Activity and distribution of intracellular carbonic anhydrase II and their effects on the transport activity of anion exchanger AE1/SLC4A1

Abstract

We have investigated the previously published 'metabolon hypothesis' postulating that a close association of the anion exchanger 1 (AE1) and cytosolic carbonic anhydrase II (CAII) exists that greatly increases the transport activity of AE1. We study whether there is a physical association of and direct functional interaction between CAII and AE1 in the native human red cell and in tsA201 cells coexpressing heterologous fluorescent fusion proteins CAII-CyPet and YPet-AE1. In these doubly transfected tsA201 cells, YPet-AE1 is clearly associated with the cell membrane, whereas CAII-CyPet is homogeneously distributed throughout the cell in a cytoplasmic pattern. Förster resonance energy transfer measurements fail to detect close proximity of YPet-AE1 and CAII-CyPet. The absence of an association of AE1 and CAII is supported by immunoprecipitation experiments using Flag-antibody against Flag-tagged AE1 expressed in tsA201 cells, which does not co-precipitate native CAII but co-precipitates coexpressed ankyrin. Both the CAII and the AE1 fusion proteins are fully functional in tsA201 cells as judged by CA activity and by cellular HCO3(-) permeability (P(HCO3(-))) sensitive to inhibition by 4,4-Diisothiocyano-2,2-stilbenedisulfonic acid. Expression of the non-catalytic CAII mutant V143Y leads to a drastic reduction of endogenous CAII and to a corresponding reduction of total intracellular CA activity. Overexpression of an N-terminally truncated CAII lacking the proposed site of interaction with the C-terminal cytoplasmic tail of AE1 substantially increases intracellular CA activity, as does overexpression of wild-type CAII. These variously co-transfected tsA201 cells exhibit a positive correlation between cellular P(HCO3(-)) and intracellular CA activity. The relationship reflects that expected from changes in cytoplasmic CA activity improving substrate supply to or removal from AE1, without requirement for a CAII-AE1 metabolon involving physical interaction. A functional contribution of the hypothesized CAII-AE1 metabolon to erythroid AE1-mediated HCO3(-) transport was further tested in normal red cells and red cells from CAII-deficient patients that retain substantial CA activity associated with the erythroid CAI protein lacking the proposed AE1-binding sequence. Erythroid P(HCO3(-)) was indistinguishable in these two cell types, providing no support for the proposed functional importance of the physical interaction of CAII and AE1. A theoretical model predicts that homogeneous cytoplasmic distribution of CAII is more favourable for cellular transport of HCO3(-) and CO2 than is association of CAII with the cytoplasmic surface of the plasma membrane. This is due to the fact that the relatively slow intracellular transport of H(+) makes it most efficient to place the CA in the vicinity of the haemoglobin molecules, which are homogeneously distributed over the cytoplasm.

Figure 1. Determination of and from the mass spectrometric records by minimizing the sum of squared deviations

Illustration of the fitting procedure used to determine at a given of 0.12 cm s⁻¹ (A) or both and (B). The sum of squares of deviations between experimental concentrations of C¹⁸O¹⁶O and their counterparts from the calculated C¹⁸O¹⁶O decay curve as obtained with the fitted values of / are plotted vs. the value of (A) or vs. both the values of and (B).

Figure 2. Co-immunoprecipitation (Co-IP) demonstrates interaction of AE1 with ANK1, but not with CAII

Western blot analysis of aliquots from cell lysates (Input; 3% of the volume of cell lysate added to the affinity gel was applied to the SDS-PAGE gel) from tsA201 cells transfected with pcDNA3–3xFlag-hAE1 expression plasmid (AE1) and pcDNA-hANK1 expression plasmid (Ankyrin) or empty expression vector (empty vector), of aliquots from supernatants of the Co-IP (Supernat.; 3% of the supernatant after centrifugation of the affinity gel was applied to the SDS-PAGE gel), and of eluates from the affinity gel (all of the eluate was applied to the SDS-PAGE gel), using anti-ANK1, anti-CAII or anti-Flag antibodies.

Figure 3. Confocal microscopy of the fluorescent fusion proteins CA II and AE1

A, confocal microscopic images of a tsA291 cell co-transfected with human CAII-CyPet and murine YPet-AE1 fusion proteins. CyPet was linked to the C terminus of CAII and YPet was linked to the N-terminus of mAE1 B, fluorescence emission intensity profiles at wavelengths appropriate for CyPet and for YPet as recorded along the red lines in A. C, confocal microscopic images of a tsA201 cell co-transfected with human YPet-AE1 and human CAII-CyPet. Comparison of A and C indicates that localization of mAE1 and hAE1 are identical. D, fluorescence control. The figure shows an overview over a patch of confluent tsA201 cells co-transfected with CAII-CyPet and YPet-AE1. Regions 1–4 represent non-transfected cells. Note the absence of non-specific fluorescence in the non-transfected cells, even with saturation of CyPet fluorescence (left panel). Non-saturated CyPet images were devoid of peripheral membrane enhancement (not shown). Arrows in right panel indicate morphologically unhealthy cells.

Figure 4. FRET measurements of CyPet- and YPet-labelled fusion proteins in tsA201 cells

A, confocal microscopic images of a tsA201 cell transfected with the doubly labelled construct of mAE1 N-terminally fused with YPet and C-terminally fused to CyPet. Illumination with 458 nm excites both CyPet (left) and YPet (right), the two dyes yielding an identical pattern. This construct gives the positive FRET signal seen in Fig. 5B. B, results of FRET experiments with tsA cells cotransfected with CAII-CyPet and YPet-hAE1 (left), with CAII-CyPet and YPet-mAE1 (centre), and a FRET control in cells transfected with YPet-mAE1-CyPet (right). This latter construct ensures proximity of the two dyes, reflected in its significantly positive FRET signal. *P < 0.01 (t test); n from left to right is 3, 7 and 7. Bars represent SE values.

Figure 5. Carbonic anhydrase activity and HCO₃⁻ permeability in native and YPet-mAE1-expressing tsA201 cells

A, intracellular CA activities measured in lysates of untreated tsA201 cells (control) and tsA201 cells transfected with YPet-mAE1 (AE1), using the mass spectrometric ¹⁸O technique. B, cellular measured in untreated tsA201 cells (control) and in tsA201 cells expressing YPet-mAE1 (AE1), in the absence or presence of DIDS (10⁻⁵ m), a strong inhibitor of AE1 transport function. DIDS showed no effect on of control cells, but inhibited the AE1-associated increase in . ns, P > 0.05, *P < 0.02, ^§P < 0.02. Number of measurements are inside the columns. Error bars represent SD values.

Figure 6. Bicarbonate permeability of tsA201 cells as a function of intracellular CA activity, A_i

tsA201 cells transfected with YPet-mAE1 (‘control’) (▾); tsA201 cells co-transfected with CAII-V143Y-CyPet and YPet-mAE1 (▴); co-transfection with WT CAII-CyPet and YPet-AE1 (♦); co-transfection with N-terminally truncated CAII-CyPet and YPet-mAE1 (•). Observed variations of are linearly related to intracellular CA activity. Correlation coefficient r= 0.90. Bars represent SE values. Number of determinations n= 17, 18, 17 and 33 (from left to right). All permeabilities under expression of the various CAII's are statistically significantly different from the mean control value (▾, transfection with YPet-mAE1 only). P values between <0.01 and <0.05.

Figure 7. Intra-erythrocytic CA activity and bicarbonate permeability of normal (control) and CAII-deficient human red blood cells (the latter from two individuals 1 and 2)

A, while normal human red cells exhibit a CA activity of 20,000, CAII-deficient red cells have an activity of ∼ 5000 due to the remaining CAI in these cells (* indicates statistically significant difference from control CA activity, P < 0.01). B, is identical regardless of whether CAII is present or completely absent. Numbers of measurements are given inside each column. Bars represent SD.

Figure 8. Time courses of bicarbonate influx, expressed as mol HCO₃⁻ (cm² membrane surface area)⁻¹, for various intra-erythrocytic CA activities, calculated with the model illustrated in Fig. S2 (upper scheme)

The numbers on the curves indicate the acceleration factors of CO₂ hydration. Therefore, 1 indicates the absence of CA activity, and 20,000 an activity of (20,000 – 1). The dependence of HCO₃⁻ influx on CA activity is most pronounced between 1 and 1000, and becomes minor between 5000 and 20,000.

Figure 9. Significance of the subcellular localization of CA for uptake of HCO₃⁻ and the associated CO₂ release in human red cells

Uppermost curves (in blue): calculated for a homogeneous intracellular distribution of CA in the cytoplasm of red cells (upper scheme in Fig. S2). Middle curves (in red): calculated for an accumulation of all the CA of the red cell in a thin (0.01 μm) layer immediately adjacent to the internal side of the red cell membrane (lower scheme in Fig. S2). Lowermost curves (black, in A and B): complete absence of CA activity inside a red cell. A, HCO₃⁻ influx after a step change in extracellular HCO₃⁻ from 25 to 35 mm. B, the efflux of CO₂ following the influx of HCO₃⁻ shown in A. C, effect of haemoglobin diffusivity on the kinetics of HCO₃⁻ exchange. The two curves are identical to the uppermost (blue) and middle (red) curves of A. The results represented by open circles were calculated for the accumulation of all red cell CA at the cytoplasmic side of the membrane, as in the lower (red) curve, with the exception that the diffusivity of haemoglobin was set to 100-fold its true value. This makes both types of CA distribution equivalent in terms of HCO₃⁻/CO₂ exchange, indicating that intra-erythrocytic Hb-facilitated proton transport causes the limitation of the process when CA is associated with the membrane.