128 CHARLES A. TIKiMAS, JR. 



laiiolod T2 phajz;e witli tritiiiin and then mixed with a hirge excess of 

 unhibeled carrier phage to prevent breakage l)y shear (hiring extraction. 

 When this 1)X.\ is streaketl on a ghiss slide and then covered witii 

 stripping fihn. the localized grains produced by the weak tritium decays 

 are seen to lie in more or less straigiit lines, the lengths of which are 

 approximately ")4 /i (Fig. 5). 



Taken togi'ther. these studies indicate tliat the DXA molecule 

 liberated from the jihages so far studied haw the length and mass that 

 woulil be expected for a single helical duplex. 



4. The Frohlcni of Wccik Points or Polynucleotide Chdin Interruptions 



Ila\-ing considered the evidence that T2 and T4 ])hage DXA is a 

 single linear duplex, one would now like to know if there are any special 

 non-nucleotide residues which unite distinct polynucleotide chains. As 

 mentioned above, stability tests and the kinetics of shear breakage seem 

 to indicate no i)referential weak points along the molecule. Therefore, if 

 non-nucleotide linkages exist, they must be at least as strong as the 

 inter-nucleotide linkages themselves. But what about interruptions in 

 the polynucleotide chain; could there be a few breaks located at special 

 or random points in the duplex? Some information on this point has been 

 obtained by sedimentation equilibrium studies done on the unbroken 

 molecules, and the polynucleotide chains produced by the denaturation 

 of the unbroken molecule (Thomas and Berns, 1961; Berns and Thomas. 

 1961). In these experiments the unbroken molecules were denatured by 

 any of three different methods, in the i)resence of 1% CHoO. The 

 formaldehyde reacts rapidly with the amino groups exposed by the 

 denaturation, and effectively prevents the reassembly of the poly- 

 nticleotide chains. This stabilization by reaction with CHoO is important 

 because the separated polynucleotide chains have a strong tendency to 

 reassociate in solution unless the amino groups are blocked. "When a 

 mixture of undenatured whole molecules and denatured whole molecules 

 is banded in CsCl density gradient (in the presence of 1% CHoO), one 

 gets two well-separated, sharp bands. From the variance of these distri- 

 butions one may calculate a molecular weight value. It turns out that the 

 calculated molecular weight of the polynucleotide chains is just one-half 

 the molecular weight of the undenatured material. This is true irrespec- 

 tive of the mode of denaturation. These results are summarized in Fig. 6. 

 This could only be the case if there were no interruptions in the poly- 

 nucleotide chains. The method should be able to detect a single interrup- 

 tion if it occurred in the middle third of the molecule. Recent findings 

 indicate that the polynucleotide chains from either end of the molecule 

 have slightly different densities — a fact which would increase the 



