BARBARA W. LOW AND JOHN T. EDSALL 



preparations will be obtained in future, and their structures 

 determined, in steadily increasing numbers. 



We now return to the more general theme of our inquiry 

 into protein structure. The structure of a native protein may 

 be considered at three different but somewhat overlapping levels. 

 The primary level is that of structure in the sense generally 

 employed by the organic chemist; it involves knowledge of the 

 number of polypeptide chains in the protein molecule, the 

 sequence of the amino acid residues in each chain, and the nature 

 and position of the cross links which bridge different chains or 

 form loops involving two different portions of the same chain. 

 Such a structure, even if completely established, may correspond 

 to a vast number of possible spatial configurations all of which 

 are compatible with the structural formula. At the second level of 

 organization we must consider the spatial configuration of 

 individual peptide chains, or segments of chains, the bond 

 distances and bond angles, and the degree of resistance to 

 rotation around various valence bonds. The possible structures 

 include the now familiar a-helix, as well as a number of other 

 helical configurations which are theoretically possible. They 

 also include the extended "j8" configurations, of which the most 

 probable are the "pleated sheets" of Pauling and Corey. Other, 

 more irregular, configurations certainly occur also. At the 

 third level, which is closely related to the second, we consider 

 the folds and turns which may, and indeed must, occur in these 

 already coiled peptide chains in order to fit the entire molecule 

 into a region of the appropriate size and shape. We may 

 briefly consider all of these problems in turn. 



Sequences and Cross Linking in Peptide Chains 



The first level of knowledge indicated above is the problem 

 which has apparently been completely solved by Sanger and 

 his collaborators for the three insulins. More recently a virtually 

 complete sequence determination has been achieved for ^- 

 corticotropin of the pig (12) with a single chain of 39 residues — 

 longer than either of the chains of insulin, but lacking in sulf- 



382 



