THE BIOCHEMISTRY OF PLANT VIRUSES 57 



These observations more or less established that the phosphate ester link 

 was tlirough C-3' or C-2'. The unfortunate migration of the phosphate, 

 of course, precluded the decision whether C-2' or C-3' was the actual original 

 site of the ester linkage, but this was later clarified by Brown and Todd 

 (1953) for the pyrimidine nucleotides by using model compounds to deduce 

 the absolute substrate sj)ecificity of pancreatic ribonuclease, which they 

 found to attack pyrimidine nucleoside 3'-phosphate benzyl esters, but not 

 the analogous 2'-compoimds. As far as the purine nucleotides were con- 

 cerned, they were shown to have phosphate at C-3' by two independent 

 methods. Whitfeld and Markham (1953) used a chemical degradation 

 method on dinucleotides, and prepared the purine nucleoside 3'-phosphates. 

 Heppel and associates (1953) degraded dinucleotides using a diesterase from 

 spleen, which does not give a 2' : 3'-cyclic nucleotide intermediate, and they 

 also obtained nucleoside 3'-phosphates from polynucleotides by this route. 



The other end of the linkage was determined in several ways. Cohn and 

 Volkin (1952) obtained nucleoside 5'-phosphates from ribonucleic acids by 

 enzymatic degradation. This method is now in general use for analytical 

 purposes. Another piece of evidence was produced by Markham and J. D. 

 Smith (1951a, 1952b), who showed that some dinucleotides, produced by 

 ribonuclease degradation from virus ribonucleic acids, had a terminal 

 nucleoside 2' : 3'-cyclic phosphate residue, and hence, by elimination, had 

 to have their C-5' involved in the internucleotide linkages. 



The further structural study of virus ribonucleic acids has been based 

 very largely on the use of pancreatic ribonuclease. This enzyme is specific 

 for pyrimidine nucleoside 3'-phosphate esters, and so will break the poly- 

 nucleotide chain in places where these compomids are present. In conse- 

 quence, one gets a series of products having the general formula: n purine 

 nucleotides -f 1 pyrimidine nucleotide. The factor n may have any value 

 from upward. Markham and J. D. Smith (1952b) isolated and identified a 

 large number of such products from digests of the turnip yellow mosaic 

 \nriis, some of them incomplete digestion products. But, as has been shown 

 by Heppel et at. (1955), the enzyme has considerable synthetic potentialities, 

 so that data obtained from partial degradation are not unequivocal. Provided, 

 however, that the system is free from other enzymes, the final products 

 are unique. Unfortunately it is by no means certain that any nucleic acid 

 has yet been obtained completely free from degradative enzymes, and so 

 the results of enzymatic degradation must be regarded with a little reserve. 

 There is, of course, not much doubt that the bulk of the data is in fact 

 relevant, but with very large molecules a little impurity may give rise to 

 artifacts, which may be difficult to detect. 



Most other enzymes have even less specificity than pancreatic ribo- 

 nuclease, and are only really suited to the attack on small polynucleotides. 



