potassium u as found in the basic amino acid fraction and no detectable amount with the other amino acids. I n obtaining the amine fraction, the monov d e n t cations are removed froni the resin with 0 . 5 s hydrochloric acid, but the di- and trivalent cations are only partially eluted. The remaining polyrnlent cations I\ hich are eluted with 6-1- hydrochloric acid usually do not interfere n ith the chromatography of aniincs. The acid form of Dowex 50 does not retain strongly acidic amino acids such as cysteic acid nnd taurine (3)and these are not purified. Such strongly acidic amino acids h a w not been reported in plants (20). It would be possible to retain these acids and other anions on the salt form of a strongly hasic resin while nonionic substances would pass through. Siich variations ~ ~ - o u lprovide d a basis for a more extensin. fractionation
scheme nhereby separate fractions with nonionic materials, anions, neutral amino acids. acidic amino acids, basic amino acids, cations, and some amines could be obtained (10. 1 7 ) . LITERATURE CITED
(1) Boulanger, P., Biserte, G., Bztll. soc. chzni. bid. 31, 696 (1949). 12) Zbid., 33, 1930 (1951). (3) Carsten, 11. E., J . Am. Chen2. SOC. 74,5954 (1952). (4) Consden, R., Gordon, ..i.H., Martin, A.J . P., Bzocheni. J . 38,224 (1944). ( 5 ) Zbid., 41, 590 (1947). (6) Drbze, 8., JIoore, S., Bigwood, E. J., Anal Chivi. d c t a 11, 554 (1954). ( i )Hems, B. A,, Page, J. E., Kaller, J. G., J . Soc. Chenz. 2nd. (Lcndon) 67, 77 (1948). (8) Him, C. H. IT.,Moore, S., Stein, IT.H., J . Bzol. Chem. 195,669 (1952).
(9) RlcKee, H. S., Uibach, G. E., d u s tralian J . Bzol. Sct. 6, 369 (1953). (10) Miller, S. I,.] J . A m . Chem. SOC.77,
2351 (1955).
(11) Moore, S.,Stein, IT. H., J . Hiol. Chem. 192, 663 (1951). (12) Ibid., 211, 907 (1954).
(13) Mueller, G. C., Bon-man, Grace, Herranen, Ailene, AXAL,CHEX.27, 1357 (1955). (14) Keuberger, il., Proc. Roy. Soc. (London) 158A, 68 (193i). (15) Piez, K. A,, Tooper, E. B., Fosdick, L. S., J . Biol. Chem. 194,660 (1952). (16) -Plnisted. P. H.. Contribs. Boiice --Thompson jnst. 19, 231 (1958). (17) Racusen, D. W.,Aronoff, S., -4rch. Biochem. Bzophys. 51,68 (1954). 118) Roverv, M., Desnuelle, P., Biochz???. et Bzophyb. Acta 8 , 450 (1952). (IF)) Stein. W. H.. Moore. 9.. J . Bzol. Chem. 190, 103 (1951). (20) Steward, F. C., Zacharius, R. >I., Pollard, J. K., Ani?.Acad. Sei. Fennzcae, Ser. A 60, 321 (1955). 121) Svnne. R. L. hl.. Biochem. J . 48, ' 429 11g51). (22) Thompson, J. F., hlorris, C. J., ANAL.CHEX 31, 1031 (1959). \--,
\
-
RECEIVEDfor review July 24, 1958. -4ccepted January 30, 1959.
Determination of Amino Acids' from Plants by Paper Chromatography JOHN F. THOMPSON and CLAYTON J. MORRIS
U. S.
Plant, Soil and Nutrition laboratory, Agriculfural Research Service,
b In the determination of plant amino acids b y quantitative paper chromatography, neutral and acidic amino acids are separated on a two-directional chromatogram. The leucines, phenylalanine, and the basic amino acids are separated b y one-directional chromatography. Proline, asparagine, and pipecolic acid are located on the paper with ninhydrin and the reaction with ninhydrin i s completed in the test tube. The remaining amino acids are reacted with ninhydrin on the paper, the color i s dissolved out and measured colorimetrically. This method i s sensitive and reproducible. A determination of P-lactoglobulin by this method i s compared with that obtained b y other procedures.
P
CHROMATOGRAPHY is useful for the identification (25, 29) and determination (7, 16,32, 33) of the protein and nonprotein amino acids of plant materials. This method has advantages over starch ( 2 7 ) , silica gel (Sq), and ion exchange resin chromatography ( I S , 19), and microbiological assay (14, SO) because it requires less sample, and the nonprotein amino acids such as pipecolic acid, 4-aminobutyric acid, asparagine, and glutamine, which APER
U. S. Department of Agriculfure,
are common in plants (28, 2Q) can be determined. This paper presents paper chromatographic methods for the determination of amino acids from plants. After chromatography, the amino acids are measured from the color produced by their reaction with ninhydrin either in the test tube after location on the chromatogram or by careful reaction on the paper and subsequent elution. Previous methods (32, 33) have been extensively modified and applied t o the analysis of all protein amino acids except hydroxyproline, cysteine, and tryptophan, while four nonprotein amino acids common to plants-asparagine, glutamine, 4-aminobutyric acid, and pipecolic acid-have been included. Tn-0-directional chromatography, as outlined here, separates most amino acids except the leucines from each other, or the basic amino acids from each other, cysteine, or asparagine. The basic amino acids are removed from the other amino acids and all are purified with ion exchange resins (31). Three chromatographic separations are performed and these are presented individually. MATERIALS
DEIONIZED WATER(31)
Ithaca, N. Y.
TVASHED FILTER PAPER.Two sizes of Whatman N o . 1 filter paper sheets, 23 x 23 or 23 X 28 inches, are clamped in a special apparatus (9)so that the fast direction of the paper is vertical. (It is vertical if a sheet of paper, folded over a glass rod, wrinkles when wet with water rather than lying smooth on the rod.) The filter paper is ivashed with 0.2 to 0.3N hydrochloric acid (approximately 40 liters of acid for 60 sheets) and then rinsed with deionized rrater until free of acid. The solutions must go through the paper evenly and not around it. The larger sheets are used without neutralization and are dried a t this point. The smaller sheets are neutralized by percolating 0.2N sodium hydroxide through tge paper until the effluent is alkaline. The base is washed from the paper with deionized water until the effluent at the bottom is neutral. The wash water must be pure; other\yise the paper will retain cations which may interfere with chromatography and color development (6). PHENOL SOLVENT MIXTURE. A technical grade of phenol is vacuum-distilled ' C. in a glass and stored as a solid at 4 container in the dark. The phenol is melted and eight volumes are mixed with three of deionized water. The phenol-water mixture is neutral and keeps at room temperature in the dark for several months without discoloring. BUTYLALCOHOL-ACETIC ACID SOLVEXT MIXTURE. Commercial grade VOL. 31, NO. 6, JUNE 1959
1031
butyl alcohol is distilled, retaining that portion distilling at 116" to 120" C. Nine volumes of butyl alcohol are mixed with one of reagent grade glacial acetic acid and 2.5 volumes of deionized water. The mixture is prepared just before use to keep ester content as lonas possible. BmzYIrtert-BUTm -4LCOHOL SOLVEST MIXTURE. Commercial grade benzyl alcohol and tert-butyl alcohol are distilled before use. Five volumes of benzyl alcohol, 5 of tert-butyl alcohol, and 2 of distilled water are shaken together. SODIUM PHOSPHATE STOCK SOLUTION. A 1 M solution of reagent grade salt is prepared in deionized water and purified by shaking with a saturated solution of dithizone in distilled carbon tetrachloride. For 1 liter of solution, 25 ml. of dithizone solution should be adequate to remove all heavy metals. For 0.01M and 0.04111 sodium phosphate, dilute with deionized water. SODIUM PHOSPHATE BUFFERp H 2.5, Concentrated phosphoric acid (857,, c.P.) is added to 0.04111 sodium phosphate solution until the p H is 2.5 (about 1 ml. of concentrated acid per liter). COLLIDINE. Commercial y-collidhe (Koppers Co., Inc., Pittsburgh, Pa.) is distilled from 5 ml. of concentrated hydrochloric acid using a fractional distillation column. T h a t portion which distills a t 169' to 170" C. (uncorr.) is used and stored in the dark.
Table 1. Effect of Various Factors on Determination of Amino Acids by Paper Chromatography
Color Produced on Unwashed as Recovery" Compared of Amino with Washed Acid after Paper after Chroma- ChromatogAmino Acid tography, % raphy, % Aspartic 101 91 Glutamic 98 83 Serine 102 85 Glycine 102 70 Threonine 97 69 Alanine 100 53 Glutamine 102 55 Methionine sulfoxide 98 35 Proline 100 49 Valine 100 58 Leucines 96 61 Phenylalanine 99 16 Tyrosine 105 26 4-Aminobutyric acid 100 44 Pipecolic 100 42 a Pure amino acids (40 y of each in 20 ~ 1 . ) were plaed with micropipet individually on origin of chromatograms and separately on sheet of chromatography paper. After chromatography, amino acids on chromatograms were reacted with ninhydrin under the conditione given. Color so obtained was compared with that obtained from unchromatographed acids.
PROCEDURE
Separation of Neutral and Acidic Amino Acids by Two-Directional Chromatography. Washed and neutralized filter paper (23 X 23 inches) is buffered by completely wetting each sheet rvith 0.01M sodium phosphate, allowing t h e excess t o drip off, a n d t h e paper t o dry. The sample is applied t o t h e corner opposite !There t h e paper was clamped in t h e n-asher. The sample should contain 40 to 60 y of amino nitrogen as determined by ninhydrin activity (36, 37) after evaporation to dryness with a n alkaline buffer (4). The sample is applied to the paper so that the initial spot is not more than 2 em. in diameter. Methionine is oxidized to the sulfoxide by placing 1 drop of 30% hydrogen peroxide in a watch glass beneath the origin for 20 to 30 minutes. The amino acids are chromatographed with the butyl alcohol-acetic acid solvent (130 nil. per t\To sheets) in the slow direction of the paper a t 18" to 20" C. The atmosphere of the chromatographic chamber is moistened by placing trays of butyl alcohol and water in the bottom. The paper is placed in the chamber, 16 to 24 hours before the solvent is added. T h e n the troughs are dry (3 to 4 days), the sheets are removed and dried in moving air for a t least 2 to 4 hours a t room temperature. The sheets are chromatographed at right angles to the original direction with the phenol solvent mixture (90 ml. per two sheets). When the solvent reaches the lower edge (24 to 36 hours),
1032
ANALYTICAL CHEMISTRY
the sheet is removed and dried overnight in flom-ing air at room temperature. Separation of Leucine, Isoleucine, and Phenylalanine by One-DirecA sheet of tional Chromatography. n-ashed unneutralized filter paper (23 x 25 inches) is buffered with 0.01M sodium phosphate. Samples are applied on a line 8 cm. from t h e shorter edge. T h e far edge is serrated n-ith 1-em. deep notches t o hasten dripping of the solvent. Chromatograms are placed in a n airtight cabinet with a pan of fresh ivater covering the bottom. Dishes Jyith a surface area equivalent to one tenth of bottom area are filled with the organic solvent and placed in the water. After equilibration for 48 hours, 200 ml. of tert-butyl alcohol-benzyl alcohol solvent mixture is added to each trough through a small hole in the top of the cabinet. K h e n the solvent is gone from the trough (6 to 7 days) the sheets are removed and dried a t room temperature for 2 days. Separation of Basic Amino Acids by One-Directional Chromatography. Washed and neutralized paper (23 X 23 inches) is cut into strips 23 cni. nide, serrated t o 1-em. depth a t one end, buffered with 0.04M sodium phosphate buffer ( p H 2.5), and dried. TKOsamples are placed o n each sheet 8 em. from t h e end and 5 cni. from each side. Eight strips are placed in four troughs in a square glass jar (Corning 431412) with a blotter around the side soaked in phenol saturated with x a t e r and four Petri dishes of R-ater in the bottom. The cover is tightly
sealed and the paper hung for 24 hours. Then 50 ml. of phenol solvent is added to each trough and after 60 hours, the paper strips are dried a t room temperature for a t least 12 hours in moving air. DISCUSSION
Separation of Neutral and Acidic Amino Acids by Two-Dimensional Chromatography. Phenol has been t h e most effective solvent for t h e separation of amino acids. Butyl alcohol-acetic acid, rather t h a n collidine-lutidine (L9,52) was chosen as the second solvent because it provides better separation of glycine and asparagine, serine and glutamic acid, glutamine and alanine. and methionine sulfoxide and 4-aminobutyric acid. The use of butyl alcohol-acetic acid also avoids (Table I) the loss of amino acids during chromatography as obtained n-ith collidine-lutidine (15, 32). The principal difficulty arising from butyl alcohol-acetic acid lies in the separation of glycine, serine, and glutamic acid. These compounds are relatirely prominent constituents of both the protein and nonprotein fraction of plants and, consequently, adequate separation is essential but sometimes difficult, particularly if one is disproportionately high. Certain features of this prccedure effect this separation. The use of butyl alcohol-acetic acid as the first solvent results in a better separation of serine and glycine in the subsequent phenol run. The complete separation of glutamic acid necessitates the extended run of the butyl alcoholacetic acid solvent which partly justifies the use of descending rather than ascending chromatography. Hoviever. the excess solvent carries the leucines and phenylalanine off the paper. The leucines and phenylalanine can be retained on the paper m-ith fair separation of glutamic acid with a double solvent run to the end of the sheet. but the leucines and phenylalanine yield distorted spots and no separation of the individual leucines. PAPER.The use of large sheets n-ith don-nward solvent f l o ~provides adequate separation of amino acids eren when present in relatirely high amounts. Because larger total amounts of amino acids can be handled, those present in relatively small amounts can be determined more accurately. This is important because the amount of amino acid determines the size of the spot (6) and thereby the completeness of separation. The filter paper sheets are acidwashed to remove alkaline contaminants which cause amino acid spots to be distorted, have beards, and react poorly n i t h ninhydrin (Table I). The paper must be washed and buffered with solutions free of heavy metals
because amino acids f(Jril1 complexes n hich are immobile and react poorly with ninhydrin ( 5 ) . The acid-n-ashing, subsequent neutralization, and buffering of the paper ensure both adequate separations and ma.tiniuni color in the ninhydrin reaction. Seutralization of the paper after acid-u ashiiig is necessary to obtain separation of tyrosine and 4-aminobutyric acid. Similarly, buffering of the paper is essential for adequate separation of glycine and glutamic acid. Of the filter papers (Kliatnian S o . 3, SR-S 507, 576, 112) tested, none yielded better separation of amino acids than Khatman KO. 1. The use of 23 X 23 rather than 18 X 22 inch sheets was preferable. Tn odirectional chromatography is preferred to a series of one-directional chromatograms ( 7 , 1 0 ) because better separations are obtained, especially of the nonprotein amino acids. Less nork is involved and less sample is required. The procedures described h a r e been designed to obtain optimum separation with the maximum amount of each amino acid. The suggested quantity of amino acids (40 to 60 y of nitrogen per chroniatogram) is suitable for the average plant sample. If any amino acid is disproportionately high it may be necessary to analyze smaller aliquots to obtain satisfactory separations. Conversely, for acids present in small amounts, it may be necessary to run larger aliquots to obtain greater accuracy. APPLICATIOXOF SAVPLE TO THE PAPER. Accurate quantitative application of sample to the paper is a problem. Small samples cannot be made up precisely to a small enough volume to use a micropipet. To aroid this, the follon-ing procedure has been developed: Sample solutions are made u p to a relatively large volume ( 5 to 10 nil.) so that 0.5 or 1.0 nil. contains the amino acids to be applied to one chromatogram. This aliquot is dried in a small test tube (12 X 75 mm.) in vacuum over solid sodium hydroxide or in flon-ing air. The dry residue is dissolved in a drop of deionized Ivater. This solution is transferred to the paper with a melting point cnapillary (1 x 100 mm.) in 20-pl. (approx.) portions with drying between applications. This operation is repeated with 2 or 3 additional drops of water and does provide quantitative transfer. CHROMATOGRAPHIC CONDITIOXS. Ill contrast t o paper chromatography of leucines and basic amino acids, a high atmospheric solvent content in the chromatography cabinet is not essential for adequate separations of neutral and acidic amino acids. Serine and glycine are better separated when the solvent content is low. However,
glutamine and methylcysteine sulfoxide (28) separate only when the level is high. The temperature is kept constant to avoid any separation of phases that might result from large temperature changes and to standardize flow rate of solvent. SOLVEXTS.Solvents are prepared as unsaturatecl solutions so that small changes in temperature cannot cause the formation of two phases. A more commonly used ( 2 , 1.2) mixture of butyl alcohol-acetic acid-water (4:1:5) prolides no better separations and has the disadvantages of requiring the separation of two phases and of being saturated. Distillation of phenol removes acidic and metallic contaminants and preservatives. Separations are impaired if the phenol is impure. Irl contrast to the results obtained Fvith basic solvents (15, 52), there is essentially no loss of amino acids in chromatography with acidic solvents (Table I) ( I O ) . This makes careful control of the distance or time of solviznt run (15) unnecessary. OTHER FACTORS. The tu-o-directional chromatograms )vi11 separate hydroxyproline, 8-alanine, honioserine, and methylcysteine sulfoxide or citrulline. Met hylcyteine sulfoxide and citrulline have similar R fvalues in these solvents. 4-Aminobutyric acid is not separated from S-methylcysteine. h1ethionine is preferably converted to its sulfoxide in pretreatment of the sample because methionine is partly oxidized during chromatography and methionine does not separate well from valine. The oxidstion of methionine under these conditions converts it quantitatively to the sulfoxide with no loss of other amino acids and no further oxidation of methionine sulfoxide to the sulfone. Separation of Leucine, Isoleucine, and Phenylalanine. Because leucine, isoleucine, and phenylalanine are difficult t o separate consistently in onedire2tional chromatography. several solvents were tried before t h e adoption of this solvent miuture. The cabinet must be saturated Tvith water vapor and a 48-hour equilibration time is recommended. The R , value of phenylalanine varies considerably with the n a t e r vapor content of the atmosahere in the cabinet. Because the R f values of these three acids are relatively l o r , it is necessary to run the solvents off the paper. Tryptophan overlaps leucine slightly in this solvent system. Other amino acids do not interfere with the leucines or phenylalanine but do not separate from one anof her. The water in the bottom of the cabinet must be changed after each run because sufficient solvent will fall froni the chromatograms to cover the water and reduce its evaporation.
'
After chromatography the sheets must be dried for 2 days a t room temperature because benzyl alcohol is relatively nonvolatile and interferes with the ninhydrin reaction. Separation of Basic Amino Acids. Satisfactory separation of basic amino acids was not obtained n-ith several published methods (7, 13; 16), although the basic solvent of Hardy, Holland, and Nayler ( 8 ) gave the necessary separations but incomplete recovery of histidine (about 50%). Consistent and marked separations can be obtained with phenol if the acids are kept in the salt form with a n acidic buffer on the paper. If the basic amino acids are pure, the p H of the solution is not critical because the buffer adequately controls this. Maintenance of high water vapor level is essential for obtaining coiisistently good separations. The procedure renders good separations of not only arginine, lysine, and histidine. but also of ornithine. The decarboxylation products of arginine, lysine, histidine, and ornithine have the same R , values as the parent compounds, but all except histamine can be separated on ion exchange resins (31). The amines n-hich are decarboxylation products of the neutral amino acids may also be in the fraction with basic amino acids ( S I ) b u t they do not interfere. K O effort has been made to separate and determine these. QUANTITATIVE DETERMINATION OF INDIVIDUAL AMINO ACIDS
One objective was t o obtain high color yields from t h e reaction of t h e amino acids with ninhydrin t o increase t h e sensitivity of t h e analysis and t o reduce the variabilitg of individual analvses by obtaining a complete reaction. T o accomplish this, most amino acids are reacted on the paper under special conditions. Pipecolic acid, asparagine, and proline present special problems (15, 20, 33) n-hich are solved by completing the reaction in the test tube. Materials and Reagents. S i n hydrin. Commercial mateiial from Kutritional Biochemical Co. or Dougherty Chemical Co. is satisfactory n ithout purification. 95y0Alcohol. Conimercial alcohol is distilled froni citric acid. Sodium Acetate Buffer pII 4.60. One mole of sodium acetate is dissolved in pure ivater and glacial acetic 1s added until the p H is 4.6 (about BO d . ) . The volume is made to 1 liter (4). Meth! 1 Cellosoli e. Commacia1 grade is distilled over ferrous sulfate. This should not be kept a t room temperature for more than 1 to 2 months ( 4 ) , but may be kept at -20" C. for 6 months. Detection Ninh! drin Solution, 0.05% in absolute methanol. Ninhydrin Reagent 1. A 2% solution of ninh>drin in 957c alcohol conVOL. 31, NO. 6, JUNE 1959
* 1033
taining 2% of a mixture of 5 volumes of glacial acetic acid and 1 of collidine is prepared just before use and is not kept more than 24 hours. Ninhydrin Reagent 2. A 2% solution is prepared in 9 parts of 95% ethvl alcohoi and 1 part bf collidin&glac&,l acetic acid (5-3 v./v.). Ninhydrin Reagent 3. Twenty milligrams of ninhydrin and 2 of hydrindantin per ml. are dissolved in a 1 to 1 mixture of methyl Cellosolve and sodium acetate buffer, p H 4.6. Ninhydrin Reagent 4. A 4% solution of ninhydrin is prepared in a mixture of 9 volumes of butyl alcohol and 1 of 6M phosphoric acid. This must be prepared within 6 hours of use but the ninhydrin can be dissolved in the butyl alcohol and kept indefinitely. Ninhydrin Reagent 5. A 4% solution of ninhydrin is prepared in a mixture of 9 volumes of butyl alcohol and 1 of licf phosphoric acid (2'6) just before use. Glacial Acetic Acid, reagent grade. Potassium Hydroxide-Borate Buffer. Thirty-two and five tenths milliliters of 6N potassium hydroxide is added to a solution of 9.28 grams of boric acid in absolute methanol and diluted to 500 ml. with methanol (4). 50% Alcohol Solvent for Dissolving Yinhydrin-Amino Acid Color from Paper. One volume of 95% alcohol is mixed with one of distilled water and to this is added 2% collidine (v./v.). Chromatogram Holder. To prevent condensation of alcohol on the paper support and dripping on the chromatogram, a special glass holder has been designed for two-directional chromatograms (Figure 1) and the chromatogram clamped a t points A and B. Strip chromatograms are attached to a similar rack modified to hold two strips. PROCEDURE
Neutral and Acidic Amino Acids. The
two-directional
Table II.
. -.
_
--
P
__
-___. -
e-
I
?-__-
7
-
PAPER'
_ _ __
~
~~~
_
is sprayed with the detection ninhydrin solution in the proline and asparagine regions and dried a t room temperature for 5 t o 15 minutes. If the spots are not visible, the chromatogram is heated for 1 to 2 minutes a t 60" C. The asparagine spot is cut out and a n equal piece (as a blank) is cut from a region where there are no amino acids. Both blank and sample paper are hung on a small hook of Nichrome wire and dipped in the potassium hydroxidtborate buffer. The wire is hung on a rod and dried a t 40" C. in moving air for 10 minutes. Each paper is cut in approximately l/d-inch square pieces into a freshly acid-washed and dry test tube (18 X 175 mm.). Five milliliters of ninhydrin reagent 3 is added to each tube and the tube is shaken until the paper is covered by the solution. The tubes are loosely capped and heated 15 minutes in boiling water, cooled, and diluted with 5 ml. of 50% alcohol solvent. Readings are made a t 570 mp and the blank reading is subtracted from the corresponding sample reading to obtain color due to asparagine. The proline spot is located by its yellow color. If it is very faint it may be necessary to illuminate the chromatogram from behind. The proline spot is cut in '/c-inch squares, and put in a clean dry test tube (18 X 75 mm.). Because the blank correction for proline is small and consistent, only one blank has to be included with a set of chromatograms and it is obtained in the same manner as the asparagine blank. Five milliliters of reagent 4 is added to the
Method of Final Color Production In test tube by procedures specified In test tube by procedures specified
ANALYTICAL CHEMISTRY
-
-
1
Figure 1 . Glass rack for holding chromatograms during color development
sprayed with 0.05% ninhydrin in absolute methanol and kept at 25" C. for 1hr. 3. Same as 2 but kept a t In test tube by procedures 60" C. for 2 min. specified 4. Paper wet with ninIn test tube by procedures hydrin reagent sohspecified tion 1 and developed like chromatograms (30 min. a t 68" to 70' C.) 5. Same as 4 but color disNo further color development solved in glacial acetic acid a Compounds (20 y ) placed on filter paper sheet.
1034
~
Effect of Various Conditions on Reaction of Proline and Pipecolic Acid with Ninhydrin
Pretreatment' 1. None 2. Paper
chromatogram
,
Maximum Amount of Color, % Pipecolic Proline acid 100 100 99
...
98
...
49
62
21
34
tube, the tube is loosely capped and heated 20 minutes in vigorously boiling rrater. The tube is cooled 5 minutes and 5 ml. of glacial acetic acid is added. After thorough mixing, the color is read a t 512 mp 0.5 t o 2 hours later. The rest of the tiyo-directional chromatogram is wetted completely with reagent 1 and heated for 30 minutes a t 68 to 70" C. in an alcohol-saturated atmosphere made anaerobic with carbon dioxide (33). The sheet must not have excess ninhydrin solution (as indicated by a sheen) when inserted into the heating chamber to avoid movement of spots during color development. The heating is accomplished in a special chamber 30 x 36 x 3 inches of stainless steel maintained a t 68" to 70" C. in a temperature-controlled box (33). A 3-inch pool of 95% alcohol containing citric acid is kept in the bottom and carbon dioxide is passed into the chamber through a perforated pipe. A plate over the pipe prevents the alcohol from splashing on the chromatogram. The carbon dioxide flow is maintained a t approximately 100 ml. per minute. The carbon dioxide is run through the chamber several hours before use or until the top of the chamber is anaerobic as tested with a glowing splinter. After heating, the paper is dried in moving air. As quickly as possible, the individual amino acid spots are cut out along with several blank areas. Each paper piece is weighed to the nearest milligram (Roller-Smith balance, RollerSmith Instrument Division) and cut in strips l / 4 x 3/4 inch and placed in a clean dry test tube (18 X 175 mm.). Ten milliliters of the 50% alcohol solvent is added and the tube is gently shaken. After 30 minutes the color is read a t 570 mp in a 1-cm. cuvette. A blank correction, based on the amount of color per milligram of blank paper, is calculated from the weight of each spot. The blank correction is subtracted from the original reading to give an absorbance due to the acid itself. The pipecolic acid spot is cut in '/,-inch squares into a test tube (18 x 175 mm.). The blanks and samples are treated as in the proline determinations. The tubes are heated in a boiling water bath for 15 minutes after the addition of 5 ml. of reagent 5. After rapid cooling, the solution is diluted with 5 ml. of glacial acetic acid. After 30 minutes in darkness, solutions are read a t 570 mp within 1 hour. The amount of amino acid in each case is determined from a standard curve obtained by running different amounts of a mixture of pure amino acids through the entire chromatographic and color development procedure.
Leucines and Phenylalanine. The color for these amino acids is developed in a n alcohol-saturated atmosphere in the same way as the neutral and acidic amino acids and blank areas are cut out in strips between t h e acids. Standard curves are obtained as before b u t the inclusion of standards with each set iinproves accuracy. Basic Amino Acids. These are reacted with ninhydrin the same as the leucines and phenylalanine using reagent 2. Inclusion of standards rvith each run is desirable. DISCUSSION
Color Production in Test Tubes.
Although the special determination of asparagine, proline, and pipecolic acid complicates the procedure, this is necessary t o reduce variability and increase sensitiyity. ilsparagine reacts poorly with ninhydrin on paper (20, 33) and maj7 react incompletely in the test tube (4, 3 7 ) . The two imino acids react better with ninhydrin under acidic conditions (95, 26). The presence of an acid reduces the blank color because amino acids and ammonia do not give a color with ninhydrin under acidic conditions (24). Because there is essentially no proline lost in locating it on the paper (Table 11, 2, 3), no interference by ammonia, and no fading of color in the dark in 24 hours, highly reproducible determinations can be made. The final color obtained from the other acids is not reduced by the color developed in locating proline and asparagine. Considering their similarity in structure, proline and pipecolic acid behave differently in their reaction with ninhydrin. Although both react under acidic conditions, the concentration of acid for optimum color and color produced are different. Whereas proline reacts readily with ninhydrin under mild conditions,. pipecolic acid requires a much longer time a t higher temperatures. Therefore, pipecolic acid can most conveniently be located on the chromatogram after quantitative development for the neutral and acidic amino acids. The color production under these conditions is poor (Table 11, 5 ) and variable. Although the loss of final color resulting from reaction on paper is not so great as for proline (Table 11, 4), it is still considerable. Fortunately, the final color is consistent and reproducible (Table 111) and higher, mole for mole, than for any other acid. The extinction coefficient for 1 pmole in 10 ml. is 4.89, as compared with a value of 2.17 (36) for the (Yamino acids. The pipecolic acid color is unstable, especially in bright light, so the samples should be kept in the dark and read within 1 hour. With a little experience, pipecolic acid can be
located accurately on the chromatogram by its position relative to proline without the development of any pipecolic acid color. This increases the sensitivity (Table 11) and reproducibility. Color Production on Paper.
Al-
though ninhydrin solutions are commonly sprayed on a chromatogram, for quantitative work i t is better t o pour the solution on the paper in a thin stream with a modified pipet t o cover the paper more evenly and thoroughly. A concentration of ninhydrin of 2% was adequate for maximum reaction (33). The inclusion of collidine and acetic acid in the ninhydrin solution causes several of the amino acids to react better (Table IV). Table Ill.
A temperature of 68" to 70" C. results in better color production than a t 60" C. (Table IV) without undue increase in the blank color or loss in reproducibility, b u t higher temperatures are unsuitable. A heating time of 30 minutes was chosen because most of the reactions are complete and longer times increase the danger of running of spots although color is higher for aspartic acid and the aromatic amino acids (Table IV). Most of the amino acids give essentially the theoretical amount of color (Table IV), the notable exceptions being aspartic acid, phenylalanine, and tyrosine. The incomplete reaction of aspartic acid and the aromatic amino acids has been obtained in other ninhydrin procedures ( 4 ) (Table 11)
Reproducibility in Determination of Pure Amino Acids a t Different Levels Determined after Chromatography
Amino Acid Aspartic Glutamic Serine Glycine Asparagine Threpnine Alanine Glutamine Histidine Lysine Arginine Methionine sulfoxide Proline Valine Isoleucine Leucine Phenylalanine Tyrosine 4-Aminobutyric Pipecolic Six determinations. Table IV.
Acid"
10
... ...
4.7 4.0
... ...
...
...
2.0 2.4
2.6
... ... *..
2.2 2.5
...
5.4 .
I
.
...
Amino Acid, y 20 30 40 60 Coefficient of Variation5 as Per Cent of Mean 2.3 ... 2.2 2.4 2.4 *.. 2.9 1.6 1.8 4.8 4.0 ... 2.5 4.0 5.5 ... 3.0 ... 3.8 3.8 0.7 3.2 ... 2.5 0.8 ... 2.0 0.9 3.0 4.2 2.7 ... ... 1.2 1.6 1.6 ... 0.6 0.8 1.o 2.5 ... 0.8 1.9 0.7 ... 4.0 3.7 0.6 2.9 ... 0.8 3.0 1.4 3.7 ... 2.7 ... 3.4 3.6 2.2 2.4 3.6 ... 2.9 ... 1.6 2.5 4.9 2.4 9.6 2.4 1.7 2.5 ... 2.4 2.2 0.6 1.7 ...
80 2.2
2.4
...
...
3.4 1.9
1.5
1.7
... , . .
...
1.8 1.7 2.2
... ... 2.4 ...
2.1
3.9
Effect of Various Factors on Reaction of Ninhydrin on Paper
Color Produced* as Per Cent of Theorye 2 3 4 5 79 89 58 io 98 ... 97 ..
6 Aspartic 53 Glutamic 84 ... Serine 81 100 ... 100 .. ... Glycine 78 100 100 100 Threonine Alanine Glutamine Histidine 90 107 Lysined 53 65 ... 97 104 65 104 Arginine 93 ... 124 Methionine sulfoxide 45 95 ... 78 si, ... Valine 88 99 ... 96 .. ... Isoleucine 86 97 ... 91 .. ... Leucine 89 98 ... 94 .. ... Phenylalanine 41 81 96 79 48 ... Tyrosine 40 80 95 78 54 ... 4-Aminobutyric 39 99 100 98 90 ... 40 y of each acid. * Col. 1. 30-min. heatingat 60" C. with ninh drin reagent 1. Col. 2. 30-min. heating at 68' to 70" C. with ninhydrin reagent 1. Jol. 3. 60-min. heating at 68" to 70" C. with ninhydrin reagent 1. Col. 4. 30-min. heating at 68" to 70" C. with no collidineacetic in ninhydrin reagent. Col. 5. 30-min. heating a t 68" to 70" C. with paper buffered at pH = 7. Coi. 6. 30-min. heating at 68" to 70" C. with ninhydrin reagent 2. Extinction coefficient for 1 pmole in 10 ml. a t 570 mb is 2.17 as determined on pure diketohydrindylidenediketohydrindamineby Troll and Cannan (35). Lysine values based on two amino groups. 1
I
.
.
I
.
(I
VOL. 31, NO. 6, JUNE 1959
1035
and a complete explanation is not available. Asparagine gives 90% of theoretical color. The theoretical amounts of color for proline and pipecolic acids have not been determined, but the methods used here are more sensitive for these imino acids than for the a-amino acids. Under the conditions used here, the basic amino acids react well with ninhydrin (Table IV, column 6). On the basis of both amino groups, lysine reacts completely on paper but only about 55% of theory in vitro (4, 37). The more complete reaction increases sensitivity without impairing accuracy. The reaction of histidine is more than can be accounted for by the amino group (Table IV) indicating that the ring reacts under these conditions. Arginine likewise renders over 100% reaction (on the basis of the amino group) indicating some reaction with the guanidino group. The color developnient procedure requires careful operation to keep the chromatogram moist while keeping the spots from running. Results have been satisfactory when a wet blotter was placed on one side of the chamber and the carbon dioxide put through a t a rate of 100 to 200 nil. per minute.
Table V. Effect of pH of 50% Ethyl Alcohol on Fading Rate of Color Produced by Reaction of Ninhydrin with Amino Acids on Paper Buffered with 0.01 M Sodium Phosphate Loss of Color in 33 Hr. as ?’ & of Initial
Solvent 50% CzHjOH, pH 5.5 50% CZHbOH-NaOH added until pH 6.9 50% CiHjOH containing 27, collidine, pH 6.9
Table VI.
Value 21 3
0.5
RESULTS
The coefficient of variation for the estimation of the various amino acids by these procedures indicates that a variation of 5% or less is obtained with all amino acids (Table 111) and less than 391, for most acids. The variability found is less than that reported
Amino Acid Composition of P-Lactoglobulin -4mino Acid aer 100 Grams of Protein. Grams
Present method
(27) (14) (8) Aspartic acid 11 , 4 11.5 11.3 11.4 Gluhrnic acid 18.5 19.1 1 8 . 4 19.5 Serine5 3.5 5.0 3.8 4.0 1.4 1.4 Glycine 1.5 1.4 5.2 Threonine” 4.7 4.9 5.8 Alanine 7.1 7.1 6.2 5.8 1.6 1.5 1.6 1.6 Histidine Lysjne 11.6 12.6 11.2 1 1 . 4 2.9 Arginine 2.5 2.9 2.9 3.1 3.2 3.1 Methionine 3.2 4.1 5.0 Proline 4.8 5.1 6.2 Valine 5.5 5.6 5.8 Isoleucine 5.9 8.4 7.3 5.9 Leucine 14.9 15.5 1 5 . 2 15.6 3.6 Phenylalanine 3.3 3,8 3.5 Tyrosine 3.7 3.6 3.9 3.8 Corrected for losses in hydrolysis (WI).
1036
The chromatogram is dried and processed as rapidly as possible after the heating period, t o prevent increase in blank color. The paper niust be handled carefully with clean hands or gloves to avoid addition of ninhydrinreactive compounds. Once the 50% alcohol solvent has been added to tubes, there is no further color development. One two-directional chromatogram can usually be handled in 30 minutes. The addition of collidine to the 50% alcohol reduces the fading rate (Table V) by maintaining the colored compound in the salt form. Background color is caused by animonia and other compounds (36) on the paper. This color increases \yith the water content of the alcohol in the color development chamber and this alcohol should be replaced weekly. On two-directional chromatograms the background color is sufficiently low and uniform that no special care in choice of blank areas is required to obtain reproducible results. I n the one-directional runs, the blank color is higher and it is important to obtain blank strips between the different amino acid spots. Attempts to remove ammonia by alkalinizing the paper resulted in greater variability and lorn color. The method of Connell, Dison, and Hanes ( 4 ) eliminates interference by ammonia but is more difficult to perform ivith lo^ amounts of amino acids and is more laborious than the methods outlined here. These methods permit the complete analysis of 24 samples with 80 man-hours of work and 12 days of elapsed time.
ANALYTICAL CHEMISTRY
(63) 11.5 19.8 3.9 1.6 4.9 6.9 1.7 13.0 2.8 3.0 5.1 5.6 5.6 15.7 3.6 3.6
(17 )
11 10 3 1 4 6
2 7 4 4 6 3
I .o 11 2 2 6 2 9 5 0 4 5 5 1 10 75 3 4 3 9
ACKNOWLEDGMENT
(30) ...
(34)
(11)
...
...
...
... ...
. .
...
1.56
4.6
...
1.5 11.1 2.8 2.5
...
5.5
1 3 4.3
...
by Roberts and Kolor (22) for a niaximum density method using seven solvents. Tyrosine shows a relatively high coefficient of variation because it streaks on chromatograms and reacts poorly with ninhydrin (Table IV). The comparatively poor reproducibility for serine and glycine is due to difficulties in separation because the sum of the two acids has a lower variation. Separations of up to 80 y of each amino acid are readily obtained except for glycine, serine, and the leucines. Separations of these acids are not satisfactory a t the 60- and 80-7 levels and this accounts for the omission of these data from Table 111. Standard curves of the amino acids (except tyrosine) deviate only slightly from Beer’s law over the ranges tested. This is to be expected, because essentially no amino acids are lost in chromatography (Table I) and most react ne11 with ninhydrin. The slopes of standard cuves can be approximately calculated from the data of Table IV, columns 2 and 6. The reason for the failure of tyrosine to give a straight line is not known, but may be related to the fact that it streaks badly a t higher levels in but)] alcohol-acetic acid. As a further test of the method, :in analysis of 6-lactoglobulin was performed (Table VI). The published data of Stein and Moore (2‘7),Brand et al. ( 3 ) ,and Keston, Udenfriend, and L e y (11) are probably the most reliable. With these values as a criterion, the chromatographic data of column 1 agree reasonably well in d l cases indicating that these procedures are reliable. The greatest deviations occur for glutainic acid, serine, and leucine. Talues in column 1 were obtained consistently and are believed to be correct for the sample of @lactoglobulin used. The evidence ( I ) for seyeral forins of @-lactoglobulin may mean somen-hat different amino acid contents depending on the source of protein.
... ...
6.1
... ,..
... ...
4.3 4.4 17.1 2.3
...
...
... ... ... ... ... 4.9 ...
7.05
...
...
...
The authors acknowledge the assistance of Rose Cering in parts of this work. W. G . Gordon very kindly donated a sample of pure P-lactoglobulin, LITERATURE CITED
(1) Aschaffenburg, R., Drewry, J., Suture 176, 218 (1955).
(2) Block, R. J., Durrum, E. L., Zmeig, G , “Manual of Paper Chromatography and Paper Electrophoresis,” chaps. 4, 5, -4cadeniic Press, New York, 1955. (3) Brand, Erwin, Saidel, L. J., Goldwater, W. H., Kassell, Beatrice, Ryan, F. J., J.Am. Chem. SOC.67, 1524 (1945). (4) Connell, G. E., Dixon, G. H., Hanes, C. S., Can. J . Biochem. Physiol. 33, 416 (1955).
(5) Crumpler, H. R., Dent, C. E., A-uture 164, 441 (1949). (6) .Fisher, R. B., Parsons, D. S., Morrison, G. A., Zbid., 161, 764 (1948). (7) Fowden, L., Biochem. J . 50, 355 (1952). (8) Hardy, T. L., Holland, D. O., Nayler, J. H. C., ANAL.CHEM.27,971 (1965). (9) Isherwood, F. A., Hanes, C. S., Biochem. J . 55, 824 (1953). (10) Kay, R. E., Harris, D. C., Entenman, C., Arch. Bwchem. Biophys. 63, 77 (1956). (11) Keston, A. S., Udenfriend, Sidney, Levy, Milton, J . Am. Chem. SOC.69, 3151 (1947). (12) Lederer, E., Lederer, M., "Chromatography-Review of Principles and Applications," Chap. 30, Elsevier, New York, 1957. (13) Levy, A. L., Chung, David, ANAL. CHEW25, 396 (1953). (14) ~, Lewis. J. C.. Shell. N. S.. Hirschniann, D. J., ' Fraenkel-Conrat, H., J . Biol. Chem. 186, 23 (1950). (15) Mansford, K., Raper, R., $nn. Botany (London) 20, 287 (1956).
(16) McFarren, E. F., ANAL.CHEM.23, 168 (1951). (17) McFarren, E. F., Mills, J. A,, Zbid., 24, 650 (1952). (18) Moore, Stanford, Spackman, D. H., Stein, W. H., ANAL.CHEM.30. 1185 (1958). (19) Moore, Stanford, Stein, W. H., J . Biol. Chem. 211, 893 (1964). (20) Porter, C. A., Margolis, D., Sharp, P., Contribs. Boyce Thompson Znst. 18, 465 (1957). (21) Rees, M. W.,Biochem. J . 40, 632 (1946). (22) Roberts, H. R., Kolor, M. G., ANAL.CHEM.29, 1800 (1957). (23) ,Roland, J. F., Jr., Gross, A. M., Zbzd., 26, 502 (1954). (24) Schlenker, F. S., Zbid., 19, 471 (1 4471
(2jj-sSchweet, R. S., J . Biol. Chem. 208. 603 (1954).' (26) Silberstein, 0. O., Adjarian, R. Thompson, J. F., A N A L . CHEM. 855 (1956). (27) Stein, W. H., &loore, Stanford Biol. Chem. 178, 79 (1949).
(28) Steward, F. C., Pollard, J. K., Ann. Rev. Plant Physiol. 8, 65 (1957). (29) Steward, F. C., Zacharius, R. ?*I., Pollard, J. K., Ann. Acad. Sn'. Fennicae Ser. '4 ZZ 60, 321 (1955). (30) Stokes, J. L., Gunness, M., Dayer, I. M., Caswell, M. C., J . Biol. Chem. 160, 35 (1945). (31) Thompson, J. F., Morris, C. J., Gerine. R. K.. ANAL.CHEM.31. 1028 (1959r (32) Thomuson, J. F., Steward,' F. C.. Plant PhyswZ: 26, 42i (1951). (33) Thompson, J. F., Zacharius, R. M., Steward. F. C.. Zbid.. 26. 375 (1951). (34) Tristiam, G. R.; B'iochek. J.' 40, 721 (1945). (35) Troll, W.,Cannan, R. K., J . Bid. Chem. 200, 803 (1953). (36) Wynn, V., A-ature 164, 445 (1949). (37) Yemm, E. W., Cocking, E. C., Snalyst 80, 209 (1955). '
RECEIVEDfor review July 24, 1958. Accepted January 30, 1959.
Thermistorized Apparatus for Differential Thermal Analysis Application for Determination of Thermograms of Nitrate Esters of Cellulose and Pentaerythritol JACK M. PAKULAK, Jr., and GUY WILLIAM LEONARD U. S. Naval Ordnance Test Station, China Lake, Calif.
b
Application of differential thermal analyses to the study of organic compounds is becoming increasingly important. Because the characteristic portions of the thermograms for most organic compounds lie in the 20" to 300" C. range which falls within the useful range of thermistors, the use of thermistors in differential thermal analysis was studied. By placing two matched thermistors with their parallel shunts in adjacent arms of a bridge, and feeding the output from the bridge to a recorder, a simple and very sensitive apparatus can be achieved for tracing differential thermograms. Its applicability is shown by thermograms for certain inorganic and organic compounds.
B
thermistors are a very sensitive means for measuring and controlling temperature (3), their substitution for thermocouples in differential thermal analysis (DTA) was investigated. The application of a thermistor as a temperature-sensing device stems from its extremely large change in ECAUSE
resistance with very small changes in temperature (6). Sensitive measurements can be made with thermistors by using a bridge circuit. Thermistors of comparatively high resistance, 100,000 ohms or greater, are used in this apparatus, so that even a t 200" C. a satisfactory resistance is maintained. A vacuum tube voltmeter was employed t o select thermistors whose resistances matched within 1% a t room temperature. The matched thermistors are located in the adjacent arms of a direct current bridge. By inserting one thermistor into a sample of a n inert reference material of aluminum oxide and the other thermistor into the sample material under investigation, the temperature differential b e t r e e n the two samples can be measured by the unbalance of the bridge. Only negligible drift is present between these two thermistors, when both are immersed in the same inert reference material and heated t o elevated temperatures a t a constant rate. The calibration of the third thermistor for temperature is obtained by a temperature-resistance relationship with standard thermom-
eters. This thermistor is used for the temperature-indicating electrode in the differential thermal analysis apparatus. The bead type of thermistor used in this work was highly stable, and may be continuously operated up to 300" C. DESIGN O F DIFFERENTIAL APPARATUS
The Western Electric thermistors used for the sensing elements have a temperature coefficient of approximately -4% per O C. at 25" C. Figure 1, -4,is a graph of resistance us. temperature of the thermistor. Because this curve has a logarithmic nature, using the thermistors directly in a measuring circuit would result in a n extremely poor temperature scale. Readings a t one end of the scale would show a very small resistance differential for a giren temperature differential; those at the opposite end would have very large resistance differentials. Selection of Shunt. T o obtain a more linear scale, t h e thermistors a r e shunted n i t h a suitable resistance. Although t h e s h u n t reduces t h e sensitivity of t h e thermistor, t h e sensitivity was more t h a n sufficient for this inVOL. 31, NO. 6, JUNE 1959
* 1037
