V. MOLECULAR MECHANISM OF MUTATIONS 227 



nitrous acid should induce only, or predominantly, base "transitions" and 

 that the deamination of G should not be mutagenic since the resulting 

 base xanthine could still pair with cytosine. To prove this Vielmetter and 

 Schuster (19601 compared at different pH's the rates of the deamination 

 of the three DXA bases with the rate, per time unit, of mutation induc- 

 tion. They found that the deamination rates of A and C decrease with 

 increasing pH as rapidly as the mutagenic rate while that of G decreases 

 about one-third as fast. This shows that the deamination of G cannot 

 be mutagenic. Since the lethal rate of nitrous acid treatment also de- 

 creases slowly with increasing pH these authors propose that the deam- 

 ination of G is lethal. This conclusion is not compelling since nitrous 

 acid exerts side reactions in addition to deamination; one can see this 

 when dC]MP is treated with nitrous acid in vitro and then chromato- 

 graphed. Furthermore, Geiduschek (19611 has shown that nitrous acid 

 causes cross-linking of DXA which would be almost certainly lethal for 

 phages. The pH dependence of all these side reactions has not yet been 

 measured. Another side reaction is the lethal effect of nitrous acid on 

 protein; for T2 or T4 phages this is a relatively small effect, since only 

 about 1 out of 10 lethal hits (for definition see Section IV,C) prevents 

 the injection of phage DXA (Harm, 1960; E. B. Freese and E. Freese, 

 1961). 



The induction of mottled plaques after nitrous acid treatment of 

 standard type phage T2 and of non-mottled mutant plaques after treat- 

 ment of phage </)X174 has been used as a further proof that T2 has 

 double-stranded and <^X174 has single-stranded DX'A (Tessman. 1959). 



2. Hydroxylamine (HA) and Hydrazine 



Both hydroxylamine (HA) and hydrazine react with certain pyrim- 

 idine bases but in spite of their chemical similarity show very different 

 chemical specificitj'. 



The effect of hydrazine on uracil has been long known (Fosse et al., 

 1924) ; it has been examined again for UMP and also for C]\IP by Baron 

 and Brown (1955). Hydrazine breaks the ring of uracil and cytosine, 

 giving rise to pyrazolone and 3-aminopyrazole respectively, while the 

 remaining urea remains attached to the sugar and in the presence of 

 water is subsequently hydrolyzed off. Treatment of RX"A by anhydrous 

 hydrazine produces "ribo-apyrimidinic acid" free of pyrimidines (Take- 

 mura, 1957) and the treatment of DX'A produces the corresponding 

 "apyrimidinic acid" (Takemura. 1959). The author finds a small amount 

 of thymine remaining on this DX'A. In contrast, treatment of the nucleo- 

 tides in a 1.0 M aqueous solution of pH 8.5 gives a reaction (decrease 

 of UV absorption at 276 m^n) with dT:\IP. U:\IP. and BUdR while 



