Fluorescent effector as a probe of the allosteric equilibrium in methemoglobin.

A fluorescent analogue of diphosphoglycerate (DPG), hydroxy-pyrenetrisulfonate (HPT), was used as a probe of the allosteric equilibrium of methemoglobin. Like DPG, HPT binds, one per tetramer, with a higher affinity to deoxyHb than to oxyHb. Once bound, the HPT fluorescence is quenched by energy transfer to the hemes. HPT can thus serve as a probe of the conformational state of the hemoglobin tetramer: a higher quenching indicates a stronger binding and therefore, more of the deoxy conformation. Since HPT binds to the same site as DPG, it can be displaced by DPG in order to determine the fluorescence intensity of the free HPT under the same conditions, to correct for the inner filter effect. The high spin ferric ligands, such as water and F, showed less fluorescence (more of the deoxy state) than low spin cyano-metHb. The aquo-metHb samples showed a reversion to the oxyHb conformation above pH 7, as expected due to the acid-alkaline transition forming hydroxy-metHb. Effectors such as bezafibrate, which do not bind to the same site as DPG, show an increase in the deoxy-like characteristics.

Heme-CO binding to tryptophan-containing calmodulin mutants.

The binding of heme-CO to genetically engineered calmodulin containing a single tryptophan residue has been studied. A tryptophan residue was integrated at one of five positions: 26 or 62 of the N-terminal, 81 in the central helix, or 99 or 135 of the C-terminal. As for the wild type, the mutant calmodulins bind four molecules of heme-CO with an average affinity of 1 microM. (i) Homotropic effect. The quenching of the tryptophan fluorescence by energy transfer to the hemes indicates that there is no preference between the N- or C-terminal pockets for heme binding. The quenching is less than expected for a binomial distribution of four sites. This could indicate a lower energy transfer rate due to a specific orientation factor. The weak quenching as a function of the number of hemes bound may also reveal a cooperativity in the heme binding; the data can be simulated assuming two pairs of sites, where each pocket shows a cooperative binding for two hemes. (ii) Heterotropic effect. As observed for the wild type, addition of melittin does not displace the hemes from the mutant calmodulins; the affinity of heme-CO for the calmodulin.melittin complex is higher than that for calmodulin alone. The affinity of heme-CO for native calmodulin is also higher in the presence of trifluoperazine.

Heme as an optical probe for studying the interactions between calmodulin and the Ca(2+)-ATPase of the human erythrocyte membrane.

The heme group was used as an optical probe to study the interactions between calmodulin and its targets: the peptide melittin and the enzyme Ca(2+)-ATPase. As already reported, melittin when present in Tris buffer binds hemin-CN which quenches the tryptophan fluorescence. Addition of calmodulin restores the fluorescence significantly accompanied by a blue shift. We show here that the recovery of fluorescence is very slow and takes about 120 min to become constant. In a hydrophobic buffer, the fluorescence spectrum of melittin is already shifted with a peak at 335 nm and intensity almost 2-fold relative to a similar concentration of melittin in Tris buffer. The quenching of tryptophan fluorescence is lesser in this buffer and further addition of calmodulin fails to restore the fluorescence. This indicates the absence of binding of calmodulin to melittin in hydrophobic conditions. Under similar conditions of hydrophobicity, hemin-CN quenches about 35% of the tryptophan fluorescence of the Ca(2+)-ATPase. The subsequent addition of calmodulin restores about half of the quenched fluorescence. The interaction of calmodulin with the Ca(2+)-ATPase even under hydrophobic conditions suggests its high specificity for the enzyme which may be expected for a physiological target.

Picosecond geminate recombination of CO to the complexes calmodulin*heme-CO and calmodulin*heme-CO*melittin.

Picosecond CO recombination kinetics have been measured after photodissociation of the artificial complexes calmodulin*heme-CO and calmodulin*heme-CO*melittin. These systems show an enhancement of the geminate fraction of kinetics relative to unbound heme-CO, due in part to fast geminate kinetics (tau=50ps for the initial phase), as well as a decrease in the rate of migration of CO away from the binding site. This indicates that calmodulin provides a complete pocket around the heme group. Rather than competing with the hemes for binding to calmodulin, the melittin seems to act as a cap to further enclose the hemes; melittin increases the affinity of calmodulin for heme-CO, but only weakly affects the CO recombination kinetics.

Asymmetric (deoxy dimer/azido-met dimer) hemoglobin hybrids dissociate within seconds.

Double mixing stopped-flow experiments have been performed to study the stability of asymmetric hemoglobin (Hb) hybrids, consisting of a deoxy and a liganded dimer. The doubly liganded [deoxy/cyano-met] hybrid (species 21) was reported to have an enhanced stability, with tetramer to dimer dissociation requiring over 100 seconds, based on a method that required an incubation of over two days. However, kinetic experiments revealed rapid ligand binding to species 21, as for triply liganded tetramers, which dissociate within a few seconds. For the present study, [deoxy dimer/azido-met dimer] hybrids are formed within 200 ms by stopped-flow mixing of dithionite with a solution containing oxyHb and azido-metHb. The dithionite scavenges oxygen, thus transforming oxyHb to deoxyHb, and the [oxy dimer/azido-met dimer] hybrid to the asymmetric [deoxy/azido-met] hybrid (species 21). After a variable aging time of the asymmetric hybrids, their allosteric state is probed by CO binding in a second mixing. As previously observed the freshly produced asymmetric hybrids bind CO rapidly as for R-state Hb. As the hybrids are aged from 0.1 to 10 seconds, the fraction of slow CO binding increases, consistent with a dissociation of the asymmetric hybrid to form the more stable deoxy Hb tetramer which reacts slowly with CO. Control experiments showed a predominantly slow phase for deoxy Hb, and fast rebinding for the symmetric hybrids. The kinetic data can be simulated with a tetramer to dimer dissociation rate for species 21 of 1.5/second at 100 mM NaCl (pH 7.2) and 1.9/second at 180 mM NaCl (pH 7.4). These values are similar to those reported for liganded Hb, as opposed to deoxy (T-state) tetramers which dissociate over four orders of magnitude more slowly. As expected from simulations of dimer exchange, the observed transition rate depends on the initial fractions of oxy- and metHb; this effect is not consistent with a slow R to T transition. These results, showing a lifetime of about one second for species 21, do not support the symmetry rule which is based on an enhanced stability of the asymmetric hybrid.

Asymmetric hemoglobin hybrids.

We have investigated the functional properties of hemoglobin (Hb) valency hybrids, and specifically whether it makes a difference if the oxidized subunits are on the same dimer (asymmetric hybrid) or not. CO recombination kinetics were used to probe the allosteric equilibrium of tetramers with two oxidized subunits. Asymmetric hybrids were prepared by mixing HbCO with a large excess of cyano-metHb; dimer exchange occurs in the order of seconds, producing a population of fully liganded [dimer-CO/dimer-CN] hybrids; these hybrids could then be photolysed to study CO rebinding to the doubly liganded [dimer-deoxy/dimer-CN] species. Before mixing, the HbCO samples typically show 30 to 50% slow phase, characteristic of deoxy (or T-state) Hb. Addition of HbCN to these samples decreased the slow fraction. The higher the ratio of HbCN to HbCO, the less slow phase was observed, with about 5% slow phase at a ratio of 10:1. This would indicate that the photoproduct [deoxy-dimer/ dimer-CN] is not predominantly in the low-affinity (T-state) conformation. We did not observe the difference between asymmetric and symmetric hybrids expected from published studies. The difference between the present flash photolysis results and the published equilibrium studies could be due to a kinetic factor: if the conversion to the T-state is slow after photolysis, then the biomolecular kinetics will reflect the pre-flash conditions of fully liganded (R-state) hemoglobin.

Allosteric transition in triply met-haemoglobin.

Methaemoglobin undergoes a transition to a T-like form at acid pH in the presence of strong effectors such as inositol hexakisphosphate (IHP), as evidenced by spectroscopic and oxidation potential measurements. Since oxygen and CO do not bind to the ferric haems, it is difficult to compare the properties of the R-met and T-met forms with those of ferrous haemoglobin. We have therefore prepared 90% oxidized samples, where the dominant signal for ligand (oxygen or CO) binding is due to tetramers with three met haems. Measurements were made of the oxygen equilibrium curves and CO rebinding kinetics after photodissociation. Without effectors, the partially oxidized samples show mainly R-state properties. Addition of IHP at acid pH induces an increase in T-state behaviour, as indicated by a lower oxygen affinity and a higher fraction of the slow bimolecular component for CO rebinding.

Analysis of hemoglobin oxygen equilibrium curves. Are unique solutions possible?

As a hemoglobin tetramer is oxygenated, it converts from a form with low ligand affinity to a high-affinity structure. This allosteric transition occurs in partially liganded molecules, typically after two or three ligands are bound. As a result of the co-operative nature of the process, the populations of the partially liganded forms are low. The relative proportions and precise properties of these intermediate substrates are therefore difficult to measure and are subject to controversy. The problem is compounded by compensating effects; for example, over-estimation of the oxygen affinity of triply liganded forms will result in under-estimation of the proportion of these species. Specifically, published values for the oxygen affinity of the triply liganded species vary for identical conditions by a factor of more than 6. In analyses based on the highest affinity values, the triply liganded species virtually disappears. However, this affinity is usually calculated from the last few per cent of the oxygenation curve, and this part of the curve is extremely sensitive to the normalization of the data. We conclude that unique solutions for the ligand affinity and substrate populations may be impossible to achieve, and that unusual mechanisms based on particular combinations of parameters, therefore, should be viewed with caution.

Intrinsic activity at the molecular level: E. J. Ariëns' concept visualized.

The concept of using affinity and intrinsic activity to analyze drug interactions with receptors has had a long history in pharmacological studies. In the simplest case, the biological response will be proportional to the amount of drug bound, i.e. its affinity. However, the biological response is also mediated by the ability of a drug when bound to exert its maximum effectiveness. This effectiveness is termed the intrinsic activity. Physicochemical processes have been thought to be at the basis of intrinsic activity. Detailed oxygen and solution binding experiments combined with X-ray crystallographic studies on allosteric effectors to hemoglobin demonstrate that these potential drug agents bind at the same site in hemoglobin with similar binding constants yet shift the allosteric equilibrium and the oxygen affinity of the T-structure by different degrees. Therefore some of the effectors with similar binding affinities for the same site exhibit varying degrees of affectiveness, i.e. they possess different intrinsic activities. The intrinsic activity of the effector is defined as the ratio of the oxygen affinity constant to the T-state with drug/oxygen affinity constant to the T-state without drug (KT+drug)/(KT control). The source of the intrinsic activity appears to be the ability of the effectors to interact with key residues such as Lys99 alpha at the binding site. These results suggest a general molecular mechanism for allosteric effector modulation of hemoglobin function that might be of use in other allosteric enzyme systems.

Very high resolution structure of a trematode hemoglobin displaying a TyrB10-TyrE7 heme distal residue pair and high oxygen affinity.

Monomeric hemoglobin from the trematode Paramphistomum epiclitum displays very high oxygen affinity (P(50)<0.001 mm Hg) and an unusual heme distal site containing tyrosyl residues at the B10 and E7 positions. The crystal structure of aquo-met P. epiclitum hemoglobin, solved at 1.17 A resolution via multiwavelength anomalous dispersion techniques (R-factor=0.121), shows that the heme distal site pocket residue TyrB10 is engaged in hydrogen bonding to the iron-bound ligand. By contrast, residue TyrE7 is unexpectedly locked next to the CD globin region, in a conformation unsuitable for heme-bound ligand stabilisation. Such structural organization of the E7 distal residue differs strikingly from that observed in the nematode Ascaris suum hemoglobin (bearing TyrB10 and GlnE7 residues), which also displays very high oxygen affinity. The oxygenation and carbonylation parameters of wild-type P. epiclitum Hb as well as of single- and double-site mutants, with residue substitutions at positions B10, E7 and E11, have been determined and are discussed here in the light of the protein atomic resolution crystal structure.

Synthesized allosteric effectors of the hemoglobin molecule: a possible mechanism for improved erythrocyte oxygen release capability in hemoglobinopathy H disease.

Patients with the nondeletion genotype of hemoglobinopathy H (HbH or beta4) disease have higher proportions of HbH and more severe tissue hypoxia than patients with the deletion genotype. Because these patients' red blood cells (RBCs) contain mainly two Hb species, HbH and HbA, the high proportion of HbA can be exploited by lowering its oxygen affinity; this would probably increase oxygen delivery to the RBCs and improve the patients' clinical phenotype. Allosteric effectors that induce a low-affinity Hb may be useful in this regard. We investigated the effect of a bezafibrate derivative, RSR-4, on the oxygen affinity of RBCs and purified hemolysates containing HbA and HbH. This allosteric effector crosses RBC membranes and binds reversibly to the alpha-chains of deoxy-Hb, decreasing hemoglobin oxygen affinity. The blood used was obtained from a patient with HbH disease (alphaTSaudi homozygote) whose HbH level was 33.5% as measured by high-performance liquid chromatography. Oxygen binding studies were performed in RBCs and purified hemolysates. RBCs incubated in the presence of 500 microM RSR-4 (2-[[[(3,5-dichloroanilino)-carbonyl]methyl]phenoxy]-2-methylpropi onic acid) in standard conditions (pH 7.4, 0.14 M NaCl, 37 degrees C) displayed an increase in their P50 value from 14.5 to 35.2 mm Hg. Oxygen binding studies in purified stripped hemolysates (pH 7.2, 0.1 M NaCl, 25 degrees C) showed that addition of both 500 microM RSR-4 and 1 mM of 2,3 diphosphoglycerate (DPG) led to an 11-fold decrease in oxygen affinity, whereas the addition of the natural effector DPG or RSR-4 alone produced a 2.7- and 5.7-fold decrease, respectively. In both cases, the oxygen equilibrium curves (OECs) were biphasic due to the presence of the noncooperative, high-oxygen-affinity HbH (beta4) component. After addition of RSR-4, the lower part of the OEC (corresponding to HbH) was not shifted compared with the upper part (corresponding to HbA). These results were confirmed by kinetic studies of CO recombination. Both experiments demonstrated that RSR-4 does not affect beta4 Hb. Our findings provide an experimental model for lowering the oxygen affinity of HbA in HbH-containing cells and suggest that the oxygen delivery capability of the latter would be thereby improved.

Tetramer-dimer equilibrium of oxyhemoglobin mutants determined from auto-oxidation rates.

One of the main difficulties with blood substitutes based on hemoglobin (Hb) solutions is the auto-oxidation of the hemes, a problem aggravated by the dimerization of Hb tetramers. We have employed a method to study the oxyHb tetramer-dimer equilibrium based on the rate of auto-oxidation as a function of protein concentration. The 16-fold difference in dimer and tetramer auto-oxidation rates (in 20 mM phosphate buffer at pH 7.0, 37 degrees C) was exploited to determine the fraction dimer. The results show a transition of the auto-oxidation rate from low to high protein concentrations, allowing the determination of the tetramer-dimer dissociation coefficient K4,2 = [Dimer] 2/[Tetramer]. A 14-fold increase in K4,2 was observed for addition of 10 mM of the allosteric effector inositol hexaphosphate (IHP). Recombinant hemoglobins (rHb) were genetically engineered to obtain Hb with a lower oxygen affinity than native Hb (Hb A). The rHb alpha2beta2 [(C7) F41Y/(G4) N102Y] shows a fivefold increase in K4,2 at pH 7.0, 37 degrees C. An atmosphere of pure oxygen is necessary in this case to insure fully oxygenated Hb. When this condition is satisfied, this method provides an efficient technique to characterize both the tetramer-dimer equilibrium and the auto-oxidation rates of various oxyHb. For low oxygen affinity Hb equilibrated under air, the presence of deoxy subunits accelerates the auto-oxidation. Although a full analysis is complicated, the auto-oxidation studies for air equilibrated samples are more relevant to the development of a blood substitute based on Hb solutions. The double mutants, rHb alpha2beta2 [(C7) F41Y/(G4) N102A] and rHb alpha2beta2 [(C7) F41Y/(E10) K66T], show a lower oxygen affinity and a higher rate of oxidation than Hb A. Simulations of the auto-oxidation rate versus Hb concentration indicate that very high protein concentrations are required to observe the tetramer auto-oxidation rate. Because the dimers oxidize much more rapidly, even a small fraction dimer will influence the observed oxidation rate.

Functional consequences of mutations at the allosteric interface in hetero- and homo-hemoglobin tetramers.

A seminal difference exists between the two types of chains that constitute the tetrameric hemoglobin in vertebrates. While alpha chains associate weakly into dimers, beta chains self-associate into tightly assembled tetramers. While heterotetramers bind ligands cooperatively with moderate affinity, homotetramers bind ligands with high affinity and without cooperativity. These characteristics lead to the conclusion that the beta 4 tetramer is frozen in a quaternary R-state resembling that of liganded HbA. X-ray diffraction studies of the liganded beta 4 tetramers and molecular modeling calculations revealed several differences relative to the native heterotetramer at the "allosteric" interface (alpha 1 beta 2 in HbA) and possibly at the origin of a large instability of the hypothetical deoxy T-state of the beta 4 tetramer. We have studied natural and artificial Hb mutants at different sites in the beta chains responsible for the T-state conformation in deoxy HbA with the view of restoring a low ligand affinity with heme-heme interaction in homotetramers. Functional studies have been performed for oxygen equilibrium binding and kinetics after flash photolysis of CO for both hetero- and homotetramers. Our conclusion is that the "allosteric" interface is so precisely tailored for maintaining the assembly between alpha beta dimers that any change in the side chains of beta 40 (C6), beta 99 (G1), and beta 101 (G3) involved in the interface results in increased R-state behavior. In the homotetramer, the mutations at these sites lead to the destabilization of the beta 4 hemoglobin and the formation of lower affinity noncooperative monomers.

Two mutations in recombinant Hb beta F41(C7)Y, K82 (EF6)D show additive effects in decreasing oxygen affinity.

Based on the properties of two low oxygen affinity mutated hemoglobins (Hb), we have engineered a double mutant Hb (rHb beta YD) in which the beta F41Y substitution is associated with K82D. Functional studies have shown that the Hb alpha 2 beta 2(C7)F41Y exhibits a decreased oxygen affinity relative to Hb A, without a significantly increased autooxidation rate. The oxygen affinity of the natural mutant beta K82D (Hb Providence-Asp) is decreased due to the replacement of two positive charges by two negative ones at the main DPG-binding site. The functional properties of both single mutants are interesting in the view of obtaining an Hb-based blood substitute, which requires: (1) cooperative oxygen binding with an overall affinity near 30 mm Hg at half saturation, at 37 degrees C, and in the absence of 2,3 diphosphoglycerate (DPG), and (2) a slow rate of autooxidation in order to limit metHb formation. It was expected that the two mutations were at a sufficient distance (20 A) that their respective effects could combine to form low oxygen affinity tetramers. The double mutant does display additive effects resulting in a fourfold decrease in oxygen affinity; it can insure, in the absence of DPG, an oxygen delivery to the tissues similar to that of a red cell suspension in vivo at 37 degrees C. Nevertheless, the rate of autooxidation, 3.5-fold larger than that of Hb A, remains a problem.

Beta-lactoglobulin binds retinol and protoporphyrin IX at two different binding sites.

Measurement of tryptophan fluorescence quenching and the excitation energy transfer from tryptophanyl residues to the bound ligand indicates that beta-lactoglobulin binds tightly to hemin and protoporphyrin IX in a ligand-to-protein stoichiometric ratio. The apparent dissociation constants of hemin-beta-lactoglobulin and protoporphyrin IX-beta-lactoglobulin complexes are 2.5 x 10(-7) M and 4 x 10(-7) M, respectively. The addition of beta-lactoglobulin (final concentration = 10 microM, phosphate buffer 50 mM, pH 7.1) to the solution containing retinol and protoporphyrin IX triggers an energy transfer between beta-lactoglobulin tryptophan and protoporphyrin IX as well as between retinol and protoporphyrin IX. The efficiency of energy transfer depends on the distance between the donor (retinol) and the acceptor (protoporphyrin IX). Using the Förster theory, a retinolprotoporphyrin IX distance of 25 A was calculated. These results indicate that retinol and protoporphyrin IX are bound to the beta-lactoglobulin monomer at two different sites.

Heme-CO as a probe of the conformational state of calmodulin.

The interaction of heme-CO with calmodulin, in the presence of calcium, leads to a complex of four heme-CO molecules per protein. No interaction was observed in the absence of calcium. The binding of heme-CO to calmodulin was monitored by the shift in the Soret absorption band from 407 to 420 nm (bound form); the four sites are not spectrally identical. The ligand CO can be photodissociated from the calmodulin-heme-CO complex and the biomolecular recombination kinetics also indicate a heterogeneous mixture. The complex does not bind oxygen reversibly. As calmodulin has only one histidine, the hemes are apparently not bound by the iron atom as in hemoglobin, but are probably loosely associated (Kd = 0.5 microM) in hydrophobic pockets which apparently open when the protein is activated by calcium.

Loss of allosteric behaviour in recombinant hemoglobin alpha 2 beta 2(92)(F8) His-->Ala: restoration upon addition of strong effectors.

In the stereochemical model proposed by Perutz [1], the Fe-His(F8) bond plays a significant role in the allosteric transition in hemoglobin and the resulting cooperativity in ligand binding. When this bond is ruptured, there is a loss in the transmission of the information concerning ligand binding; examples are Hb(NO)4 in the presence of inositol hexakisphosphate (IHP), or nickel substituted Hb hybrids which, despite being liganded, exhibit deoxy-like properties. To study the effects of the loss of the iron proximal histidine bond, we have engineered the alpha 2 beta 2(F8)H92A recombinant Hb. The replacement of the highly conserved proximal histidine F8 residue by an alanine results in a low affinity for the heme group and a loss of the allosteric properties; kinetics of CO recombination after photodissociation show only the rapid bimolecular phase, characteristic of the high affinity R-state. However, a significant amount of deoxy (T-state) kinetics are observed after addition of external effectors such as IHP. The iron-histidine bond is apparently crucial for the heme-heme interaction, but the allosteric equilibrium may still be influenced by external constraints.

Modulation of the oxygen affinity of cobalt-porphyrin by globin.

We have combined two extreme effects which influence the oxygen affinity to obtain a cobalt-based oxygen carrier with an affinity similar to that of human adult hemoglobin (HbA). The goal was to obtain an oxygen transporter with a lower oxidation rate. Exchange of the heme group (Fe-protoporphyrin IX) in Hb with a cobalt-porphyrin leads to a reduction in oxygen affinity by over a factor of 10, an oxygen affinity too low for use as a blood substitute. At the other extreme, certain globin sequences are known to provide a very high oxygen affinity; for example, Hb Ascaris displays an oxygen affinity 1000 times higher than HbA. We demonstrate here that these opposing effects can be additive, yielding an oxygen affinity similar to that of HbA, but with oxygen binding to a cobalt atom. We have tested the effect of substitution of cobalt-porphyrin for heme in normal HbA, sperm whale (SW) Mb (Mb), and high affinity globins for leghemoglobin, two trematode Hbs: Paramphistomum epiclitum (Pe) and Gastrothylax crumenifer (Gc). As for HbA or SW Mb, the transition from heme to cobalt-porphyrin in the trematode Hbs leads to a large decrease in the oxygen affinity, with oxygen partial pressures for half saturation (P(50)) of 5 and 25 mm Hg at 37 degrees C for cobalt-Pe and cobalt-Gc, respectively. A critical parameter for Hb-based blood substitutes is the autoxidation rate; while both metals oxidize to an inactive state, we observed a decrease in the oxidation rate of over an order of magnitude for cobalt versus iron, for similar oxygen affinities. The time constants for autoxidation at 37 degrees C were 250 and 100 h for Pe and Gc, respectively.

Heparin coated poly(alkylcyanoacrylate) nanoparticles coupled to hemoglobin: a new oxygen carrier.

A new generation of drug delivery systems based on heparin-poly(isobutylcyanoacrylate) copolymers has been developed to carry hemoglobin. These copolymers spontaneously form, in water, nanoparticles with a ciliated surface of heparin. These nanoparticles maintain the heparin antithrombotic properties and inhibit complement activation. One ml of nanoparticle suspension can be loaded with up to 2.1mg of hemoglobin, which preserves its ligand binding capacity. This work constitutes the first demonstration of hemoglobin loaded on nanoparticle surface, rather than being encapsulated. With a size of 100 nm, these drug delivery systems make suitable tools in the treatment of thrombosis oxygen deprived pathologies.

Protection by lazaroids of the erythrocyte (Ca2+, Mg2+)-ATPase against iron-induced inhibition.

The calmodulin-stimulated (Ca2+, Mg2+)-ATPase (calmodulin-ATPase) of the erythrocyte membrane is susceptible to oxidative stress induced by heme and non-heme iron. There is a time-and concentration-dependent inhibition of the calmodulin-ATPase activity when the erythrocyte membranes are treated with either iron or hemin. In the present study, the calmodulin-ATPase has been used as a model system to evaluate the protective effects of a vitamin E analog (U83836E) and two 21-aminosteroids (U74500A and U74389G) against calmodulin-ATPase inhibition induced by iron and hemin. The drugs, lazaroids from Upjohn, can significantly protect the enzyme against iron-induced inhibition and also causes a decrease in the formation of thiobarbituric acid reactive species, with an IC50 of 0.4 microM for the drug U83836E and 4 microM for the drug U74500A. The 21-aminosteroid U74389G does not restore iron-inhibited calmodulin-ATPase activity under similar conditions. At higher concentrations (> 100 microM) all three drugs inhibit the calmodulin-ATPase activity. None of the drugs tested can restore hemin-inhibited calmodulin-ATPase activity.