Identifying the conformational state of bi-liganded haemoglobin

Abstract

While most researchers agree on the global features of cooperative ligand binding to haemoglobin (Hb), the internal mechanisms remain open to debate. This is not due to inaccurate measurements, but is rather a consequence of the cooperative ligand binding that decreases the equilibrium populations of the partially liganded states and makes observation of the transitions between these substates more difficult. For example, the equilibrium population of the doubly liganded tetramers is typically less than 5% of the total Hb. As a result many models with widely varying mechanisms may fit the oxygen equilibrium curve, but may not be consistent with observations of other parameters, such as ligand-binding kinetics or subunit association equilibria. The wide range of methods and models has led to divergent conclusions about the properties of specific substates. One notable debate concerns the properties of the doubly liganded forms. The simple two-state model predicts a shift in the allosteric equilibrium based on the number of ligands bound, but not on their distribution within the tetramer. From studies of dimer-tetramer equilibria of various pure and hybrid forms, it was concluded that a tetramer with two ligands bound on the same alpha beta dimer (species 21, an asymmetric hybrid) shows an enhanced tetramer stability, similar to singly liganded Hb, relative to the other three types of doubly liganded tetramers which resemble the triply liganded forms [Ackers et al. (1992). Science 255: 54-63]. The implications of this model and the relevant experiments will be reviewed here.

Confirmation of a unique intra-dimer cooperativity in the human hemoglobin ?1?1half-oxygenated intermediate supports the symmetry rule model of allosteric regulation

“…This finding does not alter the consensus distribution Z Hb , which has not required modification to accommodate the new databases after its formulation in 1992. This is in stark contrast to opposing arguments that ligand rearrangement casts significant doubt on, or actually refutes, the SR model 20–24, 26, 33. These viewpoints are also addressed in the present study, since the recurring theme of equating the CNmet ij ΔG c distribution with the SR model underlies these arguments as well.…”

“…This was perhaps best illustrated by a study18 in which, as a test of the SR model, a more complex 30‐state model was formulated and then evaluated by attempting to fit O 2 binding curves with CNmet microstate free energies (which did not fit), rather than with the relevant consensus Fe/FeO 2 microstate values (which had been shown to fit the experimental isotherms in figure 3 of Ackers et al5). While the failure of that 30‐state model has been cited repeatedly as evidence against the SR model,19–26 the actual quantitative consistency between SR model parameters and the tetrameric O 2 binding curves of native Hb has been demonstrated extensively 12, 13, 27. The fundamental distinction between individual analog ij ΔG c values and their consensus ij ΔG c distribution will thus be considered further in relation to new data of the present study.…”

“…8 or 9 using the standard relationship: while the traditional Hill coefficient n H is: The four experimentally determined Adair parameters K i provide much less stringent constraints for testing molecular models than the nine microstate constants of which they are comprised 1. It was nevertheless important to evaluate the accuracy of microstate‐generated Adair constants in predicting the Y vs. [O 2 ] isotherm of native HbO 2 measured under solution conditions of the present study16 because of the claims18–26 that the SR model parameters would not be consistent with the binding curve of native human Hb, Fe/FeO 2 (see Discussion). In contrast to these claims, the isotherm calculated using the 9 experimental microstate constants (Table II) is found to be in close conformity with the measured tetramer binding isotherm, as shown in Figure 4b.…”

“…The assessment of any such work without consideration of the entire database reduces the inherently complex Hb problem to that of a simple “litmus test,” e.g., of the SR model as opposed to the MWC model. Applying a litmus test is faster and easier than conducting a complete analysis: any single technique or data subset that fails to reproduce the consensus distribution of doubly ligated microstates may thus be deemed inconsistent with the SR model,18, 19, 21, 22, 25, 26, 31, 32 or, in the extreme, to actually refute the SR model 18, 19–26, 31, 32. A complicating factor that voids the reliability of the litmus approach is the possibility that tertiary and quaternary components of cooperative free energy may compensate to yield different patterns of free energy (such as 21 ΔG c ≈ 22 ΔG c ≈ 23 ΔG c ≈ 24 ΔG c ) under a given set of conditions.…”

“…Thus, by definition, the CNmet characteristics were in agreement with the original Z Hb and with the SR partition function,1, 29 as were the other analog databases, while each exhibited some variations or anomalies. The value of 21 ΔG c for the CNmet analog was never proposed to be equated with the value of 21 ΔG c for HbO 2 , contrary to representations18, 19–26, 30–32: the proposal instead was to compare 21 ΔG c values and all other relevant data to Z Hb 1, 5, 12, 29…”