RIBONUCLEIC ACIDS 



position, with rupture of the 5' Hnkage and thus of the chain. 

 The cycHc esters would subsequently be hydrolyzed randomly 

 to the mixture of the 2' and 3' nucleotides already found. 



Although this mechanism was inherently logical, it rested on 

 certain assumptions which it became necessary to verify (as the 

 assumptions of the older hypothesis had not been). The most 

 important of these was that a cyclic 2':3' intermediate existed. 

 However, this postulate was promptly verified by synthesis of 

 these compounds (9), by isolation of them from partial hydroly- 

 zates (46,47), and by the testing of both synthetic and isolated 

 phosphodiesters. All this work appeared within the span of one 

 year (1951-1952) and, together with the following correlations, 

 completely established the 3 ',5' (or 2 ',5') hypothesis. 



The newer structure, to begin with, needs no assumptions 

 of preferential alkali lability of one phosphate bond over the 

 other. Preferential splitting of the 5' linkage rests on the 

 demonstration of the cyclic 2':3' intermediate. The possibility 

 of forming such intermediates simultaneously and randomly 

 throughout the polynucleotide chain gives a logical basis to the 

 observations that RNA is almost completely reduced to nucleo- 

 tides by mild alkali treatment, whereas DNA, with the same 

 3 ',5' backbone but no 2' hydroxyl, is completely resistant to 

 mild alkali. Preferential splitting of the 3 ' (or 2 ') link by enzymes 

 to yield 5' nucleotides is demonstrably caused by the specificity 

 of the diesterase enzymes used which also produce 5' deoxy- 

 nucleotides from DNA (14,34). The presumed specificity of 

 action has been further verified by the use of synthetic and iso- 

 lated smaller substrate molecules (22,49). The older observa- 

 tion that RNA polynucleotides reduce periodate only after 

 dephosphorylation could now easily be explained as brought 

 about by the removal of end 3' phosphates leaving terminal 

 nucleosides esterified only in the 5' position. Furthermore, the 

 production of nucleosides and 3 ',5' (or 2 ',5') diphosphonucleo- 

 sides in equal amounts in the diesterase degradation of poly- 

 nucleotide material (19,20) could be accounted for only by a 

 3 ',5' (or 2 ',5') structure. 



465 



