CHEMICAL METHODS IN THE ANALYSIS OF ANTIGENIC STRUCTURE 281 



It was natural to suppose that the polysaccharide component isolated from B. antliracis 

 was the capsular material; but Tomcsik and Szongott (1932, 1933) report that this is 

 not the case. They state that the polysaccharide is a somatic component common to the 

 virulent, capsulated rough-colony-forming type, and to the avirulent, non-capsulated, 

 smooth-colony-forming type. The capsular substance is a protein-like substance. 

 Ivanovics and his colleagues identified the capsular substance as a polypeptide which 

 on hydrolysis yielded c?(-) -glutamic acid almost in the amount to be expected if the starting 

 substance were built up of glutamic acid only. This antigenicaUy distinct capsular poly- 

 peptide was also found in B. mesentericus and certain other members of the Bacillus group 

 (Ivanovics and Erdos 1937, Ivanovics and Bruckner 1937a, b, 1938). 



There is another way in which chemical methods may be applied in the investigation 

 of antigenic structure. Bacteria may be treated with various chemical solvents, or other 

 reagents, with or without heat, and the effect of this treatment on their immunological 

 behaviour may be studied (see White 1927, 1928, 1929, 1932, 1933). 



As an example of an indirect method of chemical analysis, we may cite the comparison 

 of the serological reactions of a natural antigen with that of a " synthetic " antigen whose 

 hapten is of a known chemical nature. Thus Goebel (1936) confirmed the importance 

 of glucuronic acid in the molecular structure of the capsular polysaccharide of Types II, 

 III and VIII pneumococci, by demonstrating the cross-reactions between the poly- 

 saccharides and antisera to glucuronic acid azo-proteins. The method is not entirely 

 specific since a group of non-identical haptens having certain chemical similarities may 

 cross-react with one another. Thus the precipitation of antibody to galactoside-azo- 

 protein by the polysaccharide of pneumococcus Type I confirms the importance of the 

 galacturonic acid identified in the polysaccharide molecule ; but antisera to azo-protein 

 antigens containing a benzene carboxylic radicle (which has never been identified in 

 Type I pneumococci) also precipitate with the polysaccharide (Goebel and Hotchkiss 1937). 

 The cross-reaction in this case appeared to depend on the reactivity of the antisera with 

 acidic groups, irrespective of the hapten radicle that bore them (see also Goebel 1940). 



The first essential for the characterization of a bacterial product is the certainty 

 that it has been synthesized by the bacterium and not derived from the culture 

 media from which it was obtained. It will be obvious that blood and other animal 

 proteins used in culture and adsorbed to the bacterium, may induce the formation 

 of antibodies that are later confused with those induced by the bacteria (see, for 

 example, Bliss 1938). 



Bailey and Raffel (1941) record that even infusion broths will substantially modify 

 the antigenicity of bacteria grown in them. In association with bacteria, non -antigenic 

 substances like agar may be antigenic (SordeUi and Mayer 1931, Morgan 19366, Sickles 

 and Rice 1938). The mechanism of this association is unknown, but Partridge and Morgan 

 (1942) have provided a model for it by making artificial antigenic complexes of agar and 

 the conjugated protein component of the endotoxin of Sh. sliigce. 



The development of culture media whose components are all known and all of relatively 

 low molecular weight wiU eliminate this source of error. (See, for example, Freeman 

 et al. 1940.) 



It should be emphasized that only in a few cases has the chemical constitution 

 of antigens been fully determined. The antigenic substances isolated from bacteria 

 are usually either large molecules, or large associations of molecules. The inter- 

 pretation of their biological activity in terms of chemical structure must always 

 be tentative in the absence of satisfactory evidence that the product is pure. 

 Unfortunately the biologist's and the chemist's conception of purity are not always 

 coincident, and each tends to use criteria borrowed from the other, without full 

 appreciation of their limitations. For instance, the fact that a substance has 



