Binding modes of L35 to alpha- and beta-semihemoglobins: structural insights into the inequivalence of alpha- and beta-subunits of hemoglobin

It has been thought for several years that the greatly lowered oxygen affinity, high cooperativity, and heterotropic modulation displayed by tetrameric human hemoglobin (Hb) was an exclusive result of the assembly of high affinity alpha(1)beta(1) dimers into alpha(2)beta(2) tetramers. However, in recent times, it has been shown that alpha- and beta-semihemoglobins, namely alpha(heme)beta(apo) and alpha(apo)beta(heme), which are dimers of Hb characterized by a high affinity for oxygen and lack of cooperativity do respond to effectors such as 2-[4-(3,5-dichlorophenylureido) phenoxy]-2-methylpropionic acid (L35), a bezafibrate (BZF) related compound, by decreasing the ligand affinity to a considerable extent (between 60- and 130-fold). In order to shed some light on the structural basis of this phenomenon, we have developed a binding mode of L35 to semihemoglobins through docking analysis using the program GRID. Molecular modelling studies did identify sites on semihemoglobins where favourable interactions with L35 can occur. We found that the effector binds differently to the two semihemoglobins exhibiting high affinity only for the alpha chain heme pocket. The proposed binding models are consistent with the experimental findings and may be rationalized in terms of different hydrophobic and hydrophilic characteristics between alpha- and beta-heme pockets of Hb.     

Heterotropic effectors control the hemoglobin function by interacting with its T and R states--a new view on the principle of allostery

Careful analyses of precise oxygenation curves of hemoglobin (Hb) clearly indicate that, contrary to the common belief, allosteric effectors exert a dramatic control of the oxygenation characteristics of the protein by binding not only to the T (unligated), but also to the R (ligated) state, in a process that is proton-driven and involves proton uptake. The most striking functional changes were obtained when the allosteric effectors were bound to the fully ligated Hb: the oxygen affinity decreased dramatically, Bohr effect was enhanced, and cooperativity of oxygen ligation was almost absent, emulating a Root effect-like behavior. However, structural analysis, such as Cys beta 93 sulfhydryl reactivity and ultraviolet circular dichroism, confirmed that the ligated Hb was in fact in the R state, despite its extremely low affinity state features. These findings provide a new global view for allosteric interactions and invoke for a modern interpretation of the role of allosteric effectors and a reformulation of the Monod-Wyman-Changeaux model for control of allosteric systems, and other complementary models as well.     

Ligand-linked structural transitions in crystals of a cooperative dimeric hemoglobin

Cooperative ligand binding in the dimeric hemoglobin (HbI) from the blood clam Scapharca inaequivalvis is mediated primarily by tertiary structural changes, but with a small quaternary rearrangement (approximately 3 degrees), based on analysis of distinct crystal forms for ligated and unligated molecules. We report here ligand transition structures in both crystal forms. Binding CO to unligated HbI crystals results in a structure that approaches, but does not attain, the full allosteric transition. In contrast, removing CO from the HbI-CO crystals results in a structure that possesses all the key low affinity attributes previously identified from analysis of HbI crystals grown in the unligated state. Subsequent binding of CO shows the reversibility of this process. The observed structural changes include the quaternary rearrangement even under the constraints of lattice interactions, demonstrating that subunit rotation is an integral component of the ligand-linked structural transition in HbI. Analysis of both crystal forms, along with data from HbI mutants, suggests that the quaternary structural change is linked to the movement of the heme group, supporting a hypothesis that the heme movement is the central event that triggers cooperative ligand binding in this hemoglobin dimer. These results show both the effects of a crystal lattice in limiting quaternary structural transitions and provide the first example of complete allosteric transitions within another crystal lattice.     

High pressure reveals that the stability of interdimeric contacts in the R- and T-state of HbA is influenced by allosteric effectors: Insights from computational simulations

The molecular details of the mechanism of action of allosteric effectors on hemoglobin oxygen affinity are not clearly understood. The global allostery model proposed by Yonetani et al. suggests that the binding of allosteric effectors can take place both in the R and T states and that they influence oxygen affinity through inducing global tertiary changes in the subunits. Recently published high pressure studies yielded dissociation constants at atmospheric pressure that showed a stabilizing effect of heterotropic allosteric effectors on the dimer interface in the R state, and a more pronounced destabilizing effect in a T state model. In the present work, we report on computational modeling used to interpret the high pressure experimental data. We show structural changes in the hemoglobin interdimeric interfaces, indicative of a global tertiary structural change induced by the binding of allosteric effectors. We also show that the number of water molecules bound at the interface is significantly influenced by binding effectors in the T state in accordance with the experimental data. Our results suggest that the binding of effectors at definite sites leads to tertiary changes that propagate to the interfaces and results in overall structural re-organizations.     

Kinetics of oxygen binding to hemoglobin A

The two-state model [Monod, J., Wyman, J., and Changeux, J. P. (1965) J. Mol. Biol. 12, 88-118] postulates a single conformational change which, in the case of hemoglobin, has been related to the structural differences between deoxy and ligated hemoglobins [Perutz, M. F. (1979) Nature (London) 228, 726-739]. In its simplest form, the model does not represent satisfactorily either the equilibrium or the kinetics of the hemoglobin-oxygen reaction. The kinetic difficulty is with the rate of dissociation from the T-state, and may be met by assuming a wide difference in behavior between alpha- and beta-subunits. Experiments with Ni-Fe hybrids, however, show almost identical rates of combination with, and dissociation from, the two types of subunit, both of which develop R-like reactions as the pH is raised, the alpha-Fe-subunits at lower pH than the beta-Fe-subunits [Shibayama, N., Yonetani, T., Regan, R. M., and Gibson, Q. H. (1995) Biochemistry 34, 14658-14667]. The reactions of oxygen with hemoglobin A and the effect of pH upon them may be represented by assuming behavior of its subunits similar to that of the Ni-Fe hybrids. In such a scheme, alpha-alpha and beta-beta interactions become important elements in cooperativity, and more than two allosteric states are required, for reconsideration of the structural basis of cooperativity.     

Effect of aromatic isothiocyanates on the functional properties of human hemoglobin. Role of the stereochemistry of the charged group

The effect of chemical modification of hemoglobin with six derivatives of benzene isothiocyanate has been studied. The negatively charged reagents (isothiocyanates of benzoic and benzenesulfonic acids) markedly inhibit the interaction of hemoglobin with allosteric effectors such as H+, Cl- and organic phosphates; the affinity for heme ligands in the absence of effectors is reduced but cooperativity is maintained, making these modified hemoglobins suitable models for a possible 'blood substitute'. The only uncharged reagent tested (isothiocyanate of benzenesulfonamide) increases the oxygen affinity of hemoglobin and affects only slightly the interaction with heterotropic ligands; its potential use as an antisickling drug is under study.     

Allosteric interactions in sipunculid and brachiopod hemerythrins

Chemical and spectroscopic consequences of allosteric interactions for ligand binding to sipunculid (Phascolopsis gouldii) and brachiopod (Lingula reevii) hemerythrins (Hrs) have been investigated. Possible allosteric effectors for homotropic effects in sipunculid Hrs have been examined, but only reduction in ligand affinity is observed without cooperativity. In contrast to sipunculid Hr, L. reevii Hr binds O2 cooperatively in the pH range 7-8 and exhibits a Bohr effect. Spectroscopic comparisons of the sipunculid and brachiopod Hrs show no significant differences in the active site structures; therefore, modulation of oxygen affinity is attributable to effects linking the site to quaternary structural changes in the octamer. Oxygen equilibria can be fit with a conformational model incorporating a minimum of three states, tensed (T), relaxed (R), and an R-T hybrid. Resonance Raman spectra of L. reevii oxyHr show a shift in the peroxo stretching frequency when the pH is lowered from pH 7.7 (predominantly R oxyHr) to pH 6.3 (a mixture of R, T, and R-T hybrid), but P. gouldii Hr does not have a frequency shift under the same conditions. In contrast to hemoglobins, ligand binding to the deoxy and met forms is noncooperative for brachiopod (and sipunculid) Hrs. It is thus suggested that conformational changes in the protein are linked to the oxidation state change that accompanies oxygenation of the coupled binuclear iron site (deoxy [FeIIFeII]----oxy [FeIIIFeIII]). The total allosteric energy expended in oxygenation is about 1.4 kcal/mol, and such a shift is possible in the relaxed-tense conversion with relatively limited constraints of the iron coordination environment via the protein quaternary structure. The mechanism of cooperativity in the binuclear copper oxygen carrier hemocyanin is discussed in light of these results.     

Properties of a recombinant human hemoglobin with aspartic acid 99(beta), an important intersubunit contact site, substituted by lysine.

Site-directed mutagenesis of an important subunit contact site, Asp-99(beta), by a Lys residue (D99K(beta)) was proven by sequencing the entire beta-globin gene and the mutant tryptic peptide. Oxygen equilibrium curves of the mutant hemoglobin (Hb) (2-15 mM in heme) indicated that it had an increased oxygen affinity and a lowered but significant amount of cooperativity compared to native HbA. However, in contrast to normal HbA, oxygen binding of the recombinant mutant Hb was only marginally affected by the allosteric regulators 2,3-diphosphoglycerate or inositol hexaphosphate and was not at all responsive to chloride. The efficiency of oxygen binding by HbA in the presence of allosteric regulators was limited by the mutant Hb. At concentrations of 0.2 mM or lower in heme, the mutant D99K(beta) Hb was predominantly a dimer as demonstrated by gel filtration, haptoglobin binding, fluorescence quenching, and light scattering. The purified dimeric recombinant Hb mutant exists in 2 forms that are separable on isoelectric focusing by about 0.1 pH unit, in contrast to tetrameric hemoglobin, which shows 1 band. These mutant forms, which were present in a ratio of 60:40, had the same masses for their heme and globin moieties as determined by mass spectrometry. The elution positions of the alpha- and beta-globin subunits on HPLC were identical. Circular dichroism studies showed that one form of the mutant Hb had a negative ellipticity at 410 nm and the other had positive ellipticity at this wavelength. The findings suggest that the 2 D99K(beta) recombinant mutant forms have differences in their heme-protein environments.     

Effectors of hemoglobin. Separation of allosteric and affinity factors.

The relative contributions of the allosteric and affinity factors toward the change in p50 have been calculated for a series of effectors of hemoglobin (Hb). Shifts in the ligand affinity of deoxy Hb and the values for 50% ligand saturation (p50) were obtained from oxygen equilibrium data. Because the high-affinity parameters (liganded conformation) are poorly determined from the equilibrium curves, they were determined from kinetic measurements of the association and dissociation rates with CO as ligand. The CO on-rates were obtained by flash photolysis measurements. The off-rates were determined from the rate of oxidation of HbCO by ferricyanide, or by replacement of CO with NO. The partition function of fully liganded hemoglobin for oxygen and CO is only slightly changed by the effectors. Measurements were made in the presence of the effectors 2,3-diphosphoglycerate (DPG), inositol hexakisphosphate (IHP), bezafibrate (Bzf), and two recently synthesized derivatives of Bzf (LR16 and L35). Values of p50 change by over a factor of 60; the on-rates decrease by nearly a factor of 8, with little change in the off-rates for the liganded conformation. The data indicate that both allosteric and affinity parameters are changed by the effectors; the changes in ligand affinity represent the larger contribution toward shifts in p50.     

Kinetics and energetics of intramolecular electron transfer in yeast cytochrome c peroxidase

The oxidation of ferric cytochrome c peroxidase by hydrogen peroxide yields a product, compound ES [Yonetani, T., Schleyer, H., Chance, B., & Ehrenberg, A. (1967) in Hemes and Hemoproteins (Chance, B., Estabrook, R. W., & Yonetani, T., Eds.) p 293, Academic Press, New York], containing an oxyferryl heme and a protein free radical [Dolphin, D., Forman, A., Borg, D. C., Fajer, J., & Felton, R. H. (1971) Proc. Natl. Acad. Sci. U.S.A. 68, 614-618]. The same oxidant takes the ferrous form of the enzyme to a stable Fe(IV) peroxidase [Ho, P. S., Hoffman, B. M., Kang, C. H., & Margoliash, E. (1983) J. Biol. Chem. 258, 4356-4363]. It is 1 equiv more highly oxidized than the ferric protein, contains the oxyferryl heme, but leaves the radical site unoxidized. Addition of sodium fluoride to Fe(IV) peroxidase gives a product with an optical spectrum similar to that of the fluoride complex of the ferric enzyme. However, reductive titration and electron paramagnetic resonance (EPR) data demonstrate that the oxidizing equivalent has not been lost but rather transferred to the radical site. The EPR spectrum for the radical species in the presence of Fe(III) heme is identical with that of compound ES, indicating that the unusual characteristics of the radical EPR signal do not result from coupling to the heme site. By stopped-flow measurements, the oxidizing equivalent transfer process between heme and radical site is first order, with a rate constant of 0.115 s-1 at room temperature, which is independent of either ligand or protein concentration.(ABSTRACT TRUNCATED AT 250 WORDS)     

Spectroscopic and equilibrium studies of ligand and organic substrate binding to indolamine 2,3-dioxygenase

The binding of a number of ligands to the heme protein indolamine 2,3-dioxygenase has been examined with UV-visible absorption and with natural and magnetic circular dichroism spectroscopy. Relatively large ligands (e.g., norharman) which do not readily form complexes with myoglobin and horseradish peroxidase (HRP) can bind to the dioxygenase. Except for only a few cases (e.g., 4-phenylimidazole) for the ferric dioxygenase, a direct competition for the enzyme rarely occurs between the substrate L-tryptophan (Trp) and the ligands examined. L-Trp and small heme ligands (CN-,N3-,F-) markedly enhance the affinity of each other for the ferric enzyme in a reciprocal manner, exhibiting positive cooperativity. For the ferrous enzyme, L-Trp exerts negative cooperativity with some ligands such as imidazoles, alkyl isocyanides, and CO binding to the enzyme. This likely reflects the proximity of the Trp binding site to the heme iron. Other indolamine substrates also exert similar but smaller cooperative effects on the binding of azide or ethyl isocyanide. The pH dependence of the ligand affinity of the dioxygenase is similar to that of myoglobin rather than that of HRP. These results suggest that indolamine 2,3-dioxygenase has the active-site heme pocket whose environmental structure is similar to, but whose size is considerably larger than, that of myoglobin, a typical O2-binding heme protein. Although the L-Trp affinity of the ferric cyanide and ferrous CO enzyme varies only slightly between pH 5.5 and 9.5, the unligated ferric and ferrous enzymes have considerably higher affinity for L-Trp at alkaline pH than at acidic pH. L-Trp binding to the ferrous dioxygenase is affected by an ionizable residue with a pKa value of 7.3.     

Altered ligand rebinding kinetics due to distal-side effects in hemoglobin chico (Lysbeta66(E10) --> thr)

Hb Chico is an unusual human hemoglobin variant that has lowered oxygen affinity, but unaltered cooperativity and anion sensitivity. Previous studies showed these features to be associated with distal-side heme pocket alterations that confer increased structural rigidity on the molecule and that increase water content in the beta-chain heme pocket. We report here that the extent of nanosecond geminate rebinding of oxygen to the variant and its isolated beta-chains is appreciably decreased. Structural alterations in this variant decrease its oxygen recombination rates without significantly altering rates of migration out of the heme pocket. Data analysis indicates that one or more barriers that impede rebinding of oxygen from docking sites in the heme pocket are increased, with less consequence for CO rebinding. Resonance Raman spectra show no significant alterations in spectral regions sensitive to interactions between the heme iron and the proximal histidine residue, confirming that the functional differences in the variant are due to distal-side heme pocket alterations. These effects are discussed in the context of a schematic representation of heme pocket wells and barriers that could aid the design of novel hemoglobins with altered ligand affinity without loss of the normal allosteric responses that facilitate unloading of oxygen to respiring tissues.     

Thermodynamic model of cooperativity in a dimeric protein: unique and independent parameters formulation.

A model of the cooperative interaction of ligand binding to a dimeric protein is presented based upon the unique and independent parameters (UIP) thermodynamic formulation (Gutheil and McKenna, Biophys. Chem. 45 (1992) 171-179). The analysis is developed from an initial model which includes coupled conformational and ligand binding equilibria. This completely general model is then restricted to focus on conformationally mediated cooperative interactions between the ligands and the expressions for the apparent ligand binding constant and the apparent ligand-ligand interaction constant are derived. The conditions under which there is no cooperative interaction between the ligands are found as roots to a polynomial equation. Consideration of the distribution of species among the various conformational states in this general model leads to a set of inequalities which can be represented as a two dimensional plot of boundaries. By superimposing a contour plot of the value of the apparent ligand-ligand interaction constant over the plot of boundaries a complete graphical representation of this system is achieved similar to a phase diagram. It is found that the parameter space homologous to Koshland-Nemethy-Filmer type of model is most consistent with both positive and negative cooperativity in this model. The maximal amount of positive and negative cooperativity are found to be simple functions of Kc, the equilibrium constant associated with the change of a subunit and ligand from the unligated to ligated conformation. It is shown that under certain limiting conditions the apparent allosteric interaction between ligands is equal to the conformational interaction between subunits. The methods presented are generally applicable to the theoretical analysis of thermodynamic interactions in complex systems.     

Oxygen equilibrium and EPR studies on alpha1beta1 hemoglobin dimer

We undertook this project to clarify whether hemoglobin (Hb) dimers have a high affinity for oxygen and cooperativity. For this, we prepared stable Hb dimers by introducing the mutation Trp-->Glu at beta37 using our Escherichia coli expression system at the alpha1beta2 interface of Hb, and analyzed their molecular properties. The mutant hybrid Hbs with a single oxygen binding site were prepared by substituting Mg(II) protoporphyrin for ferrous heme in either the alpha or beta subunit, and the oxygen binding properties of the free dimers were investigated. Molecular weight determination of both the deoxy and CO forms showed all these molecules to be dimers in the absence of IHP at different protein concentrations. Oxygen equilibrium measurements showed high affinity and non-cooperative oxygen binding for all mutant Hb and hybrid Hb dimers. However, EPR results on the [alpha(N)(Fe-NO)beta(M)(Mg)] hybrid showed some alpha1beta1 interactions. These results provide some clues as to the properties of Hb dimers, which have not been studied extensively owing to practical difficulties in their preparation.     

Evidence for the existence of a low spin complex in acidic methemoglobin: its structure and formation.

By means of electron spin resonance and magneto-optical rotation, specific low spin complexes in acidic methemoglobin are obtained. The formation of these complexes is explained by a specific stereochemical arrangement of the distal histidine in the absence of allosteric effectors inducing the formation of a low spin ligand at room temperature. At low temperature, however, the distal histidine is directly bound to the heme iron. As the formation of the low spin complexes depends on allosteric effectors it is suggested that via the distal histidine the affinity of heme iron ligands is modified.     

Involvement of the distal histidine in the low affinity exhibited by Hb Chico (Lys beta 66----Thr) and its isolated beta chains.

Hemoglobin (Hb) Chico (Lys beta 66----Thr at E10) has a diminished oxygen affinity (Shih, D. T.-b., Jones, R. T., Shih, M. F.-C., Jones, M. B., Koler, R. D., and Howard, J. (1987) Hemoglobin 11, 453-464). Our studies show that its P50 is about twice that of Hb A and that its cooperativity, anion, and Bohr effects between pH 7 and 8 are normal. The Bohr effect above pH 8 is somewhat reduced, indicating a small but previously undocumented involvement of the ionic bond formed by Lys beta 66 in the alkaline Bohr effect. Since the oxygen affinity of the alpha-hemes is likely to be normal, that of the beta-hemes in the tetramer is likely to be reduced by the equivalent of 1.2 kcal/mol beta-heme in binding energy. Remarkably, both initial and final stages of oxygen binding to Hb Chico are of lowered affinity relative to Hb A under all conditions examined. The isolated beta chains also show diminished oxygen affinity. In T-state Hb A, Lys(E10 beta) forms a salt bridge with one of the heme propionates, but comparison with other hemoglobin variants shows that rupture of this bridge cannot be the cause of the low oxygen affinity. X-ray analysis of the deoxy structure has now shown that Thr beta 66 either donates a hydrogen bond to or accepts one from His beta 63 via a bridging water molecule. This introduces additional steric hindrance to ligand binding to the T-state that results in slower rates of ligand binding. We measured the O2/CO partition coefficient and the kinetics of oxygen dissociation and carbon monoxide binding and found that lowered O2 and CO affinity is also exhibited by the R-state tetramers and the isolated beta chains of Hb Chico.     

Structural and thermodynamic aspects of cooperativity in the homodimeric hemoglobin from Scapharca inaequivalvis

The homodimeric cooperative hemoglobin from the mollusk Scapharca inaequivalvis displays an unusual subunit assembly with respect to vertebrate hemoglobins. The intersubunit contact region is formed by the two heme-carrying E and F helices, which bring the two hemes in contact with each other. At variance with tetrameric vertebrate hemoglobins, the ligand binding is not accompanied by a significant quaternary transition. The major ligand-linked changes are tertiary and are limited to the heme pocket and subunit interface. These unique structural features of HbI are not easily reconciled with the classical thermodynamic models used to describe cooperative ligand binding in vertebrate hemoglobins. The lack of distinct quaternary states and the absence of allosteric effectors suggested that cooperativity in HbI is entirely homotropic in origin. Thereafter, high resolution X-ray crystallographic data displayed the preferential binding of water molecules at the intersubunit interface in the unliganded protein with respect to the liganded one. These ordered water molecules were thus proposed to act as heterotropic effectors in HbI. The contribution of specific water binding to the observed cooperativity in HbI is discussed in the framework of the enthalpy-entropy compensation effect emerging from previous accurate equilibrium oxygen binding measurements.     

Quaternary structures of intermediately ligated human hemoglobin a and influences from strong allosteric effectors: resonance Raman investigation

The Fe-histidine stretching (nu(Fe-His)) frequency was determined for deoxy subunits of intermediately ligated human hemoglobin A in equilibrium and CO-photodissociated picosecond transient species in the presence and absence of strong allosteric effectors like inositol(hexakis)phosphate, bezafibrate, and 2,3-bisphosphoglycerate. The nu(Fe-His) frequency of deoxyHb A was unaltered by the effectors. The T-to-R transition occurred around m = 2-3 in the absence of effectors but m > 3.5 in their presence, where m is the average number of ligands bound to Hb and was determined from the intensity of the nu(4) band measured in the same experiment. The alpha1-beta2 subunit contacts revealed by ultraviolet resonance Raman spectra, which were distinctly different between the T and R states, remained unchanged by the effectors. This observation would solve the recent discrepancy that the strong effectors remove the cooperativity of oxygen binding in the low-affinity limit, whereas the (1)H NMR spectrum of fully ligated form exhibits the pattern of the R state.