Concentration-dependent Effects of Anions on the Anaerobic Oxidation of Hemoglobin and Myoglobin*

The redox potentials of hemoglobin and myoglobin and the shapes of their anaerobic oxidation curves are sensitive indicators of globin alterations surrounding the active site. This report documents concentration-dependent effects of anions on the ease of anaerobic oxidation of representative hemoglobins and myoglobins. Hemoglobin (Hb) oxidation curves reflect the cooperative transition from the T state of deoxyHb to the more readily oxidized R-like conformation of metHb. Shifts in the oxidation curves for Hb A0 as Cl− concentrations are increased to 0.2 m at pH 7.1 indicate preferential anion binding to the T state and destabilization of the R-like state of metHb, leading to reduced cooperativity in the oxidation process. A dramatic reversal of trend occurs above 0.2 m Cl− as anions bind to lower affinity sites and shift the conformational equilibrium toward the R state. This pattern has been observed for various hemoglobins with a variety of small anions. Steric rather than electronic effects are invoked to explain the fact that no comparable reversal of oxygen affinity is observed under identical conditions. Evidence is presented to show that increases in hydrophilicity in the distal heme pocket can decrease oxygen affinity via steric hindrance effects while increasing the ease of anaerobic oxidation.

The globin structure controls the redox potential of the heme site of hemoglobins (Hbs) 1 and myoglobins (Mbs), protects them from rapid oxidation, and thereby allows for reversible oxygen binding. Studying the redox properties of such systems allows us to gain insight on how anion-induced alterations fine tune the equilibrium between the oxidized and the reduced forms of Hbs. Our results extend earlier studies that have brought to light many possible modes of altering the electronic and ligand binding properties of the active site iron atoms .
Although a large body of literature describes mechanisms involved in controlling the oxygen affinities of Hbs and Mbs, there are still unanswered questions with regard to how globin structure controls the redox potential (E1 ⁄2 ) of the active site, a parameter that describes the propensity of the site to donate or accept electrons. Evaluating and predicting trends in the redox potentials of Hbs and Mbs, as the composition of both the globin and the medium are changing, is a complex and challenging problem, but one that is fundamental to understanding the way nature is able to accomplish oxygen transport, oxygen storage, and electron transfer reactions in highly polar environments.
Oxygenation and oxidation-reduction studies of Fe(II)/Fe(III) centers in diverse Hbs have uncovered parallels between these two processes (22). Notably, the shift from the T (deoxy Fe(II)) to the R (oxygenated) or R-like (met Fe(III)) conformation of the Hb tetramer underlies the cooperativity observed in both oxygenation and oxidation processes. Structural changes that stabilize either the T or the R conformation are typically reflected by parallel alterations of both oxidation and reduction processes (1,3,22). In many Hbs, heterotropic effectors such as protons, anions, and carbon dioxide, although bound at spatially remote sites, influence the oxygenation process and have been shown to affect the oxidation process as well (22)(23)(24)(25)(26). Some anionic effectors, known to bind at the ␤-␤ interface in the central cavity of Hb, are capable of influencing the oxygen affinity and the redox potential of the heme groups that are located several Å away.
Anion effects on Hb function have traditionally been explained in terms of preferential binding to the low affinity (T state) quaternary conformation (27,28). Chiancone and coworkers (23,29,30) showed, using NMR techniques, that Cl Ϫ binds with different affinities to deoxy and oxyHb and that the binding is proton-dependent. Their results indicated the presence of two major classes of binding sites, differing in Cl Ϫ affinity. The high affinity sites of deoxyHb showed a 10-fold higher Cl Ϫ affinity (about 100 M Ϫ1 ) than in oxyHb (about 10 M Ϫ1 ). These sites are distinctly different from the low affinity sites (about 0.1 M Ϫ1 ) that have approximately the same Cl Ϫ affinity for both oxy and deoxyHb.
Precise pH-stat measurements of Cl Ϫ effects on liganded and deoxyHb confirmed the fact that both forms of Hb bind Cl Ϫ but with different affinities (31). These affinity differences have a direct bearing on the magnitude and often disputed origin of the Bohr effect (31)(32)(33)(34)(35). The N-terminal valine residues and positively charged residues of the ␤ chain 2,3-diphosphoglycerate binding site have been implicated in preferential anion binding by deoxyHb, although recent x-ray and chemical studies of Perutz et al. (36) found no clear indications of specific Cl Ϫ binding sites in either bovine or human hemoglobin. We recently showed that Cl Ϫ effects on oxygen binding derive in part from steric hindrance effects that are not present in oxidation processes (22). The situation with regard to Cl Ϫ effects on Hb function is thus more complex than can be explained by anion binding at a few specific sites. Lack of awareness of the rather dramatic changes in ease of anaerobic oxidation of the heme that occur as anion levels are altered between 0 and 0.4 M, as documented in this report, may have led to apparently conflicting results in laboratories examining Hb function under what appeared to be similar experimental conditions.
We have previously shown that comparison of oxidation and oxygenation curves can help discriminate between the electronic and steric consequences that result from changes in globin structure (22). In this paper we present the results of systematic studies of the anaerobic oxidation of a selected set of structurally distinct Hbs and Mbs as a function of anion type and concentration. Our aim is to further clarify the impact these anions have on the electronic properties of the active site of these proteins.
Anaerobic oxidation curves were obtained to document the influence of a broad concentration range of anionic effectors (organic polyphosphates, chloride, nitrate, and perchlorate) on the oxidation of various Hbs (dolphin Hb, horse Hb, spot fish Hb, Hb Presbyterian, and Hb A 0 ) and Mbs (sperm whale Mb and its carbamylated variant, horse Mb and its carbamylated variant, and Aplysia Mb). The oxidation curves reveal intriguing aspects of the concentration dependence of anion effects on the electron affinity of the heme groups of these representative Hbs and Mbs and highlight the importance of a careful choice of background electrolyte and buffer system when studying these heme proteins.
Our earlier comparisons of oxygenation and oxidation processes led us to suggest a new paradigm of Hb function, in which anion-induced restrictions of the conformational fluctuations of the Hb molecule play a large and previously unrecognized role in control of oxygen affinity (26). The concentration-dependent anion effects presented in this report further support this paradigm and show that anion-induced changes in the hydrophilicity of the heme pocket could account for the dramatic contrast between anion effects on oxidation and oxygenation processes.
EXPERIMENTAL PROCEDURES Ru(NH 3 ) 6 Cl 3 (Strem Chemical Co.; Ͼ99%), NaNO 3 (Mallinckrodt; Ͼ99%), KNO 3 (Sigma; Ͼ99%), MOPS (Sigma; Ͼ99%), HEPES (Sigma; Ͼ99%), KCl (Fisher; Ͼ99%), IHP (Sigma; 98%), and platinum (52 mesh gauze; Fisher; 99.95%) were used as received. Horse Mb, carbamylated horse Mb, sperm whale Mb, carbamylated sperm whale Mb, Aplysia Mb, Hb A 0 , bottlenose dolphin Hb, and horse Hb were prepared by the ammonium sulfate method (37). Samples were stripped of organic phosphates by a two-step passage through mixed bed resin, using a high salt elution to displace tightly bound phosphates followed by a low salt elution to remove excess salt. All samples except horse Hb (two major Hbs) and dolphin Hb (showing single electrophoretic bands) were subjected to chromatographic purification with a fast performance liquid chromatography system and stored in liquid nitrogen at 1-3 mM (in heme). Sample concentrations and oxidation states were determined spectrophotometrically using published extinction coefficients (38). Samples containing spectrally detectable levels of hemichrome were discarded. The relative levels of oxidized Mb or Hb (metMb or metHb) and oxygenated Mb or Hb (oxyMb or oxyHb) were determined by spectral analysis using a Cary model 2300 UV-visible spectrophotometer.
For spectroelectrochemistry, the electrochemical mediator Ru(NH 3 ) 6 Cl 3 was dissolved in a 0.05 M MOPS buffer or 0.05 M HEPES buffer solution adjusted to pH 7.1 or 7.5, respectively, to give a concentration of 4.5-5.5 mM. Ru(NH 3 ) 6 3ϩ/2ϩ has a E1 ⁄2 of 0.214 V relative to the normal hydrogen electrode (NHE; all potentials reported in this manuscript are standardized relative to the NHE scale), which makes it suitable as a mediator for Mb and Hb electrochemistry. It creates no spectral or allosteric interferences as shown previously (25). MOPS and HEPES were selected as the buffers for their noncomplexing nature and stability, as well as the absence of spectral and electrochemical inter- For each experiment, a solution containing 1 mM Ru(NH 3 ) 6 Cl 3 and 0.05 M MOPS or HEPES at pH 7.1 or 7.5 with specific concentrations of anions in a 5-ml pear-shaped flask was connected to a vacuum line for repeated pump purging with N 2 , followed by addition of Mb or Hb and additional pump purging with gentle swirling to minimize bubbling. Final concentrations were typically 0.06 -0.08 mM in heme.
Spectroelectrochemical experiments were carried out in an anaerobic optically transparent thin layer electrode cell made in-house of a 1 ϫ 2-cm piece of 52-mesh platinum gauze placed between the inside wall of a 1-cm-path length cuvette and a piece of silica glass held in place by a small Tygon spacer positioned so as not to interfere with the spectral measurement (25). This assembly results in an optical path length of 0.03 cm, as calculated from the absorbance of the fully oxidized spectrum of Mb/Hb at ca. 554 nm (3). The cell was kept anaerobic by capping the cuvette with a septum that allowed no air to enter but permitted a continuous flow of N 2 . A platinum wire connected to the platinum gauze working electrode was inserted through the septum. A Pasteur pipette salt bridge plugged at the bottom with an agar gel was prepared so as to connect the Ag/AgCl reference (Bioanalytical Systems Inc.) electrode to the working electrode. The salt bridge solution was composed of 0.2 M KCl in 0.05 M MOPS at pH 7.1 or 0.2 M KCl in 0.05 M HEPES at pH 7.5 and was degassed and then flushed with N 2 for 1 h. The optically transparent thin layer electrode cell was purged with N 2 for 15 min prior to injecting the protein solution.
In a typical experiment, about 0.5 ml of the working solution was injected at the bottom of the optically transparent thin layer electrode cell via a gas tight syringe. The cell was then placed in the temperaturecontrolled cell holder of a CARY 2300 UV-visible spectrophotometer held at 20°C and linked to a PAR model 75 potentiostat. Spectroelectrochemistry was carried out from 340 to 700 nm, with specific emphasis on the Soret region. Absorbance changes were monitored at 410 nm (metMb) or 406 nm (metHb) and 435 nm (deoxyMb) or 430 nm (de-oxyHb). The absorbances of the fully oxidized (A o ) and fully reduced (A r ) Mb and Hb were obtained by applying a potential of ϩ400 mV and Ϫ250 mV (versus NHE), respectively, and the absorbance was recorded when the system reached equilibrium (no more change in absorbance at the fixed potential and wavelength). For each experiment, the path length was determined using the Soret band absorbance. The concentration was determined independently by spectral analysis after addition of the oxygen scavenger sodium dithionite (2 mg/ml) to the unused sample solution.
A typical increment of 20 mV was applied to the system starting at approximately ϩ300 mV down to Ϫ120 mV (versus NHE). At each applied potential, the absorbance was monitored until no change was detected. Although most experiments were performed going from fully oxidized to fully reduced Mb, the system was shown to be reversible under our experimental conditions (i.e. the Nernst plot can be generated in either the oxidation or reduction direction and equilibrium is achieved within 30 -40 min at each applied potential). Nernst plots were then derived from the observed changes in absorbance as was described previously (25). Fig. 1a sets the stage for comparative studies by illustrating a set of oxidation curves, presented as Nernst plots, for three distinct Mb systems. These plots illustrate redox differences associated with a first coordination shell effect. Monomeric Aplysia Mb exhibits an E1 ⁄2 value 75 mV more positive than that of sperm whale Mb and horse Mb. The Nernst plots have slopes of unity, as we previously reported for similar noncooperative one-electron transfer systems (3,39). A valine residue at position E7 in Aplysia Mb replaces the more common distal histidine present in both sperm whale and horse Mbs (40). The crystal structures show the presence of a water molecule in the distal heme pocket of both horse and sperm whale Mbs, stabilized by the distal histidine residue (3, 40 -43). The presence of this water molecule as a sixth ligand in the oxidized form of the protein is consistent with the negative shift in E1 ⁄2 observed for these two Mbs relative to Aplysia Mb, which does not possess this water ligand. Fig. 1b and results shown in Table I illustrate how globin differences in Hb can alter the shape and position of the Nernst plots. Nernst plots for horse Hb, bottlenose dolphin Hb, and human Hb (Hb A 0 ) have mid-point (n1 ⁄2 ) slopes greater than 1, consistent with a cooperative redox process. The redox curves of dolphin Hb and horse Hb are equivalent and shifted to lower potentials relative to Hb A 0 , indicating their increased ease of oxidation. The crystal structures and primary sequences of these three Hbs show that they have few differences in the active site region (3,44,45). Interestingly, the presence of a serine at the position ⌭14␤ in dolphin and horse Hbs has been identified (3,46). This residue, outside the first coordination sphere of the iron, replaces an alanine found in Hb A 0 . The presence of this serine appears to create a more hydrophilic distal heme pocket and thereby increases the ease of active site Hb oxidation. These intrinsic differences in redox potential may be distinguished from those brought about by allosteric effectors, as documented in the figures that follow. A quantitative summary of E1 ⁄2 values under various experimental conditions is reported in Tables I and II. Fig. 2 illustrates the effects of relatively low levels of anions (0.2 M Cl Ϫ or P i , and 1:10 heme:IHP) on the oxidation of (a) Hb A 0 , (b) bottlenose dolphin Hb, and (c) spot fish Hb. The shifts in redox potential (E1 ⁄2 ) illustrated indicate that low levels of anions stabilize the reduced state (resulting in decreased ease of oxidation). These Nernst plots also illustrate the strong allosteric effect of IHP on the redox behavior of Hb. This highly charged anion brings about a strong positive shift of the E1 ⁄2 , with a negligible shift of the initial stages of oxidation and large shifts of the final stages of oxidation. This result is consistent with T state stabilization, where the more highly charged phosphates are typically stronger effectors than the monovalent Cl Ϫ anion. The decreased n1 ⁄2 is consistent with an incomplete transition of metHb to the R state. The results of Fig. 2 are paralleled by anion-induced decreases in the O 2 affinities of these Hbs (data not shown) and demonstrate that the effects of low levels of anionic effectors on the redox and oxygenation process are not species-specific. Fig. 3 takes the general observations in Fig. 2 one step further by documenting the alterations of the Nernst plots for Hb A 0 in the presence of varied levels of Cl Ϫ . Three types of behavior can be observed in Fig. 3 and will be discussed further. First, as shown in Fig. 3a, Hb A 0 becomes less easily oxidized as Cl Ϫ levels increase from 0 to 0.2 M. Over this range of anion levels there is a progressive shift of the mid-point of the curves that defines the redox potential of the protein (E1 ⁄2 ). There is also a shift to a more positive potential in the final stages of oxidation (where log[ox]/[red] Ն 1), showing that the R-like state of metHb has been destabilized. There is no change in the initial stages of oxidation (where log[ox]/[red] Ͻ Ϫ1), showing that the electronic properties of the T state of deoxyHb are unaltered by low levels of Cl Ϫ . These alterations are consistent with a simple two state MWC model, where Cl Ϫ acts as a simple allosteric effector and causes redox shifts as it binds preferentially to the central cavity of Hb in its T state. As a consequence of anion-induced shape changes, the sigmoidal curves flatten, with somewhat lower n1 ⁄2 values, as the concentration of Cl Ϫ increases.

RESULTS
An unexpected result of this intensive study of anion effects on Hb is shown in Fig. 3b. There is, rather remarkably, a reversal of the trend in E1 ⁄2 as Cl Ϫ levels are increased from 0.2 to 0.4 M. The entire Nernst plot shifts toward a more easily oxidized species. Both top (R state) and bottom (T state) parts of the curve shift toward more negative potentials. It is significant that similar shifts are associated with an increased hydrophilicity of the heme environment (Fig. 1b).
Finally, as shown in Fig. 3c, between 0.4 and 1.0 M [ Cl Ϫ ] a shift in the final stage of oxidation is observed but in the opposite direction as compared with Fig. 3a. That is, the top of the curves (where log [ox]/[red] Ն 1) shift to a more negative potential, suggesting a shift in the relative stability of the R-like state. It is noteworthy that over this concentration range the slopes of the curves (n1 ⁄2 values) increase as the concentration of Cl Ϫ increases. Fig. 4a plots the redox potential (E1 ⁄2 ) of Hb A 0 and representative Mbs as a function of varied levels of three anionic effectors (Cl Ϫ , NO 3 Ϫ , and ClO 4 Ϫ ). We show that the concentrationdependent reversal of E1 ⁄2 observed for Cl Ϫ interactions with Hb also occurs for other small anionic effectors. The same trends are evident for all three effectors, with sensitivity to anion concentration decreasing in the order NO 3 Ϫ Ͼ Cl Ϫ Ͼ Ͼ ClO 4 Ϫ . This ease of anaerobic oxidation reversal is clearly associated with the cooperative Hb system, since, as also shown in Fig. 4a, this trend is not observed in the case of various monomeric Mbs. The relatively small and progressive decrease in E1 ⁄2 for these Mbs is like that shown in Fig. 3c for Hb exposed to Cl Ϫ concentrations of 0.4 to 2.0 M, indicating that the larger anioninduced shifts observed in tetrameric Hbs are superimposed on the background effects seen with the noncooperative Mb systems.
The inset of Fig. 4a shows that there is a competition between IHP and Cl Ϫ for sites on Hb A 0 as Cl Ϫ levels are raised above 0.2 M. This is a significant result with respect to the globin binding site responsible for the reversal of redox potential at the active site, because the dominant IHP binding site is known to be in the 2,3-diphosphoglycerate binding region, in the central cavity between the two ␤ chains (47). Although IHP is a much stronger allosteric effector (and produces larger redox shifts), the presence of Cl Ϫ gradually attenuates the IHP effect until the system returns to the behavior observed when no IHP is present. Competition between IHP and Cl Ϫ or superimposition of an opposing low affinity Cl Ϫ effect onto the "classical" allosteric effect was previously reported in oxygenbinding (48,49).
Because the dimer-tetramer equilibrium for Hb A 0 is anion concentration-dependent, we looked at whether dimer formation could be the cause of the observed variation in E1 ⁄2 between 0.2 and 0.4 M Cl Ϫ ([heme] ϭ 80 M). We investigated both the 10% oxidized level where hemoglobin is mostly in a T state conformation (a tightly associated tetramer) and the 50% oxidized level where the protein is present as a mixture of R-like and T conformations. Fig. 4b represents the same set of data as illustrated in Fig. 4a only for the 10% oxidized level of Hb A 0 . Comparison of these two sets of data reveals identical trends, reinforcing the idea that tetramer-dimer dissociation is not responsible for this unusual E1 ⁄2 reversal (23,29). It is important to note as well that the same pattern of concentration-dependent anion effects are observed in Spot Hb, where previous studies have shown that dimer formation is greatly reduced compared with that in Hb A 0 and negligible even at 0.5 M CO-Hb (in heme) (50). Finally, the difference in extinction coefficients for the Hb dimer and tetramer would preclude the maintenance of all five isosbestic points that we observe during the course of our spectroelectrochemical titrations. These findings support our assertion that the trends observed are not linked to a tetramer-dimer equilibrium position but to a change in the electronic properties of the hemoglobin tetramer. Fig. 5a compares the influence of Cl Ϫ on Hb A 0 , and Hb Presbyterian (G10␤ Asp 3 Lys). The latter Hb variant possesses an additional positive charge in its central cavity over Hb A 0 and was therefore chosen to evaluate the influence of increasing electrostatic interactions around the ␤-␤ interface (48, 51). Fig. 5a shows that the dramatic reversal of electron affinity that occurs with Hb A 0 at about 0.2 M Cl Ϫ also occurs in Hb Presbyterian but at about 0.6 M Cl Ϫ . This implies that more Cl Ϫ is required to compensate for the extra positive charges in the central cavity in this variant. The redox properties of the variant are also significantly more sensitive to the presence of the anion, so that there is a larger maximum redox shift before reversal. Fig. 5b illustrates that structurally different Hbs can show anion selectivity in their redox potential shifts. In the highly pH-dependent (Root effect) Hb of the Spot fish, Leiostomus xanthurus, the phosphate effect on the redox potential exceeds that associated with Cl Ϫ (also see Fig. 2c and Table I). A similar anion selectivity is also seen in oxygen binding studies. 2 This Root effect fish Hb is fully responsive to phosphates but has a blocked ␣-N terminus and thus lacks one of the residues of importance in oxygen-linked Cl Ϫ binding to human Hb (52). DISCUSSION This study of anion effects on the redox behavior of representative Hbs and Mbs provides further insight into the steric and electronic consequences of anion-globin interactions. The concentration-dependent effects observed underscore the need for care in comparing the functional properties of heme proteins, because the nature and type of anions in the experimental medium can have a dramatic effect on the redox process and other measures that probe the electronic properties of the active site iron atoms of these proteins.
We show in Fig. 5a that the extra positive charges in the central cavity of Hb Presbyterian has the effect of strengthening the Cl Ϫ effect on its redox properties, as well as increasing the level at which the reversal of electron affinity occurs. The   (5) 138 (6) reverse trend in E1 ⁄2 with anion concentrations above 0.2 M is unequivocally a result of the tetrameric structure of Hb because this phenomenon is not observed for the Mbs. Because low anion concentrations stabilize the T structure and shift the E1 ⁄2 positive, it follows that a reversal of this trend at higher anion concentrations is either due to filling lower affinity anion binding sites or to an increase in the hydrophilicity of the heme pocket (e.g. through greater solvent exposure) or both. We have drawn attention to a reversal of the trend in active site E1 ⁄2 that occurs in Hb A 0 when Cl Ϫ levels are raised above 0.2 M. Very similar patterns, with maxima near 0.2 M Cl Ϫ at neutral pH, were observed in earlier studies of the influence of Cl Ϫ on the pH dependence of Hb oxygenation. To explain these earlier results, the T state of deoxyHb was inferred to have a higher Cl Ϫ affinity than the R state of liganded Hb (31). In support of this conclusion, Chiancone and co-workers (23,29,30) showed, using NMR techniques, that Cl Ϫ binds differently to deoxyHb and oxyHb and that the binding is proton-linked. The NMR studies showed the presence of binding sites on deoxyHb that have a 10-fold higher Cl Ϫ affinity (about 100 M Ϫ1 ) than in oxyHb (about 10 M Ϫ1 ). Some, but not all, high affinity Cl Ϫ binding observable by NMR was lost in the presence of the competitive anion IHP that binds in the central cavity of the Hb tetramer.
It is reasonable to conclude that our redox studies show the similar pattern of Cl Ϫ concentration dependence as found by the workers cited above because deoxyHb has a higher Cl Ϫ affinity than R-like metHb. The shifts of E1 ⁄2 as Cl Ϫ levels increase from 0 to 0.2 M (as shown in Fig. 3a) are classic representations of allosteric effects brought about by preferen-tial binding of anions to deoxyHb. As Cl Ϫ levels are raised from 0.2 to 0.4 M at neutral pH, the occupancy of lower affinity Cl Ϫ binding sites on metHb (possibly aided by increased hydrophilicity of the heme pocket) reverses the trend in E1 ⁄2 of the heme (Fig. 3b). This effect is seen with other small anions, with sensitivity to anion concentration decreasing in the order NO 3 Ϫ . We note that stronger binding to deoxy (T state) Hb and greater redox shifts are associated with the polyphosphate anion IHP than for Cl Ϫ or other small anions. The reversal observed with small anions was not seen with the highest levels of IHP used in our studies (about 12 mM) but might occur at higher IHP levels.
Both Hb and Mb show small, progressive changes in redox potential as Cl Ϫ or NO 3 Ϫ levels are raised above 0.4 M. These redox changes are thus not unique to Hb tetramers. Redox changes in this range of anion concentrations are expected to accompany tertiary level changes in subunits of the Hb tetramer. The anion binding responsible for these redox shifts appears to be correlated with the low affinity class of Cl Ϫ binding sites, documented by NMR techniques as having about 100-fold lower Cl Ϫ affinity than the relatively high affinity sites on R state Hb and 1000-fold lower Cl Ϫ affinity than even higher affinity sites on deoxyHb (23).
It is significant that the redox shifts observed for Mbs and Hbs at high anion levels (Figs. 3b and 4a) are like those associated with more hydrophilic heme pockets (Fig. 1). This similarity supports the concept that anion-globin interactions are capable of creating a more hydrophilic environment for the heme group. This was hypothesized by Caughy and co-workers (53,54) in previous studies on anion effects on auto-oxidation of air-equilibrated Hb. These workers proposed that Cl Ϫ is a weak nucleophile that mediates the release of superoxide and could be responsible for metHb and superoxide formation under physiological conditions. A general (weaker) electrolyte effect on the autoxidation process documented by Caughy and coworkers (53) was proposed to have its origin in an electrolyte induced relaxation of the heme pocket allowing freer access of water to the oxygen binding site. Elegant studies using sitedirected mutagenesis methods to modify the heme pocket of Mb led Brantley and co-workers (55) to a similar conclusion, i.e. that the rate of auto-oxidation could be dramatically increased by increasing the polarity of the heme pocket or by increasing the net anionic charge at the protein surface in the vicinity of the heme.
The anion-induced reversal of Hb E1 ⁄2 trends as Cl Ϫ levels are raised above 0.2 M is not seen in oxygen binding curves, where the dominating consequence of increased anion concentration can be inferred to be largely a steric rather than an electronic effect. The sharp contrast between anion effects on oxidation and oxygenation processes lends additional support to our hypothesis (26) that anion-induced restrictions of the frequency or extent of conformational fluctuations of the Hb molecule play a large role in control of its oxygen affinity. In light of the work cited above, it is tempting to speculate that Cl Ϫ levels above 0.2 M make water more accessible or more tightly held in the heme pocket, thereby accounting for the greater ease of oxidation. The resulting increase in hydrophilicity in the distal heme pocket could decrease oxygen affinity via steric hindrance effects, while increasing the ease of oxidation, as is experimentally observed. Further studies are underway to explore this possible explanation for the concentration-dependent anion effects observed.