58 R. MAEKHAM 



Methods for studying such molecules by means of enzymes are listed in the 

 above reference (Heppel et al., 1955). 



The chemical degradation of polynucleotides is even more difficult, but 

 may be made to yield dinucleotides (Merrifield and Woolley, 1952; Cohn and 

 Markham, 1956), but only in small yield. A promising method is that of 

 Whitfeld (1954), which is essentially a procedure for stepwise degradation. 

 Unfortunately it too has to rely on an enzpnatic step in order to produce a 

 polynucleotide chain ending in a nucleoside residue with free — OH groups 

 at C-2' and C-3'. The enzyme used, human prostate phosphatase, can be 

 prepared essentially free from detectable nucleases, but the possibility of 

 nuclease action on a long chain is not predictable with absolute certainty. 

 Having removed the terminal phosphate residue, the nucleic acid is then 

 subjected to the action of sodium periodate. This reagent, which is remarkably 

 specific, oxidizes the cis-glycol at the terminal nucleoside residue to a dialde- 

 hyde. Such a compound has a — C=0 group in the jS-position in relation to 

 the phosphate ester linkage at C-5', and the latter is thus rendered very 

 labile. On exposure to a pH of 10, the phosphate group (and the rest of the 

 chain) undergoes ehmination. The net effect of this reaction is that the residue 

 of the terminal nucleotide drops off the chain, and leaves a chain reduced 

 in length at one end by one nucleotide residue. The phosphatase treatment 

 may then be repeated, followed by the oxidation, and so on. This procedure 

 has been followed quite successfully on small polynucleotides. Of course, 

 for a ribonucleic acid with a chain length comparable with that of, say, the 

 tobacco mosaic virus nucleic acid, an absolute minimum of 100 mg. of material 

 would be needed for the procedure, and the manipulations become exceedingly 

 complex. In addition the tobacco mosaic virus nucleic acid has sufficient 

 plant ribonuclease attached to it to break the chain spontaneously, so that 

 the results to be expected are not too hopeful. 



Another method for determining the ends of ribonucleic acid chains is 

 dependent upon the specificity of pancreatic ribonuclease. Because of this 

 specificity any purine nucleotides, either 2', 3', or 2' : 3' cychc ones, which 

 may be liberated, or any small polynucleotide terminated by such a group 

 must have been at one end of the chain, which, of course, is the same end 

 as that detached by Whitfeld's method. These groups are, in fact, readily 

 recognized (Markham and J. D. Smith, 1952c), but their quantity and their 

 distribution are such that they must be artifacts, or else the ends of the 

 nucleic acids must be extremely varied. The former would appear to be the 

 more likely. 



A modification of the above treatment is removal of the terminal phosphate 

 groupings by means of phosphatase, followed by alkaline hydrolysis. Any 

 nucleoside residues liberated must represent chain ends. 



One method which is free from the objections mentioned above is that due 



