SARA E. BRANHAM 717 



there is a possibility of specificity among them. Nevertheless, the types of carbo- 

 hydrates with which we have been long familiar seem to be entirely devoid of any 

 antigenic activity. The production of anti-hemolysins with a purified hemolytic ex- 

 tract of Amanita phalloides has been reported by Ford.' This research has often been 

 quoted as the outstanding example in which antibody production may be incited by 

 means of an antigen free from protein. This active component has been described by 

 Abel and Ford^ as a derivative falling into the group of saponin-like glucosides. This 

 material, however, can scarcely be classified as a glucoside of the saponin type, not 

 only because such glucosides are non-antigenic, but also because the saponins contain 

 no nitrogen in their molecules. The "glucoside" which Abel and Ford described con- 

 tained nearly 11 per cent of nitrogen. 



If one must assume that this unique complex is indeed a nitrogenous pentose- 

 containing glucoside, as Abel and Ford have classified it, the simplest derivative of 

 this type (the one containing the highest theoretical percentage of nitrogen) contains 

 8.6 per cent nitrogen, 39.2 per cent oxygen, and 44.2 per cent carbon: 



CH.OH - CH - CHOH - CHNH, - CHOCH3 . 



I Q 1 



The derivative of Abel and Ford contained 48.9 per cent carbon, 10.8 per cent nitro- 

 gen, and 32.3 per cent oxygen. On the basis of these analytical figures it is obvious 

 that the carbon and nitrogen content of their compound is too high, and the oxygen 

 content too low, to permit it to fall into the class of nitrogenous carbohydrate de- 

 rivatives.^ 



Although Abel and Ford considered their antigenic hemolysin free from protein, 

 they based their contention on the assumption, first, that protein is quantitatively 

 ehminated with metaphosphoric acid and uranyl acetate; and, second, that the result- 

 ing filtrate gave no biuret test. Neither of these reagents are quantitative precipi- 

 tants for proteins, nor can a negative biuret test, sensitive at best to one part of 

 protein in ten thousand parts of water, be accepted as evidence that such filtrates 

 are free from proteins, when one must take into account that antibody response has 

 been secured with amounts of protein as small as o.oooooi gm. 



Abel and Ford have shown that when their hemolytic antigen was dialyzed in 

 collodion bags, part of the glucoside passed through the dialyzing membrane. Al- 

 though the glucosides recovered both from within and without the dialyzing sac ap- 

 peared to be identical, that which had passed through the membrane was not hemo- 

 lytic. Unfortunately, no test for antigenicity of this non-hemolytic fraction was made. 

 Hence it is difficult to accept the conclusions of these investigators that the active 

 hemolytic principle of Amanita phalloides is a nitrogenous glucoside capable of pro- 

 ducing antibodies. It seems, therefore, that this often-quoted example of a non- 

 protein antigen should be subjected to a more detailed and rigorous chemical in- 

 vestigation. 



' Ford, W. W.: J. Pharmacol, b' Exper. Thcrap., 2, 145. 1910-u. 



^ Abel, J. J., and Ford, W. W.: /. Biol. Cliem., 2, 273. 1907. 



' For consultation, I am indebted to Dr. Milton T. Hanke and Dr. Walther F. Goebel. 



