BARBARA W. LOW AND JOHN T. EDSALL 



The important question, however, is the relevance of these 

 stereochemical models to the actual insulin structure. The 

 x-ray study of acid insulin sulfate crystals does offer rather 

 persuasive geometrical evidence that a good part at least of the 

 insulin molecule is in a regular configuration. Arndt and Riley's 

 study gives even further evidence of this. Unfortunately, the 

 1.5 A spacing characteristic of the a-helix structure has not been 

 found, despite intensive search in x-ray photographs of wet 

 crystals of acid insulin sulfate. However, the nature of the inter- 

 chain packing between a-helices could reduce the intensity of 

 this reflection very markedly, so that negative observations of 

 this kind carry little weight. It is a little difficult, however, to 

 dismiss so easily the study of Elliott (42) , using polarized infrared 

 radiation, on air dried crystals of acid insulin sulfate. For the 

 absorption bands corresponding to the C=0 and N — H 

 vibrations marked dichroic effects would be expected from an 

 arrangement of a-helices stacked parallel to the a axis. The 

 dichroism should be reduced somewhat if a part of one helix 

 were tilted at an angle. But the complete absence of obvious 

 dichroism in these crystals is difficult to explain. At this stage 

 in such studies of protein structures, however, these observations 

 should probably be considered merely as anomalies, requiring 

 further investigation. It is interesting to note that when the 

 cystine linkages in insulin are oxidized the A and B chains both 

 appear to be in the /3 configuration (5). Schellman (98) regards 

 this as strong evidence for the favorable effect of interchain 

 covalent bonds in holding a molecule in the coiled a configura- 

 tion. Linderstr0m-Lang and his co-workers have studied the 

 kinetics of the deuterium exchange of insulin with D2O. Linder- 

 str0m-Lang (60b) reports that a certain number of the hydrogen 

 atoms exchange very slowly. The actual number of slowly 

 exchanging hydrogen atoms would suggest that only those 

 regions of the A and B chains between the two interchain cystine 

 linkages are in a regular coiled configuration with strong intra- 

 chain hydrogen bonds, for dissolved insulin. The imide hydro- 

 gens in the intrachain disulfide loop of the A chain are probably 

 free to exchange readily with deuterium. 



424 



