Chromatographic Separation and Determination of Straight-Chain Saturated Monocarboxylic Acids C, through Cl0 and Dicarboxylic Acids C,, through CI6 GERALDINE B. CORCORAN Research Division, Armour and Co., Chicago 9,
111. ( 7 ) proved unsuitable as a supporting medium for the aqueous phase of the chromatogram because of its adsorptive properties. Very pure gel ( 3 ) exhibits similar properties, wggesting that the adsorption is connected with the structure of the gel rather than with the presence of impurities. The method described here for the preparation of a nonadsorbent silicic acid is similar to the method of Isherwood (g), and is based on the fact that treatment n-ith 10S hydrochloric acid a t room temperature for a period of 24 to 36 hours eliminates its ability to adsorb organic acids. Mallinckrodt's silicic acid is suspended in 10-V hydrochloric acid for the afore-ment'ioned period. The supernatant liquid is then decanted and the silicic acid washed with water to remove hydrochloric acid. When acid-free, the gel is washed n-ith absolute methanol until the filtrate is neutral to litmus. It is then washed with anhydrous ether and finally dried in vacuo over phosphorus pentoside. Two to three days and several changes of phosphorus pentoxide are necessary before the silicic acid is sufficiently dry. I t is then stored in airtight glass jars. Preparation of the Aqueous Phase. ;1 2-12 stock solution of the glycine stationary phase is prepared and st'ored in a refrigerator. The buffer at pH 2 is prepared by adding 0.5X hydrochloric acid to part of this solution, using the Beckman Model G pH meter and external electrodes. Inasmuch as pH is critical it is necessary to restandardize the meter before each titration.
A partition chromatographic method for the accurate analysis of mixtures of monocarbovylic acids CI through CIOand dicarboxylic acids C11 through Cle has been developed. The stationary phase is 2M aqueous glycine on silicic acid adjusted to the desired pH with either 0 . 3 hydrochloric acid or concentrated sodium hydroxide; the mobile phase is 1-butanol-chloroform in giadient elution. A series of three columns, with stationary phases of pH 2, 8.4, and 10, is employed for analysis of mixtures of monocarboxylic acids Ci through CIOand a series of two columns, with stationary phases of pH 8.5 and 9, is necessary for analysis of mixtures of dicarboxylic acids CIIthrough Cia. Excellent resolution and analytical accuracy within 3 ~ 1 %are obtained for all components.
A
SURVEY of existing methods of analysis for mixtures of fatty acids disclosed that no simple reproducible method had been devised which would separate a homologous series of monocarboxylic acids differing by one methylene group. Moyle's (4)phosphate-buffered column and Zbinovsky's (9) Cellosolvewater-Skellysolve B-n-butyl ether method proved to be unsatisfactory for quantitative work because of poor resolution and incomplete recovery. Other methods for separation of acids differing by two methylene groups by Silk ( 8 ) , Peterson ( 6 ) , Nijkamp ( 5 ) ,and Howard ( I ) were not adjustable to separation of acids differing by one methylene group. A need for such a method applicable on a small scale has long been felt. Current work in this laboratory required a method which would accomplish this purpose with an accuracy approaching 100%.
C-4.5.6.7.9.9.10
2.0 Cd
1.0
x
APPARATUS AND REAGESTS
C.,
The following apparatus is needed: Beckman Model G pH meter with external electrodes. Technicon fraction collector equipped with drop counter, Technicon, Inc. Glass chromatographic columns, 18 mm. in inside diameter, 600 mm. long with delivery tube and stopcock. Glass plunger. Automatic 10-ml. microburet, Scientific Glass Apparatus Co. Dry nitrogen, high purity. Reducing gage equipped with double needle valve. Volumetric pipets. Analytical balance. Vacuum desiccator.
005 004.
c.2 rJ-
0 2. 01 0 Y
4t .
0 L
I2
16
20
24 29 32 36 FRACTION NUMBER
40
a4
48
52
56
SO
Figure 1. Analysis of monocarboxylic acids C1 through C ~using O glycine stationary phase at pH 2
The stationary phases higher in pH than G 5 are prepared by adding concentrated sodium hydroxide to the 2Af glycine solution. One drop of a 50% solution of Arquad H T (trimethyloctadecylammonium chloride) in isopropyl alcohol is added to each 150 ml. of the aqueous glycine as a fungicide and bactericide. Preparation of the Column. In preparing the rolumn, 22 ml. of the appropriate glycine phase and 25 grams of the hydrochloric acid-treated silicic acid are ground together in a beaker for 5 minutes, using a test tube as a pestle. Seventy-five milliliters of the first eluent, 1% l-butanol-99yo chloroform, are introduced gradually with stirring to form a smooth slurry. -4 pad of glass wool inserted in the neck of the chromatographic column retains the silicic acid. The slurry is added in small increments and packed with a glass plunger, and the excess solvent is allowed to drain. The usual precautions are observed t o eliminate entrapped air.
These reagents are needed: 1-Butanol, C.P. grade. Chloroform, reagent grade. Silicic acid, Mallinckrodt chromatography grade, 100 mesh. Standard methanolic sodium hydroxide, approximately 0.03
_-n n i v.
tn
Glycine buffer, Pfanstiehl Chemical Co. aminoacetic acid. Adjust aqueous solution to proper pH with concentrated sodium hydroxide or 0.5N hydrochloric acid. m-Cresol purple, 1% in 95% methanol. PROCEDURE
Preparation of Silicic Acid. Silicic acid as prepared by Mallinckrodt according to the method of Ramsey and Patterson
168
169
V O L U M E 2 8 , N O . 2, F E B R U A R Y 1 9 5 6
pounds per square inch), sufficient to maintain a drop rate of approximately 106 drops per minute. A total of 200 ml. of each of three eluents is used, the concentrations of butanol being 1%, lo%, and 25% in chloroform. Sixty fractions are collected and each fraction is titrated with approximately 0.03N sodium hydroxide in absolute methanol, with 1% m-cresol purple in 95y0 methanol used as the indicator. A stream of nitrogen is used for agitation during the titration to exclude carbon dioxide from the air. The time necessary for the completion of one column is approximately two hours. RESULTS ARD DISCUSSIOR
FRACTION NUMBER
Figure 2. Analysis of monocarboxylic acids C1 through CIOusing glycine stationary phase at pH 8.4
0
a/
I
c.9
c-IO
I cn
n
1
0
Analysis of a mixture of the series of ten monocarboxylic acids from formic through capric is effected by the use of three columns with pH 2, 8.4, and 10 (Figures 1, 2, and 3). Because of the relative strength of formic acid, 2 is the optinium pH a t which it could be recovered in this system. Acetic and propionic acids are also recovered separately at this pH but acids from Cathrough GOare eluted in one peak (Figure 1). Figures 2 and 3 illustrate the effect of increasing the p H of the stationary phase. The more strongly acidic constituents of the series are retained on the column. I n each case, the same quantity of solvent and ratio of butanol to chloroform are maintained.
4
8
12
'6
20
2*
28
I
32
36
40
44
48
52
I
60
56
FRACTION NUMBER
Analysis of monocarboxylic acids CI through C ~ using O glycine stationary phase at pH 10
Figure 3.
0
Preparation of Sample and Operation of Column. Chloroform solutions of essentially pure monocarboxylic acids were prepared and a composite solution of all the acids was made, varying the concentration of each to simplify identification upon elution.
Table I . Results of Chromatographic Analysis of Rronocarboxylic Acids CI through CIO hlonocarboxylic Acids Cl
Charge Meq. Mg.
0.0512 0.0334 0.0774 0.0872 0.0444 C, c, 0.0152 c0.0198 0.0293 CS 0.0236 c9 0 0168 ClO TOTALS 0.3683
c2 Ca C4
-
2.36 2.05
5 72
5.02 4.53 1.76 2.58 4.23 3.74 2.89
Recovery Meq. Mg. 0.0514 0.0331 0.0779 0,0578 0.0450 0.0153 0.0197 0.0298 0.0244 0.0168 0.3712
2.37 1.99
5.77
5.09 4.59 1.77 2.56 4.29 3.85 2.89 35 1;
- 34.88
Difference, Mg. +0.01 -0.06 +0.05 f0.07
4
6
12
16
20
24
28
36
32
40
44
48
52
I
60
56
F R A C T I O N NUMBER
Figure 4.
Analysis of monocarboxylic acids Cl through Cs using glycine stationary phase at pH 6.5
The versatility of the method is demonstrated in Figures 4 and 5. By varying the pH of the stationary phase, it is possible to resolve any desired group of the afore-mentioned acids. I t is necessary only to determine the correct pH a t which the>- may be recovered and resolved. I n general, as molecular weight increases and acid strength decreases the pH of the column must be increased. The mixture chromatographed at pH 6.5
+0.06 +O.Ol -0.02 +0.06
+o.
11
0
2 5 % 1 . S U O 3 . 7 ) ~ CW'L3
10*.I~WOY.907.CHCL3-
An aliquot of the composite sample equivalent to a maximum of 35 mg. of monobasic acid is carefully pipetted onto the top of the column and allowed to percolate into the silicic acid. Simultaneously, the Technicon fraction collector is set in operation and collection of 400-drop or 10-ml. fractions is begun. After the sides of the tube are washed several times with small quantities of the first eluent and the washings are allowed to percolate into the column, the tube is filled with the 1% l-butanol-99% chloroform eluent and nitrogen pressure is applied (about 2
c.7
0
4
8
12
16
20
24
28
32
F R I C T I O N NUMSER
36
40
46
48
'
52
'
16
1 so
Figure 5. Analysis of monocarboxylic acids CI through C ~ using O glycine stationary phase at pH 9.6
170
ANALYTICAL CHEMISTRY
(Figure 4) contained acids formic through valeric. Formic acid remained on the column. The data obtained by analyzing the mixture of monocarboxylic acids formic through capric are shown in Table I.
Table 11. Results of Chromatographic Analysis of Dicarboxylic Acids C11 through Cle Dicarboxylic Acids Cll Cia ClS c 1 4
CIS CIS
Charge
Meq.
Mg,
0.0066
0.72 0.76 1.48 3.58 2.50 2.52 14,54
0.0066 0 . 0367 0.0276 0.0184 0.0176 __ TOTALS 0.1136
-
Recovery Meq. Mg. 0.75 0.0070 0.0073
0.0367 0.0270 0.0180
Difference, hIg.
0.84 4.48 3.49 2.45 2.52 14.53
0.0176 -_ 0.1136
series, an average titration error of 0.0016 ml. per fraction is indicated, which is within experimental error. Because dibasic and monobasic acids are often present in the same mixture, it was necessary to determine the interferences contributed by dibasic acids. 1 typical mixture of C, with C? through C12 dibasic acids was prepared and chromatographed with the monobasic acids a t the same pH ranges necessary for complete analysis of the monobasic acids (Figures 6, 7 , and 8). At pH 2 (Figure 6) the Cq through Clo monobasic acids were eluted in one peak in the enme position as when mono acids alone were present.
+0.03 +0.08 0 -0.07 -0.05 0
The optimal load needed to obtain perfect resolution appears to be approuimately 35 mg. On a total of 20 determinations the recovery slightly exceeded the charge. This may be attributed partially to the fact that solvents were not redistilled and partially to experimental error. Inasmuch as a total of 180 10-m1. fractions is collected for complete analysis of the C1 through Clo
I
1
C.3 MCNO ACiD
C - 4 MONO AClO
0 21
4
8
FRLCTION NUM0ER
Figure 8. Analysis of a mixture of monocarboxylic acids CI through Clo and dicarboxylic acids C4 and CI through CU using glycine stationary phase at pH 8.5
The dicarhox) lic acids ivere eluted in one peak folloxed by CI, C2, and C1 monobasic acids which were completely resolved. The threshold volumes of these lower molecular n-eight acids are shifted as much as 220 ml. by the presence of dibasic acids, necessitating the collection of 78 fractions in order to elute formic acid completely.
FRACTION NUMBER
Figure 6. Analysis of a mixture of monocarboxylic acids C1 through Clo and dicarboxylic acids C4 and C7 through Clz using glycine stationary phase at pH 2
C-15
-
1% I . B U O H . ~ ~ ~ C W C L ~
030.
-
10%I-BUOH-~OZ cncL,
25% I - B U O H - ~ S C~ H C L ~
-
c-IS
c- I3 4
I 020.
__ 2 5 % IbBUOH-75%
1 0 %1 . ~ u o n . 9 0CHCL ~
CHCL
c.14
-
0 0
2
010.
0 c.4 M O N O ACID c.11 DI ACID
C.5 MONO AClO
4
6
1 I2
I6
20
24
28
32
36
40
44
48
52
56
0
e c-3 MONO 4 C l O
R
C.12 01 bClD
Figure 9. Analysis of dicarboxylic acids C13 through CIS using glycine stationary phase at pfI 9
ACID
0.1
FReiCTlON NUMBER
Figure 7. Analysis of a mixture of monocarboxylic acids C1 through CIOand dicarboxylic acids C, and C7 through Clz using glycine stationary phase at pH 8.4
rlt pH 8.4 (Figure 7 ) , the threshold volumes of the monocarboxylic acids combined with dicarboxylic acids are the same as those with only monocarboxylic acids present. C12 dicarboxylic acid, however, is eluted with Cj monocarboxylic acid;
V O L U M E 28, N O . 2, F E B R U A R Y 1 9 5 6 and C, dicarboxylic acid with C, monocarboxylic acid. T o separate the monobasic from the dibasic acids, it is necessary t o increase the p H of the stationary phase t o 8.5 and change the butanol ratios slightly (Figure 8). The change in p H of 0.1 gives excellent resolution of both mono- and dicarboxylic acids. It \vas apparent that the glycine method was equally effective iri +eparating both mono- and dicarbo\ylic acids; therefore, an attompt was made to determine the range of dicarboxylic acids which could be separated effectively. The available supply of higher molecular weight dibasic acids included those from C1, through C16. Excellent resolution and recovery are obtained n i t h a column at pH 9 for the CI3 to cl6 dicarboxylic acids (Figure 9). CII and Cl2 acids are resolved on a column a t p H 8.5 (Figure 8). The conflicts of C1, to cl6 dibasic acids with the monobasic acid peaks were not determined because of the extremely short supply of dibasic acids. Data pertaining to recovery of dibasic acids are shown in Table 11. The author believes this method is applicable t o the deter-
171 mination of dicarboxylic acids through C ~ merely O by increasing the p H of the stationary phase. ACKKOW LEDGMENT
The author is indebted to Stewart Leslie for preparation of the drawings. LITERATURE CITED
(1) Howard. G. a,, and Nartin, A. J. P.. Biochem. J . 46, 532 (1950). (2) Isherwood, F. A , , Ibid.,40, 688 (1946). (3) Kargin, V. d.,and Rabinoyitch, d. J., T i a n s . Faraday SOC.31, 284 (1935). (4) AIoyle, V., 1-aldwin, E., and Scaresbrick, R., Bwchem. J. 43, 308 (1948). (5) Nijkamp, H . J., A n a l . Chim. A c t a 5, 325 (1951). (0) Peterson, 31. H., and Johnson, 11. J., J . B i d . Chem. 124, 775 (19481. (7) R a k e y , L. L., and Patterson, T. I., J . Assoc. O B c . A g r . C'imnists 31, 139 (1948) ( 8 ) Silk, A l . H., Biochern. J . 56,400 (1954). (9) Zbinovsky, V., ~ ~ N I LCHEV. . 27, 764 (1955). RECEIVED f o r review August 29, 1955.
.\cceptrd Kovernber 2 5 , 1955.
Chromatography of AI pha-Keto Acid 2,4=Dinitrophenylhydrazones and Their Hydrogenation Products ALTON MEISTER' and PATRICIA A. ABENDSCHEIN Laboratory o f Biochemistry, National Cancer Institute, National Institutes of Health, Bethesda,
The 2,4-dinitrophen>lhjdrazonesof a series of 35 a-keto acids (including the a-keto acid analogs of most of the common naturally occurring amino acids) have been prepared and studied in several paper chromatographic sjstenis. The hydrazones (as prepared, or after elution from paper chromatogranis) were hy drogenated under pressure in the presence of platinum oxide cataljst. IIy drogena tion of the h j drazones of the a-lieto acid analogs of all the amino acids studied (except cysteine) gave the corresponding amino acids, which were identified chromatographicallj In sei era1 instances, more than one amino acid resulted from hydrogenation of a single keto acid hjdrazone. The present results emphasize the value of combining chromatographic and hydrogenation procedures. The information presented here should serve as a useful basis for the identification of' a wide variety of a-keto acids.
.
A
LTHOUGH considerable attention has been given to the chromatographic identification of amino acids, relatively little chromatographic information has been collected concerning their a-keto analogs. Several a-keto acids have been found in nature-e.g., pyruvic, glyoxylic, oxalacetic, a-ketoglutaric, or-ketoisovaleric, a-keto-r-methyleneglutaric-and it is probable that application of more sensitive procedures xi11 disclose the presence of others. A fe:v ,tudies on the paper chromatographic behavior of free a-keto acids have appeared (13, 27, S I ) . However, because of the instability of many a-keto acids during isolation procedures, and the fact that relatively large quantities of the free acids are usually required for identification on the chromatograms, a number of investigators ( 1 , 4, 20, 28, 69) have resorted to the use of the corresponding 2,l-dinitrophenylhydrazone derivatives. Although chromatography of the 2,1-dinitrophenylhydrazones 1 Present .4ddress, Department of Biochemistry a n d Nutrition, Tufts l n i v e r s i t y School of RIedicine, Boston, Rlass.
Md.
of a-keto acids is often a valuable tool for the identification and even for the quantitative determination (3, 5, 12, 23, 26) oa-keto acids, several difficulties exist. For example, the chrof matographic behavior of certain 2,4-dinitrophenylhydrazones is very similar; furthermore, under some circumstances an a-keto acid 2,klinitrophenylhydrazone ma>- give rise to two spots 011 one-dimensional paper chromatograms (and often to four spots on tivo-dimensional chromatograms). This is probably due t o the presence of the syn- and anti-hydrazones (IO,20, 2 3 ) . Kuionen ( 1 1 ) used a procedure for the hydrogenation of the hydrazones using aluminum amalgam, followed by paper chromatography of the resulting amino acids. Ton-ers, Thompson, and Stelvard ( 2 5 )independently developed a similar method based on catalytic hydrogenation ivith platinuni oxide. These techniques have proved of value in the identifieation of several a-keto acids present in blood, urine ( 1 1 ) , and certain plant tissues ( 2 5 ) . Studies in this laboratory on transamination and related problems have necessitated the use of a large number of m-keto acids, which have been prepared by synthetic organic techniques and by enzymatic oxidative deamination of the Corresponding amino acid isomers (15, 1 8 ) . I n the course of these investigations the authors have had occasion to use paper chromatography for the identification of a-keto acid hydrazones, and to carry out hydrogenation of these derivatives. This paper reports paper chromatographic studies of the 2,bdinitrophen>-lhydrazones of 35 a-keto acids, and of the products of hydrogenation of these compounds. This is a considerably larger series of aketo acids than has previously heen available for such study, and includes the a-keto analogs of most of the common naturally occurring amino acids. It may therefore be expected that the data presented here xi11 be useful to those concerned with the identification of a-keto acids. XlETHODS
References t o the methods of preparation of the a-keto acids, and the solvents used for crystallization of the 2,kiinitropheni 1-
