, S., and Stein, W. H., J . Biol. Chem., 1948, 176, 367. -- , Ibid., 1949, 178, 63. MacFadyen, D. A., Zbid., 1950, 186, 1. MacFadyen, D. A., and Fowler, N., Zbid., 1950, 186, 13. Grassmann, W., and von Amim, K., Ann. Chew., 1934, 509, 288. Atkinson, R. 0.. Stuart, R. G., and Stuckey, R. E., .4izaZysf, 1950, 75, 447. Fowden, L., Biochem. J., 1951, 48, 327. Naftalin, L., Kature, 1948, 161, 763. Eckert, .H. W., J . Biol. Chem., 1943, 148, 203. THE EDINBURGH AND EAST OF SCOTLAND COLLEGE OF AGRICULTURE March, 1951