ON THE NATURE OF HEMOPROTEIN REACTIONS 



titration experiments (47). For hemoglobin this last method 

 gives A//° = 6.5 kcal./mole, supporting the hypothesis that the 

 group with /jKo is the glyoxalinium NH *" of a histidine residue 

 to which the heme is joined. Studies of the slight variation in 

 absorption spectrum of ferrimyoglobin fluoride with pH and 

 temperature give AH^ = 2 ± 2 kcal./mole and A^**^ = — 22 e.u. 

 for its linked group (17). These values more nearly resemble 

 those for the ionization of a carboxylic acid than for the NH ^ 

 of histidine. However, before concluding from these differences 

 that the heme is joined to a diff'erent amino acid residue in 

 hemoglobin and myoglobin more confirmatory data for other 

 complexes would be desirable, but in addition there are good 

 reasons for doubting whether such a direct conclusion is valid. 

 Firstly, hemoglobin from different species shows Bohr eff'ects of 

 different magnitudes, and sea lamprey hemoglobin (with 

 apparently only one heme per molecule) has an even larger 

 Bohr effect than most mammalian hemoglobins (36). Secondly, 

 if hemoglobin is reconstituted from heme and native globin the 

 Bohr effect is much smaller (47). These facts mean either that 

 the heme is joined to different amino acid residues in the various 

 hemoglobins, and to a different residue in reconstituted hemo- 

 globin compared with the original, or else that slight differences 

 in structure or configuration about an identical residue to 

 which the heme is joined in all these molecules are responsible 

 for the variable behavior. The former explanation carries with 

 it the assumption that reversible combination with oxygen is a 

 property shared by several amino acid residues, the latter, that 

 it is a very special property. For this reason the latter explanation 

 is to be preferred, and so if the acid strength of an ionizable 

 group, in an identical amino acid residue to which the heme is 

 attached in all hemoglobins, is dependent upon the structure 

 or configuration of adjacent parts of the protein fabric, then the 

 thermodyamic data for its ionization are clearly composite, and 

 cannot be used to identify the group by straightforward com- 

 parison with corresponding data for its ionization in the isolated 

 amino acid or simpler peptides. The contrast between the data 



363 



