METABOLISM OF AZOTOBACTER 269 



product, whereas the other species produce large quantities of butyric 

 acid. Some of these products may be accounted for as f ollows : 



It is known that when sugars, such as glucose, levulose and man- 

 nose are acted upon by alkalies, there are produced a great many 

 products, some of which are formic, carbonic, oxalic, lactic, pyruvic 

 tartronic, malic, malonic, tartaric, ribonic, saccharic, and gluconic 

 acids in addition to many other either more or less complex com- 

 pounds. We can readily conceive that the Azotobacter brings about 

 a somewhat similar reaction, the stages, however, being more nicely 

 governed, because of enzymes. Many of the products would be 

 oxidized to carbon dioxid and water with the liberation of energy 

 necessary for the endothermic nitrogen reaction; others readily 

 react with the resulting nitrogen compounds. We are completely in 

 the dark as to what this first nitrogen compound is, but we know that 

 the A~otobacter possess the power of changing nitrates or nitrites 

 under appropriate conditions into ammonia. Up to date it has been 

 impossible to detect nitrate formation; it is not impossible that 

 nitrates are formed and utilized by intracelltilar enzymes. By using 

 nitrates, nitrites or ammonia, we can offer a rough explanation of 

 protein anabolism. 



The endothermic reaction, 



2N+2H 2 O = NH 4 NO 2 , 



may take place and the ammonia thus formed may react with the 

 decomposition products of the sugars, pyruvic acid for instance, 

 with the formation of alanin which Lipman considered as one of 

 the first products: 



CH 3 CO COOH + NH 3 = CH 3 CHN COOH + H 2 O 

 CH 3 CHN COOH + H 2 = CH 3 CHNH 2 COOH 



or with glyoxylic acid forming glycocoll : 



HCO COOH + NH 3 = HCNH COOH + H>O 

 CHNH COOH + H 2 = CHaNHs COOH 



By similar reactions other amino-acids may be formed. More- 

 over, Windas and Knoop have shown that rnethylimadazol may be 



