92 



shown that the synthesis of certain peptide antibiotics of well- 

 defined sequence is brought about by a group of specific proteins 

 (enzymes) without the help of nucleic acids. However, the peptide 

 antibiotics that are synthesized are relatively simple and molecules 

 of this size would not be capable of catalyzing the synthesis of com- 

 plex specific proteins. Certainly, the enzymes involved are very com- 

 plex and are themselves synthesized with the help of nucleic acids. 

 We do not think, therefore, that this system is a good model for 

 protein self-replication. It is dangerous to be dogmatic about general 

 schemes of self-replicating peptide cycles, but we suspect that they 

 all suffer from the same problem: simple peptides lack sufficient 

 specificity, while large peptides are too hard to make so that it would 

 be impossible to close the cycle. Furthermore, even if one self- 

 replicating cycle existed, it is hard to see how it could evolve to 

 greater complexity. 



Nucleic acids first- There are a number of theories in which 

 polynucleotide replication is proposed to have preceded the synthesis 

 of ordered polypeptides. Proponents of such theories emphasize that 

 nonenzymatic complementary replication of polynucleotides seems 

 plausible in light of the known interactions between nucleotide 

 bases. We shall see that there is already a substantial body of experi- 

 ments supporting the idea that a preformed polynucleotide can 

 direct the synthesis of a complementary oligonucleotide according to 

 the previously described Watson-Crick pairing rules. The required 

 "preformed" nucleic acid may itself have arisen originally in a non- 

 ordered joining reaction of mononucleotides; such reactions are 

 already known. The polymerization of nucleotides, just as in the 

 related reaction of amino acids to form a polypeptide, effectively 

 requires the removal of one molecule of water for each addition of a 

 nucleotide to the growing chain. Therefore, it is not surprising to 

 find that successful polymerizations have usually employed drying 

 conditions, the addition of "condensing" reagents, or the removal of 

 water in a prior step to form an activated monomelic nucleotide. The 

 addition of a catalyst, such as a metal ion, can increase the yield of 

 polynucleotides quite appreciably. It has been shown that divalent 

 metal ions enhance the formation of oligoadenylates from an 

 activated derivative of adenylic acid. In the case of lead ion a 56% 

 yield of oligomer was formed, while in the absence of this metal ion 



