II. CHEMISTRY 115 



I'rcatnuMit of loucoN'oriii with mineral acid t'ollowod l)y oxidation with 

 alkaHiie |ieiniangaiuite gave 2-amino-4-liydroxypteri(Hne-()-carh()\ylic a(;id 

 in 68% yield. Since no 2-amino-4-hydroxypteridine-(),7-dicarboxylic acid 

 was found, the existence of a carbon-to-carhon linkage to the 7 position of 

 the tetrahydropteridinc ring seems ((uite unhkely. This narrowed the in- 

 \estigation to the remaining possibilities of a shift of the formyl group, or 

 a new ring formation involving the 5 or 8 position. From inspection of 

 molecular models it seems more logical to choose the 5 position because of 

 its proximity to the 10-formyl group. 



Further evidence for the location of the formyl group on the 5 position 

 has been provided by the study of numerous model compounds. The 

 properties exhibited by 2,4-diamino-(i-hydroxy-5-formamidopyrimidine**'' 

 91 , 92 ^YQ found^" to be strikingly similar to leucovorin and other mono- 

 formyltetrahydropteridines, some of which are discussed below^ in more 

 detail. Thus, the 5-formamidopyrimidine is stable to h3''drolysis upon heat- 

 ing for 1 hour in 0.1 A^ sodium hydroxide solution. It does not react with 

 nitrous acid in the cold, but, after hydrolysis of the formyl group with 

 mineral acid at room temperature, reaction with 1 mole of nitrous acid 

 occurs. Since the 2- and 4-amino groups in this model pyrimidine are inert 

 to formylation and nitrosation, it is logical to assume that the 2 and 8 

 ]wsitions of the tetrahydropteridinc are similarly unreactive; this again 

 points to the 5 position of leucovorin as the most probable site for the 

 formjd group. 



Both 2-amino-4-hydroxy-G-methylpteridine (XX)-^ and 2-amino-4-hy- 

 dioxy-(),7-dimethylpteridine (XXI) can be reduced and formjiated to give 

 the corresponding 5-formyl-5,6,7,8-tetrahydropteridines (XXIP" and 

 XXIIP^). As in the case of 2 .4-diamino-6-hydroxy-5-formamidopyrimidine, 

 XXII is stable in 0.1 A" sodium hydroxide solution for 1 hour at 95° and 

 shows no wave in the polarograph at pH 9; however, after treatment with 

 mineral acid, the characteristic tetrahydropteridinc wave is found. There is 

 also no reaction with nitrous acid until the formyl group has been removed 

 by acid hydrolysis, at which time I mole is consumed. Reaction of the 

 2-amin() group in this instance appears unlikely, since it has been shown'-^ 

 that XX does not react with nitrous acid under similar conditions. The 

 ultraviolet absorption cur\'es of both XXII and XXIII are also similar to 

 that of leucovorin, and, as in the conversion of 5,G.7,8-tetrahydro PCiA to 

 leucovorin, XXII and XXIII exhibit a much higher intensity of absorj)- 

 tion than the corresponding unformylated tetrahydropteridines. 



Another experiment which presents almost unequivocal proof that the 

 formyl group is located in the 5 rather than the 8 position involves the 



9' W. Traul)e and H. W. Du.lloy, Her. 46, 3839 (1913). 

 ^~ W. Wilson, ,/. Chcm. Soc. 1948, 1157. 



