40 JOHN H. NORTHROP 



2. Antigen (or its decomposition products) acts as a co-enzyme^" in 

 this reaction and causes the formation of a slightly different protein, in 

 the same way that the substrate causes the formation of the adaptive 

 enzymes (page 34). The resulting protein reacts with the co-enzyme 

 (antigen) which led to its formation since products of enzyme reactions 

 in general react with the enzyme which produced them. 



3. The antibody is formed almost entirely in the cells where proteino- 

 gen is synthesized, since the quantity of proteinogen present in the blood 

 at any time is probably small. 



These assumptions account for the following facts : 



1. Antibodies are related to normal serum proteins and react with 

 the antigen which led to their formation. 



2. Small quantities of antigen can produce indefinite amounts of 

 antibody. 



3. Antibody production may continue after all antigen has disap- 

 peared from the circulation but not after all decomposition products of 

 the antigen have disappeared (Harris and Ehrich, 1946). In this respect 

 it differs from Burnet's and Sevag's hypotheses. 



4. The antibody may be related to one or several normal proteins, 

 depending on which autocatalytic reaction is affected by the antigen. 



5. More antibodies will not be formed when antiserum is injected 

 into an animal, since antibody formation itself is not assumed to be 

 autocatalytic. This prediction also differs from Burnet's hypothesis. 



6. Antibodies are not formed in any appreciable amount in the absence 

 of cells. The proteinogen is formed in the cells, since it is only in the 

 cells that the necessary mechanism exists for the energy supply required 

 for synthesis. It will be transformed into the various special proteins of 

 the circulation as soon as it comes in contact with them. The formation 

 of antibodies will therefore occur principally in or near the synthesizing 

 cells. All the various special proteins, including antibodies, are assumed 

 to be in equilibrium with the proteinogen and hence are in equilibrium 

 with each other. Therefore it should be possible to prepare antibodies 

 in vitro by adding antigens to blood. The quantity of proteinogen ac- 

 tually present is assumed to be very small, so that the rate of reaction 

 may be slow. There is the further technical difficulty that whole blood 

 cannot be used (cf. page 39). Production of antibodies in this way has 

 been frequently reported (cf., for instance, Ostromuislenskii, 1916; 

 Pauling, 1940). 



The hypothesis does not account in any simple way for the observa- 



12 Sevag (1945) assumes that the antigens themselves act as enzymes. 



