Protein kinase C phosphorylation disengages human and mouse-1a P-glycoproteins from influencing the rate of activation of swelling-activated chloride currents

Abstract

1. Whole-cell, swelling-activated Cl⁻ currents, I_Cl(swell), were characterized in Chinese hamster ovary (CHO) cells and found to exhibit time-dependent inactivation at depolarizing potentials, tamoxifen and dideoxyforskolin sensitivity, and an anion permeability sequence: SCN⁻ > I⁻ > Br⁻ > Cl⁻ > F⁻ > gluconate⁻.

2. CHO cells permanently transfected with either the human MDR1 or mouse mdr1a cDNAs demonstrated an increased rate of activation of I_Cl(swell) compared with parental cells or those permanently transfected with the mouse mdr1b cDNA. However, no differences in the magnitude of the currents were observed at steady state.

3. Pretreatment with the phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA) did not affect I_Cl(swell) in MDR1 or mdr1a permanently transfected CHO cells. In contrast, pretreatment with TPA reduced I_Cl(swell) in MDR1(G185V)-expressing transfected NIH3T3 fibroblasts. Subsequently, the CHO cell lines were shown to contain significantly reduced levels of protein kinase C (PKC), suggesting that PKC concentrations might be limiting in these cell lines, at least under whole-cell patch clamp conditions.

4. Addition of purified PKC to the pipette solution, followed by a pretreatment with TPA, reduced the rate of I_Cl(swell) activation in human Pgp- and mouse Pgp1a-expressing CHO cells to the levels observed in parental and mouse Pgp1b-expressing cells. This confirms that PKC is limiting in these cells under whole-cell, patch clamp conditions. Furthermore, these results suggest that PKC-mediated phosphorylation of human Pgp and mouse Pgp1a disengages the influence which these Pgps have on I_Cl(swell).

5. These studies also demonstrate a functional distinction between the two mouse homologues, Pgp1a and Pgp1b. Although both can function as drug efflux pumps, only Pgp1a can act like human Pgp to influence I_Cl(swell).

The exit of Cl⁻ and K⁺ ions from the cell, followed by the consequent loss of water, is the most widespread mechanism for restoring cell volume in response to hyposmotic solution (Hoffmann, Simonsen & Lambert, 1993). Although, the molecular identity of the channels involved in regulatory volume decrease (RVD) is unclear, it is known that the well-characterized swelling-activated anion conductance, I_Cl(swell), plays a role in many cell types (Valverde et al. 1996).

The multidrug resistance P-glycoprotein (Pgp) has been shown to modulate I_Cl(swell) in several cell types (Valverde, Diaz, Sepulveda, Gill, Hyde & Higgins, 1992; Hardy, Goodfellow, Valverde, Gill, Sepulveda & Higgins, 1995; Bond, Higgins & Valverde, 1997; Miwa, Ueda & Okada, 1997) and, as a consequence, alters the ability of these cells to undergo RVD (Valverde et al. 1996). In humans, Pgp is encoded by the MDR1 gene. In rodents, the gene encoding Pgp is duplicated and, like their human counterpart, both the mdr1a and mdr1b gene products (Pgp1a and Pgp1b, respectively) are ATP-dependent drug transporters with only minor differences in their substrate and inhibitor specificities (Yang, Cohen, Greenberger, Hsu & Horowitz, 1990; Tang-Wai, Kajiji, DiCapua, de Graaf, Roninson & Gros, 1995). However, despite the high degree of sequence identity between these two mouse homologues, significant differences in their abilities to influence I_Cl(swell) have been reported (Valverde et al. 1996). The mouse Pgp1a, but not mouse Pgp1b, increases the rate of activation of I_Cl(swell) and its sensitivity to hyposmotic shock without affecting the maximum current which can be elicited (Valverde et al. 1996; Bond et al. 1997). The fact that Pgp1b does not influence I_Cl(swell) provides strong evidence that the effects of Pgp1a on I_Cl(swell) are not simply a result of drug selection or protein overexpression which could lead to membrane crowding. Perhaps significantly, the region of greatest sequence divergence between mouse Pgp1a and Pgp1b is in the ‘linker’ region, which is located between the two halves of the molecule and includes sites for protein kinase C (PKC)-mediated phosphorylation (Devault & Gros, 1990).

Phosphorylation of human Pgp by PKC can modify its ability to activate I_Cl(swell), at least in some cell types. Activation of PKC reduces the currents elicited 2-3 min after exposure to hyposmotic solution in human Pgp-expressing cells, but not in cells expressing a mutant form of human Pgp which cannot be phosphorylated (Hardy et al. 1995). However, in studies using other Pgp-expressing cell types, PKC activation was found not to exert an effect on I_Cl(swell) at steady state (Tominaga, Tominaga, Miwa & Okada, 1995).

To resolve these apparent discrepancies, and to understand the role of PKC phosphorylation of Pgp in the modulation of I_Cl(swell), we have taken advantage of a series of matched CHO (Chinese hamster ovary) cell lines transfected with the human MDR1 (LR73-MDR1), or murine mdr1a (LR73-mdr1a) or mdr1b (LR73-mdr1b) genes. The data obtained show that human Pgp and mouse Pgp1a, but not mouse Pgp1b, influence the rate of activation of I_Cl(swell). Further-more, phosphorylation of human Pgp and mouse Pgp1a by PKC appears to disengage this effect, such that activation of I_Cl(swell) in cells expressing the phosphorylated protein is indistinguishable from that measured in cells where human Pgp or mouse Pgp1a are not expressed. These data clarify the circumstances under which PKC-mediated phosphorylation of Pgp modulates channel activity, and highlight a functional distinction between the two mouse homologues, Pgp1a and Pgp1b.

METHODS

Cells and cell growth

CHO cells (LR73), and derivatives permanently transfected with the mouse mdr1a gene (LR73-mdr1a), or the mouse mdr1b gene (LR73-mdr1b) were provided by Dr Phillipe Gros (McGill University, Montreal, Canada) (Devault & Gros, 1990). LR73 cells permanently transfected with the human MDR1 gene (LR73-MDR1) were provided by Dr Igor Roninson (University of Illinois, Chicago, IL, USA) (Tang-Wai et al. 1995). Immunological analysis using an anti-Pgp antibody showed that each of the transfected cell lines expresses large and equivalent amounts of the respective Pgps (Tang-Wai et al. 1995; Valverde et al. 1996). NIH3T3-MDR1(G185V) cells are NIH3T3 fibroblasts permanently transfected with human MDR1 cDNA containing the G185V mutation which increases the relative resistance to colchicine (Pastan, Gottesman, Ueda, Lovelace, Rutherford & Willingham, 1988). Cells were maintained as described previously (Valverde et al. 1992, 1996).

Electrophysiology

Swelling-activated Cl⁻ currents were measured at room temperature (20-23°C) using the whole-cell recording mode of the patch-clamp technique as described previously (Valverde et al. 1992, 1996). The electrode resistance was 3-4 MΩ and series resistance was checked throughout the experiments. Intracellular (pipette) solution contained (mM): 140 NMDGCl, 1.2 MgCl₂, 1.0 EGTA, 2 ATP, 10 Hepes (pH 7.4); 280 mosmol kg⁻¹. The isosmotic bathing solution contained (mM): 140 NMDGCl, 0.5 MgCl₂, 1.3 CaCl₂, 10 Hepes (pH 7.4), 20 D-mannitol; 300 mosmol kg⁻¹. The standard 20% hyposmotic bathing solution contained (mM): 105 NMDGCl, 0.5 MgCl₂, 1.3 CaCl₂, 10 Hepes (pH 7.4); 220 mosmol kg⁻¹. For anion substitution studies, 105 mM NaCl, NaBr, NaI, NaSCN, NaF, or sodium gluconate replaced NMDGCl in the standard hyposmotic bathing solution. Junction potentials which developed while testing the anion permeability were measured as described previously (Neher, 1992) for each experimental solution and used to correct for the reversal potential of membrane currents. The rate of current increase was calculated as the slope of the increase in current in the first 2-3 min after hyposmotic challenge. Results are shown as means ± standard error of the mean (s.e.m.). Statistical significance was determined using Student's t test. To study the effects of PKC upon swelling-activated Cl⁻ currents, 175 ng ml⁻¹ (specific activity, 0.24 units ml⁻¹) of purified rat brain PKC (Calbiochem) was added to the pipette solution.

Western blotting

Proteins were extracted from cell suspensions sonicated in ice-cold lysis buffer (10 mM KCl, 1.5 mM MgCl₂, 10 mM Tris-HCl (pH 7.4)) containing the protease inhibitors aprotinin (5 μl ml⁻¹), leupeptin (20 μg ml⁻¹), and phenyl methysulphonyl fluoride (PMSF) (348 μg ml⁻¹). Samples were centrifuged to remove cell debris and incubated at 70°C for 3 min in Laemmli sample buffer (10% glycerol, 2% SDS, 5%β-mercaptoethanol, 0.0025% Bromophenol Blue, 60 mM Tris-HCl (pH 6.8)). Proteins were quantified using a modified micro Lowry assay (Peterson, 1977), separated by electrophoresis on a 10% SDS-polyacrylamide gel, and transferred to a nitrocellulose membrane (Hybond-ECL) in transfer buffer (25 mM Tris-HCl (pH 8.3), 193 mM glycine, 20% methanol, 0.04% SDS). Membranes were probed with the anti-PKC-α monoclonal antibody (Transduction Laboratories, Lexington, KY, USA) (1:1000) and immunodetection was by enhanced chemiluminescence (Amersham International).

RESULTS

Characterization of I_Cl(swell) in CHO cells

Swelling-activated Cl⁻ currents were characterized by whole-cell, patch clamp procedures in LR73 cells which did not express Pgp. Following exposure to a 20% hyposmotic solution an outwardly rectifying current appeared which showed time-dependent inactivation kinetics at depolarizing voltages. This conductance had an anion permeability sequence (P_X/P_Cl): SCN⁻ (1.53 ± 0.04) > I⁻ (1.33 ± 0.01) > Br⁻ (1.20 ± 0.01) > Cl⁻ (1.0) > F⁻ (0.56 ± 0.01) > gluconate⁻ (0.22 ± 0.02), calculated as in Diaz, Valverde, Higgins, Rucareanu & Sepulveda (1993) (Fig. 1). These currents were blocked by tamoxifen and 1,9-dideoxyforskolin with IC₅₀ (±s.e.m.) values of 1.54 ± 0.38 μM and 38 ± 15 μM, respectively. Together, the above data show that the LR73 cell line contains a swelling-activated anion current which has properties similar to I_Cl(swell) studied in many other cell types (Sheppard & Valverde, 1997).

Figure 1. Effect of anion replacements upon the swelling-activated currents in LR73 cells.

The I-V relationships shown were obtained after perfusing the cell under study in a 105 mM NaX hyposmotic solution. During the experiment different solutions in which the main anion (Cl⁻) was replaced with another anion (X⁻) were perfused across the cell. A, mean reversal potential (V_rev) ±s.e.m. of X⁻ (n ≥ 5): I⁻, -10.5 ± 0.64 mV; F⁻, 13.3 ± 0.48 mV; and gluconate⁻, 34.8 ± 1.86 mV. B, mean V_rev±s.e.m. of X⁻ (n = 6): Br⁻, -4.65 ± 0.28 mV; and SCN⁻, -10.5 ± 0.64 mV. The membrane potential was held at 0 mV and currents were generated by ramped voltage steps from -80 to +80 mV over a 1 s time interval.

Effect of 12-O-tetradecanoylphorbol-13-acetate (TPA) stimulation of PKC on I_Cl(swell)

As shown previously (Valverde et al. 1992, 1996), cells expressing human Pgp or mouse Pgp1a demonstrated larger I_Cl(swell) currents than non-Pgp-expressing cells when measured 2-3 min after hyposmotic shock (Fig. 2, top panel). The phorbol ester TPA, which activates PKC, has previously been shown to reduce I_Cl(swell) in NIH3T3 and HeLa cells expressing human Pgp (Hardy et al. 1995). In contrast, we found that TPA had no detectable effect on I_Cl(swell) in LR73 cells expressing either mouse Pgp1a or Pgp1b (P > 0.05; Fig. 2A). In simultaneous experiments, TPA reduced I_Cl(swell) in NIH3T3-MDR1(G185V) cells (P = 0.007; Fig. 2A), as reported previously. It should be noted that the G185V mutation in MDR1 does not influence its ability to modulate channel activation. The different effects of TPA on these two cell lines suggested either that human and mouse Pgps differ in their ability to influence I_Cl(swell), or that other factors which are necessary for the transduction of the TPA signal may differ between the two cell types.

Figure 2. Levels of PKC-α expression and the modulation of I_Cl(swell) by Pgp after pre-treatment with phorbol esters.

Top panel, mean whole-cell currents generated at +80 mV 3 min after exposure to a 20% hyposmotic shock with (▪) or without (□) pretreatment with 100 nM TPA for 2 min. Cells were clamped at 0 mV and currents generated by pulsing to +80 mV for 500 ms. Values represent the means ±s.e.m. for LR73 (n = 4), LR73-mdr1a (n = 6), LR73-mdr1b (n = 6), NIH3T3 (n = 4) and NIH3T3-MDR1(G185V) (n = 5). Bottom panel, crude cell lysates from LR73, LR73-mdr1a and LR73-mdr1b, NIH3T3 and NIH3T3-MDR1(G185V) cells were separated on a 10% SDS-polyacrylamide gel, transferred to a nitrocellulose membrane, and probed with a monoclonal antibody against PKC-α. Each lane contains 20 μg of total protein. Molecular weight markers were β-galactosidase (116 kDa) and bovine serum albumin (66 kDa).

Since TPA stimulates PKC to phosphorylate Pgp, differences in the levels of PKC between the LR73 and NIH3T3 cell lines were investigated. Figure 2 shows, in the bottom panel, a Western blot of crude cell extracts from the parental NIH3T3 and LR73 cells and their derivatives expressing the various Pgp homologues, probed with an anti-PKC-α antibody. PKC-α is the isoform associated with phosphorylation of Pgp (Ahmad, Safa & Glazer, 1994). LR73 and its derivative cell lines contained substantially reduced levels of PKC-α compared with the NIH3T3 and NIH3T3-MDR1(G185V) cell lines.

Addition of purified PKC restores TPA sensitivity to I_Cl(swell)

The above data suggest that the low levels of PKC-α in the LR73 cell lines may limit any effect of TPA on I_Cl(swell). To test this hypothesis, purified rat brain PKC was added to the patch pipette (175 ng ml⁻¹; specific activity, 0.24 units ml⁻¹). When PKC was included in the patch pipette, pretreatment with TPA (100 nM) for 2 min significantly reduced the rate of activation of I_Cl(swell) in cells expressing mouse Pgp1a (P = 0.002 for LR73-mdr1a vs. LR73-mdr1a + TPA and PKC; Fig. 3). In fact, the rate of activation was reduced to the basal rate seen in mouse Pgp1b- and non-Pgp-expressing cells (Fig. 3B). Inclusion of PKC in the patch pipette had no effect on I_Cl(swell) in cells expressing mouse Pgp1b or parental LR73 cells pretreated with TPA (P > 0.05; Fig. 3B), underlining the inability of mouse Pgp1b to influence I_Cl(swell). Although the addition of PKC and TPA to cells expressing mouse Pgp1a reduced the rate of activation of I_Cl(swell), the magnitude of currents achieved at steady state (10 min) were not significantly different from those observed in cells which were not exposed to PKC and TPA (P > 0.7; Fig. 3A, hyposmotic steady state).

Figure 3. The effect of purified PKC on I_Cl(swell) in CHO cells expressing mouse Pgp1a or Pgp1b.

A, current traces from LR73-mdr1a cells (upper panels) and LR73-mdr1a cells with purified rat brain PKC added to the pipette solution and pretreated with 100 nM TPA for 2 min (lower panels). These cells were held at 0 mV and pulsed to +80 mV for 200 ms while (from left to right) bathed in isosmotic solution, 2 min after switching to a 20% hyposmotic bathing solution, 5 min after switching to a 20% hyposmotic bathing solution and finally, once the currents had reached a steady-state level. B, initial rates of peak current increase measured at +80 mV for LR73, LR73-mdr1a and LR73-mdr1b cells under control conditions (□) or after a 2 min pretreatment with 100 nM TPA and PKC added to the pipette (▪). LR73 (n = 10), LR73 + TPA and PKC (n = 6), LR73-mdr1a (n = 10), LR73-mdr1a + TPA and PKC (n = 8), LR73-mdr1b (n = 7), LR73-mdr1b + TPA and PKC (n = 5).

Human Pgp influences the rate of activation of I_Cl(swell)

As reported previously, PKC-mediated phosphorylation of human Pgp can downregulate I_Cl(swell) (Hardy et al. 1995). In order to investigate if this downregulation is also due to a decrease in the rate of activation of I_Cl(swell), the effects of TPA-induced PKC stimulation were studied in LR73-MDR1 cells. Human Pgp upregulated the rate of activation of I_Cl(swell) in LR73-MDR1 cells (LR73 vs. LR73-MDR1, P = 0.0006; Fig. 4), similar to the effects of mouse Pgp1a. Pretreatment with TPA, or the inclusion of purified PKC alone in the patch pipette, did not significantly affect the rate of activation (P > 0.7; Fig. 4 inset). However, pretreatment with TPA prior to hyposmotic stimulation reduced the rate of activation of I_Cl(swell) when purified PKC was added to the patch pipette (P = 0.008 for LR73-MDR1 vs. LR73-MDR1 + TPA and PKC; Fig. 4, inset). Thus, human Pgp and mouse Pgp1a have similar effects on the rate of channel activation and, in both cases, these effects can be disengaged by PKC-mediated phosphorylation.

Figure 4. The effect of purified PKC on I_Cl(swell) in human Pgp-expressing CHO cells.

Increases in Cl⁻ currents obtained after exposure to a 20% hyposmotic solution added at 0 min. Peak currents measured at +80 mV were recorded each minute from cells voltage clamped at 0 mV and pulsed to -80 mV (200 ms) and +80 mV (200 ms) each second using an alternating pulse protocol. Values represent mean whole-cell currents ±s.e.m. obtained after each minute from parental (○; n = 10), human Pgp-expressing cells (♦; n = 5), and human Pgp-expressing cells with purified rat brain PKC (specific activity, 0.24 units ml⁻¹) added to the patch pipette and pretreated with TPA beginning at -2 min (▪; n = 5). Inset: initial rates of peak current activation measured at +80 mV in LR73-MDR1 cells (n = 5) are compared with cells pretreated with 100 nM TPA (n = 5), cells with purified rat brain PKC added to the pipette (n = 6), and cells which had PKC added to the pipette and were pretreated with TPA (n = 5).

DISCUSSION

The multidrug resistance Pgp can modulate the rate of activation of cell swelling-activated Cl⁻ channels, although the mechanism remains unclear (Valverde et al. 1992, 1996; Gill, Hyde, Valverde, Diaz, Sepulveda & Higgins, 1992; Bond et al. 1997; Miwa et al. 1997). We have previously shown that these differences are not due to differences in swelling between the cell types (Valverde et al. 1996). Phosphorylation of the linker region of human Pgp is known to affect its ability to influence I_Cl(swell), at least in some cell types (Hardy et al. 1995). However, in other studies no effect of PKC activation on I_Cl(swell) was observed (Tominaga et al. 1996; Miwa et al. 1997). In order to understand the regulatory relationship between Pgp phosphorylation and I_Cl(swell) better, we studied a series of matched cell lines transfected with human Pgp or the two mouse homologues Pgp1a and Pgp1b.

In this study, expression of mouse Pgp1a was shown to increase the rate of activation of I_Cl(swell), as reported previously (Valverde et al. 1996; Bond et al. 1997). Similar to mouse Pgp1a, human Pgp was also found to increase the rate of channel activation. In contrast, mouse Pgp1b had no detectable effect on channel activation. In this context, mouse Pgp1b provides an ideal internal control, showing that the effects of human Pgp and mouse Pgp1a on the rate of channel activation are not simply a consequence of increased levels of protein in the membrane, or of the drug present during selective cell growth of MDR cells. Furthermore, this provides the first clear functional distinction between the mouse homologues, Pgp1a and Pgp1b. Although both Pgp1a and Pgp1b act as ATP-dependent drug transporters, the finding that Pgp1a and Pgp1b have distinct spatial and temporal patterns of expression (Croop et al. 1989; P. Romano, A. E. O. Trezise & C. F. Higgins, unpublished observations) implies that they have distinct functional roles. Differences in their abilities to influence I_Cl(swell) may provide an explanation for the gene duplication in rodents and for the divergence of the two homologues after the duplication event.

I_Cl(swell) in the Pgp-expressing LR73 cell lines was unaffected by pretreatment with TPA, in contrast to that seen in NIH3T3-MDR1(G185V) cells where pretreatment with TPA reduced I_Cl(swell) (this paper; Hardy et al. 1995). This difference correlated with lower levels of PKC-α expression in the LR73 cell lines. Although PKC-α could still be detected in the LR73 cell lines, dialysis of the intracellular compartment when achieving whole-cell configuration may dilute the PKC concentration below that necessary for the phosphorylation of Pgp. Consistent with this, when purified PKC was added to the pipette solution, the activation of PKC by phorbol esters reduced the rate of channel activation in both human Pgp- and mouse Pgp1a-expressing cells but did not affect the maximum current that could be elicited. Importantly, the rate of channel activation was reduced to a rate equivalent to that seen in non-Pgp-expressing cells. Thus, phosphorylation of Pgp prevents Pgp from influencing I_Cl(swell). As the phosphorylation state of Pgp modulates its ability to influence the channel, and as the phosphorylation state is likely to differ in different cell types, it is perhaps not surprising that Pgp affects channel activation in some cell types but appears to have little effect in others. In fact, in an independent study using human epidermoid KB cells, although PKC activation did not influence I_Cl(swell), the addition of the PKC inhibitor H-7 increased the rate of activation of I_Cl(swell) above that seen in control cells (Miwa et al. 1997). This observation suggests that, in this cell type, the basal level of PKC activity is sufficient to phosphorylate Pgp and, thus, further activation of PKC has no additional effect on I_Cl(swell).

The linker region of Pgp is phosphorylated by PKC in vivo and in vitro (Chambers, Pohl, Raynor & Kuo, 1993; Orr et al. 1993), yet phosphorylation of Pgp has no detectable effect on its drug transport properties (Germann et al. 1996; Goodfellow et al. 1996) and only minor effects on its ATPase activity (Szabóet al. 1997). The finding that phosphorylation of the linker region appears to disengage Pgp from the channel now suggests a function for PKC phosphorylation of Pgp. With this in mind, it is interesting to note that the major sequence divergence between the mouse Pgp1a and Pgp1b is in the phosphorylatable ‘linker’ region. Elucidation of the molecular basis of the mechanism by which Pgp influences the rate of channel activation, and by which this is ‘uncoupled’ by PKC must await identification of the channel protein itself.