PROTEIN STRUCTURE 



atom of the C1-C2 residue to the aCo, and closing the ring by 

 forming aCo— /^C.— CH.— CHo— N bonds. 



N 



CH2 

 \CH./ 



/3CH. 



This would eliminate one hydrogen bond and lengthen another — 

 it would not, however, completely disrupt the helical configura- 

 tion. 



On the other hand, if the /3-carbon were in the alternative 

 position in Figure 5, equivalent to a left-handed helix, the 

 proline ring could not be closed. The presence of a proline 

 residue would completely disrupt the helical sequence in this 

 latter configuration. Proline residues thus provide a possible 

 stereochemical reason for discontinuity in left-handed helix 

 structures. The photographs of scale models (Figure 8) illustrate 

 the two difTerent stereochemical effects of an L-proline residue 

 in left-handed and right-handed a-helix structures. In heino- 

 globin it is suggestive that the number of proline residues per 

 globin (half- molecule) is eleven. Proline residues and the three 

 end groups are together almost adequate in number to account 

 for the number (17) of crystallographic chains pictured in the 

 model if the proline residues do occupy the required positions. 

 Proline residues in left-handed helix structures would certainly 

 lead to a region of irregular chain configurations and could 

 initiate corner turning. Howev'er, the partial sequences which 

 have been established for ribonuclease (discussed later) place 

 two of the four proline residues in a short terminal sequence of 

 15 residues. There are 124 residues in the chain (48) and the 

 turns must occur at fairly regular interv^als along the chain to 

 give a relatively compact molecular shape, such as is required 

 by the physical properties of the molecule. This suggests that 

 residues otlier than proline must sometimes be responsible for 

 initiating a reversal of chain direction. An economical and 



411 



