PROTEIN STRUCTURE 



(35) have prepared poly-7-benzyl-L-glutamate of widely varying 

 degrees of polymerization, each preparation being of highly 

 uniform chain length. In weakly interacting solvents such as 

 chloroform, these exist in the a form if the molecular weight is 

 high; strongly hydrogen-bonding solvents, such as dichloroacetic 

 acid, lead to unfolding of the polymer into a randomly coiled 

 configuration, regardless of molecular weight. Very low mo- 

 lecular weight polymers, in weakly interacting solvents, exist 

 partly in the /3 form, which dissociates upon dilution to give a 

 solvated form. All this is qualitatively in accord with the theo- 

 retical discussion outlined above, but the most interesting field 

 of study — that of the configuration of large water-soluble poly- 

 peptides in aqueous solution — is only just beginning. 



What is the relation of the considerations we have just 

 discussed to the total structure of a globular protein? Obviously 

 no simple answer can suffice. A complex of polypeptide chain 

 and side-chain interactions will determine the three-dimensional 

 structure of each single molecule. A single molecule is itself in 

 this respect comparable to a crystal structure. Once the 

 crystal structure of a compound has been established and the 

 atomic parameters determined the analysis carries its own 

 persuasive conviction. The whole range of pertinent intra- 

 and inter-molecular packing principles is seen to be nicely 

 and effectively balanced. We may expect a similar satisfaction 

 when the three-dimensional structure of a globular protein 

 molecule is determined. The detailed crystal structures, even 

 of relatively simple organic compounds, however, have never 

 yet been determined simply by correct guesswork, even when the 

 probable packing principles were recognized. Many crystal 

 structures, such as those of penicillin (32) and threonine (100), 

 took two years to solve although full use was made of all the evi- 

 dence available, both chemical information and diffraction data. 



The brilliant work of Pauling and Corey has established one 

 of the outstanding packing principles in protein stereochemistry. 

 We know that there are regions of a-helix configuration in many 

 proteins. It would, however, be unrealistic to conclude that 



419 



