MUTAGENESIS 155 



Zamenhof: Yes, that is the reason. Just how many things do you 

 have to change at once to change mutability of any particular spot? 

 Is it directed by one neighbor, two neighbors, or the whole cluster? 



Benzer: It might be feasible to do something like this in dealing 

 with reversions, using linked mutants or something you can turn on 

 or off, something with an intermediate phenotype. It might be possible. 



Lederberg: You do have these incomplete reversions, where you 

 know you have more or less wild type, and presumably at least one 

 base substitution, and one could inquire whether this in itself has had 

 any influence on the spectrum of the nearby loci. I'm not proposing 

 that you do it. I really don't see how it can tell you that it is a string 

 of adenines or what. 



Benzer: Since the bonds between adenine and thymine are weaker 

 than the bonds between guanine and cytosine, if you have a stretch of 

 adenine-thymine pairs, it will be much easier to make a purine-purine 

 pair, and thereby a mutation, than in a part of the chain which is held 

 together much more tightly. That would illustrate one type of neighbor 

 interaction. 



Atwood: To return to your conjecture that the hot spots are left 

 over from a trend toward elimination of hot spots in evolution, I 

 think the idea that you wouldn't expect to go to completion in this 

 trend is more likely since the hot spot depends on an array more or 

 less around the region — I mean, we don't know how far from it, but 

 it depends on the entire sequence, and so does the rest of the phenotype 

 that is subject to selection. Then you always have to have a com- 

 promise. If it is bad for the thing to have a hot spot, it is also bad for 

 it to change the sequences so as to avoid the hot spot, and you're going 

 to end up with a compromise between the number of hot spots and 

 the phenotype. 



Benzer: Yes. 



Lederberg: Or, alternatively, if you had the wild type so hot that 

 you had ten thousand out of a million mutants at one culture, it 

 would take a selective advantage on the part of the wild phenotype to 

 maintain itself against that mutation. 



Demerec: I'm wondering if you are familiar with Streisinger's work 

 or his claim that he found five different states in a single site. 



Benzer: Yes. 



Goodgal: How many mutants did he test? 



Benzer: He seems to have four distinguishable mutants at the same 

 site. 



Goodgal: Out of how many? 



