BARBARA W. LOW AND JOHN T. EDSALL 



covalent cross links between them. This, for instance, appears 

 to be the case with horse hemoglobin, for which Porter and 

 Sanger (87) have shown the presence of 6 A^-terminal amino 

 groups per molecule, and therefore, by inference, of 6 different 

 peptide chains. The work of Ingram (51), already mentioned, 

 on the sulfhydryl groups of horse hemoglobin has shown that 

 6 sulfhydryl groups are present, but no disulfide cross links. 

 Therefore, since there is also no evidence whatever of phosphate 

 cross linkages in hemoglobin, other types of bonds must be 

 invoked to explain the stability of the molecule. Some of these 

 may be the relatively weak linkages which Steinhardt and 

 Zaiser (102) have shown to be broken on acid denaturation 

 with the release of 36 acid and basic groups which are nontitra- 

 table in the native molecule. This, however, cannot be the 

 complete explanation, since the acid denaturation is not accom- 

 panied by a splitting of the hemoglobin into subunits. Although 

 concentrated urea causes a splitting of hemoglobin into two 

 subunits, each of half the original molecular weight, these half 

 units remain as stable entities even in concentrated urea solution 

 (109). There is no indication of the further breakdown that 

 would occur if the molecule were split into 6 constituent peptide 

 chains. In any case, the presence of large numbers of unreactive 

 acid and basic groups in native hemoglobin, which become 

 available on denaturation as shown by Steinhardt and Zaiser, 

 points to a very special set of stabilizing factors in the native 

 hemoglobin structure. The further characterization of this 

 structure, and its correlation with the rest of our knowledge of 

 hemoglobin, should lead to most important advances. Hemo- 

 globin is probably the most thoroughly investigated of all the 

 proteins, yet it still presents most baffling problems; a synthesis 

 of all our knowledge on the subject is badly needed but has not 

 yet emerged. 



In some proteins the phenolic hydroxyl groups of tyrosine 

 residues appear to be cross linked with other groups, perhaps by 

 hydrogen bonding with ionized carboxyl groups of aspartic or 

 glutamic acid residues. The simplest analogue for such hydrogen 



390 



