260 THE ANTIOEN-ANTIBODY BEACTIONS 



Similarly with the fractionation of a native antigen ; the loss of specificity 

 that follows the removal of a recognizable chemical compound does not necessarily 

 mean that the compound as such is the determinant ; and even if the specificity 

 is restored when the compound is reintroduced the essential determination of 

 specificity lies in the restoration of the relations between the compound and the 

 remainder of the molecule. 



Full proteins are usually antigenic, though among the larger mammals the 

 antibody response to protein antigens of other mammalian species is partly con- 

 ditioned by the function as well as the structure of the protein. As Harington 

 (1940) has pointed out, the metabolic proteins, like serum globulin, are fully 

 antigenic ; the storage proteins, like lens proteins and casein, less so ; and hormonal 

 proteins, like insulin and thyroglobulin, apparently shared by a large number 

 of animal species, have little or no antigenic power. 



The early successes in the study of protein antigens, the early failures to demon- 

 strate antigenic response to the injection of carbohydrates and fats, and the fact 

 that most of the known antigens contained at least a substantial amount of protein, 

 led to the supposition that all antigenicity was conferred by protein, and that 

 in the alleged examples of non-protein antigens, the substances were contaminated 

 with chemically undetectable, but immunologically active, traces of protein. It 

 is in many cases difficult to exclude the possibility of antigenic contaminants in 

 a purified preparation. The scanty work, for example (see Marrack 1938), on 

 the antigenicity of lijiins, suffers from this defect. There is another way in which 

 a substance may be erroneously called antigenic. It may combine with the native 

 proteins of the animal into which it is introduced, and convert them into " foreign " 

 proteins that stimulate antibody formation. We shall discuss examples of this 

 instructive phenomenon in connection with hypersensitivity (Chapter 51). But 

 carbohydrate substances have been prepared whose protein content is so low that 

 the possibility of an effective antigenic stimulus may be neglected, but which act 

 as full antigens, and the unique efficacy of proteins in conferring full antigenicity 

 can no longer be upheld. 



Among the proteins, gelatin is not antigenic. It lacks tyrosine and tryptophan, and 

 has only a little phenylanine, a fact which suggests that the aromatic radicles are neces- 

 sary for antigenicity (see Wells 1925). Hopkins and WormaU (19336) failed to make 

 gelatin antigenic by introducing aromatic radicles. Hooker and Boyd (1933), however, 

 demonstrated that antibodies were formed to a diazo arsanilic acid gelatin complex, and 

 Glutton, Harington and YuiU (1938) had a similar success with glucosido-tyrosyl gelatin. 

 The absence of an aromatic radicle is not, however, the sole reason for non-antigenicity. 

 Insulin, which is a fuU protein, containing aromatic radicles, is in no sense a full antigen. 

 Various workers have reported anaphylactic sensitization with insulin preparations (Barral 

 and Roux 1931, Lewis 1937, Bernstein, Kirsnerand Turner 1938), and Wasserman, Broh- 

 Kahn and Mirsky (1940) record the production of complement-fixing antibodies in the 

 rabbit. Insulin nevertheless does not stimulate frankly precipitating antibodies, but the 

 introduction of a glucosido-tyrosyl residue converted insulin into a fuU antigen (Glutton, 

 Harington and YuiU 1938). That tyrosine and other aromatic radicles are not aU-important 

 is also shown by the masking of original specificity that occurs when determinants are linked 

 to amino-acid residues containing no aromatic rings. From the alterations induced in 

 serum pseudoglobulin by different degrees of substitution of azo-atoxyl, Haurowitz, 

 Sarafian and Schwerin (1941) concluded that specificity depended on a determinant 

 arrangement of tyrosine, free amino and perhaps other groups xyn the molecular surface. 

 In native lens proteins, the sulphydryl groups are apparently important determinants 

 (Ecker and Pillemer 1940). 



