Stable, polarised, functional expression of Kir1.1b channel protein in Madin-Darby canine kidney cell line

Abstract

1. The family of Kir1.1 (ROMK) channel proteins constitute a secretory pathway for potassium in principal cells of cortical collecting duct and thick ascending limb of Henle's loop. Mutations in Kir1.1 account for some types of Bartter's syndrome.

2. Here we report that stable transfection of Kir1.1b (ROMK2) in Madin-Darby canine kidney (MDCK) cell line results in expression of inwardly rectifying K⁺ currents and transmonolayer electrical and transport properties appropriate to Kir1.1 function. When grown on permeable supports, transfected monolayers secreted K⁺ into the apical solution. This secretion was inhibited by application of barium to the apical membrane, or by reduction in expression temperature from 37 to 26°C. However, whole-cell voltage clamp electrophysiology showed that K⁺ conductance was higher in cells expressing Kir1.1b at 26°C.

3. To investigate this further, Kir1.1b was tagged with (EGFP), a modification that did not affect channel activity. Protein synthesis was inhibited with cycloheximide. Spectrofluorimetry was used to compare protein degradation at 37 and 26°C. The increased level of Kir1.1b at the plasma membrane at 26°C was due to an increase in protein stability.

4. Confocal microscopic investigation of EGFP-Kir1.1b fluorescence in transfected cells showed that the channel protein was targeted to the apical domain of the cell.

5. These results demonstrate that Kir1.1b is capable of appropriate trafficking and function in MDCK cell lines at physiological temperatures. In addition, expression of Kir1.1b in MDCK cell lines provides a useful and convenient tool for the study of functional activity and targeting of secretory K⁺ channels.

Potassium secretion in the kidney takes place in the distal segment of the nephron via both low and intermediate conductance ATP-sensitive K⁺ channels (Frindt & Palmer, 1989; Wang et al. 1997). One type of these channels, Kir1.1b (ROMK2) is likely to constitute the major secretory pathway for potassium in principal cells of the cortical collecting duct (Palmer et al. 1997). These low conductance channels are characterised by high, voltage independent open probability, mild inward rectification, high K⁺ selectivity and pH sensitivity (Ho et al. 1993; Tsai et al. 1995). They are inhibited by barium (Ba²⁺) but not tetraethylammonium (Ho et al. 1993) and ATP has a dual regulatory effect; low concentrations are necessary to avoid channel run-down (via PKA phosphorylation) but higher concentrations result in channel inhibition (McNicholas et al. 1994; Xu et al. 1996). When expressed in oocytes of Xenopus laevis, the properties of Kir1.1 channels are not absolutely identical to the native secretory K⁺ channel (Ho, 1998; Ruknudin et al. 1998). Moreover, it has been demonstrated that single channel conductance, K⁺ selectivity, as well as ATP and sulphonylurea sensitivity are all modulated when Kir1.1 is co-expressed with members of the ATP-binding cassette family of proteins such as cystic fibrosis transmembrane conductance regulator (CFTR) (McNicholas et al. 1996; Ho, 1998; Ruknudin et al. 1998), and the sulphonylurea receptor (Ammala et al. 1996). Both of these proteins are expressed in the distal nephron (Crawford et al. 1991; Morales et al. 1996; Beesley et al. 1999), prompting speculation that the properties of the native channel may reflect an interaction between Kir1.1 and ABC proteins in vivo (Ho, 1998; Ruknudin et al. 1998). Cells of the distal nephron express inwardly rectifying K⁺ channels in their basolateral membranes (Wang et al. 1994). However, it is clear that these channels are not Kir1.1, since this protein is expressed only at the apical membrane in native tissue (Xu et al. 1997; Mennit et al. 1997). Recently, Kir4.2, a K⁺ channel protein with 47 % homology to Kir1.1 has been localised to the basolateral membrane of cells in the distal convoluted tubule (Ito et al. 1996). Little is known regarding the factors that determine membrane targeting of K⁺ channels in epithelial cells. Mutations resulting in loss of function in the Kir1.1 gene have been described in several kindred genes that are affected by the antenatal form of Bartter's syndrome (Simon et al. 1996). At least one of these mutations results in abnormal membrane trafficking of the protein (Schwalbe et al. 1998). Recently, it has been suggested that Kir1.1a is biosynthetically labile and incapable of efficient trafficking to the plasma membrane of MDCK cells at 37°C, suggesting that Kir1.1 channels may require other factors, such as association of a surrogate subunit for appropriate biosynthetic processing (Brejon et al. 1999). In this study we report that stable transfection of Kir1.1b in Madin-Darby canine kidney (MDCK) cell line strain I (Barker & Simmons, 1981) results in cell monolayers that at physiological temperature secrete potassium and display electrical properties consistent with expression of Kir1.1b protein at the apical plasma membrane. In these cell lines, Kir1.1b is appropriately sorted to the apical membrane when tagged with enhanced green fluorescent protein (EGFP) at its N-terminus and functions normally. These new cell lines are likely to be useful models for studying the mechanisms determining vectorial membrane trafficking of renal K⁺ channels, as well as the basis for some forms of renal K⁺ channelopathies.

METHODS

Cell culture

High resistance (strain I) MDCK cells (gift from N. L. Simmons, Newcastle, UK) were grown in a 1:1 (v/v) mixture of Dulbecco's modified Eagle's medium with 4.5 g (l glucose)⁻¹ and F12 (Ham) supplemented with 5 mm Hepes, 25 mm NaHCO₃ and 10 % fetal calf serum. Cells were maintained at 37°C in a 95 % air-5 % CO₂ atmosphere.

Cell transfection

DNA encoding Kir1.1b was subcloned into pcDNA3 (Invitrogen) or pEGFP-C2 (Clontech) to produce either Kir1.1b or an N-terminal fusion protein of Kir1.1b and EGFP. These vector constructs were confirmed by sequencing. DNA was purified (Nucleobond AX: Macherey-Nagel, Biogene Ltd, Kimbolton, Cambridgeshire, UK) and transfections were carried out with SuperFect (Qiagen) according to manufacturer's instructions. Stable transfectants were selected with geneticin at 300 μg μl⁻¹ and the resulting cell lines were named MDCKR2 when expressing untagged Kir1.1b and MDCKER2 when expressing EGFP-Kir1.1b. Control cells, transfected with pcDNA3 or pEGFP-C2 (expressing EGFP alone) were called MDCKP and MDCKE, respectively.

Expression of the appropriate mRNAs in each cell line was confirmed by RT-PCR using primers and protocols described previously (Beesley et al. 1998).

Whole-cell clamp

Cells were grown under standard culture conditions to subconfluence on glass cover slips and placed in a Perspex bath on the stage of an inverted microscope (Olympus IX70). Standard techniques (Wang et al. 1989) were employed to investigate whole-cell K⁺ currents. Briefly, voltage protocols were applied from an IBM-compatible computer equipped with a Digidata 1200A interface and pCLAMP6 software, (Clampex: Axon Instruments). Current recordings were made using an EPC-7 amplifier (List Instruments) and currents saved directly onto the hard disk of the computer following low-pass filtering at 5 kHz. Cell membrane potential was held initially at −40 mV and then stepped to potentials between +10 and −100 mV in 10 mV steps. Mean steady-state currents were derived at the end of each potential step using pCLAMP software (Clampfit). Whole-cell conductance was calculated over the linear part of the I–V curves, from −40 mV to −100 mV, and reversal potentials (V_rev) of ohmic and rectifying currents were calculated by linear or polynomial regression as appropriate. The bath contained a high Na⁺ solution that contained (mm): 135 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 5 glucose, 10 Hepes (titrated to pH 7.4 with 1 m NaOH) and 10 mannitol. When 5 mm BaCl₂ or 2 mm TEA-Cl (tetraethyl ammonium chloride) were added to the bath, mannitol was omitted as appropriate to maintain isosmolality. The pipette solution was a high K⁺ solution which contained (mm): 140 KCl, 2 MgCl₂, 2 Na₂ATP, 1 EGTA and 10 Hepes (titrated to pH 7.4 with 1 m KOH). The osmolality of all solutions was adjusted to 290 mosmol (kg H₂O)⁻¹ with water or mannitol as appropriate.

K⁺ transport

Potassium transport was demonstrated as described previously (White et al. 1992). Cells were seeded on 10 mm transwell filters (Nunc) at a density of 4 × 10⁵ cells cm⁻² and grown to confluence in normal cell culture medium (5 ml in the basolateral side and 300 μl in the apical compartment). Cell culture medium was changed every day. For potassium transport experiments, cells were incubated for 12 h in a sulphate free, buffered saline solution containing (mm): 130 NaCl, 4.2 KCl, 1.2 CaCl₂, 0.6 MgCl₂, 18 mannitol, 25 NaHCO₃, 5 glucose, 5 Hepes (pH to 7.4 with 1 m NaOH). When BaCl₂ was added, an isosmotically equivalent amount of mannitol was omitted to maintain osmolality (290 mosmol (kg H₂O)⁻¹). After a 12 h incubation, apical and basolateral solutions were sampled and potassium concentration was measured by flame photometry (Corning, Costar, High Wycombe, Buckinghamshire, UK). Net transmonolayer K⁺ movement was expressed as Δ[K⁺] in millimolar.

Intracellular K⁺ concentration

Intracellular K⁺ concentration was measured by a method adapted from Montrose and Murer (Montrose & Murer, 1986). Cells were plated on six multiwell plates (1 × 10⁶ cells (well)⁻¹) and incubated for one day in standard culture conditions. Cells were washed twice in a buffer containing 10 mm Tris-HCl pH 7.0, 150 mm NaCl and then lysed in 0.5 ml of a buffer containing 10 mm Tris-HCl pH 7.0 and 1 % Triton-X100. After centrifugation at 14000 g for 15 min at 4°C, total protein from the supernatant was determined as described below, and [K⁺] was measured by flame photometry. Intracellular K⁺ concentration was expressed as millimoles K⁺ per gram of protein in cells grown at 37°C and in cells switched for 12 h to 26°C.

Transmonolayer electrical properties

Cells were grown to confluence on 25 mm transwell filters (Nunc) under standard culture conditions. Filters were mounted in a modified Ussing chamber, using a solution containing (mm): 130 NaCl, 4.2 KCl, 1.2 CaCl₂, 0.6 MgCl₂, 5 Hepes, 5 glucose, pH 7.4 and osmolality 290 mosmol (kg H₂O)⁻¹. Both transmonolayer potential difference (V_tm; mV) and short circuit current (I_tm; μA) were monitored with a voltage-current clamp unit (VCC600: Physiologic Instruments, CA, USA). In order to calculate transmonolayer resistance (R_tm; kΩ cm²) a standard pulse of 50 μA was applied, and the resultant change in voltage was recorded.

Western blot

Confluent MDCK monolayers growing at 26°C on a T75 tissue culture flask were lysed in 1 ml of ice cold RIPA buffer (1 % Nonidet P40, 0.5 % sodium deoxycholate, 0.1 % SDS, 1 % protease cocktail inhibitor (Sigma) in PBS) and then scraped off and passed through a 21 gauge needle. The resulting homogenate was cleared by centrifugation at 15000 g for 20 min at 4°C. The protein concentration of the resulting lysate was determined using the Biorad detergent-compatible protein assay. Samples were prepared for electrophoresis by adding × 2 concentrated loading buffer (250 mm Tris (pH 10.9), 2 % 2-mercaptoethanol, 2 % SDS, 20 % glycerol, 0.01 % bromophenol blue) and boiling for 5 min. Electrophoresis was performed on a 7 % SDS-polyacrylamide gel with 15 μg of protein loaded in each lane. The proteins were subsequently blotted onto a 0.45 μm nitro-cellulose membrane (Biorad) which was then blocked in PBS containing 0.1 % Tween (PBS-T) plus 5 % milk powder for 1 h at room temperature. Subsequent antiserum incubations and washes (3 × 10 min between each step) were performed in PBS-T containing 1 % milk powder. The blot was incubated first with a 1:500 dilution of anti-GFP monoclonal antibody (Clontech), and then with a 1:2500 dilution of donkey anti-rabbit antibody conjugated to horseradish peroxidase (Amersham) for 1 h at room temperature. Antiserum binding to the blot was subsequently visualised on hyperfilm using the ECL system (Amersham).

Spectrofluorimetric determination of EGFP-Kir1.1 stability

In order to measure EGFP-Kir1.1b fluorescence, cells were plated in six multiwell plates, incubated overnight at 37°C to allow for attachment of cells, and then transferred to 26°C for 12 h to induce accumulation of EGFP-Kir1.1b. Synthesis of proteins was blocked with cycloheximide (CX) added to a final concentration of 20 μg cm⁻³ (Lukacs et al. 1993). After incubation at 26 or 37°C for times ranging from 0 to 10 h, cells were lysed in 300 μl of a buffer containing 50 mm Tris, 2 % Triton X-100, 150 mm NaCl, 5 mm EDTA and 0.1 mg ml⁻¹ phenylmethyl-sulfonyl fluoride, pH 7.5. Cells were incubated in lysis buffer on ice for 30 min and then triturated. Lysates were cleared by centrifugation at 14000 g for 10 min at 4°C. Protein was quantified using a detergent-compatible protein assay (Bio-Rad). For spectrofluorimetric assay, 200 μl of lysate were added into 2.8 ml of water and fluorescence was read (488_ex, 505_em) in an LS-5 Luminescence Spectrometer (Perkin Elmer). The relative level of fluorescence for each sample was calculated according to eqn (1), where Fs_(x), Fs₍₀₎, Fr_(x) and Fr₍₀₎ are fluorescence of the sample (s) or of MDCKR cells (r) at times x or 0, and Ps_(x), Ps₍₀₎, Pr_(x) and Pr₍₀₎ are milligrams of protein in the sample(s) or MDCKR cells (r) at times x or 0.

(1)

In order to assess the effectiveness of CX to block protein synthesis in MDCK cells, cells grown in six multiwell tissue culture plates were transiently transfected with pTRE or pTRE-Luc (Clontech) using SuperFect (Qiagen), and then incubated in the presence or absence of CX for between 0 and 10 h. Luciferase activity was measured using a Luciferase Assay System (Promega).

Confocal microscopy

Cells were grown to confluency on 23 mm transparent cell culture inserts (Falcon) under standard culture conditions. MDCKE cells were fixed in 4% paraformaldehyde, and MDCKER2 cells were fixed in 7 methanol:3 acetone (v/v) at −20°C. Cell nuclei were counterstained with 5 μg ml⁻¹ propidium iodide (PI) after treatment with 1 mg ml⁻¹ ribonuclease A in phosphate buffered saline at pH 7.0. Fluorescence was visualised using a TCS 4D confocal microscope (Leica) equipped with a Kr/Ar laser and two filter sets (488_ex, 520_em; 514/568_ex, 590_em).

RESULTS

Kir1.1b transfected cells express functional channels at 37°C

When analysed by RT-PCR, Kir1.1b transfected MDCK cells (MDCKR2) produced a product of the appropriate size that was absent from mock transfected MDCK cells (MDCKP: data not shown). In whole-cell clamp experiments performed at 37°C, Ba²⁺-sensitive currents in MDCKP cells (n = 5) were small (0.07 ± 0.25 nS; n = 5) and their reversal potential was approximately −10 mV. In contrast, MDCKR2 cells (n = 11) exhibited Ba²⁺-sensitive conductance of 3.11 ± 0.82 nS with a reversal potential (V_rev) of −69.54 ± 2.38 mV (Fig. 1A). The calculated permeability ratio P_K:P_Na was 57.15 ± 17.21.

Figure 1.

Mean (± s.e.m.) whole cell Ba²⁺-sensitive currents in transfected and non-transfected MDCK cell lines

I–V plots from MDCKP and MDCKR2 cells at 37°C (A) and at 26°C (B) and MDCKE and MDCKER2 cells at 37°C (C) and 26°C (D).

Incubation at 26°C increases functional channel expression

Whole-cell clamp experiments revealed that incubation at 26°C increased the magnitude of Ba²⁺-sensitive currents in MDCKR2 cells, but not in MDCKP cells. In MDCKP cells (n = 5), Ba²⁺-sensitive conductance was 0.02 ± 0.06 nS but in MDCKR2 cells it was 49.28 ± 10.69 nS (P < 0.05; n = 5) with a V_rev of −65.62 ± 5.61 mV (n = 9; Fig. 1B). The calculated P_K:P_Na was 61.55 ± 17.08, which was not significantly different from that at 37°C.

Electrical properties of MDCK monolayers expressing Kir1.1b

When MDCKP cells were grown on tissue culture inserts they produced monolayers with an R_tm of 1.6 ± 0.5 kΩ cm². In MDCKR2 monolayers, R_tm was 1.1 ± 0.4 kΩ cm². The mean V_tm of MDCKP monolayers was −0.8 ± 0.4 mV, a value not statistically different from zero. In marked contrast, V_tm of MDCKR2 was orientated lumen positive (2.2 ± 0.6 mV, n = 6 for all groups; P < 0.01) consistent with secretion of a cation. Accordingly, I_tm was also reversed for MDCKR2 monolayers (−0.25 ± 0.15 vs. 2.1 ± 0.16 μA cm⁻²; P < 0.05; n = 6 for both groups).

MDCK monolayers expressing Kir1.1b secrete K⁺

MDCKP monolayers grown on tissue culture inserts displayed no net movement of K⁺ (Fig. 2A). After 12 h incubation at 37°C, Δ[K⁺] = −0.01 ± 0.01 mm (not statistically different from zero). In contrast, monolayers expressing Kir1.1b (MDCKR2) secreted K⁺ (Fig. 2A). The Δ[K⁺] was 0.58 ± 0.03 mm (P < 0.01; n = 6). Addition of 1 mm Ba²⁺ to the apical solution inhibited K⁺ secretion across MDCKR2 monolayers (Fig. 2A: Δ[K⁺] = 0.30 ± 0.03 mm; P < 0.001; n = 6). This secretion was not inhibited by TEA (data not shown) a differential pharmacological sensitivity that is characteristic of Kir1.1b. Apical Ba²⁺ had no effect on MDCKP monolayers (Δ[K⁺] = −0.04 ± 0.01 mm). Addition of 1 mm Ba²⁺ to the basolateral solution (Fig. 2A) increased K⁺ secretion by MDCKR2 monolayers (Δ[K⁺] = 0.72 ± 0.05 mm; P < 0.05; n = 6), but was without effect on MDCKP monolayers (0.02 ± 0.02 mm; n = 6). In separate experiments performed with normal tissue culture medium (Fig. 2B) rather than the modified saline, the level of K⁺ secretion in MDCKR2 monolayers was effectively doubled (Δ[K⁺] = 1.71 ± 0.06; P < 0.01; n = 6). This was presumably due to the presence of various culture media factors present in serum that may modulate the transport properties of the monolayers. Interestingly, in the presence of complete media, MDCKP monolayers also displayed a very small, but nonetheless statistically significant level of net K⁺ secretion (Δ[K⁺] = 0.06 ± 0.01 mm; change significantly different to zero, P < 0.001). Under these conditions, K⁺ secretion by MDCKR2 monolayers was temperature dependent. Paradoxically however, MDCKR2 monolayers incubated at 26°C secreted K⁺ less effectively at the lower temperature (Δ[K⁺] = 0.29 ± 0.01 mm at 26°C) despite the higher K⁺ current expression demonstrated by the whole-cell clamp experiments. Incubation at 26°C had no effect on the measured Δ[K⁺] of MDCKP monolayers (Fig. 2B: Δ[K⁺] = 0.06 ± 0.01 mm at 37°C vs. 0.06 ± 0.01 mm at 26°C; n = 6).

Figure 2.

Potassium secretion by transfected monolayers

A, change in apical potassium concentration for MDCKR2, MDCKP, MDCKER2 and MDCKE in BSS (□) with apical (▪) or basolateral () BaCl₂. B, MDCKP (▪) and MDCKR2 (□) in normal cell culture medium at 26°C or 37°C (groups labelled with ^* are not significantly different from each other, ANOVA, P < 0.001). Error bars represent s.e.m. (n = 6 for all groups).

In order to characterise the mechanism for the decreased K⁺ secretion by MDCKR monolayers at 26°C, we investigated the possibility of a decrease in the chemical driving force for K⁺ exit at the lower temperature. At 37°C, intracellular K⁺ was found to be 3.03 ± 0.09 mmol K⁺ (g protein)⁻¹. Incubation at 26°C reduced intracellular K⁺ by almost 30 % (2.15 ± 0.06 mmol K⁺ (g protein)⁻¹; P ≤ 0.001, n = 6).

Tagging Kir1.1b with EGFP does not affect channel properties

The results obtained from monolayers of MDCK cells transfected with an EGFP-tagged Kir1.1b (MDCKER2) were qualitatively similar to those obtained from cells expressing wild-type Kir1.1b (MDCKR2 cells). At 37°C, whole-cell Ba²⁺-sensitive conductance was 11.34 ± 4.27 nS, V_rev was −74.4 ± 1.2 mV and the calculated P_K:P_Na was 72.1 ± 12.5 (n = 11). These values were not different from those obtained from MDCKR2 cells (expressing wild- type Kir1.1b) at 37°C (see above). As observed with wild- type Kir1.1b, whole-cell Ba²⁺-sensitive conductance was also increased by expression at 26°C (Fig. 1D) compared with cells expressing at 37°C (Fig. 1C). At 26°C, whole-cell Ba²⁺-sensitive conductance was 26.5 ± 10.61 nS (n = 14), V_rev was −73.4 ± 1.5 mV and the calculated P_K:P_Na was 77.0 ± 17.0 (n = 14). These values were also similar to those obtained from MDCKR2 cells at 26°C, in which V_rev was −65.62 ± 5.61 mV (n = 9; Fig. 1B) and the calculated P_K:P_Na was 61.55 ± 17.08 (n = 9). Overall, these results are consistent with the idea that N-terminal tagging of Kir1.1b with EGFP has no effect on channel function.

In contrast, MDCKE cells expressed small Ba²⁺-sensitive conductances, which were not temperature sensitive (0.31 ± 0.19 and 0.24 ± 0.16 nS at 37 and 26°C, n = 6).

MDCKER2 monolayers displayed qualitatively similar characteristics to MDCKR2 (V_tm = 7 ± 1.2 mV, R_tm = 7.8 ± 3.0 kΩ cm², I_tm = 0.87 ± 0.18 μA cm⁻², n = 6) while MDCKE monolayers did not develop significant V_tm (−0.2 ± 0.06 mV) or I_tm (−0.05 ± 0.02 μA cm⁻²), despite high transmonolayer resistance (R_tm = 3.2 ± 1.6 kΩ cm²; n = 6 for all). MDCKER2 monolayers also secreted potassium. Mean Δ[K⁺] was 0.82 ± 0.01 mm. This secretion was inhibited by apical Ba²⁺ (Δ[K⁺] = 0.34 ± 0.01 mm; P < 0.001; n = 6), but was stimulated by basolateral Ba²⁺ (Δ[K⁺] = 1.43 ± 0.07 mm; P < 0.001; n = 6). Neither apical nor basolateral Ba²⁺ had any effect on MDCKE monolayers (Fig. 2A).

Incubation at 26°C stabilises Kir1.1b

Both cells expressing untagged and tagged Kir1.1b expressed higher K⁺ currents when incubated at 26 than at 37°C. We investigated whether this might be due to alterations in protein stability by spectrofluorimetric analysis of cell lysates. At 37°C, the levels of EGFP-Kir1.1b expression were not distinguishable from endogenous (auto) fluorescence of Kir1.1b-expressing cells; (279 ± 7 (MDCKER) vs. 261 ± 13 (MDCKR) fluorescence units (f.u.) (mg protein)⁻¹; n = 6). However, in cells expressing EGFP-Kir1.1b at 26°C, fluorescence was increased over threefold (923 ± 40 f.u. (mg protein)⁻¹; P < 0.001; n = 6). Altering temperature had no significant effect on the fluorescence of cells either expressing untagged Kir1.1b (MDCKR) or mock transfected cells (data not shown).

To verify the ability of CX to block protein synthesis in MDCK cells, we tested its ability to inhibit synthesis of luciferase (Fig. 3A). MDCK cells transiently transfected with pTRE (containing no Luc gene) as expected, did not express significant luciferase activity (136 ± 6 luciferase units (l.u.); n = 6). Cells transfected with pTRE-LUC expressed a high level of luciferase activity (23770 ± 1379 l.u.; n = 6) that was completely abolished when cells were incubated in the presence of CX (47 ± 19 l.u.; P < 0.05; n = 6). The effect of CX was reversible, since when cells were washed and incubated for a further 12 h in CX-free medium, luciferase activity was restored (65755 ± 5204 l.u., n = 6). This shows also that CX did not inhibit cell transfection.

Figure 3.

Determination of protein relative stability using cycloheximide

A, luciferase activity, expressed as luciferase units (l.u.), of (a) MDCK cells transfected with pTRE (a plasmid that does not contain the luciferase gene) after 12 h incubation in normal medium; (b) cells transfected with pTRE-Luc (a vector containing the luciferase gene) after 12 h incubation in normal medium; (c) cells transfected with pTRE-Luc after 12 h incubation in medium containing CX, and (d) cells transfected with pTRE-Luc after 12 h incubation in medium containing CX and then a further 12 h in medium free of CX (groups labelled with ^* are not significantly different from each other, ANOVA, P < 0.001). B, changes in relative fluorescence of MDCKER2 and MDCKE cells with time. MDCKER2 cells were incubated at 26 and 37°C in the presence or absence of CX. MDCKE cell lines were incubated at 37°C with CX. All cells were pre-incubated at 26°C for 12 h before the experiment started in order to allow for accumulation of EGFP-ROMK2. Error bars represent s.e.m. (n = 5).

To evaluate dynamic changes in levels of EGFP-Kir1.1, cells were incubated for a period of 10 h at 26 or 37°C, either in the presence or absence of CX (Fig. 3B). Fluorescence declined more rapidly in cells incubated at 37°C than in cells incubated at 26°C. Within the cells grown at 37°C, those without CX reached equilibrium, while in those with CX the decline in fluorescence continued. Conversely, in cells grown at 26°C without CX, no significant change in fluorescence occurred, while in those with CX, fluorescence declined. These results clearly show that the increased level of expression of EGFP-Kir1.1b in MDCKER2 cells incubated at 26°C was due to an increase in protein stability. For cells expressing EGFP alone (MDCKE) incubated at 37°C in the presence of CX, there was again no significant change in fluorescence. The increasing trend in fluorescence, being due to a decrease of total protein rather than to an increase of EGFP, demonstrated the extreme stability of this non-endogenous protein in mammalian cells. To determine whether all the fluorescence in MDCKER2 cell was due to EGFP-Kir1.1b protein and did not result from synthesis of EGFP, RT-PCR and Western blot analysis were performed. RT-PCR with total RNA from MDCKER2 cells produced only a product of appropriate size (data not shown). Moreover, Western blotting using a monoclonal antibody to EGFP revealed expression of a protein of the calculated size for EGFP-Kir1.1b (∼66 kDa) in MDCKER2 cells. MDCKE cells produced a band of the appropriate size for EGFP (∼26 kDa) that was absent in MDCKER2 cells (Fig. 4), showing that only the fusion protein and not EGPF alone contributed to the fluorescence of MDCKER2 cells.

Figure 4.

Protein expression in transfected cells, using a monoclonal antibody raised against EGFP

A single, major band at 26 kDa was detected in MDCKE cells, and at 66 kDa in MDCKER2 cells. No significant bands were detected in untransfected (MDCK) cells.

Apical location of EGFP-Kir1.1b

Confocal microscopy of MDCKER2 cells grown on permeable tissue culture inserts at 26°C showed that fluorescence was localized to the apical pole and was completely excluded from the cell nuclei (Fig. 5A). In MDCKER2 cells grown at 37°C, fluorescence was not consistently detectable, despite the fact that functional channels were present in the plasma membrane (Fig. 1C) and these monolayers secreted K⁺. In contrast, in MDCKE cells, fluorescence was distributed throughout the cytoplasm and nuclei (Fig. 5B).

Figure 5.

Confocal fluorescent micrographs of transfected MDCK cells grown on permeable supports

Localization of EGFP-Kir1.1b in MDCKER2 cells (A) and EGFP in MDCKE cells (B). Panels 1 to 7 are sections in the x-y plane, at 1.0 μm increments and panel 8 is in the x-z plane. For MDCKER2 cells, nuclei were counterstained with propidium iodide. EGFP-Kir1.1b is localized to the apical pole while EGFP was distributed throughout the cell. Panels 1–7 measure 43 × 43 μm.

DISCUSSION

Strain I MDCK cells are able to form a tight epithelium resembling that of the cortical collecting duct (Simmons, 1981). In this cell line, Na⁺ diffuses across the apical membrane and is actively extruded across the basolateral membrane via the Na⁺-K⁺-ATPase. The apical membrane is effectively impermeable to K⁺ (Aiton et al. 1982), making MDCK cells a good candidate for expression of an apical K⁺ channel. The electrochemical gradient across the apical membrane favours diffusion of K⁺ out of the cell. However, this does not occur because in MDCK cells only the basolateral membrane possesses significant permeability to K⁺ (Aiton et al. 1982). Consistent with the low level of expression of K⁺ channels in the plasma membrane, MDCK cells did not exhibit net transepithelial K⁺ movement, or significant whole-cell Ba²⁺-sensitive currents. However, upon transfection with Kir1.1b cDNA, MDCK cells expressed large whole-cell Ba²⁺-sensitive conductance consistent with expression of functional Kir1.1b channels. Under these conditions, Kir1.1b was clearly the dominant membrane conductance, since the reversal potential shifted in these cells to a value close to the theoretical reversal potential for K⁺ under the ionic conditions of our experiments (−84 mV). As expected, cells expressing Kir1.1b were able to secrete K⁺ and this secretion was inhibited by the K⁺ channel blocker Ba²⁺, a result consistent with expression of Kir1.1b at the apical membrane. Basolateral Ba²⁺ activated K⁺ secretion, probably by inhibition of K⁺ exit at the basolateral membrane, thereby increasing the driving force for K⁺ exit across the apical membrane.

In order to confirm that Kir1.1b was appropriately targeted in MDCK cells, we tagged Kir1.1b at its N-terminus with EGFP. Tagging Kir1.1b at its N-terminal appeared to have no significant effect on macroscopic channel function at least as judged by the whole-cell K⁺ conductance properties, which were similar to cells expressing untagged Kir1.1b. In agreement with our observations, a recent study (Flagg et al. 1999) reported that the properties of Kir1.1a (ROMK1)-mediated macroscopic currents were unaffected by GFP tagging when expressed in Xenopus oocytes. We of course, cannot discount the possibility that the EGFP tag might exert subtle effects on single-channel properties that cannot be resolved by measurement of macroscopic currents; however, a definite answer awaits more extensive electrophysiological analysis in these new cell lines. Although the observed potassium secretory activity was higher in monolayers expressing EGFP-tagged Kir1.1b than in cells expressing wild-type Kir1.1b, this most likely reflects differences in the level of expression of the proteins between clonal lines (B. Ortega & S. J. White, unpublished observations).

It has been reported that expression at reduced temperatures may stabilise proteins (Ljunggren et al. 1990). Moreover, the most common mutation (ΔF508) in cystic fibrosis produces a defect in trafficking of CFTR, which can be alleviated by reduction in expression temperature (Denning et al. 1992). For these reasons, we were interested in the effect that reduced temperature might have on expression of Kir1.1b. In MDCKR cells, Ba²⁺-sensitive conductance was increased approximately 15-fold upon incubation at 26°C. However, despite this increase in Kir1.1b-mediated conductance; K⁺ secretion was reduced. This paradoxical finding can be explained by several possible mechanisms. Firstly, a reduction in the driving force for K⁺ secretion, due to the lowered intracellular K⁺ concentration in MDCKR2 cells incubated at 26°C. This most likely occurs via inhibition of the activity of the Na⁺-K⁺-ATPase at the reduced temperature, since the activity of the enzyme is markedly temperature dependent (Brodie & Sampson, 1985). Secondly, as has been demonstrated previously in the collecting duct, it is possible that inhibition of the Na⁺-K⁺-ATPase resulted in protein kinase C-mediated inhibition of the secretory K⁺ channel activity (Wang et al. 1993). Thirdly, it is likely that other enzymatic components required for normal function of Kir1.1b in the intact cell were also compromised at the lower expression temperature.

Whole-cell clamp experiments demonstrated larger Kir1.1b-mediated currents in the plasma membrane of MDCKER2 cells grown at 26°C than those grown at 37°C. It is clear that expression of Kir1.1b was increased at 26°C, because EGFP-Kir1.1b fluorescence was much higher in cells grown at the lower temperature. We therefore utilised the fluorescent fusion protein both to verify localization of the protein and also to compare Kir1.1b stability at the two temperatures. At 37°C, EGFP-Kir1.1b had a half-life of between 3 and 6 h and this was increased greatly at 26°C. Even after 10 h in the presence of cycloheximide, levels of the fusion protein had decreased only by approximately 30 %, in comparison with an 80 % reduction in cells at 37°C. A previous study has shown that a C-terminal Bartter's mutation results in the production of Kir1.1 protein that has altered susceptibility to proteolytic degradation (Schwalbe et al. 1998). In experiments not reported here (B. Ortega, L. Robson & S. J. White, unpublished observations), we have found that a non-Bartter's mutation in the C-terminal domain of Kir1.1b, also increases greatly the stability of the protein at 37°C. The notion that alterations in channel stability may play a role in regulating the amount of Kir1.1b present at the plasma membrane as a mechanism of controlling K⁺ secretion clearly requires further investigation.

Our suggestion that the stability of Kir1.1b is increased at 26°C is supported by a recent study (Brejon et al. 1999) in which a pulse chase labelling method was used to estimate the stability of a 6 amino acid epitope- (AU1)-Kir1.1a fusion protein when expressed in MDCK cells. These authors were unable to detect any Kir1.1a-mediated currents in cells grown at 37°C. The authors argue that this could be due to the lack of putative associated surrogate proteins necessary for appropriate trafficking. However, this is clearly not the case in MDCK cells, since as shown in the present study, Kir1.1b is appropriately targeted and is functional at 37°C. A more likely explanation for the inability to observe functional Kir1.1 reported by Brejon et al. is that currents were measured in the absence of ATP, a condition known to enhance greatly Kir1.1b run-down (McNicholas et al. 1994).

In summary, we have shown that stable expression of the distal nephron K⁺ channel Kir1.1b in MDCK cells produces potassium-secreting cell lines in which the protein is functional and is targeted appropriately at physiological temperatures. These cell lines should be useful tools with which to study mechanisms determining trafficking of Kir1.1 as well as possible interactions between Kir1.1 and other proteins such as CFTR (Ho, 1998). In addition, it is likely that the monolayers will prove useful models for primary screening of pharmacological agents that might be effective in alleviating some disorders of Kir1.1 trafficking that may be the cause of some forms of Bartter's syndrome.