VARIETIES OF ANTIBODY GLOBULIN 245 



Many attempts have been made to dissociate antibodies from an antigen- 

 antibody complex in a protein-free condition. The dissociation of the antigen- 

 antibody compound may be accomplished in a variety of ways, and certain of the 

 antibody solutions so obtained have contained very little jDrotein (see Landsteiner 

 and Jagic 1903, Muir 1903, Bail and Tsuda 1909, Spat 1910, Kosaki 1918, Huntoon 

 and Etris 1921, Huntoon et al. 1921, Locke and Hirsch 1925), but in none of these 

 instances has it been satisfactorily demonstrated that the amount of protein 

 present was insufficient to account for the antibody action observed (see Eagle 

 1930). Similar attempts have been made by adsorption on to various reagents 

 other than the specific antigen, followed by elution with various fluids at different 

 pH levels. Some of these attempts have been claimed as successful (see Frankel 

 and Olitzki 1930, Olitzki and Frankel 1931, Frankel 1932, Olitzki 1932), but these 

 claims have not been confirmed by subsequent workers (see Marrack 1934, Rosen- 

 heim 1935). 



Northrop (1941) and Rothen (1941) prepared purified diphtheria antitoxin from 

 toxin-antitoxin precipitates by digestion with trypsin. The purified antitoxin had 

 a molecular weight in the region of 90,000. By fractional precipitation Northrop 

 isolated a crystalline antitoxic protein which appeared to be antigenically distinct 

 from normal horse proteins. 



Later studies of the effect of electrolytes on specific aggregation (see p. 215) 

 led to the preparation of dissociated antibody solutions of a high degree of purity 

 (Heidelberger and Kendall 1936, Heidelberger, Grabar and Treffers 1938). Heidel- 

 berger and his associates were able to purify both rabbit and horse antisera by 

 salt dissociation of agglutinated pneumococci. Types I, II and III. Dissociated 

 and native antibody behaved alike in serological reactions. Dissociated antibody 

 had the properties of a globulin. The molecular weight of dissociated rabbit 

 antibody, calculated from its sedimentation rate in the ultracentrifuge, was, like 

 that of antibody in native antiserum and of globulin in normal serum, in the 

 region of 150,000 to 196,000. 



Varieties of Antibody Globulin. 



Horse antibody to the Type III pneumococcus has a molecular weight over 

 4 times that of the corresponding rabbit antibody. A similar component was 

 found in other preparations of horse antisera, but not in normal horse serum 

 (Heidelberger and Pedersen 1937). Later, Kabat (1939) showed that these heavier 

 antipneumococcal globulins, of molecular weight 910,000 to 930,000, were formed 

 also in the cow and pig. 



These antibodies of high molecular weight differ in certain respects. For instance, 

 " heavy " antibody to pneumococcal polysaccharide will not fix guinea-pig or human 

 complement in the presence of homologous antibody, whereas ox antibody of similar 

 weight does so (Heidelberger and Treffers 1941). Heavy antibodies do not appear to be 

 quite so firmly constituted as antibodies of a lower molecular weight, for they are dis- 

 aggregated by relatively mild treatment with barium hydroxide without much reduction 

 of precipitating power. Not all horse antibody has this high molecular weight ; for 

 example, diphtheria, scarlet fever, tetanus and CI. welchii antitoxins from the horse, when 

 diluted, contain no abnormally heavy protein molecules (Fell, Stern and CoghiU 1940). 



The significance of the large antibody molecules is not clear, but it is evident that 

 response to immunization on the part of certain animals is more than the production 

 of normal globulins modified only in configuration ; a new type of globuhn may in fact 

 be formed. 



