The interpretation of the results in these experiments is difficult on account of the variations in the physical capacity of different strains of hemolytic streptococcus to agglutinate and absorb agglutinin. Even when the antigens for agglutination and absorption are standardized and sera are carefully titrated variations in agglutination occur with different strains and sera. Variations may occur because of differences in the titer of the immune sera employed. In a large series of reactions with the same sera and strains, errors in interpretation depending on these factors can be eliminated because strains which agglutinate poorly will consistently give weak reaction with all sera, and weak sera will agglutinate all strains poorly. after such variations
See PDF for Structure
in agglutination are eliminated, if a strain varies grossly in agglutination with different sera, or the titer of different sera varies with the same strain as in Tables I to III, variations which occur must represent antigenic dissimilarities.
We have previously commented on the dissimilarity of the individual strains in the scarlatinal and erysipelas groups. Although the strains within each of the groups are related they are seldom identical. This individuality of the strain was indicated by the fact that absorption by heterologous strains failed to absorb the agglutinin for the strain homologous with the serum. This strain individuality is again apparent in Tables I to III. Doses of heterologous strains equivalent to the unit absorptive dose which has already been defined, fail to remove the agglutinin for the homologous strain. This dose is sufficient however, to absorb the agglutinin for practically all heterologous strains. It appears from this observation that the strain specificity dominates the group specificity.
Between the unit absorptive dose and a dose which is approximately 0.2 to 0.3 of this unit, is a zone of great variation in absorption and agglutination. Below this zone absorption is complete for few strains. The differences in titer of various strains and absorbed sera in the zone where variations occur are very definite. For example in Table II, Strain E IX agglutinates well in Serum II absorbed with Strains E I and E III but poorly in serum absorbed with Strain E II. Numerous similar examples may be found. In a few instances heterologous strains agglutinate nearly as well as the homologous strains in these absorbed sera. Such strains probably resemble the homologous strains closely. These resemblances are not confined to the erysipelas strains but occur among the scarlet strains which agglutinate in erysipelas sera. Apparently it is impo—ssible to distinguish such scarlatinal strains from erysipelas strains by absorption of agglutinin.
The relationship shown by these absorption tests is additional evidence in favor of the mosaic nature of the antigen. Strains of streptococcus of the erysipelas groups are apparently composed of several agglutinogenic fractions. If the antigen of a strain is made up of the fractions A, B, C, D, and E, serum for that strain would contain A, B, C, D, and E agglutinins. If this serum were absorbed with the homologous strain or an identical heterologous strain the agglutinin would be completely absorbed. On the other hand a qualitative or quantitative difference between the two strains would be apparent in the absorption and agglutination reactions. Such differences are evident in Tables I to III. This serum would agglutinate any strain containing any one or several of the fractions, but absorption of the serum by such a strain would leave fractions of the agglutinin in the serum. This absorbed serum would agglutinate the homologous strain. Heterologous strains would agglutinate if fractions remained in the serum corresponding to their antigenic components. Hence the reactions in the absorbed serum would depend on the qualitative and quantitative relationships between the component fractions in the serum, the absorbing strain, and the strains agglutinated.
This conception of multiple antigens is not new in bacteriology. Durham (3) explained the reactions in the colon-typhoid group of bacilli on this basis. More recently Durand and Sedallian (4), and Andrewes, Derick, and Swift (5) have expressed the opinion that the agglutination reactions with hemolytic streptococcus can only be accounted for in this way. In many respects the reactions observed with the colon-typhoid group of Gram-negative bacilli resemble those experienced with the erysipelas and scarlatinal groups of hemolytic streptococcus. Apparently we have exhausted the possibilities of studying these groups further by agglutination and absorption. Our knowledge regarding the specificity and relationship of the antigenic fractions must come from the study of fractions isolated and refined by chemical methods.