392 KK'II AHI) SCIIWKKT AND .loHN UISHol' 



witliout U could also code, tlit- degeneracy must he even greater. On the 

 other hand, if all codes contain U, the composition of DNA and mes- 

 senger KNA does not corres])ond (does not contain sufficient U) and it 

 must he supposed that i)art of these iiolymicleotides does not code, e.g., 

 is nonsense or is used for some other function. 



An important impetus in suppoi't of the validity of the code letters 

 has come from studies of amino acid replacements in mutant jiroteins. 

 In particular, the nitrous acid-induced mutants are of interest. They 

 are jiroduccd hy nitrous acid treatment of infective TMV-RNA. After 

 infection and rei)lication, mutants are isolated and the protein structure 

 determined (see Chapter X). It is considered that most if not all of the 

 changes are due to the action of nitrous acid on adenine and cytosine, 

 which yield guanine and uracil, respectively, in the new RNA. The 

 amino acid changes being known, it is possible to compare these with 

 the change in the code words derived from cell-free systems. Tiie data 

 have been compiled from the litei-ature in both laboratories and the 

 agreement is good. Speyer et ol. (1962) record that 11 of 16 replace- 

 ments agree. For example, a common mutation results in threonine 



> isoleucine. The codes are UAC > UAU, which is a C to U 



change and agrees with the action known for nitrous acid. Important 



exceptions are the aspartic acid > serine, and serine > leucine 



replacements, which involve either change of two bases or an A > U 



change based on UoC for serine. However, if a second code word for 

 serine were UCG, some of these discrepancies would disappear. Matthaei 

 et al. (1962) include UCG as a second code word for serine, but their 

 published data do not lend strong support to this assignment. In general, 

 then, these results support the proposed code words, since the combined 

 effect of the known action of nitrous acid and the proposed code words 

 are so restrictive to the possible amino acid changes that the agreement 

 found must have significance. 



This is reinforced by the conclusions reached by Smith (1962) in a 

 study of amino acid replacements in human hemoglobins. The question 

 studied is whether known amino acid changes in the abnormal hemo- 

 globins are compatible with the change of a single base in the synthetic 

 code words. In this case there are no restrictions to the kind of base 

 change involved and in fact almost every type must occur to account 

 for the amino acid changes. Smith predicted the code words for four 

 amino acids based on known ones. For example, glutamic acid had not 

 yet been coded; in fact, the data with synthetic polynucleotides is still 

 weak for this amino acid (Speyer et al., 1962). But since glutamic acid 

 could be replaced !)y valine (UoCi), lysine (UA_,), and glycine (UGo) 

 in various hemoglobins, the code word for glutamic acid must contain 



