In the studies on the mechanism of immunity reactions, Umezawa (1) criticized the comparative merits of two-stage theory of Bordet and the lattice theory proposed by Marrack (2) . His results were reasonably explained by the latter.
It is an essential assumption of the lattice theory that antibodies and antigens must be multivalent to each other, in order to form a framework structure of antigen-antibody complex. One molecule of antibody must possess at least two sites with which it combines antigens. Meanwhile the two-stage theory can explain the formation of precipitates if the antibody is monovalent. The multivalency of antigens was ascertained experimentally by many authors (3-5) . While the problem of the valency of antibody, though it holds sway over the justification of these two theories, had not been decided directly by the time when I undertook the following experiment (1946) . The multivalency of antibody was ascertained by my experiment as follows.
When a rabbit is immunized with horse serum globulin (G) coupled with a simple substance of known chemical structure (a) as a determinant group, it is always observedd that heterogenous antibodies are produced in its serum. If antibodies are bivalent, they may be grouped into next three types, if such are actually produced.
(i) Both combining sites are specific to (a) .
(ii) Both combining sites are specific to (G) .
(iii) One site of combination is specific to (a), while the other is specific to (G) .
If, on the contrary, antibodies are monovalent, their single combining site will be
(i') specific to (a), or
(ii') specific to (G), or
(iii') specific to the linkage of (a) and (G) .
Suppose simple hapten (A) containing the above group (a), is added to the antiserum. Then in the former case, it should bind with antibodies of the types (i) and (iii), while in the latter case with (i') and (iii'), but no precipitate will be formed in both cases.
Then add horse serum globulin to the mixture. If antibodies are bivalent, antibodies of type (ii) are of course precipitated with globulin molecules. But also those of type (iii), only one combining site of which binds with (A), will be taken up into the framework of antigen-antibody complexes with the remaining site, and hence the formed precipitates will contain the hapten (A) . Meanwhile if antibodies are monovalen, those of type (iii'), not to speak of those of type (i'), can not bind with (G), on account of the spacial expulsion of previously bound hapten (A) . Therefore precipitates will consist of (G) and (ii') alone, and substance (A) will not be found in them, except those adsorbed on them non-specifically.
Thus. when a dye or an easily detectable substance is employed as hapten (A), the valency of antibody can be decided clearly from the colour or analysis of the precipitates. Upon this basis, I made the following experiment. Antibodies were found to possess at least two combining sites and form framework structured precipitates with antigens.
抄録全体を表示