Replication of DNA in Vitro 



287 



random. The nonrandomness of base se- 

 quences is also supported by experimental 

 results which reveal that: 70% of the bases 

 are distributed so that three or more pyrim- 

 idines (and hence purines) occur in succes- 

 sive linear nucleotides; linear sequences of 

 five successive T's exist; and 4.9% of a 

 given DNA contains sequences of eight or 

 more pyrimidines in succession. 



Is the nearest-neighbor frequency the same 

 when native DNA is used as primer-template 

 as when DNA synthesized from this native 

 DNA is used? With synthesized calf thy- 

 mus DNA as primer-template, the nearest- 

 neighbor frequencies of the newly synthe- 

 sized product are essentially the same (for 

 example, CG is .011 while GC is .042) as 

 they are in the product formed using calf 

 thymus DNA as primer-template (above). 

 Consequently, as revealed by nearest-neigh- 

 bor analysis, the products of synthesis are 

 identical when native DNA and when DNA 

 synthesized from native DNA serve as 

 primer-template. 



A + T 

 The — — L — ratio is 1.25 for calf thymus 

 C + G y 



DNA and 1.29 for B. subtilis. Even though 

 these base ratios are very similar, it is un- 

 likely that one of these is a molecular poly- 

 ploid of the other (see p. 265 ). In fact, the 

 dinucleotide sequences determined for DNA 

 synthesized from the bacterial DNA (for ex- 

 ample, CG is .050 and GC, .061) are quite 

 different from those of calf thymus (.016 

 and .044, respectively). Other work gen- 

 erally shows that in higher plants and ani- 

 mals CG is lower than expected on a random 

 nucleotide sequence, whereas the reverse 

 occurs in bacteria. It should also be noted 

 that different normal or neoplastic tissues of 

 the same individual give nucleotide neighbor 

 frequencies which are not demonstrably dif- 

 ferent. We conclude, therefore, that each 

 type of natural DNA has unique and repro- 

 ducible dinucleotide sequences, not pre- 

 dictable from its base composition. 



De Novo Synthesis of DNA 



The double-stranded polymer of A and T 

 appearing de novo, referred to earlier, can 

 be used as primer-template to study its di- 

 nucleotide sequences. Only the AT and TA 

 sequences are found; thus, it appears that 

 A and T occur in perfect alternation in a 

 strand, forming what is called a copolymer 

 of AT, or dAT(d-AT). The de novo syn- 

 thesis of a dAT, sufficiently long to serve 

 as template for extended synthesis, requires 

 a lag period of a few hours, during which 

 </APPP, TPPP, Mg++, and E. coli DNA 

 polymerase are incubated together. It has 

 been hypothesized that during the lag period, 

 the E. coli DNA polymerase catalyzes the 

 de novo formation of single strands from 

 single deoxyriboside 5'-triphosphates. Once 

 the strand is started, the short strand or 

 oligodeoxyribotide can serve as a primer to 

 lengthen itself. In the presence of d/KPPP 

 or TPPP alone, dAT adds one or two units 

 of <r/AP or TP per chain in a limited reac- 

 tion. Neither dCP nor dGP is incorporated 

 into dAT when dCPPP and or dGPPP are 

 added to the substrate, whether or not 

 JAPPP and TPPP are also present. It is 

 clear, therefore, that both limited and ex- 

 tended synthesis of dAT require base-pair- 

 ing. It is possible that a limited reaction 

 involves a single strand of dAT which is 

 folded so as to base-pair with itself except 

 at the nucleotide end. So, once the dAT 

 strand is long enough, the polymerase can 

 use it as a template for base-pairing syn- 

 thesis. Note that no lag period occurs in 

 limited or extended base-pairing syntheses. 

 In the absence of pre-existing DNA and 

 alter a lag period. E. coli polymerase — in 

 the presence of Mg+ + and high concentra- 

 tions of dCPPP and dGPPP — catalyzes the 

 de novo formation of another double- 

 stranded polymer containing only C and G. 

 Nearest-neighbor analysis shows only two 

 dinucleotide sequences, CC and GG. 

 Clearly, this polymer, called dGdC (or 



