286 



CHAP II R 21 



produces deoxj riboside 3'-monophosphates 



In breaking the chain at 5'. Consequently, 

 the labeled phosphate ( P* ) is found joined 

 at the }' position of the deoxyriboside just 

 anterior to the one with which it entered 

 the ON. A strand. The digest is then an- 

 alyzed to see how frequentl) P* is part of 

 </ A y-P*. T 3'-P*, dC 3'-P*, and dG 3'-P*. 

 If the P* were originally in </AP*PP. we 

 would then know the relative linear fre- 

 quencies of TA, AA, CA. and GA. By 

 carrying out this procedure three more times, 

 labeling a different one of the triphosphates 

 each time, the relative frequency of all six- 

 teen sequences can be determined. 



Such nearest-neighbor analyses have been 

 made of the DNAs synthesized using a num- 

 ber of different preformed DNAs. As al- 

 ready mentioned, the DNA isolated from 

 particles of <£X174 is single-stranded; chem- 

 ical analysis reveals its base content to be 

 A = .246; T = .328; C = .185; and G = .242. 

 It synthesis requires the formation of com- 

 plementary base pairs, an extended reaction 

 carried out with <£X174 DNA and all four 

 triphosphates labeled and stopped after 20% 

 synthesis has occurred, should produce com- 

 plementary, labeled DNA composed of 

 .328 A; .246 T; .242 C; and .185 G. The 

 values found are exactly those expected. If 

 both old and new strands are used as tem- 

 plates, one expects and finds after a 600% 

 synthesis that A = T and C = G. More- 

 over. A and T are expected to have fre- 

 quencies of M.328 -f .246), or .287, with 

 C and G = .224. Again experimentation 

 confirms expectation. 



One can perform the 20% synthesis on 

 four occasions, each time labeling a different 

 triphosphate. In this way, nearest-neighbor 

 analyses are made, and the relative frequen- 

 cies of all sixteen dinucleotide sequences de- 

 termined. All sequences are found and in- 

 clude, for instance. .054 GA; .064 TC; .052 

 CT; and .069 AG. Nearest-neighbor anal- 

 yses can also be made after 600% synthesis. 



II the complementary strand is synthesized 

 in the same direction as the template strand, 

 from the results of the 20 r ; syntheses one 

 expects that GA CT ' .(.054 + .052 ) 



.053; and TC AG = .067. If. on the 

 other hand, the two complementary strands 

 are synthesized in opposite directions, the 

 expectations are GA = TC ' •_. (.054 + 

 .064) = .059; and CT = AG = .061. The 

 values observed (.058 GA; .065 TC; .064 

 CT; and .058 AG) clearly follow those ex- 

 pected for complementary strands synthe- 

 sized in vitro in opposite directions, as do 

 the values obtained for the other dinucleo- 

 tide sequences. 



Chemical analysis of Mycobacterium phlei 

 DNA yields .162 A; .165 T; .335 C; and 

 .338 G. If extended syntheses permitting 

 nearest-neighbor analyses are performed, one 

 can determine the relative amount of base X 

 incorporated into DNA from the sum of the 

 separate frequencies with which XA. XT, 

 XC, and XG occur. When X is, in turn, 

 A, T. C, and G, the relative frequencies are 

 found to be .162, .164, .337. .337. respec- 

 tively. Thus, what is already demonstrated 

 via chemical analyses is independently 

 proved via nearest-neighbor analysis — 

 namely, that the product of an extended 

 synthesis has the same base frequencies as 

 the natural two-stranded DNA used as 

 primer-template. 



The question of whether or not the se- 

 quences of bases along a strand is random 2 

 can be decided by using calf thymus DNA 

 as primer-template and determining all the 

 dinucleotide frequencies in the product. 

 Among the frequencies observed, CG is .016 

 and GC. .044. Had these two dinucleo- 

 tides occurred with equal frequency, the 

 hypothesis of a random nucleotide sequence 

 in a strand would have been supported. 

 Since the frequencies are clearly different, 

 however, base sequence in a strand is not 



-' In this connection see also J. H. Spencer and 

 E. Chargaff (1963). 



