31. SYNTHESIS OF POLYNUCLEOTIDES 139 



inhibited by 50 % by the pyrophosphate. Under these conditions when la- 

 beled pyrophosphate is used, its incorporation into the terminal pyrophos- 

 phate groups of the four nucleosides-triphosphates is found. The rate of 

 this exchange reaction is comparable to the synthetic rate in the absence of 

 the added pyrophosphate. Like the synthetic reaction, the exchange reac- 

 tion is dependent upon the presence of DXA. DXA alone in the presence 

 of an excess of inorganic pyrophosphate and the polymerase undergoes only 

 a small amount of pyrophosphorolytic cleavage. 



There is a striking difference between the synthetic and pyrophosphoro- 

 lytic reaction in respect of the effect of omitting one or more of the deoxy- 

 ribonucleoside-5'-triphosphates from the reaction mixture. Whereas syn- 

 thesis is reduced to a fraction of a per cent (see below), the reversal is 

 affected relatively little. In other words, appreciable exchange of the pyro- 

 phosphate grouping occurs when only one of the nucleoside-5'-triphosphates 

 is present. 



The above findings have been interpreted to mean 86 that whereas pyro- 

 phosphorolysis of a long tightly coiled hydrogen-bonded molecule is slow 

 and difficult, single or a few nucleotide units freshly added to the preexist- 

 ing chain are readily pyrophosphorolyzed. The stabilization effect must be 

 a cumulative one since it would appear that the selection of the incoming 

 nucleotide is governed by prior hydrogen bonding (see below). 



d. The Base Composition of the Product as a Function of the DNA Used as 

 Primer^ 



The extended studies of Chargaff and co-workers (Volume I, Chapter 

 10) on the analytical composition of samples of DNA from a great variety 

 of sources showed the remarkable generalization that purine content always 

 equals the pyrimidine content. Although among the purines, the adenine 

 content may differ greatly from that of guanine and similarly, among the 

 pyrimidines, thymine from cytosine, the total equivalence (adenine + 

 guanine = cytosine + thymine) is maintained by the fact that adenine 

 content equals that of thymine and, again, guanine corresponds in amount 

 to cytosine. The base equivalence provided originally the strongest chemical 

 support for the Watson-Crick 84 postulate for DXA structure. 



Using DXA of different origins (bacterial, thymus, T2 phage) as primers 

 in their system, Romberg and co-workers have found that the base com- 

 position of the enzymically synthesized DXA always resembles that of the 

 primer used. Thus, in the synthetic products, adenine and thymine were 

 present in approximately equal amounts and similarly the guanine content 



90 I. R. Lehman, S. B. Zimmerman, J. Adler, M. J. Bessman, E. S. Simms, and A. 

 Romberg, Proc. Natl. Acad. Sci. U. S. 44, 1191 (1958). 



