226 PYRIDOXINE AND RELATED COMPOUNDS 



shift to shorter wave lengths. Figure 1 illustrates the absorption spectra 

 of pyridoxine at various pH values. ^^ 



The pK (base) value of pyridoxine was found to be 6.2 X iO"^".''* 



C. CONSTITUTION 



1. Introduction 



Only a short time elapsed between the isolation of the vitamin in crystal- 

 line form and the recognition of its chemical constitution by two independ- 

 ent groups of chemists, Stiller, Keresztesy, and Stevens in the United 

 States, and Kuhn, Wendt, and Westphal in Germany. ^^-^^ 



2, Proof of Structure 

 a. Stiller, Keresztesy, and Stevens 



The establishment of the structure of pyridoxine by Stiller, Keresztesy, 

 and Stevens^^ was based upon the study of the properties of the vitamin as 

 isolated and its methoxy derivative. The elemental y analysis of pyridoxine 

 (as the free base) gave the empirical formula of CsHnNOs . Theelectrometric 

 titration curve of the hydrochloride showed only one break, indicating 

 that the salt is a monohydrochloride and that its formula must be 

 CgHuNOrHCl.^^ Analysis showed it contained one C-methyl group. O- 

 methyl and N-methyl groups were absent. Its failure to react with ni- 

 trous acid, its phenolic reaction with ferric chloride, and its ultraviolet 

 adsorption characteristics pointed to the probability that the vitamin was 

 a derivative of /3-hydroxy pyridine. Comparison of the ultraviolet absorp- 

 tion of pyridoxine with that of 2-methyl-3-hydroxy-5-ethylpyridine con- 

 firmed this point. ^^ 



Methylation of the base with diazomethane resulted in the formation of 

 the methyl ether, C9H13O3N, melting point 101 to 102°. The absorption 

 spectrum showed a single maximum at 2800 A. which was not altered b}^ pH 

 changes. Oxidation with permanganate yielded two products. One was a 

 dibasic acid which crystallized with one molecule of water, C9H9O5N, melt- 

 ing point 208 to 209° (dec). The other was a lactone, C9H9O3X, melting 

 point 209 to 210° (dec). 



Since the dibasic acid gave a negative test with ferrous sulfate, it ap- 

 peared that neither of the carboxjd groups was in the a-position of the 

 pyridine ring. Furthermore, the test with resorcinol was positive, and there- 

 fore the two carboxyl groups must be attached to adjacent carbon atoms. 

 Thus the dibasic acid could be represented by either of two structures. 



" R. Kuhn and G. Wendt, Her. 72B, 305 (1939). 



18 R. Kuhn, H. Andersag, K. Westphal, and G. Wendt, Ber. 72B, 309 (1939). 



'9 R. Kuhn, G. Wendt, and K. Westphal, Ber., 72B, .310 (1939). 



