ANALYTICAL CHEMISTRY the compound to benzaldehyde benzylhydrazone, which produces the second wave. If this proposed mechanism were correct, in alkaline solutions one should obtain azobenzyl (C6HoCHzX=N-CHzC&H,). I n analogy to azobenzene one might expect azobenzyl t o reduce below - 1.0 volt polarographically. HOKever, both waves observed for benzalazine in alkaline media (Table I) are considerablv more negative. The proposed mechanisms should be verified by controlled potential electrolysis by which the products could be isolated and identified. Some of this work is now in progress. From the similarity of half-wave potentials of benzaldehyde and benzalazine, it might be assumed that what one was really reducing a t the dropping mercury electrode with solutions of benzalazine was actually benzaldehyde formed by hydrolysis. I n order to chow that the reaction rvas a reduction of the benzalazine itself, several tests were made. The classical aldehyde qualitative tests (silver mirror and 2,4-dinitrophenylhydrazine) showed no aldehyde to be present in the benzalazine solutions. I n polarographic work it was observed that a t a p H of 6 there was a slight separation of half-wave potentials betlveen benzalazine solutions and benzaldehyde solutions. Ultraviolet absorption spectra measured for old and new solutions of tribenzaldehrde(bis)methylhydrazone and solutions of benzalazine, benxaldehyde dimethylhydrazone, and benzaldehyde shoa ed that the characteristic absorption peak a t 246 mp for benzaldehj-de was not present in the benzalazine and benzaldehyde dimethylhydrazone solutions, but was present to a very small extent in new tribenzaldehyde(his)methylhydrazone solutions and \vas very strong in old Folutions. Thus, it appears that the benzalazine
molecule rather than benzaldehyde is the species which undergoes reduction a t the dropping mercury electrode. ACKNOWLEDGMENT
The authors wish to express their thanks to H. IT.Kruse for the helpful suggestions and advice throughout this work. This paper is published by permission of W, B. McLean, Technical Director of the U. S. Naval Ordnance Test Station, China Lake, Calif. LITER4TURE CITED ( 1 ) Britton, H . T.
S.,Robinson, R. .I., J . Chem. Soc. 1931, 1946. (2) Fisher, R. A , , “Statistical Methods for Research Workers,” 10th ed., Oliver and Boyd, Edinburgh, 1946. (3) Harries, C.,Haga, T., Ber. 31, 62 (1898).
(4) Hatt, H. H., “Organic Syntheses,” Coll. T’ol. 11, p p ~6-395, Wiley, New York, 1947. (5) Kolthoff, I. 11..Lingane, J. J., “Polarography,” 2nd ed.. p. 70, Interscience Publishers, Sew York, 1952. ( 8 ) I b i d . , p. 374. ( i )Lingane, J. J., J . Am. Chem. SOC.67, 1916 (1945). (8) Sernst, W.Z . , Phusik. Chem. 2, 613 (1888). (9) Pasternak, R., Hela. Chim. Acta. 31, 753 (1948). (10) Todd, David, J . Am. Chem. SOC.71, 1353-5 (1949). (11) Tokuoka, >I., Collection Czechoslov. Chem. Communs. 7, 392 (1935). (12) Whitnack, G. C.,Sielson, J. 11,,Gantz, E. 9. C., J . -4m. Chem. SOC. 76,4711-14 (1954). R E C E I ~ Efor D review August 29, 1955. .4ccepted February 9, 1956. Dirision of Analytical Chemistry, 128th meeting, .iCS, Minneapolis, Ninn., September 1955.
Separation of Acetylated Neomycins B and C by Paper Chromatography S. C. PAN and JAMES D. DUTCHER The Squibb institute for M e d i c a l Research, N e w Brunswick,
A semiquantitative method for distinguishing and determining neomycins B and C is presented. The procedure is directly applicable to fermentation samples or, preferably, after a preliminary purification by adsorption and elution on an ion exchange resin. The method is based on the separation by paper chromatography of the neutral A‘-acetyl derivatives of neomycins B and C and their detection through conversion to the N-chloro derivative followed by color development with starch-potassium iodide spray reagent.
T
HE use of paper chromatography as a criterion for the homogeneity of neomycin preparations has been reported from several laboratories ( 4 , 5, IO), but subsequent to the actual isolation and characterization of neomycin B and neomycin C ( d ) , it was found that none of these methods was capable of resolving a mixture of the two neomycins. The most recent report (8) still admits the inability t o resolve the neomycins satisfactorily by paper chromatographic procedures. I n view of the chemical nature of the neomycins ( d ) , these two antibiotics may be considered as iosmeric oligosaccharides with a number of amino groups. Conceivably, they differ from each other only in minor structural features, possibly a difference in the configuration of one of the glycosidic linkages. Experiences in this laboratory indicate that the basicity of these two neomycin isomers probably dominates their paper chromatographic behavior and can account
N. J.
for the difficulty encountered in their resolution. It was therefore reasoned that if the basicity were neutralized by acetylation, the minor structural difference might be able to manifest itself as a difference in the mobility on a paper chromatogram. This proved to be the case. EXPERIMEHTAL
Detection of N-Acetyl Neomycins. For the reasons just discussed, paper chromatography of the crystalline h’-acetyl neomycins ( 2 ) was tried. Because the ,V-acetyl neomycins are biologically inactive and do not respond to the ninhydrin reagent, an effective method for detecting them on paper chromatograms had to be developed first. I t was found that a modification of the method developed by Rydon and Smith ( 7 ) for detecting peptides and amino acids was very effective for this purpose. The modified procedure employs the following reagents. Sodium hvoochlorite. Add 1 Dart of Clorox (5.25% sodium hy- ochloriteh water) to 20 partiof water. &hyl alcohol, 95%. Starch-iodide reagent. Mix equal volumes of a 1% soluble starch solution and a 1 % potassium iodide solution. The starch solution should not be used for longer than 1 week. The procedure consists of three steps. The developed paper chromatogram is sprayed with the sodium hypochlorite. When dry, it is ?prayed with 95% ethyl alcohol. Finally, after the ethyl alcohol has evaporated, the chromatogram is sprayed with starch-iodide reagent. The acetvlated neomycins show up as deep blue spots again& a colorless background.
837
V O L U M E 28, NO. 5, M A Y 1 9 5 6 T h e principles involved consist of converting a substituted amide with hypochlorite into the chloramide, destroying the excess hypochlorite with a mild reducing agent, and oxidizing the iodide t o iodine by the chloramide. Other mild reducing agentse.g., a 5y0 glucose solution or a 1% formaldehyde solution-also destroy the excess hypochlorite without attacking the chloramide. T n the original method of Rydon and Smith ( 7 ) chlorine gas was employed and the excess chlorine later removed b\ arration. SOLVENT_
S-acetyl derivatives. Originally ( 2 ) these had been prepared by acetylating the free bases in a methanol solution x i t h acetic anhydride. It was found that S-acetylation could be readily carried out in buffered aqueous solutions.
To 1 ml. of a solution containing 1 to 2 mg. of neomycin per milliliter, is added 0.2 ml. of a 4.5%' sodium acetate solution or 0.1 ml. of 3M dipotassium hydrogen phosphate. Then 0.1 ml. of acetic anhydride is added dropwise JTith shaking. When the resulting solution is spotted directly on paper and developed as described above, a satisfactory chromatogram is obtained as shown in Figure 2 Figure 2 also shoxs the chiomatogram of a sample acetjlated in the absence of the buffer as well as that of the unacetylated neomycins. The acetylation in the absence of buffer apparently led to the formation of a mixture of partially aceylated products. Evperiments 1% ith different sodium acetate concentrations demonstrated that 0.351 is the minimum required to ensure complete acetylation. Use of dipotassium hydrogen phosphate a t the same concentration (0.3.hl) was found to be more advantageous, because a slight interference with the spot for .?--acetyl neomycin C by sodium acetate could be eliminated.
FRONT
Solvenl FtQnl
f l y 0.29
Rfs0.19
ORIGIN
Rf10.29
B
c
BtC
R f 2 0.19
Figure 1. Paper chromatograms of N-acetyl neomycins B and C B. C.
\-;\cetyl neomycin B .Y-.icet> I neomycin C
This procedure has two disadvantages: The background is also somewhat blue and, in addition t o amides and peptides, many other compounds including amino and ammonium compounds also show up as blue spot$. T h e present modification eliminates these disadvantages and, with a few exceptions, is specific for monosubstituted amides. K i t h glycylglycine, as little as 0.3 y on a spot 12 mm. in diameter can be detected. Use of benzidine or tolidine in place of the starch-iodide reagent, as reported by Reindal and Hoppe ( 6 ) , is, of course, also effective but the color is not as deep. Development of Paper Chromatogram. Because the S-acetyl neomycins were expected to behave like neutral oligosaccharidea, the solvent system (butyl alcohol-pyridine-water, 60 to 40 to 30) developed by Jeanes, Wise, and Dimler (3) for sugars was first selected for test. Descending chromatography on Whatman S o . 1 paper for an overnight period-i.e., approximatelj. 16 hours lor a 40-em. run-effected a satisfactory resolution of the S-acetyl neomycins. A diagram of the paper chromatogram thus obtained is shon-n in Figure 1. The R/ of S-acetyl neomycin B is equal to 0.29 and the Rj of S-acetyl neomycin C is equal to 0.19. Repeated tests showed that it was essential to prepare the solvent mixture fresh every day because re-use often led to unsatisfactory separation. The chromatographic procedure was also shown to be scnsitive; a spot approximately 13 mm. in diameter can be obtained x i t h 6 y of N-acetyl neomycin B or C,although quantities a.s small as 2 t o 4 y can also be detected. Acetylation of Neomycins. I n order t o apply the method directly to neomycin samples, including fermentation broths, a procedure was developed for converting the neomycins to the
Origin
Figure 2.
Acetylation procedure as studied by paper chromatography
hlixtur? of S-acetyl neomycin B and .%--acetyl neomycin C: Mixture of neomycin B and neomycin C. acetylated with sodium acetate as buffer C . Rlixture of neomycin B and neomyrin C, acetylated with dipotassium hydrogen phosphate as buffer D . Mixture of neomycin B and neomycin C, acetylated in absence of buffer E. Mixture of neornvcin B and neomvcin C. not acetvlated F. Fernitmtation broth sample acetylated directly with procedure giren a, b . Traces of color a t a and b also appeared in acetylation mixture containing no neomycin with dipotassium hydrogen phosrihate as buffer A.
B.
The last example in Figure 2 represents a fermentation sample m-hich has been acetylated by the procedure described above. It may be seen that there are a number of rompounds other than the neomycins in the fermentation sample v-hich also appear in the chromatogram. However, the characteristic spots of *Vacetyl neomycins B and C are readil) visible Later experiments s h o m d that the colored streak due to compounds other than neomyrins can be eliminated if the neomycins in the fermentation sample are first puiified b:- means of .Imberlite IRC-50 (9) This paper chromatographic procedure is, therefore, applicable to the detection of neomycins B and C in fermentation samples. SEAZIQUANTITATIVE DETERiVIKATION
The areas of the spots, as determined by the method of cutting out and weighing, xere found to be proportional to the logarithm of the sample weight xithin the range from 3 to 50 y ( 1 ) . How-
ANALYTICAL CHEMISTRY ever, because the position and slope of the standard curve are someahat variable, it has been found necessary to include standard samples of the pure X-acetyl neomycins in each run. I n this manner it is possible to estimate the percentage of neomycins B and C in a given sample. The error in the value for the percentage of the minor component, usually neomycin C, is of the order of =tt30%. Although the method as developed permits only a semiquantitative estimation, it is useful for evaluating the ratio of neomycin C to neomycin B in various preparations and fermentation broths. LITERATURE CITED
(2) Dutcher, J. D., Hosrtnsky, S . , Donin, 11.N., Wintersteiner, O., J . Am. Chem. SOC.73, 1384 (1951). (3) Jeanes, A . , Wise, C. S., Dimler. R. J., ANAL.CHEM.23, 416 (1951). (4) Leach, B. E., De Vries, U’.H., Selson, H. A., Jackson, I+’ G., Evans, J. S., J . Am. Chem. Soc. 73, 2797 (1951). (5) Regna, P., Murphy, F., Ibid., 72, 1045 (1950). (6) Reindal, F.. Hoppe, W., Ber. 87, 1103 (1954). (7) Rydon, H. N., Smith, P. W. G., S a t w e 169, 922 (1952). (5) Saito, A., Schaffner, C. P., Abstract, p. 98, IIIrd International Coneress of Biochemistrv. Brussels. 1955. (9) St. John, C. V., Flick, D: E., Tepe, J. B., ANAL.CHEX 23, 1255 (1951). (10) Waksman, S. A., “Neomycin,” p. 75, Rutgers University Press, Xew Brunswick, 3.J., 1953.
RECEIVED for review December 28, 1955. ilccepted February (1)
Block, R. J.. Le Strange, R., Zweig, G., “Paper Chromatography,” p. 38, Academic Press, New York, 1952.
11. 1956.
Division of Agricultural and Food Chemistry, 128th hfeeting, ACS. hlinneapolis, Rlinn., September 1955.
Partition Chromatography of Aliphatic Acids Quantitative Resolution of Normal Chain Even Acids from CIS to
624
F. A. VANDENHEUVEL and D. R. VATCHER Fisheries Research Board of Canada, Halifax, Nova Scotia, Canada
Reversed-phase chromatography on a silane-treated silicic acid column, with 2,2,4-trimethylpentane (isooctane) as stationary phase and aqueous methanol as eluting solvent, allows the quantitative resolution of mixtures of even-numbered, normal-chain saturated fatty acids in the range CIZto CN. Less than 50 mg. of mixed acids are needed. The elution and titration of the seven acids require about 5 hours. Individual components representing 10% or more of the mixture are An determined with an error not exceeding 2.57‘. almost comparable degree of accuracy is attainable for minor components (1 to 3% of the total mixture), when a procedure involving a complementary analysis is followed. A device for automatically and continuously changing the composition of the eluting solvent and a semiautomatic, motor-driven microburet are described. The method is applicable to the determination of the relative proportion of each carbon series represented in a fat sample and the relative proportion of saturated to unsaturated acids in each carbon series.
I
N 1951 the present authors undertook to devise a quantitative chromatographic method for the estimation of even saturated fatty acids from CI2 to C L ~which could be used to analyze fats of marine origin. It was known that two mixtures of saturated acids could be obtained from such fate, the first resulting from complete hydrogenation of the sample and the second, from removal of all unsaturated components through oxidative degradation. Each of these mixtures would be chromatographed and the results combined to yield the respective proportions of saturated and unsaturated acids within each carbon series. Although no detailed information concerning individual unsaturated components would be obtained, a true picture of the fat structure would emerge. One proposed application of such a procedure was the study of changes occurring in the fatty acid composition of fat in fishes and marine mammals during their life cycles. As this would require routine analysis of many samples, detection of relatively small changes, and estimation of small amounts of certain acids, a highly standardized and accurate method was needed.
Only the methods of Howard and Martin (8) and of Boldingh (3) were knoa-n a t the time this investigation was undertaken. Mealorub powder could not be secured t o try the method of Boldingh, the other method was found unsatisfactory over the range Ci2 to (221. Successive modifications of the latter have led to the method described herein. Others have described separations of even saturated acids of high molecular weight (9, 10, 14). These procedures covered sections of the range C1?to Cz4; the recoveries were not fully quantitative, .4pplications of the method of Howard and Martin involving combination of two complementary chromatograms (4,6, 11, 12, 16) all had one point of similarity with the proposed method: One of the chromatograms used was that of a mixture of saturated acids derived from the original mixture, either by complete hydrogenation (11,15) or by elimination of unsaturated components through oxidative degradation (4,12). Homever, they all used the intact original mixture of acids for the complementary chromatogram to secure detailed information on all acid components. Moderately quantitative results were obtained with simple fats (4,6, 12) or with simple fractions from fats (11, 16). Silk and Hahn showed that the chromatographic method used is semiquantitative only (14) in estimating the even saturated acids from CIS to Czt. It is believed that these applications could be improved by the use of the more accurate method described below. I n separating the unsaturated acids from the original mixture, the method of Schuette and Dal Nogare (13) is used to eliminate acid oxidation products completely, thereby avoiding possible interference in the chromatogram of the saturated coniponents. OUTLINE OF METHOD
The present method has one basic feature in common with that of Howard and Jlartin (8) for the separation of acids from Clz t o C1,. The column material is treated with dichlorodimethylsilane to render it univettable by polar solvents. It has much in common with a method proposed by Vandenheuvel and Hayes (16) for the analysis of the monocarboxylic acids from CZ to Clr: the type of Trubore glass column used; the commercial grade silicic acid used to prepare the column mixture; the proce-
