Direct endosomal acidification by the outwardly rectifying CLC-5 Cl(-)/H(+) exchanger

Abstract

The voltage-gated Cl(-) channel (CLC) family comprises cell surface Cl(-) channels and intracellular Cl(-)/H(+) exchangers. CLCs in organelle membranes are thought to assist acidification by providing a passive chloride conductance that electrically counterbalances H(+) accumulation. Following recent descriptions of Cl(-)/H(+) exchange activity in endosomal CLCs we have re-evaluated their role. We expressed human CLC-5 in HEK293 cells, recorded currents under a range of Cl(-) and H(+) gradients by whole-cell patch clamp, and examined the contribution of CLC-5 to endosomal acidification using a targeted pH-sensitive fluorescent protein. We found that CLC-5 only conducted outward currents, corresponding to Cl(-) flux into the cytoplasm and H(+) from the cytoplasm. Inward currents were never observed, despite the range of intracellular and extracellular Cl(-) concentrations and pH used. Endosomal acidification in HEK293 cells was prevented by 25 microm bafilomycin-A1, an inhibitor of vacuolar-type H(+)-ATPase (v-ATPase), which actively pumps H(+) into the endosomal lumen. Overexpression of CLC-5 in HEK293 cells conferred an additional bafilomycin-insensitive component to endosomal acidification. This effect was abolished by making mutations in CLC-5 that remove H(+) transport, which result in either no current (E268A) or bidirectional Cl(-) flux (E211A). Endosomal acidification in a proximal tubule cell line was partially sensitive to inhibition of v-ATPase by bafilomycin-A1. Furthermore, in the presence of bafilomycin-A1, acidification was significantly reduced and nearly fully ablated by partial and near-complete knockdown of endogenous CLC-5 by siRNA. These data suggest that CLC-5 is directly involved in endosomal acidification by exchanging endosomal Cl(-) for H(+).

Figure 1. Properties of CLC-5 expressed in HEK293 cells

A, whole-cell currents recorded from untransfected cells or those expressing CLC-5-EYFP or unfused CLC-5 as indicated. Cells were held at −30 mV and 10 ms test pulses were applied at 10 mV increments between −100 mV and +100 mV as shown by the voltage protocol below the representative traces. B, mean (±s.e.m., n= 6–7 cells) current density–voltage relationships of cells expressing CLC-5-EYFP or CLC-5 compared to untransfected HEK293 cells. C, tail currents were evoked by 3 ms pre-pulses to +200 mV followed by 5 ms test pulses from +200 mV to −200 mV at 10 mV increments. The dotted line denotes the zero current level (0).

Figure 2. Effects of changing extracellular and intracellular [Cl⁻] on whole-cell CLC-5 currents

Cells were held at −30 mV and 10 ms test pulses were applied at 10 mV increments to between −100 mV and +200 mV. A, representative current families recorded using the conditions indicated by the whole-cell patch clamp schematic diagrams denoting the intracellular and extracellular [Cl⁻] in mm. B, mean (±s.e.m., n= 7 cells) current density–voltage relationships using an extracellular solution containing 140 mm Cl⁻ and varying the intracellular [Cl⁻], in mm as indicated in the key. C, mean (±s.e.m., n= 6–7 cells) current density–voltage relationships using an intracellular solution containing 42 mm Cl⁻ and varying the extracellular [Cl⁻], in mm as indicated in the key. The continuous lines are Boltzmann functions (eqn (3)) fitted directly to the mean current density–voltage plots, using the predicted current reversal potential (eqn (2)).

Figure 3. Effects of changing extracellular and intracellular pH on whole-cell CLC-5 currents

Cells were held at −30 mV and 10 ms test pulses were applied at 10 mV increments to between −100 mV and +200 mV. A, representative current families from whole-cell recordings using conditions indicated by the whole-cell patch clamp schematic diagrams denoting the intracellular and extracellular pH as indicated. B, mean (±s.e.m., n= 6–7 cells) current density–voltage relationships when the extracellular solution had pH 7.4 and the intracellular pH was varied, as indicated. C, mean (±s.e.m., n= 6–7 cells) current density–voltage relationships when the intracellular solution had pH 7.4 and the extracellular pH was varied, as indicated. The continuous lines are Boltzmann functions (eqn (3)) fitted directly to the mean current density–voltage plots.

Figure 4. CLC-5 activation curves with the various extracellular and intracellular [Cl⁻] and pH

The conductance G was calculated using the predicted reversal potential (eqn (2)), the Boltzmann function fitted to the conductance–voltage relationship, and data presented relative to the maximum conductance G_max. A, mean (±s.e.m.) activation curves when the extracellular [Cl⁻] was varied, as indicated, in mm. The continuous lines are Boltzmann functions fitted to the mean data. B, mean activation curves when the intracellular [Cl⁻] was varied, as indicated. C, the mean (±s.e.m.) V_1/2 values returned by the Boltzmann fits to the individual data sets plotted against extracellular and intracellular [Cl⁻]. Mean activation curves when the extracellular (D) and intracellular (E) pH were varied. F, mean activation V_1/2 with the different extracellular and intracellular pH.

Figure 5. Cl⁻/H⁺ exchange through CLC-5 enhances endosomal acidification

A, endosomal pH, which is regulated by the activity of both v-ATPase and CLC-5, was measured using ratiometric pHluorin targeted to the luminal face of the endosomal membrane as a VAMP2-pHluorin fusion protein. B, confocal images showing VAMP2-pHluorin (green, upper inset image) and HA-CLC-5 (red, lower inset image) co-localised (yellow, main image) in fixed HEK293 cells (scale bar = 10 μm). C, endosomal pH (mean ±s.e.m., n > 20 cells) measured using VAMP2-pHluorin transfected into HEK293 cells alone (Ctrl) or with HA-CLC-5 (CLC-5). Where indicated (+baf), cells were treated with 25 μm bafilomycin-A1 for 1 h prior to and during the experiment to inhibit v-ATPase, or were untreated (–baf). The dotted line represents the pH of the bathing medium (pH 7.4). Statistical significance is shown with #P < 0.05 vs. CLC-5 (–baf), and *P < 0.05 vs. control (–baf) (ANOVA). D, current density–voltage relationships of whole-cell recordings from HEK293 cells expressing wild-type (WT), E211A, or E268A CLC-5-EYFP. Cells were held at −30 mV and 10 ms test pulses were applied at 10 mV increments between −100 mV and +100 mV. E, confocal images showing co-localisation (yellow, main images) of VAMP2-pHluorin (green, inset images) with either E211A or E268A mutant HA-CLC-5 (red, inset images) in fixed HEK293 cells (scale bar = 10 μm). F, endosomal acidification (mean ±s.e.m., n≥ 18 cells) in HEK293 cells transfected with VAMP2-pHluorin alone (Ctrl) or with either E211A or E268A mutant HA-CLC-5. Where indicated, cells were treated with 25 μm bafilomycin-A1 (+baf) for 1 h prior to and during the experiment to inhibit v-ATPase, or were untreated (–baf). The dashed line represents the pH of the bathing medium (pH 7.4). *P < 0.05 (Student's t test).

Figure 6. Effects of CLC-5 knockdown on endosomal acidification in OK cells

OK cells were either mock transfected (M), transfected with control (C), or transfected with CLC-5-targeted siRNA1, -2 and -3 (s1, s2, s3, respectively). A, knockdown of endogenous CLC-5 in OK cells by siRNA determined by qRT-PCR 72 h post-transfection of CLC-5-targeted or control siRNA duplexes as indicated. The abundance of CLC-5 cDNA was normalised to that of β-actin and expressed relative to the ratio in mock transfected cells (mean ±s.e.m., n= 4 experiments). *P < 0.05 vs. mock-transfected cells. Products from the qRT-PCR reaction resolved by gel electrophoresis are shown below. B, reduced CLC-5 protein levels in similarly treated OK cells determined by Western blotting. Equal amounts of protein from OK cell lysates were loaded into each lane and proteins detected by anti-CLC-5 and anti-TRAM1 are shown in the top and bottom Western blot, respectively. Data are representative of three blots. C, endosomal pH (mean ±s.e.m., n≥ 16 cells) measured from OK cells that were treated as in A and co-transfected with VAMP2-pHluorin. Where indicated, cells were treated with 25 μm bafilomycin-A1 (+baf, filled bars) for 1 h prior to and during the experiment to inhibit v-ATPase, or were untreated (–baf, open bars). The dashed line represents the pH of the bathing medium (pH 7.4). *P < 0.05 –baf vs.+baf and #P < 0.05 vs. corresponding treatments in mock transfected cells. D, correlation between CLC-5 expression (from A) and endosomal acidification (from C) in bafilomycin-A1-treated (filled symbols) and untreated (open symbols) cells that were mock transfected (squares), or transfected with control (circles) or with siRNA1 (triangles), siRNA2 (inverted triangles), or siRNA3 (diamonds) CLC-5-targeted siRNA. The continuous fitted lines are second-order polynomials and the dashed line represents the pH of the bathing medium (pH 7.4).

Figure 7. Proposed roles of CLC-5 in endosomal acidification: Cl⁻ shunt (A) versus direct acidification (B)

CLC-5 (light grey shape) and v-ATPase (dark grey shape) are depicted in the endosomal membrane. In A, two Cl⁻ ions are transported into the vesicle lumen in exchange for a single proton. This provides an anion shunt for the pumping of protons into the endosome by v-ATPase since the 2Cl⁻ to 1H⁺ gives a net influx of 3 negative charges. However, for every two H⁺ pumped into the endosome by v-ATPase, one is removed by CLC-5. In B, CLC-5 exchanges two Cl⁻ ions from the endosomal lumen for a proton from the cytoplasm. This leads to endosomal acidification directly by CLC-5 and in parallel with v-ATPase. The compensatory shunt conductance is provided by an unidentified ion channel or transporter (open circle).