Search

The triplet probe eosin-5-maleimide (EMA) is a specific inhibitor of anion transport mediated by the erythrocyte membrane protein, band 3. It was previously shown that the eosin moiety is located close to the anion binding site when EMA is covalently bound to band 3 [Pan, R.-j., and Cherry, R. J. (1995) Biochemistry 34, 4880-4888]. In the present study the electrostatic properties and membrane sidedness of the EMA binding site of band 3 were further investigated by triplet state quenching. A series of stable nitroxyl free radicals, which are characterized by different charges, and I- were used as the quenchers. Time-resolved laser spectroscopy was employed to measure the triplet lifetime of EMA. It was found that the quenching reaction between the quenchers and band 3-bound EMA follows a linear Stern-Volmer plot. The quenching rate constants (Kq) of the quenchers are in the order of NH3+-TEMPO (Kq = 6.34 x 10(6) M-1 s-1) > TEMPO-Choline+ (Kq = 2.18 x 10(6) M-1 s-1) > TEMPO (Kq = 1.13 x 10(6) M-1 s-1) > I- (Kq = 2.46 x 10(5) M-1 s-1) > pyrroline-COO- (Kq = 2.18 x 10(4) M-1 s-1). Experiments with resealed ghosts and inside-out vesicles revealed that negatively charged quenchers can only access the EMA binding site from the extracellular side of the membrane while the positively charged quenchers acted from the cytoplasmic side. The ionic strength dependence of the quenching rate constants and the effects of pH on the quenching reaction were also studied. For both TEMPO-Choline+ and I-, the Kq values decreased as the ionic strength increased, but quenching by TEMPO was independent of the ionic strength variation over the same range. It was also found that at lower pH, the I- quenching rate constant increases but the TEMPO-choline+ quenching rate constant decreases. In both cases, the dependence of quenching on pH exhibited an apparent pKa of about 6.5, which suggests the involvement of one or more histidine residues. This notion gained further support from the finding that modification of His residues of band 3 by DEPC reduced I- quenching at pH 6. On the basis of these results, it is proposed that eosin is located in the anion transport channel such that it is accessible from both sides of the membrane. Histidine residues, which have previously been proposed to lie in the anion channel, probably are located on either side of the eosin probe where they contribute to electrostatic interactions which determine the Kq values for the charged quenchers.

Rotational diffusion of eosin-5-maleimide-labeled band 3 was measured in erythrocyte membranes at pH 9.4-10.4. Band 3 was found to be more mobile in this pH range than at pH 7.5. Similar results were obtained with spectrin-actin-depleted membranes, where it was further shown that ankyrin is the only detectable protein released from the membrane at pH 10. Further experiments were performed at pH 7.5 to investigate the effects of rebinding purified ankyrin and/or band 4.1 to ghosts stripped of skeletal proteins. Ankyrin was found to reduce band 3 rotational mobility, but band 4.1 had no effect. A fluorescence binding assay revealed that fluorescein isothiocyanate-labeled ankyrin had similar binding parameters to those reported previously using 125I labeling. Finally, the rotational mobility of purified band 3 reconstituted into lipid bilayers was determined before and after ankyrin binding. The results of these reconstitution experiments were globally analyzed, assuming the existence of two populations of band 3 with different correlation times. The faster correlation time is consistent with that expected for either dimers or compact tetramers of band 3. Ankyrin binding reduces the proportion of band 3 contributing to the faster component. This result demonstrates that ankyrin promotes the association of band 3 into more slowly rotating complexes independently of any other components of the erythrocyte membrane. It has been reported that ankyrin contains two binding sites for band 3 [Michaely, P., & Bennett, V. (1995) J. Biol. Chem. 270, 22050-22057]. The results of the present study are thus explained by the ability of ankyrin to cross-link band 3 into larger diameter complexes. Cross-linking by ankyrin in part accounts for the slow components in the anisotropy decays of band 3 in the erythrocyte membrane. Other factors which probably influence band 3 aggregation include the membrane "fluidity" and protein concentration.

The interaction between eosin-5-maleimide (EMA), an inhibitor of the anion transport protein, band 3, and I-, a transportable substrate, was investigated by fluorescence quenching. The Stern-Volmer plot for the quenching reaction between EMA-labeled band 3 and I- exhibits downward curvature both in human erythrocyte ghosts and in purified band 3. The quenching reaction is insensitive to the viscosity of the bulk phase. The shape of the Stern-Volmer plot becomes more linear with increasing temperature. Following the approach of Blatt et al. [(1986) Biophys. J. 50, 349-356], we have developed a binding-diffusion model which is in good agreement with the quenching data. The model supposes that EMA is located in a compartment or "pocket" in band 3 which is separate from the bulk phase and contains a binding site or sites for the quencher. Quenching of band 3-bound EMA fluorescence by I- is inhibited by DIDS and by the transportable anions Cl-, HCO3-, and Br-. Analysis of these experiments yields dissociation constants for the anions which are in reasonable agreement with those determined from transport kinetics and by NMR. We thus deduce that the quencher binding site is the anion binding/transport site on band 3. We propose that EMA is located in the wall of the anion access channel such that it does not inhibit anion binding. The methods described in this report should facilitate detailed studies of anion binding to the transport site on band 3 under a variety of experimental conditions.

The triplet lifetime and absorption anisotropy decay of eosin-labeled band 3 was measured in resealed erythrocyte ghosts. Membrane potentials were generated by the addition of valinomycin in the presence of a K+ gradient. Neither negative nor positive membrane potentials had any detectable effect on the rotational diffusion of band 3 nor on the eosin triplet lifetime. The membrane potential did, however, affect quenching of the eosin triplet state by I- and TEMPO (2,2,6,6-tetramethylpiperidine-N-oxyl). Quenching was enhanced by a negative membrane potential (negative inside) and reduced by a positive membrane potential. In addition, it was found that a negative membrane potential enhanced the efficiency of eosin labeling of band 3 in intact erythrocytes. A positive membrane potential had the opposite effect. These results indicate that the eosin binding site on band 3 becomes more accessible to the extracellular aqueous phase in the presence of a negative membrane potential and less accessible in the presence of a positive membrane potential. Quenching by I- and TEMPO of the triplet state of eosin-labeled band 3 was further investigated as a function of pH. Quenching by TEMPO and its dependence on membrane potential were relatively insensitive to pH. In contrast, the rate of quenching by I- showed a marked decrease over the range pH 5.5-9.5. Moreover, the effect of a negative membrane potential on I- quenching also varied with pH. These results are discussed on the supposition that the eosin probe is located in the anion access channel of band 3.(ABSTRACT TRUNCATED AT 250 WORDS)

Rotational mobility of the lipoyl domain of a number of 2-oxoacid dehydrogenase complexes was investigated by transient dichroism after the domain had been specifically labeled with the triplet probe eosin-5-maleimide. Complexes investigated included pyruvate dehydrogenase complexes from Bacillus stearothermophilus, ox heart, and Escherichia coli (in which the E2 component had been genetically engineered to contain one lipoyl domain) and 2-oxoglutarate dehydrogenase complexes from ox heart and E. coli. Measurements were also performed with ox heart pyruvate dehydrogenase and 2-oxoglutarate dehydrogenase complexes specifically labeled on E1. Anisotropy decays were recorded in glycerol-buffer solutions of varying viscosity and at different temperatures. For E2-labeled complexes, the decays were found to be multiexponential, and the fastest correlation time was considerably shorter than expected for tumbling of the whole complex. This fast correlation time was absent from E1-labeled complexes and was assigned to independent motion of the lipoyl domain. Plots of the fast correlation time against eta/T showed a surprisingly weak dependence on viscosity and extrapolated to a time of 30-40 microseconds at zero viscosity. To explain this result, a model is proposed in which the lipoyl domain is in equilibrium between "free" and bound states. The time of 30-40 microseconds is shown to correspond to 1/koff, where koff is the rate constant for dissociation of the domain from binding sites on the complex. This dissociation phenomenon only contributes to the anisotropy decay when the viscosity of the solution is sufficiently high to slow the tumbling of the whole complex to times that are long in comparison to 1/koff.

Spectrin was purified from human erythrocytes and labeled with the triplet probe eosin-5-maleimide. Rotational diffusion of spectrin was investigated by observing transient dichroism following flash excitation of the probe. Measurements were performed at 4 degrees C in solutions of varying viscosity and with spectrin rebound to spectrin/actin-depleted erythrocyte membranes. In solution, complex anisotropy decays were observed which could not be satisfactorily fitted by the equations for a rod-shaped molecule of appropriate dimensions. When spectrin was rebound to the erythrocyte membrane, a decay in the anisotropy was still present but was markedly less sensitive to solution viscosity and flatter at longer times. In order to overcome the objection that the cytoskeleton is only partially reconstituted when spectrin is rebound, a method was developed for labeling spectrin with eosin-5-maleimide in situ. Anisotropy decays for these labeled membranes exhibited features similar to those obtained for spectrin labeled in solution and subsequently rebound. Taken together, the results provide good evidence for segmental motion of spectrin when incorporated into the erythrocyte cytoskeleton. Upon increasing the temperature, the initial anisotropy ro for both rebound and in situ labeled spectrin decreases, and above 30 degrees C the measured anisotropies are small. Thus, at physiological temperature the probe is almost completely randomized by motions with correlation times less than 10 microseconds.

Ca2+,Mg2+-dependent adenosine 5'-triphosphatase (ATPase) in sarcoplasmic reticulum vesicles is labeled with the triplet probe, 5-iodoacetamidoesin. Rotational mobility of the ATPase is investigated by measuring flash-induced transient dichroism of the eosin probe. The absorption anisotropy measured 20 mus after the exciting flash is found to be small at 37 degrees C but increases considerably with decreasing temperature and upon fixation with glutaraldehyde. A purified Ca2+,Mg2+-dependent ATPase preparation partially depleted of membrane lipids exhibits similar properties. The low value of the anisotropy at 37 degrees C is due to the existence of a fast motion which in part is assigned to independent segmental motion of the protein. This internal flexibility of the ATPase may have considerable significance for the functional properties of the enzyme. At times longer than 20 mus, the anisotropy decays with a time constant which varies from approximately 90 mus at 0 degrees C to approximately 40 mus at 37 degrees C. This decay is assigned to rotation of the ATPase about an axis normal to the plane of the membrane. There is some evidence for self-aggregation of the protein at lower temperatures.

A new approach to the study of molecular protein interactions in biological membranes is presented. The technique is based on measuring the rotation of a membrane protein in the presence and absence of specific antibodies directed toward a purported complex partner. As a first illustration of the method, the putative association of band 3 with glycophorin A in the human erythrocyte membrane was investigated. The rotational diffusion of band 3 was strongly reduced following cross-linking of glycophorin A with divalent antibodies. However, little or no effect on band 3 rotation was produced by monovalent antiglycophorin A Fab fragments, antispectrinor nonspecific antibodies, ruling out major effects on band 3 mobility due to steric hindrance, unspecific antibody adsorption, or transmembrane interactions involving spectrin. It is concluded that immobilization of band 3 by antiglycophorin A antibodies is directly caused by cross-linking of a preexisting band 3-glycophorin A complex in the human erythrocyte membrane.

Band 3 rotation in the human erythrocyte membrane is measured by observing flash-induced dichroism of eosin probes. The decay of the absorption anisotropy is found to be strongly dependent on temperature. The results are analyzed on the assumption that rotation of band 3 only occurs about the membrane normal. It is deduced that both fast and slowly rotating forms of band 3 coexist in the membrane. The equilibrium between these forms is temperature dependent, the slowly rotating species becoming increasingly dominant as the temperature is reduced. Plots of the fractional distribution of the different species against temperature show a marked change of slope at around 37--40 degrees C. The effects are essentially reversible over the range 1--45 degrees C and independent of the presence of the spectrin--actin network. The results could be due to temperature-dependent protein--protein associations mediated either by a protein conformational change or by lipid phase segregation. In further experiments, the cholesterol content of the erythrocyte membrane is varied by incubation with lipid vesicles. No significant changes in the rotational diffusion of band 3 are observed following variation of membrane cholesterol/phospholipid mole ratios over the range 0.34--1.66. This is a surprising result in view of the well-known effects of cholesterol on lipid fluidity.

Experiments have been performed to test the validity of measurements of slow rotational diffusion using eosin isothiocyanate as a probe. When eosin-protein conjugates are covalently bound to Sepharose 4B, the dichroism of the flash-induced absorption transient does not decay with time. The magnitude of the dichroism is smaller than for eosin in solid solution. These results, together with measurements with highly viscous solutions, demonstrate the existence of a rapid but restricted independent motion of the eosin probe. This causes a partial loss of pf dochroism leaving a residual dichroism whose time dependence is determined by the rotational motion of the protein. Rotational correlation times in the microsecond-millisecond time range have been determined for eosinprotein conjugates dissolved in glycerol-water mixtures of varying viscosity. The measured values are in good agreement with theoretical predictions and show the correct dependence on the viscosity of the medium and the size and state of association of the protein.

Bacteriorhodopsin (BR) was incorporated into large unilamellar dimyristoyl- and dipalmitoylphosphatidylcholine vesicles (100--300-nm radius). The effect of this intrinsic membrane protein on the order and dynamics of the lipid and on the cooperativity and transition temperature (Tc) of the gel to liquid-crystalline phase transition was investigated as a function of the lipid:protein ratio (L/BR). The lipid phase transition induces protein segregation. Above Tc, bacteriorhodopsin is in the monomeric state. Below Tc, BR is aggregated in the same hexagonal lattice as in the purple membrane (PM). In this reconstituted system, BR has a photochemical cycle similar to that in the PM and is active as a light-driven proton pump. The lipid phase transition which was monitored by using the steady-state anisotropy of the fluorescent probe 1,6-diphenyl-1,3,5-hexatriene (DPH) broadens with decreasing L/BR but occurs at approximately the same Tc. Below Tc, the fluorescence anisotropy of DPH is quite high (0.35) and independent of the L/BR. Above Tc, however, the anisotrophy increases markedly with decreasing L/BR. It was recently pointed out that the fluorescence anisotropy of probes like DPH contains information not only on the dynamics (correlation times) but also on the order parameters of the lipids [Heyn, M. P. (1979) FEBS Lett. 108, 359--364]. The most likely explanation of the observed increase in anisotropy above Tc is that the perturbation of the lipid bilayer caused by the incorporation of BR leads both to an increase in order and to a slowing of the rotational diffusion of the lipids (increased viscosity). In agreement with this latter dynamical effect, the rotational diffusion constant of BR itself decreases above Tc with decreasing L/BR. Above Tc, the membrane viscosity as determined from the rotational diffusion constant of BR is at least 1.5 times larger than that obtained from the fluorescence depolarization of DPH. The formation of the BR lattice as a function of temperature was followed by using the circular dichroism (CD) exciton effect together with measurements of the rotational diffusion of BR. Both methods show similar transition curves for the protein crystallization whose midpoints, however, occur several degrees below Tc.

Slow rotational diffusion may be investigated by measuring the decay of dichroism of flash-induced transient absorption changes of suitable probes. The preparation of the covalent ""triplets" probe eosin isothiocyanate is described together with investigations of spectroscopic properties of eosin-protein conjugates. Triplet state lifetimes of air-equilibrated solutions of eosin-protein conjugates are in the order of 10 mus, demonstrating that the probe is protected from oxygen quenching by the protein. Experiments with the sucrase-isomaltase complex from small intestine show that its enzymatic activity is little affected by binding up to 2 mol of eosin/mol of protein.

Band 3, the anion transporter of the human erythrocyte membrane, has been purified and reconstituted into phospholipid vesicles of varying composition. Rotational diffusion of band 3 in these vesicles was measured by observing flash-induced transient dichroism of an eosin triplet probe covalently bound to the protein. In egg phosphatidylcholine vesicles of high lipid/protein ratio at temperatures well above the gel to liquid-crystalline phase transition, the absorption anisotropy decays to a constant value of 12 +/- 1% (expressed as a percentage of the initial anisotropy). However, higher values of the residual anisotropy and a slower decay are observed upon decreasing the temperature, even though the lipids remain in a liquid-crystalline phase. A similar effect is observed upon decreasing the lipid/protein ratio at constant temperature. It is concluded that self-association of band 3 occurs which is dependent on temperature and protein concentration in the bilayer. It is very probable that similar effects occur in the erythrocyte membrane, where a strong temperature dependence of band 3 rotational mobility is also observed. Reconstitution of band 3 into dioleoylphosphatidylcholine vesicles yields results similar to those obtained with egg phosphatidylcholine. When dimyristoylphosphatidylcholine is used for reconstitution, band 3 is immobilized below the lipid phase transition and remains partially associated above the transition. A precise analysis of the anisotropy decay curves is hampered by the presence of multiple rotating species.(ABSTRACT TRUNCATED AT 250 WORDS)

The cholesterol/phospholipid mole ratio (C/P) in the human erythrocyte membrane was varied by incubating cells with liposomes. The rotational mobility of band 3 proteins was measured in these membranes by observing flash-induced transient dichroism of the triplet probe eosin maleimide. Measurements were performed with membranes in which associations of band 3 with cytoskeletal proteins were removed by mild proteolysis with trypsin. It was found that decreasing C/P resulted in a more rapid decay of the flash-induced anisotropy. The anisotropy decay curves were analyzed by curve-fitting procedures, which indicated the existence of different sized small aggregates of band 3. The changes in the decay curves with varying C/P can be explained by an effect of cholesterol on the size distribution of these aggregates. The experiments suggest a possible role of cholesterol in regulating associations between integral membrane proteins.