224 ROBERT L. SINSHEIMER 



mine might be expected to favor enolization of 5-bromouracil as compared 

 to thymine; 2-aminopurine can act similarly. 



Nitrous acid is believed to act by oxidative deamination of cytosine 

 or adenine residues. Such deamination will cause erroneous pairing at the 

 next DNA replication. 



All three of these agents, then, are expected to cause replacement of 

 adenine-thymine by guanine-5-hydroxymethylcytosine pairs and vice versa. 

 Hence, each of these mutagens should be very effective agents to produce 

 reversions of mutations caused by any of this mutagenic group, and this 

 is found to be the case. 5-Bromouracil can induce reversion of 98% of 

 the rll mutants induced by 2-aminopurine, 95% of its own mutations, 

 and 87% of those induced by nitrous acid 159 (as these agents increase 

 the mutation rate by two or more orders of magnitude, the background 

 of spontaneous mutations can be largely ignored). In a few cases, the mu- 

 tations are apparently more complex. It is of interest that mutations in- 

 duced by 5-bromouracil tend to be more easily induced to revert by 2- 

 aminopurine, and vice versa, although exceptions are known. 156 



The action of pH 5 at 45° is believed to result in occasional loss of a 

 purine ring from the DNA chain. It might be expected that, during repli- 

 cation, this gap in the template could be paired by any of the four nucleo- 

 tides available. However, 77% of such pH 5 mutants are reversible by 

 5-bromouracil or 2-aminopurine. 159 As mentioned, these agents are believed 

 to be able to interchange adenine-thymine and guanine-5-hydroxymethyl- 

 cytosine pairs but presumably could not interchange a purine-pyrimidine 

 pair for a pyrimidine-purine pair. Hence, it would appear that on most 

 occasions the gap in the template, left by the loss of a purine, is paired 

 by a pyrimidine nucleotide. 



The nature of proflavine mutagenesis is obscure. Proflavine increases 

 the mutation rate at the rll locus some thirtyfold. The mutants produced 

 are relatively homogeneous as to spontaneous reversion rate. It is of con- 

 siderable interest that most proflavine-induced mutants, and spontaneous 

 mutants, are almost entirely nonreversible by 5-bromouracil or 2-amino- 

 purine and hence involve a different kind of modification of nucleic acid 

 than those induced by the latter agents. Similarly, 5-bromouracil muta- 

 tions are not reversible by proflavine although proflavine mutants are self- 

 reversible. 156, 164 No coincidences were observed between the sites of 55 

 proflavine-induced mutants and 67 5-bromouracil-induced mutants. 161 Pro- 

 flavine and most spontaneous mutations may then represent replacement 

 of a purine-pyrimidine pair by a pyrimidine-purine pair. 



It is of interest that Astrachan and Volkin 165 have demonstrated chro- 



164 S. Brenner, Brookhaven Symposia in Biol. 12, 74 (1959). 



165 L. Astrachan and E. Volkin, J. Am. Chem. Soc. 79, 136 (1957). 



