90 MUTATIONS 



found that the 5-bromouracil-induced mutant frequency does not 

 increase when bacteria are subjected to 20 cycles of 5-bromouracil 

 incorporation and replacement by thymine; furthermore, when bac- 

 teria are grown in 5-bromouracil and tested on media containing 

 thymine only, the induced mutant frequency was the same as when 

 tested on medium containing 5-bromouracil. 



Auerbach: I mean, when you added more thymine, did you get 

 fewer mutations? 



Benzer: It is technically difficult to distinguish between mutations 

 produced by the bromouracil going in or by the bromouracil being 

 in, because you get so many mutations during the incorporations step. 



Auerbach: I mean something much more radical. Could not bromo- 

 uracil act from outside the DNA? Certainly, caffeine must do so, but 

 could not bromouracil do the same? Must it act by incorporation? 



Freese: There is no evidence that proves that it acts by incorpora- 

 tion. 



Auerbach: Except that it is the most plausible interpretation. 



Freese: Yes, this is the most plausible. 



Goodgal: Just one comment. You don't really mean that concentra- 

 tion has no effect on the mutation rate? 



Freese: I didn't say that. I said the amount of incorporation. 



Goodgal: Yes, but then there must be some level at which the fre- 

 quency would depend on the amount of incorporation, isn't there? 



Freese: Let me tell you about an experiment which one could think 

 of, which one cannot do at the present time, but may be able to do in 

 the future, with DNA polymerase. It would be to add both thymine 

 and bromouracil and extremely little cytosine. In this case, you would 

 preferentially make mistakes by which the bromouracil is incorporated 

 in place of the cytosine, and you would prevent the incorporation of 

 the bromouracil by adding much thymine. In this case, you should get 

 many mutations if you could prove that you have mutations in the 

 DNA, but you should not get much incorporation. 



Zamenhoj: I could suggest an experiment which would follow along 

 the lines that you have mentioned. If one incorporates 5-bromouracil 

 into DNA having transforming activity (19), and then subject it to 

 Meselson's density gradient centrifugation (72), one eventually comes 

 out with three fractions (peaks) ; that is, one corresponding to mole- 

 cules without any bromouracil, another one with only one strand 

 labeled, and the third one with two strands labeled. Now, the specific 

 transforming activity of each of these peaks is practically the same. 

 It makes no difference for the transforming activity whether there is 



