CHEMICAL DETERMINATION OF PYRIDOXINE. REACTIONS IN PURE SYSTEMS

Transcription

1 CHEMICAL DETERMINATION OF PYRIDOXINE. REACTIONS IN PURE SYSTEMS BY MELVIN HOCHBERG, DANIEL MELNICK, AND BERNARD L. OSER (From the Food Research Laboratories, Inc., Long Island City, New York) (Received for publication, May 17, 1944) Critical studies have been conducted in our laboratories on various methods for the determination of pyridoxine. Our experiences with the microbiological method have been presented elsewhere (14). The present paper is concerned with the chemical determination of pyridoxine based upon its coupling reaction with 2,6dichloroquinonechloroimide. For a better understanding of the factors modifying the reaction and how these should be controlled, tests were first conducted with pure systems. The results of these studies are now presented. In the following paper (5) the application of the method to pharmaceutical preparations and biological materials is described. The phenolic nature of pyridoxine has been the basis for chemical de termination through coupling reactions with 2,6dichloroquinonechloroimide (15, ll), diazotized sulfanilic acid (16, l), the Folin-Denis reagent (16), and ferric chloride. However, even the chloroimide reagent, the most specific of these, reacts to form dyes with many other phenols, amines, and unrelated compounds (11). The reliability of each of the chemical methods for pyridoxine has been largely dependent upon the removal from the test extract of such interfering substances. Principle of Present Method The vitamin is coupled with 2,6dichloroquinonechloroimide in a strongly buffered alcoholic solution to yield a blue pigment, which is estimated CHnOH (1) + Cl-Nd=O ---f Cl Pyridoxine 2, B-Dichloroquinonechloroimide CHPOH Blui pigment

2 11 DETERMINATION OF PYRIDOXINE phot,ometrically. The reaction between the vitamin and the reagent is believed to occur according to Equation 1 (4,12). Potential errors due to other compounds coupling with the reagent are eliminated by conducting the reaction concurrently in the presence of an excess of borate, which under proper conditions renders the pyridoxine non-reactive without affecting the reaction of other coupling compounds. The structural formula of the pyridoxine-borate complex is believed (12) to be Reagents- Chbroimide Reagent-1 mg. of recrystallized 2, B-dichloroquinonechloroimide are dissolved in 25 cc. of isopropanol. The solution is stored in a glass-stoppered bottle in the refrigerator, and portions are withdrawn as needed. The reagent should not be kept for more than 1 month, and should be discarded sooner if a pink discoloration develops. The solid reagent is first purified by dissolving 1 gm. in 5 cc. of acetone and precipitating by the gradual addition of small amounts of water while stirr+g. The crystals are collected on a Buchner funnel, rapidly air-dried by suction, and then stored in a sealed bottle in the refrigerator. The recrystallized reagent is stable for more than 6 months under such conditions of storage. Ammonia-Ammonium Chloride Solutdon-16 gm. of ammonium chloride, c.p., are dissolved in 7 cc. of water and 16 cc. of concentrated ammonia water (approximately 27 per cent) added. The solution is diluted with water to 1 liter. Boric Acid Solution-5. gm. of powdered boric acid, c.p., are dissolved in 1 cc. of distilled water. Pyridozine HydrochZorid@-1 mg. of the crystalline vitamin are dissolved in 1 cc. of.1 N hydrochloric acid. This stock solution is stable for at least 3 months if stored in the refrigerator in an amber bottle. Working standards are prepared daily by dilution of the stock solution. Color Development-A direct reading photoelectric calorimeter is most satisfactory for the measurementcs.3 To a tube containing 1 cc. of pyridoxine solution, 5 cc. of isopropanol, 2 cc. of ammonia-ammonium chloride 1 The crystalline compound may be prepared by the method described by Gibbs (4) or purchased from the Eastman &dak Company, Chemical Sales Division, Rochester, New York, or from the Organic Products Company, Ne,w York. 2 Obtained from Merck and Company, Inc., Rahway, New Jersey. In the present study, the terms pyridoxine and pyridoxine hydrochloride are used interchangeably, and all values are expressed as the hydrochloride. 3 The Evelyn photoelectric calorimeter obtained from the Rubicon Company, Philadelphia, was employed in these studies.

3 HOCHBERG, MELNICK, AND OSER 111 solution, and 1 cc. of boric acid solution, 1 cc. of the chloroimide reagent is added, and the instrument is set at 1 per cent transmission 6 seconds later.4 The color is developed in 6 seconds in a similar tube containing water in place of the boric acid solution. Readings are taken with the 62 rn/* filter. Though pyridoxine is destroyed by irradiation, especially in alkaline solution (6,7), exposure to ordinary laboratory light for 3 hours produces nomeasureable destruction of the vitamin even in colorless solutions. Only in the case of the standard solution of pyridoxine, which may be used Reaction time (Seconds) FIG. 1. Rate of formation and fading of the dye formed in the reaction between pyridoxine and 2,6-dichloroquinonechloroimide. Curve a, 1 y of pyridoxine per tube; Curve b, 2 y of pyridoxine per tube. repeatedly for months, should special precautions be taken to avoid light destruction. Typical time reaction curves showing the rate of formation of the colored complex are presented in Fig. 1. The curves for 1 and 2 y of pyridoxine were corrected for the small, progressive changes in the reagent blank before being plotted. The graph demonstrat,es that t&e dye forms rapidly during the 1st minute and is sufficiently stable at 6 seconds for precise measurement. When a Verona1 buffer (11) is used in the same water-isopropanol system, 2 minutes are required for maximal color development. In the water-butanol system of Scudi (11) the coupling reaction is completed only after 4 minutes. 4 The solutions are added in the order mentioned to prevent the precipitation of salts.

4 112 DETERMINATION OF PYRIDOXINE In Fig. 2 are plotted data indicating the proportionality of the photometric density of the pigment to initial concentration of pyridoxine, exactly 6 seconds after the addition of chloroimide reagent. It is apparent that the reaction obeys Beer s law. When readings are taken at any time other than in the stage of maximal color development, deviations from the linear relationship between color and pyridoxine content may be found. Reproducibility and Specifiity of LWethod-More than 2 measurements over a period of a year on pure solutions containing 1 and 2 y of pyridoxine per tube, with different sets of buffer and reagent solutions, some freshly prepared, others 2 to 3 weeks old, have given values agreeing with those plotted in Fig. 2 to within ~1.1 per cent a.verage deviation. In all VrCdoxine concentration (Meg. pep tube) FIG. 2. Proportionality of photometric density of pigment to tion of pyridoxine. initial concentra- cases, the same batch of recrystallized 2,6dichloroquinonechloroimide was employed for making the reagent. Experimental evidence has been obtained proving the specificity of the method for pyridoxine. The vitamin reacts to form a dye with 2,6-dichloroquinonechloroimide but fails to do so in the presence of boric acid. Other compounds which couple with the reagent react to the same degree in the presence or absence of the boric acid. The photometric densities of the dyes due to pyridoxine and to interfering compounds are additive. These points are clearly illustrated in Table I. Only the diacetyl derivative of pyridoxine, 2-methyl-3-hydroxy-4, B-bis- (acetoxymethyl)pyridine showed an appreciable difference in reaction in

5 HOCHBERG, MELNICK, AND OSER 113 the presence and absence of boric acid. The compound is reputed to be biologically equivalent to pyridoxine on a molar basis, owing to the ease of removal of the acetyl groups from the molecule (17). The sample was assayed microbiologically (14) in our laboratory and found to possess 3 per TABLE I Validity of Borate Blank in Determination of Pyridoxine in Presence of Other Compounds React2 inf 7 with, 8, GDichloroquinonechloroimide Reagent - - Photometric densityt at 62 II+ I ncremenl I nsctiva- Compound Aniline Phenol &Naphthol Naphthylamine hydrochloride Catechol Resorcinol 3-Hydroxypyridine 2,4-Dimethyl3-hydroxy-5- hydroxymethylpyridine hydrochloride 2-Methyl3-hydroxy-4,5- bis(acetoxymethyl)pyridine 2-Methyl-3-amine-4-ethoxy methyl-5-aminomethylpyridine dihydrochioride monohydrate 2-Methyl-3-hydroxy-4- methoxymethyld-hydroxymethylpyridine {mount per tube 11,&o ( 8.1 (=) + () + 1 Test 6 ir$ 1Oypf p,:i& acid dpy m ox1ne ine and boric acid () (b) (4 Cd) : I I I 13.! o.ooo I I I due to 17 of PY[;p 3) - (.I (6) tion of 1crement sy boric acid pe? GCAI * The weights of the pyridoxine analogues taken are equivalent on a molar basis to 1 y of pyridoxine. These compounds were kindly supplied by Dr. K. Folkers of Merck and Company, Inc., Rahway, New Jersey. t The calorimeter was set at 19 per cent transmission with the reagent blank cent of the potency of pyridoxine on a molar basis, a value in very close agreement with that derived from the chemical test. This may have been due to the presence of free pyridoxine in the sample. The acid-hydrolyzed material gave much higher values, equivalent to 75 to 85 per cent of pyridoxine when subjected to both microbiological and chemical assay. The

6 114 DETERMINATION OF PYRIDOXINE other analogues of pyridoxine had no appreciable stimulatory action on the yeast, Saccharomyces cerevisiae, before or after acid hydrolysis (5). They likewise failed to show any significant differences in reaction with the reagent in the two buffers. Biological assays on the rat have shown these compounds to possess practically no vitamin Be activity (17). Bijeseken (3) reported that boric acid forms complexes with 1,2-dihydroxy compounds. Scudi (11) found this to be true also in the case of pyridoxine. In natural extracts other compounds may compete for the boric acid. However, under the conditions of our test, there is always an excess of borate present for complete inactivat,ion of the pyridoxine (5). Thus, in routine assays of a large variety of biological materials, it has been found in every case that added pyridoxine WELT rendered more than 97 per cent non-reactive in the presence of borate. As a result of the tests recorded in Table I, it has been considered expedient in testing biological materials to set the instrument at 1 per cent transmission with a reaction tube containing the test extract and boric acid, 6 seconds after the addition of the reagent. The pigment formed in testing the same extract in the absence of borate is then determined photometrically. By this procedure only the additional light absorption due to the dye formed by the pyridoxine is measured. Increment Procedure for Evaluating Photometric Density of Dye Formed by Pyridoxine-In the presence of relatively high concentrations of some interfering coupling compounds, the reaction between pyridoxine and 2,6- dichloroquinonechloroimide may be partially or completely inhibitsed within the period of measurement. Illustrative data with two such interfering compounds are presented in Table II. In the presence of sufficient resorcinol or naphthylamine hydrochloride, the photometric density due to a given amount of pyridoxine is much less than when the reaction is carried out in pure solution. Thus, evaluation of the photometric density due to pyridoxine, by interpolation on a reference curve derived from pure solutions or the vitamin after correction for the borate blank, would yield erroneously low values. However, the photometric density per unit of pyridoxine (K value) in a given solution is a constant. The interferences noted are undoubtedly due to the removal of much of the 2,6-dichloroquinonechloroimide by the high concentrations of the resorcinol or naphthylamine. Despite this, the additional concentration of pyridoxine is insufficient to affect significantly the residual chloroimide, so that the K value for pyridoxine remains constant in each case. Consequently, the addition of the standard in the form of an increment to an aliquot of the test solution containing the inhibitory substance offers a means for automatically correcting for the interferences noted. It is recognized that such a procedure constitutes a difference method

7 HOCHBERG, MELNICK, AND OSER 115 involving a sacrifice of precision. This, however, is not serious in view of the excellent reproducibility of the color development. The constancy of the K value has been repeatedly confirmed in tests of biological materials TABLE Interference of Relatively High Concentrations of Other Coupling Compounds with Reaction of Pyridoxine and 2,6-Dichloroquinonechloroimide; Correction of Such Interference by Increment Procedure II None Resoroinol Coupling Naphthylamine hydrochloride compound Y Added pyridoxine Y rest solution - - Same + boric acid K valuet )f pyridoxine increment * The calorimeter was set at 1 per cent transmission with the reagent blank. Readings were taken 6 seconds after the addition of the reagent to the solutions. t The K value is the increase in photometric density per 1 y of pyridoxine per tube. (5). The advantages of the increment or internal standard procedure in assaying a wide variety of materials have also been recognized in the determination of riboflavin (8), niacin (9), and vitamin A (1).

8 116 DETERMINATION OF PYRIDOXINE DISCUSSION As a result of the present studies, a procedure has been evolved for the determination of pyridoxine in test materials. This involves coupling of the vitamin with the 2,6-dichloroquinonechloroimide reagent in a single phase system, correction for the presence of other compounds yielding a blue color by the use of borate which renders the pyridoxine non-reactive, and evaluation of the resultant color by comparison with the increment in photometric density produced when a known quantity of the added vitamin is allowed to react in the same test extract. There are other features of this procedure which make it superior to published methods based on the same coupling reaction. The rate of the reaction is dependent upon the nature of the solvent, its water content, its salinity, and the kind and concentration of the bases present (4,ll). The earlier procedures (11, 2) involve color development in a weakly buffered, generally a-phase water-butanol system, without control of the salinity of the extract tested. The color developments are time-consuming, nonlinear, and non-reproducible in replicate determinations. These procedures, therefore, require a series of standards with each determination. The method described in this report involves a simple quantitative coupling reactionin a l-phase system. The use of a strong ammonia-ammonium chloride buffer of high basicity and salinity eliminates completely the interferences reported (4, 11) in similar reactions, due to the different kinds and variable amounts of bases and salts in the test extracts and the degrees to which they are buffered. The reaction between the vitamin and reagent is now very rapid and more than 3 times as sensitive as before (11). Probably the most serious objection to the procedures previously reported (11, 2), including those with other phenol reagents (1, IS), is the absence of a quantitative blank to correct for the presence of interfering phenols, amines, and other compounds which couple with chloroimide. Scudi has reported (11) unsuccessful attempts to apply a quantitative borate blank for this purpose. The value of his blank was restricted to the qualitative recognition of interfering compounds. When present in test materials, these had to be removed. In most cases (13) this was not possible, requiring the quantitative use of a qualitative blank, as a first approximation, rather than none at all. All these earlier methods depend for specificity entirely upon the preparation of extracts of biological materials free from interfering, coupling compounds. Our experiences with these procedures indicate that this is never attained with complete recovery of added pyridoxine in the assay of natural products. A compensating error may operate to some extent in these methods, since no provisions are made to correct for the possible presence of compounds which inhibit the reaction between pyridoxine and reagent.

9 HOCHBERG, MELNICK, AND OSER 117 SUMMARY By studies conducted with pure systems the basis is established for a chemical method for the determination of pyridoxine. The vitamin couples with 2,6-dichloroquinonechloroimide in a l-phase system. The use of a strong ammonia-ammonium chloride buffer of high basicity and salinity eliminates completely the interference due to the different kinds and amounts of bases and salts in the test solutions and the degrees to which the latter are buffered. The color reaction between the vitamin and reagent reaches maximal intensity in 6 seconds and is more than 3 times as sensitive as previously attained. By the incorporation of a borate blank, the reaction is made specific for pyridoxine. The influences of compounds which affect the rate or extent of formation of the pigment and possibly its stability have been eliminated by the increment or internal standard procedure. This involves evaluation of the color due to the pyridoxine in a test solution by means of the increment in photometric density resulting from the reaction of an added, known quantity of the vitamin. BIBLIOGRAPHY 1. Bina, A. F., Thomas, J. M., and Brown, E. B., J. Biol. C&m., 146, 111 (1943). 2. Bird,. D., Vandenbelt, J. M., and Emmett, A. D., J. Biol. Chem., 142,317 (1942). 3. Boeseken, J., Rec. truv. chim. Pays-Bus, 4, 553, 568, 578 (1921). 4. Gibbs, H. D., J. Biol. Chem., 72, 649 (1927). 5. Hochberg, M., Melnick, D., and Oser, B. L., J. Biol. Chem., 166, 119 (1944). 6. Hochberg, M., Melnick, D., and Oser, B. L., J. Biol. Chem., 166, 129 (1944). 7. Hochberg, M., Melnick, D., Siegel, L., and Oser, B. L., J. BioZ. Chem., 146, 253 (1943). 8. Hodson, A. Z., and Norris, L. C., J. BioZ. Chem., 131,621 (1939). 9. Melnick, D., Cereal Chem., 19, 553 (1942). 1. Oser, B. L., Melnick, D., and Pader, M., Ind. and Eng. Chem., Anal. Ed., 16, 724 (1943). 11. Scudi, J. V., J. BioZ. Chem., 139,77 (1941). 12. Scudi, J. V., Bastedo, W. A., and Webb, T. J., J. BioZ. Chem., 136,399 (194). 13. Scudi, J. V., Buhs, R. P., and Hood, D. B., J. BioZ. Chem., 142, 323 (1942). 14. Siegel, L., Melnick, D., and Oser, B. L., b. BioZ. Chem., 149, 361 (1943). 15. Stiller, E. T., Keresztesy, 6. C., and Stevens, J. R., J. Am. Chem. Sot., 61, 1237 (1939). 16. Swaminathan, M., Nature, 146, 78 (194). 17. Unna, K., Proc. Sot. Exp. BioZ. and Med., 43, 122 (194).

10 CHEMICAL DETERMINATION OF PYRIDOXINE. REACTIONS IN PURE SYSTEMS Melvin Hochberg, Daniel Melnick and Bernard L. Oser J. Biol. Chem. 1944, 155: Access the most updated version of this article at Alerts: When this article is cited When a correction for this article is posted Click here to choose from all of JBC's alerts This article cites references, of which can be accessed free at tml#ref-list-1