WW781 binds reversibly to red blood cell AE1 and inhibits anion exchange by a two-step mechanism, in which an initial complex (complex 1) is rapidly formed, and then there is a slower equilibration to form a second complex (complex 2) with a lower free energy. According to the ping-pong kinetic model, AE1 can exist in forms with the anion transport site facing either inward or outward, and the transition between these forms is greatly facilitated by binding of a transportable substrate such as Cl. Both the rapid initial binding of WW781 and the formation of complex 2 are strongly affected by the conformation of AE1, such that the forms with the transport site facing outward have higher affinity than those with the transport site facing inward. In addition, binding of Cl seems to raise the free energy of complex 2 relative to complex 1, thereby reducing the equilibrium binding affinity, but Cl does not compete directly with WW781. The WW781 binding site, therefore, reveals a part of the AE1 structure that is sensitive to Cl binding and to transport site orientation, in addition to the disulfonic stilbene binding site. The relationship of the inhibitory potency of WW781 under different conditions to the affinities for the different forms of AE1 provides information on the possible asymmetric distributions of unloaded and Cl-loaded transport sites that are consistent with the ping-pong model, and supports the conclusion from flux and nuclear magnetic resonance data that both the unloaded and Cl-loaded sites are very asymmetrically distributed, with far more sites facing the cytoplasm than the outside medium. This asymmetry, together with the ability of WW781 to recruit toward the forms with outward-facing sites, implies that WW781 may be useful for changing the conformation of AE1 in studies of structure–function relationships.

Introduction

The AE1 (band 3) protein in human red blood cells catalyzes a very tightly coupled one-for-one exchange of Cl for HCO3 that functions together with the enzyme carbonic anhydrase to increase the carrying capacity of the blood for CO2 (Knauf 1989; Jennings 1992b; Timmer and Gunn 1999). Data from a wide variety of experiments, and particularly from demonstrations that each AE1 molecule can transport one and only one anion outward across the membrane under conditions where simultaneous transport of an anion inward is not possible (half-turnover experiments; Jennings 1982; Jennings et al. 1998), strongly indicate that AE1 works by a ping-pong mechanism, as initially proposed by Gunn and Fröhlich 1979. In this model, the AE1 protein can exist in two different forms, one (Ei) in which the transport site faces the inside (that is, the cytoplasm) and another (Eo) in which the site faces the external medium. To explain the tightly coupled one-for-one exchange of anions exhibited by this system, it is proposed that the conformational change from Ei to Eo or vice versa can only take place at a significant rate when a suitable substrate anion, such as Cl or bicarbonate, is bound to the transport site to create the corresponding ECli or EClo forms (see Fig. 4).

In attempting to relate this kinetic mechanism to structural changes in the AE1 protein, two questions are central. First, what changes in protein conformation take place during the transporting conformational change from ECli to EClo, or vice-versa, and second, what is there about binding of certain substrates that lowers the transition state free energy so as to make the transporting conformational change take place at an appreciable rate? This latter effect is apparent even for slowly transported substrates, such as iodide, because the rate of reorientation of the EIi or EIo complex is still many orders of magnitude faster than the rate of the Ei to Eo transition, which is at least 10,000× slower than the rate of Cl exchange (Knauf et al. 1977; Knauf 1989). Although the AE1 protein from various species has been cloned and sequenced, and although there is considerable structural information both from electron diffraction (Wang et al. 1994) and from chemical probe and mutagenesis experiments (Jennings 1992a, Jennings 1995; Müller-Berger et al. 1995; Chernova et al. 1997; Fujinaga et al. 1999; Tang et al. 1999), there is still no information on these important points.

Oxonol dyes, such as WW781 {[3-methyl-1-p-sulfophenyl-5-pyrazolone-(4)]-[1,3-dibutylbarbituric acid]-pentamethine oxonol}, have proven to be very useful chemical probes for investigating the Cl exchange system (Raha et al. 1993). One oxonol, diBA(5)C4 [bis-(1,3-dibutylbarbituric acid)-pentamethine oxonol], is the most potent known inhibitor of the transport system (Knauf et al. 1995). Although it competes with the disulfonic stilbene, 4,4′-dinitro-stilbene-2,2′-disulfonate (DNDS), whose binding is strongly affected by Cl (Fröhlich 1982), Cl does not interfere with the binding of diBA(5)C4. The high affinity of certain oxonols for AE1 suggests that these compounds have very specific interactions with some part of the AE1 protein. If the regions of AE1 with which the oxonols interact are involved in the transporting conformational change, it is reasonable to suppose that oxonols might act as reporters of such changes.

WW781, because of its sulfonic acid moiety, should not penetrate into the interior of red blood cells at an appreciable rate (George et al. 1988). Thus, it is particularly useful as a probe, because its actions can be assigned to external-facing sites. WW781 binds to AE1 by a two-step mechanism, in which an initial complex (complex 1) is rapidly formed, followed by much slower conversion to a second complex (complex 2) in which the WW781 is more tightly bound (Raha et al. 1993). Although this adds complexity to the kinetics of WW781 interaction with AE1, the existence of these two binding steps provides additional opportunities to see whether any of the various aspects of WW781 binding is affected by the AE1 conformation.

In the present work, we have used external WW781 binding, measured by the inhibitory effect on Cl exchange at 0°C, to probe possible changes in AE1 conformation caused by substrate binding and by the reorientation of the transport site from the inward- to the outward-facing form. Brief reports of some of these experiments have been presented in abstract or summary form (Knauf et al. 1990, Knauf et al. 1992; Mendoza et al. 1990).

Materials And Methods

Nomenclature

Because of the complicated nature of the ping-pong model and the need to specify the various forms of AE1 with WW781 bound (see Fig. 4 and Fig. 5), we will use the following standardized nomenclature.

Forms of AE1 and other molecules, such as WW781 and Cl, are indicated in roman type, while dissociation constants and rate constants are in italic.

Dissociation constants for binding equilibria are indicated by Ka,b, where a is the molecule bound (e.g., Cl) and b is the form of AE1 to which it is bound; e.g., Eo. For simplicity, in the subscripts, anions are shown without a − superscript, and WW781 is designated as W. For example, the dissociation constant for binding of Cl to Eo is designated KCl,Eo.

Rate constants for conversion of one form of AE1 to another are designated ka,b, where a is the initial form of AE1 and b is the form resulting from the conversion.

The total of both inward- and outward-facing forms (e.g., Eo and Ei) is designated without a subscript; e.g., E. Dissociation constants for binding to both forms are indicated in a similar manner (e.g., KCl,E). The total of all forms of the AE1 protein is simply designated AE1.

The complex (complex 1) formed by the initial rapid binding of WW781 to an AE1 conformation is designated by • between W and the form of AE1; e.g., W•Eo. The second complex, which is formed more slowly, complex 2, is designated by a -; e.g., W-Eo. The total of the two complexes (e.g., W•Eo + W-Eo) is indicated with no intervening symbol; e.g., WEo. Dissociation constants for WW781 binding after equilibration with both complex 1 and complex 2 are indicated by an eqW subscript; e.g., KeqW,Eo. The dissociation constant for WW781 binding to form complex 1 is designated with a 1W subscript; e.g., K1W,Eo. For general cases, where a dissociation constant can be taken to mean either that for the formation of complex 1 or that at equilibrium, a plain W subscript is used; e.g., KW,Eo. The constant describing the equilibrium ratio between complex 2 and complex 1 for a particular form of AE1 (e.g., Eo) is defined as K2/1W,Eo. Apparent dissociation constants for binding to all forms of AE1 are designated without a second subscript; e.g., K1W or KeqW.

Materials

All chemicals were reagent grade, except for WW781, which was either supplied by Dr. J.C. Freedman (SUNY Health Science Center at Syracuse, Syracuse, NY), or else purchased from Molecular Probes. WW781 stocks were dissolved in ethanol. The total ethanol concentration in all flux and pretreatment media was maintained at 1% vol/vol.

Cell Preparation

Blood was obtained, with heparin as anticoagulant, from apparently healthy volunteers, with informed consent. The blood was washed three times in 150 KH (150 mM KCl, 20 mM HEPES, 24 mM sucrose, pH 6.9 at room temperature with KOH), and the white cell layer was removed during the washes. Cells were made up to 50% hematocrit.

Treatment with WW781 and Flux Measurement

For some experiments, cells were pretreated at a hematocrit of 0.25% with WW781 in 10-ml syringes with 400-μl Eppendorf microfuge tubes attached, as described previously (Raha et al. 1993). Cells were then loaded with 36Cl by incubation in 2–10 μCi/ml 36Cl (ICN Chemical and Radioisotope Division) for 10 min at 0°C. For other experiments, cells were loaded with isotope without pretreatment. 36Cl exchange was begun by resuspending the packed, isotope-loaded cells in 30 ml of ice-cold medium containing various concentrations of WW781 that, for the pretreated cells, were the same as those used for pretreatment. Timed samples were taken by a rapid filtration method, and the rate constant for Cl exchange, k, was determined from the rate of appearance of 36Cl in the medium with time (Knauf and Brahm 1989), by fitting a straight line to the plot of ln[P − Pt] versus time, where Pt is the cpm in an aliquot of cell-free medium at a particular time, and P represents the cpm in a similar aliquot after complete isotope equilibration. Intracellular Cl content was determined as previously described (Knauf and Mann 1986), and the flux was calculated as the product of the rate constant (s−1) times the intracellular Cl content (mmol/kg dry solids).

When extracellular [Cl] was lower than intracellular [Cl], Cl was replaced by sucrose or a mixture of sucrose and K3citrate (with 200 mM sucrose and 25 mM citrate taken as equivalent to 150 mM KCl), as indicated. For experiments in which the external [Cl] was much lower than the [Cl] in the cells, appropriate corrections for the ratio of the amount of Cl in the internal and external compartments were made, as described in detail elsewhere (Knauf et al. 1989). For some experiments, cells were loaded with various [Cl] by using nystatin, as described (Raha et al. 1993). The final washes and fluxes for nystatin-treated cells were in solutions analogous to 150 KH, except with the [Cl] indicated.

Determination of Parameters

The strategies for determining dissociation constants for various conformations of AE1 were as previously described and used in our laboratory (Knauf and Brahm 1989; Knauf et al. 1989). The concentrations of WW781 required to half-inhibit Cl exchange either initially (K1W) or after completion of the binding reactions (KeqW) were determined from fits of the flux (J) versus [WW781] data to the equation for single-site inhibition, which for KeqW is:

 
\begin{equation*}J={J_{{\mathrm{u}}}}/{ \left \left({1+ \left \left[{\mathrm{WW781}}\right] \right }/{K_{{\mathrm{eqW}}}}\right) \right }{\mathrm{,}}\end{equation*}
1

where Ju is the flux with no inhibitor present. Nonlinear least-squares fits were done either with Enzfitter (Elsevier Biosoft) or with Origin (Microcal) software.

A problem, particularly when WW781 is allowed to equilibrate with cells, is that the inhibitory potency is so large that the concentration required for half inhibition is comparable to the concentration of AE1 molecules in the cell suspension, even at the very low hematocrits used. As noted previously (Raha et al. 1993), depletion of the WW781 in the medium because of binding to AE1 can be minimized by keeping the concentration of WW781 high and the hematocrit low. There is, however, a region of WW781 concentrations near the KeqW at which it is difficult or impossible to control the free concentration of WW781 in the medium. Thus, we were forced to take data at higher concentrations, where inhibition is very large, as well as with no inhibitor present (control) (see Fig. 11 for example). To avoid artifacts in fitting such data to , because of the gap in the data at low [WW781], we made several measurements of the control flux, Ju, averaged these, made this a fixed parameter in , and then fitted the equation with only one variable parameter, KeqW (or K1W for initial inhibition conditions), with each point weighted equally. For consistency, this was done for all data presented here, even for conditions where experiments could be done at concentrations near that which causes half inhibition.

Results

Determination of Dissociation Constant for Complex 1 (K1W)

We have previously shown that WW781 (W) binds to AE1 by a two-step mechanism (Raha et al. 1993), which can be described by Scheme I,

 
formula
(Scheme I)

where K1W is the dissociation constant for formation of the initial, rapidly formed complex (complex 1), kW•AE1,W-AE1 is the forward rate constant for formation of the second complex, and kW-AE1,W•AE1 is the backward rate constant for conversion of complex 2 back to complex 1. We have previously shown that, for cells with 150 mM Cl inside and outside, kW•AE1,W-AE1 has a value of ∼1.1 min−1 and kW-AE1,W•AE1 is 0.15 min−1. These values give a half-time for formation of complex 2, at saturating concentrations of WW781, of 0.693/(kW•AE1,W-AE1 + kW-AE1,W•AE1), or 33 s, and the rate should be much slower at WW781 concentrations below K1W. To measure K1W, therefore, we exposed cells to WW781 at 0°C and measured the 36Cl efflux for short times. Under these conditions, there should be little of complex 2 formed, and so the extent of inhibition should largely reflect the amount of complex 1, and hence should depend on the WW781 concentration and the value of K1W. We have previously observed (Raha et al. 1993) that as complex 2 is formed the rate constant for 36Cl efflux gradually decreases, and this is reflected in a decrease in slope of the logarithmic plot of appearance of 36Cl in the medium versus time. We have previously shown (Raha et al. 1993) that, for cells with 5 mM Cl inside and outside exposed to WW781 at the start of the flux measurement, if the flux samples are taken at short times, then there is no substantial deviation of the logarithmic plot from a straight line, indicating that K1W can be measured from inhibition under these conditions without much effect from the formation of complex 2. Fig. 1 shows that this is also true for cells with 150 mM Cl inside and outside. The plots were very linear, with correlation coefficients (r2) usually >0.98. Note, however, that the data at the highest WW781 concentration, shown in Fig. 1, •, does show some systematic deviation toward a concave-downward curve, as expected if some of complex 2 is being formed. The data thus indicate that measurements at short times can give a very good approximation for K1W, but that the apparent values of K1W obtained by this method may be underestimated because of the unavoidable formation of complex 2 during the time required for sufficient radioactivity to appear in the medium for an accurate measurement of the flux.

Fig. 2 shows the unidirectional Cl exchange flux, J, as a function of WW781 concentration under these conditions for cells suspended in 150 mM Cl medium. Note that the decrease in flux as a function of WW781 concentration fits very well to the equation for single-site inhibition, as indicated by the least squares best fit line drawn through the data. The half-inhibition constant from such data should be a good approximation of K1W and certainly gives an upper limit for K1W, because any formation of complex 2 during the flux measurement will cause additional inhibition of transport, corresponding to a higher inhibitory potency and hence a lower value of K1W.

Fig. 3 shows K1W values calculated from several similar experiments in cells that had been equilibrated with media with different Cl concentrations by the nystatin technique (except for those in 150 mM Cl). If WW781 inhibits anion exchange by competing with Cl for the transport (substrate) site, then K1W should increase with increasing [Cl]. In contrast to this prediction, if anything there is a decrease in K1W with increasing [Cl], although the slope of a best-fit straight line (solid line) relating K1W to [Cl] is not significantly different from zero (P = 0.14). The mean values of K1W at various [Cl] are shown in Table, together with SEM and number of experiments. In no case are the K1W values for different [Cl] significantly different (P ≥ 0.26).

Effect of AE1 Conformation on K1W

Because WW781 is not a competitive inhibitor, it may in principle bind to any of the different forms of the anion exchange protein, with the transport site facing inward or outward and unloaded or loaded with Cl (Fig. 4), and each of these forms may bind WW781 by a two-step process, as shown in Scheme I and Fig. 4. If these transport-related changes in AE1 conformation affect the binding of WW781, the affinities of WW781 for the various conformations may be different.

To see whether this is the case for the initial binding step (Fig. 5 A), we measured the effect of short exposure to WW781 on the Cl exchange flux and used Dixon plot intersection techniques (Knauf and Brahm 1989; Knauf et al. 1989) to determine the affinity of WW781 for certain forms of AE1. As previously shown (Fröhlich and Gunn 1986; Knauf and Brahm 1989), the dissociation constant for the binding of an inhibitor such as WW781 to Eo is given by the negative of the intersection point of Dixon plot (1/J versus [WW781]) lines for cells with the same [Cli] and different external Cl concentrations, [Clo]. If measurements are made at early times, the intersection point will largely reflect the dissociation constant for WW781 binding to Eo to form the first complex, complex 1, which is designated K1W,Eo (Fig. 5 A).

Fig. 6 shows data for a typical experiment. The line for fluxes measured in 150 mM Cl medium (▵) intersects that for fluxes in 4 mM Cl medium (○) at a point whose x value is equal to −K1W,Eo, the dissociation constant for binding of WW781 to Eo. It can be seen that this value is much smaller than the concentration required for half-inhibition of the Cl flux in 150 mM Cl medium under similar conditions (K1W), given by the negative of the x intercept for the 150 mM Cl line, which lies far to the left of the y axis in Fig. 6. As previously shown (Knauf et al. 1992), the K1W in 150 mM Cl reflects an average of the dissociation constants for binding of WW781 by all of the forms of AE1, weighted according to their prevalence in 150 mM Cl medium. The fact that this value is much larger than the value of K1W,Eo indicates that the affinity of WW781 for Eo is greater than for the mixture of the other forms of AE1 present in 150 mM Cl medium. This conclusion is further reinforced by the significant difference (P = 0.003) between the K1W value in 150 mM Cl (Table) and the average K1W,Eo value (Table), as well as by the fact that the K1W value for 2 mM [Clo] medium (Table), in which the relative amount of Eo is increased, is significantly smaller than the K1W value for 150 mM [Clo] (P < 0.0001).

Because of the rapid equilibration of ECli and EClo, and because the ratio of EClo to ECli is fixed by the ratio of kECli,EClo to kEClo,ECli (see Fig. 4 and Fig. 5 A), it is not possible to determine the value of K1W,EClo for WW781 binding to the EClo form by any flux technique. This can be done, however, for the Eo form that is associated with a slowly transported anion, such as iodide. As shown in Fig. 6, the dissociation constant for EIo, designated K1W,EIo, is given by the negative of the x value of the intersection point for the data with 4 mM Cl media either without (○) or with (•) iodide (6 mM) present (Knauf and Brahm 1989). This value is only about twice as large as K1W,Eo, suggesting that binding of an external anion to Eo, even a large anion such as I, has a relatively small effect on the binding affinity for WW781.

Similar techniques can be used to measure K1W,Ei and K1W,EIi, as shown in Fig. 7, except that in this case the external Cl concentration is maintained constant at 20 mM. Note that the dissociation constant for initial WW781 binding to Ei, K1W,Ei, given by the negative of the x coordinate of the intersection point of the line for 20 mM [Cli] (○) and that for 50 mM [Cli] (▵), is much larger than K1W,Eo (compare with Fig. 6; note difference in abscissa scale). The corresponding value for K1W,EIi, given by the intersection point of the lines for cells with 20 mM [Cli] and either no (○) or 30 mM (•) [Ii], is similar to K1W,Ei, suggesting little effect of binding of even a large anion on the WW781 affinity for Ei.

Table shows the average K1W values for various forms of AE1 obtained from a number of experiments similar to those shown in Fig. 6 and Fig. 7. The only statistically significant difference in the K1W values is that K1W,Ei is significantly larger (P = 0.03) than K1W,Eo. This is true despite the fact that the K1W,Ei values show considerably greater scatter than the values of K1W,Eo. This is mainly due to the fact that K1W,Ei is larger, so the intersection point occurs at more negative values of WW781, which necessarily are farther from the positive values at which actual measurements of fluxes can be done, but the greater difficulty of loading cells with different [Cl] and the fact that at low [Cli] the ratio of [Cli]/[Clo] often deviates somewhat from unity also may contribute to the scatter in K1W,Ei. The mean K1W,EIo value is somewhat larger than K1W,Eo, but the difference is not significant (P = 0.18), and the single value of K1W,EIi is very similar to the K1W,Ei value, suggesting that binding of anions to either Eo or Ei has only a relatively small effect on the affinity of AE1 for the initial binding of WW781.

As shown by Fröhlich 1982 and Fröhlich and Gunn 1986, the average dissociation constant for an inhibitor binding to both of the unloaded forms of AE1, Ei, and Eo, designated K1W,E for the initial binding step, is given by:

 
\begin{equation*}K_{1{\mathrm{W,E}}}=\frac{1+A}{ \left \left({1}/{K_{1{\mathrm{W,Ei}}}}\right) \right + \left \left({A}/{K_{1{\mathrm{W,Eo}}}}\right) \right }{\mathrm{,}}\end{equation*}
2

where A is the asymmetry factor for the unloaded transporter (without anions bound), defined as the ratio Eo/Ei with [Cli] = [Clo] (Knauf 1979; Knauf and Brahm 1989). If [Cli] = [Clo], then the K1W value should approach K1W,E as [Cl] → 0, because under these conditions the only forms of the transporter present are Ei and Eo. If we take the y intercept of the best-fit straight line in Fig. 3 as a reasonable estimate for K1W,E, we can calculate the K1W,Ei value that would be required to give the observed intercept K1W,E value, assuming the mean value of K1W,Eo shown in Table and various values of the asymmetry factor, A. A plot of K1W,Ei versus A is shown in Fig. 8. Note that for A values of 0.05–0.1, corresponding to the values obtained by other measurement techniques (Knauf and Brahm 1989; Gasbjerg and Brahm 1991), the calculated K1W,Ei values are very similar to the measured values shown in Table. For A values over 0.15, however, the calculated value of K1W,Ei rises sharply, and for A values > 0.187, K1W,Ei approaches infinity. Thus the data are consistent with the observed asymmetry for unloaded transport sites as measured by other techniques, and they show further that no value for K1W,Ei can be obtained that is consistent with the K1W,E and K1W,Eo data if A > 0.187, providing further evidence that the unloaded transport sites are very asymmetrically distributed, with far more AE1 molecules in the Ei form than in the Eo form.

Determination of Equilibrium Dissociation Constant, KeqW

To measure the apparent dissociation constant after WW781 comes into equilibrium with both complex 1 and complex 2, designated KeqW (Fig. 5 B), cells were exposed to WW781 for at least 4 min before Cl exchange fluxes were measured, as described in materials and methods. On the basis of the rate constants measured for the WW781 complex formation in 150 mM [Cl] medium, this should be sufficient time for WW781 inhibition to approach its maximum value. Indeed, it was found that 1 or 2 h of pretreatment produced no significant increase in inhibition (decrease in the apparent KeqW value) as compared with the standard pretreatment protocol, and that the plots of appearance of isotope versus time in media with the same [WW781] as in the pretreatment medium were nearly linear (data not shown), as expected if little further inhibition of Cl exchange occurs during the flux measurement.

To see whether or not inhibition under these conditions still fits to a one-site binding model, fluxes were plotted against the WW781 concentration and the data were fitted to the equation for single-site inhibition. Data for cells loaded with 600 mM [Cl] are shown in Fig. 9, plotted in the form of a Dixon plot (1/J versus [WW781]). Although there is some scatter at the highest WW781 concentrations, where the fluxes are smallest, in general the fit to a straight line is very good, indicating that equilibrium inhibition can be well described by a single-site model. The straight line through the data in Fig. 9 was drawn using the parameters obtained from a nonlinear fit of the single-site inhibition equation () to the original flux versus WW781 plot. The negative of the x intercept of this line gives the apparent KeqW under these conditions.

Fig. 10 shows the results of several such experiments at different Cl concentrations. Note that in this case the concentrations required for half-inhibition of Cl exchange (KeqW) are much lower than those shown in Fig. 3 (K1W), where exposure to WW781 was for very short times. In this case, the slope of a best-fit straight line (solid line) is significantly different from zero (P = 0.008), indicating that Cl binding does affect the equilibrium of WW781 with complex 2.

Table shows the comparison of the mean values of KeqW with those of K1W. For every case except the 600-mM Cl data, the difference is significant, and in the latter case it approaches significance (P = 0.06). Particularly for the low [Cl] condition, the KeqW is substantially lower than K1W, demonstrating that the equilibrium between complex 1 and complex 2 is strongly in favor of complex 2 under these conditions.

Effect of AE1 Conformation on Equilibrium Affinity for WW781

As for the initial binding, the equilibrium binding dissociation constants of the various forms of AE1 for WW781 may differ (Fig. 5 B). For the equilibrium binding, the dissociation constants are designated KeqW,Eo, KeqW,EClo, etc. To see whether or not the AE1 conformation affects the equilibrium affinity for WW781, we first varied external [Cl], which should affect the fraction of Eo relative to the other forms of AE1. As shown in Table, the KeqW with 2 mM [Cl] outside was significantly lower than that with 150 mM [Cl] outside, indicating that the Eo form has a higher affinity for WW781 than the mixture of the other forms of AE1 that prevails in 150 mM [Cl] medium.

To determine KeqW,Eo, the dissociation constant for equilibrium binding to Eo, the effects of WW781 were measured in cells with ∼150 mM [Cl] inside and different external [Cl] concentrations. Combined data for four such experiments are shown in Fig. 11. In this case, particularly at the low external Cl concentration (○), inhibition is very pronounced even at the lowest WW781 concentrations (Fig. 11 A). Although it was not possible to obtain data between 0 and ∼0.1 μM WW781, because the binding to AE1 causes depletion of the WW781 in the medium under these conditions (Raha et al. 1993), the data at higher concentrations fit quite well to the equation for single-site inhibition, as shown by the linearity of Dixon plots of the data (Fig. 11 B). Thus, it was possible to determine KeqW,Eo values from the negative of the x value of the intersection point of the Dixon plot lines for different [Clo], by using the parameters obtained from fits to the flux versus [WW781] data to calculate the intersection point. As expected from the KeqW data, KeqW,Eo is significantly lower than K1W,Eo (Table). Similar data were obtained with iodide added to the low [Cl] external medium to determine KeqW,EIo (Table).

As shown in Fig. 4, for each form of AE1, an equation for the two-step binding can be written, analogous to the equation for the total AE1 present. For Eo (Scheme II),

 
formula
(Scheme II)

where kW•Eo,W-Eo and kW-Eo,W•Eo are the rate constants for formation and breakdown of complex 2 (Scheme II). The constant describing the equilibrium ratio between complex 2 and complex 1 for Eo is defined as:

 
\begin{equation*}K_{{2}/{1{\mathrm{W,Eo}}}}=\frac{ \left \left[{\mathrm{W}}-{\mathrm{E}}_{{\mathrm{o}}}\right] \right }{ \left \left[{\mathrm{W}}{\cdot}{\mathrm{E}}_{{\mathrm{o}}}\right] \right }|_{eq}=\frac{k_{{\mathrm{W}}{\cdot}{\mathrm{Eo,W}}-{\mathrm{Eo}}}}{k_{{\mathrm{W}}-{\mathrm{Eo,W}}{\cdot}{\mathrm{Eo}}}}{\mathrm{.}}\end{equation*}
3

K2/1W,Eo can be determined from K1W,Eo and KeqW,Eo in a manner analogous to equation 6 of Raha et al. 1993, as follows:

 
\begin{equation*}K_{{2}/{1{\mathrm{W,Eo}}}}= \left \left({K_{{\mathrm{1W,Eo}}}}/{K_{{\mathrm{eqW,Eo}}}}\right) \right -1{\mathrm{.}}\end{equation*}
4

From the data in Table, K2/1W,Eo is 28, corresponding to a free energy difference, ΔG0 [= exp(−K2/1W,Eo/RT)], between complex 2 and complex 1 (for the Eo form of AE1) of −7.6 kJ/mol. As in the case of the K1W values, binding of iodide to Eo causes an apparent increase in KeqW that is not statistically significant, but which corresponds to a K2/1W,EIo value of 12.5, suggesting that binding of iodide to Eo destabilizes complex 2 and thus decreases the equilibrium ratio of complex 2 to complex 1.

From the y intercept of the straight line fit to the KeqW values in Fig. 10, it is possible to obtain a value for KeqW,E, the equilibrium dissociation constant for binding to both forms of AE1 without Cl bound (Eo and Ei), with [Clo] = [Cli], which is 0.046 μM. As for the K1W values, this can be used together with the value of KeqW,Eo to calculate values of KeqW,Ei that are consistent with the ping-pong model (see ). As shown in Fig. 12, the minimum value of KeqW,Ei consistent with the model is equal to KeqW,E, 0.046 μM, and the maximum value of A is 0.0926. Thus, the equilibrium binding data provide further evidence for at least 10-fold asymmetry in favor of Ei versus Eo, in agreement with other determinations of this parameter (Knauf and Brahm 1989; Gasbjerg and Brahm 1991).

Although it is not possible to measure the dissociation constants for WW781 binding to the individual Cl-loaded forms of AE1, ECli, and EClo, it is possible to estimate a value for the dissociation constant, KeqW,ECl (see ), for the combination of EClo and ECli, whose ratio is a constant determined by the value of the asymmetry factor for Cl-loaded forms of AE1, ACl (= kECli,EClo/kEClo,ECli = EClo/ECli) (Knauf 1989). Calculations using , together with the mean value of the KeqW at 600 mM [Cl] and values of KCl,E, the apparent Cl dissociation constant with [Cli] = [Clo], ranging from 40 to 65 mM, give values for KeqW,ECl of 0.166–0.185 μM. Similar calculations for K1W,ECl range from 0.524 to 0.529 μM. The ratio of complex 2 to complex 1 for the Cl-loaded forms (K2/1W,ECl) is 1.8–2.2, as compared with 14.6 for the unloaded forms (K2/1W,E). Cl binding thus appears to reduce the tendency of WW781 to form the second complex with AE1.

An alternate method for determining KeqW,E and KeqW,ECl involves fitting all of the K1W or KeqW data to , which gives the half-inhibitory concentration as a function of KW,E, KW,ECl, and [Cl], under conditions where [Cli] = [Clo]. If this is done for the K1W data in Fig. 3, assuming that KCl,E = 50 mM (Gasbjerg and Brahm 1991), the dashed line is obtained, with K1W,E = 0.99 μM and K1W,ECl = 0.51 μM. When the calculations of Fig. 8 are repeated with the higher value of K1W,E, the minimum value of K1W,Ei rises to 0.99 μM and the maximum value of A ≤ 0.129.

Similar calculations for the KeqW data in Fig. 10 (without the lowest point at 150 mM [Cl]) (dashed line) give KeqW,ECl = 0.19 μM and KeqW,E = 0.033 μM, which corresponds to the minimum possible value of KeqW,Ei. Consistent results are obtained with A ≤ 0.134.

From the estimates of KeqW,ECl, it is further possible to calculate the values of KeqW,ECli that are compatible with the other measured parameters. This requires some knowledge of KeqW,EClo, which cannot be directly measured. However, if KeqW,EClo lies somewhere between the value of KeqW,Eo and KeqW,EIo, the value with I bound (which seems reasonable because the smaller Cl ion would be expected to have less effect on WW781 binding than does the larger iodide), then calculations of KeqW,ECli as a function of ACl can be done using , as shown in Fig. 13. The minimum value of KeqW,ECli consistent with the data is 0.17–0.19 μM (depending on which KeqW,ECl value is used), considerably higher than the minimum value of KeqW,Ei (0.033–0.046 μM). Regardless of which value of KeqW,EClo is chosen, there is some value of ACl at which KeqW,ECli becomes infinite; that is, no consistent solution can be found. Thus, (based on the assumption that KeqW,EClo is within 1 SEM KeqW,EIo value) the data provide evidence that ACl < 0.28, consistent with previous nuclear magnetic resonance (NMR) experiments indicating that ACl is similar to A (Liu et al. 1996). Furthermore, the fact that calculations assuming that KeqW,EClo = KeqW,Eo (Fig. 13, leftmost dotted line) only give consistent solutions with values of ACl much lower than that estimated by NMR suggests that external Cl binding probably does increase the equilibrium dissociation constant for WW781 binding to Eo; i.e., KeqW,EClo > KeqW,Eo.

Measurements of the KeqW,Ei and KeqW,EIi parameters by techniques similar to those used for the K1W parameters in Fig. 7 are difficult because the effects of WW781 on the fluxes are so much larger than those of either [Cli] or [Ii] that the Dixon plot lines tend to be nearly parallel and the intersection points are difficult to determine with any precision. The two experiments that were performed to measure KeqW,EIi, however, gave values of 0.21 and 0.35 μM (data not shown), which, while not statistically different from the corresponding KeqW,Eo and KeqW,EIo values (Table), are reasonable in light of the constraints imposed on KeqW,ECli (≥0.17 μM) according to the analyses in Fig. 13.

Discussion

Accuracy of Initial and Equilibrium Estimates of WW781 Affinities

The methods used to distinguish the effects of initial binding of WW781 from the effects after equilibrium is reached have the disadvantage that, since some complex 2 is formed even at early times, apparent values of K1W obtained from early time samples are always at least slightly contaminated by additional inhibition caused by the second binding step. The measurements of equilibrium parameters have the opposite problem in that the extent of the second binding step will be underestimated if the time of exposure to WW781 is not sufficient to ensure that equilibrium has been attained. Fortunately, the effects of these errors are in opposite directions, resulting in underestimates of K1W and overestimates of KeqW. Since KeqW < K1W, such errors will tend to bring the measured K1W and KeqW values closer together than they actually are. The fact that K1W is over nine times larger than KeqW for [Cl] ≤ 150 mM (Table) argues that the method, while not completely accurate, is at least able to detect a considerable difference in the initial and equilibrium affinities for WW781. The significant effect of increasing [Cl] on KeqW (Fig. 10), but not on K1W (Fig. 3) provides further evidence that even this first-order method is sufficient to distinguish different effects on the two parameters.

Time Course of Binding

In general, the differential equations describing the time course of the binding and transport inhibition are exceedingly complex, since each form of AE1 has its own individual parameters, as shown in Fig. 4. If, however, the forms of AE1 without WW781 equilibrate rapidly with each other, as well as with their corresponding complex 1 forms, then the time course of complex 2 formation should be described by a single rate constant. This seems to be the case, at least at 150 mM [Cl], where our previous analysis was consistent with a simple exponential approach to equilibrium for both the “on” and “off” binding reactions (Raha et al. 1993). This apparent rate constant, however, will depend on the relative abundance of the different forms of AE1, together with the magnitude of their individual rate constants, so the time course of binding is expected to be affected by changes in AE1 distribution caused, for example, by changes in [Cli] and [Clo]. Further experiments, using techniques that permit direct measurements of binding, are needed to explore this possibility.

Preference for Forms of AE1

Both for the initial binding and equilibrium association, WW781 exhibits a strong preference (by >10-fold) for the Eo form of AE1 as compared with the Ei form. Previous brief reports of this work (Knauf et al. 1990, Knauf et al. 1992; Mendoza et al. 1990) gave a value of >100 for the ratio of the equilibrium dissociation constants for Ei and Eo, but this was based on a particular estimate of the loaded-site asymmetry ratio, ACl, which gave a higher value of KeqW,Ei (see Fig. 12). The present data, which only give a lower limit for KeqW,Ei, do not preclude the possibility that the KeqW,Ei/ KeqW,Eo ratio could be as large as 100, but they provide no evidence for a ratio >12 (Table). Even this is a larger selectivity factor than that exhibited by any other inhibitor (Knauf et al. 1992), except the disulfonic stilbenes such as DNDS (Fröhlich 1982; Knauf et al. 1993). In contrast to the disulfonic stilbenes, where binding of Cl greatly diminishes the apparent affinity for the inhibitor, in a manner that resembles competitive inhibition (Shami et al. 1978; Fröhlich 1982), for the initial binding step of WW781 there was no significant effect of [Cl] (Fig. 3), and for the equilibrium binding (Table) we observed only about a fourfold increase in the dissociation constant for binding to the Cl-loaded forms (KeqW,ECl) as compared with the unloaded forms (KeqW,E).

Conclusions about the WW781 affinity of the Cl-loaded forms, based as they are on the [Cl] dependence of the KeqW, must be tempered by the possibility that binding of Cl to the inhibitory modifier site may also affect the affinity for WW781. Although the dissociation constant for binding of Cl to the modifier site (over 300 mM; Dalmark 1976; Gasbjerg and Brahm 1991) is much higher than that (around 50 mM; Gasbjerg and Brahm 1991) for binding to the transport site, the inevitable scatter in measurements of dissociation constants (because the parameter measured, Cl exchange inhibition, is a less-than-linear function of the dissociation constant) makes it impossible from our experiments to precisely determine the Cl affinity of the site that affects WW781 binding. Models in which the change in WW781 affinity is assumed to be due to binding to the transport site, however, indicated by the dashed lines in Fig. 3 and Fig. 10, fit the observations reasonably well. Nevertheless, the possibility that modifier site binding of Cl causes some or all of the change in WW781 affinity cannot be excluded.

Although the binding of an oxonol analogue of WW781, diBA(5)C4, is mutually exclusive with the disulfonic stilbene, DNDS (Knauf et al. 1995), the different effects of [Cl] on the affinity of DNDS (Fröhlich 1982) as compared with diBA(5)C4 (Knauf et al. 1995) and WW781 (Fig. 3 and Fig. 10) indicate that, if the binding sites overlap, they do so only partially. Thus, the fact that WW781 binding is affected both by the orientation of the transport site (Ei versus Eo) and by the binding of Cl suggests that an additional part of the AE1 protein, probably adjacent to but distinct from the disulfonic stilbene binding site, is affected by the conformational changes that occur during substrate binding and transport-site reorientation. This in turn implies that these conformational changes are not simply local reorientations of a few amino acids, but likely involve more global changes in protein structure such as have been observed with other proteins such as hemoglobin and alcohol dehydrogenase.

Effects of WW781 Binding on AE1 Conformation

Because of the preference for AE1 forms with the transport site facing outward (Eo and EClo), the addition of even as little as 1 μM of WW781 should cause a pronounced reorientation of AE1 toward the outward-facing conformations. Calculations for various cases based on values for A and ACl consistent with the present and previous data are shown in Fig. 14. With 5 mM [Cl] inside and outside, AE1 is highly biased toward the Ei form, and >93% of AE1 is in some inward-facing form. Addition of WW781 causes this to shift so that ∼60% is in the outward-facing form. A Cl gradient across the membrane, with 150 mM inside and 2 mM outside, normally results in ∼56% of the sites facing outward, but with WW781 this is increased to >98%. Even with a much smaller gradient, such as 150 mM [Cli] and 30 mM [Clo], the preference for outward-facing forms is >80% in the presence of WW781 (data not shown). With symmetric 150 mM Cl concentrations, where >90% of the sites face inward, WW781 causes rearrangement toward a more uniform distribution among the various forms, with ∼55% of the transport sites facing outward.

Because of this property, WW781 may be a useful tool for changing the orientation of sites, for studies aimed at detecting differences between the inward and outward-facing conformations, one of the central questions that must be answered if AE1 anion exchange is to be understood at the molecular level. The reorientation by WW781 toward outward-facing sites may also help to explain the early observation by Freedman and Novak 1983 that WW781 appears to cause an increase in net Cl permeability, PCl. Considered together with the observation of Fröhlich et al. 1983 that PCl increases when the fraction of AE1 in the Eo conformation is increased by lowering [Clo], the skewing of the AE1 distribution to outward-facing forms by WW781 may explain the increase in PCl.

Information About the Asymmetry of the AE1 Mechanism

The data presented here add to a growing body of evidence that the AE1 transport sites are very asymmetrically distributed. The conclusion that A ≤ 0.1 is in good agreement with most of the data from flux measurements and other techniques (Knauf and Brahm 1989; Gasbjerg and Brahm 1991), although some of the data that give higher estimates of A (Hautmann and Schnell 1985; Knauf and Brahm 1989) are incompatible with this result. The idea that ACl ≤ 0.28 agrees with 35Cl NMR binding measurements, which give an average ACl value of ∼0.1, based on a measured external Cl dissociation constant (Ko) of ∼40 mM in eosin-5-maleimide–treated cells (Liu et al. 1996). The values of A and ACl used to calculate Fig. 14 predict a slightly lower Ko of ∼33 mM, nearer to the value of 23 ± 9 mM from more recent improved NMR measurements (Kennedy, S.D., C. Wu, and P.A. Knauf, unpublished data). For both the Cl-loaded and unloaded forms of AE1, therefore, the Gibbs free energy for the form with the transport site facing inward seems to be ∼5 kJ/mol lower than that for the outward-facing form.

The data presented here demonstrate the usefulness of an oxonol inhibitor to distinguish the various conformations of AE1 and to both measure and alter the distribution among the various forms. Further work is needed to define the structural requirements that confer conformational selectivity on WW781, and to determine whether or not other oxonols with different structures exhibit the same or different conformational preferences.

Acknowledgments

The authors gratefully acknowledge the invaluable assistance of Mr. Alvin Law in analyzing data and preparing the figures in this paper and the excellent technical assistance of Ms. Jacqueline Brescia. The assistance of Dr. J.C. Freedman in providing initial samples of WW781, as well as the suggestions of Dr. Alan Weinstein (Cornell University Medical College, New York, NY) concerning the nomenclature, are also appreciated.

This study was supported by a grant from the National Institute of Diabetes, Digestive and Kidney Diseases (R01-DK-27495).

Dependence of KW on [Cl] with [Cli] = [Clo]

KW is defined as the concentration of WW781 that causes 50% inhibition of transport. Depending on whether experiments are done under initial binding conditions or equilibrium conditions, KW can be thought of as either K1W or KeqW.

KCl,E, the concentration of Cl that gives half-maximal Cl exchange, with [Cli] = [Clo], is given in in Gasbjerg and Brahm 1991 and in Liu et al. 1996 (Eq. A1):

 
\begin{equation*}K_{{\mathrm{Cl,E}}}={K_{{\mathrm{Cl,Ei}}} \left \left(1+A\right) \right }/{ \left \left(1+A_{Cl}\right) \right }{\mathrm{,}}\end{equation*}
A1

where A is the asymmetry ratio (Eo/Ei) for the forms of AE1 with the transport site unloaded and with [Cli] = [Clo], and ACl is the asymmetry ratio (EClo/ECli) for the Cl-loaded forms of AE1. A = (kECli,ECloKCl,Eo)/(kEClo,ECliKCl,Ei) and ACl = kECli,EClo/kEClo,ECli, where the various constants are defined as described in materials and methods and Fig. 4 and Fig. 5. If we define ECl as the sum of ECli plus EClo and E as the sum of Ei + Eo, then their ratio is given by:

 
\begin{equation*}{ \left \left[{\mathrm{ECl}}\right] \right }/{ \left \left[{\mathrm{E}}\right] \right }={ \left \left(1+A_{{\mathrm{Cl}}}\right) \right \left \left[{\mathrm{Cl}}^{{\mathrm{-}}}\right] \right }/{ \left K_{{\mathrm{Cl,Ei}}} \left \left(1+A\right) \right \right }={ \left \left[{\mathrm{Cl}}^{{\mathrm{-}}}\right] \right }/{K_{{\mathrm{Cl,E}}}}{\mathrm{.}}\end{equation*}
A2

KW can be defined as an apparent dissociation constant:

 
\begin{equation*}K_{{\mathrm{W}}}= \left \left[{\mathrm{W}}\right] \right { \left \left( \left \left[{\mathrm{E}}\right] \right + \left \left[{\mathrm{ECl}}\right] \right \right) \right }/{ \left \left( \left \left[{\mathrm{WE}}\right] \right + \left \left[{\mathrm{WECl}}\right] \right \right) \right }{\mathrm{,}}\end{equation*}
A3

where W is WW781, WE represents the sum of the E forms associated with WW781; that is, the sum of WEi plus WEo, and WECl represents the sum of WECli + WEClo (Fig. 4 and Fig. 5).

We can also define the apparent dissociation constant for binding of WW781 to the unloaded forms of AE1 as:

 
\begin{equation*}K_{{\mathrm{W,E}}}={ \left \left[{\mathrm{E}}\right] \right \left \left[{\mathrm{W}}\right] \right }/{ \left \left[{\mathrm{WE}}\right] \right {\mathrm{,}}}\end{equation*}
A4

and that for the Cl-loaded forms as:

 
\begin{equation*}K_{{\mathrm{W,ECl}}}={ \left \left[{\mathrm{ECl}}\right] \right \left \left[{\mathrm{W}}\right] \right }/{ \left \left[{\mathrm{WECl}}\right] \right {\mathrm{.}}}\end{equation*}
A5

Substituting , and into A3, we obtain:

 
\begin{equation*}K_{{\mathrm{W}}}={K_{{\mathrm{W,E}}} \left \left(1+{ \left \left[{\mathrm{Cl}}^{{\mathrm{-}}}\right] \right }/{K_{{\mathrm{Cl,E}}}}\right) \right }/{ \left \left(1+{ \left \left[{\mathrm{Cl}}^{{\mathrm{-}}}\right] \right K_{{\mathrm{W,E}}}}/{K_{{\mathrm{Cl,E}}}K_{{\mathrm{W,ECl}}}}\right) \right }{\mathrm{.}}\end{equation*}
A6

This equation was used to determine the dashed lines in Fig. 3 and Fig. 10. It can be solved for KW,ECl if KW,E is known as well as the KW for any given [Cl]:

 
\begin{equation*}K_{{\mathrm{W,ECl}}}={K_{{\mathrm{W}}}K_{{\mathrm{W,E}}} \left \left[{\mathrm{Cl}}^{{\mathrm{-}}}\right] \right }/{ \left K_{{\mathrm{W,E}}} \left \left[{\mathrm{Cl}}^{{\mathrm{-}}}\right] \right +K_{{\mathrm{Cl,E}}} \left \left(K_{{\mathrm{W,E}}}-K_{{\mathrm{W}}}\right) \right \right }{\mathrm{.}}\end{equation*}
A7

This equation was used to calculate KW,ECl values from the KW values with 600 mM [Cl].

As shown by Fröhlich 1982 and Fröhlich and Gunn 1986, and as is obvious from , the KW extrapolated to zero [Cl] is equal to KW,E, which is given by (see equations 39, 35a, and 35b of Fröhlich and Gunn 1986; note that their a/b = A):

 
\begin{equation*}K_{{\mathrm{W,E}}}={ \left \left(1+A\right) \right }/{ \left \left({1}/{K_{{\mathrm{W,Ei}}}}+{A}/{K_{{\mathrm{W,Eo}}}}\right) \right }{\mathrm{,}}\end{equation*}
A8

which corresponds to for K1W,E. Similarly, KW,ECl is given by (see equation 42b of Fröhlich and Gunn 1986; note that their K2 = ACl) ():

 
\begin{equation*}K_{{\mathrm{W,ECl}}}={ \left \left(1+A_{{\mathrm{Cl}}}\right) \right }/{ \left \left({1}/{K_{{\mathrm{W,ECli}}}}+{A_{{\mathrm{Cl}}}}/{K_{{\mathrm{W,EClo}}}}\right) \right }{\mathrm{.}}\end{equation*}
A9

If KW,ECl is known, KW,ECli can be calculated as a function of ACl and KW,EClo, as in Fig. 13, as follows:

 
\begin{equation*}K_{{\mathrm{W,ECli}}}={K_{{\mathrm{W,ECl}}}}/{ \left \left(1+A_{{\mathrm{Cl}}}-{A_{{\mathrm{Cl}}}K_{{\mathrm{W,ECl}}}}/{K_{{\mathrm{W,EClo}}}}\right) \right }{\mathrm{.}}\end{equation*}
A10

References

References
Chernova
M.N.
,
Jiang
L.
,
Crest
M.
,
Hand
M.
,
Vandorpe
D.H.
,
Strange
K.
,
Alper
S.L.
Electrogenic sulfate/chloride exchange in Xenopus oocytes mediated by murine AE1 E699Q
J. Gen. Physiol.
109
1997
345
360
[PubMed]
Dalmark
M.
Effects of halides and bicarbonate on chloride transport in human red blood cells
J. Gen. Physiol.
67
1976
223
234
[PubMed]
Freedman
J.C.
,
Novak
T.S.
Membrane potentials associated with Ca-induced K conductance in human red blood cellsstudies with a fluorescent oxonol dye, WW781
J. Membr. Biol.
72
1983
59
74
[PubMed]
Fröhlich
O.
The external anion binding site of the human erythrocyte anion transporterDNDS binding and competition with chloride
J. Membr. Biol.
65
1982
111
123
[PubMed]
Fröhlich
O.
,
Gunn
R.B.
Erythrocyte anion transportthe kinetics of a single-site obligatory exchange system
Biochim. Biophys. Acta.
864
1986
169
194
[PubMed]
Fröhlich
O.
,
Leibson
C.
,
Gunn
R.B.
Chloride net efflux from intact erythrocytes under slippage conditions. Evidence for a positive charge on the anion binding/transport site
J. Gen. Physiol.
81
1983
127
152
[PubMed]
Fujinaga
J.
,
Tang
X.-B.
,
Casey
J.R.
Topology of the membrane domain of human erythrocyte anion exchange protein, AE1
J. Biol. Chem.
274
1999
6626
6633
[PubMed]
Gasbjerg
P.K.
,
Brahm
J.
Kinetics of bicarbonate and chloride transport in human red cell membranes
J. Gen. Physiol.
97
1991
321
350
[PubMed]
George
E.B.
,
Nyirjesy
P.
,
Basson
M.
,
Ernst
L.A.
,
Pratap
P.R.
,
Freedman
J.C.
,
Waggoner
A.S.
Impermeant potential-sensitive oxonol dyesI. Evidence for an “on–off” mechanism
J. Membr. Biol.
103
1988
245
253
[PubMed]
Gunn
R.B.
,
Fröhlich
O.
Asymmetry in the mechanism for anion exchange in human red blood cell membranes. Evidence for reciprocating sites that react with one transported anion at a time
J. Gen. Physiol.
74
1979
351
374
[PubMed]
Hautmann
M.
,
Schnell
K.F.
Concentration dependence of the chloride selfexchange and homoexchange fluxes in human red cell ghosts
Pflügers Arch.
405
1985
193
201
Jennings
M.L.
Stoichiometry of a half-turnover of band 3, the chloride transport protein of human erythrocytes
J. Gen. Physiol.
79
1982
169
185
[PubMed]
Jennings
M.L.
,
Anion transport proteins
Seldin
D.W.
,
Giebisch
G.
The KidneyPhysiology and Pathophysiology
1992
503
535
Raven Press
New York, NY
a
Jennings
M.L.
,
Cellular anion transport
Seldin
D.W.
,
Giebisch
G.
The KidneyPhysiology and Pathophysiology
1992
113
145
Raven Press
New York, NY
b
Jennings
M.L.
Rapid electrogenic sulfate-chloride exchange mediated by chemically modified band 3 in human erythrocytes
J. Gen. Physiol.
105
1995
21
47
[PubMed]
Jennings
M.L.
,
Whitlock
J.
,
Shinde
A.
Pre–steady state transport by erythrocyte band 3 proteinuphill countertransport induced by the impermeant inhibitor H2DIDS
Biochem. Cell Biol.
76
1998
807
813
[PubMed]
Knauf
P.A.
Erythrocyte anion exchange and the band 3 proteintransport kinetics and molecular structure
Curr. Top. Membr. Transp.
12
1979
249
363
Knauf
P.A.
,
Kinetics of anion transport
Raess
B.U.
,
Tunnicliff
G.
The Red Cell Membrane
1989
171
200
Humana Press
Clifton, NJ
Knauf
P.A.
,
Brahm
J.
,
Functional asymmetry of the anion exchange protein, capnophorineffects on substrate and inhibitor binding
Fleischer
S.
,
Fleischer
B.
Methods in Enzymology. Vol. 173
1989
432
453
Academic Press
New York, NY
Knauf
P.A.
,
Fuhrmann
G.F.
,
Rothstein
S.
,
Rothstein
A.
The relationship between anion exchange and net anion flow across the human red blood cell membrane
J. Gen. Physiol.
60
1977
363
386
[PubMed]
Knauf
P.A.
,
Law
F.
,
Hahn
K.
An oxonol dye is the most potent known inhibitor of band 3–mediated anion exchange
Am. J. Physiol. Cell Physiol
269
1995
C1073
C1077
Knauf
P.A.
,
Mann
N.A.
Location of the chloride self-inhibitory site of the human erythrocyte anion exchange system
Am. J. Physiol. Cell Physiol.
251
1986
C1
C9
Knauf
P.A.
,
Mendoza
N.S.
,
Spinelli
L.J.
The conformation of the human red blood cell anion exchange protein, band 3, strongly affects binding of the oxonol dye WW781
FASEB J.
4
1990
A819
(Abstr.)
Knauf
P.A.
,
Restrepo
D.
,
Liu
S.J.
,
Raha
N.M.
,
Spinelli
L.J.
,
Law
Y.
,
Cronise
B.
,
Snyder
R.B.
,
Romanow
L.
Mechanisms of substrate binding, inhibitor binding, and ion translocation in band 3 and band 3–related proteins
Prog. Cell Res.
2
1992
35
44
Knauf
P.A.
,
Ries
E.A.
,
Romanow
L.A.
,
Bahar
S.
,
Szekeres
E.S.
DNDS (4,4′-dinitro-stilbene-2,2′-disulfonate) does not act as a purely competitive inhibitor of red blood cell band 3–mediated anion exchange
Biophys. J.
64
1993
A307
(Abstr.)
Knauf
P.A.
,
Spinelli
L.J.
,
Mann
N.A.
Flufenamic acid senses conformation and asymmetry of human erythrocyte band 3 anion transport protein
Am. J. Physiol. Cell Physiol.
257
1989
C277
C289
Liu
D.
,
Kennedy
S.D.
,
Knauf
P.A.
Source of transport site asymmetry in the band 3 anion exchange protein determined by NMR measurements of external Cl− affinity
Biochemistry.
35
1996
15228
15235
[PubMed]
Mendoza
N.S.
,
Spinelli
L.J.
,
Knauf
P.A.
WW781 noncompetitively inhibits red blood cell Cl− exchange by a 2-step mechanism which is sensitive to substrate binding and transport site reorientation
Biophys. J.
57
1990
96a
(Abstr.)
Müller-Berger
S.
,
Karbach
D.
,
König
J.
,
Lepke
S.
,
Wood
P.G.
,
Appelhans
H.
,
Passow
H.
Inhibition of mouse erythroid band 3–mediated chloride transport by site-directed mutagenesis of histidine residues and its reversal by second site mutation of Lys 558, the locus of covalent H2DIDS binding
Biochemistry.
34
1995
9315
9324
[PubMed]
Raha
N.M.
,
Spinelli
L.J.
,
Knauf
P.A.
WW781, a potent reversible inhibitor of red cell Cl− flux, binds to band 3 by a two-step mechanism
Am. J. Physiol. Cell Physiol.
265
1993
C521
C532
Shami
Y.
,
Rothstein
A.
,
Knauf
P.A.
Identification of the Cl− transport site of human red blood cells by a kinetic analysis of the inhibitory effects of a chemical probe
Biochim. Biophys. Acta.
508
1978
357
363
[PubMed]
Tang
X.-B.
,
Kovacs
M.
,
Sterling
D.
,
Casey
J.R.
Identification of residues lining the translocation pore of human AE1, plasma membrane anion exchange protein
J. Biol. Chem.
274
1999
3557
3564
[PubMed]
Timmer
R.T.
,
Gunn
R.B.
Inducible expression of erythrocyte band 3 protein
Am. J. Physiol. Cell Physiol.
276
1999
C66
C75
Wang
D.N.
,
Sarabia
V.E.
,
Reithmeier
R.A.F.
,
Kuhlbrandt
W.
Three-dimensional map of the dimeric membrane domain of the human erythrocyte anion exchanger, band 3
EMBO (Eur. Mol. Biol. Organ.) J.
13
1994
3230
3235

Abbreviations used in this paper: diBA(5)C4, bis-(1,3-dibutylbarbituric acid)-pentamethine oxonol; DNDS, 4,4′-dinitro-stilbene-2,2′-disulfonate; Ei, inside form of the AE1 protein; Eo, outside form of the AE1 protein; NMR, nuclear magnetic resonance; WW781, [3-methyl-1-p-sulfophenyl-5-pyrazolone-(4)]-[1,3-dibutylbarbituric acid]-penta-methine oxonol.