ANALYTICAL CHEMISTRY further check, carbon dioxide from combustion of the derivative was assayed as barium carbonate in some cases. The selfabsorption curves for the four compounds are shown in Figure 3 and the combustion data in Table I. RESULTS
The .self-absorption curves for organic compounds were found to fit well with the barium carbonate curve by adjusting the scale. In Figure 3 the barium carbonate curve of Figure 2 is shown normalized to 1000 counts per minute a t 5 mg. and the experimental points for the organic compounds are shown normalized in the same way. The fit is good from 1 to 7 mg. (0.06 to 4.4 mg. per sq. em.) for all the compounds. Over the range mentioned the same self-absorption factors may be used for all the compounds studied, The specific activity (counts per minute per millimole) of barium carbonate obtained from a given substance, adjusted for self-absorption to 5 mg. per planchet, has a value very near to that obtained (counts per minute per millimole carbon) from
a derivative of the original substance similarly adjusted (Table I). The correction factors which are applicable to gas-flow counting under the conditions specified, near 5 mg. on the self-absorption curves, are considerably smaller than those previously reported for end-window Geiger-Muller tubes, and are, in fact, close to unity. LITERATURE CITED
(1) (2) (3) (4)
Crane, R. K., and Ball, E. G., J . BioZ. Chem., 188, 819 (1951). Glendenin, L. E., and Solomon, A. K., Science, 112, 623 (1950). Robinson, C. V., Ibid., 112, 198 (1950). Solomon, A. K., Gould, R. G., and Anfinsen, C. B., Phys. Rev.,
72,1097 (1947). ( 5 ) Topper, Y. J., and Hastings, A. B., J.BioZ. Chem., 179,1255 (1949). (6) Van Slyke, D. D., and Folch, J., Ibid., 136, 509 (1940). (7) Van Slyke, D. D., Steele, R., and Plazin, J., Ibid., 192, 769 (1951). (8) Yankwich, P. E., and Weigl, J. W., Science, 107, 651 (1948). R E C E ~ V Efor D review June 1, 1954. Accepted November 12, 1954. This investigation was supported by a contract with the U. S. -4tomic Energy Commission, and by the Eugene Higgins Trust through Harvard University.
Qualitative Determination of Amino Acids in Protein Hydrolyzates by Circular Paper Chromatography IRWIN ORESKES Department
of
and
ABRAHAM SAIFER
Physical Chemistry, Division o f Laboratories, Jewish Chronic Disease Hospital, Brooklyn 3,
A circular chromatographic technique for the identification of amino acids in protein hydrolyzates uses phenol and butanol-acetic acid as the solvents, isatin and ninhydrin as the color reagents. R f values compare favorably with those obtained by other investigators using buffered filter papers. The small size of apparatus and the speed and simplicity of the method make it very useful for routine analysis of amino acids. Quantitative applications of the technique are now under investigation.
I
N A previous publication (18)studies of the physical factors
that may influence circular paper chromatography in a closed system were discussed. The data obtained in this work supported the claims made by Rutter (16, f7) with respect to the many advantages of this simple chromatographic technique over the more conventional methods. While the method has been criticized by some investigators (IS)as being suitable for only simple amino acid systems, it has been shown by Giri and coworkers (5, 6) to be applicable to the complex amino acid mixtures found in protein hydrolyzates. In their latest publication dealing with the application of the circular chromatographic method to the determination of the 19 amino acids present in acid hydrolyxates of casein and edestin, Giri and Rao (8) employ a rather elaborate elution technique in conjunction with a four-solvent system. Such a system has little or no advantage over the multisolvent systems for the separation of complex amino acid mixtures-e.g., protein hydrolyzates-by one-dimensional paper sheet chromatography proposed by Fowden (3, 4 ) , McFarren and Mills (ff), and Redfield and Guzman Barron ( 1 4 ) . More recently a unidimensional procedure suitable for the determination of the amino acid content of protein hydrolyzates has been published by Roland and Gross (15). These authors employ a two-solvent system for the separation of 16 out of the 19 amino acids usually found in such mixtures, but require two additional solvent systems with developing times of 40 and 64 hours, respectively, for the separation of tryptophan and histidine. The circular chromatographic technique described in this paper
N. Y.
will permit identification of all of the 19 amino acids usually found in protein hydrolyzates, except for leucines, which are determined as a pair. This is achieved by running two separate chromatograms for each hydrolyzate with phenol (0.1% ammonia) and butanol-acetic acid as the two developing solvents. After acetone washing to remove phenol and drying, each chromatogram is cut into halves, one half being dipped into ninhydrin and the other in isatin as the two coloring reagents. The usefulness of isatin as a color reagent that permits the rapid identification of a specific amino acid in a band containing several others has been described (19). This entire determinat,ion of the 19 amino acids can be carried out within 18 to 24 hours. In addition to simplicity and compactness of apparatus, temperature control, speed and sharpness of separations, control of the rate of solvent flow, and removal of test samples during (or after) development, the circular chromatographic method described here has a high order of reproducibility of the R/ values for the individual amino acids, which compares favorably with the best of such values previously reported ( 1 0 ) . REAGENTS AND APPARATUS
Developing solvents were phenol, reagent grade (70’%), reagent isopropyl alcohol (5%), and distilled water (25%) by weight (gU). A mixture of 1-butanol, acetic acid, and water (40-10-5070 by volume), according to Partridge ( f d ) , is well shaken and the lower layer which forms is discarded. Ammonia, O.l%, was used in all phenol runs. Amino acid solutions were made up according to Levy and Chung (9) in two solutions. Solution A contained 5 millimoles er liter each of leucine, phenylalanine, tryptophan, valine, proi n e , hydroxyproline, threonine, glycine, aspartic acid, and lysine in 0.1N hydrochloric acid. Solution B contained 5 millimoles per liter each of isoleucine, methionine, tyrosine, alanine, glutamic acid, serine, arginine, histidine, and cystine in 0.1N hydrochloric acid. The standard solutions were kept refrigerated when not in use. The solution to be chromatogramed was neutralized with an equal volume of 0,LV sodium hydroxide, just before application to the paper. Color reagents used were isatin, 0.2% isatin in acetone containing 4% glacial acetic acid, and ninhydrin, 0.25% ninhydrin in acetone, These color reagents were applied t o the chromatograms by dipping. Whatman No. 1 filter paper, 24 cm. in diameter, was used in all runs.
V O L U M E 2 1 , NO. 5, M A Y 1 9 5 5
855
Culture dishes and covers (18) were used for developing the chromatograms. METHOD
The method of preparing and spotting the chromatograms as well as calculating the Rj values has been described (18). For each determination, two separate chromatograms were spotted with 20 kl. of neutralized standard solutions A and B. One was developed with the phenol solution in the presence of 20 ml. of 0.1% ammonia solution placed in a 25-ml. beaker inside the culture dish. When the solvent front reached a radial distance of 8 cm., the chromatogram was removed and washed twice with anhydrous acetone (technical), which removed all traces of the phenol solvent. This procedure obviated the necessity of air drying and gave cleaner backgrounds and sharper bands. The second chromatogram was developed with the butanolacetic acid solution and upon completion allowed to air-dry. In this case washing with acetone was of no advantage, as the butanol evaporated quickly and had no deleterious effects upon the amino acids, as has phenol. The completed chromatograms were then cut in half. One half was dipped in the isatin color reagent and the other half in the ninhydrin reagent. The sections were allowed to air-dry and develop color a t room temperature for 3 to 4 hours, or, alternatively, they were heated for 10 minutes a t 100" C. in order to hasten color development. The culture dishes used in this work were found to be sufficiently air-tight for solvents such as phenol. In the case of butanol and other volatile solvents, evaporation could be prevented by placing the culture dishes inside ordinary plastic bags, which effectively prevented evaporation of solvent from the paper. The hydrolyzate of human serum albumin was prepared by refluxing 125 mg. (0.5 ml.) with 25 ml. of 6N hydrochloric acid for 24 hours. (The 25% albumin solution was obtained from the American Red Cross through the courtesy of J. N. Ashworth. I t was a t least 95% pure by electrophoretic analysis.) The hydrochloric acid was removed by evaporation to dryness in vacuo a t 3 5 O C. (1). The resultant film of amino acids was taken up in 12.5 ml. of 0.1N hydrochloric acid, giving a concentration of 1.6 mg. of nitrogen per ml. As in the case of the standard solutions, an aliquot was neutralized with an equal volume of 0.1N sodium hydroxide just prior to use. Forty- and 8O-pl. aliquots of the neutralized hydrolyzate were found suitable for spotting the chromatograms. The larger aliquot was necessary to detect methionine and other amino acids present in low concentrations. RESULTS
Determination and Reproducibility of Rj Values. Twenty-four chromatograms containing the amino acids shown in Table I were run in the phenol solvent in the manner described above. The bands obtained were numbered beginning with the innermost one. The identities of the amino acids in each band were determined by chromatograming the standard mixture, to which an excem of one amino acid was added. Its position was then located by noting which band was darker or thicker. This procedure was used to locate the position of each amino acid, and thereby determine the order of separation shown in Table I. Table I. Reproducibility of R, Values Obtained in Standard Mixtures of Amino -4cids Employing Phenol (0.lY' Ammonia) as Developing Solvent Band KO. Amino Acids Present Mean Rj 2 S.D. Cystine Aspartic acid 3 Glutamic acid Serine 4 Glycine 5 Threonine G Alanine, tyrosine 7b 6 Hydroxyproline 9 Tryptophan 10 Histidine, valine, methionine 11 Leucines, lysine, phenylalanine, arginine 12 Proline Band sometimes splits into two bands. Occasionally this pair shows some separation. 1
0.18 i. 0.023
2n
0.24 i 0 . 0 2 0
'
a
b
0.35 1 0 . 0 2 2 0.43 3t 0.024 0.50 i 0 . 0 2 1
0 . 5 5 i 0.022 0 . 6 4 & 0.016
0 . 7 1 3tO.010 0 . 7 4 i 0.019 0.60 i 0 . 0 1 3 0.86 3t 0 , 0 1 1 0 . 9 2 3t 0.008
Table 11. Reproducibility of R/ Values Obtained in Standard Mixtures of Amino Acids Employing Butanol-Acetic Acid-Water (40:10:50) as Developing Solvent Band No. Amino Acids Present Mean Rf S.D.
*
Cystine 0.18 f 0.026 Lysine 0.22 &0.023 b Ilktidine 0 . 2 3 zt 0 . 0 2 3 C Arginine 0 . 2 5 f 0.023 3 Aspartic acid, glycine, serine 0.31 f 0.018 4 Hydroxyproline 0 . 3 3 1 0.022 5 Glutamic acid, threonine 0.36 f 0.018 6 Alanine 0 . 4 0 1 0.018 7 Proline 0 . 4 4 f 0.026 Sb a Tyrosine 0 . 4 6 f 0.021 b Tryptophan 0 . 4 7 f 0.021 9 Methionine 0 . 5 5 f 0.020 10b a Valine 0.62 f 0.016 b Phenylalanine 0.66 f 0.016 11 Leucines 0 . 7 3 f 0.014 These three amino acids travel close together. At high concentrations these two amino acids tend t o overlap. 1 2a a
a
b
iin identical series of 24 runs was performed using butanolacetic acid as the developing solvent (Table 11). The mean Rj of a component for a single run is obtained as the average of six different measurements (18). The average R, values and their standard deviation for the 24 determinations are given for phenol in Table I and for butanol in Table 11. Qualitative Identification of Amino Acids in Protein Hydrolyzates. The scheme for the qualitative identification of the amino acids in a protein hydrolyzate as developed in this laboratory employs two circular chromatograms run simultaneously. One is developed in phenol and the other in the butanol solvent, the former being washed in acetone after development. Half of each chromatogram is then dipped in isatin and the other half in ninhydrin. If the standard mixtures are assumed to represent a hypothetical protein hydrolyzate containing all 19 amino acids, their identification could then be made in the following manner: On the phenol-ninhydrin chromatogram the following amino acids separate as pure bands and can be easily identified: cystine, aspartic acid (blue-green), glutamic acid, serine, glycine, threonine, and tryptophan (brown). Hydroxyproline and proline are identifiable as pure bands on the phenol-isatin portion. In this system alanine and tyrosine migrate together, but as only tyrosine reacts with isatin, it is also identifiable on the isatin portion. On the butanol-ninhydrin chromatogram, cystine, arginine, alanine, methionine, and the leucines separate as pure bands However, valine and phenylalanine tend to overlap a t high concentrations. The identification is made on the butanol-isatin portion, where only the outer part of the band reacts, showing the presence of phenylalanine. Valine is identified by a positive ninhydrin reaction and a negative isatin reaction of the inner part of the band. With approximately 10 y of valine and phenylalanine, resolution usually occurs which permits their direct identification. Lysine and histidine travel close together in butanol and their identification is sometimes difficult with ninhydrin. However, in isatin the lysine gives a light lavender shade and the histidine a dark blue-gray color, thus making their identification possible. Except where otherwise noted, the ninhydrin colors are shades of purple and the isatin colors are bluegreen. In a number of cases isatin allows identifications to be made even where two or more components migrate close together or overlap on the chromatogram. This eliminates the necessity for complete resolution of all the components. As it is possible to cut out several sectors from circular chromatdgrams, many other color reactions can be used to confirm the presence of specific amino acids (1, ?, 20) on a single chromatogram. The application of this technique to the acid hydrolyzate of human serum albumin is shown in Figures 1 and 2. Except for the presence of alanine, the amino acids found in this hydrolyzate were the same as those reported by other investigators ( 2 , 21 ).
ANALYTICAL CHEMISTRY
,856 DISCUSSION
As stated hy MeFarren (lo), conventional, two-dimensional paper sheet chromatography suffersfrom the difficulties of poorly defined spots, irreproducible RJ values, inseparahility of some of the amino acids, and the manipulation of large paper sheets. This makes onedimensional Chromatography inherently more suitahle for the quantitative analysis of the amino acids present in protein hydrolyzates. It was essentially for this reason that a number of investigators (S, 4,8, 10, 11, 14, 16) have employed unidimensional paper chromatography using multisolvent developing systems.
Figure 2. Circular chromatogram of hydrolyzate of human serum albumin developed with butanol-acetic acid
Figure 1. Circular chromatogram of hydrolyzate of human serum albumin developed with phenol
based primarily on sequence and color rather than on the R, values. The quantitative determination of the amino acids present in various biological materials, the instrumentation involved, and the techniques employed will be discussed in a subsequent publication. ACKNOWLEDGMENT
The authors wish to acknowledge the aid and encouragement given them by Bruno W. Volk, director of lahoratories, throughout this study, and t o express appreciation to Renee Eisner for the typing and editing of the manuscript and to Herbert Fischler for the photographic illustrations. Circular chromatography h s one major advantage over other chramtLtographic systems in which the amino acids are resolved an spots: the separation of the components as narrow hands (1 t o 4 mm. in width), SO that substances can he identified, even though their R j values differ by only 0.01 to 0.02unit. This makes possible the separation and identification of such groups as lysinehistidine-arginine and valinephenylalanine, which travel close together in this system. I n general, RJ values reported in the literature show great wriability due to variations in pH, temperature, and filter paper used. McFarren (10) in an extensive study on the separation of amino acids using buffered filter papers reported R j values and their standasd deviation at a number of pH's and in a variety of solvents. Using phenol as the solvent, and by controlling pH and temperature ( 2 2 O =k '1 C , ) , he obtained standard deviations of the amino acid R, values ranging from +0.01 to +0.05. The present work was done with unbuffered papers and at room temperature, and without close temperature control. Even so, the largest standard deviation was zt0.026 for cystine and proline. This indicatei that with the circules chromatographic method, R, values can he obtained with excellent reproducibility, if the fxtors previously investigated are carefully controlled (18). The reproducibility is as good as or better than that obtained with b d e r e d papers and the extra step of preparing the buffered papers is not required. However, where the amino acids travel closely together-e.g., basic amino acids-the identification should he
LITERATURE CITED
(1) Block. R. J., ANAL.CHEM., 22, 1327 (1950). (2) Brand, E..Ann. N . Y. Amd. Sci., 47, 187 (1946). (3) Fowden, L.,B i o c h a . J . (Londm).48, 327 (1951). (4) I i A . , 50, 355 (1952). (5) Giri, K.V.. Nature, 171, 1159 (1953). (6) Giri, K. V., Krishnamurthy, K.. and Venkitasubrsmanyan, T. A,. J . Indian Inst. Sei., 34,209 (1952). (7) Giri, K. V., and Nagabhushsnam. A,, Naturwissensehaften. 39, 548 (1952). (8) Giri, K. V., and Rao, N. A. N., J . Indian Inst. Sci:, 35, 343 I\ I, U F 1y\I . "u
Levy. A. L., and Chung. D., ANAL.CHEM.,25,396 (1953). McFwren. E. F., Ibid.. 23, 168 (1951). McFarren, E.F.,and Mills, J. A,, Ibid.. 24, 650 (1952). Partridge, S. M.,Biochem. J . (London), 42, 238 (1947). Proom. H..and Woiwood, 0. J.. Nature, 171, 42 (1953). Redfield. R. R.,and Gusman Barron, E. S., A x h . Biochem. and Biophys.. 35, 443 (1952). Roland, J. F., Jr., and Gross, A. M.. ANAL. CaEm.. 26, 502 (1954). Rutter, L.,Analyst. 75, 37 (1950). Rutter. L.,N a t u ~ e 161, , 435 (1948). Saifer,A,, and Oreskes, I.. ANAL.C ~ M .25, , 1539 (1953). Saifer, A,, and Oreskes, I., Science, 119, 124 (1954). Toennies, G., and Kolb, J. J., ANAL.CHEM.,23,823 (1951) Tristram, G. R., Advances in Protein C h a . . 5, 83 (1949). Rmervso f 01 review April 29, 1954. Accepted November 12, 1954. Prssentzd before the Division of Biologi~elChemistry at the 126th Meeting of the AWEBICAN CX~MICAI. SOCIETY, New York. N. Y., 1954. Aided by Grant B-285 from the U. S. Publia Health Service.
