K+-Driven Cl−/HCO3− Exchange Mediated by Slc4a8 and Slc4a10

Slc4a genes encode various types of transporters, including Na+-HCO3− cotransporters, Cl−/HCO3− exchangers, or Na+-driven Cl−/HCO3− exchangers. Previous research has revealed that Slc4a9 (Ae4) functions as a Cl−/HCO3− exchanger, which can be driven by either Na+ or K+, prompting investigation into whether other Slc4a members facilitate cation-dependent anion transport. In the present study, we show that either Na+ or K+ drive Cl−/HCO3− exchanger activity in cells overexpressing Slc4a8 or Slc4a10. Further characterization of cation-driven Cl−/HCO3− exchange demonstrated that Slc4a8 and Slc4a10 also mediate Cl− and HCO3−-dependent K+ transport. Full-atom molecular dynamics simulation on the recently solved structure of Slc4a8 supports the coordination of K+ at the Na+ binding site in S1. Sequence analysis shows that the critical residues coordinating monovalent cations are conserved among mouse Slc4a8 and Slc4a10 proteins. Together, our results suggest that Slc4a8 and Slc4a10 might transport K+ in the same direction as HCO3− ions in a similar fashion to that described for Na+ transport in the rat Slc4a8 structure.


Introduction
The regulation of pH is critical for physiological processes such as synaptic activity and transepithelial transport.Carbonic anhydrases and H + -transporting proteins (i.e., Na + /H + exchangers and H + -ATPases) are key proteins involved in pH homeostasis, which often are functionally coupled with other ion-transporting proteins such as Na + -HCO 3 − cotransporters, Cl − /HCO 3 − exchangers and Na + -dependent Cl − /HCO 3 − exchangers [1,2].With the sole exception of Slc4a11, which does not transport HCO 3 − [3-6], all the other members of the Slc4a gene family encode for HCO 3 − transporters with different activities (i.e., Na + -HCO 3 − cotransporters, Na + -independent and Na + -dependent Cl − /HCO 3 − exchangers) [7].The importance of Slc4a transporters in pH homeostasis can be inferred from pathologies associated with SLC4A1-5, SLC4A7 and SLC4A10 gene mutations [8,9].Due to the large Na + -inward electrochemical gradient displayed by the majority of mammalian cells, electroneutral Na + -driven Cl − /HCO 3 − exchangers are believed to mediate HCO 3 − influx under physiological conditions.However, Na + -driven inward HCO 3 transport might not be a common theme for all Slc4a members.For instance, intracellular [Cl − ] at the resting state is reduced in acinar secretory cells from Slc4a9 −/− (Ae4) salivary glands, suggesting that Slc4a9 promotes Cl − influx rather than HCO 3 − influx [10].A more recent study showed that Slc4a9 is in fact a cation-driven Cl − /HCO 3 − exchanger displaying a cation selectivity, Na + = K + [11].Therefore, it is likely that other Slc4a members that mediate Na + -driven Cl − /HCO 3 − exchange might also transport K + ions.In the present study, we evaluated the cation selectivity of the Na + -dependent transporters Slc4a8 and Slc4a10 by using a functional fluorescence approach.We found that both Slc4a8 and Slc4a10 display significant Cl − /HCO 3 − exchanger activity in the presence of either Na + or K + ions.Molecular dynamics simulations, based on the available Slc4a8 structure confirm that Slc4a8 has the ability to coordinate K + ions in a similar way to how Na + ions are coordinated within the transporter's structure.

Expression of Mouse Slc4a8 and Slc4a10 in CHO-K1 Cells
We first evaluated by indirect immunofluorescence the expression of the Myc-DDK tagged versions of mouse Slc4a8 and Slc4a10 clones in CHO-K1 cells.Slc4a8-and Slc4a10tagged proteins were localized as punctate signals within the intracellular space as well as in the cell periphery (Figure 1, green fluorescence).
Int. J. Mol.Sci.2024, 25, x FOR PEER REVIEW 2 of 14 HCO3 − influx under physiological conditions.However, Na + -driven inward HCO3 − transport might not be a common theme for all Slc4a members.For instance, intracellular [Cl − ] at the resting state is reduced in acinar secretory cells from Slc4a9 −/− (Ae4) salivary glands, suggesting that Slc4a9 promotes Cl − influx rather than HCO3 − influx [10].A more recent study showed that Slc4a9 is in fact a cation-driven Cl − /HCO3 − exchanger displaying a cation selectivity, Na + = K + [11].Therefore, it is likely that other Slc4a members that mediate Na + -driven Cl − /HCO3 − exchange might also transport K + ions.
In the present study, we evaluated the cation selectivity of the Na + -dependent transporters Slc4a8 and Slc4a10 by using a functional fluorescence approach.We found that both Slc4a8 and Slc4a10 display significant Cl − /HCO3 − exchanger activity in the presence of either Na + or K + ions.Molecular dynamics simulations, based on the available Slc4a8 structure confirm that Slc4a8 has the ability to coordinate K + ions in a similar way to how Na + ions are coordinated within the transporter's structure.

Expression of Mouse Slc4a8 and Slc4a10 in CHO-K1 Cells
We first evaluated by indirect immunofluorescence the expression of the Myc-DDK tagged versions of mouse Slc4a8 and Slc4a10 clones in CHO-K1 cells.Slc4a8-and Slc4a10tagged proteins were localized as punctate signals within the intracellular space as well as in the cell periphery (Figure 1, green fluorescence).To visualize the plasma membrane, transfected CHO-K1 cells were stained with wheat germ agglutinin (WGA) conjugated to Alexa Fluor TM 633 (red fluorescence).Some overlap between WGA-and Slc4a8-(Figure 1, upper panels) or Slc4a10-(Figure 1, middle panels) associated fluorescence was observed, confirming the membrane expression of the recombinant proteins.
The specificity of the anti-DDK antibody was verified by the absence of immunoreactivity in CHO-K1 cells transfected with the empty vector (Figure 1, lower panels).

Cl − Fluxes Are Associated with Slc4a8 and Slc4a10 Expression in CHO-K1 Cells
There is no consensus regarding the activity encoded by Slc4a8 and Slc4a10; some reports describe Slc4a8 and Slc4a10 as Na + -HCO 3 − cotransporters, while other studies show that Slc4a8 and Slc4a10 are Na + -driven Cl − /HCO 3 − exchangers [12][13][14][15].To assess whether Cl − fluxes are associated with the activities of Slc4a8 and Slc4a10, we quantified changes in fluorescence intensity in cells loaded with the Cl − indicator SPQ.As seen in Figure 2A,B, a loss in normalized fluorescence (black circles) occurred in response to a reduction in the external Cl − concentration in Slc4a8-and Slc4a10-expressing cells (~20%) that was substantially higher than that observed in non-transfected CHO-K1 cells (~5%, dashed line).We interpret the loss in fluorescence intensity as an increase in the rate of Cl − efflux from the cells.Moreover, Slc4a8-and Slc4a10-mediated Cl − efflux was dependent on HCO 3 − as little, if any, Cl − efflux was seen in Slc4a8-and Slc4a10-expressing cells measured under HCO 3 − -free conditions (open circles).The reduction in normalized fluorescence intensity and the change in flow rates was comparable to that obtained from control cells measured under HCO 3 − -containing solutions (dashed lines).that was substantially higher than that observed in non-transfected CHO-K1 cells (~5%, dashed line).We interpret the loss in fluorescence intensity as an increase in the rate of Cl − efflux from the cells.Moreover, Slc4a8-and Slc4a10-mediated Cl − efflux was dependent on HCO3 − as little, if any, Cl − efflux was seen in Slc4a8-and Slc4a10-expressing cells measured under HCO3 − -free conditions (open circles).The reduction in normalized fluorescence intensity and the change in flow rates was comparable to that obtained from control cells measured under HCO3 − -containing solutions (dashed lines).1) and then switched to a HCO3 − -containing/low Cl − solution (Solution B, Table 1).We next evaluated whether Slc4a8 and Slc4a10 activities display a cation dependence.To assess whether HCO 3 − fluxes are associated with the activities of Slc4a8 and Slc4a10, we quantified changes in fluorescence intensity in cells loaded with the pH − indicator BCECF-AM.Figure 3 shows that no alkalinization was observed in response to external Cl − reduction in Slc4a8-and Slc4a10-expressing cells when N-Methyl-D-Glucamine (NMDG) was the main cation present in the external solution.In contrast, a robust alkalinization was evident when external NMDG was switched to either Na + or K + ions in the low Cl − solution.Alkalinization induced by Na + or K + in cells expressing Slc4a8 and Slc4a10 occurs in a concentration-dependent fashion (Figure 3B,D).Hill function analyses of dose-response experiments revealed EC 50 values of 38.5 mM (Na + ) and 40.8 mM (K + ) for Slc4a8, while for Slc4a10, values of 54.5 mM (Na + ) and 58.0 mM (K + ) were obtained.

Slc4a8-and Slc4a10-Mediated K + Transport in CHO-K1 Cells
To directly assess whether potassium ions (K + ) are being transported by Slc4a8 and Slc4a10, we employed the K + -selective probe PBFI-AM to measure intracellular potassium concentrations ([K + ] i ) during stimulated exchanger activity.Due to the slight preference for K + over Na + ions (~1.5 times) displayed by PBFI, the experiments were conducted under Na + -free conditions to eliminate potential interference of the K + signal by Na + ions, as outlined by Minta and Tsien in 1989 [16].The experimental design, depicted in Figure 4A, is supported by our previous work [11].In brief, cells were loaded with PBFI-AM and exposed to a low-chloride ([Cl − ]), Na + -free, and high-potassium ([K + ]) external solution to deplete cellular chloride (Figure 4A, condition i).If Slc4a8 and Slc4a10 facilitate K + transport, it would be anticipated that the combination of inwardly directed Cl − and outwardly directed K + gradients (achieved by increasing [Cl − o and decreasing [K + ] o ) would induce Cl − influx in exchange for K + − HCO 3 − efflux (Figure 4A, condition ii).As shown in Figure 4B,C, a sustained reduction in [K + ] i was observed when these gradients were imposed on Slc4a8-and Slc4a10-expressing cells.In contrast, minimal or no K + efflux was detected in non-transfected cells, strongly supporting the proposition that Slc4a8 and Slc4a10 indeed facilitate the transport of K + ions.   1) and then Cl − uptake (and K + efflux) was induced by switching to an external solution containing high Cl − and low K + (solution C, Table 1).Red dashed lines in (B,C) show the activity displayed by non-transfected cells (n = 16).The experiments shown in (B,C) were performed under Na + -free conditions to prevent Na + contamination of the PBFI signal.
Results are presented as the mean ± SEM.

Molecular Dynamics Supports K + Occupancy at the Na + Coordination S1 Site in the Cryo-EM rSlc4a8 Structure
In this study, we have shown evidence supporting K + (or Na + )-driven Cl − /HCO3 − exchanger activity by the murine Slc4a8 (mSlc4a8) and Slc4a10 (mSlc4a10) transporters.In the absence of structural data for murine orthologs of Slc4a8 or Slc4a10 proteins, we turned to the accessible cryo-EM structure of Slc4a8 from Rattus norvegicus (PDB: 7rtm),  1) and then Cl − uptake (and K + efflux) was induced by switching to an external solution containing high Cl − and low K + (solution C, Table 1).Red dashed lines in (B,C) show the activity displayed by non-transfected cells (n = 16).The experiments shown in (B,C) were performed under Na + -free conditions to prevent Na + contamination of the PBFI signal.Results are presented as the mean ± SEM.

Molecular Dynamics Supports K + Occupancy at the Na + Coordination S1 Site in the Cryo-EM rSlc4a8 Structure
In this study, we have shown evidence supporting K + (or Na + )-driven Cl − /HCO 3 − exchanger activity by the murine Slc4a8 (mSlc4a8) and Slc4a10 (mSlc4a10) transporters.In the absence of structural data for murine orthologs of Slc4a8 or Slc4a10 proteins, we turned to the accessible cryo-EM structure of Slc4a8 from Rattus norvegicus (PDB: 7rtm), recently solved in its outward-facing (OF) state [17].
In the rat Slc4a8 structure (rSlc4a8), Na + and CO 3 2− ions are anchored in the S1 site, which is located at the center of the protein.The Na + position is coordinated mainly by residues D800, T804, T847, and the CO 3 2− anion [17].The rSlc4a8 shares 97% sequence identity with mSlc4a8 and 74% with mSlc4a10.In addition, the residues D800, T804, and T847 in rSlc4a8 as well as others from the S1 site are conserved in mSlc4a8 (D798, T802 and T845, Figure S1) and mSlc4a10 (D831, T835, and T878, Figure S1).
Using molecular dynamics (MD) simulations, we explore whether K + could also occupy the natural coordination S1 site for Na + in the cryo-EM rSlc4a8 structure.We replaced Na + by K + in the presence of CO 3 2− .As result, we obtained a rSlc4a8 structure coordinating the K + -CO 3 2− pair, which was used as input for three independent 100 ns MD simulations.In the stable portion of the three trajectories (last 50 ns in Figure S2), the distances between K + and Na-coordinating residues (D800, T804, and T847) were computed to explore the putative coordination of K + ion at the S1 site.
We found that K + remains in contact with D800, T804, and T847 along all MD simulations (Figure 5) and that the K + -residue distances (Figure 6C) are comparable to those reported for Na + (Figure 6D) [17].In addition, we evidenced that K + and CO 3 2− are interacting with an average distance of 3.7 Å (Figure 6C) and that they sustain interactions (Figure 7A) that do not dissociate from the S1 site during 50 ns MD simulations (Figure 7B,C).Distances between each ion to S1 site were calculated from the Cα atom of residue A846 (Figure S1).Using molecular dynamics (MD) simulations, we explore whether K + could also occupy the natural coordination S1 site for Na + in the cryo-EM rSlc4a8 structure.We replaced Na + by K + in the presence of CO3 2− .As result, we obtained a rSlc4a8 structure coordinating the K + -CO3 2− pair, which was used as input for three independent 100 ns MD simulations.In the stable portion of the three trajectories (last 50 ns in Figure S2), the distances between K + and Na-coordinating residues (D800, T804, and T847) were computed to explore the putative coordination of K + ion at the S1 site.
We found that K + remains in contact with D800, T804, and T847 along all MD simulations (Figure 5) and that the K + -residue distances (Figure 6C) are comparable to those reported for Na + (Figure 6D) [17].In addition, we evidenced that K + and CO3 2− are interacting with an average distance of 3.7 Å (Figure 6C) and that they sustain interactions (Figure 7A) that do not dissociate from the S1 site during 50 ns MD simulations (Figure 7B,C).Distances between each ion to S1 site were calculated from the Cα atom of residue A846 (Figure S1).Residues at the S1 site involved in K + coordination.Ribbon representation of (A) MD equilibrated structure of rat Slc4a8 where K + replaces Na + at S1 site from 100 ns MD running and (B) K + interactions at S1 site with the residues originally described coordinating Na + in [17].The residue side chains are shown as sticks, CO3 2− is shown as ball and sticks and, K + is shown as a blue sphere.The transmembrane helices, TM8 and TM10, contributing to ion coordination are labeled.(C).The average distances between K + and residues D800, T804, T847 and the CO3 2− ion computed from MD simulations are indicated (referred as r1, r2, and r3).The K + -residue distances are measured to side chain atoms, in particular, the oxygen atoms (OD1 and OD2) from the carbonyl group of D800 and the oxygen (OG1) from hydroxyls of T804 and T847.The distance between K + and CO3 2− was computed between their center of mass.(D) Na + contacting distance with residues D800, T804, T847, and the CO3 2− ion was taken from Wang et al. [17].Interaction between K + and CO3 2− at the S1 site.Distance between (A) K + and center of mass of CO3 2− , (B) K + to S1 site, and (C) CO3 2− to S1 site along three 100 ns MD simulations (referred to as r1, r2, and r3).The S1-ion distance is defined as the distance between each ion and the Cα atom of residue A846, similar to the definition used by Wang et al. [17].The residue A486 belongs to the S1 site.
In addition, we measured the ion-S1 distances (K-S1 and Na-S1), which correspond to the distances between one permanent ion (Na + or K + ) and the residue A846, providing evidence of whether ions dissociate from their pocket.The A846 residue belongs to S1 site, and the ion-S1 distances are computed between the ion center of mass and the Cα atom of residue A846 (Figure 7B,C), similar to the definition used by Wang et al. [17].

Discussion
Slc4a8 expression is associated with Na + -driven Cl − /HCO3 − exchanger activity [14].Additionally, apart from its role as a Na + -HCO3 − symporter with Cl − self-exchange activity [15], Slc4a10 has been shown to exhibit sodium-driven HCO3 − /Cl − exchange activity [13,18].In summary, the literature suggests that Slc4a8 and Slc4a10 could be involved in Figure 6.Residues at the S1 site involved in K + coordination.Ribbon representation of (A) MD equilibrated structure of rat Slc4a8 where K + replaces Na + at S1 site from 100 ns MD running and (B) K + interactions at S1 site with the residues originally described coordinating Na + in [17].The residue side chains are shown as sticks, CO 3 2− is shown as ball and sticks and, K + is shown as a blue sphere.The transmembrane helices, TM8 and TM10, contributing to ion coordination are labeled.(C).The average distances between K + and residues D800, T804, T847 and the CO 3 2− ion computed from MD simulations are indicated (referred as r1, r2, and r3).The K + -residue distances are measured to side chain atoms, in particular, the oxygen atoms (OD1 and OD2) from the carbonyl group of D800 and the oxygen (OG1) from hydroxyls of T804 and T847.The distance between K + and CO 3 2− was computed between their center of mass.(D) Na + contacting distance with residues D800, T804, T847, and the CO 3 2− ion was taken from Wang et al. [17].

Figure 6.
Residues at the S1 site involved in K + coordination.Ribbon representation of (A) MD equilibrated structure of rat Slc4a8 where K + replaces Na + at S1 site from 100 ns MD running and (B) K + interactions at S1 site with the residues originally described coordinating Na + in [17].The residue side chains are shown as sticks, CO3 2− is shown as ball and sticks and, K + is shown as a blue sphere.The transmembrane helices, TM8 and TM10, contributing to ion coordination are labeled.(C).The average distances between K + and residues D800, T804, T847 and the CO3 2− ion computed from MD simulations are indicated (referred as r1, r2, and r3).The K + -residue distances are measured to side chain atoms, in particular, the oxygen atoms (OD1 and OD2) from the carbonyl group of D800 and the oxygen (OG1) from hydroxyls of T804 and T847.The distance between K + and CO3 2− was computed between their center of mass.(D) Na + contacting distance with residues D800, T804, T847, and the CO3 2− ion was taken from Wang et al. [17].In addition, we measured the ion-S1 distances (K-S1 and Na-S1), which correspond to the distances between one permanent ion (Na + or K + ) and the residue A846, providing evidence of whether ions dissociate from their pocket.The A846 residue belongs to S1 site, and the ion-S1 distances are computed between the ion center of mass and the Cα atom of residue A846 (Figure 7B,C), similar to the definition used by Wang et al. [17].

Discussion
Slc4a8 expression is associated with Na + -driven Cl − /HCO3 − exchanger activity [14].In addition, we measured the ion-S1 distances (K-S1 and Na-S1), which correspond to the distances between one permanent ion (Na + or K + ) and the residue A846, providing evidence of whether ions dissociate from their pocket.The A846 residue belongs to S1 site, and the ion-S1 distances are computed between the ion center of mass and the Cα atom of residue A846 (Figure 7B,C), similar to the definition used by Wang et al. [17].

Discussion
Slc4a8 expression is associated with Na + -driven Cl − /HCO 3 − exchanger activity [14].Additionally, apart from its role as a Na + -HCO 3 − symporter with Cl − self-exchange activity [15], Slc4a10 has been shown to exhibit sodium-driven HCO 3 − /Cl − exchange activity [13,18].In summary, the literature suggests that Slc4a8 and Slc4a10 could be involved in regulating intracellular pH and chloride concentrations, thereby implying their contribution to the modulation of synaptic transmission through ion homeostasis mechanisms.
Synaptic glutamate release critically depends on the activity of the Na + -driven Cl − /HCO 3 − exchanger Slc4a8 [19].Moreover, the absence of Slc4a10 in knockout mice disrupts GABAergic transmission, with glutamatergic transmission unaffected [20].The importance of Slc4a10 in synaptic function is underscored by revealing impaired acid extrusion in humans with mutations in the SLC4A10 gene [21].Together, Slc4a8 and Slc4a10 influence neurotransmission at the presynaptic level, where they modulate neurotransmitter release through a mechanism that likely depends on sodium-coupled bicarbonate transport.
Our findings demonstrate that the activities associated with Slc4a8 and Slc4a10 involve HCO 3 − -dependent Cl − flux (Figure 2).Furthermore, a decrease in external Cl − concentration led to an elevation in intracellular pH in a HCO 3 − -dependent manner in cells expressing Slc4a8 and Slc4a10, indicating Cl − /HCO 3 − exchanger activity (Figure 3).Intriguingly, both Na + and K + ions were found to drive Cl − /HCO 3 − exchange in cells expressing Slc4a8 and Slc4a10, suggesting that either Na + or K + may be transported in the same direction as HCO 3 − ions (Figure 3).Additional experiments conducted in cells expressing Slc4a8 and Slc4a10 and loaded with the K + indicator PBFI-AM revealed K + fluxes associated with the expression of Slc4a8 and Slc4a10 (Figure 4), indicating that K + ions might be transported by a mechanism similar to that shown for Na + ions transported by the rat Slc4a8 protein [17].
Utilizing the rat Slc4a8 structure [17], we conducted MD calculations, which provide evidence supporting K + binding to the S1 site in rSlc4a8, indicating that favorable interactions for Na + coordination are also favorable to K + binding.Given the conservation of residues in mSlc4a8, mSlc4a10 and rSlc4a8, we anticipate that mSlc4a8 and mSlc4a10 could also bind K + in the sodium coordination pocket using residues D798, T802, and T845 in mSlc4a8, and residues D831, T835, and T878 in mSlc4a10 (equivalent counterparts of D800, T804, and T847 in rSlc4a8, respectively).Moreover, our calculations showed that there is an interaction between K + and CO 3 2− at the S1 site (3.7 Å distance).Collectively, molecular dynamics calculations support the notion of K + occupying a cation transport pathway, thereby mediating Cl − /K + -HCO 3 − (or CO 3 2− ) exchanger activity of mSlc4a8 and mSlc4a10 transporter proteins.
This K + dependence exhibited by Slc4a8 and Slc4a10 is reminiscent of that described for Ae4, a Cl − /HCO 3 − exchanger driven by either Na + or K + [11].However, it is important to note that this non-selective cation feature is not conserved in all Na + -driven Cl − /HCO 3 − exchangers, as evidenced by the lack of K + driving anion exchange in the squid Na + -driven Cl − /HCO 3 − exchanger (sqNDCBE) [22].

Figure 1 .
Figure 1.Slc4a8 and Slc4a10 localization in CHO-K1 cells.Immunofluorescence studies in CHO-K1 cells transfected with plasmids encoding Myc-DDK tagged versions of Slc4a8 (upper panels) and Slc4a10 (middle panels).Transfected cells with the empty plasmid were used as negative control (lower panels).Plasma membrane was labeled with WGA conjugated to Alexa Fluor 633 TM (shown in red, left panels), followed by protein detection using an anti-DYKDDDDK antibody (shown in green, middle panels).Pictures in the right panels were generated by merging WGA, DYKDDDDK and DAPI channels.Scale bar = 10 µm.

Figure 1 .
Figure 1.Slc4a8 and Slc4a10 localization in CHO-K1 cells.Immunofluorescence studies in CHO-K1 cells transfected with plasmids encoding Myc-DDK tagged versions of Slc4a8 (upper panels) and Slc4a10 (middle panels).Transfected cells with the empty plasmid were used as negative control (lower panels).Plasma membrane was labeled with WGA conjugated to Alexa Fluor 633 TM (shown in red, left panels), followed by protein detection using an anti-DYKDDDDK antibody (shown in green, middle panels).Pictures in the right panels were generated by merging WGA, DYKDDDDK and DAPI channels.Scale bar = 10 µm.

Figure 2 .
Figure 2. Slc4a8 and Slc4a10 mediate Cl − /HCO3 − exchanger activity.Cl − fluxes in Slc4a8-and Slc4a10expressing cells loaded with SPQ.Electroporated CHO-K1 cells were perfused with HCO3 − -containing/high Cl − solution (Solution A, Table1) and then switched to a HCO3 − -containing/low Cl − solution (Solution B, Table1).Fluxes obtained from cells electroporated with Slc4a8-((A), n = 9) or Slc4a10 ((B), n = 9) encoding plasmids are shown in black circles.Fluxes obtained from non-electroporated cells are shown as red dashed lines and are the same in A and B (n = 11).Experiments under HCO3 −free conditions were also performed on Slc4a8 ((A), open circles; n = 5) and Slc4a10 ((B), open circles; n = 5) electroporated cells, which were initially perfused with HCO3 − -free/high Cl − solution (Solution E) and then switched to HCO3 − -free/low Cl − solution (Solution F).Fluxes in each condition are shown as the mean ± standard error of mean (SEM).
Figure 2. Slc4a8 and Slc4a10 mediate Cl − /HCO3 − exchanger activity.Cl − fluxes in Slc4a8-and Slc4a10expressing cells loaded with SPQ.Electroporated CHO-K1 cells were perfused with HCO3 − -containing/high Cl − solution (Solution A, Table1) and then switched to a HCO3 − -containing/low Cl − solution (Solution B, Table1).Fluxes obtained from cells electroporated with Slc4a8-((A), n = 9) or Slc4a10 ((B), n = 9) encoding plasmids are shown in black circles.Fluxes obtained from non-electroporated cells are shown as red dashed lines and are the same in A and B (n = 11).Experiments under HCO3 −free conditions were also performed on Slc4a8 ((A), open circles; n = 5) and Slc4a10 ((B), open circles; n = 5) electroporated cells, which were initially perfused with HCO3 − -free/high Cl − solution (Solution E) and then switched to HCO3 − -free/low Cl − solution (Solution F).Fluxes in each condition are shown as the mean ± standard error of mean (SEM).

Figure 2 .
Figure 2. Slc4a8 and Slc4a10 mediate Cl − /HCO 3 − exchanger activity.Cl − fluxes in Slc4a8-and Slc4a10-expressing cells loaded with SPQ.Electroporated CHO-K1 cells were perfused with HCO 3 −containing/high Cl − solution (Solution A, Table 1) and then switched to a HCO 3 − -containing/low Cl − solution (Solution B, Table 1).Fluxes obtained from cells electroporated with Slc4a8-((A), n = 9) or Slc4a10 ((B), n = 9) encoding plasmids are shown in black circles.Fluxes obtained from nonelectroporated cells are shown as red dashed lines and are the same in A and B (n = 11).Experiments under HCO 3 − -free conditions were also performed on Slc4a8 ((A), open circles; n = 5) and Slc4a10 ((B), open circles; n = 5) electroporated cells, which were initially perfused with HCO 3 − -free/high Cl − solution (Solution E) and then switched to HCO 3 − -free/low Cl − solution (Solution F).Fluxes in each condition are shown as the mean ± standard error of mean (SEM).

Figure 3 .
Figure 3. Slc4a8 and Slc4a10 display Na + -and K + -dependent Cl − /HCO3 − exchanger activity.(A,C) Time course showing that Na + and K + elicited Cl − /HCO3 − exchange in Slc4a8-(A) and Slc4a10-expressing cells (C) loaded with BCECF.Cells were perfused with NMDG-containing high Cl − solution (Solution (C), Table1) followed by NMDG-containing low Cl − solution (Solution (D), Table1) and finally switched to Na + (Solution B, black circles; n = 10 for Slc4a8 and Slc4a10) or K + -containing solution (Solution G, open circles; n = 9 for Slc4a8 and n = 10 for Slc4a10).Fluxes obtained from nonelectroporated cells are shown as red dashed lines and are the same in A and C (n = 16).(B,D) Slc4a8 (B) and Slc4a10 (D) activities were measured as the alkalinization rate in response to a reduction in external Cl − concentration using Na + -and K + -containing solutions (change from solution A to solutions 5, 25, 50, 100, 125 and 150 mM [see Table 2 for ion composition of the Na + -and K + -containing solutions]).The normalized activities at different Na + (black circles) and K + concentrations (open circles) were plotted against the cation concentration and then fitted to a Hill function (shown as black lines; see Methods for the equation used).Values correspond to the average ± SEM.
Figure 3. Slc4a8 and Slc4a10 display Na + -and K + -dependent Cl − /HCO3 − exchanger activity.(A,C) Time course showing that Na + and K + elicited Cl − /HCO3 − exchange in Slc4a8-(A) and Slc4a10-expressing cells (C) loaded with BCECF.Cells were perfused with NMDG-containing high Cl − solution (Solution (C), Table1) followed by NMDG-containing low Cl − solution (Solution (D), Table1) and finally switched to Na + (Solution B, black circles; n = 10 for Slc4a8 and Slc4a10) or K + -containing solution (Solution G, open circles; n = 9 for Slc4a8 and n = 10 for Slc4a10).Fluxes obtained from nonelectroporated cells are shown as red dashed lines and are the same in A and C (n = 16).(B,D) Slc4a8 (B) and Slc4a10 (D) activities were measured as the alkalinization rate in response to a reduction in external Cl − concentration using Na + -and K + -containing solutions (change from solution A to solutions 5, 25, 50, 100, 125 and 150 mM [see Table 2 for ion composition of the Na + -and K + -containing solutions]).The normalized activities at different Na + (black circles) and K + concentrations (open circles) were plotted against the cation concentration and then fitted to a Hill function (shown as black lines; see Methods for the equation used).Values correspond to the average ± SEM.

Figure 3 .
Figure 3. Slc4a8 and Slc4a10 display Na + -and K + -dependent Cl − /HCO 3 − exchanger activity.(A,C) Time course showing that Na + and K + elicited Cl − /HCO 3 − exchange in Slc4a8-(A) and Slc4a10expressing cells (C) loaded with BCECF.Cells were perfused with NMDG-containing high Cl − solution (Solution (C), Table 1) followed by NMDG-containing low Cl − solution (Solution (D), Table 1) and finally switched to Na + (Solution B, black circles; n = 10 for Slc4a8 and Slc4a10) or K + -containing solution (Solution G, open circles; n = 9 for Slc4a8 and n = 10 for Slc4a10).Fluxes obtained from non-electroporated cells are shown as red dashed lines and are the same in A and C (n = 16).(B,D) Slc4a8 (B) and Slc4a10 (D) activities were measured as the alkalinization rate in response to a reduction in external Cl − concentration using Na + -and K + -containing solutions (change from solution A to solutions 5, 25, 50, 100, 125 and 150 mM [see Table 2 for ion composition of the Na + -and K + -containing solutions]).The normalized activities at different Na + (black circles) and K + concentrations (open circles) were plotted against the cation concentration and then fitted to a Hill function (shown as black lines; see Methods for the equation used).Values correspond to the average ± SEM.

Figure 4 .
Figure 4. Slc4a8− and Slc4a10−dependent K + -fluxes.(A) Experimental design used to assess K + transport by Slc4a8 and Slc4a10.(B,C) Slc4a8-((B), black circles, n = 7) and Slc4a10-expressing cells (C, black circles, n = 9) loaded with PBFI-AM were depleted of Cl − by incubating the cells with low Cl − solution and high K + (solution G, Table1) and then Cl − uptake (and K + efflux) was induced by switching to an external solution containing high Cl − and low K + (solution C, Table1).Red dashed lines in (B,C) show the activity displayed by non-transfected cells (n = 16).The experiments shown in (B,C) were performed under Na + -free conditions to prevent Na + contamination of the PBFI signal.Results are presented as the mean ± SEM.

Figure 4 .
Figure 4. Slc4a8− and Slc4a10−dependent K + -fluxes.(A) Experimental design used to assess K + transport by Slc4a8 and Slc4a10.(B,C) Slc4a8-((B), black circles, n = 7) and Slc4a10-expressing cells (C, black circles, n = 9) loaded with PBFI-AM were depleted of Cl − by incubating the cells with low Cl − solution and high K + (solution G, Table1) and then Cl − uptake (and K + efflux) was induced by switching to an external solution containing high Cl − and low K + (solution C, Table1).Red dashed lines in (B,C) show the activity displayed by non-transfected cells (n = 16).The experiments shown in (B,C) were performed under Na + -free conditions to prevent Na + contamination of the PBFI signal.Results are presented as the mean ± SEM.

Figure 5 .
Figure5.K + occupancy at the S1 site in the cryo-EM rSlc4a8 structure.Distances between K + and the residues from S1 Na + -coordination site were computed within a stable range along three 100 ns MD simulations (referred to as r1, r2, and r3).Residues for distance calculation include D800 and its carbonyl side chain atoms OD1 and OD2 (A,B), and the respective hydroxyl OG1 side chain atoms from T804 (C), and T847 (D).

Figure 5 .
Figure5.K + occupancy at the S1 site in the cryo-EM rSlc4a8 structure.Distances between K + and the residues from S1 Na + -coordination site were computed within a stable range along three 100 ns MD simulations (referred to as r1, r2, and r3).Residues for distance calculation include D800 and its carbonyl side chain atoms OD1 and OD2 (A,B), and the respective hydroxyl OG1 side chain atoms from T804 (C), and T847 (D).

Figure 6 .
Figure 6.Residues at the S1 site involved in K + coordination.Ribbon representation of (A) MD equilibrated structure of rat Slc4a8 where K + replaces Na + at S1 site from 100 ns MD running and (B) K + interactions at S1 site with the residues originally described coordinating Na + in[17].The residue side chains are shown as sticks, CO3 2− is shown as ball and sticks and, K + is shown as a blue sphere.The transmembrane helices, TM8 and TM10, contributing to ion coordination are labeled.(C).The average distances between K + and residues D800, T804, T847 and the CO3 2− ion computed from MD simulations are indicated (referred as r1, r2, and r3).The K + -residue distances are measured to side chain atoms, in particular, the oxygen atoms (OD1 and OD2) from the carbonyl group of D800 and the oxygen (OG1) from hydroxyls of T804 and T847.The distance between K + and CO3 2− was computed between their center of mass.(D) Na + contacting distance with residues D800, T804, T847, and the CO3 2− ion was taken from Wang et al.[17].

Figure 7 .
Figure 7. Interaction between K + and CO3 2− at the S1 site.Distance between (A) K + and center of mass of CO3 2− , (B) K + to S1 site, and (C) CO3 2− to S1 site along three 100 ns MD simulations (referred to as r1, r2, and r3).The S1-ion distance is defined as the distance between each ion and the Cα atom of residue A846, similar to the definition used by Wang et al. [17].The residue A486 belongs to the S1 site.

Figure 7 .
Figure 7. Interaction between K + and CO3 2− at the S1 site.Distance between (A) K + and center of mass of CO3 2− , (B) K + to S1 site, and (C) CO3 2− to S1 site along three 100 ns MD simulations (referred to as r1, r2, and r3).The S1-ion distance is defined as the distance between each ion and the Cα atom of residue A846, similar to the definition used by Wang et al. [17].The residue A486 belongs to the S1 site.

Figure 7 .
Figure 7. Interaction between K + and CO 3 2− at the S1 site.Distance between (A) K + and center of mass of CO 3 2− , (B) K + to S1 site, and (C) CO 3 2− to S1 site along three 100 ns MD simulations(referred to as r1, r2, and r3).The S1-ion distance is defined as the distance between each ion and the Cα atom of residue A846, similar to the definition used by Wang et al.[17].The residue A486 belongs to the S1 site.

Table 1 .
Solutions used in this study.

Table 2 .
Solutions used in cation selectivity experiments.

Table 2 .
Solutions used in cation selectivity experiments.