F. SANGER 



the peptides from enzymic digests, by estimating the ammonia 

 liberated from them by acid hydrolysis. Certain difficulties 

 were encountered owing to the small amounts of peptides 

 available, but in general results by the two methods were con- 

 current. 



The Disulfide Bridges 



The oxidized insulin fractions contained cysteic acid resi- 

 dues in place of the three cystine residues of the original insulin. 

 In order to find the distribution of the disulfide bridges of these 

 cystine residues, intact insulin was submitted to partial hydroly- 

 sis. The resulting cystine-containing peptides were fractionated 

 from one another and oxidized, and the cysteic acid peptides so 

 obtained were identified. Since the latter peptides could be 

 completely separated from other peptides by ionophoresis at 

 pH 3.3 (Figure 3), it was unnecessary to obtain the cystine 

 peptides free of non-cystine-containing peptides, provided the 

 former could be separated one from another. 



In the initial experiments concentrated HCl was used for 

 hydrolysis, and a large number of cystine peptides were identified 

 in this way. However, the results could not be interpreted in 

 terms of a unique structure for insulin, and it appeared that the 

 disulfide bonds had become mixed up during the procedure. 

 This was found to be due to an interchange reaction of one 

 disulfide with another : 



R1SSR2 + R3SSR4 ' RiSSRs + R2SSR4, etc. 



This reaction was studied using as model system a mixture 

 of cystine and bis-DNP-cystine. On interchange they gave rise 

 to mono-DNP-cystine, which was water-soluble and could thus 

 be estimated independently of the ether-soluble bis-DNP-cystine. 



DNP-CyS CyS DNP-CyS 



I + I ^=^2 I 



DNP-CyS CyS CyS 



A similar interchange took place in neutral solution. This 



450 



