CHEMICAL BONDS IN NUCLEIC ACIDS 421 



Two other structural types (XX and XXI, or structures containing the 

 features of both) would also yield mononucleotides by the hydrolytic mech- 

 anism under discussion. In each case, however, hydrolysis would necessarily 

 have to proceed stepwise by removal of mononucleotide units from that 

 end of the polynucleotide chain bearing a free hydroxyl group at C2' or Cs- . 

 Structures of type XVI, embodying a 3 ',5 '-linkage, could in theory be 

 degraded by simultaneous attack at m.any points in the chain. Relevant 

 chemical evidence, although scanty, is in favor of simultaneous rather than 

 stepwise attack.*^ (Compare, however, Magasanik and Chargaff.^') Mer- 

 rifield and WooUey*^ showed that controlled acid hydrolysis of yeast ribo- 

 nucleic acid yielded a variety of small oligonucleotides, some of which 

 they separated by ion-exchange chromatography and were able to charac- 

 terize. Production of molecules of this type shows that structural features 

 in type XVI, in which 3',5'(or 2',5')-liiikages are present, must exist in 

 ribonucleic acids as distinct from such structures as XX and XXI, where 

 Cs' is not involved in the internucleotidic linkage and where hydrolysis, 

 which requires a free hydroxyl at C2' or Cs- , can only occur at a terminal 

 linkage; in other words C5. must be involved as one of the internucleotidic 

 linkage points. The isolation of nucleoside-5 '-phosphates after enzymic 

 hydrolysis of ribonucleic acids to be discussed later also indicates that Cs- 

 is one of the major linkage points. 



The above considerations led Brown and Todd^^ to postulate the general 

 structure XVI, involving a recurring 3',5'-phosphodiester linkage for the 

 polynucleotide sequence in ribonucleic acids, as the one which accounts 

 most satisfactorily for the chemical evidence. They pointed out, however, 

 that a 2 ',5 '-linkage would be equally admissible on the known facts of 

 hydrolysis. Preference for the 3 ',5 '-structure was expressed mainly on the 

 ground of analogy with deoxyribonucleic acids where a 2',5'-linkage is 

 structurally impossible; justification for this preference has since come from 



" R. B. Merrifield and D. W. Woolley, J. Biol. Chem. 197, 521 (1952). 



