514 A. TSUGITA AND H. KRAHNKEL-CONRAT 



It is evident that this scheme sets strict liiiiilalioiis on the possible 

 results of deaniination. Thus, certain amino acids might be replaced by 

 three others but none by more. Any amino acid ever observed to result 

 from an exchange could only be exchanged to two or one other. Finally, 

 no amino acid could result from more than three others, and those in this 

 latter group would never be rejjlaced. Wittmann has relied on this 

 scheme in the interpretations of his lesults, l)iit he finds some exchanges 

 that do not fit, and we have observed others. 



As previously discussed, we believe that in UNA, as contrasted to 

 DNA, the deaniination of cytosine is the greatly predominating muta- 

 genic event. Only much more rarely might the deaniination of A cause 

 an A— >G shift, and other base exchanges may arise with even lower 

 frequency during replication of the deaminated as well as the brominated 

 or methylated KNA, and attributable to random mistakes in replication 

 in a manner similar to natural mutation. Thus, the octet scheme ajijiears 

 too rigid to allow for all possible events, and not truly expressive of the 

 predominant role of the C-»U change in RNA deaniination. 



Focusing our attention on this predominant result of deaniination, 

 we have arranged our observed amino acid exchanges schematically in 

 groups headed by triplets of decreasing cytosine content, allocating 

 amino acids to triplets on the basis of the recent coding findings discussed 

 above (see Fig. 8). The implication in such an evaluation of our exchange 

 data is that they are the consequence of single step mutations, which is 

 probably the case for most single exchanges, but can in no individual 

 case be stated with assurance. 



Further, one must consider whether the coding triplets as observed 

 with E. coll would be of the same "language" as those in tobacco plant 

 cells, since there exists the possibility that different codes might pertain 

 in different organisms. This question appears to be answered by the 

 results reported in a preceding section where it was show'n that TAIV- 

 RNA acted as messenger for the production of material resembling the 

 TMV protein by the same E. coli system, which has been used to obtain 

 the coding data with synthetic polynucleotides. Thus, the coding 

 language, whether it is universal or not, appears to be very similar for 

 E. coli and tobacco, and the correlations shown in Fig. 8 are thus 

 justified. 



The data presented make it appear probable that proline, threonine 

 and aspartic acid or its amide are coded as C2U; leucine, serine, and 

 alanine as CU2; and phenylalanine as U.-,. These hypotheses agree in part 

 with the results obtained by one and/or the other group studying the 

 amino acid incorporation in the E. coli system. Cases which do not agree 

 with these data (e.g., Tyr-^Pho, with tyrosine being coded by an 



