Several proteins belonging to the ATP-binding cassette superfamily can affect ion channel function. These include the cystic fibrosis transmembrane conductance regulator, the sulfonylurea receptor, and the multidrug resistance protein P-glycoprotein (MDR1). We measured whole cell swelling-activated Cl− currents (I Cl,swell) in parental cells and cells expressing wild-type MDR1 or a phosphorylation-defective mutant (Ser-661, Ser-667, and Ser-671 replaced by Ala). Stimulation of protein kinase C (PKC) with a phorbol ester reduced the rate of increase inI Cl,swell only in cells that express MDR1. PKC stimulation had no effect on steady-stateI Cl,swell. Stimulation of protein kinase A (PKA) with 8-bromoadenosine 3′,5′-cyclic monophosphate reduced steady-stateI Cl,swell only in MDR1-expressing cells. PKA stimulation had no effect on the rate ofI Cl,swellactivation. The effects of stimulation of PKA and PKC onI Cl,swell were additive (i.e., decrease in the rate of activation and reduction in steady-stateI Cl,swell). The effects of PKA and PKC stimulation were absent in cells expressing the phosphorylation-defective mutant. In summary, it is likely that phosphorylation of MDR1 by PKA and by PKC alters swelling-activated Cl− channels by independent mechanisms and that Ser-661, Ser-667, and Ser-671 are involved in the responses ofI Cl,swell to stimulation of PKA and PKC. These results support the notion that MDR1 phosphorylation affectsI Cl,swell.
- multidrug resistance protein
- adenosine 5′-triphosphate-binding cassette proteins
- cystic fibrosis transmembrane conductance regulator
- sulfonylurea receptor
- chloride channels
the atp-binding cassette (ABC) superfamily of membrane proteins includes proteins such as the cystic fibrosis transmembrane conductance regulator (CFTR), the sulfonylurea receptor (SUR), and the multidrug resistance protein P-glycoprotein (MDR1). Most members of the ABC superfamily expressed in mammalian cells have two membrane-spanning regions, each followed by a nucleotide binding fold (27). A linker region that is a phosphorylation target in CFTR [regulatory (R) domain; see Ref. 16] and MDR1 (mini-linker domain; see Ref. 19) joins the two membrane spanning region-nucleotide binding domain tandems. The CFTR R domain is required for proper CFTR regulation by protein kinases (11, 16).
MDR1 is a plasma membrane ATPase that extrudes seemingly unrelated hydrophobic compounds from the cell interior, thus conferring resistance to a large variety of drugs (19). In addition to its pump function, MDR1 may also serve as a membrane transport regulator or modifier. A number of studies suggest that MDR1 expression alters membrane transport processes, as evidenced by changes in intracellular pH (4, 38, 43), cell membrane depolarization (25, 47), and an increase in Na+ channel activity (49).
Some members of the ABC family have been proposed to regulate the activity of ion channels and other membrane proteins (e.g., CFTR, MDR1) or form part of ion channels as subunits (e.g., SUR). CFTR is an ion channel (5, 6) that has also been implicated in the regulation of both outwardly rectifying Cl−channels and epithelial Na+channels (39, 41). Activation of protein kinase A (PKA) has been shown to activate outwardly rectifying Cl− channels and inhibit epithelial Na+ channels only in the presence of CFTR (14, 41). SUR seems to be a subunit of ATP-sensitive K+ channels [KATP; formed by SUR and inwardly rectifying K+ channels (Kir)]. SUR confers sulfonylurea sensitivity to Kirand is involved in the response to nucleotides (1). Several studies have shown a relationship between swelling-activated Cl− currents (I Cl,swell) and MDR1 expression (2, 32, 35, 45, 46). Most importantly, protein kinase C (PKC)-mediated phosphorylation of MDR1 was shown to either inhibitI Cl,swell (24) or reduce its activation rate (7). Although recent studies contradict these observations (35), Hardy et al. (24) showed that mutation of all consensus PKC phosphorylation sites in the MDR1 mini-linker domain abolished the effect of phosphorylation. However, only Ser-661, Ser-667, and Ser-671 of the eight Ser and Thr in the MDR1 mini-linker domain are phosphorylated by PKC (9, 19). PKA phosphorylates two of the PKC sites, Ser-667 and Ser-671, and also Ser-683, a site not phosphorylated by PKC (9, 19). There is no information on a possible role of regulation ofI Cl,swell by PKA-mediated phosphorylation of MDR1. With respect to a possible regulation of drug transport, the current evidence does not support a major role of MDR1 phosphorylation (18, 20). However, an increase in the apparent affinity to some drugs (i.e., verapamil, vinblastine, and rhodamine 123) of the drug-stimulated ATPase activity by MDR1 phosphorylation has been suggested (42).
The aims of this study were to determine whether1) activation of PKC altersI Cl,swell in MDR1-expressing cells, 2) stimulation of PKA affectsI Cl,swell in MDR1-expressing cells, and 3) Ser-661, Ser-667, and/or Ser-671 of MDR1 are required for the effects of PKA and PKC stimulation onI Cl,swell.
MATERIALS AND METHODS
Mouse fibroblast cells (BALB/c-3T3; a gift from Dr. E. B. Mechetner) were grown in DMEM supplemented with 10% fetal bovine serum (GIBCO, Rockville, MD) and 1% (vol/vol) streptomycin-penicillin (1 unit penicillin-1 μg streptomycin; GIBCO) at 37°C in 5% CO2. The cells transfected with MDR1 cDNA (BALB-MDR1) and MDR1-3SA cDNA (BALB-MDR1-3SA) were grown in the continuous presence of the antibiotic geneticin (G418; GIBCO) at a concentration of 600 μg/ml.
Generation of MDR1- and MDR1-3SA-expressing clones.
BALB/c-3T3 cells were transfected with either wild-type MDR1 or mutant MDR1 (MDR1-3SA) cDNA in the vector pLK444M. This vector contains a β-actin promoter and was derived from pLK444 (a gift of Dr. P. Melera; see Ref. 12). Site-directed mutagenesis was used to substitute Ser-661, Ser-667, and Ser-671 with Ala, and both MDR1 and MDR1-3SA contain six histidine residues at the COOH-terminal end. This results in a modified MDR1 that has drug-stimulated ATPase activity and transports drugs (Ref. 31; see Fig. 1). Detailed information on the cDNA engineering and development of the cell lines is described elsewhere (A. Castro, J. Horton, and G. Altenberg, unpublished observations). Functional expression of MDR1 in the plasma membrane was determined by measuring the efflux of the MDR1 substrate rhodamine 123, as previously described (3, 46).
BALB/c-3T3, BALB-MDR1, and BALB-MDR1-3SA were detached the day of the experiment by exposure to a Ca2+-free PBS (in mM: 137 NaCl, 8 sodium phosphate, 1.5 potassium phosphate, and 2.7 KCl, pH ∼7.4) containing 0.5 mM EDTA. The cells were allowed to recover for ∼30 min at room temperature (22–23°C) in HEPES-buffered solution (HBS; in mM: 135 NaCl, 5 KCl, 1 MgCl2, 2 CaCl2, 7.8 glucose, and 5 HEPES-NaOH, pH 7.4, osmolality ∼280 mosmol/kg). The cells were detached because when grown at low confluence they spread out and their thinness makes them difficult to patch. Detachment of MDR1-expressing BALB/c-3T3 cells did not affect MDR1 functional expression in the plasma membrane, i.e., rhodamine 123 efflux was similar in attached and detached cells (data not shown). The isolated cells were placed in a plastic chamber (volume ∼700 μl) and allowed to settle to the bottom and attach (∼20 min). The chamber was mounted on the stage of an inverted microscope (Nikon Diaphot, Nikon, Tokyo, Japan).
Whole cell recordings were performed at room temperature while the cells were superfused at 2–3 ml/min with the experimental solutions in an ∼500-μl-volume chamber. The pipette and bath solutions [isosmotic (ISO) and 22% hyposmotic (HYPO)] were designed to have Cl− as the main conductive ion. Their chemical compositions (in mM) were: ISO, 140N-methyl-d-glucammonium chloride (NMDG-Cl), 1.3 CaCl2, 0.5 MgCl2, 10 HEPES, and 7.8 glucose, pH ∼7.4, ∼275 mosmol/kg; HYPO, same as ISO, except that NMDG-Cl was reduced to 105 mM (∼215 mosmol/kg); pipette solution, 140 NMDG-Cl, 1.2 MgCl2, 10 HEPES, 2 ATP, and 1 EDTA, pH ∼7.0, ∼270 mosmol/kg. To prevent cell swelling and activation ofI Cl,swell in ISO, the pipette solution was diluted 5–10% with distilled water. The glass pipettes were pulled with a multistage P-87 Flaming-Brown micropipette puller (Sutter Instruments, San Rafael, CA) and fire polished. Pipette tip resistances were 3–5 MΩ (nearly symmetrical NMDG-Cl solutions in the pipette and bath).
After obtaining a 5-GΩ seal between the pipette and the cell membrane, the patch was ruptured by applying negative pressure or negative pressure plus a large voltage pulse (1.2 V for 0.5–3 ms). Currents were measured in the whole cell configuration by the patch-clamp technique (22) using an Axopatch 200A amplifier (Axon Instruments, Foster City, CA), sampled at 20 kHz, and filtered at 2 or 5 kHz. The holding voltage was 0 mV. The Cl− currents were measured from −80 to +80 mV at 20-mV steps, 40 ms after the start of the voltage pulse, as indicated. Pulse generation and data collection and analyses were performed using pCLAMP 6 (Axon Instruments). The measured currents are expressed relative to apparent membrane capacitance. The access resistance and apparent membrane capacitance were estimated as described by Lindau and Neher (30). Access resistance values ranged from 7 to 10 MΩ, and the apparent membrane capacitances were 31 ± 1 pF for BALB/c-3T3 cells (n = 35), 30 ± 1 pF for BALB-MDR1 cells (n = 58), and 29 ± 2 pF for BALB-MDR1-3SA cells (n = 42).
Cell volume measurements.
Detached BALB-MDR1 cells (see above) were placed on glass coverslips mounted in a Leiden microincubator (Medical Systems, Greenvale, NY). The cells were loaded for ∼1 h with 15 μM 5-chloromethylfluorescein diacetate (CMFDA; Molecular Probes, Eugene, OR; a fluorescent dye not transported by MDR1; see results) at room temperature in HBS. After CMFDA loading, the cells were placed on the stage of an inverted microscope (Nikon Diaphot) coupled to a confocal laser scanning video system (Odyssey, Noran Instruments, Middleton, WI). Superfusion with CMFDA-free ISO solution (see above for composition) was initiated, and cell fluorescence was measured (excitation at 495 nm, emission at >535 nm). After recording of cell fluorescence for 6–8 min under basal conditions, the superfusate was changed to the experimental solution, i.e., ISO plus 200 nM phorbol 12-myristate 13-acetate (PMA). Changes in cell fluorescence due to changes in cell volume were calibrated by exposing control cells to either 11% NMDG-Cl HYPO solution (NMDG-Cl was reduced by 17.5 mM; ∼245 mosmol/kg) or 11% NMDG-Cl HYPER (sucrose was added to NMDG-Cl ISO to obtain ∼305 mosmol/kg). An increase in cell fluorescence denotes an increase in CMFDA concentration due to cell shrinkage, and a fall in cell fluorescence denotes cell swelling.
Data are presented as means ± SE. Statistical differences were calculated using Student’s t-test for paired or unpaired data, as appropriate, and were considered significant at P < 0.05 (two-tailed analysis).
BALB/c-3T3 mouse cells transfected with wild-type or mutant MDR1 cDNA express functional MDR1 in the plasma membrane.
We chose to carry out the experiments on MDR1-expressing cell lines that had never been exposed to chemotherapeutic agents to exclude the possibility that exposure to chemotherapeutic agents, and not MDR1 expression itself, could account for any of the unique properties of these cells (32, 33). Cell lines expressing either wild-type MDR1 or mutant MDR1, without selection with chemotherapeutic agents, were generated by transfecting BALB/c-3T3 cells with MDR1 cDNA (BALB-MDR1 cells) or MDR1-3SA cDNA (BALB-MDR1-3SA). Both BALB-MDR1 and BALB-MDR1-3SA cells displayed significant rhodamine 123 unidirectional efflux, whereas the rhodamine 123 efflux from the parental cells was ∼10-fold slower than that of the MDR1-expressing BALB/c-3T3 cells (Fig. 1). The rhodamine 123 efflux denotes MDR1-mediated transport, as previously shown (3). Both wild-type and mutant MDR1-expressing cell lines have equivalent levels of expression of MDR1 and MDR1-3SA in the plasma membrane, and MDR1, but not MDR1-3SA, is phosphorylated in vitro by PKC (A.F. Castro, J.K. Horton, C.G. Vanoye, and G.A. Altenberg, unpublished observations). In addition, we previously showed that MDR1 expression is undetectable in the parental cells (8, 23). Figure 1 shows that BALB-MDR1 and BALB-MDR1-3SA cells express functional MDR1 at the plasma membrane, whereas BALB/c-3T3 does not.
Expression of MDR1 or MDR1-3SA does not alter the magnitude, activation rate, or current-voltage (I-V) relationship of ICl,swellin BALB/c-3T3 cells.
Representative records of Cl− currents under basal conditions (ISO) and after activation by cell swelling (HYPO) are shown in Fig. 2. Whole cell Cl− currents were measured 2 min after breaking of the seal in ISO solution and then 6 min after exposure to HYPO solution. Figures 2 and 3show that the three cell lines exhibited sizable swelling-activated Cl− currents with similar outward rectification. The ratio of the absolute currents measured at 40 ms, at +80 and −80 mV, used to assess the degree of rectification, was not significantly different among all cell lines and was on average 1.25 ± 0.02 (n = 35). The three cell lines showed reversibility ofI Cl,swell when the bath solution was changed back to ISO (data not shown). Figure4 shows the swelling-activated currents measured, as a function of time, in parental and in MDR1- and MDR1-3SA-expressing cells. As seen in Fig. 4, expression of either wild-type or mutant MDR1 did not alter the activation rate ofI Cl,swell in BALB/c-3T3 cells. Figures 2-4 show that expression of MDR1 or MDR1-3SA does not alter the magnitude, activation rate, orI-Vrelationship ofI Cl,swell in BALB/c-3T3 cells. Moreover,I Cl,swell under control conditions (absence of PKA and PKC stimulation) is similar in MDR1- and MDR1-3SA-expressing cells.
Inhibition of ICl,swellby activation of PKC and PKA.
Protein kinase stimulation was carried out by exposure to either 200 nM PMA (a membrane-permeant PKC activator) or 1 mM 8-bromoadenosine 3′,5′-cyclic monophosphate (8-BrcAMP; a membrane-permeant PKA activator). These concentrations are sufficient to attain maximal stimulation (26).
Figure 5 Ashows the swelling-activated Cl− currents measured at +80 and −80 mV in BALB/c-3T3, BALB-MDR1, and BALB-MDR1-3SA cells under control conditions and after PKC activation. The cells were exposed to ISO plus PMA for 8–10 min before the whole cell configuration was obtained. About 2 min after the breaking of the seal, the bath solution was changed to HYPO plus PMA for ∼6 min. Activation of PKC by exposure to PMA reducedI Cl,swell only in the MDR1-expressing BALB-MDR1 cells, i.e., PMA did not affectI Cl,swell in either BALB/c-3T3 or BALB-MDR1-3SA cells. These results indicate that MDR1 expression is required for the inhibitory effect of PKC onI Cl,swell. In addition, one or more of Ser-661, Ser-667, and Ser-671 are required for this inhibition. Because Ser-667 and Ser-671 are also substrates for PKA, we tested whether PKA activation reproduces the PKC effect. Figure5 B showsI Cl,swellmeasured at +80 and −80 mV in BALB/c-3T3, BALB-MDR1, and BALB-MDR1-3SA cells under control conditions and after PKA activation, by a protocol similar to that used to study the effect of PKC stimulation but with 8-BrcAMP instead of PMA. The results show that activation of PKA inhibitsI Cl,swell in BALB/c-3T3 cells expressing MDR1 but not in the parental or mutant MDR1-expressing cells. Hence, inhibition ofI Cl,swell in BALB/c-3T3 cells by either PKA or PKC activation requires MDR1 expression, and Ser-661, Ser-667, and/or Ser-671 are required for this effect. In addition, because PKA and PKC have Ser-667 and Ser-671 as common substrates, one or both residues are essential for the inhibition ofI Cl,swell by MDR1 phosphorylation.
The effects of both PKA and PKC stimulation were always present but were somewhat variable in magnitude when experiments performed months apart were compared (e.g., compare the decreases in the rate ofI Cl,swellactivation by PMA at 6 min in Figs. 5 and6). For this reason, experimental and control studies were always contemporaneous.
PKA and PKC affect ICl,swell by different mechanisms.
Recently, phosphorylation of MDR1 by PKC was shown to reduce the activation rate ofI Cl,swellfollowing exposure to hypotonic solutions (7). However, other investigators have not found this effect (35). After examination of the time course ofI Cl,swellactivation, we noticed that inhibition of the current by PKC activation was stronger at early times after reduction of the bath osmolality. This observation suggested that phosphorylation of MDR1 could alter the activation rate ofI Cl,swell but not its final magnitude. To test this possibility, we exposed BALB-MDR1 cells treated with either PMA or 8-BrcAMP to HYPO solution for a longer period. We stopped acquiring data 12–14 min after switching the bath to HYPO because membrane blebs appeared in the patched cells after that time, suggesting uncoupling of the membrane from the cytoskeleton. Figure 6 A shows that phosphorylation of MDR1 by PKC slows down the activation ofI Cl,swell without altering its steady-state value. In contrast, activation of PKA reduced the steady-state level ofI Cl,swell in BALB-MDR1 cells without modifying the activation rate (Fig.6 B). As shown in Fig.6 C, the addition of both PMA and 8-BrcAMP generated a “mixed” effect, i.e., both activation rate and final magnitude ofI Cl,swell were reduced. In summary, PKA and PKC stimulation produced distinct effects on I Cl,swell, and these effects were abolished by replacing Ser-661, Ser-667, and Ser-671 with Ala residues.
The delay of ICl,swellactivation in BALB-MDR1 by PKC stimulation is not caused by cell shrinkage.
One possible explanation for the delay in the activation ofI Cl,swell in MDR1-expressing cells treated with PMA is that MDR1 phosphorylation causes a decrease in cell volume upon stimulation of PKC. This putative reduction in cell volume could explain the delay in the onset ofI Cl,swell, i.e., exposure to HYPO during the first ∼2 min would return the cell volume to the level before PMA exposure, without activation ofI Cl,swell. To test this possibility, changes in cell volume were monitored before and during exposure to 200 nM PMA using the fluorescent dye CMFDA as a marker for cell water volume. Pilot experiments showed that the decrease in fluorescence during the time course of the studies was due to photobleaching of the fluorophore (9 ± 1%,n = 30), with negligible probe efflux in both BALB/c-3T3 and BALB-MDR1 cells (data not shown). These observations indicate that the CMFDA-glutathione conjugate is not an MDR1 substrate. The conjugate could be a substrate for the multidrug resistance associated protein (MRP1), but BALB/c-3T3 cells are devoid of MRP1 (8). Fluorescence of the CMFDA-glutathione conjugate was monitored in cells exposed to either ISO or ISO plus PMA for 10 min. Fluorescence, normalized to the average value in ISO after correction for photobleaching, was 1.00 ± 0.01 in ISO (n = 22) and 0.99 ± 0.02 in ISO plus PMA (n = 8). These results show that PMA did not elevate cell fluorescence (as expected for cell shrinkage), thus ruling out a decrease in cell volume as the mechanism for the delay inI Cl,swellactivation. The validity of the fluorescent probe-based methods to assess changes in cell volume has been previously established (44) and was confirmed by measuring CMFDA fluorescence in control cells exposed to anisotonic solutions. Exposure to 11% HYPO reduced cell fluorescence by 12 ± 3% (n = 15), and exposure to 11% HYPER increased cell fluorescence by 11 ± 5% (n = 6).
Inhibition of ICl,swellin BALB-MDR1 cells by PKA activation is not due to ATP release.
It has been suggested that ABC proteins can regulate other transporters by providing a pathway for efflux of ATP, which can activate or inhibit other transporters via activation of purinergic receptors (reviewed in Ref. 11). In this context, phosphorylation of CFTR by PKA activates outwardly rectifying Cl−channels via ATP efflux (39). Because external ATP reduces the magnitude ofI Cl,swell in several cell lines (40), including BALB-MDR1 cells (Fig.7 A), it is then possible that phosphorylation of MDR1 by PKA causes ATP release and this extracellular ATP subsequently reduces the magnitude ofI Cl,swell. To test this possibility, BALB-MDR1 cells were exposed to 8-BrcAMP in the continuous extracellular presence of hexokinase (0.5 U/ml; Sigma, St. Louis, MO), glucose, and Mg2+. Under these conditions, hexokinase catalyzes the phosphorylation of glucose by ATP, thus scavenging any secreted ATP. In CFTR-expressing cells, this experimental maneuver has been shown to prevent the activation of outwardly rectifying Cl− channels via purinergic receptors (39). Figure 7 B showsI Cl,swellmeasured at +80 mV in BALB-MDR1 cells both under control conditions and in the presence of 8-BrcAMP or 8-BrcAMP plus hexokinase. Addition of external hexokinase did not block the PKA-induced effect. This proves that external ATP is not involved in the inhibition ofI Cl,swell by PKA in BALB-MDR1 cells. Addition of external hexokinase alone did not affect I Cl,swellin BALB-MDR1 cells.I Cl,swellmeasured at +80 mV, 6 min after exposure to HYPO bath, was 83 ± 25 pA/pF (n = 5) in control conditions and 122 ± 10 pA/pF (n = 5) in the presence of hexokinase.
Current evidence indicates that expression of MDR1 regulates or modifies swelling-activated Cl− channels (7, 24, 35, 45,46). The main results supporting such a role for MDR1 are1) the modulation of the volume sensitivity and/or the activation rate of swelling-activated Cl− channels by MDR1 expression (7, 32, 35, 45) and 2) the inhibition ofI Cl,swell or a decrease in its activation rate caused by PKC activation in MDR1 (and mdr1a)-expressing cells (7, 24). The PKC effect was abolished by mutating all the consensus PKC phosphorylation sites in the mini-linker domain of MDR1 to Ala (24), indicating that phosphorylation of MDR1 is part of the mechanisms by which PKC activation affectsI Cl,swell. The present results provide additional evidence that phosphorylation of MDR1, not MDR1 expression per se, alters the activity of swelling-activated Cl−channels.
It has been shown by several investigators that expression of MDR1 alters the activation rate ofI Cl,swell by either increasing it (7, 32, 35, 45) or decreasing it (33) and that this effect is reversed by PKC activation (7, 45). However, MDR1 expression does not always alter the activation rate ofI Cl,swell (Ref.15; see Fig. 4). It is possible that the rate at whichI Cl,swell is activated in MDR1-expressing cells, vis à vis the parental, non-MDR1-expressing cells, is dependent on experimental factors that include the way cells are prepared for the experimental procedure (23) and exposure of MDR1-expressing cells to chemotherapeutic agents (32,33).
In our hands, activation of PKC by PMA affects the activation rate ofI Cl,swell in an MDR1-expressing cell line, not its final magnitude, thus confirming observations by Bond et al. (7). We also demonstrated that phosphorylation of residues Ser-661, Ser-667, and Ser-671 in the MDR1 mini-linker domain is sufficient for the effect of PKC stimulation. Moreover, because our cell lines were never exposed to chemotherapeutic drugs, our results rule out a nonspecific effect of PKC related to cell exposure to those agents (33). However, the effects of MDR1 phosphorylation on swelling-activated Cl− channels are controversial because other investigators did not find any effects of PKC stimulation onI Cl,swell in MDR1-expressing cells (35). Among the possible explanations for this discrepancy are the following. 1) The data shown by Miwa et al. (35) were obtained at steady state, when no effect is expected (see Fig. 4).2) The MDR1-dependent effects of PKC stimulation on swelling-activated Cl− channels may well vary from cell to cell, depending on the presence and location of specific PKC isoenzymes as well as on the specific molecule underlyingI Cl,swell. In this context, although ClC-3 has been tentatively identified as a swelling-activated Cl−channel (13), it seems thatI Cl,swell is underlain by different channels in different cells (46), and it is possible that MDR1 may influence only some of these channels. The observations on the association of Kir channels with SUR support this hypothesis. SUR1 has been shown to modify the function of some (e.g., Kir 6.1, Kir 6.2) but not all (e.g., Kir 2.1, Kir 4.3) Kir channels (1).
The present results demonstrate for the first time that stimulation of PKA affects swelling-activated Cl− channels selectively in MDR1-expressing cells. Exposure to a membrane-permeable analog of cAMP reduced the magnitude ofI Cl,swellelicited by hyposmotic swelling. This effect was observed only in the cells expressing MDR1 and was abolished by substitution of Ser-661, Ser-667, and Ser-671 of MDR1 with Ala. The results in the parental cells devoid of MDR1 are in agreement with the notion that the PKA pathway does not regulate swelling-activated Cl− channels (36). In contrast, it has been shown that activation of myocardialI Cl,swell is blocked by PKA-mediated phosphorylation (21). It has been proposed (36) that such a mechanism could form part of a negative feedback ofI Cl,swellactivation because some cells accumulate cAMP during swelling (48). The possibility that MDR1 expressed in the myocardium (17) is responsible for the modulation ofI Cl,swell remains to be explored.
Interestingly, our results show that the MDR1-related effects of PKA and PKC onI Cl,swell are different. PKA stimulation reduces steady-stateI Cl,swell, whereas PKC stimulation reduces the rate of activation of the currents by swelling, without affecting their steady-state level. Moreover, the effects of PKA and PKC stimulation are different and additive (reduction of speed of response plus decrease in steady-state currents). The latter result indicates that the two effects are independent. It has been shown that MDR1 is phosphorylated by PKC at Ser-661, Ser-667, and Ser-671 (probably at the same residues in vitro and in vivo; Refs. 9, 19), whereas PKA phosphorylation occurs at Ser-667, Ser-671, and Ser-683 (only in vitro data are available; see Ref. 19). Because the effects of stimulation of PKA and PKC are different and additive and require phosphorylation of Ser-661, Ser-667, and/or Ser-671, it is likely that phosphorylation of Ser-661 is involved in the response to PKC stimulation and that Ser-683 is involved in the response to PKA stimulation. However, additional mutagenesis studies are required to confirm this hypothesis.
Other ABC proteins, besides MDR1, regulate ion channels. CFTR is a Cl− channel that regulates outwardly rectifying Cl−channels, epithelial Na+ channels, and KATP channels (14, 34, 41), whereas SUR regulates certain Kirchannels (1). There is little information on the mechanism by which ABC proteins alter the activity of ion channels. ATP secretion has been implicated in the regulation of outwardly rectifying Cl− channels by CFTR (39), but ATP secretion dependent on CFTR is controversial (29, 37), and it cannot explain our results. External ATP blocksI Cl,swell, and ATP scavenging by hexokinase had no effect on theI Cl,swellreduction by PKA stimulation. Direct interaction between CFTR and the α-subunit of the epithelial Na+channel has been suggested to mediate the inhibition of Na+ channels by phosphorylation of CFTR by PKA (28). The relationship between SUR and Kir also supports protein-protein interaction as the mechanism for regulation by ABC proteins. It has been established that the effects of sulfonylureas (inhibition) and Mg-ADP (stimulation) on Kirchannels are mediated by SUR. SUR probably mediates the effects of sulfonylureas and Mg-ADP on Kirchannels via protein-protein interaction, because SUR and Kir channels form functional oligomers (KATP channels), likely consisting of four Kir and four SUR molecules (1). The precise mechanism of alteration in swelling-activated Cl−channels by phosphorylation of MDR1 is unknown. MDR1-associated transport of substrates with stimulatory or inhibitory effects on Cl− channels is unlikely because the magnitude ofI Cl,swell is independent of MDR1 expression (see Ref. 46). Efflux of ATP can also be ruled out from our studies. On the basis of the available information on the effects of CFTR and SUR on ion channels, we favor the hypothesis that the regulation of swelling-activated Cl− channels mediated by MDR1 phosphorylation is via protein-protein interaction.
Our conclusions are as follows. 1) MDR1 phosphorylation alters the function of swelling-activated Cl− channels. This modulation could depend on the cell type, e.g., differences in expression and compartmentalization of PKC isoenzymes and different Cl− channels underlyingI Cl,swell.2) The effects of PKA and PKC on swelling-activated Cl−channels by phosphorylation of MDR1 require phosphorylation of one or more of Ser-661, Ser-667, and Ser-671. These residues are located in the MDR1 mini-linker domain (10), homologous to the R domain of CFTR (the R domain is a target for modulation of CFTR activity by phosphorylation; see Ref. 11). 3) The effects of PKA and PKC are different and additive. Stimulation of PKC reduces the rate of the increase inI Cl,swellfollowing exposure to hyposmotic solution, whereas stimulation of PKA reduces the magnitude ofI Cl,swell, without affecting the speed of the response.4) The simplest explanation for our results is that phosphorylation of Ser-661 is involved in the response to PKC stimulation, whereas Ser-683 is involved in the response to PKA stimulation. This conclusion is based on the observations that Ser-667 and Ser-671 are phosphorylated by both PKA and PKC, that PKA also phosphorylates Ser-683, and that PKC also phosphorylates Ser-661.
The recent observation that both SUR and CFTR increase the sensitivity of KATP to sulfonylureas suggests that several members of the ABC superfamily can associate with the same channels (e.g., KATP channels). Therefore, understanding how MDR1 phosphorylation affects swelling-activated Cl−channels may provide useful information on the mechanism(s) by which ABC proteins modulate ion channels.
We thank Drs. S. A. Weinman and J.-T. Zhang for comments on a preliminary version of this paper and K. Spilker for technical assistance. BALB/c-3T3 cells were generously provided by Dr. E. Mechetner. The vector pLK444 was a gift of Dr. P. Melera.
Address for reprint requests: G. A. Altenberg, Dept. of Physiology and Biophysics, University of Texas Medical Branch, 301 University Blvd., Galveston, TX 77555-0641.
This work was supported by a grant from Searle Research and Development, a grant-in-aid from the American Heart Association (Texas Affiliate), and National Institutes of Health Grants CA-72783 and DK-08865.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
- Copyright © 1999 the American Physiological Society