Drug Resistance in Myco. tuberculosis 245 



Let us assume that isoniazid in high concentrations blocks 

 the formation of a specific peroxidase (almost certainly not 

 catalase) essential for the growth of the tubercle bacillus, at 

 any rate in conditions in which peroxides are produced. 

 Mutants occur naturally, at a rate of about 1 in 10^, which 

 are deficient in the gene or genes controlling the synthesis of 

 this peroxidase. Such mutants (1) cannot grow in media in 

 which peroxides are formed, (2) can grow in such media pro- 

 vided the peroxides are destroyed, e.g. by haemin (Fisher, 

 1954; Knox, 1955) or catalase (Middlebrook, 1954; Knox, 

 1955), (3) could grow in media in which peroxides were not 

 formed, (4) must be logically, and are in fact, isoniazid- 

 resistant because they lack the peroxidase-forming enzyme 

 system which is ex hypothesi the system vulnerable to isoni- 

 azid. The DNA of such mutants is abnormal — either because 

 one specific DNA molecule or part of a molecule is lacking or 

 because normal DNA is replaced by an abnormal DNA. 

 Whatever the mechanism, the result is a DNA lacking the 

 ability to control peroxidase synthesis. 



The second type of isoniazid-resistant culture is one 

 induced by isoniazid. Whereas in high concentrations 

 isoniazid completely blocks the formation of a specific perox- 

 idase and therefore the only survivors will be cells deficient 

 in the vulnerable system, in lower concentrations isoniazid 

 can induce the formation, by an adaptive process, of an 

 abnormal peroxidase system. It is even possible that the 

 abnormality might be in the protein part of the peroxidase- 

 producing gene itself. If so, then the stable highly resistant 

 cell would have an abnormal or missing gene and therefore 

 might lack the protein and so the enzyme ; while the unstable 

 pseudomutant would possess normal DNA but, under the 

 influence of isoniazid, would produce an abnormal protein. 

 If this were actually part of the gene it might be reduplicated 

 for some time even in the absence of isoniazid, but eventually 

 a return to normal protein synthesis would occur. A test of 

 such a hypothesis might be made by extracting transforming 

 substances from tubercle bacilli, but our efforts in this direction 



