STRUCTURAL AND CHEMICAL ARCHITECTURE OF HOST CELLS 185 



to consist of nucleotides linked by phosphodiester bonds from the Cg-liydroxyl 

 of one nucleoside sugar residue to the Cs-hydroxyl of another nucleoside 

 sugar, as depicted in Fig. 26. 



Gross analysis of base composition, largely a result of the exploitation of 

 paper chromatography,^ has produced a number of suggestive generaliza- 

 tions for all DNA so far analyzed and some less satisfactory generalizations 

 for most RNA. In samples of DNA, bases containing 6-amino groups exist 

 in amounts equivalent to bases containing 6-hydroxyl groups; in addition, 

 purines = pyrimidines, adenine = thymine, and guanine = cytosine. These 

 results have suggested a complementarity of purine and pyrimidine bases, 

 producing interaction by hydrogen bonding along parallel nucleotide chains. 

 The model based on these equivalences was developed by Watson and Crick 

 (1953) (also see Pauling and Corey, 1956) and is presented in Fig. 27. It has 

 been amply supported by X-ray analysis of many DNA preparations, as 

 summarized by Wilkins (1956). This general structure has proved to be a 

 useful frame of reference for consideration of the properties of DNA, particu- 

 larly its denaturation and degradation (Doty, 1956). Of at least equal interest 

 has been the hypothesis that the existence of complementary chains in DNA 

 provides a mechanism for the duplication of DNA. It is supposed that the 

 separated chains can serve as templates for the organization of new comple- 

 mentary polynucleotide chains. The evidence concerning this hypothesis wiU 

 be considered in a later section. 



Base analyses for RNA samples of entire cells have also revealed an ap- 

 parent equivalence of 6-amino groups and 6-hydroxyl groups, although not 

 of individual bases (Elson and Chargaff, 1955). A number of samples of viral 

 R.NA, e.g., the UNA of turnip yeUow mosaic virus, provide exceptions to 

 even this pairing rule. X-ray crystaUograpliic study of RNA samples has not 

 been as fruitful as for DNA, although it has been suggested that the X-ray 

 patterns suggest a DNA-Hke structure for RNA (Rich and Watson, 1954; 

 Crick, 1957a). Evidence for the hydrogen bonding of polynucleotide chains of 

 RNA have stemmed in largest part as a result of studies on the biosynthesis 

 of ribose polynucleotides, to be discussed below. Despite the absence of com- 

 plementary bases in RNA for pairing in the Watson-Crick model, in which 

 two polynucleotide chains run in opposite directions, it is possible to devise 

 other paired structures, as, for example, the model of Donohue and Stent 

 (1956) in which a duplex may be built up of two polynucleotide chains of 

 identical base sequence. 



Evidently such a structure of RNA provides a mechanism for replication, 

 even as does the Watson-Crick model for DNA. That the requirement for a 



^ In recent years, paper chromatography and ion exchange techniques have revealed 

 trace elements of new bases in both RNA and DNA (Davis and Allen, 1957; Dumi and 

 Smith, 1955). 



