II. REPLICATION OF DNA IN CHROMOSOMES 77 



on the segregation of tritium-labeled DNA in whole cells of Eschenchia 

 coli. Cells of E. coll 15T" were exposed to thymidine-H-^ of high specific 

 activity for less than one division time and then removed from the label 

 and allowed to divide a number of times. The distribution of labeled 

 DNA was very heterogeneous in clones derived from single young cells. 

 The pattern of segregation indicated that two large DNA-containing 

 units regularly separate at cell division. These units may then be per- 

 petuated intact for many successive cell divisions, but are subject to 

 fragmentation at a finite probability. These observations are consistent 

 with the view that the two subunits represent the single chromosome, 

 which replicates in a semi-conservative fashion, but is subject to an 

 occasional sister chromatid exchange as observed in higher forms. 



These results certainly rule out any dispersive mechanism of replica- 

 tion, but they cannot be taken as definitive proof of a semi-conservative 

 scheme at the molecular level. Because of the uncertainties concerning 

 the structure and organization of chromosomes it was conceivable, 

 although highly unlikely, that the two subunits of DNA in a chromo- 

 some represented a higher level of organization than that of the molecule 

 of DNA. 



B. MOLECULAR LEVEL 



The first evidence that replication might be semi-conservative at the 

 molecular level was provided by Levinthal (1956). In an autoradio- 

 graphic study of the distribution of phosphorus-32-labeled DNA from 

 phage T2, he obtained evidence that about 40% of the DNA, which 

 remained in one piece when the phage was broken by osmotic shock, 

 was replicated semi-conservatively. The particles obtained from cells 

 infected with P^--labeled phage yielded DNA fragments after similar 

 treatment with one-half as much P^- as the original phage. Doubts 

 concerning the validity of the results were raised when other methods of 

 analysis failed to reveal these large pieces (see Chapter III ; also Davison 

 and Levinthal, 1961). It now appears that the large pieces were probably 

 some sort of artifact and therefore the significance of the original experi- 

 ments remains unclear. However, more convincing results at the molec- 

 ular level were soon presented. 



Meselson et al. (1957) published a technique for separating DNA 

 molecules of different densities in a cesium chloride gradient at high 

 centrifugal force. They soon applied the technique to a study of the 

 distribution of labeled DNA during replication in Escherichia coli 

 (Meselson and Stahl, 1958). The cells were grown for a number of 

 generations in N^^-labeled medium. The heavy DNA of these cells could 

 be banded separately from the regular N'^-containing DNA (Fig. 3). 



