430 D. M. BROWN AND A. R. TODD 



characterized several di nucleotides from acid hydrolysates of yeast ribonu- 

 cleic acid, is of considerable interest. For structural determination they 

 used methods similar to those just described except that intestinal phos- 

 phatase, inhibited with arsenate, was used instead of snake venom diesterase 

 to demonstrate the C 5 /-linkages. Since the formation of their dinucleotides 

 depended on a random fission of the polynucleotide chain, they obtained 

 products containing purine and pyrimidine residues isomeric with those 

 found in ribonuclease digests, but in which the terminal phosphoryl group 

 was attached to the purine nucleoside residue and hence could be further 

 degraded by ribonuclease. Moreover, they were able to separate dinucleo- 

 tides which were isomeric with one another by virtue of the position (a or 

 h) of the terminal phosphoryl group, in accordance with prediction from 

 the postulated mechanism of chemical hydrolysis, ^^ in which phosphoryl 

 migration must always accompany fission of the diester linkages between 

 individual nucleoside residues in the ribonucleic acid molecule. 



7. Evidence for 3'(6)-Linkages in Ribonucleic Acids 



In discussing the evidence so far presented on ribonuclease action, the 

 3',5'-linked polynucleotide structure has been assumed. However, since 

 degradation with ribonuclease proceeds via intermediates with cyclic 2',3'- 

 phosphate groupings it is clear that, as in the case of chemical hydrolytic 

 studies, no differentiation can be made between C2' and Ca- as a linkage 

 point in the original ribonucleic acids. These could be either 2', 5'- or 3',5'- 

 linked structures, or might contain both types of linkage. Clearly, finality 

 regarding the linkage requires a decision between C2' and Cs-. 



A solution to this problem was found in the study of the action of 

 enzymes on simple esters of the mononucleotides. Brown and Todd®' ex- 

 amined the action of ribonuclease on the benzyl esters of the a and b iso- 

 mers of the pyrimidine nucleotides. Cytidine benzyl phosphate b and uri- 

 dine benzyl phosphate b were converted by ribonuclease into cytidylic acid 

 b (cytidine-3 '-phosphate) and uridylic acid b (uridine-3 '-phosphate), hy- 

 drolysis proceeding in each case by way of the intermediate cyclic 2',3'- 

 phosphates. The benzyl esters employed in these experiments were oriented 

 by catalytic hydrogenation to the parent nucleotides, a process unlikely to 

 involve phosphoryl migration. The methyl and ethyl esters were similarly 

 hydrolyzed by ribonuclease. The isomeric esters of the pyrimidine a nu- 

 cleotides were completely unaffected, nor had ribonuclease any action on 

 esters of either a or 6 isomers of the purine nucleotides. Not only does this 

 further justify the belief in specificity of ribonuclease for pyrimidine nu- 

 cleotide ester linkages in ribonucleic acids, but it leads to important con- 

 clusions about the linkage position. Since, as shown above, all the inter- 

 nucleotidic linkages involving pyrimidine nucleotide residues are apparently 



