MOROWITZ 149 



An alternative method of investigating the ordering relations is to give up 

 dealing with pure proteins and examine mixtures of related proteins in an effort 

 to elucidate statistical relations. Thus we may inquire into the frequency dis- 

 tribution of N-terminal amino acids, C-terminal amino acids, amino acids made 

 N terminal by tryptic digestion, or any other experimentally available cut. We 

 may look at the entire collection of proteins from a single cell type, or any 

 subgrouping of proteins definable by some experimental procedure. It must be 

 kept in mind that we are seeking general laws for given collections of proteins. 



The experimental procedures are derivable from the techniques of sequence 

 analysis [4]. In particular, use is made of the fluorodinitrobenzene reaction and 

 the Edman degradation [5]. The application of these techniques must be modi- 

 fied, as we are dealing with a collection of proteins rather than a single protein. 



Since detailed experimental protocols are being published elsewhere, we shall 

 only outline the method here [6]. E. coli cells are grown on a defined medium 

 and randomly labeled with a single radioactive amino acid (leucine) or in the 

 case of sulfur 35 label they are randomly labeled in three amino acids, cysteine, 

 cystine, and methionine. The cells are then washed, and extracted with ethanol, 

 ether, and hot trichloroacetic acid. The residual material containing most of 

 the cell protein and little else is subjected to oxidation with performic acid. 

 One portion of this material is subjected to fluorodinitrobenzene end group 

 analysis or conversion to the phenylthiohydantoin derivatives; a second portion 

 is digested to completion with trypsin and then subjected to the same end 

 group analysis. Quantitative end group analysis is carried out by counting the 

 spots on paper chromatograms that result from N-terminal amino acids and 

 interior amino acids. 



We thus determine the frequency of occurrence of cysteine (including 

 cystine), methionine, and leucine in their positions, N terminal and following 

 the tryptic cut (which has a high degree of specificity for the carboxyl end of 

 arginine and lysine). The results of these experiments are shown in table 1. 



To interpret these results, let us first consider what results would have been 

 obtained if there were no ordering rules. That is, suppose that each protein 

 sequence was completely independent and any ordering was equally probable. 

 If there were a large number of different protein species, on a statistical basis 

 the amino acids would be expected to be randomly ordered. On the basis of 



