An existing method for choline analysis .... placed in micro-Kjeldahl flasks and ... Table I. Factors for Estimating Quaternary Nitrogen Compounds from Periodide.
The use of the term fi, in contrast t o fl, has not been entirely satisfactory. This is especially obvious in comparing the experimental deviations obtained b y using Equations VI and VII. The possibility exists [as first pointed out to the authors by Kieselbach ( S ) ] that the use of Ep,v,fl rather than Cv,fi would prove more satisfactory in Equation VII. This follows from the fact t h a t fl is a slowly varying function differing only slightly from unity a t any time. The use of unity in place of fl, Equation VI, is, of course, very satisfactory. This explanation would constitute a n argument in favor of the predominance of the E term. Some other factor might, however, be responsible for the deviations. The plate height contribution of the injection and detection devices cannot be held responsible, because the trend from larger to smaller values would be reversed,
and this effect is as large as 10% only in extreme cases (as shown b y measuring the peak width for short, blank columns, and extrapolating to zero length), and is usually much smaller. ,4 great deal more experimental and theoretical work is needed on this subject. ACKNOWLEDGMENT
Assistance from the University of Utah Research Fund for the purchase of equipment is acknowledged. LITERATURE CITED
(1) Baker, W. J., 2nd Biannual Inter-
national Gas Chromatography Symposium, Michigan State University, 1959. (2) Bohemen, J., Purnell, J. H., “Gas Chromatography,” D. H. Desty, ed., p. 6, Butterworths, London, 1958. (3) Giddings, J. C., Nature 184, 357 (1959).
(4) Glue$auf, E., “Gas Chromatography, D. H. Desty, ed., p. 33, Butterworthe, London, 1958. (5) Glueckauf, E., “Vapour Phase Chromatography,” D. H. Desty, ed., p. 29, Butterworths, London, 1957. (6) Golay, AI. J. E., “Gas Chromatography,” D. H. Desty, ed., pp. 36, 65, Butterworths, London, 1958. (7) James, A. T., Martin, A. J. P., Biochem. J . 50, 679 (1952). (8) Kieselbach, R., E. I. du Pont de Nemours 8i Co., Wilmington, Del private communication. (9) LittlSt,u-ood, A. B., “Gas Chromatography, D. H. Desty, ed., p, 35, Butterworths, London, 1958. (10) Stewart, G. H., Seager, S. L., Giddings, J. C., -4NAL. CHEW31, 1735 i1959). (11) van Deemter, J. J., Zuiderweg, F. J., Klinkenberg, A,, Chem. Eng. Sci. 5 , 271 (1956). ~
RECEIVED for review November 12, 1959. Sccepted April 4, 1960. Work supported by the U. S. Atomic Energy Commission under Contract AT-(11-1)-748.
Spectrophotometric Determination of Betaines and Other Quaternary Nitrogen Compounds as Their Periodides JOSEPH S. WALL, DONALD D. CHRISTIANSON, ROBERT J. DIMLER, and FREDERIC R. SENTI Northern Regional Research laborafory, Peoria, 111.
b A rapid sensitive procedure was required for the detection and determination of betaines and other quaternary nitrogen compounds in the effluent of chromatographic columns employed for their separation from mixtures. An existing method for choline analysis was modified to determine additional substances by spectrophotometric measurement of their periodide derivatives. This was possible because ethylene dichloride solutions of the periodides of all quaternary nitrogen compounds tested including betaine, trigonelline, carnitine, stachydrine, and y-butyrobetaine absorbed in the ultraviolet with maxima a t 365 and 295 mMu. Acidification of the medium in which the periodides of the betaines were prepared by reaction with iodine resulted in their more complete precipitation. Iodine concentration, certain anions, temperature, and time also affected periodide yields. HE DEVELOPMENT of column chroT m a t o ,uraphic methods for the ready separation of many quaternary nitrogen compounds has created the need for a general procedure t o detect and to
870
ANALYTICAL CHEMISTRY
determine compounds of this class in the effluent fractions. Rerle ( 1 4 ) has summarized the numerous procedures employed for the analysis of individual quaternary nitrogen compounds; however, none of these methods directly fulfilled the requirements of speed, simplicity, and general application necessary for use in column chromatography. Friedman et al. (7’) have described certain of the difficulties encountered upon employing some of these existing methods in chromatographic analysis and emphasized the need of better methods for this purpose. Iodine is a sensitive precipitating agent for choline and yields similar water-insoluble complexes with other quaternary nitrogen compounds (11). Thiosulfate titration of the combined iodine has been preferred for measuring the yields of periodides of choline (6), betaine ( 2 ) ,and trigonelline (9). Appleton et al. (1) observed that the periodide of choline in ethylene dichloride solution exhibits an absorption maximum in the ultraviolet a t 365 mp and applied this finding to the determination of choline. A similar absorption spectrum has been reported for tetramethylammonium periodide (3). Studies were under-
taken to ascertain if periodides of other quaternary nitrogen compounds exhibit absorption in the ultraviolet and whether such absorption might provide a basis for their determination. Experiments r e r e also carried out to determine the conditions necessary for the precipitation of the periodides in such yields as to enable the determination of microgram amounts of the bases. MATERIALS AND METHODS
Standard Quaternary Nitrogen Compounds. Choline chloride, betaine hydrochloride, tetramethylammonium chloride, and thiamine chloride were Eastman Grade reagents, dried over desiccant. ?-Butyrobetaine and DL-carnitine were obtained from the International Minerals and Chemicals Corp., and ergothioneine. HC1 from the California Foundation for Biochemical Research. Trigonelline, HC1 ( I O ) , stachydrine. HC1 (6), and N-methylnicotinamide chloride ( 8 ) mere prepared according to references cited. All compounds were recrystallized from suitable solvents until melting points agreed with literature values. The identity and purity of these substances were further confirmed by chro-
12-
10-
1.5-
20.8 0
v
-
Y I
g0.b-
u z
=
ocl na
U
m
2 0 4U
0 2-
-
t
bl
,,
, I
100
71 05
I
I
I
I
110
120
130
140
Y B I T A l H t HCI
Figure 1 . Absorbances in 5 ml. of ethylene dichloride solution of periodides of betaine obtained under different conditions
instographic procedures (4). Concentrated aqueous solutions of these substances containing 20 mg. of compound per milliliter were prepared and stored a t 4" C. T o prepare standard curves, these solutions were appropriately diluted in the desired acid solution. Reagents. T o prepare iodine in potassium iodide solution (KI-12 reagent), 15.7 grams of iodine a n d 20.0 grams of potassium iodide, both analytical reagent grade, mere made u p to 100 ml. and treated as described by Appleton et al. ( 1 ) . The ethylene dichloride used was Eastman Grade (White Label). Analytical Method. Appleton's method ( 1 ) was modified t o Dermit aiialysis of Additional compounds h d to facilitate more rapid hmdling of a large number of samples. T o standard 12-ml. tapcred, 1ieavy-n.slled borosilicate glass centrifuge tubes was transferred 0.5 ml. of the solution to be :inRlyzed. This sample contained 10 to 100 y of the quaternary nitrogen compound in 1S sulfuric or 1 to 6-Y !iydrochloric acid solution. -1lthough the presence of tlie acid was essential, t1.e kind and concentr:ition could be \.nried as indicated, thcrcby permitting Jirect analysis of acidic effluents of iwious chromatographic sl.stenis. The st:indard curves for individual quaternnr'y nitrogen compounds were establi4Icd using acidic solutions conipar:lble in con~positiont o those on which r1,c mcrhod w s used. Blanks \\.ere 21-0 prcpared by employing 0.5 nil. of nolution s i n d a r to that analyzed but rlei.oid of quaternary nitrogcn compr,\Indj. ' l l i c - reaction mixtures were liv;Jt cold clwing d l steps Ic3ding t o the i.2cistion of the periodides. .is rnanj. as 50 tubes containing s:~mples for snal!$s u w e pliced in a rnck niid held in a cold w t e r bath a t 0' to 4" C. for 10 minutes. ' l h n 0.2 X I . of cold IiI-12 reagent n a s aclded and the, contents of thc tubes n-erc mixed : i t i d t!icri cooled i n the cold bath for 50 minurc.?. Cr\~stalliz~ltion or^ the pericldidc; \vas h j t e n c d by scratching tlie tube nxlk and stirring the rcaction solution at frequent jntervals during the cooling period using tapcrcd rods
-
o r
0O L
Figure 2. Absorption spectra of periodides of quaternary nitrogen compounds and of iodine in potassium-iodide reagent in 5 ml. of ethylene dichloride solution
prepared from 3-nim. glass rod. The tubes \vere centrifuged a t the maximum speed in a refrigerated centrifuge (International, Model PR 1, Head No. 811, 4000 r.p.m.) for 5 minutes. Upon removal of each tube from the centrifuge, the supernatant was aspirated b y a glass tube drawn to a vexy fine tip. Care was taken to avoid disturbing the precipitate a t this step, and a n incandescent lamp aided in seeing it. To the precipitate were added 5.0 or 10.0 ml. of ethylene dichloride and the mixture was stirred to effect solution. The absorbances a t 365 mp of the solutions were read in a Beckman Model DU spectrophotometer using 1.0-cm. cells.
Calculations. Standard curves were plotted of t h e absorbances of t h e recovered precipitate dissolved in ethylene dichloride us. t h e a m o u n t in micrograms of t h e original known quaternary nitrogen compound. Absorbances of solutions of precipitates from 10- to 50-7 amounts of a quaternary nitrogen compound were usually determined in 5.0 ml. of ethylene dichloride; precipitates from 20to 100-7 quantities generally gave satisfactory absorbances in 10.0 ml. of ethylene dichloride. Greater precision is attained with the larger amounts but with an attendant loss in sensitivity. Figure 1, A , shon-s a standard curve obtained with betaine periodides prepared in 1 .OAT sulfuric acid and dissolved in 5.0 nil. of ethylene dichloride. Each concentration was run in duplicate. The absorbance curves do not pass through the origin because of the small but measurable solubility of the periodides in the reaction mixtures and the resulting loss of periodide upon removal of the supernatant. The amount lost through solubility is equal to the horizontal intercept of the curve. The unknown concentrations of quaternary nitrogen compounds may be calculated directly from the curves or by using an equation.
The curves ale essentially linear and satisfy the espreision T. = my b (I),nherein:
+
y = absorbance at 365 mM
ni
=
b
=
z
=
inverse of slope of curve (micrograms of quaternary nitrogen compound as periodide in the ethlyene dichloride per absorbance unit) intercept with z axis (solubility of periodide in reaction solution) total quaternary nitrogen compound
Values for nz and b for a number of quaternary nitrogen compounds are assembled in Table I. These values were obtained by graphical analysis of the points obtained n i t h duplicate samples in each of t n o series of determinations. K h e n the techniquer of the procedure R ere carefully performed, excellent replication of the standard curves n-as achieved n i t h a precision to xk2 y. Preparation and Analysis of Purified Periodide. To characterize chemical composition and properties more fully, larger amounts of t h e pcriodides, freed from a n y rctaincd r a g e n t solution, were prepared. T o 100 mg. of the quaternary nitrogen compound contained in 20 ml. of 1N sulfuric acid in a 40-nd. centrifuge tube, were added 8 ml. of the KI-12 reagent and the mixture was treated in a manner similar to that of the analytical procedure. The isolated periodides n-ere then u-ashed by stirring in the centrifuge tube with 3 ml. of cold 1N sulfuric acid followed by centrifuging in the cold and by aspirating the supernatant; this process was repeated three times. The periodides were then dissolved in 50 ml. of absolute ethyl alcohol. The complexed iodine of the periodides was determined by titrating 1-ml. aliquots of the ethyl alcohol solution n i t h thiosulfate, starch solution serving as a n indicator. For nitrogen determinations aliquots of the alcoholic-periodide solutions were placed in micro-Kjeldahl flasks and water was added. Alcohol and comVOL. 32, NO. 7,JUNE 1960
871
Table I.
Factors for Estimating Quaternary Nitrogen Compounds from Periodide Absorption"
Periodide Formed in
periodide Solubility y Original Compound in 0.7-M1. Reaction Solution
Purifiedc 1 . ON H2S04 l . 0 N HCI 2 . 5 N HCl
0 5 6"
Volume Ethylene Dichloride, M1.
bb,
Quaternary Nitrogen Compound Carnitine, HCl
Betaine. HC1
Purified €I20
1 ONH2SOd 1 ON HC1 2 5iVHC1
Trigonelline. HCl
I
0 93 1 5 5
Purified 1.ON HzSO, 4 . O N HC1
0 4
5
10 mb,
Original Compound per Unit Absorbance a t 365 Mp of Periodide y
38.4 42.0 44.0 46.0
80.0 82.0 84.0
7 0 0 0 0
58 0 59 0 600
30 50 30 30 31
,..
...
5
34.6 34.0 37.0
67.0
Purified 2 5N HCl
0 4
34.0 37.0
74.0
Stachydrine.HCI
Purified 2 . 5 N HC1
0
35.1 36.0
... 71.0
Choline chloride
Purified 1 . O N H~SOI 2 5N HCI
0 2 3
27. ;i
27.9 29 0
54 0 55.0
Purified 2.5N HC1
0 2
21.1 23.0
47.0
r-Butyrobetaine. HCl
I
Tetramethylammonium chloride
5
74.0
...
...
Thiamine chloride 6.0-V HCl 0 233.3 ... a Values should be redetermined to allow for any variations in experimental conditions. As defined in Equation I. Purified Deriodide was isolated, analyzed, and dissolved in ethylene dichloride for absorbance measurements as detailed in tkxt. '
plexed iodine were removed by boiling the solution. Nitrogens were then determined upon the remaining aqueous solution of quaternary nitrogen compound iodides by the micro-Kjeldahl procedure using a mercuric oxide catalyst. EXPERIMENTAL
Absorption Spectra a n d Absorbance of Periodides. Appropriate aliquots of t h e purified periodides dissolved in ethyl alcohol were diluted 500-fold in ethylene dichloride and t h e absorbances of the solutions were measured at various wave lengths in the ultraviolet against a n ethylene dichloride blank. T h e ultraviolet absorption spectrum of 0.025 ml. of the KI-I2 reagent (an amount in excess of that retained in the blank) dissolved in 5.0 ml. of ethylene dichloride was also determined. I n Figure 2 the absorption spectra for periodides of the betaine-type compounds -betaine, trigonelline, and carnitine-are shown with that of choline periodide and of the KI-I2 reagent. Each of the quaternary nitrogen compounds exhibits similar absorption spectrum with maxima in the ultraviolet a t 365 and 295 mp. Similar absorption spectra were also 872
ANALYTICAL CHEMISTRY
observed with periodides prepared from y-butyrobetaine, stachydrine, A'-methylnicotinamide, ergothioneine, and thiamine. These curves are like those reported for choline periodide by Xppleton ( 1 ) and for tetramethylammonium periodide b y Buckles, Yuk, and Popov (3). Following the practice of dppleton et al. (1)' the absorbance a t 365 rather than at 295 mp, the greater peak, was measured in the analytical procedure, thereby minimizing absorption due to molecular iodine present in the sohtion and permitting the use of Corex cells and a tungsten lamp in the spectrophotometer. Absorption measurements a t 365 mp were made on various concentrations of these periodides in ethylene dichloride. An essentially linear relationship was observed betn een periodide concentration and absorbance. Figure 1, B , shows the absorbance a t 365 mp of the purified betaine periodide In ethylene dichloride plotted against it6 betaine content as determined by nitrogen analysis and expressed as the hydrochloride. The micrograms of quaternary nitrogen compounds contained in amounts of purified periodides having unit absorbance a t 365 mp when dissolved in 5.0 ml. of ethylene dichloride are listed in Table I. Good
agreement occurs between the values for weight of quaternary nitrogen con~poundper absorbance unit obtained with the purified periodides and those obtained in the analytical procedure employing 1N sulfuric acid and correcting for solubility losses. Webster ( I S ) has st,at,ed that the relationship between the absorbance of the periodide solution and the choline concentration obtained in his analytical procedure for choline does not follow Beer's law. However! a reproduction of his data in graphical form gives a linear curve intercepting the horizontal axis (choline concentratioiii. Both the slope and the intercept agree with corresponding values in the present' experiments (Table I). This correlstion suggests that in Kebster's procedure, as well as the authors', the inability to obtain a linear curve passing through the origin resulted from solubility losses rather than from the failure of the periodide derivative to obey Beer's law. Table I1 lists the molar absorbances a t 36.5 nip(, based on nitrogen content, of the purified periodides of the various quaternary nitrogen compounds in cthylene dichloride solution. The molar abfiorhances of the different periodides were nearly identical in value. T l i ~ the nature of the quaternary nitrogen compound eserbs little influence upon either the shape of the spectral curve or the magnitude of the absorbance of their periodide derivatives. I3uckles, Yuk, and Popov (3) have postulated that the absorbance is due to the triioclide ion. This nould explain not only the absorbance characteristics of the prriodides, but also the similar ultraviolet absorbancc spectrum of the KI-I, reagent in c,thylcne dichloride (Figure 2 ) . Composition of Periodides. Attempts were made to isolate, waterwash, and d r y over desiccant t h e periodides of the betaine-type conipounds. The dried materials reyealed variable iodine contents, evidence for the estreme lability of the periodides especially t h a t of trigonclline. T'ariations in iodine content iwre reflected in variations in absorbances. Therefore, alcoholic solutions were analyzed d i r e d y for iodine and nitrogen and the ratios of the complexcd iodine atoms to the nitrogen were calculated. These values are tabulated in Table 11, which also gives the probable formula of the periodide ion associated il-ith the quaternary base. Choline and tetrametliyl~ninionium periodides gave analyses conforming to the known ennxiodides. The betaine-type compounds gave iodine-nitrogen ratios indicating their occurrence primarily as t,he heptaiodides. These formulas cont'rast with those ascribed to t'he isolated periodides of betaine ( 2 ) and of trigonelline (9)
bl- other investigators who found lower iodine contents. Influence of pH on Yield of Periodides. When periodides from such betaine-type compounds as betaine and trigonelline were prepared in neutral solutions, low yields resulted. Stanek (12) demonstrated that the formation of periodide precipitates from betaine-type compounds is pH-sensitive as contrasted with that from quaternary amines. I n Figure 3 are compared the yields of choline arid betaine periodides, as indicated by their absorption a t 365 mp in ethylene dichloride solut'ion, when the p H of the init,ial solution was adjusted to various values. .4t low concentrations of betaine, RE also observed by Stanek, the periodide was precipitated only from highly acidic solutions. -kt a concent'ration of 1 0 s sulfuric acid, iodine was precipitated. I n strongly alkaline solutions liypoiodite is formed which reacts with the choline. The difference in yields of betaine aiid choline periodides in avidic and basic media has been used by Stjanek (12) and can be used in the present method for their differential deterniination in the presence of one another. Two fuctors contribute to the increased yield of betaine-periodide precipitate v\-hcn 1 S sulfuric acid is the reaction medium rather than water. Oiic is the lower solubility of the periodide in acid as measured by the horizontal intercept of curve A compared with curve C of Figure 1. The other is the increased association of iodine with betaine in the acid solution as shown by the increased slope of A compared with C, Figure 1. Both factors can be ascribed to the reduced ionization of the carboxyl group of betaine in acid solution. The resulting increased positive charge on the betaine molecule favored periodide formation: while the decreased polarity of the betaine periodide diminished its solubility. Effect of Time a n d Temperature on Periodide Precipitation. I n t.he procedure of -4ppleton et al. (1) the K I - I r choline reaction solution is cooled at 4" C. for 20 minutes. Applying these conditions to betaine-type compounds did not give reproducible values. Studies \yere, therefore, made on the influence of reaction time on periodide yields. The time for maximum precipitation varied, being great'est for carnitine and diminishing Tvith decreasing polarity of the substances in the order: betaine, 7-hutyrobetaine, and trigonelline. Precipitation of the periodide of each compound was accelerated b y frequently scratching the side of the reaction t'ube and st'irring its contents. Eighty minutes was the minimal time that allowed complete precipitation of carnitine periodide under the described condi-
1 \
\ \ \
HCI 4 0 Y
p H O f REACTION SOLUTION
Figure 3. Effect of p H on yields of betaine and choline periodide Absorbances determined in 5.0 ml. of ethylene dichloride
tion?. The high solubilities of certain of the periodides, such as those of carnitine and betaine, require that the reaction mixture be maintained a t low temperatures during this step and during subsequent centrifugation prior t o isolation of the precipitate. Effect of Anions and Iodine Concentration on Periodide Yields. Blood and Cranfield (.2) reported that the presence of chloride ions decreased the recovery of betaine periodide and used sulfuric acid in their procedures. Because, in ion exchange chromatographic methods for separating quaternary nitrogen compounds, hydrochloric acid solutions have been used advantageously as eluents (4, 7 ), the influence of chloride ion upon periodide yields was studied. Performing the reaction in 1N hydrochloric acid resulted in a very slightly smaller yield of carnitine and betaine periodides than in 1 S sulfuric acid (Table I). An even loner yield of precipitate was obtained in 2.5X hydrochloric acid. The action of chloride ion in decreasing periodide yield was more pronounced after shorter reaction times than those employed in the analytical procedure. The presence of chloride ion probably reduced the rate of crystallization of the periodide b y the formation of chlorideiodine coinplexes with the quaternary nitrogen compounds. The longer re-
Table
II.
action period and use of standards run in similar hydrochloric acid concentration permit satisfactory determinations of carnitine, betaine, and other quaternary nitrogen compounds in hydrochloric acid solution. Another factor that influenced the rate of periodide precipitation was the iodine concentration of the reagent. When the iodine concentration was increased to 17.5 grams of iodine per 100 ml. of water and 20 grams of potassium iodide, maximum yield of carnitine periodide n a s obtained in a shorter reaction time. Increasing the iodine concentration to 20.0 grams of iodine per 100 ml. resulted in precipitation of iodine with a resultant increase of the blank to excessive values under the prescribed conditions. The use of 17.5 grams of iodine per 100 ml. of reagent increased the precipitation of certain other nonquaternary nitrogen compounds as detailed in a subsequent section. Because this reaction reduced the specificity of the procedure for quaternary nitrogen compounds, the loner concentration of iodine (15.7 grams per 100 ml.) was generally employed in the survey of chromatographic fractions. Solubility of Periodides in Ethylene Dichloride. Because of their tendency t o lose iodine t h e periodides should be dissolved in ethylene dichloride immediately after their isolation. T h e periodides were dissolved easily, except for thiamine when its concentration was at t h e higher lerels of t h e method's range. This lower solubilitv of thiamine periodide is probably due t o the polar amine hydrohalide group present in the molecule. Thiamine periodide tends to decompose in ethylene dichloride to yield free iodine. Determinations of thiamine by this procedure are, therefore. unsatisfactory. Effect of Other Nitrogen Compounds upon Determination of Quaternary Compounds a s Periodides. Stanek (11) reported that a large number of nitrogenous compounds, including adenine, nicotinic acid, histidine, and others, gave precipitates with iodine. During ion exchange chromatography of substances in extracts of biological materials these compounds
Composition and Absorbance of Periodides
Observed Ratio Atoms Complexed Periodide from Iodine/Nitrogen Choline chloride 7.85 Tetramethylammonium chloride 8.05 Betaine.HC1 6.05 -~Butyrobetaine.HCl 6.40 DkCarnitine.HC1 6.22 Stachydrine.HC1 6.00 Trigonelline.HC1 5.40
Periodide Ion I 8 1I81
-
I,I-
1,I I,I I,I IC1-
Molar Absorbance X 10-8 in Ethylene Dichloride 25.5 25.9 25.0 26.5 25.7 25.5 25.2
VOL. 32, NO. 7, JUNE 1960
873
may be eluted with certain quaternary nitrogen compounds-nicotinic acid with trigonelline (4). Experiments were performed to determine the extent of precipitation of these compounds with the KI-12reagent. Histidine dihydrochloride did not yield a precipitate with the KI-I2 reagent at, levels up to 150 y in 0.5 ml. of 4.ON hydrochloric acid. Sicotinic acid did not give a periodide precipitate at comparable levels when the reagent contained 15.7 grams of iodine per 100 ml. When the iodine content was increased to 17.5 grams per 100 ml., a measurable precipitate was obtained with 80 y of nicotinic acid. Adenine a t low levels gave a n appreciable precipitate n i t h the reagent. As in the case of thiamine, the resulting precipitate dissociated in ethylene dichloride to yield free iodine and the resulting solution had a low absorption at 365 mp. These and other nitrogenous constituents contained in extracts of corn did not interfere with the deter-
mination of the quaternary nitrogen compounds in the extracts b y the present procedure after their separation b y ion exchange chromatography (4). ACKNOWLEDGMENT
The authors thank the International Minerals and Chemical Corp. for samples of y-butyrobetaine and DLcarnitine arid G. S. Fraenkel of the Department of Entomology, University of Illinois, for a gift of DL-carnitine also used in these studies. LITERATURE CITED
(1) Appleton, H. D., La Du, B. N., Jr., Levy, B. B., Steele, J. hf., Brodie, B. B., J . Biol. Chem. 205,803 (1953). ( 2 ) Blood, J. W., Cranfield, H. T., Analyst 61,829 (1936). (3) Buckles, R. E., Yuk, J. P., Popov, A. I., J. Am. Chem. SOC. 74. 4379 (1952j. (4) Christianson, D. D., Wall, J. S., Dimler, R. D., Senti, F. R , ANAL. CHEM.32,874 (1960). (5) Cornforth, J. K.,Henry, A. J., J . Chem. SOC.1952,601.
(6) Ericson, B. N., Avrin, I., Teague, D. M., WTTilliams, H. H., J . Biol. Chem. 135,671 (1940). ( 7 ) Friedman, S., McFarlane, J. E., Bhattacharyya, P. K., Fraenkel, G. S., Arch. Biochem. Biophys. 59,484 (1955). (8) Huff, J., Perlsweig, W. A., J . Biol. Chem. 150,395 ( 1943). (9) Nottbohm, F. E., hlayer, F., Z. Untersuch. Lebensm. 61,202 (1931). (10) Sarett, H. P., Perlzweig, K., Levv. E. D., J . Biol. Chem. 135. 483 ii940). ’ (11) Stanek, V., 2. physiol. Chem. 47, 83 IlROfi). (12) %d.;48, 334 (1906). (13) Webster, G. R., Biochim. et Biophys. Acta 20,432 (1956). (14) Werle. E.. “hfodern Methods of . Plant Analyses,” ed. by K. Paech and hl. W. Tracy, Vol. 4, pp. 517-623, Springer-Verlag, Berlin, 1955. RECEIVEDfor review October 14, 1959. Accepted February 12, 1960. Presented in part before the Division of Agricultural and Food Chemistry, 133rd Meeting, ACS, San Francisco, Calif., April 1958. Mention of firm names or trade products does not imply that they are endorsed or recommended by the Department of Agriculture over other firms or similar products not mentioned.
Separation and Determination of Quaternary Nitrogen Compounds and Other Nitrogenous Substances by Ion Exchange Chromatography Application to Analysis of Corn Extracts DONALD D. CHRISTIANSON, JOSEPH S. WALL, ROBERT J. DIMLER, and FREDERIC R. SENTI Norfhern Regional Research laboratory, Peoria, 111.
)A general procedure was developed for quantitatively separating mixtures of naturally occurring quaternary nitrogen compounds, including betaine, choline, stachydrine, trigonelline, and thiamine, b y chromatography on a column of a sulfonated polystyrene resin in the acid form. Amino acids and certain other nitrogen compounds were also separated in this system. Use of appropriate methods permitted analysis for all of these substances in the effluent fractions of a single chromatographic separation, Basic principles governing the chromatographic behavior of quaternary nitrogen compounds on sulfonated resins in the acid form were demonstrated by their elution sequence. The applicability of the procedure to biological materials was shown by its use in characterizing and determining the free nitrogenous substances in extracts of corn grain.
874
ANALYTICAL CHEMISTRY
Q
nitrogen compounds as well as free amino acids and amides are present in corn grain as demonstrated by paper chromatography of ethyl alcohol-water extracts. Because the characterization and determination of these substances required a more suitable procedure, the use of ion exchange chromatography was investigated. Although successful separations of specific quaternary nitrogen compounds on ion exchange resins have been reported, the methods employed were not designed for general application t o many of these commonly occurring substances in biological extracts. Friedman et al. (9) quantitatively separated carnitine and betaine on Dowex 60 in the hydrogen form but did not determine choline in their system. However, Pilgeram and coworkers (19) employed a similar procedure to resolve the choline, serine, and ethanolamine in phospholipide hydrolUATERNARY
ysates. A cation exchange resin was also used by Aronoff ( 1 ) to isolate trigonelline. It was observed in the present experiments that the pattern of elution of quaternary nitrogen compounds from sulfonated polystyrene resins with acid eluents was similar to that of the related amino acids. Chromatography of these substances then was tried under the conditions for separating amino acids on Dowex 50 in the hydrogen form as described by Stein and Moore (20). This procedure resolved many quaternary nitrogen compounds and permitted simultaneous separation of these substances and amino acids in one chromatographic run. Certain basic aromatic compounds can also be resolved in this system ( 2 2 ) . Quaternary nitrogen compounds in effluent fractions were determined conveniently by a recently described procedure (23).
