810 Artalyst, December, 1968, Vol. 93, fie. 810-816 Simple, Rapid Quantitative Determination of Amino-acids by Thin-layer Chromatography BY MARY E. CLARK (Department of Organismic Biology, University of California, Irvine, California 92664, U.S.A .) A simple, rapid method for the quantitative determination of complex mixtures of amino-acids is described. After separation on thin layers of cellulose mounted on flexible plastic sheets, the chromatograms are sprayed with ninhydrin and developed under controlled conditions. The spots are cut out and eluted with 2 ml of 50 per cent. propyl alcohol and the optical density at 570 nm determined with a microspectrophotometer. From 4 to 6 chromato- grams can be eluted and read in 1 day. An accuracy of about 5 nmoles is obtained over a range of 12.5 to 50 nmoles.A t higher concentrations, accuracy is within A10 per cent. Standard graphs are reproducible for at least 5 months. DURING an investigation of the composition of the free amino-acid pools of marine inverte- brates, a quick and reliable method for quantitative analysis was sought, Although mixtures of several amino-acids can be separated rapidly by thin-layer chromatography, its application has been severely limited because of the difficulties involved in making the method quantita- tive. Previously described methods involve scraping the coloured spots from glass plates, and either their subsequent elution for absorption spectrophotometryl or their analysis by reflectance spectroph~tometry.~~~ Both methods require considerable time in the handling of individual spots.The present paper describes a method in which commercially available thin layers of cellulose supported on a flexible plastic backing are used. After separation of the amino-acids and development of the coloured ninhydrin reaction products, the spots are cut out with scissors and eluted directly, in a few minutes, with a small volume of solvent. The optical density is then determined with a microspectrophotometer, the sensitivity and reproducibility of the method being of the same order as those previously described. EXPERIMENTAL STANDARD AMINO-ACID MIXTURE- Throughout the development of this method a single batch of a standard mixture of 17 amino-acids, obtained from CalBiochem (Kit No. AA-5), was used. This contained 2.5 +_ 0.05 pmoles per ml of each amino-acid.The graphs obtained by using this mixture were later compared with a known mixture of amino-acids, made up to resemble the free amino-acid pool of an experimental animal and with a concentration factor of 30 between the least and most concentrated amino-acids. As shown in Table 11, amino-acids of the known mixture gave values within the error appropriate to their concentration levels. It would thus appear that the CalBiochem mixture provided a suitable standard. SEPARATION OF THE AMINO-ACIDS- Samples were spotted on to 20 x 20-cm sheets of MN-Polygram Cell 300 (MN 300), which is a cellulose powder produced by Machery-Nagel and Co. (Duren), and supported on inert sheets of poly(ethy1ene terephthalate). To increase the reproducibility of separation, spots were always placed at the same corner of each chromatogram, in relation to its position in the package as supplied.Spots were applied with Drummond “Microcap” capillary pipettes, which were rinsed once with distilled water to ensure complete transfer of the contents. * Present address : W. M. Keck Engineering Laboratories, California Institute of Technology, Pasadena, California 91109, U.S.A. 0 SAC and the author. A hot air stream was directed on to the spot during application.CLARK 81 1 Chromatograms were run in Kontes Chromaflex developing tanks, with the solvent system described by Jones and Heath~ote.~ To increase further the separation of the larger spots, the times were slightly increased; thus, solvent I, which was isopropyl alcohol - formic acid - water (40 + 2 + lo), was run for 3Q hours and solvent 11, t-butyl alcohol - ethyl methyl ketone - ammonia solution - water (25 + 15 + 5 + 5 ) ) was run for 3 to 35 hours.Even so, with amounts of each amino-acid greater than about 25 nmoles, arginine and lysine overlap considerably. Between solvents I and I1 the chromatograms were dried for at least 5 minutes in a stream of cold air, and a line was then scored just below the zone of yellow impurities to separate them from the remainder of the chr~matogram.~ On removal from solvent I1 the chromatograms were subjected to a cold air stream for at least 20 minutes, and then stored between sheets of clean paper. COLOUR DEVELOPMENT- Before treating the chromatograms with ninhydrin they were once again subjected to a stream of cold air for 20 minutes to remove any ammonia absorbed from the atmosphere during storage.In our laboratory, where developing chromatograms could not be isolated from other chemical work, this proved the best insurance against high background colour. Chromatograms were sprayed with 7 to 8 ml of freshly prepared 2 per cent. ninhydrin in absolute ethanol,5 as this amount permits saturation without dripping. As soon as the ethanol was evaporated in a cold air stream, the chromatograms were moved to a dark cupboard, under conditions of constant temperature and humidity. The importance of the physical conditions for reproducible colour development of ninhydrin-reaction products in paper chromatography has been demonstrated by Wellington, who found that, for most amino-acids, 20" C and 35 to 40 per cent.relative humidity gave maximum colour produc- tion.696 Reproducible, though less intense, colour was produced under other conditions, provided they were constant. The effects of temperature and humidity on intensity of colour production on thin layers was not investigated here, reliance being placed on the air conditioning in the laboratory to maintain constant conditions. The temperature in the cupboard in which the chromatograms were developed remained between 23.3" and 26.1" C and the relative humidity between 60 and 64 per cent. over a period of several weeks. The chromatograms were developed for 24 to 30 hours, as this length of time was found to yield a maximum reaction on paper.5 ELUTION OF SPOTS- As elution entails destruction of the chromatogram, it is useful to have a record to indicate which spots were overlapping.Each spot was outlined with a blunt pencil and identified. In addition, background spots were drawn in at this time. The background varies across the chromatogram, especially in the direction of the second solvent. This may be caused either by impurities in the cellulose layer or by impurities in the solvents (or by both sources). It is convenient to divide the amino-acids into four groups, with a single background spot for each group. This spot was drawn the same size as the largest spot in the group, and when the other spots in this group were cut out, sufficient background was added so that all the spots were the same size as the background spot.The chromatograph was recorded with a Polaroid camera and Polaroid Land Projection film, Type 146-L. Fine-tipped dissecting scissors were used to cut out individual spots, care being taken to avoid finger-prints. To prevent chipping of the cellulose layer and loss of coloured material it was necessary to cut in straight lines. Each spot was placed in a 1.2 x 10-cm test-tube. If the spots were large, they were cut in half so that they were small enough to be completely covered by 2 ml of eluting fluid, and the pieces placed in the tubes with the cellulose layers facing outwards for efficient elution. Two millilitres of 50 per cent. propyl alcohol in distilled water were transferred by pipette into each tube. Elution required 20 minutes, after which each tube was shaken vigorously and left for 20 minutes to allow the cellulose particles to settle.DETERMINATION OF OPTICAL DENSITY- The optical density at 570 nm was determined for each spot, including the blanks, after setting the instrument at 100 per cent. transmission for 50 per cent. propyl alcohol. Micro- cuvettes and a conventional spectrophotometer can be used, or, for more rapid analysis,812 CLARK : SIMPLE, RAPID QUANTITATIVE DETERMINATION [Arta&st, Vol. 93 a Gilford Microspectrophotometer 300 is convenient. Both have been used in obtaining the standard graphs in the present study. In the latter method, about 1.5 ml of the coloured eluate is carefully decanted into a small vial, avoiding transfer of cellulose particles, prior to taking a reading.The optical density of the appropriate blank was then subtracted from that of each spot. CALBIOCHEM STANDARD MIXTURE- As is well known, equimolar amounts of the various amino-acids do not produce the same amount of coloured reaction product with ninhydrin. Although proline appears to give a more purplish colour on thin layers of cellulose than on paper, the amount of colour is insufficiently reproducible to yield quantitative results, and no proline values are included here. The results of analyses for three different amounts of the standard mixture are shown in Table I. The number of replicates (.n) at each concentration is given at the top of each column, and for all three amounts, replicates were made over a period of 5 months. As is to be expected, the percentage error decreases with increasing amounts of amino- acids. This is primarily caused by the uncertainties introduced by variable amounts of background colour.Thus, those amino-acids which produce less colour with ninhydrin, such as aspartic acid, glycine, cystine, histidine and tyrosine, tend to give a higher error than more chromogenic amino-acids. The variability of leucine, isoleucine and phenylalanine is also caused by background effects. These amino-acids lie in the upper right-hand quadrant of the chromatogram where the background is always deeper and more variable. It will be seen that, at concentrations ranging from 12.5 to 50 nmoles, the absolute error for most amino-acids remains at about the same level, between 3 and 5 nmoles. At the highest concentration, the total load on the chromatogram was 0.85 pmoles of amino-acid. When the mean optical density values for the individual amino-acids are plotted against concentration, rectilinearity (indicating agreement with Beer's law) is obtained for all amino- acids, except threonine, isoleucine and leucine (Fig.1). Even for these three amino-acids, reproducible graphs are obtained, and thus such graphs can be used as standards for the determination of amino-acids in unknown samples. RESULTS 0.500 0.400 r).zci? 03909 0.100 10 20 30 40 50 0.500 C 1 0,400 0.300 0.200 0.100 10 20 30 40 50 Nanomoles Fig. 1. Optical density veisus concentration graphs of several amino-acids. Values a t 12-5, 25 and 50 nmoles are taken from Table I. Values at 5 nmoles are based on 1 determination; those a t 37.5 are the mean of 2 deter- minations.Threonine and leucine do not give straight-line graphs. (The graph for isoleucine resembles that of leucine.) All other amino-acids follow Beer's lawTABLE . I MEAN OPTICAL DENSITIES, STANDARD DEVIATIONS, PERCENTAGE ERROR AND ABSOLUTE ERROR FOR VARYING CONCENTRATIONS OF CALBIOCHEM AMINO-ACID MIXTURE 1/2 Cystine Aspartic acid Glutamic acid Arginine Lysine . . Glycine . . Serine . . Histidine Alanine Tyrosine.. Valine . . Methionine Threonine Isoleucine Leucine Phen ylalanine 7 x . . 0.054 . . 0.051 * . 0.110 . . 0.093 . . 0.099 . . 0.053 . . 0.102 . . 0.050 . . 0.121 . . 0.071 . . 0.111 . . 0.106 . . 0.113 . . 0.083 . . 0.058 . . 0.070 12.5 nmoles n = 5 Standard deviation f 0.034 f 0.007 f 0.015 f 0.025 f 0.024 f 0.015 f 0.022 f 0.026 f 0.013 f 0.024 f 0,022 f 0.012 & 0-040 f 0-025 f 0-018 f 0.008 Error, per cent.63 14 14 27 24 28 22 44 11 34 20 11 35 30 31 11 - X = Mean optical density. n = Number of determinations. 3: nmoles f 7.9 f 1.8 f 1.8 f 3.4 f 3.0 f 3.5 f 2.8 f 5.5 f 1.4 f4.3 f 2.5 1.4 f 4.4 f3.8 f 3.9 f 1.4 25.0 nmoles n = 6 r A Standard Error, deviation per cent. 0.120 f 0.039 33 0.106 f 0.028 26 0.201 f 0-050 25 0.231 f 0.039 17 0.111 f 0*028 25 0-197 f 0.046 23 0.146 f 0.025 17 0.228 f 0.009 4 0.148 f 0.034 23 0.211 f 0.034 16 0.204 f 0.042 21 0.204 f 0,017 13 0.183 f 0.029 16 0.148 f 0.032 22 0.140 & 0.026 19 0-200 f 0.023 12 I f nmoles f 8.3 f 6.5 f 6.3 f 4.3 f 3.0 f 6.3 f 5-8 k4.2 f 1.0 f5.8 f 4.0 f 5.3 f 3.3 f 4.0 f 5.5 f4.8 50.0 nmoles n = 4 A I \ Standard Error, U deviation per cent.f nmoles rn 0.212 f 0.017 8 -fs 4.0 0.456 f 0.053 12 f 6.0 0.477 f 0.076 16 f 8.0 4 0.418 f 0.022 5 f 2.5 M 0.263 & 0.014 5 3: 2-5 0.512 & 0.019 4 5 2.0 0 0.286 f 0.018 6 f 3.0 9 0 0.479 f 0.038 8 f 4.0 0.455 f 0-017 4 f 2.0 0.388 f 0.029 7 0.177 f 0.036 20 flO*O c3 0.421 f 0.039 9 f 4.5 5 0.184 & 0.021 11 f 5.5 5 0.421 f 0.063 15 f 7.5 gj 0-396 f 0.066 17 f 8-5 0.288 & 0.009 3 f 1.5 $ 4814 CLARK : SIMPLE, RAPID QUANTITATIVE DETERMINATION [Autahyd, VOl. 93 ARTIFICIAL MIXTURE RESEMBLING THE FREE AMINO-ACIDS OF A POLYCHAETE WORM- In preliminary investigations, the composition of the free amino-acid pool of the poly- chaete worm Stauronereis rudolphi (Delle Chiaje) was determined on de-proteinised and de- salted aliquots of whole-body homogenates.(Details of the preparative procedures will be described elsewhere.) Nineteen amino-acids were identified (excluding the amino-acid derivative taurine, which does not give reproducible colour reactions by the present method and must be determined by alternative procedures.) These ranged in concentration from 0.5 mmoles per kg of body water (histidine) to 15 mmoles per kg of body water (aspartic acid). An artificial mixture was then prepared (Table 11) and analysed by the method described above. A chromatogram of this mixture is shown in Fig. 2. The results of two individual chromatograms of this mixture are shown in Table 11. The difference between the amount found and that expected falls, in most instances, near or within the range of error expected, as determined from the percentage error of the nearest amount used in preparing the standard graphs (Table I).Arginine and lysine spots overlap with each other and with asparagine, which gives a weak, brown colour with ninhydrin, but this latter error can be obviated by hydrolysing the mixture before it is used for a chromato- gram. There were no standard graphs for asparagine nor tryptophane, and these were read on the lysine and threonine graphs, respectively. For tryptophane this resulted in a systematic error, as the colour for tryptophane under these conditions was about half that of threonine. It should also be noted that the CalBiochem standard mixture contained cystine (1-25 pmoles per ml), whereas the artificial mixture contained cysteine, which yields a much weaker colour with ninhydrin.Aspartic acid consistently seems to give slightly high values, and alanine consistently slightly low ones. This may be a result of an interaction during migration of the amino- acids, when those present in disproportionately high concentration exert a salt-like effect on the movement of some of the other amino-acids. The total recovery of amino-acids was within 10 per cent. of that expected. DISCUSSION The method described here provides a rapid means of quantitatively determining small amounts of amino-acids. It requires about 3 days to carry out the whole procedure, of which only the third day, when the spots are eluted and read, requires continuous attention from the researcher.It was found possible to elute and read as many as six chromatograms in 1 day. It should be emphasised that during the development of this technique, every effort was made to treat replicate chromatograms on different days at each stage of the procedure, so that maximum variability of the conditions would be observed. In this way, it was possible to establish the true variation in reproducibility in our laboratory. In addition, all results were obtained on complex mixtures of at least 17 amino-acids; thus the range of error reported is conservative in all respects. It is probable that some of the error, especially at low concentrations, could be reduced by carrying out all the steps following the evaporation of the second solvent in a separate room kept scrupulously free from ammonia and other chemical fumes.There appear to be only two earlier methods for the quantitative determination of amino-acids with thin-layer chromatography. Frodyma and Frei,2s3 who use reflectance spectrophotometry, found that when silica gel on glass plates was used scanning did not give sufficiently reproducible results, and that it was therefore necessary to scrape the spots off the glass surface before determinations were made. By running simultaneous standards they achieved a 5 per cent. reproducibility, within a range of about 5 to 200 nmoles for various amino-acids. However, as the percentage reflectance is not linear with concentration nor, for several amino-acids, is it linear even with the square root of the concentration, analysis of complex mixtures necessitates running a complete set of standard graphs with each unknown.The other method is that recently published by Bondivenne and Busch,l who used cellulose 300 MN spread on glass plates and after colour development scraped the individual spots off the glass and eluted the colour through sintered-glass filters. As this required 4 ml of eluant, the sensitivity of their method is about one half that of the present method. The smallest amount of the four amino-acids analysed by these workers was 50 nmoles, at which level the error ranged from 3.7 to 8.5 per cent. When a chromatogram was made of theFig. 2. The chromatogram of an artificial mixture of amino-acids resembling the composition of the free amino-acid content of the pool of the polychaete worm, Stauronereis rudolphi.Solvent I was run in the vertical direction, solvent I1 in the horizontal To face page 8 14)TABLE I1 Cysteine . . Aspartic acid Glutamic acid Arginine . . Asparagine Lysine . . Glycine . . Serine .. Histidine . . Alanine . . Tyrosine . . Valine . . Methionine Tryptophane Threonine Isoleucine Leucine . . Phenylalanine Total . . .. .. ,. . . .. .. .. .. .. .. .. ,. .. .. . . .. .. .. . . RECOVERY OF KNOWN MIXTURE OF AMINO-ACIDS FROM TWO THIN-LAYER CHROMATOGRAMS Artificial mixture of amino-acids, mmoles per litre . . 2.0 . . 15.0 . . 10.0 * . 2.0 . . 5.0 . . 5.0 . . 10.0 . . 2.0 . . 0.5 . . 10.0 . . 1.0 . . 1.0 . . 1.0 . . 1.0 * , 1.5 * . 1-0 . . 1.5 . . 1.6 . . 71.0 nmoles found 1.0 71.6 43.6 13.8 5.6 12.8 40.2 9.4 1.8 37.0 6.4 4.4 4.6 1.8 4-4 4.4 7-6 5.8 276.2 nmoles expected 8.0 60.0 40.0 8.0 20.0 20.0 40.0 8.0 2.0 40.0 4.0 4.0 4.0 4.0 6.0 4.0 6-0 6-0 284.0 Difference found, nmoles + 11.6 + 3.6 - 7*0* 4- 5*8* - 14.6 - 7*2* + 0.2 + 1.4 - 0.2 - 3.0 $.2.4 + 0.4 + 0.6 - 2.2* - 1.6 + 0.4 + 1.6 - 0.2 Recovery, per cent. . . 97.3 * Spots where large errors were expected-see text for explanations. Approximate difference expected, nmoles f 5.0 f 6.0 f 2.4 f 2.2 -f: 2.4 & 4.4 f 1.8 0.9 4 1.6 f 1.4 f 0.8 f 0 - 4 - I f2.1 f 1.2 4 1.9 f 0.7 nmoles found 10.0 168.2 102.1 24.8 15.2 35.0 87.2 18.2 1-4 89.7 13.2 7.6 9.6 4.8 11.4 12.2 16.6 10.4 637.6 89.8 nmoles expected 20.0 150.0 100.0 20.0 50.0 50.0 100.0 20.0 5.0 100.0 10.0 10.0 10.0 10.0 15.0 10.0 15.0 15.0 710.0 Difference found, nmoles - 10*0* + 18.2 + 2.1 + 4.8* - 34.8* - 15*0* - 12.8 - 1.8 - 3.6 - 10.3 + 3.2 - 2.4 - 0.4 - 5*2* - 3.6 + 2.2 + 1-6 - 4.6 Approximate 's difference expected, 2 nmoles & 6.6 2 f15.0 $j f 6.0 f 3.4 2 f 4.6 f 2.2 ;p 4 f 4.0 M f 3.4 f 2.0 X f 1.1 & 5.3 & 3.0 f 4.7 f 11.7 cd 0 - 2 I?816 CLARK four amino-acids together, in amounts exceeding 100 nmoles of each, the error increased, ranging from 7.8 to 13.5 per cent.for five replicate determinations. It would thus appear that the present method offers a more rapid analysis, with somewhat improved accuracy and sensitivity. Recently Heathcote and Washington' have published a sensitive method based on paper chromatography, in which they state that, for most amino-acids, determination of amounts of from 1 to 250nmoles is possible with an over-all accuracy of _+lo per cent.It is not clear if this high degree of accuracy at lower concentrations could be obtained in complex mixtures, but it would appear that their method is more reproducible at lower concentrations than the present one. It entails, however, a rather elaborate, prolonged procedure for eluting the spots. The staining reagent used by these workers was chosen for its increased sensitivity, which is about five times that of ninhydrin alone. However, as the elution procedure required 10 ml for each spot, optical densities were equivalent to those obtained here. In addition, as with all methods in which paper chromatography is used, the time required in the solvent systems is about three times that necessary for separation on thin layers. When sufficient material is available to carry out several chromatograms on the same sample, better accuracy can be expected with the present method and this may prove less time-consuming than the more elaborate method of Heathcote and Washington, which would be the method of choice when sample size is limited. This work was supported by USPHS Grant 12889, awarded to Dr. Grover C . Stephens, in whose laboratory the work was done. I am most grateful to Dr. Stevens for making this study possible, and for his encouragement throughout. REFERENCES 1. 2. 3. 4. 5. 6. 7. Bondivenne, R., and Busch, N., J . Chromat., 1967, 29, 349. Frodyma, M. M,, and Frei, R. W., Ibid., 1964, 15, 501. - - , Ibbid., 1965, 17, 131. Jonis, K., and Heathcote, J. G., Ibid., 1966,24, 106. Wellington, E. F., Can. J . Chem., 1952,30, 581. - , Ibid., 1953,31, 484. Heathcote, J. G,, and Washington, R. J., AnaZyst, 1967, 92, 627. Received Muy 6th. 1968