Guillermo A. Altenberg, Carlos G. Vanoye, Ernest S. Han, Joachim W. Deitmer, and Luis ..... Acknowledgments-We are grateful to Drs. J. A. Belli, K. H. Cowan,.
Vol. 269, No. 10, Issue of March 11, pp. 7145-7149, 1994 Printed in U.S.A.
THEJOURNAL of BIOLCXICAL CHEMISTRS Q 1994 by The American Society for Biochemistry and Molecular Biology, Inc.
Relationships between Rhodamine 123 Transport, Cell Volume, and Ion-channel Functionof P-glycoprotein* (Received for publication, July 12, 1993, and in revised form, December 8, 1993)
Guillermo A. Altenberg, Carlos G. Vanoye, Ernest S. Han, Joachim W. Deitmer, and Luis ReussS From the Department of Physiology and Biophysics, University of Texas Medical Branch, Galueston, Texas 77555-0641
The P-glycoprotein (Pgp),a plasma membrane protein overexpressed in multidrug-resistant tumor cells, is thought to be both an ATPase that actively exports cytotoxic drugs and aC1- channel activated by cell swelling. The partial reversal of multidrug resistance by C1transport blockers suggests a possible role for C1- in Pgp-mediated drug transport. We used multidrug-resistant Chinese hamster fibroblasts and human breast cancer cells expressing Pgp to study the roles of C1- (and also Na+ and HCO&02) on Pgp-mediated efflux of the fluorescent dye rhodamine 123 (R123). In Pgp-expressing Chinese hamster fibroblasts, exposed to isosmotic solutions, the unidirectionalefflux of R123 was not measurably changed by a -60-minremoval of C1- (or by exposure to Na+-free, or nominallyHCO;/C02-freemedium); short term (2-3min) ion substitutions were also ineffective. In human breast cancer cells transfected with human mdrl cDNA, hyposmotic solutions activated a C1- current but had no effect on the Pgp-mediated unidirectional efflux of R123. Additionally, in human breast cancer cells, the intracellular presence of R123 did not prevent activation of the C1- current by hyposmotic solution. The lack of detectable effect of removal of C1-, Na+, or HC05 on Pgp-mediated R123 transport rules out direct coupling between substrate transport and transportof either of these ions by Pgp. The persistence of Pgp-mediated R123 efflux in osmotically swollen cells indicates that activation of the Pgp-associated C1current does not hinder the Pgp pump function. The lack of effect of R123 on swelling-activated C1- current denotes that Pgp-mediated transport of organic substrates andPgp-associated C1- currents canoccur at the same time in a single cell. Theseresults underscore the dissociation between Pgp-mediated active drug transport and electrodiffusive C1- transport.
Multidrug resistance caused by overexpression of P-glycoprotein (Pgp)’ is generally attributed to enhanced drug efflux that results in reduced intracellular steady-state levels of cytotoxic drugs (1, 2). It is thought that Pgp is botha drug efflux pump and a C1- channel activated by cell swelling (3-7). Furthermore, it has been suggested that the drug-transport and C1- channel Pgp functions can be dissociated, inasmuchas cells
* This work was supported in part by a grant from the John Sealy Memorial Endowment Fund and by Grant RR07205 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Towhom correspondence and requests for reprints should be addressed: Dept. of Physiology and Biophysics,University of Texas Medical Branch, Galveston, TX 77555-0641. Tel.: 409-772-5444; Fax: 409772-3381. The abbreviations used are: Pgp, P-glycoprotein; R123, rhodamine 123; V P , verapamil; NMDG, N-methyl-D-glucammonium.
transfected with a mutant form of Pgp t h a t does not confer C1- currents activatedby cell multidrug resistance still display swelling ( 5 ) .These results suggest that Pgp-dependent C1- currents can occur in the absence of drug transport, but they do not establish whether or notC1- has a role in drug transport. The observation of partial reversal of multidrug resistance by C1transport inhibitors (8)suggests the possibility of a role forC1-, either direct or indirect. For instance, direct couplingcould be by active C1--drug cotransport, and an indirect role of C1- could be by allosteric modification of Pgp, resulting in the appropriC1- transport ate conformation for drug transport. However, blockers are notspecific (9),and thus theireffect on multidrug resistance could be unrelated to C1- transport. In addition, some C1- channels are insensitive to C1- channel blockers (10, 111, although they may have similar properties (such as conductance, gating, and voltage dependence) to those of blockersensitive channels. The present studies were designed to establish: ( a ) whether Pgp-mediated extrusionof the fluorescent substrate rhodamine obligatory coupling to C1-, Na+, or 123 (R123) involves HCO;/CO2 transport, ( b ) whether C1- current activation by hyposmotic swelling, presumed to put the Pgp in a “channel configuration” ( 5 , 12) preventsor alters drug transport,and (c) whether the presence of Pgp substrate alters the activationof the C1- current by cell swelling. EXPERIMENTALPROCEDURES General Procedures-Experiments were performed on subconfluent monolayers of wild-type (V79)and highly multidrug-resistant Chinese hamster lung fibroblasts (LZ-8)(13-19, and two human breast cancer cell lines (16, 171, drug-sensitive (MCF-7) and multidrug-resistant, transfected with human mdrl cDNA(BC19/3).Culture conditions were as previously described (17). Western blots using the monoclonal antibody C219 have shownthat Pgp is present in both Chinese hamster cell lines, but the levels are =13-fold greater in LZ-8 compared to V79 cells (15). Levels of gene amplification and expression are also greater in LZ-8 cells ( = E - and =39-fold for DNA and mRNA, respectively) (15). MCF-7 cells do not express Pgp, as evaluated by Western blot and mdrl RNA analysis (18, 191, immunofluorescence (201, and lack of effects of verapamil ( I T ) on R123 emux and steady-state accumulation (17) (see also “Results”). Similar studies in BC19/3 cellshave shown expressionof functional Pgp (16, 17, 20). Correlations between levels of Pgp expression and either drug transport or putative channel function of Pgp have been shownin several cell types, by measurements of mRNAlevels, plasma membrane Pgp expression, drug transport, and whole-cell currents (7, 15, 18).2 However, these studies do not prove cause-effectrelationships. In most experiments, we used a HCO;/CO,-buffered solution with the followingcomposition (in mM):115NaC1,25 NaHC03, 5 KCl, 1 MgCl,, 2 CaCl,, 1.5 sodium phosphate, and 7.8 glucose, equilibrated with 95% 02,5% COz,pH7.42-7.43. The solutions bufferedwith HEPES contained (in mM): 135NaC1, 5 KCl, 1 MgCl,, 2 CaCl,,7.8 glucose, 5 HEPES, equilibrated with air, titrated with NaOH toa pH of 7.42-7.43. Ion replacements were isomolar; Na’ was replaced with N methyl-D-glucammonium (NMDG) or tetramethylammonium, and C1G. A. Altenberg, C.G. Vanoye,J. K. Horton, and L. Reuss, submitted for publication.
7 145
A
B
[CI--freelj
lvplIvpJ
0.00
(min)
Is Ct-independent
P-glycoprotein-mediated Dartsport Drug
7146
time (min)
time
FIG.1. Effect of transient C1- removal on the decay of R123 fluorescence (FRLz9). A, unidirectional efflux was started in control solution (126~ Cl-1; l l during ~ the time indicatedby the top bar, all C1in the superfusate was replaced mole for mole with cyclamate. The line represents the non-linear least squaresoffit a single exponential tothe data. E , unidirectional efflux was started under control conditions; verapamil (10PM) was added during the periods indicatedby the top bars. with cyclamate or sulfate (sucrose was added t o maintain the osmolality constant, at -280 mosmkg). The basic pipette-filling solution had the following composition (in m):140 NMDG-C1, 1.2 MgCl,, 10 HEPES, 1 EGTA, 2 ATP, and 0.5 GTP. The isosmotic NMDG-Cl bath solution contained (inm):140 NMDG-Cl, 1.3 CaCl,, 0.5 MgCl,, 10 HEPES, and 7.8 glucose. In the hyposmotic solutions, the [NaCl] was reducedby 40 mM and the [NMDG-Cl] by 35 m, respectively. The pH of these solutions was 7.4Ck7.43. Substitution ofC1- by gluconate was isomolar. Details on the chambers and perfusion system have been published previously (17, 21). R123 andVP were purchased from Sigma. Fluorescence Microscopy-The decay of intracellular rhodamine 123 fluorescence ( F R 1 2 3 ) was measuredby quantitative fluorescence microscopy, as previously described (17)., Rate constants for FRlZ3 decay (k) were estimated from fits of Equation 1 to the data
TABLEI Effects of changes in solution ioniccomposition on the rate constant for R123 efflux Values were determined cells in superfused with HCO;/ C0,-buffered solution, HEPES-buffered, HEPES-buffered Na+-free, or HEPES-buffered C1--free solutions. Cells were loaded with R123 for 60-90 min in each experimental solution. Rate constants are in min". Data are means-c S.E. of 5-30 experiments. Results with different ion substitutes (see "Experimental Procedures") did not differ and were pooled. Experiment
v79
LZ-8
HCO;/CO, HEPES Na' removal C1- removal
0.219 -c 0.012 0.214 f 0.041 0.198 2 0.041 0.228 2 0.108
2.298 f 0.456 2.696 -c 0.290 2.613 2 0.495 2.296 f 0.242
LZ-8
EC 1 9/3
ISOSMOTIC
HYPOSMOTIC
a L
_]lo0
pA
10 rns
ISOSMOTIC
FIG.2. Effects of exposure to hyposmotic solution on C1- currents in multidrugresistant Chinese hamster lung fibroblasts (LZ-8) andmultidrugresistant humanbreastcancer cells (BC19/3).The main salt in pipette and bath solutions was NMDG-C1. Pulse protocol was as follows: holding voltage was 0 mV, and voltages
ranged from -80 to 80 mV in 20-mV steps. Records are averagesof 10 sweeps. Upward deflections denote current flow from cell to bath (outward current), which correspondst o C1- flux from bath to cell interior. where Fois the background fluorescence, andFRIZ3 mBx is FRIz3 a t t = 0. Currents were measuredat theend of the pulse (longer pulses revealed Electrophysiology-Whole-cell currents were measured in the bro- that a steady state had been reached a t 40 ms). Current recordsLZ-8 in ken-patch configuration as described by Hamill et al. (22), using a List and BC19/3 cells are shown in isosmotic solution (top records),at 6 min EPC-7 amplifier (List Medical, Darmadst, Germany). In most experiof exposure to hyposmotic solution (middle records), and at 10 min of so that C1- was the exposure to isosmotic solution (bottom records). The current was actiments, the bath and pipette solutions were chosen only conductive ion (5). The pipettes were pulled with P-87 a Flaming- vated by exposure to hyposmotic solutiononly in BC19/3 cells. Brown micropipette puller (Sutter Instruments Co., San Rafael, CA) and fire-polished. Pipette resistances ranged from 3 to7megohms placement with HEPES). In all instances there was no meas(NMDG-Cl solutions in pipette and bath). The pulse protocol is given in urable change in R123 efflux. the legend to Fig. 2. Pulse generation and datacollection and analysis Table I summarizes the effects of long term (60-90 min) were camed out with a computer, using commercial software (PClamp, removal of CI-, Na+, or HCO; on R123 efflux in V79 and LZ-8 Axon Instruments, Inc., Foster City, CA). cells, expressed as first-order rate constants calculated from Statistics-Datashown are means -c S.E. Statistical comparisons were done by Student's t tests for paired or unpaired data as appropri- non-linear least squaresfits of single exponentials to the data. The rate constants are about 10-fold greater in LZ-8 cells. In p was less than0.05. ate. Differences were considered significant when Fix123 = F
a + F R I Z S m&
-Ii
(Eq. 1)
both cell lines, the ion substitutions had no significant effects on R123 efflux. Multidrug-resistant Chinese Hamster Fibroblasts Do Not ExRemoval of Na+, GI-, or HCO, lC0, Have No Detectable press C1- CurrentsActivated by Cell Swelling-Fig. 2 illusEffect on the Unidirectional Efflux of Rhodamine 123-The in the whole-cell current wasmeasdecay in FR123 upon removal of the probe from the external trates an experiment which compartment denotes a unidirectional efflux under zero-trans ured in an LZ-8 cell before, during, and after exposure to a conditions.' Hereafter, we refer t o the F~123decay as R123 hyposmotic (approximately -22%, lowering [NMDG-ClI) soluefflux. Short term ion-replacement experiments were carried tion, with isosmotic NMDG-C1 in the pipette throughout. The out with V79 cells because the R123 efflux is slower than that cell displayed a C1- current (as demonstrated by anion substiwhich was not increased by exposure in LZ-8 cells, allowing for paired comparisons of data obtained tutions; data not shown), during brief changes in external solution. As seen in Fig. l A , to hyposmotic solution. Results from several such experiments C1- removal (cyclamate replacement) did not alter R123 efflux are summarized inFig. 3A. In rare occasions, exposure of LZ-8 from V79 cells. As previously shown (171, the F ~ p - 3decay fol- or V79 cells to hyposmotic solution elicited a reversible C1lows a single-exponential function, as shown by the curve su- current activation. Because of this variability, we turned to perimposed on the raw data. Fig. 1B demonstrates that W human breast cancer cells for studies involving activation of reversibly inhibits the R123 efflux, as expected if this flux is C1- currents. Human Breast Cancer Cells Bansfected with Human mdrl mediated by Pgp. In similar experiments we removed Na' (replacement with cDNAExhibit Swelling-activated GI- Currents-Fig. 2 illusNMDG or tetramethylammonium) or HCO;/COz (isohydric re- trates theeffect of a transient exposure to hyposmotic medium RESULTS
P-glycoprotein-mediated Drug ll-ansport Is Ct-independent
7147
B
A
400
1
-1
1
’
50
100
J -400
-400
D
C 400
-
I/T fi
HYPO
200 -1 00 1
/
-50 1
r
50
f-
IS0
’ v 100
-100
-50
(mv)
-200
- -400
-
-400
FIG.3. Average whole-cell I-V relationships in multidrug-resistantChinese hamster lung fibroblasts (LZ-8) andin human breast cancer cells, both sensitive (MCF-7)and multidrug-resistant(BC19/3).Data shown are means2 S.E. of the currents (I) at 40 ms of the onset of the voltage pulse (V). In panels A, B , and C , open symbols denote isosmotic NMDG-CI in the bath ( B O ) and filled symbols denote 6 min of exposure to hyposmotic NMDG-Cl in the bath(HYPO). A, LZ-8 cells (n = 5). There was no change in current levels. B , MCF-7 cells (n = 6). There was no change in current levels. C, BC19/3 cells (n = 11). There was significant activation of the current; the activation was reversible (see Fig. 2). D , selectivity of the currentactivated by hyposmotic solution in BC19/3 cells (n = 4); after a 6-min exposure to hyposmotic NMDG-Cl solution (HYPO,open symbols), NMDG-Cl was replaced with NMDG-gluconate (hyposmotic) for 4 min (filled symbols). The abolishment of the outward current indicates that the current is camed byC1-. The reversal potential, assuming no cation permeation, yields a permeability ratio (ClH gluconate) of 49.
on whole-cell C1- currents in multidrug-resistant(BC19/3) human breast cancer cells. In contrast with the result in LZ-8 cells, the hyposmotic solution increased reversibly the wholecell current.Data from severalexperimentsin MCF-7 and BC19/3 cells are summarized inFig. 3 (B and C),respectively. In isosmotic NMDG-Cl, whole-cell currents of similar magnitudes are observed in both cell lines. Upon exposure to hyposmotic medium, the whole-cell currents increased only in the multidrug-resistant cells BC19/3. Fig. 30 shows that replacement of bath C1- with gluconate abolishesthe outward current. This indicates that the swelling-activated current is carriedby C1-, i.e. it is not a non-selective cation current. Similar experimentsindicatethatthecurrentsmeasuredin isosmotic NMDG-Cl in MCF-7 and BC19/3 cells are also carried by C1(data not shown). Interestingly, the swelling-activated C1- current in mdrl-transfected fibroblasts and lungcarcinoma cells is outward-rectifying (5, 7). In contrast, in BC19/3 cells we observed outward-rectifying, inward-rectifying,or linear currentvoltage (I-V) relationships. These results indicate thatat least two kinds of channels are activated by swelling in BC19/3 cells (the linear I-V relationship could represent the sum). At this time it is notknown which one, if any, of these channels is the Pgp itself.
We also assessed whole-cell currents using a HEPES-buffered mixtureof KC1and potassium gluconatein the pipette and HEPES-buffered PSS in the bath, to try to approach more physiologic K’ and C1- gradients across the cell membrane ([Cl-I = 35 mM, pH 6.7-6.8; Refs. 17 and 23). In these experiments, the cell-swelling-activated C1- current was 119 T 41 pA at -80 mV, similar to that found using NMDG-Cl in both pipette and bath. Swelling of Pgp-expressing Human BreastCancer Cells Does Not Alter the UnidirectionalEflux of Rhodamine 123-To test whether activationof the putative channelfunction of Pgp has an effect on its pump function, we compared the unidirectional R123 efflux in BC19/3 cells in isosmotic and hyposmotic solutions. Panel A in Fig. 4 depicts a n experiment in which transient exposure to hyposmotic solution (-22%) had no measurable effect on R123 efflux. To test whether the Pgp can be “locked” in configurations that exclusively express one function, i.e. either channel or pump, we carried out two sets of experiments, also illustrated in Fig. 4 (panels B and C).For the experimentsshown in panel B , we first exposed BC19/3 cells to hyposmotic solution for 5 min, loaded them withR123, and then measured the efflux, all in hyposmotic solution (Fig. 4 B , HYPO). After completing this
7 148
\
P-glycoprotein-mediated Drug Dansport Is CP-independent
FIG.4. Effects of exposure to hyposmotic solutionon R123 efflux and effects of R123 on swellingactivated whole-cell currents inBC19/3cells. A, R123 efflux was started in isosmotic medium;during the time indicated by the top bar, the osmolality was reduced by 22%. The l i n e superimposedon the data is a single-exponential fit. B , unidirectional R123 efflux in BC1913 cells exposed to hyposmotic solution (HYPO) or isosmotic solution (ZSO). Verapamil ( t o p bars) was added at a concentration of-100 1.1~.The records are representative of three experiments. C , rate constant (k)for R123 efflux in BC19/3 cells (n = 4) and MCF-7 cells (n = 5), exposed to isosmotic solution ( I S O ) or hyposmotic solution (HYPO). - and + signs under the BC19/3 cell bars indicate absence and presence of 100 p~ VF', respectively.D , whole-cell I-V relationships in BC19/3 cells ( n = 4).Protocol was as described for Fig. 3C,but 100 1.1~R123 was added to the pipette solution. Open symbols, isosmotic NMDG-Cl in the bath (ISO);filled symbols, 6 min of exposure to hyposmotic NMDG-Clthein bath (HYPO). There is significant activation of the current (cf Fig. 3C).
1 .oo
n
0.75
?
0
v
n 0.50
B 0.75
-
n 0.50
-
FF
? o
v
N
c
a
LL
G
0.25
0.00 ~.
1
:
I
:
I
.
2
0
4
6
-
0.4
-
0.2
-
0.0
-
0
time (rnin)
0.6
0.25
0.00 -
I
~
n
-
-
N
-
1.00
10
30
40
50
time (rnin)
D 300
1
T
.-c E
v y
procedure, the cells were returned to isosmotic solution for 10-30 min, and loaded again with R123 to measure theefflux in isosmotic solution (Fig.4B, I S 0 ). There wasno effect of cell swelling on FRIz3decay. VP inhibited R123 efflux in hyposmotic and isosmotic media. A statistical summary of these experiments is presented in Fig. 4C. The rate constants for R123 efflux in BC19/3 cells exposed to isosmotic and hyposmotic media did not differ and were virtually abolished by VP in both isosmotic and hyposmotic media, denoting that theR123 efflux is mediated by Pgp in both media.As also shown in Fig. 4C, the rate constantsfor R123 efflux from MCF-7 cells are very small, consistent with the undetectable degree of Pgp expression in these cells (16,18, 19).We have previously shown that VP has no effect on R123 efflux in MCF-7 cells (17). In summary, hyposmotic swelling has no effects on R123 efflux, although under these conditions C1- currents are activated (Figs. 2 and 3). During the period of exposure to hyposmotic solution, BC19/3 cells remain swollen (20). Presence of Rhodamine 123 in the Pipette Solution Does Not Prevent C1- Current Activation byHyposmoticSwelling-In fibroblasts transfected with human mdrl cDNA, exposure to Pgp substrates was reported to prevent C1- current activation by swelling (5).In Fig. 40 we show the results of experiments in which cells were voltage-clamped (whole-cell configuration) with R123 added to the pipette solution, and then exposed to hyposmotic solution. The presence of R123, which as demonstrated in Fig. 1 is a Pgp substrate, did not prevent the activation of the C1- current by exposure to thehyposmotic solution (compare Figs. 3C and 40). DISCUSSION
Our first aim was totest whether the pumpfunction of Pgp involves flux coupling between organicsubstrate andinorganic ions. ATP-dependent, Pgp-mediated drug transport has been previously demonstrated in membranevesicles in thenominal
absence of Na+, C1-, and HCO;/C02 (4, 23, 241, but these results do not rule out flux coupling. For instance, it is possible that the ATP-dependent drug transport measured in membrane vesicles is only a small fraction of that in intact cells because drug transportis coupled to inorganic-ion transport or requires specific modulators. Prolonged removal of C1-, Na+, and HCO;/C02 had no measurable effects on R123 efflux. Hence, we rule out flux coupling between HCO;/C02, Na+, or C1-, and R123 transport inPgp-expressing multidrug-resistant cells. Furthermore, transient ion removal, which changes the transmembrane concentration gradient and, insome cases, the membrane voltage, the intracellularpH, and other parameters, was also ineffective. In other words, R123 efflux is neither directly (via cotransport or exchange) nor indirectly (e.g. via cell membrane voltage) influenced C1-, by Na+, or HCO;/C02. In V79 and LZ-8 cells, removal of HCO;/C02 rapidly elevates intracellular pH (171, while C1- removal elevates intracellular pH relatively slowly and to a lesser degree, and Na+ removal produces intracellular a~idification.~ absence The of detectable changes in R123 efflux upon any of these maneuvers is consistentwith our previous conclusion that intracellular pHchanges are notnecessary for multidrug resistance and do not influence R123 efflux (17). It hasbeen recently suggestedthat thePgp has a dual function, namely drug pump and C1- channel (5,7). In addition, it has been reported that C1- channel blockers reverse multidrug resistance (8), suggesting association between C1- and drug transport via Pgp. The present experiments indicate thatC1removal has no effects on R123 efflux in Chinese hamster multidrug-resistant cells. However, it is possible that the pump and channel functions of Pgp are related when the currents are activated by cell swelling. Hence, we carried out experiments in G. A. Altenberg, C. G. Vanoye, E. S. Han, G. Young, and L. Reuss, unpublished observations.
P-glycoprotein-mediated Drug Dansport Is Ct-independent
7 149
D. Glass and F. Bavarian for human breast cancer cells expressing Pgp and in which swell- preliminary version of the manuscript, and ing does activate a C1- current. We addressed thequestion of a technical assistance. relationship between the two putative Pgp functions in two REFERENCES ways. First, we investigated whether channelactivation alters 1. Beck, W. T. (1992) in Molecular and Cellular Biology of Multidrug Resistance pump function, and second, we tested whether organic-subin %nor Cells (Roninson, I. B.,ed) pp. 215-227, Plenum Press, New York strate transport prevents the increaseC1-incurrent elicited by 2. Roninson, I. B., Pastan, I., and Gottesman, M.M. (1992) in Molecular and cell swelling. In response to the first question, activation of C1Cellular Biology of Multidrug Resistance in %nor Cells (Roninson, I. B., ed) pp. 91-106, Plenum Press, New York currents by exposure to hyposmotic solutions did not alterR123 3. Ambudkar, S. V., Lelong, I. H., Zhang, J., Cardarelli, C. 0..Gottesman, M. M., efflux. Hence, we suggest that reversal of multidrug resistance and Pastan, I. (1992) Proc. Natl. Acad. Sci. U. S. A. 89,847243476 4. Doige, C. A,, Yu, X., and Sharom, F. J. (1992) Biochim. Biophys. Acta 1109, by C1- channel blockers is not mediated by interference with 149-160 plasma membrane C1- transport. For instance, bicuculline, a 5. Gill, D. R., Hyde, S. C., Higgins, C. F., Valverde, M. A,, Mintenig, G . M., and Sephlveda, F. V. (1992) Cell 71, 23-32 y-aminobutyric acid receptor antagonist that blocks C1- con6. Sarkadi, B., Price, E. M., Boucher, R. C., Germann, U. A,, and Scarborough, G . ductance, may reverse multidrug resistance by a direct interA. (1992)J. Biol. Chem. 267,48544858 action with Pgp (8). In response to the second question, the 7. Valverde, M. A., Diaz, M., Seprilveda, F. V., Gill, D. R., Hyde, S. C., and Higgins, C. F. (1992) Nature 366, 830-833 intracellular presence of R123 did not preventactivation of C18. Shalinsky, D. R., Jamshidipour, R., Las, G . , Christen, R. D., and Howell, S. B. currents by cell swelling. Under the same experimentalcondi(1993) Proc. Am. Assoc. Cancer Res. 34,315 9. Pon, D. J., Flezar, M., Lister, D. L., and Heisler, S. (1993) Eul: J. Pharmacol. tions, some Pgp substrates do inhibit swelling-activated cur246, 119-127 rents, while others have no e f f e ~ tTaken .~ together, these two 10. Copello, J., Heming, T. A,, Segal,Y., and Reuss, L. (1993)J. Gen. Physiol. 102, sets of results argue against the notion that the Pgp can be 177-199 “locked”in eithera drug-transporting or an ion-conductive con- 11. Anderson, M. P., Sheppard, D. N., Berger, H. A,, and Welsh, M. J. (1992) Am. J. Physiol. 263, Ll-LI4 figuration. 12. Higgins, C. F. (1993) C u m B i d . 6,684-687 In summary, our results demonstrate that there is no flux 13. Howell, N., Belli, T. A., Zaczkiewicz, L. T.,and Belli, J. A. (1984) Cancer Res. 44,4023-4029 coupling between C1-, Na+, andHCO, on one hand and R123 14. Roninson, I. B., Abelson, H. T., Housman, D. E., Howell, N., and Varshavsky, on the other, that there are no indirect effects of removal of A. (1984) Nature 309, 626-628 these ions on Pgp-mediated substrate transport, that activa- 15. Sognier, M. A,, Zhang, Y., Eberle, R. L., and Belli, J. A. (1992) Biochem. Pharmacol. 44, 1859-1868 tion of the Pgp-associated C1- current by cell swelling does not 16. Fairchild, C. R., Moscow, J. A., OBrien, E. E., and Cowan, K. H. (1990) Mol. Pharmacol. 37,801409 impair substrate transport by Pgp, and that substrate transG . A., Young, G., Horton, J. K., Glass, D., Belli, J. A,, and buss, L. port does not prevent C1- current activation. Hence, these re- 17. Altenberg, (1993)Proc. Natl. Acad. Sci. V. S. A. 90,9735-9738 sults do not supportthe proposal that Pgp molecules can 18. Nakagawa, M., Schneider, E., Dixon, K. H., Horton, J., Kelley, K., Morrow, C., and Cowan, K. H. (1992) Cancer Res. 52,6175-6181 “switch” between drug-transporting and ion-conductive states K., Vaughan, M. M., Slocum, H. K , Fredericks, W. J., Chen, Y.-f., Ar(5, 12). Rather, our observations appear to have two possible 19. Tbth, redondo, M. A,, Harstrick,A,, Karakousis, C., Baker, R. M., and Rustum., Y. explanations. Either the putative channel and pump functions M. (1992)Am. J . Pathol. 140, 1009-1016 are virtuallyexclusive and only a few Pgp molecules are chan- 20. Altenberg, G .A., Deitmer, J. W., Glass, D. C., and Reuss, L. (1994) Cancer Res., in press nels, or Pgp in the channel configuration can also transport 21. Altenberg, G . , Copello, J., Cotton, C., Dawson, K., Segal, Y., Wehner, F., and Reuss, L. (1990)Methods Enzymol. 192, 650-683 organic substrates. Acknowledgments-We are grateful to Drs. J. A. Belli, K. H. Cowan, and C. R. Fairchild for providing the cell lines used in these studies. We thank Drs. J. K. Horton, J. Navarro, andS. A. Weinman for discussions, Drs. E. Bello-Reuss, J. M. Russell, and J.-T. Zhang for comments on a
22. Hamill, 0.P., Marty, A., Neher, E., Sakmann, B., and Sigworth, F. J. (1981) pfliigers Arch. 391,85-100 23. Kamimoto, Y., Gatmaitan, Z., Hus, J., and Arias, I. M. (1990) J. Biol Chem. 284, 11693-11698 24. Horio, M., Gottesman, M. M., and Pastan, I. (1988) Proc. Natl. Acad. Sei. U. S. A. 85,358C3584
