86 ERWIN CHARGAFF 



trend of the distribution of 5-methylcytosine in consecutive fractions 

 obtained by the fractionation of deoxyribonucleic acid from calf thymus, 

 wheat germ, and rye germ. In all instances, the relative abundance of 

 5-methylcytosine with respect to cytosine varied in such a manner that 

 the percentage contribution of the minor pyrimidine was higher in the 

 earlier fractions. In no case could one have claimed that the two 6-amino 

 pyrimidines replaced each other at random within the polynucleotide 

 chain. I have discussed, in more detail, the significance of the finding of a 

 disproportionate distribution of cytosine and 5-methylcytosine in view 

 of current replication hypotheses at a previous occasion [10]. 



An even more impressive diflference between the ways in which 

 methylcytosine and cytosine are selected for insertion into the poly- 

 nucleotide emerges from the study of the frequency of these pyrimidines 

 as solitary and bunched nucleotides [37]. As will be seen in Fig. 2, a 

 remarkably large proportion of methylcytosine, and a surprisingly small 

 proportion of cytosine, exist as solitary units. These two 6-aminopyrimi- 

 dines can hardly share a common pattern of nucleotide sequence. What is, 

 however, remarkable is that the mole fractions of 5-methylcytosine and 

 thymine appearing as solitary units are nearly equal in most instances. One 

 must conclude that, whereas there is no evidence that the two 6-amino 

 pyrimidines are treated indiscriminately by the selection mechanism, there 

 is some indication that no such distinction is exercised in regard to the two 

 5-methylpyrimidines. In the terms defined above, cytosine and 5-methyl- 

 cytosine are pleromers, but not homotopes; thymine and 5-methylcytosine 

 appear to be homotopic with respect to the sequence purine-pyrimidine- 

 purine, although they are not pleromeric. 



SELECTION OF A FRAUDULENT ANALOGUE 



In the instance under discussion, the incorporation of 5-bromouracil 

 into the deoxyribonucleic acid of E. coli, the problem of replacement can 

 be posed directly. In the case of a natural satellite, such as 5-methylcytosine, 

 the complement appears to be fixed rigidly by the cell, whereas no obvious 

 cellular mechanism, except the death of the cell, can be seen that would 

 limit the extent of incorporation of the halopyrimidine which the cell 

 presumably cannot but treat as an impostor. If the cellular apparatus is 

 inveigled into treating the fraudulent 6-ketopyrimidine as if it were 

 thymine, does it also insert it as such indiscriminately into the newly 

 forming deoxyribonucleic acid ? This does not appear to be the case. A 

 glance at Table VI will show that the deoxyribonucleic acids made by the 

 cell in the absence and in the presence of bromouracil are vastly diflFerent 

 with regard to their sequence characteristics. Not only is there no sign of 

 an equal relative utilization of thymine and bromouracil for common 



