Electron capture by FeIII and (FeO2) centres in haemoglobin, and the absence of subsequent electron transfer from (FeO2)– centres to FeIII. An electron spin resonance study

Abstract

Haemoglobin molecules containing Fe^III and FeO₂ porphyrin units have been exposed to ⁶⁰Co γ-rays at 77 K. Electron ejection within the protein molecules is followed by efficient electron capture at both iron centres, giving a loss in the Fe^IIIg= 6 feature and a gain in features assigned to (FeO₂)^– units. Features in the g= 2 region of the spectrum are assigned to protein electron-loss centres. The results confirm that electrons in these proteins are mobile and relatively long-lived, despite their ability to react with amide groups and various side groups. Furthermore, a method is developed which gives relative yields of the Fe^II and (FeO₂)^– units with considerable accuracy and it is concluded that e^– capture by FeO₂ is ca. 2.5 times as probable as capture by Fe^III. Also, storage at 77 K for several months resulted in no detectable intramolecular electron transfer to give FeO₂ and Fe^II units despite the resultant gain in stability. This result is contrasted with a recent study involving electron transfer from ³Zn^II to Fe^III in Zn–Fe hybrids, which occurs quite rapidly at 77 K (k_t≈ 0.8 s^–1). We conclude that in our case electron-transfer rates are greatly reduced because the energy difference is smaller than for the ³Zn^II case. Various alternatives based on non-statistical distribution of O₂ in the protein tetramers are considered, but we conclude that these cannot explain the different results. Reasons for the greater ease of electron-addition to FeO₂ units relative to Fe^III units are discussed. The electron-loss centres are identified as nitrogen-centred radicals, formed from the peptide backbone by proton transfer to carbonyl oxygen atoms. That these centres become rapidly immobilized is established by the fact that when pure deoxyhaemoglobin (Fe^II) is irradiated, there is no detectable formation of Fe^III units. From an analysis of the β-proton hyperfine coupling in these N-centred radicals it is shown that the majority must be formed within the α-helical regions of this protein, the conformations requiring that A(βH)≈ 38 G. However, a considerable number are formed in random-coil regions; these give small β-proton splittings.