202 S. S. COHEN 



In addition, if there were such a simple relation, it might be imagined that 

 the KNA or even ribonucleoprotein of a cell elaborating protems containing 

 an abnormal concentration of some amino acids would display related abnor- 

 mahties in nucleic acid composition. However, the silk gland of the silk moth 

 Philosamia ricini contains proteins having about 45 % glycme. The RNA of 

 this gland possessed a base composition very much like the RNA of the gut 

 of the same animal, which had very different types of protein (Matthews, 

 1957). Similarly, although the RNA of turnip yeUow mosaic virus contains 

 37 % cytidyHc acid, the protein of this virus did not possess a particularly 

 abnormal gross amino acid composition. 



Although the ribonucleoprotem of the silk gland of the silkworm, Bombyx 

 mori, has an amino acid composition unlike silk fibroin, which contains 

 about 80 % of glycine, alanine, and serine, these amino acids are, neverthe- 

 less, iV-terminal in the nucleoprotein (Shimura et al., 1956). It is conceivable 

 that only a part of the ribonucleoprotein isolated may act as template for the 

 fibroin molecvde and that one should not expect to detect a stoichiometric 

 relationship between the final, dissociated, accumulated product and the 

 working sequence of nucleotides which may exist as only a small part of a 

 template. 



As discussed by Gamow et al. (1955), more than one nucleotide must be 

 used to code each amino acid, since there are only 4 nucleotides for 20 amino 

 acids. A variety of triplet codes have been devised in which a set of three 

 nucleotides determines a single amino acid, but a number of these were easily 

 excluded (Gamow et al., 1955). Brenner (1957) has now demonstrated that 64 

 overlapping triplets, made possible by the choices among 4 nucleotides taken 

 3 times, are insufficient to code the known amino acid sequences. This result 

 also appears to exclude the concept of a 1 : 1 relation of amino acid to nucleo- 

 tide along a single duplex of one chain each of polynucleotide and polypep- 

 tide, even as suggested by the analytical results. 



A number of other approaches to the problem have been set forth. In one 

 complex projjosal, the use of nonoverlapping triplets of nucleotides as dif- 

 fusible trinucleotides is suggested to code for and be associated with a single 

 amino acid, the entire sterically specific complex of which might then attach 

 to an RNA template, polymerize ammo acids, and dissociate polypeptide 

 (Crick et al., 1957). There is no a priori reason at present to exclude complex 

 mechanisms on this question, Occam's razor notwithstanding. Certainly the 

 discovery of the role of soluble RNA in transferring amino acids to the micro- 

 somal fraction appears consistent with this proposal. 



VIII. Conclusion 



This survey, as fragmentary and lacking in depth as it may be, provides 

 the biochemical framework within and on which the modern virologist 



