Search

Previous studies on bovine hemoglobin (HbBv) have suggested amino acid substitutions, which might introduce into human hemoglobin (HbA) functional characteristics of HbBv, namely a low intrinsic oxygen affinity regulated by Cl(-). Accordingly, we have constructed and characterized a multiple mutant, PB5, [beta(V1M + H2 Delta + T4I + P5A + A76K)] replacing four amino acid residues of HbA with those present at structurally analogous positions in HbBv, plus an additional substitution, beta T4I, which does not occur in either HbBv or HbA. This 'pseudobovine' hemoglobin has oxygen binding properties very similar to those of HbBv: the P(50) of HbA, PB5 and HbBv in the absence of Cl(-) are 1.6, 4.6 and 4.8 torr, respectively, and in 100 mM Cl(-) are 3.7, 10.5 and 12 torr, respectively. Moreover, PB5 has 3-fold slower autoxidation rate compared to HbA and HbBv. These are desirable characteristics for a human hemoglobin to be considered for use as a clinical artificial oxygen carrier. Although the functional properties of PB5 and HbBv are similar, van't Hoff plots indicate that the two hemoglobins interact differently with water, suggesting that factors regulating the R to T equilibrium are not the same in the two proteins. A further indication that PB5 is not a functional mimic of HbBv derives from PB5(control), a human hemoglobin with the same substitutions as PB5, except the beta T4I replacement. PB5(control) has a high oxygen affinity (P(50)=2.3 torr) in the absence of Cl(-), but retains the Cl(-) effect of PB5. The Cl(-) regulation of oxygen affinity in PB5 involves lysine residues at beta 8 and beta 76. PB4, which has the same substitutions as PB5 except beta A76K, and PB6, which has all the substitutions of PB5 plus beta K8Q, both have a low intrinsic oxygen affinity, like HbBv and PB5, but exhibit a decreased sensitivity to Cl(-). Since HbBv has lysine residues at both beta 8 and beta 76, these results imply that Cl(-) regulation in HbBv likewise involves these two residues. The mechanism responsible for the low intrinsic oxygen affinity of HbBv remains unclear. It is suggested that residues peculiar to HbBv at the alpha(1)beta(1) interface may play a role.

In the crystal structure of human T-state hemoglobin with a sebacyl residue cross-linking the two beta-subunit Lys82,s (DecHb), the Fe atoms of the alpha-subunit hemes are found to be oxidized with a water molecule bound. The three-dimensional structure and heme geometries were compared to those of deoxyhemoglobin and other partially and fully oxidized hemoglobins [R. Liddington, Z. Derewenda, E. Dodson, R. Hubbard, G. Dodson, J. Mol. Biol. 228 (1992) 551]. The heme geometries of the alpha-subunits are consistent with those observed in oxidized structures. The proximal histidines of the alpha-subunits move toward the heme plane shifting the F-helix and FG-corner in a manner observed for partially oxidized human hemoglobin. This supports the hypothesis that these perturbations may precede the T- to R-state transition. Circular dichroism studies comparing DecHb and natural human hemoglobin in the deoxy and CO ligated forms confirm that the conformations of the deoxy forms are identical, but the ligated forms have slight differences in the solution structures. DecHb is found to be more resistant to autoxidation than natural hemoglobin. The time course of autoxidation of DecHb shows that it is virtually absent for the first 1500 min followed by a rapid increase. Thus, the discovery of the oxidation of the alpha-subunits in the deoxy-crystals is quite unexpected. The data confirm that ligation of the alpha-subunits precedes that of the beta-subunits. This may suggest a low ligand affinity of the alpha-diligated form of hemoglobin.

The dimer-tetramer association constants of several recombinant human hemoglobins (in the CO form) have been measured by differential gel filtration. Recombinant human hemoglobin prepared from recombinant beta-chains, and mutant hemoglobins where the substitution was on the surface, beta(Thr4-->Asp), in the heme pocket, beta(Val67-->Thr), at the 2,3-DPG binding site, beta(Val1-->Met+His2del), had a twofold smaller association with respect to natural hemoglobin. In a mutant at the alpha 1 beta 2 interface, beta(Cys93-->Ala), the association constant was decreased three-fold. Conversely, in a mutant at the alpha 1 beta 1 interface, beta(Cys112-->Gly), the association constant was two- and four-fold increased with respect to natural and recombinant human hemoglobin. These differences are energetically very small, consistent with the correct folding of the recombinant hemoglobins. The stabilization of the tetrameric structure by a mutation at the alpha 1 beta 1 interface indicates that structural changes in this interface can be propagated through the protein to the alpha 1 beta 2 interface and, thereby, exert an effect on the allosteric equilibrium.

Measurements of oxygen binding to bovine hemoglobin have been carried out over the temperature range 15-37 degrees C at pH 7.33. The standard enthalpy of oxygenation after correction for the heat of oxygen solution and of the Bohr protons is found to be -7.1 or -7.2 kcal/mol in the presence of 0.1 M chloride or bromide, respectively. This value is well below the -14.4 kcal/mol determined for human hemoglobin under identical experimental conditions. As reported by Fronticelli et al. (C. Fronticelli, E. Bucci and A. Razynska, J. Mol. Biol. 202 (1988) 343), the preferential binding of anions by bovine hemoglobin recognizes the various halides. Measurements at various temperatures reveal that this is true only above 25 degrees C. The halide recognition and the less exothermic enthalpy of oxygenation of bovine hemoglobin are probable due to oxygen-linked hydrophobic effects that are larger in bovine than in human hemoglobin.

Bovine red cells do not contain appreciable amounts of 2,3-diphosphoglycerate (2,3-DPG). Bovine hemoglobin, however, has a particular sensitivity to chloride ions and as a result it can attain oxygen affinity values lower than those measured for human hemoglobin in the presence of 2,3-DPG. The interaction of bovine hemoglobin with anions is modulated by the hydrophobic characteristics of the protein. Comparison of the hydropathy plots of primate and ruminant hemoglobins indicates constant regions of opposite hydrophobicity, which have fixed amino acid differences. A model is proposed for explaining the regulation of oxygen affinity by chlorides, as an alternative to the classic modulation by 2,3-DPG.

Perturbation of sarcoplasmic reticulum ATPase with the nonionic detergent C12E8 is modulated by the amount of free Ca2+ present in the solvent prior to the addition of detergent. CD measurements show that the enzyme exists in solution in two different conformations that react differently with the detergent. They probably represent the free enzyme, and its complex with Ca2+. On this assumption, titrations with increasing amounts of Ca2+ produced data superimposable on curves obtained measuring Ca2+ bound to sarcoplasmic reticulum vesicles.

We used front-face illumination to examine the steady-state and time-resolved emission from the intrinsic tryptophan emission of human hemoglobin (Hb). Experimental conditions were identified which eliminated all contributions of scattered light. The sensitivity obtained using front-face optics was adequate to allow measurement of the wavelength-dependent frequency response of the emission to 2 GHz. The intensity decays displayed pico- and nanosecond components in the emission at all wavelengths from 315 to 380 nm. The contribution of the picosecond component decreased from 72 to 37% over this range of wavelengths. Frequency-domain measurements were used to calculate the time-resolved emission spectra and decay-associated emission spectra. These spectra indicate that the picosecond components of the emission display maxima near 320 nm, whereas the nanosecond components are centered at longer wavelengths near 335 nm. The nanosecond components appear to be due to residual impurities which remain even in highly purified samples of Hb. However, we cannot eliminate the possibility that some of these components are due to Hb itself.

We have used time-resolved fluorescence spectroscopy for following the unfolding of apomyoglobin in urea and guanidine hydrochloride (GdnHCl). The data have been compared with those obtained using classical techniques such as CD and steady-state emission spectroscopy. Both the average intensity of the lifetimes and the size of the librational cone of the fluorophores, as measured by time-resolved fluorescence, increased with denaturant concentration and their changes largely preceded the modifications detectable with CD and the shift of the maximum of emission spectra. The data indicate that the changes in the local environments of the tryptophans were completed when the global modification monitored by CD and the emission spectra was still minimal. This suggests that an initial event in the denaturation of apomyoglobin is localized at the tryptophan residues. The correlation times of native apomyoglobin showed the rotational diffusion characteristics of a rigid rotor. In 3.6 M GdnHCl and 7.5 M urea, where the secondary structure is practically absent, the correlation times of the two systems became very short, as expected from the motion of a flexible polymer. In GdnHCl, under conditions of partial unfolding, it was not possible to detect the presence of native totally folded molecular species.

Continuing our investigation of the relationships between internal motions and functional properties of soluble and membrane-bound proteins we have explored the lifetimes and correlation times associated with the fluorescence emission of fluorescein-labeled Ca2+-dependent ATPase of sarcoplasmic reticulum. The emission was characterized by two lifetime components near 1.8 and 4.1 ns, probably due to exposure of the probe to environments of different polarities. The time-dependent anisotropy showed the presence of two correlation times near 0.8 and 6.6 ns. The shorter correlation time was due to motions of the probe around its point of attachment on the surface of the protein. The longer correlation time indicated the presence of internal motions of the protein. Both lifetimes and correlation times were insensitive to temperature between 2 and 10 degrees C. They were also insensitive to addition and removal of 100 microM free Ca2+.

Human hemoglobin and its isolated beta-subunits were denatured by addition of HCl so as to reach final pH values ranging from 2.0 to 3.2. The beta-subunits were alkylated in both the beta 93 and beta 112 cysteines; this treatment makes the beta-subunits monomeric. The kinetics of acid denaturation of the two proteins was followed spectropolarimetrically in the millisecond time range, measuring the changes in circular dichroism at 225 nm. At all pH values, in both systems, the decay of ellipticity could be simulated by two exponentials. The initial ellipticity values of the solutions, obtained by extrapolation at zero time, were those expected for the native proteins. The rates of denaturation were lower in the hemoglobin system than in the isolated monomeric beta-subunits. The data suggest that in the tertiary structure of hemoglobin and beta IAA there are different domains which unfold at different rates upon exposure to acid.