72 ERWIN CHARGAFF 



nucleic acids must be imprinted on the sequential arrangement of the 

 monomeric constituents and prompted us to search for ways to approach 

 this difficult task. 



The general problem is similar to that of the structure of proteins, but 

 immensely more forbidding. The very great length of the chains composed 

 of a much smaller number of different monomers raises obstacles that 

 render improbable an unambiguous solution by existing methods. More- 

 over, the available procedures are not yet as refined as in the case of the 

 proteins : few specific enzymes, no generally applicable method for marking 

 the end groups. On the other hand, the remarkable difference in the stability 



V 



fjy, hi.,, .TV-, 



.0.^ ^ N 



)'\ 



;a"!ir'%;ii-<^'i4- 





"..: 



Fig. I . Schematic representation of a fragment of a double strand of a deoxy- 

 ribonucleic acid. The purines (A, adenine; G, guanine) are depicted in black, the 

 pyrimidines (C, cytosine; T, thymine) in white. 



of the glycosidic bonds holding the purines and the pyrimidines offered a 

 novel possibility that we have exploited fully. 



Our first attempts in the direction of utilizing a specific enzyme were 

 concerned with the problem of a recognizable repeating unit. 



STEPWISE DEGRADATION BY DEOXYRIBONUCLEASE 



When crystalline pancreatic deoxyribonuclease acts on a highly poly- 

 merized preparation of a deoxyribonucleic acid, it is possible to conduct 

 the experiment in such a manner as to collect the products detached 

 gradually in a dialyzable form and to separate them from the enzyme- 

 resistant core [7, 25]. The latter is characteristically different in composition 



